Library homepage

  • school Campus Bookshelves
  • menu_book Bookshelves
  • perm_media Learning Objects
  • login Login
  • how_to_reg Request Instructor Account
  • hub Instructor Commons
  • Download Page (PDF)
  • Download Full Book (PDF)
  • Periodic Table
  • Physics Constants
  • Scientific Calculator
  • Reference & Cite
  • Tools expand_more
  • Readability

selected template will load here

This action is not available.

Statistics LibreTexts

9.1: Null and Alternative Hypotheses

  • Last updated
  • Save as PDF
  • Page ID 23459

The actual test begins by considering two hypotheses . They are called the null hypothesis and the alternative hypothesis . These hypotheses contain opposing viewpoints.

\(H_0\): The null hypothesis: It is a statement of no difference between the variables—they are not related. This can often be considered the status quo and as a result if you cannot accept the null it requires some action.

\(H_a\): The alternative hypothesis: It is a claim about the population that is contradictory to \(H_0\) and what we conclude when we reject \(H_0\). This is usually what the researcher is trying to prove.

Since the null and alternative hypotheses are contradictory, you must examine evidence to decide if you have enough evidence to reject the null hypothesis or not. The evidence is in the form of sample data.

After you have determined which hypothesis the sample supports, you make a decision. There are two options for a decision. They are "reject \(H_0\)" if the sample information favors the alternative hypothesis or "do not reject \(H_0\)" or "decline to reject \(H_0\)" if the sample information is insufficient to reject the null hypothesis.

\(H_{0}\) always has a symbol with an equal in it. \(H_{a}\) never has a symbol with an equal in it. The choice of symbol depends on the wording of the hypothesis test. However, be aware that many researchers (including one of the co-authors in research work) use = in the null hypothesis, even with > or < as the symbol in the alternative hypothesis. This practice is acceptable because we only make the decision to reject or not reject the null hypothesis.

Example \(\PageIndex{1}\)

  • \(H_{0}\): No more than 30% of the registered voters in Santa Clara County voted in the primary election. \(p \leq 30\)
  • \(H_{a}\): More than 30% of the registered voters in Santa Clara County voted in the primary election. \(p > 30\)

Exercise \(\PageIndex{1}\)

A medical trial is conducted to test whether or not a new medicine reduces cholesterol by 25%. State the null and alternative hypotheses.

  • \(H_{0}\): The drug reduces cholesterol by 25%. \(p = 0.25\)
  • \(H_{a}\): The drug does not reduce cholesterol by 25%. \(p \neq 0.25\)

Example \(\PageIndex{2}\)

We want to test whether the mean GPA of students in American colleges is different from 2.0 (out of 4.0). The null and alternative hypotheses are:

  • \(H_{0}: \mu = 2.0\)
  • \(H_{a}: \mu \neq 2.0\)

Exercise \(\PageIndex{2}\)

We want to test whether the mean height of eighth graders is 66 inches. State the null and alternative hypotheses. Fill in the correct symbol \((=, \neq, \geq, <, \leq, >)\) for the null and alternative hypotheses.

  • \(H_{0}: \mu \_ 66\)
  • \(H_{a}: \mu \_ 66\)
  • \(H_{0}: \mu = 66\)
  • \(H_{a}: \mu \neq 66\)

Example \(\PageIndex{3}\)

We want to test if college students take less than five years to graduate from college, on the average. The null and alternative hypotheses are:

  • \(H_{0}: \mu \geq 5\)
  • \(H_{a}: \mu < 5\)

Exercise \(\PageIndex{3}\)

We want to test if it takes fewer than 45 minutes to teach a lesson plan. State the null and alternative hypotheses. Fill in the correct symbol ( =, ≠, ≥, <, ≤, >) for the null and alternative hypotheses.

  • \(H_{0}: \mu \_ 45\)
  • \(H_{a}: \mu \_ 45\)
  • \(H_{0}: \mu \geq 45\)
  • \(H_{a}: \mu < 45\)

Example \(\PageIndex{4}\)

In an issue of U. S. News and World Report , an article on school standards stated that about half of all students in France, Germany, and Israel take advanced placement exams and a third pass. The same article stated that 6.6% of U.S. students take advanced placement exams and 4.4% pass. Test if the percentage of U.S. students who take advanced placement exams is more than 6.6%. State the null and alternative hypotheses.

  • \(H_{0}: p \leq 0.066\)
  • \(H_{a}: p > 0.066\)

Exercise \(\PageIndex{4}\)

On a state driver’s test, about 40% pass the test on the first try. We want to test if more than 40% pass on the first try. Fill in the correct symbol (\(=, \neq, \geq, <, \leq, >\)) for the null and alternative hypotheses.

  • \(H_{0}: p \_ 0.40\)
  • \(H_{a}: p \_ 0.40\)
  • \(H_{0}: p = 0.40\)
  • \(H_{a}: p > 0.40\)

COLLABORATIVE EXERCISE

Bring to class a newspaper, some news magazines, and some Internet articles . In groups, find articles from which your group can write null and alternative hypotheses. Discuss your hypotheses with the rest of the class.

In a hypothesis test , sample data is evaluated in order to arrive at a decision about some type of claim. If certain conditions about the sample are satisfied, then the claim can be evaluated for a population. In a hypothesis test, we:

  • Evaluate the null hypothesis , typically denoted with \(H_{0}\). The null is not rejected unless the hypothesis test shows otherwise. The null statement must always contain some form of equality \((=, \leq \text{or} \geq)\)
  • Always write the alternative hypothesis , typically denoted with \(H_{a}\) or \(H_{1}\), using less than, greater than, or not equals symbols, i.e., \((\neq, >, \text{or} <)\).
  • If we reject the null hypothesis, then we can assume there is enough evidence to support the alternative hypothesis.
  • Never state that a claim is proven true or false. Keep in mind the underlying fact that hypothesis testing is based on probability laws; therefore, we can talk only in terms of non-absolute certainties.

Formula Review

\(H_{0}\) and \(H_{a}\) are contradictory.

  • If \(\alpha \leq p\)-value, then do not reject \(H_{0}\).
  • If\(\alpha > p\)-value, then reject \(H_{0}\).

\(\alpha\) is preconceived. Its value is set before the hypothesis test starts. The \(p\)-value is calculated from the data.References

Data from the National Institute of Mental Health. Available online at http://www.nimh.nih.gov/publicat/depression.cfm .

Have a language expert improve your writing

Run a free plagiarism check in 10 minutes, generate accurate citations for free.

  • Knowledge Base
  • Null and Alternative Hypotheses | Definitions & Examples

Null & Alternative Hypotheses | Definitions, Templates & Examples

Published on May 6, 2022 by Shaun Turney . Revised on June 22, 2023.

The null and alternative hypotheses are two competing claims that researchers weigh evidence for and against using a statistical test :

  • Null hypothesis ( H 0 ): There’s no effect in the population .
  • Alternative hypothesis ( H a or H 1 ) : There’s an effect in the population.

Table of contents

Answering your research question with hypotheses, what is a null hypothesis, what is an alternative hypothesis, similarities and differences between null and alternative hypotheses, how to write null and alternative hypotheses, other interesting articles, frequently asked questions.

The null and alternative hypotheses offer competing answers to your research question . When the research question asks “Does the independent variable affect the dependent variable?”:

  • The null hypothesis ( H 0 ) answers “No, there’s no effect in the population.”
  • The alternative hypothesis ( H a ) answers “Yes, there is an effect in the population.”

The null and alternative are always claims about the population. That’s because the goal of hypothesis testing is to make inferences about a population based on a sample . Often, we infer whether there’s an effect in the population by looking at differences between groups or relationships between variables in the sample. It’s critical for your research to write strong hypotheses .

You can use a statistical test to decide whether the evidence favors the null or alternative hypothesis. Each type of statistical test comes with a specific way of phrasing the null and alternative hypothesis. However, the hypotheses can also be phrased in a general way that applies to any test.

Here's why students love Scribbr's proofreading services

Discover proofreading & editing

The null hypothesis is the claim that there’s no effect in the population.

If the sample provides enough evidence against the claim that there’s no effect in the population ( p ≤ α), then we can reject the null hypothesis . Otherwise, we fail to reject the null hypothesis.

Although “fail to reject” may sound awkward, it’s the only wording that statisticians accept . Be careful not to say you “prove” or “accept” the null hypothesis.

Null hypotheses often include phrases such as “no effect,” “no difference,” or “no relationship.” When written in mathematical terms, they always include an equality (usually =, but sometimes ≥ or ≤).

You can never know with complete certainty whether there is an effect in the population. Some percentage of the time, your inference about the population will be incorrect. When you incorrectly reject the null hypothesis, it’s called a type I error . When you incorrectly fail to reject it, it’s a type II error.

Examples of null hypotheses

The table below gives examples of research questions and null hypotheses. There’s always more than one way to answer a research question, but these null hypotheses can help you get started.

*Note that some researchers prefer to always write the null hypothesis in terms of “no effect” and “=”. It would be fine to say that daily meditation has no effect on the incidence of depression and p 1 = p 2 .

The alternative hypothesis ( H a ) is the other answer to your research question . It claims that there’s an effect in the population.

Often, your alternative hypothesis is the same as your research hypothesis. In other words, it’s the claim that you expect or hope will be true.

The alternative hypothesis is the complement to the null hypothesis. Null and alternative hypotheses are exhaustive, meaning that together they cover every possible outcome. They are also mutually exclusive, meaning that only one can be true at a time.

Alternative hypotheses often include phrases such as “an effect,” “a difference,” or “a relationship.” When alternative hypotheses are written in mathematical terms, they always include an inequality (usually ≠, but sometimes < or >). As with null hypotheses, there are many acceptable ways to phrase an alternative hypothesis.

Examples of alternative hypotheses

The table below gives examples of research questions and alternative hypotheses to help you get started with formulating your own.

Null and alternative hypotheses are similar in some ways:

  • They’re both answers to the research question.
  • They both make claims about the population.
  • They’re both evaluated by statistical tests.

However, there are important differences between the two types of hypotheses, summarized in the following table.

To help you write your hypotheses, you can use the template sentences below. If you know which statistical test you’re going to use, you can use the test-specific template sentences. Otherwise, you can use the general template sentences.

General template sentences

The only thing you need to know to use these general template sentences are your dependent and independent variables. To write your research question, null hypothesis, and alternative hypothesis, fill in the following sentences with your variables:

Does independent variable affect dependent variable ?

  • Null hypothesis ( H 0 ): Independent variable does not affect dependent variable.
  • Alternative hypothesis ( H a ): Independent variable affects dependent variable.

Test-specific template sentences

Once you know the statistical test you’ll be using, you can write your hypotheses in a more precise and mathematical way specific to the test you chose. The table below provides template sentences for common statistical tests.

Note: The template sentences above assume that you’re performing one-tailed tests . One-tailed tests are appropriate for most studies.

If you want to know more about statistics , methodology , or research bias , make sure to check out some of our other articles with explanations and examples.

  • Normal distribution
  • Descriptive statistics
  • Measures of central tendency
  • Correlation coefficient

Methodology

  • Cluster sampling
  • Stratified sampling
  • Types of interviews
  • Cohort study
  • Thematic analysis

Research bias

  • Implicit bias
  • Cognitive bias
  • Survivorship bias
  • Availability heuristic
  • Nonresponse bias
  • Regression to the mean

Hypothesis testing is a formal procedure for investigating our ideas about the world using statistics. It is used by scientists to test specific predictions, called hypotheses , by calculating how likely it is that a pattern or relationship between variables could have arisen by chance.

Null and alternative hypotheses are used in statistical hypothesis testing . The null hypothesis of a test always predicts no effect or no relationship between variables, while the alternative hypothesis states your research prediction of an effect or relationship.

The null hypothesis is often abbreviated as H 0 . When the null hypothesis is written using mathematical symbols, it always includes an equality symbol (usually =, but sometimes ≥ or ≤).

The alternative hypothesis is often abbreviated as H a or H 1 . When the alternative hypothesis is written using mathematical symbols, it always includes an inequality symbol (usually ≠, but sometimes < or >).

A research hypothesis is your proposed answer to your research question. The research hypothesis usually includes an explanation (“ x affects y because …”).

A statistical hypothesis, on the other hand, is a mathematical statement about a population parameter. Statistical hypotheses always come in pairs: the null and alternative hypotheses . In a well-designed study , the statistical hypotheses correspond logically to the research hypothesis.

Cite this Scribbr article

If you want to cite this source, you can copy and paste the citation or click the “Cite this Scribbr article” button to automatically add the citation to our free Citation Generator.

Turney, S. (2023, June 22). Null & Alternative Hypotheses | Definitions, Templates & Examples. Scribbr. Retrieved March 12, 2024, from https://www.scribbr.com/statistics/null-and-alternative-hypotheses/

Is this article helpful?

Shaun Turney

Shaun Turney

Other students also liked, inferential statistics | an easy introduction & examples, hypothesis testing | a step-by-step guide with easy examples, type i & type ii errors | differences, examples, visualizations, what is your plagiarism score.

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • View all journals
  • My Account Login
  • Explore content
  • About the journal
  • Publish with us
  • Sign up for alerts
  • Open access
  • Published: 11 December 2018

Recent origin and evolution of obesity-income correlation across the United States

  • R. Alexander Bentley 1 , 2 ,
  • Paul Ormerod 3 &
  • Damian J. Ruck 1 , 2  

Palgrave Communications volume  4 , Article number:  146 ( 2018 ) Cite this article

42k Accesses

21 Citations

276 Altmetric

Metrics details

  • Archaeology
  • Social policy

From a gene-culture evolutionary perspective, the recent rise in obesity rates around the Developed world is unprecedented; perhaps the most rapid population-scale shift in human phenotype ever to occur. Focusing on the recent rise of obesity and diabetes in the United States, we consider the predictions of human behavioral ecology (HBE) versus the predictions of social learning (SL) of obesity through cultural traditions and/or peer–to–peer influence. To isolate differences that might discriminate these different models, we first explore temporal and geographic trends in the inverse correlation between household income and obesity and diabetes rates in the U.S. Whereas by 2015 these inverse correlations were strong, these correlations were non-existent as recently as 1990. The inverse correlations have evolved steadily over recent decades, and we present equations for their time evolution since 1990. We then explore evidence for a “social multiplier” effect at county scale over a ten-year period, as well as a social diffusion pattern at state scale over a 26–year period. We conclude that these patterns support HBE and SL as factors driving obesity, with HBE explaining ultimate causation. As a specific “ecological” driver for this human behavior, we speculate that refined sugar in processed foods may be a prime driver of increasing obesity and diabetes.

Similar content being viewed by others

null hypothesis about poverty

Twitter (X) use predicts substantial changes in well-being, polarization, sense of belonging, and outrage

Victoria Oldemburgo de Mello, Felix Cheung & Michael Inzlicht

null hypothesis about poverty

Globally representative evidence on the actual and perceived support for climate action

Peter Andre, Teodora Boneva, … Armin Falk

null hypothesis about poverty

Ideological differences in the expanse of the moral circle

Adam Waytz, Ravi Iyer, … Jesse Graham

Introduction

In the United States, where adult obesity prevalence rates have been rising since the 1970s (Kranjac and Wagmiller, 2016 ), about two-thirds of adults are now overweight and over 100,000 U.S. deaths per year are attributed to obesity (Ogden et al., 2014 ). With obesity rates having tripled in many U.S. states over the past 25 years, this rise in obesity prevalence has accelerated. In 1990, about 11% of a typical U.S. state population was obese and no state had more than 15% obesity in its adult population. By 2015, U.S. obesity rates had more than doubled, with several states above 35% adult obesity and no state below 20% obesity in the population (Centers for Disease Control and Prevention, 2017a ). In one generation, the change has been so dramatic that the obesity rate in any U.S. state in 2015 would have been an extreme outlier in the U.S. in 1990.

From a gene-culture evolutionary perspective, the recent rise in obesity rates, occurring across the Developed world (Goryakin et al., 2017 ), is unprecedented. In the past, human niche construction evolved over a time scale of centuries or millennia (Creanza and Feldman, 2016 ; Milot et al., 2011 ). For example, the evolution of lactase persistence among Neolithic populations of central Europe was rapid in evolutionary terms but nevertheless took place over thousands of years, in coevolution with the intensification of dairying economies (Brock et al., 2015 ; Gerbault et al., 2013 ). In contrast, industrially–processed foods have transformed Western human diets in less than a century. Not only has this made calories and junk food abundant and inexpensive in high-income countries, but there appear to be other effects such as reduction of gut microbiome diversity (Smits et al., 2017 ; Muscogiuri et al., 2018 ).

In the simplest view, obesity in Developed economies is a result of over-abundance of inexpensive food calories combined with decreases in daily physical activity in the industrialized world and its built environment (Mattson et al., 2014 ; Mullan et al., 2017 ). Negative energy balance is not the only factor, however, and with heterogeneity across socioeconomic groups, the specific causes for the rapid and recent increase in U.S. obesity remain unclear (Cook et al., 2017 ; Dwyer–Lindgren et al., 2013 ; Flegal et al., 2016 ).

One thing that is clear in high-income countries is that, despite decades of economic growth, obesity disproportionately affects the poor—the “poverty–obesity paradox” (Hruschka and Han, 2017 ). The proportion of obese individuals in industrialized nations now correlates inversely with median household income. This phenomenon is called the “reverse gradient” because it is the reverse of the pattern in developing countries, where higher income correlates with higher body mass. In the United States and other developed countries, lower income households tend to have higher rates of obesity (Hruschka, 2012 ; Subramanian et al., 2011 ). In 2015, over 35% of the population was obese in U.S. states where median household incomes were below $45,000 per year, whereas obesity was less than 25% of state populations where median incomes were above $65,000 (Centers for Disease Control and Prevention, 2017c ). Similarly in Europe today, poor individuals are 10% to 20% more likely to be obese (Salmasi and Celidon, 2017 ). This pattern is unique to Developed economies; within China, for example, an inverse correlation between income and obesity/diabetes is observed only in the most economically developed regions (Tafreschi, 2015 ).

Cultural evolution potentially offers a less proximate, more ultimate explanation for the recent rise in obesity. Evolutionary approaches to behavioral change include human behavioral ecology and cultural evolutionary theory; the former tends to prioritize optimality of adaptive behavior while the latter tends to prioritize social learning. Generally speaking, human behavioral ecology (HBE) emphasizes the plasticity of human physiology and behavior, by which individuals minimize risk to survival and optimize their long-term reproductive payoffs (Higginson et al., 2017 ). As wealth mitigates survival risk, HBE predicts a positive correlation between BMI and wealth, as humans have evolved to store calories as insurance against future famine or food shortage (Shrewsbury and Wardle, 2012 ; Higginson et al., 2017 ; Tapper, 2017 ). In the poorest 80% of the world’s societies, body mass index (BMI) generally increases with household wealth (Subramanian et al., 2011 )—except below about 400 USD per capita, when poverty is such that BMI is uniformly low (Hruschka et al., 2014 ). In high-income countries, HBE predicts greater obesity among the poor, partly because humans have evolved behavioral “rules” that lead to overeating in rich environments and partly because poorer people have more immediate risks and concerns than outweigh long-term mortality risk of being obese (Dittmann and Maner, 2017 ; Dohle and Hafmann, 2017 ; Higginson et al., 2017 ; Mani et al., 2013 ; Smith, 2017 ).

The HBE hypothesis predicts that obesity has recently evolved in strong correlation with both the food environment and with income/wealth. The “Insurance Hypothesis” (Nettle et al., 2017 ) uses HBE to explain why the reverse is true in Developed countries where extreme BMI (obesity) is more frequent among the poor. Under the Insurance Hypothesis (IH), “individuals should store more fat when they receive cues that access to food is uncertain” (Nettle et al., 2017 ). Poor people in high-income countries receive such cues, as they experience more stress and greater existential risk for multiple reasons. Prominent among these risks is malnutrition, yet empty calories are still inexpensively available as processed foods and sugar-sweetened beverages (Bray et al., 2004 ; Johnson et al., 2007 ; Jürgens et al., 2005 ; Bocarsly et al., 2010 ). The IH is consistent with observations of women in high-income countries, who are more likely to be obese when confronted by food insecurity (Nettle et al., 2017 ). An alternative explanation, however, occurs at the societal level in high-income countries, where heavier women tend to marry into poorer households due to through “anti-fat discrimination” in marriage (Hruschka, 2012 ; Hruschka and Han, 2017 ).

In contrast, social learning (SL) explanations emphasize learned behavior in groups: behaviors are inherited from parents and learned socially from contemporaries through the generations of family traditions or community cultures (Bentley et al., 2016 ; Colleran and Mace, 2015 ; Colleran, 2016 ). Dietary habits are often determined as much by cultural traditions as they are by nutritional needs and family economics (Anderson and Whitaker, 2010 ; Anderson, 2012 ; Hughes et al., 2010 ; Lhila, 2011 ; Mata et al., 2017 ; Redsell et al., 2010 ; Vizireanu and Hruschka, 2018 ). Cultural factors may therefore underlie local differences in obesity and diabetes rates, which exhibit effects of local neighborhood and its built environment (Alvarado, 2016 ; Carroll et al., 2016 ; Mullan et al., 2017 ; Kowaleski-Jones et al., 2017 ), family size (Datar, 2017 ), ethnic group and age group (Cook et al., 2017 ).

Under SL, obesity may also increase through social influence. A widely-discussed argument, first presented by Christakis and Fowler ( 2007 ), is that obesity “spreads” through social influence in networks of family and friends (Christakis and Fowler, 2013 ). Relatedly, recent modeling and experimental studies show how a minority group can initiate rapid change in social conventions, provided the minority reaches a ‘critical mass’ (Centola et al., 2018 ; Couzin et al., 2011 ). Under SL, therefore, a new behavior can become a new social norm relatively quickly, if obesity were indeed a new social norm.

The alternative to the social-learning explanation is homophily, if obesity clusters in social networks merely because those clusters are similar people in the same environments (Shalizi and Thomas, 2011 ). Homophily could be viewed either as similar behavior derived from shared cultural ancestry or else as similar behavior that reflects adaptations to similar environments. Outside of carefully monitored conditions (Centola et al., 2018 ; Hobaiter et al., 2014 ), however, it is difficult if not impossible to distinguish social influence from homophily, even if obesity is observed to cluster in social networks, without a fine–grained temporal dimension to the data (Christakis and Fowler, 2013 ; Shalizi and Thomas, 2011 ; Thomas, 2013 ).

Unlike the small-scale social network study of obesity versus specific friends and kin members (Christakis and Fowler, 2007 ), this study examines annual, population-scale obesity rates aggregated by U.S. county. If the aggregated data are time–stratified, however, we can still attempt to test the SL hypothesis. We will use multiple measures (obesity, leisure, income) and ten years of county-scale data to assess any “social multiplier” effects. The social multiplier effect is identified when the rate of behavior among a group is greater than what would be predicted based on individual–scale variables alone. Identification of groups is a problem in the empirical literature on social interactions (Blume et al., 2011 ), but useful proxies have based on Zip codes (Corcoran et al., 1992 ) and census tracts (Weinberg et al., 2004 ).

A study of crime rates (Glaeser et al., 2003 ), for example, used statistics of individuals to predict crime rates and regressed those on crime rates in groups. This is how the social multiplier was defined at the county level, specifically by comparing the coefficient b in the regression,

with the coefficient b' in its group counterpart

where ω ig denotes the choice of individual i in county g and x i is a vector of observable individual-specific characteristics; the social multiplier is defined as the coefficient ratio, b '/ b (Blume et al., 2011 ).

In this approach, estimating the social multiplier requires an estimate of individual-level rates, which do not exist in aggregated data. Faced with this problem, Glaeser et al., ( 2003 ) used nationwide arrest rates by age that, when combined with demographic data, provided a predicted level of crime in each neighborhood. These predicted rates were then regressed actual crime rates at the county level, yielding a coefficient of 1.7 at the county level, which was their estimate of the social multiplier at that scale of aggregation (Glaeser et al., 2003 ).

The data we use here are aggregated by U.S. county annually: over three thousand county-level time series of obesity, leisure and income rates over ten-year period (2004 to 2013). This amounts to ten sets of annual data, on several variables, for 3110 U.S. counties. If, controlling for the effects of household income, we find that lack of physical activity, or leisure rate, has a disproportionate effect on obesity rate in 2013 compared with 2004, then there may be support for the social multiplier effect. We have only aggregated statistics but we have the advantage of a time series. In principle there exists an individual-level, effectively physiological, connection between lack of exercise (leisure) and obesity rates, which we assume remains constant through time. The correlation between obesity and leisure rates should reflect this individual relationship as a baseline, plus any social multiplier effects over time.

In other words, change in the leisure–obesity correlation between 2004 and 2013 ought to reflect the social multiplier effect. As there are also unobservable connections between leisure and obesity, however, we follow the cautious approach of Glaeser et al., ( 2003 ), who “take these results warily, as they may well overstate the true social multiplier.”

To investigate whether obesity increased in the classic S –shaped pattern consistent with social learning, we carried out regression analysis on the annual data for each state over the 1990–2016 period. A simple linear increase in obesity rate since 1990 serves as our null hypothesis, with the alternative hypothesis being a non-linear time trend. If the null hypothesis of non-linearity could not be rejected, by implication an S–curve would not be present in the data.

We applied two separate but complementary approaches. First, we tested the null of linearity against a general non–linear alternative, using the methodology of local linear regression (Cleveland and Devlin, 1988 ). We used the “loess” command in R. Local linear regression fits simple linear models to localized subsets of the data to describe the deterministic part of the variation in the data, point by point, without specifying a global functional form. An input parameter in the loess command (“span”) allows the “equivalent number of parameters” (ENP) to be varied. ENP serves as a measure of the non–linearity of the series. The approach enables ANOVA tests to be carried out of the null of linearity against a range of non-linear alternatives. Local linear regression is a powerful approach, but does not yield a specific functional form.

For the second approach, we tested the null of linearity against a specific non–linear functional form, namely that of a classic adoption curve (Bass, 1969 ; Bentley and Ormerod, 2010 ; Henrich, 2001 ). We have:

where F t,i is the obesity rate in year t of U.S. state i , μ i is the chance in state i that a person becomes obese (i.e., BMI of 30 or above) through individual behavior and q i is the probability within state i , where F t,i are already obese, that a person becomes obese through social influence. This ODE can be solved for F t,i ,

where M is the magnitude of change and F 0,i is set as the obesity rate of U.S. state i in year 1990. This equation can be fitted to the obesity data, with the goodness of fit reported as the adjusted R 2 statistic, defined as:

where v is the total number of explanatory variables in the model (not including the constant term), and n is the sample size. The linear model has v  = 2 parameters (slope and intercept), whereas the Bass model has v  = 3 parameters ( p , q , and M ).

Here we focus on an under–studied, but revealing question: in the U.S., how has the correlation between household income and obesity changed in the past 25 years? In the U.S., obesity and diabetes rates currently have a strong negative correlation with household income (Hruschka, 2012 ). These correlations have been demonstrated cross-sectionally but not longitudinally, however, and therefore it is not possible to establish their causality (Boden and McLeod, 2017 ). In industrialized economies, the increase in obesity prevalence has been fastest among low income levels, as fast as tripling within a generation among certain subpopulations.

Annual, age-adjusted data on obesity rates at the county level, for years 2004 to 2015, were obtained from the publicly accessible archive maintained by the Center for Disease Control and Prevention ( www.cdc.gov/diabetes/data ). For these county-level data, we also make use of the CDC age-adjusted estimates, (Klein and Schoenborn, 2001 ), in which rates are age adjusted to the 2000 U.S. standard population using age groups, 20–44, 45–64, and 65 or older (Centers for Disease Control and Prevention, 2017b ). Older obesity data at the state level since 1990 were obtained from the annual reports of the Trust for America’s Health and the Robert Wood Johnson Foundation (stateofobesity.org). For analyzing the time-series of state-level obesity rates, we cautiously added data from the years 1991 and 1998 from a different source (Mokdad et al., 1999 ) to examine more closely any potential non-linear change in the 1990s.

Due to missing data at the county level, we excluded Alaska from all analyses at county level, while including Alaska for state-level analysis. We also used five years of CDC data on age-adjusted annual CDC diabetes rates, 2009–2013, in U.S. counties (Centers for Disease Control and Prevention, 2017d ). The estimates of diabetes rates are derived through telephone surveys, normalizes the data using population data from the US Census, and smooths the estimates such that three years of data are averaged in each annual estimate (Centers for Disease Control and Prevention, 2017c ).

Importantly, we use age–adjusted rates for both obesity and diabetes in our analysis, and so we do not include the age profile of an area as an explanatory factor. This adjustment has already been carried out by the CDC in the data which we use. By using age–adjusted data, we minimize the effect of demographics in our results. To anticipate, we also note for reference that our results were essentially the same when we used data that were not age–adjusted.

Estimates of leisure–time physical inactivity come from the CDC Behavioral Risk Factor Surveillance System, a system of health-related telephone surveys, which began in 1984 with 15 U.S. states, and now collects data in all 50 states through over 400,000 adult interviews each year (Centers for Disease Control and Prevention, 2017c ). The “leisure” statistic indicates the fraction of population who are designated as physically inactive, meaning they answered “no” to the question, “During the past month, other than your regular job, did you participate in any physical activities or exercises such as running, calisthenics, golf, gardening, or walking for exercise?”

Food desert data were recently made available through the Food Access Research Atlas (FARA) project (Rhone et al., 2017 ). The estimates are derived from the 2010 US Census and the 2010–2014 American Community Survey, in which census tracts are categorized by median income, vehicle availability, and Supplemental Nutrition Assistance Program (SNAP) participation (Rhone et al., 2017 ). To this geographic dataset are added two 2015 lists of supermarkets, supercenters, and large grocery stores to represent sources of affordable and nutritious food (Rhone et al., 2017 ). The FARA records for each U.S. Census tract the number and share of people more than a certain distance to a supermarket: in urban areas, that specified distance is half a mile or 1 mile, whereas in rural areas the distances are 10 miles or 20 miles (Rhone et al., 2017 ). Also recorded is whether the population in the census tract has overall low access to vehicles. Census tracts are also designated as rural, urbanized (over 50,000 people) or urban cluster (2500 to 50,000 people); for the purposes of estimating the urban/rural ratio of a county, we counted both urbanized and urban cluster tracts as being urban. Because food deserts are defined quite differently for urban (0.5 mile) versus rural (10 miles) counties in the FARA, we consider just those counties whose populations we calculated as at least seven–eighths (87.5%) urban, totaling n  = 250 counties across the U.S.

We find the reverse gradient has only existed for less than thirty years. In the U.S. in 1990, when population–scale obesity rates were about a third of what they are today, there was no correlation between income and obesity or diabetes. The inverse correlations between income and diabetes/obesity rates have developed only within the past thirty years. By 2015, the correlation was stronger than ever: in states where median household incomes were below $45,000 per year, like Alabama, Mississippi and West Virginia, over 35% of the population was obese, whereas obesity was less than 25% of state populations where median incomes were above $65,000, such as in Colorado, Massachusetts or California.

In the U.S., there is considerable geographic heterogeneity in obesity prevalence. Figure 1 shows maps of obesity rates and diabetes rates by county in 2013 (Table 1 ). By 2013, the reverse gradient in the U.S. was pronounced; the simple correlation between obesity and ln (income) across n  = 3110 U.S. counties was r  = −0.486 and between ln (income) and diabetes was r  = −0.531. All reverse gradients are better determined against the logged income data than against median income itself. Plots of ln (income) against both diabetes and obesity (Fig. 2 ), aggregated at county level, reveal mild degrees of non–linearity in each of the relationships.

figure 1

Prevalence of adult a obesity and b diabetes in 2013, mapped at the scale of U.S. county for CDC age-adjusted figures

figure 2

Reverse gradients across U.S. counties in 2013 (in all states except Alaska) between the natural logarithm of household income and rates of a obesity (regression slope = −9.66) and b diabetes (slope = −4.93). Also shown are correlations between prevalence of physical inactivity (“leisure”) versus c obesity (slope = 0.643) and d diabetes (slope = 0.288). Correlation values within each state are listed in Table 1

State-level data on annual obesity and diabetes rates, available for years 1990 through present, together with age–adjusted and inflation–adjusted income data from the U.S. Census, show how the reverse gradient in the U.S. changed over twenty–five years (Fig. 3 ). For the year 1990 (data from n  = 43 states for this first year of data), the Pearson correlation between state-level obesity and the natural log of median household income was r  = −0.240 [−0.502, 0.061], which is not significant ( p  = 0.116).

figure 3

Negative gradient between household income and obesity and diabetes rates. Scatterplots showing a Obesity vs. ln (income) by state, 1990 and 2015; b Diabetes vs. ln (income) by state, 1990 and 2015. For 1990 (blue), the slopes are −2.39 for obesity and −0.75 for diabetes; for 2015, the slopes −16.26 for obesity and −8.18 for diabetes. Panels c , d show the change in these correlations over time, between ln (income) and c Obesity and d Diabetes, at both state (solid) and county (dashed) levels

Twenty–five years later, a strong inverse correlation had developed between median household income and rates of obesity and diabetes (Figs. 3a, b ). In 2015, the correlation between ln (income) and obesity rate across all 50 states was r  = −0.697 [−0.816, −0.522], which is highly significant ( p  < 0.00001); even limited to the 43 states also recorded in 1990, the correlation in 2015 still yields r  = −0.699 [−0.824, −0.508] ( p  < 0.00001).

A similar change is evident for the reverse gradient involving diabetes and ln (income); insignificant for 1990 ( r  = −0.090 [−0.380, 0.216], p  = 0.566) and highly significant by 2015 with r  = −0.706 [−0.823, −0.532] ( p  < 0.00001). Again, if we restrict the 2015 data to the 42 states available in 1990 (one state fewer than in the obesity data), the correlation between ln (income) and diabetes rate yields r  = −0.684 [−0.816, −0.483] ( p  < 0.00001).

Figures 3a, b show the actual and fitted values of the regressions of obesity and diabetes, respectively, on the log of income in both 1990 and 2015. In 1990, neither slope was significantly different than zero at the state level. In the regressions using 2015 state data, the slope coefficients for obesity are −0.163 ± 0.024 ( R 2  = 0.476) and for diabetes −0.080 ± 0.011 ( R 2  = 0.491). Figure 4 shows how the slope coefficient in these regressions has evolved, using data in 1990, 1995, 2000, 2003, and then annually from 2005 onwards.

figure 4

a Slopes and b intercepts of the reverse gradients between household income and obesity and diabetes rates, 1990–2015

Figure 5 shows how, at 5–year intervals from 1990 to 2015, the rate of growth in obesity or diabetes per year was inversely related to median household income. The temporal evolution of these reverse gradients, for the rates of obesity and diabetes, respectively, can be described by the following equations:

where X t is the natural logarithm of median household income in year t . The colored lines in Fig. 5 show how well Eqs 6a and 6b represent the actual reverse gradients of ln (income) versus obesity and diabetes rates, respectively, across 25 years of evolution of these negative gradients. The evolving slope coefficients imply that above an annual income level of $250,000 for obesity and $150,000 for diabetes, any further increases in income have negligible effects in term of further reducing obesity and diabetes (Fig. 5c ).

figure 5

Evolution of the reverse gradients for a obesity and b diabetes at 5–year intervals from 1990 to 2015. Colored lines show how the time-evolution of these gradients can be described by the equations in Eqs 6 a and 6b , which yields c an approximated annual change as a function of household income

Available data at the U.S. county level are available only from year 2004, so do not capture the start of this phenomenon, but these data include not only diabetes and obesity rates but also a “leisure” statistic derived from self–reported activity levels. Across all 3110 counties in any given year, the leisure statistic correlates best with both obesity but also with income, reflecting the feedback between income, health habits and obesity (Table 2 ). For the leisure statistic in 2013, for example, the relationship with both obesity and diabetes is strongly positive and linear (Figs. 2c, d ) with r  = 0.719 [0.701, 0.735] for leisure versus obesity, and r  = 0.686 [0.667, 0.704] for leisure versus diabetes.

We find that each year 2004 to 2013 the leisure and income statistics were a good predictor of ( r 2  > 0.5) obesity rates at the county level (Table 2 ). Notably, the respective regression coefficient on leisure grew from about 0.40 to 0.63 (Table 2 , Fig. 6 ). This means that if we applied the 2004 regression coefficient, the actual obesity rate would be 1.56 times our prediction based on 2013 leisure rate. Hence following Glaeser et al., ( 2003 ), we estimate the social multiplier as at least 1.56, since by 2004 there had already been a decade of sharp increase in obesity rates.

figure 6

Change in multiple regressions, U.S. counties, 2004–2013. For regressions predicting obesity rates, the open circles (with blue curve) show the coefficient on the leisure statistic; black filled circles show coefficient on ln (income). The dashed line shows the trend through the black filled circles ( R 2  = 0.36). Data are shown in Table 2

This steady increase change in coefficient (Table 2 ) is not due to change in the leisure rate, which rose and fell: the average county rate increased from 25.3% in 2004 to a peak of 26.9% in 2009 and then fell to 24.7% by 2013. We are not aware of an individual–level reason why the relationship between lack of exercise and obesity will have changed in ten years. We therefore posit 1.56 as a measure of social multiplier effect at county level.

For the time series of state–level obesity rates from 1990 to 2016, regression analyses serve to rule out any strong social–shaped pattern in obesity rise for any of the U.S. states. Figure 7 shows typical examples. Using the local linear regression approach (see Methods), the null hypothesis of linearity could not be rejected for seven of the 46 states examined (the few we excluded lacked data points for the early 1990s): Connecticut; Delaware; Iowa; Louisiana; Maine; Vermont; Wisconsin; Wyoming. For a further seven states, linearity could be rejected at the standard 5 per cent level, but only when the alternative exhibits a mild degree of non-linearity, with the Equivalent Number of Parameters being just 2.33: Idaho, Illinois; Kentucky; North Dakota; Oregon; Virginia; West Virginia. In the adoption-curve analysis, this favors “r–curves” of individual learning (Fig. 7 ). Adjusted r -squared values (Table 3 ) are strong where the individual parameter p is the same magnitude as the social parameter q , and fit almost as well with the “social” parameter, q , set to zero (Table 3 ) to represent pure individual learning.

figure 7

Rise in state-level obesity rates in four different states, showing the fit of the adoption curve (Eq. 4 ) with the social parameter, q , set to zero (solid red), as well as a linear fit (dashed black). See Table 3

Lastly, the effect of food deserts is also evident, among a subset of U.S. counties (Fig. 8 , Table 4 ). As described in our Methods, from the FARA data we consider just those counties whose populations we calculated as at least seven eights (87.5%) urban, totaling n  = 250 counties across the U.S. In these urban counties, the regression of obesity rates versus the share of population without access to supermarkets within half a mile yields Pearson’s r  = 0.292 [0.175, 0.401], which is significant ( p  < 0.00001). For diabetes, however, the correlation with this food desert variable in 2013 is not significantly different from zero ( p  = 0.223). For these same urban counties, food deserts among low income census tracts has stronger correlation with obesity (Fig. 8b ), with r  = 0.563 [0.472, 0.642], and now a significant correlation with diabetes rate r  = 0.462 [0.359, 0.544]; both correlations are highly significant ( p  < 0.000001).

figure 8

Food desert index versus obesity rate for 250 urban counties in the U.S. in 2013. In a the food desert index is the share of the urban population living a half mile or more from a supermarket. The blue line shows the regression (Pearson’s r  = 0.292 ± 0.11). In b the food desert index is the share of low-income population living a half mile from supermarket. Blue line shows the regression (Pearson’s r  = 0.563 ± 0.08)

Food deserts are, of course, closely related to income. We confirm the validity of the individual correlations between obesity and diabetes and income and leisure in Table 2 by running simple multiple regressions of obesity and diabetes on income and leisure. The results are set in Table 5 . Although each overall fit could be slightly improved with mild non–linearities, the simple regressions show that both ln (income) and leisure have significant effects on obesity and diabetes.

Our analyses here use data aggregated from both sexes, as the same correlations for each sex were quite similar to the results for both sexes aggregated together. In 2013 for example, among the 43 states with sufficient data points for men and women (excluding AK, CT, DE, HI, NH, and RI), the reverse gradient between income and diabetes is significant for both sexes at p  < 0.01 in 30 states. In some states, however, there are small differences. In eight states the reverse gradient in 2013 is significant at p  < 0.01 for one sex and p  < 0.05 for the other. In Arizona, New Mexico, New Jersey and Utah, the reverse gradient is significant for men but not for women, whereas in Massachusetts, it is significant for women but not for men.

Here we have explored the origins and development of the inverse correlation between household income and obesity/diabetes rates in the U.S. We used data on mean household incomes and rates of obesity and diabetes at the level of U.S. state, which date back to 1990, as well as county level statistics that offer larger sample sizes and higher spatial resolution but only extend back the early 2000s.

Using age–adjusted U.S. data on mean household incomes and rates of obesity and diabetes—at state level since 1990 and county level since the mid 2000s—we found that the reverse gradient originated and evolved over a period of about 25 years. Here, we report that this reverse gradient did not exist in the U.S. in 1990 but has increased markedly since then.

Specifically, across the U.S. states by 2015 there were highly significant correlations between ln (income) and state–level rates of obesity ( r  = −0.697, p  < 0.00001) and diabetes ( r  = −0.706, p  < 0.00001), whereas in 1990 neither correlation was yet evident. By 2013, the age-adjusted prevalence of obesity in the U.S. was 35% among men and 40% among women (Flegal et al., 2016 )—across all age adult groups, obesity rates among U.S. women have been 4% to 5% higher than among men (Arroyo–Johnson and Mincey, 2016 ). Since 1990, this change was continual, such that we determine equations for the linear development of these reverse gradients since 1990.

In the U.S., an inverse correlation between obesity and median household income (logarithm) developed from nonexistent in 1990 into a steep inverse correlation, known as the “reverse gradient”, 25 years later. The reverse gradient involving diabetes prevalence also developed in the U.S. over the same period; this lagged the reverse gradient with obesity in 1990 but had fully caught up with it after 2010 (Fig. 3 ). Both facets of the reverse gradient developed remarkably quickly in the U.S., in about one generation.

Any hypotheses for the recent rise in obesity must account for the obesity epidemic having emerged only in the last few decades in the U.S. We observed that over ten years, the regression coefficient on leisure in predicting obesity increased from 0.40 to 0.63. We interpret this to be a social multiplier effect, as a physiological effect ought to have had that same coefficient through time. Our social multiplier estimate of 1.56 is close to that of crime and U.S. wages, both about 1.7, estimated by Glaeser et al., ( 2003 ) at similar scales of aggregation. While this is some evidence for social influence, these estimates might overstate the true social multiplier, due to correlation between demographics and unobservable elements (Glaeser et al., 2003 ).

Our simple S-curve approach was more illuminating than we had expected (see Kandler and Powell, 2018 ), because they were unexpectedly linear or perhaps slightly r-shaped (sensu Henrich 2001 ), i.e., the adoption patterns do not show much evidence for social (S-shaped) diffusion. A plausible explanation for the steady, 25-year increase in state–level obesity rates is a higher-obesity younger generation progressively entering the adult cohort.

The recent origin of the reverse gradient appears to favor the Insurance Hypothesis of HBE. The time series reveal only weak evidence of social learning in most states, insufficient to falsify the HBE hypothesis that obesity increased due to individual responses to a changing nutritional/economic environment. Additionally, there is a lack of evidence for a deep cultural history to obesity. While economic development may be a prerequisite for the reverse gradient (Tafreschi, 2015 ), the U.S. and Europe possessed developed economies for a century before the reverse gradient materialized. In Western Europe, there was still no reverse gradient as of 2008 (García Villar and Quintana-Domeque, 2009 ). The second is the fact that the reverse gradient developed smoothly over time, as described by Eq 6, which indicates the close relationship between income levels and propensity toward obesity. These observations are consistent with the Insurance Hypothesis, which is predicated on an evolved tendency for lower-income people to perceive risks in their local environment and over–compensate through excessive calorie intake.

There are alternatives to the Insurance Hypothesis, as an explanation based upon the behavioral responses of individuals. At the scale of economic geography, a significant factor is food deserts, where “easy geographic access to fast-food outlets and convenience stores encourages individuals to consume foods that are high in energy and saturated fats” (Mullan et al., 2017 ). Over 50 million people, almost 18% of the U.S. population, live in low–income areas without convenient access to a supermarket (Rhone et al., 2017 ). In high–income, highly urbanized countries, diabetes correlates positively with the percentage living in urban areas (Goryakin et al., 2017 ).

Some research has focused on the effect of highly processed foods, which typically contain much more added sugar than unprocessed foods (Lhila, 2011 ; Bocarsly et al., 2010 ; Jürgens et al., 2005 ; Martínez Steele et al., 2016 ; Stanhope et al., 2009 ). Excessive sugar intake, which may be addictive (Avena et al., 2008 ) is a causal factor in diabetes (Hu and Malik, 2010 ; Shang et al., 2012 ; Cornelsen et al., 2016 ) and may also be a causal factor in high obesity rates (Basu et al., 2013 ; Hu and Malik, 2010 ; Shang et al., 2012 ).

Americans have consumed refined sugar since the nineteenth century, however, so the question remains why the obesity increase, and the reverse gradient happened only in the past three decades. One possible explanation is the recent introduction of high fructose corn syrup (HFCS) into the food economy. Fructose, which decreases insulin sensitivity in obese people (Stanhope et al., 2009 ), has been used in commercial sugar–sweetened beverages since about 1970. That said, the trend might be due to unobserved individual–level effects, such as the more leisure, the more HFCS drinks consumed. The timing is suggestive; Fig. 9 shows a timeline of the increase in contribution of refined sugar and HFCS to U.S. diet, together with the increase in U.S. obesity rate. While overall sugar consumption rose gradually in the 20th century, from 12% of U.S. food energy in 1909 to 19% by the year 2000, the use of high fructose corn syrup in the U.S. increased from virtually zero per capita in 1970 to over 60 pounds per capita annually in the U.S. in 2000 (Gerrior et al., 2004 ), about half of total sugar consumption. HFCS became the main sweetener in soft drinks. By 2016 in the U.S., sweetened beverages constituted over 7% of household food expenditures and over 9% of expenditures for low-income households in the SNAP program (Garasky et al., 2016 ).

figure 9

A timeline of the increase in contribution of refined sugar and high fructose corn syrup (HFCS) to U.S. diet, together with the increase in U.S. obesity rate. The data for sugar, dairy and HFCS consumption per capita are from USDA Economic Research Service (Johnson et al., 2009 ) except for sugar consumption before 1967, which are historical estimates (Guyenet et al., 2017 ). Obesity data (% of U.S. adult population) are from the Robert Wood Johnson Foundation’s Trust for America’s Health (stateofobesity.org). Total U.S. television advertising data are from the World Advertising Research Center (www.warc.com). The y –axis on the left covers all data series except advertising expenditures, which uses the y –axis on the right

The metabolic effects of HFCS include complications of glucose metabolism, lipid profile and insulin resistance (Pereira et al., 2017 ; Johnson et al., 2016 ; Bocarsly et al., 2010 ; Bray et al., 2004 ; Jürgens et al., 2005 ). HFCS as the driver of obesity and diabetes epidemics would be consistent with HBE in the general sense of human physiology having evolved around a diet containing little sugar and no refined carbohydrates. Hunter–gatherers generally do not exhibit obesity, diabetes, or cardiovascular disease (Kaplan et al., 2017 ). The HFCS explanation is also consistent with the Insurance Hypothesis, in that poor families are most subject to food scarcity (Hernandez, 2015 ) and HFCS-sweetened beverages predominate the food economy of poor regions of the U.S.

Conclusions

In conclusion, we find a steady increase, since 1990, in the “reverse gradient” or negative correlation between median household income and both obesity and diabetes rates. In 1990, there was no correlation across the US between either obesity and income or diabetes and income, yet by 2015 strong negative correlations existed across and within U.S. States. We have determined equations for the continual development of these reverse gradients over the past 25 years.

To explain this change, we find evidence in support for both HBE and “social multiplier” effect, a balance similar to empirical studies of other human behavior (Aral et al., 2009 ). We ascribe more weight to the HBE explanation, in that evolved mechanisms that increase fat storage in response to resource scarcity should promote obesity in high-income countries, where the poor have greater exposure to junk food and other cheap calories including processed sugars (Hill et al., 2017 ; Wells, 2017 ).

We speculate that the rapid increase in consumption of high fructose corn syrup (HFCS) may have been a key driver. The obesity and diabetes epidemics could be driven by the commercial oversupply and widespread marketing of inexpensive high-sugar foods, especially HFCS–sweetened beverages (Johnson et al., 2007 ; Song et al., 2012 ; Basu et al., 2013 ).

A fuller explanation of the timing and geography of the obesity epidemic will require the specific history of societal–level factors. Besides the suggestive temporal concurrence between obesity, food deserts and HFCS–sweetened beverages, additional clues lie in the considerable variation in the strength and evolution of the reverse gradient within different states of the U.S. This marked geographic variation in the slope of the reverse gradient indicates that government health policies can mitigate the effect of socioeconomic disparities. To explore the scale of these drivers, future work would review and compare state level health policies versus how the negative gradient evolved in those states.

Data availability

All data we used for this study are publicly–accessible, aggregated data. The datasets analyzed during the current study are available in the Dataverse repository: https://doi.org/10.7910/DVN/SMTX3X . These datasets were derived from the following public domain resources: Age–adjusted data on obesity rates at the county and state level for years 2004 to 2015 (Centers for Disease Control and Prevention, 2017d ), as well as diabetes rates for 2009-2013 (Centers for Disease Control and Prevention, 2017d ), are available from CDC. Available at: www.cdc.gov/diabetes/data/countydata/countydataindic State-level obesity rates since 1990 were obtained from the annual reports of the Trust for America’s Health (Robert Wood Johnson Foundation). Available at: stateofobesity.org/adult-obesity. Estimates of “leisure” (physical inactivity) are available from the CDC Behavioral Risk Factor Surveillance System (Centers for Disease Control and Prevention, 2017a ). These CDC data come from health-related telephone surveys, which began in 1984 with 15 U.S. states, and now collects data in all 50 states through over 400,000 adult interviews each year. Available at: www.cdc.gov/diabetes/data/countydata/countydataindic. Food desert designations for the U.S. were recently made available through the Food Access Research Atlas (FARA) project (Rhone et al., 2017 ). The estimates are derived from the 2010 US Census and the 2010-2014 American Community Survey, in which census tracts are categorized by median income, vehicle availability, and SNAP participation. Available at: www.ers.usda.gov/webdocs/publications/82101/eib-165.pdf?v=42752 .

Alvarado SE (2016) Neighborhood disadvantage and obesity across childhood and adolescence: Evidence from the NLSY children and young adults cohort (1986-2010). Soc Sci Res 57:80–98

Article   PubMed   Google Scholar  

Anderson PM (2012) Parental employment, family routines and childhood obesity. Econ Hum Biol 10:340–351

Anderson SE, Whitaker RC (2010) Household routines and obesity in us preschool-aged children. Pediatrics 125:420–428

Aral S, Muchnik L, Sundararajan A (2009) Distinguishing influence-based contagion from homophily-driven diffusion in dynamic networks. Proc Natl Acad Sci USA 106(51):21544–21549

Article   ADS   CAS   PubMed   PubMed Central   Google Scholar  

Arroyo–Johnson C, Mincey KD (2016) Obesity epidemiology trends by race/ethnicity, gender, and education: National Health Interview Survey, 1997–2012. Gastroenterol Clin North Am 45(4):571–579

Article   PubMed   PubMed Central   Google Scholar  

Avena NM, Rada P, Hoebel BG (2008) Evidence for sugar addiction: behavioral and neurochemical effects of intermittent, excessive sugar intake. Neurosci Biobehav Rev 32:20–39

Article   CAS   PubMed   Google Scholar  

Bass FM (1969) A new product growth model for consumer durables. Manag Sci 15:215–227

Article   MATH   Google Scholar  

Basu S, Yoffe P, Hills N, Lustig RH (2013) The relationship of sugar to population-level diabetes prevalence: An econometric analysis of repeated cross-sectional data. PLoS One 8(2):e57873

Bentley RA, Ormerod P (2010) A rapid method for assessing social versus independent interest in health issues. Soc Sci Med 71:482–485

Bentley RA, Brock WA, O’Brien MJ (2016) Evaluating reproductive decisions as discrete choices under social influence. Philos Trans R Soc B 371:20150154

Article   Google Scholar  

Blume LE, Brock WA, Durlauf SN, Ioannides YM (2011) Identification of social interactions. Handb Social Econ 1B:853–963

Bocarsly ME, Powell ES, Avena NM, Hoebel BG (2010) High-fructose corn syrup causes characteristics of obesity in rats: Increased body weight, body fat and triglyceride levels. Pharmacol Biochem Behav 97(1):101–106

Article   CAS   PubMed   PubMed Central   Google Scholar  

Boden J, McLeod G (2017) Epidemiological foundations for the insurance hypothesis: Methodological considerations. Behav Brain Sci 40:e109

Bray GA, Nielsen SJ, Popkin BM (2004) Consumption of high–fructose corn syrup in beverages may play a role in the epidemic of obesity. Am J Clin Nutr 79(4):537–543

Brock WA, O’Brien MJ, Bentley RA (2015) Validating niche–construction theory through path analysis. Archaeol Anthropol Sci 8:819–837

Carroll SJ, Paquet C, Howard NJ, Coffee NT, Taylor AW, Niyonsenga T, Daniel M (2016) Local descriptive norms for overweight/obesity and physical inactivity, features of the built environment, and 10–year change in glycosylated haemoglobin in an Australian population-based biomedical cohort. Soc Sci Med 166:233–243

Centers for Disease Control and Prevention (2017a) Behavioral risk factor surveillance system. http://www.cdc.gov/brfss . Accessed on 20 Oct 2016

Centers for Disease Control and Prevention (2017b) Behavioral risk factor surveillance system: methods and references for county–level estimates and ranks and state–level modeled estimates. http://www.cdc.gov/brfss . Accessed on 20 Oct 2016

Centers for Disease Control and Prevention (2017c) Data and statistics, diabetes. http://www.cdc.gov/diabetes/data . Accessed on 20 Oct 2016

Centers for Disease Control and Prevention (2017d) Data and statistics, overweight and obesity. http://www.cdc.gov/obesity . Accessed on 20 Oct 2016

Centola D, Becker J, Brackbill D, Baronchelli A (2018) Experimental evidence for tipping points in social convention. Science 360:1116–1119

Article   ADS   CAS   PubMed   Google Scholar  

Christakis NA, Fowler JH (2007) The spread of obesity in a large social network over 32 years. N Engl J Med 357:370–379

Cleveland WS, Devlin SJ (1988) Locally weighted regression: an approach to regression analysis by local fitting. J Am Stat Assoc 83:596–610

Christakis NA, Fowler JH (2013) Social contagion theory: examining dynamic social networks and human behavior. Stat Med 32:556–577

Article   MathSciNet   PubMed   Google Scholar  

Colleran H (2016) The cultural evolution of fertility decline. Philos Trans R Soc B 371:20150152

Colleran H, Mace R (2015) Social network-and community-level influences on contraceptive use: evidence from rural Poland. Proc R Soc B 282:20150398

Cook WK, Tseng W, Tam TC, John I, Lui C (2017) Ethnic-group socioeconomic status as an indicator of community–level disadvantage: A study of overweight/obesity in Asian American adolescents. Soc Sci Med 184:15–22

Article   ADS   PubMed   PubMed Central   Google Scholar  

Corcoran M, Gordon R, Laren D, Solon G (1992) The association between men’s economic status and their family and community origins. J Hum Resour 27:575–601

Cornelsen L, Adams J, Gasparrini A, Iskander D, Knai C, Mytton O, Petticrew M, Scott C, Smith R, Thompson C, White M, Cummins S (2016) Impact of a levy on sales of sugar–sweetened beverages within a national chain of restaurants: Interrupted time-series analysis. Lancet 388:S15

Couzin ID, Ioannou CC, Demirel G, Gross T, Torney CJ, Hartnett A, Conradt L, Levin SA, Leonard NE (2011) Uninformed individuals promote democratic consensus in animal groups. Science 334(6062):1578–1580

Creanza N, Feldman MW (2016) Current opinion in genetics and development. Lancet 41:85–92

CAS   Google Scholar  

Datar A (2017) The more the heavier? Family size and childhood obesity in the U.S. Soc Sci Med 180:143–151

Dittmann AG, Maner JK (2017) A life–history theory perspective on obesity. Behav Brain Sci 40:e115

Dohle S, Hafmann W (2017) Toward a mechanistic understanding of the impact of food insecurity on obesity. Behav Brain Sci 40:e116

Dwyer–Lindgren L, Freedman G, Engell RE, Fleming TD, Lim SS, Murray CJL, Mokdad AH (2013) Prevalence of physical activity and obesity in U.S. counties, 2001–2011: A road map for action. Popul Health Metr 11:7

Flegal KM, Kruszon–Moran MS, Caroll MD, Fryar CL, Ogden CL (2016) Trends in obesity among adults in the United States, 2005 to 2014. JAMA 315(21):2284–2291

Garasky S, Mbwana K, Romualdo A, Tenaglio A, Roy M (2016) Foods typically purchased by snap households. United States Department of Agriculture, Food and Nutrition Service. http://www.fns.usda.gov/research-and-analysis . Accessed on 17 Feb 2017

García Villar J, Quintana-Domeque C (2009) Income and body mass index in Europe. Econ Hum Biol 7:73–83

Gerbault P, Roffet–Salque M, Evershed R, Thomas M (2013) How long have adult humans been drinking milk? Int Union Biochem Mol Biol 65:983–990

Gerrior S, Bente L, Hiza H (2004) Nutrient Content of the U.S. Food Supply, 1909–2000. United States Department of Agriculture

Glaeser EL, Sacerdote BI, Scheinkman JA (2003) The social multiplier. J Eur Econ Assoc 1(2-3):345–353

Goryakin Y, Rocco L, Suhrcke M (2017) The contribution of urbanization to non–communicable diseases: evidence from 173 countries from 1980 to 2008. Econ Hum Biol 26:151–163

Guyenet S, Landen J, Houser RF, Baghdady G (2017) Sugar consumption in the U.S. diet between 1822 and 2005. OnlineStatBook Project, case study 27, http://www.onlinestatbook.com . Accessed on 28 Apr 2017

Henrich J (2001) Cultural transmission and the diffusion of innovations. Am Anthropol 103(4):992–1013

Hernandez DC (2015) The impact of cumulative family risks on various levels of food insecurity. Soc Sci Res 50:292–302

Higginson AD, McNamara JM, Dall SRX (2017) Towards a behavioural ecology of obesity. Behav Brain Sci 40:e118

Hill S, Proffitt Leyva R, DelPriore D (2017) Predicting human adiposity—sometimes—with food insecurity: Broaden the model for better accuracy. Behav Brain Sci 40:e119

Hobaiter C, Poisot T, Zuberbühler K, Hoppitt W, Gruber T (2014) Social network analysis shows direct evidence for social transmission of tool use in wild chimpanzees. PLoS Biol 12(9):e1001960

Article   PubMed   CAS   PubMed Central   Google Scholar  

Hruschka DJ (2012) Do economic constraints on food choice make people fat? A critical review of two hypotheses for the poverty–obesity paradox. Am J Hum Biol 24(3):277–285

Hruschka DJ, Hadley C, Brewis A (2014) Disentangling basal and accumulated body mass for cross–population comparisons. Am J Phys Anthropol 153:542–550

Hruschka DJ, Han SY (2017) Anti-fat discrimination in marriage more clearly explains the poverty–obesity paradox. Behav Brain Sci 40:e120

Hu FB, Malik VS (2010) Sugar–sweetened beverages and risk of obesity and type 2 diabetes: epidemiologic evidence. Physiol Behav 100(1):47–54

Hughes CC, Sherman SN, Whitaker RC (2010) How low–income mothers with overweight preschool children make sense of obesity. Qual Health Res 20:465–478

Johnson ML, Distelmaier K, Lanza IR, Irving BA, Robinson MM, Konopka AR, Shulman GI, Nair KS (2016) Mechanism by which caloric restriction improves insulin sensitivity in sedentary obese adults. Diabetes 65:74–84

CAS   PubMed   Google Scholar  

Johnson RJ, Segal MS, Sautin Y, Nakagawa T, Feig DI, Kang DH, Gersch MS, Benner S, Sánchez–Lozada L (2007) Potential role of sugar (fructose) in the epidemic of hypertension, obesity and the metabolic syndrome, diabetes, kidney disease, and cardiovascular disease. Am J Clin Nutr 86:899–906

Johnson RK, Appel LJ, Brands M, Howard BV, Lefevre M, Lustig RH, Sacks F, Steffen LM, Wylie–Rosett J (2009) Dietary sugars intake and cardiovascular health: a scientific statement from the American Heart Association. Circulation 120:1011–1020

Jürgens H, Haass W, Castañeda TR, Schürmann A, Koebnick C, Dombrowski F, Otto B, Nawrocki AR, Scherer PE, Spranger J, Ristow M, Joost HG, Havel PJ, Tschöp MH (2005) Consuming fructose–sweetened beverages increases body adiposity in mice. Obes Res 13:1146–1156

Kandler A, Powell A (2018) Generative inference for cultural evolution. Philos Trans R Soc B 373(1743):20170056

Kaplan H, Thompson RC, Trumble BC, Wann LS, Allam AH, Beheim B, Frohlich B, Sutherland ML, Sutherland JD, Stieglitz J, Rodriguez DE, Michalik DE, Rowan CJ, Lombardi GP, Bedi R, Garcia AR, Min JK, Narula J, Finch CE, Gurven M, Thomas GS (2017) Coronary atherosclerosis in indigenous south American Tsimane: a cross–sectional cohort study. Lancet 389:1730–1739

Klein RJ, Schoenborn CA (2001) Age adjustment using the 2000 projected U.S. population. Healthy People Stat Notes 20:1–10

Google Scholar  

Kowaleski-Jones L, Brown BB, Fan JX, Hanson HA, Smith KR, Zick CD (2017) The joint effects of family risk of obesity and neighborhood environment on obesity among women. Soc Sci Med 195:17–24

Kranjac AW, Wagmiller RL (2016) Decomposing trends in adult body mass index, obesity, and morbid obesity, 1971–2012. Soc Sci Med 167:37–44

Lhila A (2011) Does access to fast food lead to super-sized pregnant women and whopper babies? Econ Hum Biol 9:364–380

Mani A, Mullainathan S, Shafir E, Zhao J (2013) Poverty impedes cognitive function. Science 341:976–980

Martínez Steele E, Baraldi LG, Louzada ML, Moubarac JC, Mozaffarian D, Monteiro CA (2016) Ultra–processed foods and added sugars in the U.S. diet: Evidence from a nationally representative cross–sectional study. BMJ Open 6:e009892

Mata J, Dallacker M, Hertwig R (2017) Social nature of eating could explain missing link between food insecurity and childhood obesity. Behav Brain Sci 40:e122

Mattson MP, Allison DB, Fontana L, Harvie M, Longo VD, Malaisse WJ, Mosley M, Notterpek L, Ravussin E, Scheer FA, Seyfried TN, Varady KA, Panda S (2014) Meal frequency and timing in health and disease. Proc Natl Acad Sci USA 111:16647–16653

Milot E, Mayer FM, Nussey DH, Boisvert M, Pelletier F, Réale D (2011) Evidence for evolution in response to natural selection in a contemporary human population. Proc Natl Acad Sci USA 108:17040–17045

Mokdad AH, Serdula MK, Dietz WH, Bowman BA, Marks JS, Koplan JP (1999) The spread of the obesity epidemic in the United States, 1991-1998. JAMA 282(16):1519–1522

Mullan B, Ntoumanis N, Thøgersen-Ntoumani C, Lipp OV (2017) It’s a bit more complicated than that: A broader perspective on determinants of obesity. Behav Brain Sci 40:e124

Muscogiuri G, Balercia G, Barrea L, Cignarelli A, Giorgino F, Holst JJ, Laudisio D, Orio F, Tirabassi G, Colao A (2018) Gut: A key player in the pathogenesis of type 2 diabetes? Crit Rev Food Sci Nutr 58(8):1294–1309

Nettle D, Andrews C, Bateson M (2017) Food insecurity as a driver of obesity in humans: the insurance hypothesis. Behav Brain Sci 40:e105

Ogden CL, Carroll MD, Kit BK, Flegal KM (2014) Prevalence of childhood and adult obesity in the United States, 2011–2012. JAMA 311:806–814

Pereira RM, Botezelli JD, da Cruz Rodrigues KC, Mekary RA, Cintra DE, Pauli JR, da Silva ASR, Ropelle ER, de Moura LP (2017) Fructose consumption in the development of obesity and the effects of different protocols of physical exercise on the hepatic metabolism. Nutrients 9:article 405

Article   CAS   Google Scholar  

Redsell SA, Atkinson P, Nathan D, Siriwardena AN, Swift J, Glazebrook C (2010) Parents’ beliefs about appropriate infant size, growth and feeding behaviour: Implications for the prevention of childhood obesity. BMC Public Health 10:7–11

Rhone A, Ver Ploeg M, Dicken C, Williams R, Breneman V (2017) Low–income and low–supermarket–access census tracts, 2010–2015. United States Department of Agriculture, Economic Information Bulletin, EIB–165. http://ageconsearch.umn.edu/record/262134

Salmasi L, Celidon M (2017) Investigating the poverty–obesity paradox in Europe. Econ Hum Biol 26:70–85

Shalizi CR, Thomas AC (2011) Homophily and contagion are genetically confounded in observational social network studies. Sociol Methods Res 40:211–239

Article   MathSciNet   PubMed   PubMed Central   Google Scholar  

Shang XW, Liu AL, Zhang Q, Hu XQ, Du SM, Ma J, Xu GF, Li Y, Guy HW, Du L, Li TY, Ma GS (2012) Report on childhood obesity in China (9): Sugar–sweetened beverages consumption and obesity. Biomed Environ Sci 25(2):125–132

PubMed   Google Scholar  

Shrewsbury V, Wardle J (2012) Socioeconomic status and adiposity in childhood: a systematic review of cross–sectional studies 1990–2005. Obesity 16:275–284

Smith TG (2017) Household–level financial uncertainty could be the primary driver of the global obesity epidemic. Behav Brain Sci 40:e128

Smits SA, Leach J, Sonnenburg ED, Gonzalez CG, Lichtman JS, Reid G, Knight R, Manjurano A, Changalucha J, Elias JE, Dominguez–Bello MG, Sonnenburg JL (2017) Seasonal cycling in the gut microbiome of the Hadza hunter-gatherers of Tanzania. Science 357:802–806

Song WO, Wang Y, Chung CE, Song B, Lee W, Chun OK (2012) Is obesity development associated with dietary sugar intake in the U.S.? Nutrition 28:1137–1141

Stanhope KL, Schwarz JM, Keim NL, Griffen SC, Bremer AA, Graham JL, Hatcher B, Cox CL, Dyachenko A, Zhang W, McGahan JP, Seibert A, Krauss RM, Chiu S, Schaefer EJ, Ai M, Otokozawa S, Nakajima K, Nakano T, Beysen C, Hellerstein MK, Berglund L, Havel PJ (2009) Consuming fructose–sweetened, not glucose–sweetened, beverages increases visceral adiposity and lipids and decreases insulin sensitivity in overweight/obese humans. J Clin Invest 119(5):1322–1334

Subramanian SV, Perkins JM, Özaltin E, Davey Smith G (2011) Weight of nations: a socioeconomic analysis of women in low– to middle–income countries. Am J Clin Nutr 93:413–421

Tafreschi D (2015) The income body weight gradients in the developing economy of China. Econ Hum Biol 16:115–134

Tapper K (2017) Children respond to food restriction by increasing food consumption. Behav Brain Sci 40:e129

Thomas AC (2013) The social contagion hypothesis: comment on ‘Social contagion theory: examining dynamic social networks and human behavior’. Stat Med 32:581–590

Article   MathSciNet   CAS   PubMed   Google Scholar  

Vizireanu M, Hruschka DJ (2018) Lay perceptions of healthy eating styles and their health impacts. J Nutr Educ Behav 50:365–371.e1

Weinberg B, Reagan P, Yankow J (2004) Do neighborhoods affect hours worked? Evidence from longitudinal data. J Labor Econ 22:891–924

Wells JCK (2017) Obesity is not just elevated adiposity, it is also a state of metabolic perturbation. Behav Brain Sci 40:e130

Download references

Author information

Authors and affiliations.

Anthropology Department, University of Tennessee, 1621 Cumberland Ave., 520 Strong Hall, Knoxville, 37996, TN, USA

R. Alexander Bentley & Damian J. Ruck

Center for the Dynamics of Social Complexity, University of Tennessee, Knoxville, 37996, TN, USA

Centre for the Study of Decision-Making Uncertainty, University College London, Gower Street, London, WC1E 6BT, UK

Paul Ormerod

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to R. Alexander Bentley .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Bentley, R.A., Ormerod, P. & Ruck, D.J. Recent origin and evolution of obesity-income correlation across the United States. Palgrave Commun 4 , 146 (2018). https://doi.org/10.1057/s41599-018-0201-x

Download citation

Received : 25 June 2018

Accepted : 13 November 2018

Published : 11 December 2018

DOI : https://doi.org/10.1057/s41599-018-0201-x

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

Quick links

  • Explore articles by subject
  • Guide to authors
  • Editorial policies

null hypothesis about poverty

Statology

Statistics Made Easy

How to Write a Null Hypothesis (5 Examples)

A hypothesis test uses sample data to determine whether or not some claim about a population parameter is true.

Whenever we perform a hypothesis test, we always write a null hypothesis and an alternative hypothesis, which take the following forms:

H 0 (Null Hypothesis): Population parameter =,  ≤, ≥ some value

H A  (Alternative Hypothesis): Population parameter <, >, ≠ some value

Note that the null hypothesis always contains the equal sign .

We interpret the hypotheses as follows:

Null hypothesis: The sample data provides no evidence to support some claim being made by an individual.

Alternative hypothesis: The sample data  does provide sufficient evidence to support the claim being made by an individual.

For example, suppose it’s assumed that the average height of a certain species of plant is 20 inches tall. However, one botanist claims the true average height is greater than 20 inches.

To test this claim, she may go out and collect a random sample of plants. She can then use this sample data to perform a hypothesis test using the following two hypotheses:

H 0 : μ ≤ 20 (the true mean height of plants is equal to or even less than 20 inches)

H A : μ > 20 (the true mean height of plants is greater than 20 inches)

If the sample data gathered by the botanist shows that the mean height of this species of plants is significantly greater than 20 inches, she can reject the null hypothesis and conclude that the mean height is greater than 20 inches.

Read through the following examples to gain a better understanding of how to write a null hypothesis in different situations.

Example 1: Weight of Turtles

A biologist wants to test whether or not the true mean weight of a certain species of turtles is 300 pounds. To test this, he goes out and measures the weight of a random sample of 40 turtles.

Here is how to write the null and alternative hypotheses for this scenario:

H 0 : μ = 300 (the true mean weight is equal to 300 pounds)

H A : μ ≠ 300 (the true mean weight is not equal to 300 pounds)

Example 2: Height of Males

It’s assumed that the mean height of males in a certain city is 68 inches. However, an independent researcher believes the true mean height is greater than 68 inches. To test this, he goes out and collects the height of 50 males in the city.

H 0 : μ ≤ 68 (the true mean height is equal to or even less than 68 inches)

H A : μ > 68 (the true mean height is greater than 68 inches)

Example 3: Graduation Rates

A university states that 80% of all students graduate on time. However, an independent researcher believes that less than 80% of all students graduate on time. To test this, she collects data on the proportion of students who graduated on time last year at the university.

H 0 : p ≥ 0.80 (the true proportion of students who graduate on time is 80% or higher)

H A : μ < 0.80 (the true proportion of students who graduate on time is less than 80%)

Example 4: Burger Weights

A food researcher wants to test whether or not the true mean weight of a burger at a certain restaurant is 7 ounces. To test this, he goes out and measures the weight of a random sample of 20 burgers from this restaurant.

H 0 : μ = 7 (the true mean weight is equal to 7 ounces)

H A : μ ≠ 7 (the true mean weight is not equal to 7 ounces)

Example 5: Citizen Support

A politician claims that less than 30% of citizens in a certain town support a certain law. To test this, he goes out and surveys 200 citizens on whether or not they support the law.

H 0 : p ≥ .30 (the true proportion of citizens who support the law is greater than or equal to 30%)

H A : μ < 0.30 (the true proportion of citizens who support the law is less than 30%)

Additional Resources

Introduction to Hypothesis Testing Introduction to Confidence Intervals An Explanation of P-Values and Statistical Significance

' src=

Published by Zach

Leave a reply cancel reply.

Your email address will not be published. Required fields are marked *

What is The Null Hypothesis & When Do You Reject The Null Hypothesis

Julia Simkus

Editor at Simply Psychology

BA (Hons) Psychology, Princeton University

Julia Simkus is a graduate of Princeton University with a Bachelor of Arts in Psychology. She is currently studying for a Master's Degree in Counseling for Mental Health and Wellness in September 2023. Julia's research has been published in peer reviewed journals.

Learn about our Editorial Process

Saul Mcleod, PhD

Editor-in-Chief for Simply Psychology

BSc (Hons) Psychology, MRes, PhD, University of Manchester

Saul Mcleod, Ph.D., is a qualified psychology teacher with over 18 years experience of working in further and higher education. He has been published in peer-reviewed journals, including the Journal of Clinical Psychology.

Olivia Guy-Evans, MSc

Associate Editor for Simply Psychology

BSc (Hons) Psychology, MSc Psychology of Education

Olivia Guy-Evans is a writer and associate editor for Simply Psychology. She has previously worked in healthcare and educational sectors.

On This Page:

A null hypothesis is a statistical concept suggesting no significant difference or relationship between measured variables. It’s the default assumption unless empirical evidence proves otherwise.

The null hypothesis states no relationship exists between the two variables being studied (i.e., one variable does not affect the other).

The null hypothesis is the statement that a researcher or an investigator wants to disprove.

Testing the null hypothesis can tell you whether your results are due to the effects of manipulating ​ the dependent variable or due to random chance. 

How to Write a Null Hypothesis

Null hypotheses (H0) start as research questions that the investigator rephrases as statements indicating no effect or relationship between the independent and dependent variables.

It is a default position that your research aims to challenge or confirm.

For example, if studying the impact of exercise on weight loss, your null hypothesis might be:

There is no significant difference in weight loss between individuals who exercise daily and those who do not.

Examples of Null Hypotheses

When do we reject the null hypothesis .

We reject the null hypothesis when the data provide strong enough evidence to conclude that it is likely incorrect. This often occurs when the p-value (probability of observing the data given the null hypothesis is true) is below a predetermined significance level.

If the collected data does not meet the expectation of the null hypothesis, a researcher can conclude that the data lacks sufficient evidence to back up the null hypothesis, and thus the null hypothesis is rejected. 

Rejecting the null hypothesis means that a relationship does exist between a set of variables and the effect is statistically significant ( p > 0.05).

If the data collected from the random sample is not statistically significance , then the null hypothesis will be accepted, and the researchers can conclude that there is no relationship between the variables. 

You need to perform a statistical test on your data in order to evaluate how consistent it is with the null hypothesis. A p-value is one statistical measurement used to validate a hypothesis against observed data.

Calculating the p-value is a critical part of null-hypothesis significance testing because it quantifies how strongly the sample data contradicts the null hypothesis.

The level of statistical significance is often expressed as a  p  -value between 0 and 1. The smaller the p-value, the stronger the evidence that you should reject the null hypothesis.

Probability and statistical significance in ab testing. Statistical significance in a b experiments

Usually, a researcher uses a confidence level of 95% or 99% (p-value of 0.05 or 0.01) as general guidelines to decide if you should reject or keep the null.

When your p-value is less than or equal to your significance level, you reject the null hypothesis.

In other words, smaller p-values are taken as stronger evidence against the null hypothesis. Conversely, when the p-value is greater than your significance level, you fail to reject the null hypothesis.

In this case, the sample data provides insufficient data to conclude that the effect exists in the population.

Because you can never know with complete certainty whether there is an effect in the population, your inferences about a population will sometimes be incorrect.

When you incorrectly reject the null hypothesis, it’s called a type I error. When you incorrectly fail to reject it, it’s called a type II error.

Why Do We Never Accept The Null Hypothesis?

The reason we do not say “accept the null” is because we are always assuming the null hypothesis is true and then conducting a study to see if there is evidence against it. And, even if we don’t find evidence against it, a null hypothesis is not accepted.

A lack of evidence only means that you haven’t proven that something exists. It does not prove that something doesn’t exist. 

It is risky to conclude that the null hypothesis is true merely because we did not find evidence to reject it. It is always possible that researchers elsewhere have disproved the null hypothesis, so we cannot accept it as true, but instead, we state that we failed to reject the null. 

One can either reject the null hypothesis, or fail to reject it, but can never accept it.

Why Do We Use The Null Hypothesis?

We can never prove with 100% certainty that a hypothesis is true; We can only collect evidence that supports a theory. However, testing a hypothesis can set the stage for rejecting or accepting this hypothesis within a certain confidence level.

The null hypothesis is useful because it can tell us whether the results of our study are due to random chance or the manipulation of a variable (with a certain level of confidence).

A null hypothesis is rejected if the measured data is significantly unlikely to have occurred and a null hypothesis is accepted if the observed outcome is consistent with the position held by the null hypothesis.

Rejecting the null hypothesis sets the stage for further experimentation to see if a relationship between two variables exists. 

Hypothesis testing is a critical part of the scientific method as it helps decide whether the results of a research study support a particular theory about a given population. Hypothesis testing is a systematic way of backing up researchers’ predictions with statistical analysis.

It helps provide sufficient statistical evidence that either favors or rejects a certain hypothesis about the population parameter. 

Purpose of a Null Hypothesis 

  • The primary purpose of the null hypothesis is to disprove an assumption. 
  • Whether rejected or accepted, the null hypothesis can help further progress a theory in many scientific cases.
  • A null hypothesis can be used to ascertain how consistent the outcomes of multiple studies are.

Do you always need both a Null Hypothesis and an Alternative Hypothesis?

The null (H0) and alternative (Ha or H1) hypotheses are two competing claims that describe the effect of the independent variable on the dependent variable. They are mutually exclusive, which means that only one of the two hypotheses can be true. 

While the null hypothesis states that there is no effect in the population, an alternative hypothesis states that there is statistical significance between two variables. 

The goal of hypothesis testing is to make inferences about a population based on a sample. In order to undertake hypothesis testing, you must express your research hypothesis as a null and alternative hypothesis. Both hypotheses are required to cover every possible outcome of the study. 

What is the difference between a null hypothesis and an alternative hypothesis?

The alternative hypothesis is the complement to the null hypothesis. The null hypothesis states that there is no effect or no relationship between variables, while the alternative hypothesis claims that there is an effect or relationship in the population.

It is the claim that you expect or hope will be true. The null hypothesis and the alternative hypothesis are always mutually exclusive, meaning that only one can be true at a time.

What are some problems with the null hypothesis?

One major problem with the null hypothesis is that researchers typically will assume that accepting the null is a failure of the experiment. However, accepting or rejecting any hypothesis is a positive result. Even if the null is not refuted, the researchers will still learn something new.

Why can a null hypothesis not be accepted?

We can either reject or fail to reject a null hypothesis, but never accept it. If your test fails to detect an effect, this is not proof that the effect doesn’t exist. It just means that your sample did not have enough evidence to conclude that it exists.

We can’t accept a null hypothesis because a lack of evidence does not prove something that does not exist. Instead, we fail to reject it.

Failing to reject the null indicates that the sample did not provide sufficient enough evidence to conclude that an effect exists.

If the p-value is greater than the significance level, then you fail to reject the null hypothesis.

Is a null hypothesis directional or non-directional?

A hypothesis test can either contain an alternative directional hypothesis or a non-directional alternative hypothesis. A directional hypothesis is one that contains the less than (“<“) or greater than (“>”) sign.

A nondirectional hypothesis contains the not equal sign (“≠”).  However, a null hypothesis is neither directional nor non-directional.

A null hypothesis is a prediction that there will be no change, relationship, or difference between two variables.

The directional hypothesis or nondirectional hypothesis would then be considered alternative hypotheses to the null hypothesis.

Gill, J. (1999). The insignificance of null hypothesis significance testing.  Political research quarterly ,  52 (3), 647-674.

Krueger, J. (2001). Null hypothesis significance testing: On the survival of a flawed method.  American Psychologist ,  56 (1), 16.

Masson, M. E. (2011). A tutorial on a practical Bayesian alternative to null-hypothesis significance testing.  Behavior research methods ,  43 , 679-690.

Nickerson, R. S. (2000). Null hypothesis significance testing: a review of an old and continuing controversy.  Psychological methods ,  5 (2), 241.

Rozeboom, W. W. (1960). The fallacy of the null-hypothesis significance test.  Psychological bulletin ,  57 (5), 416.

Print Friendly, PDF & Email

Contributions of Community Education in the Eradication of Poverty among Communities in Rivers State...

Olori Christian N., Dimkpa Ezimma E., Olori Gloria I.

null hypothesis about poverty

  • Publication A-Z index
  • Browse by subject
  • Submit Manuscript

null hypothesis about poverty

  • EDUCATION Home
  • Special Issues
  • Conferences

null hypothesis about poverty

Contributions of Community Education in the Eradication of Poverty among Communities in Rivers State, Nigeria

null hypothesis about poverty

1 Department of Adult & Non-Formal Education, Federal College of Eduation (Technical) Omoku, Rivers State

2 Institute of Education, Rivers State University of Science & Technology, Port Harcourt, Rivers State

  • Related Content
  • About the Authors
  • Follow the Authors

This is a descriptive survey research study that sets to examine the extent community education has contributed to the eradication of poverty among communities in Rivers State of Nigeria. Two research questions and one null hypothesis were developed to guide the study. The population of the study was 1,672 respondents made up of 270 married men, 517 married women and 885 youths in there selected local government areas of the state. A sample of 428 respondents was drawn from the population using a purposive sampling technique. A-15 item structured questionnaire weighted on a 4-point rating scale was the data collecting instrument, face validated by three validates from the University of Nigeria, Nsukka. Descriptive statistics of means and inferential statistics of analysis of variance (ANOVA) were used for data analysis. Findings of the study revealed that the extent community education has contributed to the eradication of poverty among communities in Rivers State was low. Inhibiting factors associated with the low level of community education were ignorance, corruption of most of the community leaders, lack of political will and low level of education. There was no significant difference in the mean responses of respondents on the extent community education has contributed to the eradication of poverty among communities in Rivers State. It was therefore recommended among others that sensitisation on empowerment of community members and establishment of more adult education programmes in the state be made.

Keywords: contributions, community, community education, poverty

Cite this article:

  • Normal Style
  • Chicago Style
  • Olori Christian N., Dimkpa Ezimma E., Olori Gloria I.. Contributions of Community Education in the Eradication of Poverty among Communities in Rivers State, Nigeria. American Journal of Educational Research . Vol. 3, No. 10, 2015, pp 1279-1283. http://pubs.sciepub.com/education/3/10/11
  • N., Olori Christian, Dimkpa Ezimma E., and Olori Gloria I.. "Contributions of Community Education in the Eradication of Poverty among Communities in Rivers State, Nigeria." American Journal of Educational Research 3.10 (2015): 1279-1283.
  • N., O. C. , E., D. E. , & I., O. G. (2015). Contributions of Community Education in the Eradication of Poverty among Communities in Rivers State, Nigeria. American Journal of Educational Research , 3 (10), 1279-1283.
  • N., Olori Christian, Dimkpa Ezimma E., and Olori Gloria I.. "Contributions of Community Education in the Eradication of Poverty among Communities in Rivers State, Nigeria." American Journal of Educational Research 3, no. 10 (2015): 1279-1283.

1. Introduction

Education is generally acknowledged as a weapon for development by various scholars. It is a tool for the transformation and empowerment of citizens in a country. Nigeria as a developing country sees education as a catalyst for empowerment which provides communities with necessary skills required for sound social living. The Federal Republic of Nigeria (FRN) [ 7 ] submits that education is an instrument par excellence for effective national development. Development in this context encapsulates a purposive change in a society that contributes to the social, political, cultural and economic wellbeing of the people without creating any disharmony. The Niger Delta Development Commission (NDDC) [ 8 ] sees development as the only way to sustain alleviation of poverty, improve community services and living standards. This view of development is traced to the economic perspective. Suffice to say that development could be viewed from different perspectives. National development on the other hand entails all activities engaged by a nation for the enhancement of standard of living. Ugwu [ 13 ] explains that the main purpose of national development is the liberation of its citizens from the shackles of poverty.

Community education involves the use of community learning resources and research to bring about community change and recognition that people can learn through with and from each other to create a better world. Ezimah [ 6 ] added that it is a process aimed at raising consciousness, spreading understanding and providing the necessary skills, including the human and material resources, for the social, economic, political and cultural development of the community. The array of definitions of community education points to the fact that it is an organised learning activities directed to communities in attempt improve their standard of living. Interestingly, it is the education for peoples’ empowerment over their own lives in order to bring about transformation and change in individuals, communities, societies and nations. Anyanwu [ 4 ] noted that the philosophy of community education is tied up with the idea of enabling people to exploit their resources and use such to increase their competence and confidence in handling their own affairs. It is therefore a process and movement for the attainment of community growth.

Accordingly, the attainment of excellence as conceived by the Federal Republic of Nigeria requires diversity in the mode of training such as formal, informal and non-formal. Community education is an aspect of non-formal education. It is a programme of adult education geared at empowering members of a community for enhanced quality of life. Akande [ 3 ] sees community education as a type of education needed to ensure the self-confidence, self-respect and personal independence as well as to safeguard human rights to achieve social equality. It is a veritable tool for the stimulation of community members to actively participate in social activities that would generate employment, increase their incomes as well as improve their quality of life. The Canadian Association for Community Education [ 12 ] sees community education as a process whereby learning is used for individual, community and global betterment. This education they added is characterised by the integrated involvement of people of all ages.

Two schools of thought have emerged in the definition of community education as conceived by Anyanwu [ 4 ] . The first sees community education as a guide to community development activities. In this school, emphasis is placed on material improvement of communities and achievement of tangible goals. The second views community education in terms of building up communities. It involves the improvement of individuals in the community. Anyanwu concluded by stating the objectives of community education as follows; to

1. Educate and motivate the people for self-help.

2. Develop responsible leadership among the people.

3. Inculcate among the members of’ a community a sense of citizenship and a spirit of civic consciousness.

4. Introduce and strengthen democracy at the grassroots level, through the creation and/or revitalisation of institutions designed to serve as instrument of local participation.

5. Initiate a self-generative, self sustaining and enduring process of growth.

6. Establish and maintain co-operative and harmonious relationships in the community and to

7. Bring about gradual and self-chosen changes in the life of a community, with a minimum of stress and disruption.

Poverty is another concept in the study that requires an explanation. It is a phenomenon that is multi-faceted, multi-dimensional and multi-disciplinary. The term has been subjected to variety of interpretations by scholars in different areas of endeavour. The Central Bank of Nigeria (CBN) [ 5 ] classifies poverty to structural, economic, social, cultural and political deprivation. As a complex and multi-dimensional phenomenon, Adehavo [ 1 ] , explains that poverty goes beyond condition of lack of resources, but extends to social inequality, insecurity, illiteracy, poor health, restricted or total lack of opportunity for personal growth and self- realisation. As a human condition, Preece [ 10 ] describes poverty in terms of sustained or chronic deprivation of resources, capabilities, choices, security and power necessary for enjoyment of adequate standard of living and other civil, cultural, economic, political and social rights. The World Bank [ 14 ] posits that it is conceived as the inability of certain person to attain a minimum standard of living. Some other writers have conceptualised the term in terms of being absolute or relative. As absolute, Ogwumike and Ozughala [ 9 ] submit that it is characterised by inadequate health facilities, poor quality of education and low life expectancy. It is relative when a household possesses per capital of less than 1/3 of the average per capital income of the country concerned. Poverty from these wide range of meaning denotes the deprivation that incapacitates an individual or group of people to effectively and freely exercise their rights on issues of personal and collective concerns. This view corroborates with Adebayo’s [ 1 ] stipulation that poverty takes various forms. This includes lack of income and productive resources sufficient to ensure sustainable livelihoods, hunger, and malnutrition, ill-health, limited or no access to education, homelessness, inadequate housing, unsafe environments and social discrimination and exclusion. It also included restrictions on or lack of participation in the decision making process in civil, social and cultural life.

Prior to the emergence of western education in Rivers State, indigenous community education was practised. Sarumi [ 11 ] explains that this education was comprehensive, effective, work-oriented and less expensive. It was also community-based, where the entire community was seen as a teacher. Thus, functional, collective, communalistic and democratic. Adeyinka [ 2 ] further reports that boys and girls were given the kind of education that enabled them to fulfill masculine and feminine responsibilities in the community. Consequently, the philosophy that undermines community education as relevant to people’s needs, aspirations and locally-based was evident. However, with the emergence of western education in the state, there seems to be a wrong notion among people on what community education actually is and therefore wonders how it could serve as a catalyst for the eradication of poverty. Akande [ 3 ] rightly observed that some of the challenges facing community education are as follows:

•  Inadequate financial allocation attributable to lack of political will on the part of the government.

•  Considerable efforts not made to recruit professionals such as adult educators, community educators and social welfare officials.

•  The non-involvement of beneficiaries by community development programmers in the sharing of decision making for the effectiveness of the programme.

•  The presence of overlapping roles among the governmental and quasi- governmental agencies involved in the provision of community education.

•  Lack of effective collaboration among the majority of NGOs involved in the promotion of community education, and

•  Incessant and protracted communal, ethnic, regional and religions conflicts.

These problems emanate from the fact that community education is multi-dimensional, and differs in scope and interpretations by various writers. However, it is doubtful whether; community education has contributed to the eradication poverty in Rivers State. Informed by this, attempt is made by the researchers to assess the level of contribution of community education in the eradication of poverty in Rivers State.

From the background, scholars have identified community education as a non-formal education programme directed at improving the quality of life of the communities. In educating the people on the use of their own resources to improve their living condition, community education is seen playing a significant role in the reduction of poverty. However, despite the contribution of community education, empirical studies to ascertain the extent of contribution of this education in the eradication of poverty in Rivers State to the best knowledge of the researchers have not been conducted. The absence of these studies in the State has provided knowledge gap, which this present study intends to fill. The problem of this study therefore is to assess the extent community education has contributed to the eradication of poverty among communities in Rivers State.

The purpose of the study was to assess the extent community education has contributed to poverty eradication among communities in Rivers State. The objectives of the study are specifically to:

1. Determine the extent of community education has contributed to poverty eradication among communities in Rivers State.

2. Find out the inhibiting factors affecting community education from eradicating poverty among communities in Rivers State.

The following research questions are posed to guide the study

1. To what extent has community education contributed to poverty eradication among communities in Rivers State?

2. What are the inhibiting factors affecting community education from eradicating poverty among communities in Rivers State?

The null hypothesis of this was tested at .05 level of significance

Ho 1 : There is no significant difference in the mean responses of respondents regarding the extent community education has contributed to poverty eradication among communities in Rivers State.

2. Methodology

The descriptive survey research design was adopted for the study. This design aimed at assessing the opinion of respondents on the extent community education has contributed to the eradication of poverty among communities in Rivers State of Nigeria. The population for the study was 1,672 respondents comprising of 270 married men, 517 married women, and 885 youths from Ogba Egbema Ndoni, Emohua and Opobo/Nkoro Local Government Areas of Rivers Slate. The sample size for the study was 418 respondents (25% of the study population). Breakdown of the sample shows that there are 68 married men, 129 married women and 221 youths. Purposive sampling technique was employed in drawing the sample size.

Data collecting instrument for the study was a structured questionnaire. This was titled ‘Questionnaire on the Contributions of Community Education in the Eradication of Poverty’ (QCCEEP). It contained a-15 item questionnaire weighted on a-4 point scale. The QCCEEP was faced validated by three validates from the Department of Adult Education and Extra-Mural Studies of the University of Nigeria. Nsukka. The reliability coefficient index of 0.79 was obtained using the Cronbach Alpha after administering 25 copies of the questionnaire to 25 respondents from lkwerre and Bonny Local Government Areas with similar characteristics with the study area.

The administration of the questionnaire was done by the researchers to the respondents in their respective local government areas. Out of the 418 copies distributed, 397 copies were duly filled and returned, giving 95% returned rate. The data collected from the respondents were analysed using the mean for the research questions. The criterion mean of 2.50 was used to accept an item, below 2.50 was rejected. The level of contribution was classified as follows:

3.50 - 4.00 Very High Extent (VHE)

2.50 - 3.49 High Extent (HE)

2.00 - 2.49 Low Extent (LE)

1 .00 - 1 .99 Very Low Extent (VLE)

The analysis of variance (ANOVA) was used to test the null hypothesis at .05 level of significance. Significant difference was found if the calculated f-ratio is less than the critical f-value. This means that the null hypothesis was accepted. On the other hand, significant difference was found if the calculated f-ratio is greater than the critical f-value. This implies that the null hypothesis was rejected.

Data on Table 1 show very low extent for respondents in items 1(1.99, 1.95, 1.89). 3(1.65, 1.52, 1.84), and 6(1.74, 1.1.52, 1.79). Item 2 has mean scores of high extent for both married men and youths, but low (2.29) for married women. Item 4 has the mean score of low extent (2.10) for married men, and very low extent (1.69, 1.80) for both married women and youths. The mean scores of high extent for respondents were accorded to items 5 and 7. With the cluster mean of 2.25, the table indicated that the extent community education has contributed to poverty eradication in Rivers State was low.

Table 2 shows that results of all the items were rated above 2.50 as agreed, except the mean score of married women in item 13 which was 2.27. With the cluster mean of 2.94, the table revealed that several inhibiting factors affecting community education from eradicating poverty in Rivers State were identified. These factors inc1uded lack of political will, ignorance, corruption, non-involvement of’ direct beneficiaries in decision making, inadequate funding of projects, low educational attainment and poor maintenance culture.

Table 3 shows that the calculated f-value (.499) is not significant at .904 level of significance, which is not significant at .05 level of significance, hence, the acceptance of the null hypothesis. This means that there was no significant difference in the mean responses of respondents regarding the extent community education has contributed to the eradication of poverty among communities in Rivers State.

Table 1. Mean ratings of respondents on the extent community education has contributed to poverty eradication

null hypothesis about poverty

Table 2. Mean ratings of respondents on inhibiting factors affecting community education from the eradication of poverty

null hypothesis about poverty

Table 3. ANOVA summary for the mean responses of married men, married women and youths on the extent of contribution of community education in the eradicating of poverty

null hypothesis about poverty

4. Discussion

Findings in research question one showed that the extent to which community education has contributed to the eradication of poverty among communities in Rivers State was low. Supporting the findings, respondents indicated that there were high rate of insecurity and illiteracy level in the state. This is not unconnected with the fact that enlightened community members tend to make positive contributions in the development of their environment. This further corroborates with Akande’s [ 3 ] stipulation that community education is capable of ensuring self-confidence, self-respect, personal independence as well as safeguarding human rights for the attainment of social equality. Anyanwu [ 4 ] added that it introduces and strengthens democracy at the grassroots level.

In research question two, findings revealed various factors inhibiting community education from the eradication of poverty among communities in Rivers State. Some of these factors as indicated by respondents included lack of political by the administrators, ignorance of most community members, corruption of some community leaders and the non-involvement of direct beneficiaries in decision making. The findings further noted that inadequate funding of projects, non-involvement of professionals, low educational attainment and poor maintenance culture were also identified as contributing factors. These factors are attributable to the fact that scholars have given different interpretations on what community education denotes. It may therefore not be surprising that community education loses its importance in the course of the diverse views. This finding is supported by Akande [ 3 ] who identifies lack of political will as one of the challenges facing community education.

The ANOVA test for the null hypothesis indicated that there was no significant difference found in the mean responses of respondents on the extent community education has contributed to the eradication of poverty among communities in Rivers State. The existence of no significant difference further shows that the respondents have similar views regarding the extent community education has contributed to the eradication of poverty.

5. Implication of the Research Findings

Findings on the low contribution level of community education in the eradication of poverty is an indication for the need to re-orientate the people about the philosophy underlying community education using the services of professionals in adult education, community education as well as social workers.

The factors inhibiting community education from eradicating poverty call for leaders with good will to promote activities of community education.

6. Conclusion

The level of contribution of community education in the eradication of poverty among communities in Rivers State was low. This low level contribution of community education was associated with several factors. These inhibiting factors included lack of political will, ignorance, corruption of some of the leaders in the community, non-involvement of beneficiaries in decision making, non-involvement of professionals in community education affairs, low educational attainment of most community members and poor maintenance culture of existing projects.

7. Recommendations

Based on the findings of this study, the following recommendations were made:

1. The services of professionals in community education affairs should be employed in the organisation and co-ordination of community education activities.

2. Sensitisation on the importance of community education should be made through awareness campaign.

3. More adult education programmes should be established to accommodate varying needs of the communities.

4. Beneficiaries of projects should be encouraged to participate in decision making affecting their well being in the community.

5. Synergy between the government and community leaders should be intensified for better implementation of policies affecting the communities.

  • Full-Text PDF
  • Full-Text ePUB
  • Citation-(RIS Format)
  • Citation-(BibTeX Format)
  • Citation-(EndNote Format)

CiteULike

  • Google Scholar
  • VIRAL HEPATITIS CONGRESS
  • JournalTOCs

Help & Contacts

  • Questionnaire

Facebook

U.S. flag

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

  • Publications
  • Account settings
  • Advanced Search
  • Journal List
  • Healthcare (Basel)

Logo of healthcare

Out-of-Pocket Health Expenditure and Poverty: Evidence from a Dynamic Panel Threshold Analysis

Associated data.

The data presented in this study are openly available in WHO and World Bank databases.

The current study investigated the association between out-of-pocket health expenditure and poverty using macroeconomic data from a sample of 145 countries from 2000 to 2017. In particular, it was examined whether the relationship between out-of-pocket health expenditure and poverty was contingent on a certain threshold level of out-of-pocket health spending. The dynamic panel threshold method, which allows for the endogeneity of the threshold regressor (out-of-pocket health expenditure), was used. Three indicators were adopted as poverty measures, namely the poverty headcount ratio, the poverty gap index, and the poverty gap squared index. At the same time, out-of-pocket health expenditure was measured as a percentage of total health expenditure. The results showed the validity of the estimated threshold models, indicating that only beyond the turning point, which was about 29 percent, that out-of-pocket health spending led to increased poverty. When heterogeneity was controlled for in the sample, using the World Bank income classification, the findings showed variations in the estimated threshold, with higher values for the low- and lower-middle-income groups, as compared to the high-income group. For the lower-income groups, below the threshold for out-of-pocket health expenditure, it had a positive or insignificant effect on poverty reduction, while it led to higher poverty above the threshold. Further, the sampled countries were divided into regions, according to the World Health Organization. Generally, improving health care systems through tolerable levels of out-of-pocket health expenditure is an inevitable step toward better health coverage and poverty reduction in many developing countries.

1. Introduction

As stated in the World Health Organization’s sustainable development goal 3.8, reaching universal health coverage and financial risk protection are important indicators to guarantee better healthy lives and higher well-being. In developing countries, health care resources are inadequate to guarantee that all individuals have equal access to necessary health needs. According to World Health Statistics [ 1 ], out-of-pocket health expenses can create financial hardship by forcing people to choose between health expenses and other necessities. Many studies have described out-of-pocket expenditure on health as catastrophic when it surpasses a certain threshold of a household’s consumption or income. Moreover, the proportion of the world’s population, which spent more than 10% of its household income on medical care, increased from 9.4% in 2000 to 12.7% in 2015, amounting to about US$927 million. The percentage of the population spending more than 25% of their family budget on health care also increased from 1.7% in 2000 to about 3% in 2015. The majority, or 87%, of the population that suffered huge out-of-pocket expenditures in 2015 were in middle-income countries. About 1 billion, or 12.9% of the population, were expected to spend at least 10% of their family budget on medical care by 2020 [ 1 ]. Financing health spending through out-of-pocket expenses has several severe consequences, such as pushing individuals and households into poverty. Most of the population who are pushed into extreme poverty, as a result of out-of-pocket healthcare expenditures, are located in less-developed countries. However, between 2000 and 2015, the number of people falling into poverty due to out-of-pocket healthcare spending (defined as US$1.90 per person per day) dropped from 123.9 million (2%) to 89.7 million (1.2%). This decline was in line with the decline in the total population living in extreme poverty. Additionally, out-of-pocket expenditure on health was the main source of economic disadvantages, such as low-income level. In particular, the increase in the world’s relative poverty, as a result of out-of-pocket health expenditure, was 110.9 million people in 2000 and 183.2 million people in 2015. Therefore, achieving universal health coverage remains an important challenge for many countries around the world.

According to the World Health Organization and the World Bank [ 2 ], universal health coverage is defined as the extent to which public health services are designed to promote health, prevent disease, and provide high-quality treatment, rehabilitation, and palliative care sufficient to be effective while ensuring that the services that they provide will not expose the user to financial difficulties. The World Health Report [ 3 ] stated that to achieve universal health coverage, governments could take action in the following ways: allocate additional resources for health, lessen financial obstacles, and raise financial risk protection through pooling and prepayment; and ensure more equitable and efficient use of the available resources. Therefore, to have effective health systems, these systems must be efficiently financed by their host countries. In general, low-income countries have a higher share of private spending on health services than middle- and high-income countries, and most of the private health expenditures are out-of-pocket.

Considerable numbers of studies have been carried out to investigate the impact on the impoverishment level of populations caused by out-of-pocket health spending [ 4 , 5 , 6 , 7 , 8 ]. It has been observed that poorer populations tend to incur higher catastrophic health spending and seek no, or lower quality of health care than less-poor populations [ 9 , 10 ]. One major issue in earlier out-of-pocket healthcare spending–poverty research concerned the reliance on the variation between specific poverty measures before and after out-of-pocket health expenses were included in total household consumption; as a primary empirical approach, see Wagstaff et al. [ 8 ]. Therefore, the previous literature on the impoverishing effects of out-of-pocket health spending has depended predominantly on survey-based or micro-level data of a specific country [ 11 ]. To date, however, few studies have investigated the association between out-of-pocket health expenditure and poverty using aggregated macro-level data [ 11 ], which is useful for providing cross-country analysis. It has also been argued that some households may spend up to a certain threshold of their budget (10% or 25%) on health care without affecting the resources left over to sustain their basic needs [ 12 ]. So far, very little attention has been paid to the idea that the relationship between out-of-pocket health spending and poverty is contingent on a certain threshold level.

There are several important areas where this study makes an original contribution to the out-of-pocket health spending literature. The main purpose of this study is to investigate the relationship between out-of-pocket health expenditure and poverty from a macroeconomic perspective. In particular, the hypothesized out-of-pocket health spending–poverty threshold nexus for 145 developed and developing countries over the period 2000 to 2017 is empirically examined. This study relies on Seo and Shin’s [ 13 ] endogenous panel threshold method, which has been recently introduced for its estimation procedure. Unlike most previous studies, poverty is measured by using three indicators, namely poverty headcount, poverty gap, and poverty gap squared. Finally, potential heterogeneity is controlled for in the sample.

The remainder of this paper proceeds as follows: Section 2 reviews the related empirical literature. Section 3 provides a brief overview of the methodology and data used in this study. Section 4 contains the results and discussion. Section 5 offers a conclusion to the paper and offers some thoughts regarding future studies.

2. Literature Review

The level of out-of-pocket health spending is mainly driven by variations in socioeconomic status, such as the income, age, and education level of households [ 14 , 15 ]. To provide policy responses, Xu et al. [ 4 ] investigated the determinants of catastrophic health spending. Their study defined out-of-pocket health payments as catastrophic if the spending on health exceeded 40% of household income. They argued that huge variations between countries regarding the percentage of households facing catastrophic out-of-pocket payments existed. Importantly, the coverage of health care services that required payment, the low ability to pay, and the nonexistence of health insurance were found to be among the essential factors that induced catastrophic health spending. Out-of-pocket payments were considered to be the major health care financing style in low-income countries and were mostly larger than government expenditure. The level of out-of-pocket expenditure tended to decline as income rose and when other forms of financing mechanisms increased [ 16 ]. Out-of-pocket healthcare payments are largely agreed to be a degrading, humble, and unsustainable manner to finance health care. The key determinants of catastrophic out-of-pocket health expenditure are poor economic conditions and low living standards [ 17 , 18 ]. Moreover, being part of a poor household can be one reason to pay for health care directly or out-of-pocket [ 7 , 8 ]. Excessive out-of-pocket payments may hinder access to health care, since they stop people from seeking necessary medical assistance and also deter them from receiving appropriate treatment [ 10 , 19 ].

Moreover, out-of-pocket expenditure on health care puts a huge burden on households and reduces their overall welfare [ 19 ], and Brown et al. [ 10 ] found that poorer households were less likely to experience catastrophic out-of-pocket payments since they avoided or delayed required health care. However, different evidence, as provided by Seeberg et al. [ 9 ], has suggested that low-income households sought medical care from less qualified providers, and, as a result, they faced catastrophic out-of-pocket spending. Another study by Ku et al. [ 20 ], conducted in Taiwan, argued that the implementation of National Health Insurance in the year 1995 contributed to reducing out-of-pocket health spending, especially for low-income households.

A large and growing body of literature has investigated the relationship between out-of-pocket health expenditure and poverty. Generally, out-of-pocket health spending was reported as a major reason for household impoverishment [ 5 , 6 , 21 , 22 ]. Poor health conditions were among the factors that contributed effectively to higher poverty rates, especially in developing countries such as Ghana [ 23 ]. Arsenijevic et al. [ 24 ] revealed that out-of-pocket health care payments were catastrophic for poor households and a leading cause of poverty in Serbia. Additionally, Koch et al. [ 22 ] showed that about 1% of households were pushed into poverty due to out-of-pocket health expenditure in Chile. Furthermore, Datta et al. [ 25 ] investigated the implications of out-of-pocket health expenditure, financial stress, and households’ impoverishment with non-communicable diseases. They found that non-communicable diseases induced higher medical expenditure, a greater probability of catastrophic out-of-pocket health spending, more financial stress, and a higher risk of impoverishment.

A study by Rashad and Sharaf [ 26 ] attempted to provide a new estimate for poverty that considered catastrophic out-of-pocket health expenditure in Egypt. Their approach relied on comparing the poverty gap and poverty line, both before and after the inclusion of catastrophic expenditures. They found that out-of-pocket payments led to 7.4% more households being pushed into poverty. Additionally, out-of-pocket health spending was strongly related to both the problems of inefficient and unequal access to health care; for example, low-income individuals who were not covered by social health insurance were more likely to have catastrophic out-of-pocket health spending than higher-income individuals covered by social health insurance [ 27 ]. Moreover, Van Minh et al. [ 7 ] noted that many poorer households faced catastrophic out-of-pocket health spending, and others were pushed into poverty. They found a modest impact from national health insurance on reducing impoverishment since it did not provide enough financial protection to households. Therefore, reducing the large share of out-of-pocket payments in health systems is a universal target for health system development. Eliminating out-of-pocket payments seems to be quite difficult, even within a health system offering free care. Kumara and Samaratunge [ 28 ] showed that free provision of health care did not stop the rising trend of out-of-pocket medical spending. However, as countries have accomplished high health coverage with well-organized health financing systems, they have found some significance in keeping certain out-of-pocket payments as incentives to have efficient health care performance. This may support the idea of the threshold effects of out-of-pocket health expenditure on poverty.

The existing literature on out-of-pocket health expenditure–poverty nexus has focused extensively on studying the issue in a single country setting, typically using survey data. Wagstaff et al. [ 8 ] and Wagstaff et al. [ 11 ] were among the few studies that used multi-country data and macroeconomic and health system indicators to investigate out-of-pocket health expenditure. Both of these studies indicated the existence of a large variation in out-of-pocket health spending across countries. Wagstaff et al. [ 11 ] revealed that out-of-pocket health expenditure at the USD 1.90 per person per day poverty line led to poverty, particularly in low-income countries. All of the studies reviewed so far have supported the hypothesis that the impoverishment impact of out-of-pocket health expenditure is asymmetric concerning certain factors such as the share of out-of-pocket health expenditure to the household’s budget or income. Yet, investigating the threshold effect of out-of-pocket health expenditure on poverty using cross-country and health system indicators remains a clear gap in the existing literature.

3. Methodology and Data

3.1. model specification and estimation technique.

To examine the relationship between out-of-pocket health expenditure and poverty, the following model is specified:

where P o v i t refers to poverty, as measured by the poverty headcount, the poverty gap, and the poverty gap squared; O O P i t is out-of-pocket health spending, as a percentage of total health expenditure; Z i t is a vector of the explanatory variables; G D P i t is the real income per capita; G H E i t is government health expenditure, as a percentage of total health expenditure; ε i t is an error term; i = 1, …, N denotes the country (N = 145 countries); and t = 1, …, T denotes the time, which is between 2000 and 2017.

As suggested in the existing literature, out-of-pocket health expenditure should be treated as an endogenous variable in its relationship with poverty. Out-of-pocket health payments can lead to higher population poverty, and at the same time, poverty may lead to increased financial risk and out-of-pocket health expenses. Therefore, it is critically important to control for the endogeneity of out-of-pocket health expenditure. Various threshold-estimating procedures can be used, such as Hansen’s [ 29 ] static panel method and Kremer’s et al. [ 30 ] dynamic panel technique. However, the methods mentioned above assume the exogeneity of the threshold regressor. Seo and Shin [ 13 ] suggested an alternative estimation procedure, which allows for the endogeneity of both regressors and the threshold variable. To estimate the non-linear relationship between out-of-pocket health expenditure and poverty in the sample of 145 countries ( Table A3 in Appendix A shows the list of countries), the dynamic panel threshold technique was used, as suggested by Seo and Shin [ 13 ]. For that, let us consider the following threshold model:

where u i is the individual-specific effect, and ε i t is the error term, which is assumed to be ε i t   ~ (0, σ 2 ). The indicator function I ( ⋅ ) indicates the regime or group, according to the threshold variable O O P i t , and γ denotes the impact of out-of-pocket health expenditure, depending on whether O O P i t is below or above the threshold level. Z i t contains a vector of the control variables, which are specified above. The impact of out-of-pocket health expenditure on poverty can be explained by β ^ 1 ( β ^ 2 ) which denote the marginal effect of out-of-pocket health expenditure on poverty in the low (high) out-of-pocket health expenditure regime, i.e., when out-of-pocket health expenditure is below (above) the threshold. Normally, out-of-pocket health expenditure is relatively high in countries where health care coverage is low. Note that all the variables are transformed into natural logarithms as the coefficients are easier to interpret, and the data will most likely follow a normal distribution.

Equation (2) is estimated using the method of Seo and Shin [ 13 ], which allows for an endogenous threshold variable and regressors and uses the generalized method of moments (GMM) estimation technique as proposed by Arellano and Bond [ 31 ]. This technique comprises two steps: first, for a given threshold ( γ ), the coefficients ( ρ , β 1 , β 2 , π i ) are estimated using the GMM estimator, as proposed by Arellano and Bond [ 31 ]. Second, the first step is repeated for the value of the threshold’s belonging in a strict subset of out-of-pocket health expenditure support, resulting in different estimates for each selected threshold. The threshold value ( γ ), which minimizes the objective function of the GMM estimator and its estimated parameters, is deemed to be the optimal threshold.

Parallel to other threshold estimating procedures, such as the Hansen [ 29 ] static method and the dynamic method of Kremer et al. [ 30 ], the Seo and Shin [ 13 ] technique has the advantage of allowing for the endogenous estimation of the threshold variable and any other regressors, such as the GDP per capita. This is applicable, in a practical sense, in this study because a higher poverty level may lead to greater out-of-pocket spending on health care due to the vulnerability and inadequacy of health coverage. To fit the properties of the GMM estimator, Equation (2) is estimated using three years of averaged data, namely from 2000 to 2017.

3.2. The Data

The World Bank frequently updates its definition of the poverty line as basic food, clothing, and housing costs change around the world. In 2015, the World Bank set the poverty line to be USD 1.90 per person per day, rather than USD 1.25 per person per day. The present study used three poverty measures based on the poverty line being set at USD 1.90 per person per day. First, the poverty headcount refers to the percentage of the population living below the national poverty line(s). Second, the poverty gap index of the World Bank is used. The poverty gap denotes the proportion by which the poor’s average income level dropped below the poverty line. Generally, a higher poverty gap indicates the increased severity of poverty in a given country. Finally, similarly to the second measure, the poverty gap squared index represents the extent of disparity among the poorer element of the population. The poverty headcount ratio, the poverty gap index, and the poverty gap squared index were retrieved from the World Bank PovcalNet [ 32 ].

Out-of-pocket health expenditure refers to individuals’ direct expenses to health care providers, excluding any prepayments for health services, such as taxes, insurance premiums, or contributions. Out-of-pocket health payments are part of health financing in all countries that rely on user fees and/or co-payments to rationalize the use of health services, advance health system efficiency, and improve the quality of their services [ 33 ]. Therefore, it is expected that a reasonable level of out-of-pocket health expenditure would be tolerable or insignificant to poverty reduction. Nonetheless, higher out-of-pocket spending, as a mean of health financing, is harmful to the eradication of poverty. Additionally, it can be argued that out-of-pocket health spending could lead to the impoverishment of households, while poverty of the population, in turn, could cause higher financial risk and out-of-pocket spending. The data for out-of-pocket health expenditure, as a percentage of total health expenditure, was obtained from the World Health Organization database [ 34 ]. Furthermore, the standard of living is an important factor that affects a nation’s poverty level; thereby, a higher quality of life is expected to reduce poverty. The growth of income per capita is one of the main sources for poverty reduction in many countries around the globe, since without an increase in economic growth, eradication of poverty remains a difficult task to be achieved [ 35 , 36 , 37 , 38 , 39 ]. Additionally, Škare and Družeta [ 40 ] argue that as economic growth takes place, poverty is reduced regardless of the level of income inequality. In this study, the real GDP per capita represented the statistical measure of the standard of living. The GDP per capita (constant USD 2010) was collected from the World Bank, World Development Indicators [ 41 ]. Lastly, the control variable of government health expenditure was measured as a share of general government expenditure, which may negatively or positively affect poverty. In general, the government is expected to affect poverty levels through its spending on important sectors such as education and health [ 42 , 43 , 44 , 45 ]. Thus, taking the impact of government expenditure on poverty reduction into consideration is important for modelling purposes. The data for government health expenditure were found in the World Health Organization database [ 34 ].

Table 1 presents the results of the descriptive statistics and correlation analysis. It is apparent from the table that the number of observations was around 870 for all the variables. What is interesting about the data in this table is that the poverty measures have large variations, as compared to the other variables, as they have the highest standard deviations. However, out-of-pocket health expenditure has a relatively low variation. These descriptive figures suggest that there are huge differences in poverty across nations, where variations concerning out-of-pocket health spending across countries exist; however, it remained comparatively lower. The results of the correlation matrix are summarized at the bottom of Table 1 . Closer inspection of the correlation coefficients shows that there is high positive linear dependency among the three measures of poverty. Most importantly, the correlation among the explanatory variables remains low, which shows that the potential of collinearity is not of concern.

Descriptive statistics and correlation.

Note: obs, St. Dev., Min, and Max, denote observation, standard deviation, minimum, and maximum, respectively. PHC, PG, and PG 2 refer to poverty headcount, poverty gap, and poverty gap squared, respectively. All the variables are expressed in logarithmic form.

Normally, in using time series variables, testing the stationarity properties of the data is a common pretest. The results of the Im et al. [ 46 ] (IPS) unit root test, shown in Appendix A   Table A1 , revealed the variables to be stationary after taking the first-difference. Fortunately, the dynamic panel threshold used by our study relies on the Arellano and Bond [ 31 ] GMM estimator that runs the regression in first-difference.

4. Results and Discussion

Regression of the dynamic panel threshold analysis was used to predict the effect of out-of-pocket health expenditure on poverty. To assess the transition effect of out-of-pocket health spending on poverty, three aggregate indicators were used as poverty measures, namely poverty headcount, the poverty gap, and the poverty gap squared. Table 2 shows the results obtained from the analysis using the Seo and Shin [ 13 ] model.

Dynamic panel threshold results—all countries.

Note: ***, ** denote 1%, 5%, respectively. Between ( ) are robust standard errors. N refers to the number of countries.

When poverty headcount was used as the dependent variable, the results showed that the threshold of out-of-pocket health expenditure was about (ln3.378), or 29.3 percent, of total health expenditure, indicating that around 56 percent of the observations lay in the high out-of-pocket regime. Below the threshold, the coefficient β ^ 1 was negative and insignificant, indicating that out-of-pocket health expenditure did not influence poverty. However, above the threshold β ^ 2 was positive and statistically non-zero, which implied that any increase in out-of-pocket health spending would lead to more poverty. These findings were in line with the expectation that out-of-pocket health spending beyond a certain threshold would contribute to greater poverty. The lagged-dependent variable of poverty headcount was positive and statistically different from zero, signifying some persistence in poverty across countries. The results revealed a negative and significant relationship between the GDP per capita and the poverty headcount. This indicated that with continuous increases in the level of income, poverty was further reduced. There was no evidence that government health expenditure influenced poverty headcount.

When the poverty gap was used as the dependent variable, the estimated out-of-pocket threshold parameter was about (ln3.390) or 29.7 percent of total health expenditure, with approximately 55 percent of the observations in the high-out-of-pocket regime. The findings revealed a negative and statistically insignificant relationship between out-of-pocket health spending and the poverty gap below the threshold. In contrast, a positive and significant relationship was found above the threshold. These outcomes suggested that with higher out-of-pocket health expenditure that, beyond a certain level, poverty would increase. Moreover, the coefficient of the lagged-dependent variable of the poverty gap was shown to be positive and significant. There was a significant negative and significant association between the income per capita and the poverty gap, which suggested the importance of improving income in eliminating poverty. However, it was found that government health expenditure had no evident effect on the poverty gap.

Turning to the results of the poverty gap squared, when used as a dependent variable, the findings showed that the threshold of out-of-pocket health expenditure was about (ln3.390), or 29.7 percent, which indicated that about 55 percent of the observations were above the threshold of out-of-pocket health spending. It was found that β ^ 1 had no significant impact on the poverty gap squared, while β ^ 2 appeared to be positive and statistically significant. These findings implied that any additional out-of-pocket health spending beyond the threshold point led to an increased poverty gap squared. The lagged-dependent variable’s dynamic term was shown to be positively related to its current poverty gap squared. Moreover, it was found that the GDP per capita coefficient was negative and significantly associated with the poverty gap squared. A possible implication was that countries with a higher level of income were more likely to reduce poverty. Nevertheless, it was found that there was no change in the poverty gap squared associated with changes in government expenditure on health.

To summarize, above the threshold point, higher out-of-pocket health expenditure led to increased poverty. In addition, GDP per capita was an important factor that resulted in higher poverty reduction. At the same time, government health expenditure had no significant impact on poverty. To check the soundness of the estimated poverty threshold models, the results of the linearity test that had the null of no threshold effects were conducted. For the poverty headcount and poverty gap, it was found that the bootstrap p -values were 0.06 and 0.04, respectively, suggesting the existence of threshold effects. However, the bootstrap p -value for the poverty gap squared model was 0.2, indicating a non-threshold effect.

To control for variation across countries, the World Bank’s income classifications were relied upon, as well as the use of a group dummy. Table 3 shows the dynamic panel results with the endogenous threshold regressor controlling for low-income and lower-middle-income countries. When poverty headcount was the dependent variable, the estimated out-of-pocket health spending thresholds for low- and lower-middle-income countries were (ln3.292), or 26.9 percent, and (ln3.877), or 48.3 percent, of total health expenditure, respectively. These thresholds indicated that approximately 83 percent of the observations were in the high-out-of-pocket regime, in the case of low-income countries, whereas about 41.5 percent of the observations were in the upper regime of out-of-pocket health expenditure in lower-middle-income countries. Below the thresholds, it was found that out-of-pocket health expenditure had negative and significant effects on poverty headcount. However, beyond the thresholds, out-of-pocket health expenditure had positive and statistically meaningful effects on poverty. Earlier findings indicated that financing health expenditure through the out-of-pocket mode helped to reduce poverty initially. However, with further spending, the effect became a poverty-increasing factor. The results revealed that the dynamic term of poverty headcount was positive and statistically significant in both the sampled low and lower-middle-income countries. Importantly, negative and significant income per capita effects on poverty headcount were found in both income groups. This particular result suggested that improvement in the standard of living was an influential factor contributing to poverty reduction in the sampled less-developed countries, while there was evidence of a negative and positive relationship between government expenditure on health and poverty headcount in the sampled low- and lower-middle-income countries, respectively, suggesting that it led to reduced poverty in low-income and higher poverty as observed in the lower-middle-income group, as a result of government health spending.

Dynamic panel threshold results—controlling for income differences.

When poverty was measured using the poverty gap, it was found that the out-of-pocket thresholds were (ln2.282), or 9.8 percent, and (ln3.679), or 39.6 percent, of total health expenditure, for the sampled low- and lower-middle-income countries, respectively. These turning points showed that nearly 96 percent of the observations were in the upper regime of out-of-pocket health spending for the sampled low-income countries. For the sampled lower-middle-income countries, around 59 percent of the observations were in the upper regime of out-of-pocket health spending. The results showed that the effect of out-of-pocket health spending on the poverty gap, below the threshold, was negative and significant for both low- and lower-middle-income countries. However, above the threshold, out-of-pocket health spending had positive and statistically meaningful effects on poverty, with a higher magnitude in the lower-middle-income group. The previous results suggested that out-of-pocket health expenditure lessened poverty initially but increased poverty eventually. Moreover, the results revealed that the lagged-dependent variable of the poverty gap was positive and significant in both the sampled low- and lower-middle-income countries. Notably, a negative and significant relationship was found between per capita income and the poverty gap in both income groups, suggesting that a higher standard of living contributed effectively to poverty reduction in the sampled less-developed countries. Lastly, the relationship between government health expenditure and the poverty gap was negative in the sample low-income and positive in the sampled lower-middle-income countries, but it was insignificant in the sampled lower-middle-income countries.

When the poverty gap squared was the dependent variable, the results showed that the thresholds of out-of-pocket health expenditure for both groups were (ln3.218), or 24.9 percent, and (ln3.877), or 48.3 percent, which indicated that about 83 percent and 41.5 percent of the observations were above the threshold of out-of-pocket health spending for both the sampled low- and lower-middle-income countries, respectively. Specifically, it was found that β ^ 1 had a negative and insignificant effect on the poverty gap squared in the low-income group, but it had a significant negative effect on poverty in the sampled lower-middle-income countries. Nevertheless, it was found that β ^ 2 had a positive and statistically significant effect on the poverty gap squared in both of the sampled groups of countries. These findings denoted that further out-of-pocket health expenditure, beyond the threshold, led to a higher poverty gap squared in both groups. The dynamic term of the poverty gap squared was positive and significantly related to its current level in both samples. Additionally, it was found that the coefficient of per capita income was negative and significantly connected to the poverty gap squared in both the sampled low- and lower-middle-income countries, implying that countries with a higher level of income were more likely to have lower poverty rates. However, it was found that government health expenditure was negatively related to the poverty gap squared in the sampled low-income countries. In contrast, it had no significant effect on poverty in the lower-middle-income group.

Beyond the threshold point, a greater level of out-of-pocket health expenditure led to a higher poverty rate, while the effect of income per capita was negatively related to different poverty measures. However, the effect of government health expenditure on poverty was ambiguous. The bottom of Table 3 shows the linearity test outcomes that had the null hypothesis of no threshold. For the sampled low- and lower-middle-income countries, it was found that all of the bootstrap p -values for the three models of poverty were less than 0.05, which suggested the validity of the threshold effects of out-of-pocket health spending on poverty.

Table 4 demonstrates the results obtained from the use of Seo and Shin’s [ 13 ] dynamic panel threshold analysis, controlling for both upper-middle-income and high-income countries. The results revealed that the estimated out-of-pocket health spending threshold was (ln3.750), or 42.5 percent, of total health expenditure for the three poverty models in the sampled upper-middle-income countries. The threshold indicated that nearly 27 percent of the observations were in the upper out-of-pocket regime in the sampled upper-middle-income countries. As shown in Table 4 , the outcomes indicated that out-of-pocket health spending had positive and significant effects on poverty below the threshold. However, above the threshold, out-of-pocket health spending had no significant effects on poverty in the sampled upper-middle-income countries. The former results showed that out-of-pocket health expenditure led to increased poverty initially. However, beyond a certain level, poverty became irresponsive to changes in out-of-pocket health spending in the sampled upper-middle-income countries. The reported results show that the lagged-dependent variable of poverty was positive and statistically significant in two out of the three models in the upper-middle-income sample. Notably, the relationship between the income per capita and poverty was negative and significant, indicating that a higher standard of living was a prominent element that lessened poverty in the sampled upper-middle-income nations. However, there was a positive relationship between government spending on health and poverty in the two models, suggesting that a higher level of poverty might result from higher government spending on health.

Dynamic panel threshold results—controlling for income differences (continue).

Table 4 shows the results of the estimated poverty models for the sampled high-income countries. The out-of-pocket threshold values were (ln2.218), or 9.2 percent, (ln3.106), or 22.3 percent, and (ln3.074), or 21.6 percent, of total health expenditure, for the three models, respectively. The threshold points illustrated that about 98.7 percent, 41.2 percent, and 43 percent of the observations were in the upper regime of out-of-pocket health spending for the three measures of poverty, respectively. (It is worth mentioning that the thresholds appeared to be invalid due to the failure of rejecting the null hypothesis of no threshold in the estimated models for the sampled high-income countries.) The findings indicated a positive relationship between out-of-pocket health expenditure and the three measures of poverty below the thresholds. The relationship between the two variables was inconsistent since it was mostly insignificant from a statistical point of view. The preceding effects proposed that greater out-of-pocket expenditure on health increased poverty in relatively developed nations. A positive and significant lagged-dependent variable was found for the poverty models in high-income countries. Moreover, there was a negative but mostly insignificant relationship between per capita income and poverty in two out of the three estimated models, which suggested that the higher standard of living in developed nations contributed less to poverty reduction. Nonetheless, a positive relationship between government health expenditure and the poverty headcount and poverty gap models was found, whereas the effect was insignificant in the poverty gap squared model.

To ensure the accuracy of the estimated threshold models for the upper-middle-income and high-income groups, the results of the linearity test are shown in Table 4 . For the sampled upper-middle-income countries, the bootstrap p -values were 0.04, 0.0, and 0.08, for the poverty headcount, poverty gap, and poverty gap squared models, respectively, suggesting the validity of the threshold effects between out-of-pocket health spending and poverty. However, for the sampled high-income countries, the bootstrap p -values for all of the models were greater than 0.1, indicating the nonexistence of a threshold effect between out-of-pocket health spending and poverty in the high-income group.

An additional analysis was conducted to control for heterogeneity across the sampled countries, using the World Health Organization’s regional classifications as the criteria. A dummy variable for each of the WHO’s six regions was used, namely Africa, the Americas, Eastern Mediterranean, Europe, South-East Asia, and Western Pacific. Table 5 and Table 6 present the threshold regression analysis results after controlling for the six regions. Starting with the African region, it was found that the estimated thresholds were between (ln3.694), or 40.2 percent, and (ln3.851), or 48.04 percent. From the results in Table 5 , it was apparent that out-of-pocket health expenditure had a positive and statistically significant effect on poverty, both below and above the turning points, which suggested that out-of-pocket health expenditure led to increased poverty in many African countries. Regarding the results for the control variables, it was found that the lagged-dependent variable and government expenditure were positively related to poverty. However, income per capita was negative and significantly affected poverty in the African states, pointing toward the importance of economic growth and improving the population’s welfare in reducing poverty.

Dynamic panel threshold results—controlling for WHO regions.

Note: ***, **, * denote 1%, 5%, 10%, respectively. Between ( ) are robust standard errors. N refers to the number of countries.

Dynamic panel threshold results—controlling for WHO regions (continue).

Turning to the empirical evidence for the Americas region, the results indicated that the threshold’s values were (ln2.487), or 12 percent, and (ln2.662), or 14.3 percent. From Table 5 , it can be observed that there was evidence of non-linear associations between out-of-pocket health expenditure and poverty. Below the threshold, it was found that β ^ 1 was negative and significantly associated with the poverty gap and poverty gap squared, whereas β ^ 2 appeared to be positive and statistically different from zero. Looking at the results for the other independent variables, a similar conclusion was found to the findings of earlier models. When the same poverty models were estimated for the Eastern Mediterranean region, it was observed that the effect of out-of-pocket health spending on poverty was statistically insignificant for both β ^ 1 and β ^ 2 , although the three models rejected the null hypothesis of no threshold effect. Moreover, most of the explanatory variables were not different from zero.

Regarding the results for the European region, as shown in Table 6 , the different values of the threshold ranged from (ln2.331), or 10.3 percent, to (ln3.444), or 31.3 percent, of total health expenditure. However, the null hypothesis of no threshold effect could not be rejected for the three models. The outcomes revealed that, below the threshold, there was a positive relationship between out-of-pocket health expenditure and poverty. However, the coefficients above the threshold levels were inconsistent since they appeared to be negatively related to poverty headcount and poverty gap squared but positively correlated to the poverty gap. In contrast, the findings of the lagged-dependent variable and government expenditure were found to be positive and statistically significant, whereas the GDP per capita had a negative and significant effect on poverty, which supported the essential role played by higher income levels on poverty reduction in the European region.

The results for the South-East Asian region are presented in Table 6 . The estimated threshold values were (ln2.661), or 14.3 percent, and (ln3.028), or 20.7 percent, of total health expenditure. The results showed an insignificant relationship between out-of-pocket health expenditure and poverty, both below and above the thresholds. However, the upper regime coefficients were positive, showing potential inverse effects of out-of-pocket health expenditure on poverty. The poverty level in the region showed some persistence due to the significance of lagged-poverty. However, the other explanatory variables were found to be insignificantly associated with poverty in the South-East Asian region.

In the final part of the analysis, the threshold model results for the Western Pacific region are shown in Table 6 . It is interesting to note that the findings for this region revealed the lowest threshold levels, as compared to the models for the other regions. These thresholds ranged from (ln1.024), or 2.8 percent, to (ln2.121), or 8.3 percent. Unsurprisingly, therefore, the coefficients in the lower regime of out-of-pocket spending were positively related to various poverty measures, while the coefficients in the upper regime had a negative sign. This result indicated that higher out-of-pocket health expenditure both increased poverty below the threshold and reduced poverty above the threshold. The previous result may have been due to the low level of out-of-pocket health expenditure in the Western Pacific region. The dynamic terms of poverty appeared to be significant in the two models. Moreover, the impact of income per capita on poverty was negative, which suggested a positive association between the income level and poverty reduction. Regarding government expenditure on health, there was slight evidence that it harmed the poverty headcount.

Table A2 in Appendix A shows the results of the robustness analysis after we included additional variables such as corruption index and income inequality. The main findings regarding the threshold effect of out-of-pocket health spending on poverty remained similar to the earlier results.

A strong relationship between out-of-pocket health payments and poverty has been reported in the previous literature. Similarly, this study found a significant association between these variables. However, the results provided more evidence of a threshold impact of out-of-pocket health spending on poverty. A moderate level of out-of-pocket health expenditure can be beneficial to poverty reduction, whereas beyond a threshold, out-of-pocket health expenditure fuels poverty. Previous studies have highlighted this issue, mainly in the context of a single country, except for Wagstaff et al. [ 8 ] and Wagstaff et al. [ 11 ]. Nevertheless, the present study has contributed to the existing literature in that regard by using macroeconomic data from multiple countries.

5. Conclusions

Reducing out-of-pocket spending on health is not only essential to protect individuals and households against any financial risk, but it is an inevitable element to eradicate poverty. The present study was designed to determine the effects of out-of-pocket health expenditure on poverty, using cross-country macroeconomic data. The use of various poverty measures in this study was also an additional distinguishing feature, as compared to the reviewed previous studies. Specifically, the threshold relationship between out-of-pocket health spending and poverty was tested on data from 145 countries, using an appropriate econometric technique. In addition, any potential heterogeneity regarding out-of-pocket health expenditure and poverty was controlled for by running different regressions, according to the World Bank’s income groups and the World Health Organization’s regions.

The findings obtained by the dynamic panel threshold regression revealed the existence of a threshold effect of out-of-pocket health spending on poverty. Below the threshold of approximately 29%, out-of-pocket health expenditure made no significant difference to the level of poverty. However, further increases in out-of-pocket health expenditure, as a percentage of total health expenditure, led to greater poverty above the threshold. Beyond the threshold mentioned above, a 1% increase in out-of-pocket health expenditure, as a percentage of total health expenditure, led to around 1.8%, 2%, and 2% increases in the poverty headcount, the poverty gap index, and the poverty gap squared index, respectively. These results supported the idea that in countries where out-of-pocket health expenditure is higher than the threshold, which is mainly in the less-developed world, individuals are pushed into poverty due to higher health financial risk.

When income differences across countries were taken into consideration, this study’s results showed relatively higher turning points in the sampled low- and middle-income countries rather than in the sampled high-income nations. In particular, in low- and lower-middle-income countries, the effects of out-of-pocket health spending helped to reduce various poverty measures. Above the thresholds, however, out-of-pocket health expenditure increased poverty. In the case of the sampled upper-middle- and high-income states, out-of-pocket health expenditure appeared to be harmful to poverty below the thresholds, while it was insignificant beyond the thresholds. Note that the results of the upper regime in the upper-middle-income groups showed relatively fewer observations (countries) in the higher out-of-pocket-regime, whereas the threshold effects were invalid in the high-income sample. The principal empirical implication of these outcomes is that in countries where the share of out-of-pocket health expenditure to total health expenditure is comparatively high, poorer household health most likely would lead to higher poverty.

To investigate the relationship between out-of-pocket health expenditure and poverty, the sampled countries were divided according to the World Health Organization’s regions. The results revealed that in four out of the six regions, the effects of out-of-pocket health spending on poverty were contingent on certain threshold levels. This showed that below the threshold, out-of-pocket health spending was negatively related to poverty, while it was positively correlated to poverty above the threshold, although the coefficients in two of the regions were not significant. The findings showed relatively low turning points regarding the European and Western Pacific regions, and out-of-pocket health expenditure was positive and negative, both below and above the threshold, respectively. In the European region, there were no threshold effects detected. However, out-of-pocket health expenditure seemed to positively impact poverty reduction, especially above the threshold in the Western Pacific region.

One of the most significant findings to emerge from this study was that the lagged-dependent variables positively influenced various poverty measures, indicating that poverty was affected by its previous level. Overall, this finding strengthened the idea that poorer households are most likely to remain in poverty compared to non-poor households. The current findings support the relevance of the living standard to poverty reduction. These findings have significant implications for understanding how the improvement of income per capita is translated to poverty reduction, while low income causes more poverty. Taken together, these findings provide support for the concept of the poverty trap. In particular, it is most likely that poorer nations with fairly low standards of living will remain in extreme poverty, on the one hand. On the other hand, improved living standards for poor countries may help them to escape the poverty trap [ 47 ]. Regarding the effect of health expenditure by governments on poverty, the findings of this study were uncertain.

The findings of this study have some important implications for future practice. The huge variations in poverty levels across the globe necessitate the significance of reprioritizing economic policies to make poverty and income inequality reduction the top item on each nation’s economic agenda. Moreover, there is an urgent need for fast structural reforms for health care systems, especially in the post-COVID-19 era. In particular, governments are responsible for eliminating any financial risk associated with health uncertainty, thus reducing excessive out-of-pocket health expenditure. Greater investment in health-related infrastructure will increase access to health care and reduce the burden of catastrophic out-of-pocket expenditure for populations, which may help to eliminate any poverty associated with financial hardship and out-of-pocket health payments. For future studies, alternative poverty lines, other than the level of USD 1.90 per person per day, might be used for comparison, since this level only accounts for extreme poverty, which suits developing nations. Lastly, future studies that aim to estimate country-specific out-of-pocket payment thresholds are recommended to examine national poverty reduction policies.

IPS panel unit root test.

Note: *** denotes 1% significance level. The first-difference critical values are −1.770, −1.840, and −2.000 at 10%, 5%, and 1%, respectively.

Robustness analysis dynamic panel threshold results—all countries.

Note: ***, ** denote 1%, 5%, respectively. Between ( ) are robust standard errors. N refers to the number of countries. The variables COR and GINI refer to corruption index (International Country Risk Guide) and gini coefficient, respectively.

List of countries.

Author Contributions

Conceptualization, A.S. and N.M.N.; Formal analysis, A.S.; Methodology, A.S.; Project administration, N.M.N.; Resources, N.M.N.; Software, A.S.; Writing—original draft, A.S.; Writing—review and editing, N.M.N. All authors have read and agreed to the published version of the manuscript.

This research received no external funding.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Data availability statement, conflicts of interest.

The authors declare no conflict of interest.

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Childhood Poverty, Adverse Childhood Experiences, and Adult Health Outcomes

Affiliations.

  • 1 postdoctoral researcher, Department of Social Welfare, Ewha Womens University, Seoul, Korea.
  • 2 interim director, professor, and doctoral program chair, School of Social Work, University of Wisconsin-Madison.
  • 3 Vilas Distinguished Achievement Professor, School of Social Work, University of Wisconsin-Madison.
  • 4 executive director, Wisconsin Child Abuse and Neglect Prevention Board, Madison.
  • PMID: 34312679
  • DOI: 10.1093/hsw/hlab018

This study aimed to consider childhood poverty in relation to a count measure of adverse childhood experiences (ACEs) as a predictor of adult health outcomes and to determine whether associations are sensitive to how childhood poverty is operationalized. A sample of 10,784 adult residents was derived using data 2014-2015 Wisconsin annual Behavioral Risk Factor Survey data, derived from the Centers for Disease Control and Prevention (CDC) Behavioral Risk Factor Surveillance System (BRFSS). Adult health outcomes (health risk behaviors, general health problems, chronic health problems, and depression) were predicted using a more conservative and severe indicator of childhood poverty, and authors tested whether observed associations were attenuated by the inclusion of an ACE count variable. Findings showed that severe indicators of childhood poverty are associated with general and chronic health problems as well as adult depression. These associations are attenuated, but remain intact, when ACEs are included in regression models. Using the CDC BRFSS data for Wisconsin, the study showed that associations between childhood poverty and adult health are sensitive to the way in which childhood poverty is operationalized. The relationship between childhood poverty and other ACEs is complex and thus warrants treating the former as a distinct childhood adversity rather than an item in an ACE summary score.

Keywords: adult health; adverse childhood experiences; childhood; poverty.

© 2021 National Association of Social Workers.

  • Adverse Childhood Experiences*
  • Behavioral Risk Factor Surveillance System
  • Outcome Assessment, Health Care
  • Risk Factors

Grants and funding

  • Wisconsin Behavioral Risk Factor Survey

Advertisement

Advertisement

Energy poverty and the convergence hypothesis across EU member states

  • Original Article
  • Open access
  • Published: 05 May 2023
  • Volume 16 , article number  38 , ( 2023 )

Cite this article

You have full access to this open access article

  • Athanasios Anastasiou   ORCID: orcid.org/0000-0003-4546-7846 1 &
  • Eftychia Zaroutieri 1  

1908 Accesses

3 Citations

2 Altmetric

Explore all metrics

Energy poverty is an emerging issue towards global affairs. Currently, the development of energy-related policies is becoming essential, with regard to new societies, social inclusion and social rights. In this paper, we examine the dynamic patterns of energy poverty among 27 EU member states between 2005 and 2020. We use the log- t regression test to investigate the convergence hypothesis, and the P&S data-driven algorithm to detect potential convergence clubs. The empirical results of energy poverty indicators are mixed, and the convergence hypothesis of the states is rejected. Instead, convergence clubs are exhibited, implying that groups of countries converge to different steady states in the long run. In view of the convergence clubs, we suggest that the affordability of heating services is potentially explained by structural conditions of housing, climate conditions and energy costs. Besides, the adverse financial and social conditions for the European households have significantly triggered the arrears on utility bills. Moreover, a significant proportion of households do not have basic sanitation services.

Similar content being viewed by others

null hypothesis about poverty

Household Energy Poverty and the “Just Transition”

null hypothesis about poverty

Is energy poverty of Eastern European countries a threat or opportunity in the European Union’s fight against climate change?

Volkan Han, Oguz Ocal & Alper Aslan

null hypothesis about poverty

Energy Security, Economic Growth, and Poverty Reduction: Empirical Evidence from Selected ASEAN Member States

Avoid common mistakes on your manuscript.

Introduction

International organizations and government agencies are deeply concerned about tackling energy issues as a pillar of achieving major development goals (Barnes et al., 2011 ; He et al., 2022 ). In the light of this consideration, the United Nations defined “Affordable, reliable, sustainable and modern energy for all” as an objective for the implementation of the 2030 Agenda. The European Pillar of Social Rights affirms that energy is an essential source which any household should be granted. Hence, energy poverty is a core issue in civil society and a priority that forms the pillar of multiple initiatives and policy implications, set by the European Commission and its stakeholders (Bouzarovski & Tirado Herrero, 2017 ). Besides, the energy sector undergoes rapid changes and challenges followed by the global concerns about climate change, shift in energy prices, social welfare and the sustainable development goals (González-Eguino, 2015 ).

Recently, an academic and policy knowledge has grown over energy poverty, arguing that the last is a social rather than an inherently environmental or economic issue, followed by global affairs (Samarakoon, 2019 ). Verily, vulnerable consumers where inevitably hit by the adverse effects of the outbreak of COVID-19 pandemic on the energy sector (Carfora et al., 2022 ; Hesselman et al., 2021 ). The rising energy prices accompanied by the pandemic made several people unable to pay their energy bills (Baker et al., 2021 ). As a result, final energy use was significantly reduced by the household sector and mainly by vulnerable groups such as lower-income households (Clark et al., 2021 ). Nevertheless, it is not only the health crisis that caused detrimental effects on energy security and conservation, but also energy geopolitics. Concurrently, Russia’s war of aggression against Ukraine and the immediate response of nations, i.e. the imposed sanctions to the country, intensified the existing energy crisis in the European Union (Siksnelyte-Butkiene, 2022 ). The burden of rising prices is heavily borne by households through the rising cost of fuels, which are essential for heating, cooling, transport and other daily needs, as well as via the decrease in disposable income, i.e. the higher share of energy costs on households’ expenditures (Biernat-Jarka et al., 2021 ). Nevertheless, cross-country differences are observed, implying that the effects of rising prices and other energy-related issues of individual welfare and prosperity are disturbed by country-specific characteristics (Carfora et al., 2022 ). Hence, energy-related issues of disposable income and the living standards of individuals should be considered both at regional and international levels.

Even though the academic community dives into the causes and effects of energy poverty, the dynamic behaviour of energy poverty across economies is ambiguous. Understanding the evolutionary patterns is essential in terms of policymaking and decision. We then throw that the investigation of energy poverty patterns among EU member states is a precondition for the European Commission to enact effective policy measures. The analysis of different aspects of energy poverty convergence, i.e. the exploitation of different indicators, is essential in the measurement of disparities. This contributes to the detection of convergence behaviour between members which is inherently affected by the definition; viz., a country may appear in one cluster when using one indicator, while it may be included in a different club when another indicator is considered. In all respects, potential convergence clubs provide an understanding of the evolution of energy poverty, given different aspects of the term. Besides, the existence of any clusters implies that energy poverty declines faster in some economies rather than the other ones. This, however, enables policymakers to design country-specific policy measures and initiatives rather than uniform policies which may be ineffective for some states.

In this framework, the purpose of this study is to investigate the dynamic patterns of energy poverty across 27 EU member states between 2005 and 2020. To be more precise, we exploit the log- t regression test, as proposed by Phillips and Sul ( 2009 ), and we examine the convergence hypothesis for the whole panel. Besides, we use the data-driven algorithm to detect potential convergence clubs of the EU economies. We are then concerned to answer the following research questions that currently originate from the energy poverty. Our research questions are as follows: Do the European economies converge to an equilibrium steady state? Given that this hypothesis is violated, do countries define potential convergence clubs? If we account for different aspects of energy poverty, in what level do the results vary?

Literature review

  • Energy poverty

The understanding of energy poverty requires a consistent and valid definition, which is innately misleading and ambiguous (Deller et al., 2021 ). Besides, an inaccurate definition of the term makes the identification of households who live under the risk of energy poverty difficult for academic community and policymakers (Pachauri & Rao, 2020 ). Moreover, both energy poverty and poverty are abstruse concepts, i.e. the first accounts for the energy-related aspect of the second (Halkos & Gkampoura, 2021a , 2021b ). Nevertheless, an acknowledged definition of energy poverty relates to the inadequate access to energy services; viz., energy-poor are the individuals/households which are unable to afford basic needs in their homes, such as cooking, heating, cooling and lighting (Castaño-Rosa et al., 2019 ; Turai et al., 2021 ). Energy poverty is a multi-dimensional concept which is captured by different aspects. The European Commission argues that a single definition of energy poverty is inexistent, although accepting that it is kindly expressed by the inability to keep homes adequately warm (Thomson & Bouzarovski, 2020 ). Energy poverty is related to low income, high-energy prices and energy inefficiencies which are technically associated to infrastructures (Brucal & McCoy, 2021 ). Verily, this issue reflects the economic and social conditions within economies, i.e. it illustrates different aspects of poverty and welfare which are attributed to the distribution of energy (Karpinska & Śmiech, 2020 ). To that end, the concept is viewed over two perspectives, that is over a micro-survey and a macro-perspective (Rademaekers et al., 2016 ). Primary indicators are designed to capture energy poverty from the view of disposable income and spending for energy services, while secondary ones reflect multi-dimensions, i.e. indicators that reflect energy prices, housing infrastructure, building stock features and poverty levels.

From a different perspective, the access to energy services is the ability that an individual acquires, in terms of social integration (Bouzarovski et al., 2020 ; Hanke et al., 2021 ). This arises from the fact that the distribution of energy across households is uneven, implying that energy poverty can increase inequalities and threaten individual welfare and health conditions (Bouzarovski et al., 2014 ; Thomson & Bouzarovski, 2018 ). To be more precise, the inability of covering daily needs, e.g. cooking, cooling, heating and lighting, is evident in several households across economies. Approximately 8% of the European population is unable to keep their home adequately warm in 2020, while large disparities exist across the EU member countries. The difficulties of households to meet basic human needs generate barriers that burden well-being and human development.

In consequence, the European Commission is concerned about tackling energy poverty and adopts long-term schemes and initiatives to mitigate energy, social and economic disparities. The global 2008 financial recession and the outbreak of 2019 health crisis had a deep and long-lasting impact on social and economic inequalities among EU countries (Anastasiou, 2009 ; Arbolino & Di Caro, 2021 ; Baldwin & Di Mauro, 2020 ; Conte et al., 2020 ; Dayrit & Mendoza, 2020 ; Zervoyianni & Anastasiou, 2007 , 2009 ; Zervoyianni et al., 2014 ). Recently, the European policy measures against poverty and income inequalities lie around the promotion of convergence to a minimum steady state. Howbeit, the understanding of disparities in economic and social sectors, followed by the aforementioned issues, is ambiguous. Precisely, the real gains from policy measures and initiatives vary across member states, despite the formation of a unified economic area (Bouzarovski & Tirado Herrero, 2017 ). From a holistic perspective, it is essential to understand how energy disparities and wider socioeconomic inequalities relate to each other, via the lens of differences in energy poverty outcomes.

  • Convergence

The concept of convergence was originally introduced to explain the evolution of per capita income, i.e. economic disparities exhibiting from the integration to international trade activities, capital accumulation, labour and capital productivity (Solow, 1956 ). In broader terms, the catch-up effect implies that, as economies exhibit higher growth rates, the per capita income is gradually transferred from the developed to developing countries. The dynamic behaviour explains how capital is accumulated in poor countries, so that a steady state between wealth and poor economies is achieved (Barro, 1991 ; Barro et al., 1991 ; Barro & Martin, 1992 ). Early studies on convergence are grounded on economic growth, while recent work is delved into the fields of energy and environmental economics. Distinctive studies focus on energy efficiency (Cheng et al., 2020 ; Liu & Li, 2018 ; Markandya et al., 2004 ; Pan et al., 2015 ; Zhang et al., 2017 ), consumption (Akram et al., 2020 ; Kim, 2015 ; Mishra & Smyth, 2017 ; Pan & Maslyuk-Escobedo, 2019 ) and intensity (Markandya et al., 2006 ; Solarin, 2019 ; Szép, 2016 ; Zhang & Broadstock, 2016 ).

The literature on convergence consists of σ-convergence, β-convergence, stochastic convergence and club convergence. Precisely, σ-convergence captures the cross-sectional equilibrium over time (viz., the variation of per capita income between economies reduces) (Quah, 1996 ), while β-convergence is attributed to the increasing growth rates of less-developed economies. The last concept implies that a steady-state equilibrium is achieved when poorer nations exhibit rapid growth, which is mainly led by an increase in marginal productivity and diminishing returns to capital, leading to a long-run convergence (Horta & Camanho, 2015 ). To this end, β-convergence examines the relation between per capita income at initial stages and following growth rates of economies (Costantini and Lupi, 2005 ). If the correlation is significant and negative, β-convergence is achieved, while positive correlation implies a divergence behaviour of poor and wealth nations. Nevertheless, given the assumption of equal speeds of convergence, homogeneous technological progress and initial income between economies, the results of convergence or divergence between rich and poor economies may be biased and inconsistent (Durlauf, 2003 ; Phillips & Sul, 2003 ). Moreover, the unobserved heterogeneity between economies disturbs the empirical estimations, due to endogeneity and omitted variable issues.

Stochastic convergence then accounts for stationarity issues, i.e. convergence between states is achieved if the per capita income of a country relative to the benchmark country is an I(0) stationary process (Lee et al., 1997 ). In other terms, stochastic convergence suggests that long-run economic growth does not depend on idiosyncratic country-specific characteristics, rather these effects are temporary (Evans, 1998 ). Howbeit, stochastic convergence is sensitive to less effective univariate tests and misspecification errors, if time series issues are not considered, e.g. structural breaks or cyclical components (Bigerna et al., 2021 ; Carrion-i-Silvestre & German-Soto, 2009 ).

Given these shortcomings, multiple steady-state equilibrium points imply the existence of convergence groups. In view of this, the rejection of convergence hypothesis over a group of countries does not necessarily imply a divergence behaviour between subgroups of economic unities. In this view, Phillips and Sul ( 2007 , 2009 ) have recently constructed a non-linear time-varying factor model, which accounts for individual and transitional heterogeneity of economies and is independent of trend stationarity assumptions of the series. A data-driven algorithm detects convergence clusters, even if the convergence hypothesis is rejected for the whole panel. The methodological approach of Phillips and Sul ( 2007 , 2009 ) has many benefits over the conventional methods of convergence. First, there is no strict assumption of the stationarity of the variables and the common factors. Moreover, the method accounts for the changing behaviour of the economies, i.e. the transition heterogeneity or divergence. Third, the method has the capability to detect possible convergence clusters between heterogeneous economies, i.e. convergence paths where countries move towards a steady state in the long run (Saba & Ngepah, 2022a , 2022b ). Fourth, the data-driven algorithm further enables the detection of club merging between the existing clusters, when the clustering procedure overestimates the real number of clusters (Saba & Ngepah, 2022a , 2022b ).

Recently, studies have mainly focused on the impacts and causes of energy poverty (Ballesteros-Arjona et al., 2022 ; Neacsa et al., 2020 ; Papada & Kaliampakos, 2018 ; Primc et al., 2019 ), rather than the convergence hypothesis. Notwithstanding, the literature on energy poverty convergence is scare, while, to the best of our knowledge, two studies have investigated the hypothesis. Huang et al. ( 2022 ) exploit the method of moments quantile regression to examine whether 28 EU countries converge to an equilibrium steady state of energy poverty, while the empirical results confirm the hypothesis. Similarly, Salman et al. ( 2022 ) investigate the hypothesis, considering 146 economies. The empirical results of the Phillips and Sul algorithm suggest that the countries showcase divergence behaviour, but convergence clusters are existent. The rationale of this article is to examine the club convergence between 27 EU countries, considering three different aspects of energy poverty.

Methodological approach

Log-t regression test.

A model of the panel data variable X it is decomposed into a systematic component ( g it ) and a transitory component ( a it ) as follows: \({X}_{it}={\alpha }_{it}+{g}_{it}=\left(\frac{{\alpha }_{it}+{g}_{it}}{{\mu }_{t}}\right){\mu }_{t}={\delta }_{it}{\mu }_{t}\) for i  = 1, 2,…, N economic unities which are observed between t  = 1, 2,…, T periods. Then, \({\mu }_{t}\) is a single common component and \({\delta }_{it}\) is a systematic idiosyncratic element, i.e. a time-varying heterogenous component which accounts for the distance between \({X}_{it}\) and the common factor \({\mu }_{t}\) . If \({\mu }_{t}\) is the common trend between individuals, then \({\delta }_{it}\) is a transition parameter as the share of \({X}_{it}\) in the common trend component. To estimate \({\delta }_{it}\) , the authors construct a relative transition parameter as \({h}_{it}=\left(\frac{{X}_{it}}{\frac{1}{N}\sum_{i=1}^{N}{X}_{it}}\right)=\left(\frac{{\delta }_{it}}{\frac{1}{N}\sum_{i=1}^{N}{\delta }_{it}}\right)\) . This parameter reflects the transitional behaviour of individual i with respect to the cross-sectional average at time t . The convergence hypothesis implies a common limit in the transition path between the economic unities so that \({h}_{it}\to 1\) for all i  = 1, 2,…, N , as t \(\to \infty\) . If \({\delta }_{it}\to \delta\) as t \(\to \infty\) , then the cross-sectional variation should converge to zero as follows: \({H}_{it}=\frac{\sum_{i=1}^{N}{\left({h}_{it}-1\right)}^{2}}{N}\to 0.\) Phillips and Sul use the log- t regression test to construct the null hypothesis of convergence by technically using the following assumption for the idiosyncratic component \({\delta }_{it}\) where \({\delta }_{it}={\delta }_{i}+{\sigma }_{i}{\xi }_{it}{L(t)}^{-1}{t}^{-\alpha }\) . Then, \({\delta }_{i}\) and \({\sigma }_{i}\) are constant; \({\xi }_{it}\) is i.i.d. with zero mean and variance 1, i.e. \({\xi }_{it}\sim (\mathrm{0,1})\) ; L ( t ) is a slowly varying function which eliminates the increase in the variance of log( t ) where L ( t ) \(\to \infty\) as t \(\to \infty\) ; and \(\alpha\) captures the decay rate, that is the rate of convergence. The null and the alternative hypotheses are then obtained as Ho: δ i  =  δ and α  ≥ 0 against H 1 : δ i  ≠  δ and α  < 0. Under the null hypothesis, the whole panel convergence is achieved. To test the null hypothesis, the authors estimate the following regression model: \(\mathrm{log}\left(\frac{{H}_{1}}{{H}_{t}}\right)-2\mathrm{log}(L(t))=\alpha +\beta \mathrm{log}(t)+{\varepsilon }_{t}\) for \(t=\) [ rT ], [ rT ] + 1,…, T , for all r  > 0, where rꞓ (0.2,0.3). An optimal value of r is 0.2 for T  ≥ 100. Then, the one-side t test of heteroscedasticity and autocorrelation is conducted for \(\beta\) coefficient. The null hypothesis of convergence, i.e. Ho: β   \(\ge\)  0, is rejected if the t -statistic of the test is below − 1.65 as the critical value at the 5% level of significance, that is if \({t}_{\widehat{\beta }}=\frac{\widehat{\beta }-\beta }{\mathrm{se}(\widehat{\beta })}<-1.65\) . Instead, the whole panel convergence is exhibited if \({t}_{\widehat{\beta }}>-1.65\) .

Club clustering algorithm

The rejection of the whole panel convergence does not strictly imply the existence of convergence clubs, i.e. subgroups of countries that converge to local steady states. Phillips and Sul ( 2007 ) hence developed an algorithm that explores potential clubs as follows:

Sort the individuals by descending order regarding the sample mean of the last period.

Find k , 2 ≤  k  ≤  N : For each k , k  + 1, k  + 2,…, k  +  j individuals which are iteratively incorporated to a core group, conduct the log- t regression test until \({t}_{\widehat{\beta }}>-1.65\) . If no k satisfies the convergence hypothesis, then no convergence subgroups are exhibited. Otherwise, find k such that k  = arg max k { tk } subject to min{ tk } >  − 1.65. The core group G k * is then defined.

Initial club membership: For every individual which is not included in the core group, add one at a time to the core group and conduct the log- t regression test, until any individuals are available. If \({t}_{\widehat{\beta }}>-1.65\) , then the individual is included in the initial core group G k * .

Recursion and stopping rule: The remaining individuals define a new subgroup, where the log- t regression test is conducted. If the convergence hypothesis is achieved, then two final subgroups are extracted. Instead, if the convergence hypothesis fails, then steps 1–3 are repeated to detect the existence of further clubs.

Club merging: When the initial clusters are determined, then for each one pair, i.e. club 1 + 2, club 2 + 3, club k  − 1 +  k , the log- t regression test is conducted. If the convergence hypothesis is accomplished, then these clubs merge into a larger group. This procedure is repeated until no groups can be merged. The final classification is completed.

Data description and analysis

To detect the convergence patterns of energy poverty across EU member counties, we employ a dataset from the EU Statistics on Income and Living Conditions (EU-SILC). Despite the prevailing term of energy poverty, a uniform definition is difficult to be deployed, implying that a fundamental understanding is essential (Day et al., 2016 ). Energy poverty is a multi-dimensional factor which is reached by several indicators and methods (Bollino & Botti, 2017 ). These indicators according to the Energy Poverty Observatory are classified into two aspects, i.e. primary and secondary indicators, as already mentioned in a previous section.

Primary indicators are designed to capture energy poverty from the view of disposable income and spending for energy services, while secondary ones reflect multi-dimensions, i.e. indicators that account for energy prices, housing infrastructure, building stock features and deprivation levels (Sy & Mokaddem, 2022 ). Precisely, the Energy Poverty Observatory (EPOV) classifies primary indicators in consensual-based and expenditure-based indicators, which originate from the Household Budget Survey (HBS) micro-data. Consensual-based indicators reflect households’ potential to meet basic human needs which are energy related, i.e. the accessibility of modern and conventional energy services that are essential for cooking, cooling, lighting or heating needs. EPOV’s primary indicators are calculated directly from the responses of individuals, i.e. the indicators are based on self-assessment. Instead, expenditure-based indicators capture the level of energy poverty based on the household income. The indicators account for the share of household’s expenditure on energy services to the total disposable income or total household expenditure, or the level of household’s energy expenditures with respect to the minimum level of basic energy services (Halkos & Gkampoura, 2021a , 2021b ). Expenditure-based indicators are determined by certain thresholds which define a level of disposable income or energy expenditures, below or beyond which households are qualified as energy-poor (Romero et al., 2018 ; Sareen et al., 2020 ). The thresholds set a minimum or a maximum level of these aggregates, so that households should spend to access modern energy services and be fairly integrated in society (Belaïd, 2022 ).

Secondary indicators of energy poverty comprise a range of variables which are extracted from the EU-SILC survey and the Building Stock Observatory (BSO). Even though these metrics are supporting indicators which can enrich the understanding of energy poverty outcomes, they are not sufficient for the measurement of energy poverty (Rademaekers et al., 2016 ). The secondary indicators are based on multiple areas, i.e. demographic factors (household size, tenure status, location), energy demand and supply (prices, tariffs), income (households in material deprivation or at risk of poverty), building efficiency (quality of building, tenure status, dwelling type) and policy intervention–based factors (social income support).

Presently, expenditure-based data coverage on energy poverty is limited; hence, HBS-based indicators are currently unavailable. Accordingly, we have access solely to consensual-based data. Hence, for our analysis, we employ a dataset with consensual-based and secondary indicators. The first primary indicator of energy poverty, from the Statistics on Income and Living Conditions (SILC) survey, pertains to the affordability of energy prices, that is, the population unable to keep their home adequately warm by poverty status as a share of the total population and expressed in percentage (EP1). Particularly, the indicator is retrieved from the “Environment and Energy” database for the “Sustainable Development indicators Goal 7”. We further exploit a different consensual-based indicator for the arrears of utility bills, i.e. the population unable to afford utility bills for the main dwelling on time, due to financial distress, as a share of the total population in percentage terms (EP2). On the other hand, we are interested in focusing on a different aspect of energy poverty that captures income conditions in terms of the affordability of certain commodities or services. In this view, we exploit a secondary indicator that embodies poverty and health risks, namely the material deprivation for the “Housing” dimension. The European Statistical Authority (Eurostat) has recently proposed the material deprivation rate as a measure of households’ ability to afford necessary or desirable items, in the matter of human well-being. Material deprivation is a newly introduced measure, which is designed to capture material living conditions of households and individuals. Withal, material deprivation rate is not a unique indicator, but it consists of several statistical indicators, each one of them corresponding to a different dimension. These indicators will potentially provide useful information on monetary poverty, which, in turn, avail policy coordination and progress monitoring for social exclusion and deprivation. Precisely, the indicators are intended to acquire information on different aspects of monetary poverty determinants, i.e. each item corresponds to an overall dimension of lifestyle deprivation relative to economic strain, durables and housing. For the aim of the study, we use material deprivation rate for the housing dimension from the SILC survey, for the households whose income is below 60% of median income. In this instance, we exploit < 0.6 M threshold, rather than > 0.6 M or total, to follow up on lower- and middle-income households in relation to the risk of poverty threshold. The material deprivation for the housing dimension corresponds to the share of individuals who are deprived in terms of materials, i.e. they reside in distressed dwellings. Eurostat defines materially deprived for housing those persons who have buildings where they suffer from at least one of the following items: (a) leaking roof, damp walls/floors/foundation or rot in window frames; (b) accommodation too dark; (c) no bath/shower; and (d) no indoor flushing toilet for sole use of the household. The indicator is calculated over the number of items, i.e. population who suffer from 0 item, 1 item, 2 items, 3 items or 4 items. However, we use the material deprivation of individuals who suffer from 3 out of the 4 items (EP3). Eventually, we collect data for 27 EU countries which are observed between 2005 and 2020 and the sample consists of 432 observations.

Table 1 represents the descriptive statistics of the energy poverty indicators. The distribution of EP indicators is positively asymmetrical, implying extreme values with respect to most frequent ones. Nevertheless, the log- t regression test and the club clustering algorithms are not sensitive to extreme values, due to the smoothing and standardization processes. The sample mean of the share of population who are unable to keep their home adequately warm is 11.5%, and the standard deviation is 12.1%, which justifies the potential asymmetries. Besides, the difference between the medians and the third quartiles (Q3) explicates the asymmetries which are either attributed to the disparities across members or the deviations within members over time.

To show the distribution and the dispersion of the data for the EP indicators within EU countries, we illustrate the boxplots in Fig.  1 . We can then detect how the values of EP indicators are spread out for each country over the reference period. Overall, the distribution of EP indicators is asymmetrical, implying the countries’ deviations over time. Evidently, the distribution of population unable to keep their home adequately warm in Bulgaria is higher relative to the EU members. Even though the affordability of heating services was decently improved in the 2010s, Bulgaria is dependent on the use of natural gas and coal for electricity generation and supply which are imported from Russia. In this view, any disruption on supply will dramatically affect the Bulgarian economy, due to the increase in electricity prices and the unexpected shortages. Verily, the dramatic rise in electricity prices makes the costs for electricity production challenging, which is followed by adverse effects on vulnerable consumers. Conversely, countries such as Austria, Estonia, Netherlands, Luxembourg, Finland and Sweden maintain lower levels of energy poverty, in favour of energy services affordability. The financial conditions of Austrian households evidence the sustained efforts of government to protect poorer households, in the matter of energy services expenditures. Moreover, Luxembourg has set initiatives for the provision of financial aid and support for low-income persons who are unable to meet energy costs, since 2009. An illustration of these programmes is the administration of zero-interest bank loans or the consulting towards a more efficient energy use, and the replacement of energy-guzzling devices with a subsidy of the acquisition cost up to 75%. Although Estonia’s households’ expenditures for heating are high, because of the weather conditions, i.e. the cold climate, the country maintains lower shares of people unable to afford heating services than other EU members. However, the inability to access adequate warming relative to the income conditions in urban areas exceeds that of rural areas. Soever, the relatively low energy costs and the stabilization of electricity prices since 2013 significantly reduce the prospective risks of energy poverty in Estonia. Considering the arrears of utility bills, Greece has experienced severe problems in meeting payment needs for mortgage; rent; utility bills including electricity, natural gas, water and waste; or hire purchase payments. The Greek economic has faced a serious disruption anent economic instability and political turmoil, followed by the Great Recession. As a result, households’ financial ability was in a ruinous state, intensifying the affordability of expenses. Croatia maintains a higher share of the population with difficulties on utility bills relative to the EU average, which evinces a limited purchasing power of Croatian residents. With respect to material deprivation of housing, the distribution of Romania is higher relative to other EU member countries, which implies that an overwhelming number of households suffer from three out four housing dimension items. Recall that the indicator we exploit for material deprivation of housing is based on three of the following items: (a) leaking roof, damp walls/floors/foundation or rot in window frames; (b) accommodation too dark; (c) no bath/shower; and (d) no indoor flushing toilet for sole use of the household. Hence, Romanian private households face severe difficulties in their dwelling. Similarly, a large portion of population in Latvia and Lithuania is considered as materially deprived in the matter of housing. Greece, Bulgaria, Latvia and Lithuania are among the EU members whose citizens are severely material deprived, in terms of dwellings issues. Howbeit, deprivation in Bulgaria has surpassed all the other members, which is followed by the fact that Bulgaria’s social policy measures have failed to reduce inequalities. Meanwhile, the estimation of costs of living is a precondition for the implication of effective policy measures and schemes for material deprivation rates of housing.

figure 1

Distribution of energy poverty indicators across EU member states (reference period 2005 – 2020)

Results and discussion

Analysis of whole panel convergence.

Table 2 illustrates the results of the log- t regression test for each indicator, respectively. Recall that the log- t regression test is a left-tailed t test in which the critical value at the 5% level of significance is − 1.65. The null hypothesis of the whole panel convergence (Ho: \(\beta \ge 0\) ) is rejected if the t -statistic under the null hypothesis, i.e. \({t}_{\widehat{\beta }}\) , is smaller than − 1.65. Hence, if \({t}_{\widehat{\beta }}<-1.65\) , we reject the convergence hypothesis at the 5% level of significance and we infer that countries diverge in the long run. The null hypothesis of the whole panel convergence for the first indicator is rejected, that is the t -statistic of the estimated regression coefficient \({\widehat{\beta }}_{1}\) is below the − 1.65 critical value, i.e. \({t}_{{\widehat{\beta }}_{1}}=-28.545<-1.65\) . Accordingly, the null hypothesis of convergence in the second and third indicators between EU countries is rejected, in that \({t}_{{\widehat{\beta }}_{2}}=-60.306<-1.65\) and \({t}_{{\widehat{\beta }}_{3}}=-4.981<-1.65\) . We then infer that the European member states exhibit a divergence behaviour related to the population unable to keep home adequately warm, the arrears on utility bills and the material deprivation for housing. The relative transition paths of energy poverty indicators illustrate the divergence behaviour of EU member states, that is countries move towards different steady states in the long run (Fig.  2 ). To be more precise, each line corresponds to the relative transition parameter of a given member state for the reference period 2005–2020. The EU economies exhibit a convergence behaviour in specific periods, that is when the relative transition curves cross, which captures the transitional behaviour. In the aggregate, the union shows disparities, withal. Therefore, the European countries have, so far, failed to jointly reach a minimum level of the individuals who are unable to keep their home adequately warm, who are in arrears with utility bills and who are materially deprived in terms of the housing dimension.

figure 2

Relative transition paths of energy poverty indicators

Analysis of the clustering algorithm for convergence clubs

Nevertheless, the rejection of the whole panel convergence in energy poverty does not entail the inexistence of individual clusters between the states. To detect potential convergence clusters, we conduct the club clustering algorithm, and we summarize the results in Tables 3 , 4 and 5 . Recall that, after the initial clusters are defined, the algorithm performs multiple iterations for potential merged clusters until the final clusters are detected. The first part of the tables shows the initial clusters, in which the numbers in brackets indicate the number of countries in each cluster. To be more precise, the countries are initially sorted by descending order, and the P&S algorithm conducts multiple log- t regression tests to detect the core group. The algorithm adds one country at a time to a core group, until the null hypothesis is not rejected, i.e. member i does not converge to the core group. Following this procedure, the log- t regression test is conducted by adding each one of the residual members in a new cluster, until the null hypothesis is not rejected. These steps are repeated right up until the initial clusters are extracted.

The initial classification of EP1 suggests three clusters (Table 3 ). Precisely, the estimated \(\widehat{{\beta }_{1}}\) coefficients of the log- t regression tests for the initial clusters 1, 2 and 3 are statistically insignificant, in that the estimated t -statistics are greater than the 5% critical value, i.e. \({t}_{\widehat{{\beta }_{1}}}=1.451>-1.65\) , \({t}_{\widehat{{\beta }_{1}}}=-0.501>-1.65\) and \({t}_{\widehat{{\beta }_{1}}}=1.688>-1.65\) . This implies that the null hypothesis of convergence for the initial clusters is not rejected; hence, 8, 10 and 9 EU member states define the first, the second and the third initial clusters, respectively. Given the identification of the initial clusters, the P&S algorithm investigates whether the initial clusters can be merged into a larger cluster. For each pair of the initial clubs, the test of club merging, i.e. the log- t regression test, is iterated until no more clubs can be merged into a larger convergence group. Precisely, each member of the initial club 2 is added in the initial club 1 and the log- t regression test is anew conducted. The test of club merging between the initial clubs 1 and 2 yields \({t}_{\widehat{\beta }}=-8.618<-1.65\) , implying the rejection of the null hypothesis, i.e. clubs 1 and 2 are not merged into a larger group. The test of club merging is repeated for the initial clubs 2 and 3, successively. The log- t regression test gives \({t}_{\widehat{{\beta }_{1}}}=-16.943<-1.65\) , and the null hypothesis is rejected, i.e. clubs 2 and 3 diverge. Finally, the initial clusters define the final groups as the club merging is not feasible. Hence, the classification suggests that three groups of countries converge to an equilibrium steady state of the population unable to keep their home adequately warm.

Table 4 shows the results of the arrears in utility bills. Against the small number of initial clubs of the affordability for heating services, the initial classification of the arrears in utility bills (EP2) evinces six convergence clubs and one divergence club. Accordingly, we observe that the initial clusters 1, 2, 3, 4, 5 and 6 comprise 5, 6, 9, 2, 2 and 2 members, respectively. The estimated \(\widehat{{\beta }_{2}}\) coefficients of the preliminary log- t regression tests are statistically insignificant, that is given the non-rejection of the null hypothesis, the regression parameters are equal or greater than zero. To that end, the algorithm provides strong evidence for the existence of six initial convergence groups of countries and one divergent member. Analytically, the estimated t -statistics for the initial clubs 1–6 are greater than the 5% critical value, i.e. \({t}_{\widehat{{\beta }_{2}}}=0.021>-1.65\) , \({t}_{\widehat{{\beta }_{2}}}=0.833>-1.65\) , \({t}_{\widehat{{\beta }_{2}}}=2.254>-1.65\) , \({t}_{\widehat{{\beta }_{2}}}=18.325>-1.65\) , \({t}_{\widehat{{\beta }_{2}}}=-1.139>-1.65\) and \({t}_{\widehat{{\beta }_{2}}}=1.153>-1.65\) . This indicates that the null hypothesis of convergence for the initial clusters is not rejected, and 5, 6, 9, 2, 2 and 2 EU member states form the first, the second, the third, the fourth, the fifth and the sixth initial clusters, individually. In turn, the algorithm performs a step-by-step club merging procedure. The test of club merging between the initial clubs 1 and 2 gives \({t}_{\widehat{{\beta }_{2}}}=-2.465<-1.65\) , which implies the rejection of the null hypothesis; viz., the initial clubs 1 and 2 diverge. As a result, we extract the final convergence club 1, which corresponds to the initial club 1. Subsequently, the test for clubs 2 and 3 suggests the rejection of convergence hypothesis, being that \({t}_{\widehat{{\beta }_{2}}}=-5.042<-1.65\) ; hence, we have the second final convergence club, i.e. club 2, which originates from the initial club 2. Identically, the algorithm performs the test for the initial clubs 3 and 4 and the results cannot reject the null hypothesis of convergence, as \({t}_{\widehat{{\beta }_{2}}}=0.640>-1.65\) . Consequently, clubs 3 and 4 are merged into a larger group with [9] + [2] = 11 members, which form the third final convergence club. In this step, the algorithm is at a critical juncture, as clubs 3 and 4 are merged. Nevertheless, the algorithm follows the steps for the initial clubs 4 and 5, giving an estimate of \({t}_{\widehat{{\beta }_{2}}}=-7.740<-1.65\) . Accordingly, the clubs diverge and we get the final convergence club 4, which corresponds to the initial club 5. At the last step, the log- t regression test is performed for the initial clubs 5 and 6, which finally suggests the rejection of convergence, that is \({t}_{\widehat{{\beta }_{2}}}=-9.973<-1.65\) . Alternatively, the clubs diverge and the algorithm estimates the final convergence club 5, i.e. the initial club 6. Ultimately, we find five convergence clubs of countries in the arrears in utility bills. Overall, the initial clusters 3 and 4 are merged into a larger group, while the remaining initial clusters fail to merge. Hence, the residual initial clusters define the final clusters, individually. The algorithm finally estimates five groups of countries that converge and one country that diverges from the others in the arrears on utility bills.

Table 5 illustrates the results of material deprivation for the housing dimension. Three convergence groups of countries are initially formed, including 5, 7 and 15 members, respectively. The log- t regression test suggests the existence of three initial clusters since the estimated \(\widehat{{\beta }_{3}}\) coefficients are statistically insignificant, i.e. the null hypothesis in favour of convergence is not rejected. Precisely, the estimated t -statistics for the initial clubs 1–3 are greater than the 5% critical value, that is \({t}_{\widehat{{\beta }_{3}}}=30.945>-1.65\) , \({t}_{\widehat{{\beta }_{3}}}=1.309>-1.65\) and \({t}_{\widehat{{\beta }_{3}}}=2.566>-1.65\) . Then, the club merging algorithm is conducted to detect whether the initial clubs merge into larger groups. The log- t regression test between the initial clubs 1 and 2 cannot reject the null hypothesis of convergence because \({t}_{\widehat{{\beta }_{3}}}=7.918>-1.65\) . Hence, the initial clusters 1 and 2 merge into another club and define the final convergence club 1 with [5] + [ 7] = 12 countries. Subsequently, the test between the initial clubs 2 and 3 suggests the rejection of convergence, given that \({t}_{\widehat{{\beta }_{3}}}=-1.902<-1.65\) . Hence, we find the second final convergence club which corresponds to the initial club 3. Finally, the union exhibits two groups of countries that converge to an equilibrium steady state in material deprivation for the housing dimension.

Identification of members

In this section, we analyse the empirical results, and we seek to interpret the patterns of convergence between the members. We should originally identify which members compose the clusters, and the distribution of energy poverty indicators between them. Table 6 delineates the EU member states in each group, as well as the level of energy poverty. Considering EP1, we find that the “High” EP1 club consists of Bulgaria, Cyprus, Greece, Ireland, Italy, Lithuania, Luxembourg and Portugal and the “Moderate” EP1 club consists of Croatia, France, Germany, Hungary, Latvia, Malta, Netherlands, Romania, Slovakia and Spain. Verily, the EU-wide survey of 2018 for the population unable to keep their home adequately warm shows that Bulgaria, Lithuania, Greece, Cyprus, Portugal and Italy lie around the upper threshold of the distribution. Moreover, the final convergence club of EP1, i.e. the “Low” EP1 cluster, incorporates the following states: Austria, Belgium, Czechia, Denmark, Estonia, Finland, Poland, Slovenia and Sweden. The EU-wide survey of 2018 argues that Austria, Finland, Estonia and Sweden are among the countries whose individuals manifested the lowest inability to afford heating service.

Against the number of convergence clubs for EP1, EP2 showcases 5 convergence groups and one divergence group. Given the number of clusters, we set a 5-point Likert scale for the volume of arrears in utility bills, viz., “Very high”, “Above average”, “Average”, “Below average” and “Very low”. Accordingly, we estimate that the very high EP2 group consists of Bulgaria, Croatia, Cyprus, Romania and Spain, while the above average cluster consists of Denmark, Finland, Hungary, Ireland, Latvia and Slovenia. However, we should clarify that utility bills apply not only to energy; rather, energy bills are the principal component of invoices paid for essential services. We observe that Bulgarian households suffer from difficulties on utility payments and these members are unable to keep energy costs down. Despite, Austria, Belgium, Estonia, France, Italy, Lithuania, Luxembourg, Malta, Poland, Portugal and Slovakia constitute the average EP2 club, that is the countries maintain moderate levels of difficulties on utility bills. Moreover, Germany and Sweden define the below average EP2 club and Czechia and Netherlands lie around the lower threshold of the distribution, i.e. very low EP2.

Instead, the results of material deprivation rate for housing are narrow, that is the EU countries formulate two convergence clubs. High EP3 club includes Croatia, Estonia, Germany, Hungary, Ireland, Latvia, Lithuania, Luxembourg, Malta, Poland, Slovakia and Sweden, rather Austria, Belgium, Bulgaria, Cyprus, Czechia, Denmark, Finland, France, Greece, Italy, Netherlands, Portugal, Romania, Slovenia and Spain join the low EP3 club. Evidently, any geography-specific patterns across energy poverty convergence clubs are hard to determine. However, we provide potential explanations in the ensuing section.

Figure  3 illustrates the spatial distribution of energy poverty convergence clubs between the EU member states. The distribution of population unable to keep home adequately warm is uniform in the European Union. In other terms, the number of countries in each cluster is approximately equal. Indeed, the population in 8, 10 and 9 out of 27 EU members indicates high, moderate and low inability to afford heating services. Hence, we infer that the share of individuals that suffer from the inability to maintain their home warm is uniformly dispersed across countries. Nevertheless, this does not apply in reference to the arrears on utility bills. Evidently, we estimate that 11 out of 27 EU countries show average difficulties on utility payments, but [5] + [6] = [11] out of 27 members indicate “very high” and “high” arrears, respectively. Concisely, we need to highlight that most of the European households face severe and moderate difficulties on utility bills. Considering the material deprivation for housing, we find that the distribution is uniform between EU countries; viz., 12 out of 27 and 15 out of 27 members indicate high and low deprivation. In view of this, the persons who mostly or slightly live under the risk of material deprivation for housing are equally distributed between the European economies.

figure 3

Spatial distribution of energy poverty convergence clubs

Descriptive statistics

We present the descriptive statistics of energy poverty indicators between convergence clubs in Table 7 . The descriptives are extracted from the original data, i.e. exclusive of any standardization or detrending processes. With respect to EP1 indicator, we observe that the average energy poverty reduces with the clubs, viz., \({\overline{\mathrm{EP}1} }_{\mathrm{c}1}>{\overline{\mathrm{EP}1} }_{\mathrm{c}2}>{\overline{\mathrm{EP}1} }_{\mathrm{c}3}\) . Nevertheless, due to the asymmetry in the original data of EP1, the sample mean is not representative for the inference. To that end, we should account for the median and the third quartile which are kindly consistent with the skewness. If we consider for the second (median) and third (Q3) quartiles, we get 22.1% and 29.3%, respectively. In other words, the values of population unable to keep their home adequately warm are below 22.1% across 50% of the observations (60 out of 120) and below 29.3% across 75% of the observations (90 out of 120) in the first club. We observe that the median and the third quartile decrease as we move from club 1 to 3. Verily, the values of the inability to afford heating services are below 0.3% across 50% of the observations (80 out of 160) and below 5.6% across 75% of the observations (120 out of 160) in the third club. This explains our reasoning behind the definition of the 3-point Likert scale from High to Low, as well as of EP2 and EP3 Likert scale points. Given the descriptives of arrears in utility bills, we find that the values of the indicator are below 17.5% across 50% of the observations (40 out of 80) and below 28% across 75% of the observations (60 out of 80) in the first convergence club, i.e. in the very high EP2 club. Instead, 50% of the values within the moderate EP2 club are below 6.3% and 75% of them are below 9.4%. Conversely, the members to which households are relieved from arrears on utility bills showcase lower values of quartiles related to the other countries. Concerning material deprivation for housing, we detect two convergence clubs, i.e. high and low EP3 clusters. At any rate, we find that 96 out of 192 observations in the high material deprivation club receive values below 2.2%, while this percentage is 0.6% for the low EP3 cluster. From a broad perspective, the quartiles indicate that material deprivation for the housing dimension in the first club is higher related to the second club.

Interpretation of relative transition paths

Population unable to keep home adequately warm (ep1).

The transitional behaviour of countries is accurately illustrated by the relative transition paths across clubs in Figs. 4 , 5 and 6 . The plot of relative transition paths is a time series plot of the relative transition parameter h it , for country i at time t . Recall that h it captures the transitional behaviour of country i with respect to the cross-sectional average at time t . As h it limits to unity, the economies converge to an equilibrium steady state; instead, the economies showcase a divergence behaviour if h it limits to zero. From a broad perspective, we observe that the relative transitional parameters between EU members in the same club converge to an equilibrium steady state in the long run.

figure 4

Relative transition paths of the EP1 indicator across convergence clubs

figure 5

Relative transition paths of the EP2 indicator across convergence clubs

figure 6

Relative transition paths of the EP3 indicator across convergence clubs

At first, we analyse the transitional behaviour of EU members for the population unable to keep their home adequately warm (see Fig.  1 ). The relative transition paths of Cyprus, Portugal and Lithuania cross over in 2010, and the countries indicate an even speed of convergence until 2013. This implies that the economies converge to a different steady state relative to the other countries in the high EP1 group (i.e. club 1) between the reference period. Instead, Cyprus and Portugal follow the same paths until 2016, while Portugal moves towards a different path. Soever, the countries converge to a steady state in the long run.

Particularly, Bulgarian households are mostly hit by energy poverty, in the matter of heating services affordability, and the economy is among the members with the lower speed of convergence related to other members in the first cluster. Individuals who reside in countries with intense climate conditions are significantly affected by temperature outcomes (Beg et al., 2002 ; Santamouris & Kolokotsa, 2015 ). Verily, the quality of life and the living standards of the vulnerable consumers are adversely affected by extreme weather conditions (Thomson et al., 2019 ).

Precisely, the affordability of energy needs for the Bulgarian households is inherently affected by temperature, that is, the country showcases extremely high and low temperatures during summer and winter, respectively (Hajdinjak & Asenova, 2019 ). As a result, individuals should spend most of their income on winter, to keep home warm, and on summer, to access cooling services. Although Portugal indicates moderate temperatures, several households suffer from the inability to maintain heating services, while the country is one of the most vulnerable economies in the European Union (Bollino & Botti, 2017 ; Bouzarovski & Tirado Herrero, 2017 ). However, this can be attributed to significant insulation issues, because of the age and the low efficiency of the residential building stock. Precisely, most of the Portuguese buildings are old and the residential buildings can hardly conserve thermal energy due to poor thermal characteristics and low efficiency (Mafalda Matos et al., 2022 ).

Moreover, Greece, Romania, Portugal and Italy are mostly hit hard by the general upward trend on electricity prices, which is followed by network charges and taxes. Instead, to provide a potential explanation on the formation of high EP1 club, we should account for the housing conditions. Several European households are suffering from severe structural problems, which, in turn, affect the ability to keep their home adequately warm (Dubois & Meier, 2016 ). Based on the EU-SILC, Portugal, Italy, Lithuania and Greece are among the EU members with the highest rates of severe housing deprivation, i.e. a significant share of population suffers from leaking roof or darkness in the dwelling, or a lack of bath or toilet. The inefficiencies in dwellings and the population growth have significantly deteriorated persons’ inability to afford energy services and housing costs. This is evident for the citizens in Luxembourg, where structural problems on dwellings are intense across households, and particularly between the youth. Moreover, the severe housing conditions are reflected in the great number of decayed buildings and the poor thermal characteristics (Karpinska & Śmiech, 2020 ).

Conversely, the number of the intersecting points of relative transition curves in the moderate EP1 club (club 2) exceeds the respective number of high EP1 club. Specifically, this implies that countries diverge or converge to different steady states at specific periods. In the early stages of energy poverty, that is between 2005 and 2008, France and Germany converge to a steady state, while the relative transition paths diverge thereon. However, Latvia and Lithuania follow the same path in the main. Even though the relative transition parameters fluctuate over the reference period, they evidently move towards the equilibrium steady state on the long run. This is certainly true in the case of Malta, to which the relative transition parameter decreases in 2007, increases between 2008 and 2012 and then follows a downward trend, i.e. the economy moves towards the panel average. Instead, Slovakia diverges from the panel average until 2012 and EP1 finally returns to the panel mean. A certain share of households in Romania, Latvia and Hungary is also suffering from problems in dwellings, but the results show that these countries maintain moderate inability to keep home adequately warm. Alternatively, the economies in the moderate EP1 club indicate a mixed performance on the ability to afford heating services.

However, Hungary maintains the lowest energy prices in the EU which contributes to the mitigation of the adverse effects on vulnerable consumers, i.e. particularly single-parent families, unemployed or single-elderly persons and large families. Besides, the issues on housing deprivation are significantly attributed to the inefficiencies in the country’s energy supply system (Bajomi et al., 2020 ). Based on the EU-SILC surveys, economies in the moderate EP1 club showcase lower rates of severe housing conditions, rather than countries in the first club. Nevertheless, a large part of the energy-poor in France, Germany, Croatia, Spain and Malta is attributed to social housing and private tenants’ tenures. In essence, disaggregated data for these countries suggest that people who live in apartment-type dwellings are considered more vulnerable, related to those who reside in semi-detached or detached dwellings. Besides, this can be partly explained by the severe effects of financial distress, the outbreak of the global health crisis and the poor energy efficiency of buildings.

Against the fact that the relative transition paths of the moderate EP1 club osculate to each other at specific periods, the paths of low EP1 club intersect per two countries. For instance, we find that Poland and Belgium, Czechia and Slovenia, Denmark and Slovenia, Austria and Estonia, as well as Austria and Finland indicate a crossover point at specific years. Overall, the third club of EP1 includes countries that prosper in terms of the access to heating services. Economies such as Austria, Belgium, Sweden and Finland indicate higher performance than the EU average, in terms of energy poverty. Particularly, just a small portion of the Finish and Swedish population reports that they are unable to keep their home sufficiently warm. This can be attributed to the lease agreements, i.e. the fact that rental payments are frequently inclusive of energy bills, implying that the median share of the households’ energy costs on total income lowers. Nevertheless, this does not necessarily entails that energy poverty reduces; instead, energy expenditures of households are not deeply affected by rising energy prices. Moreover, Belgium, Sweden, Finland and Estonia set disconnection protection schemes during heating and winter periods. As a result, the population unable to keep home adequately warm significantly reduces.

We then infer that the inability to keep home adequately warm across convergence clubs is potentially attributed to structural conditions of housing, weather conditions (i.e. temperature outcomes) and energy costs. Ultimately, economies of which households are hit hard by the inability to keep home adequately warm, i.e. countries in the high energy poverty club, are severely affected by soaring energy prices and partially affected by weather conditions. Individuals who reside in countries that signal moderate energy poverty are also distressed by energy prices, but to a lesser extent withal. Howbeit, households in these economies showcase severe problems on buildings stock, i.e. the households suffer from structural problems which intensify the inability to access heating services. In contrast, economies that prosper in terms of the ability to afford heating services set disconnection protection schemes, that is support income and social schemes, in favour of the access to energy services. Besides, the median share of energy expenditures in households’ income is below the EU median, which is mainly attributed to the inclusion of energy bills on rental payments.

Arrears on utility bills (EP2)

Against the lower number of convergence clubs for the inability to keep home adequately warm, EU countries form a greater number of clusters in terms of the arrears on utility bills (Fig.  2 ). Foremost, we observe that the relative transition paths of the EU members within the third club, i.e. middle arrears on utility bills, indicate the highest number of intersecting points related to the other clubs. To be more precise, this implies that even though countries in the middle club converge to different steady states at specific periods, they converge to an equilibrium state in the long run. Nevertheless, just 4 out of 27 EU members form the low and very low convergence clubs, which signifies that a large part of the European households struggle with arrears on utility bills. Given the relative transition paths of the very high EP2 club (i.e. club 1), Bulgaria and Croatia converge to a steady state in 2012, but follow different paths thereon. At the early stages of transitional behaviour, that is in 2010, the relative parameters fluctuate; instead, the transitional parameters indicate a deterministic trend beyond that year. Howbeit, Bulgaria showcases the lowest speed of convergence with respect to the other members in the club. The financial crisis caused severe effects on employment and disposable income, and therefore, it triggered the arrears on utility bills across Croatian, Romanian, Spanish, Bulgarian and Cypriot households. In tandem, the arrears on utility bills in Romania are significantly affected by climate conditions, that is the country experiences cold climates which raise the energy consumption for heating services. Hence, the energy bills prove less affordable for households. Evidently, we observe that Bulgaria and Cyprus are included in the first convergence club of both EP1 and EP2 indicators. Alternatively, we suggest that these economies are worst-performing among the EU member states, in terms of energy poverty. Instead, the countries in the high EP2 club indicate a greater speed of convergence on average, rather than members in the first club. The adverse effects of financial crisis deteriorated the living and income conditions of Irish and Slovenian households, as a large part of population is at a higher risk of energy poverty. In contrast, economies which form the average EP2 club maintain social policies that mitigate the difficulties on utility bills. France and Italy are substantially setting similar policies that notably reduce the arrears on utility bills. Besides, progressive policies such as social tariffs for energy consumption are considerable in the Belgian economy. Despite, Germany, Sweden, Czechia and Netherlands display the least arrears on utility bills; viz., the economies perform better than the EU average. The notably low levels of difficulties on utility bills for the Swedish and German population are potentially attributed to the policy measures, in favour of energy efficient buildings.

Material deprivation rate on housing (EP3)

Interestingly, the relative transition parameters for the material deprivation on housing dimension formulate two convergence clubs, i.e. low and high EP3 clubs. The relative transition paths across countries smoothly converge to the equilibrium steady state. Howbeit, the relative parameters in the second club indicate extreme values for 2019. Verily, at the early stages of transitional behaviour, the curves are being osculated, i.e. the economies showcase a convergence behaviour, but they significantly diverge in 2019, while moving towards the equilibrium state in 2020. The extreme values of the relative transition parameters originate from the large values of France, Belgium, Czechia, Finland, Spain, Austria, Netherlands, Greece and Italy, as well as from the least negative values of Romania, Latvia, Lithuania and Bulgaria. The occurrence of the outliers implies that a large part of the European population suffers severely from structural problems in their homes. Several households across the union are found to lack basic sanitary facilities or struggle with problems such as leaking roofs and damp walls. Overall, households undergo profound issues on the efficiency of basic sanitary facilities and other problems in infrastructural terms.

The share of cost of heating services in the EU households’ final expenditures is high, and as a result, it threatens vulnerable consumers. Consumer vulnerability, energy poverty and the disconnections are among the major challenges of the European members. Recently, energy is an essential source for heating and cooling services, regarding the diverse climates in the European Union. If energy resources in households become unaffordable, the social and economic impacts are severe. Latest arguments of international energy agents and the European Commission consider low income, low energy efficiency on households/dwellings and high energy prices to be the fundamental causes of energy poverty. In view of these considerations, we summarize three broader factors, i.e. the economic, the social and the political perspective of energy poverty.

The effect of institutions and the regulatory framework is exemplified in the Third Energy Package of the European Commission that enters into force in 2009. Verily, the package has been criticised about the regulatory solutions in favour of the private investments in new technologies on the electricity sector. Concurrently, the increasing competition, the uncertainties and the various regulatory failures of the electricity markets affect price signals, i.e. energy prices or the short-term markets. Footnote 1 To that end, the households’ affordability for heating services is significantly affected by uniform policies or country-specific policies. This implies that, instead of the existence of uniform energy policies, the responses of governments in the electricity market’s challenges vary across the European members. The economic state, the policy framework and the socio-political systems have a considerable impact on households’ income, through austerity measures, consumer-related income schemes and energy price regulations. Moreover, the regressive effects of these challenges on vulnerable consumers are strictly associated with the effectiveness of the initiatives and the compensation mechanisms that are exerted by each state. Besides, the rate and the degree by which a European economy comforts into the European regulation system are different across EU countries.

In contrast, a current debate on energy poverty and vulnerability revolves around the definitions of the terms. Verily, in the amendments of 22 June 2022, the European Parliament argued that no standard EU-level definition of energy poverty is existent. As a result, some policy measures and initiatives on energy poverty may prove ineffective in alleviating energy poverty. Besides, the inexistence of a uniform definition impedes the collection of transparent and comparable data. In other terms, an inconclusive dispute over which part of the population is considered vulnerable adversely affects the outcomes of the undertaken initiatives and the level of support on certainly poor households. Energy inequalities are indirectly associated with the provision of an exact and clear definition of energy poverty. In view of this, the core issue is to identify who really needs to receive support, in terms of the access to clear and affordable energy services, similar income support schemes or other aspects of energy poverty. Despite the vagueness of the definition, the European Parliament states that policy measures, such as social tariffs or income support schemes alleviate poverty outcomes in favour of the energy-poor households, but in the short run. Nevertheless, while these measures are effective on short-term poverty, they are unable to support governments’ schemes to fight against poverty in the long run. In terms of the government perspective, the provision of insufficient, or even the lack of sufficient information to consumers about energy use and conservation may also cause adverse effects (Halkos & Gkampoura, 2021a , 2021b ). Economies with liberalised markets set initiatives and campaigns to build consumers’ awareness of energy issues. These issues include the efficient exploitation of energy services and products as well as the comparison of energy prices. Hence, consumers’ ignorance of energy-related issues slows down the prosecution of regulatory measures and strategies (Nussbaumer et al., 2012 ).

The underlying argument in favour of energy poverty alleviation is the adoption of structural measures, i.e. initiatives related to building renovation in terms of comfortable and energy-efficient households and dwellings. Household characteristics such as size, year of residential construction, location and building quality directly affect energy poverty outcomes. The quality and basic amenities such as household equipment, the number of rooms and the existence of energy-efficient appliances affect the energy use (Thomson & Snell, 2013 ) and, accordingly, the energy bills and the share of energy payments on housing costs. In many cases, a higher number of rooms impede the household’s ability to keep home adequately warm, particularly for low-income households (Thomson & Snell, 2013 ). Meanwhile, energy poverty outcomes differ between rural and urban areas, i.e. the location may inherently affect the share of energy-poor to the total population. The needs of rural citizens are more income oriented rather than energy oriented, i.e. urbanization increases the energy dependence as the economies focus on economic growth, while residents in rural areas may not be prioritised in spending their disposable income on unaffordable energy services. Moreover, individuals who reside in inefficient and very old homes or use older equipment indicate higher energy consumption rather than people who own energy-efficient appliances and equipment. The problems with infrastructure such as dwellings not comfortably cool during summer or warm during winter, leaking roofs, damp walls/floors/foundations, dark accommodation, problems with indoor flushing toilet for sole use and lack of heating or cooling equipment severely affect energy use. Alternatively, these issues generate energy losses, i.e. lower-quality inefficient appliances and the lack of technical possibilities of dwellings raise the energy expenditures of households and, in turn, energy bills. Thereby, energy-inefficient dwellings generate more expenditures rather than the energy-efficient dwellings, implying that lower-income households, which are unable to buy high-quality equipment, spend more money on energy services. Hence, poor housing conditions and the aforementioned structural issues affect the quality of households and the energy expenditures.

Given the income perspective, the economic conditions of states significantly affect the energy-poor. Lower-income households face several difficulties in affording modern energy services, which, in turn, affect their living conditions and well-being (Owusu et al., 2016 ). Alternatively, poor households are mostly deprived of monetary sources to cover energy needs, i.e. they are unable to afford cooling and heating services or even to cover daily mobility needs (Lucas et al., 2016 ). Nevertheless, energy poverty does not imply income poverty, i.e. these terms are not equivalent. Although energy and income poverty both reflect economic deprivation and poor living conditions, it is accepted that they measure different aspects of poverty (Thomson et al., 2019 ).

Currently, soaring energy prices have been one of the most significant drivers of energy poverty, and the energy market affairs have raised energy and economic disparities. The liberalisation, the geopolitical risks and the rivalry of energy market systems significantly affect the distribution, the aggregate demand and supply, the tariffs and the government intervention on energy resources and, consecutively, the outcomes of energy poverty (Trinomics, 2020 ). Hence, market rivalry, tariff-specific preferences and tariff choices may disturb the formation of energy prices, in favour of consumers’ protection (Dobbins et al., 2015 ). The energy price shock on the European economy has already deteriorated human well-being and economic welfare, while the expected consequences in the long run are destructive. According to the European Commission’s State of the Energy Union 2021, the steep increase in wholesale electricity prices, followed by rising gas prices, has caused detrimental effects on individuals. Verily, 31 million European citizens are considered as energy-poor in 2021, owing to energy prices spike. In the interim, the Russian invasion of Ukraine has caused a decay in global energy markets, i.e. an erratic surge in crude oil prices. Although the effects of domestic energy prices may vary due to country- and income-specific characteristics, they continue to be regressive. Ultimately, energy poverty is a multi-dimensional concept that needs to be acknowledged not only on the grounds of infrastructure, income and energy prices, but also from the political perspective.

Energy poverty is an emerging issue towards global affairs. Currently, the shaping of energy-related policies is becoming essential, in the view of new societies. Given the Sustainable Development Goals, nations are recently bounded by uniform strategies and initiatives in favour of the energy-poor population. Overall, international energy agencies and governments seek to address energy-related issues that threaten sustainability. To that end, economies set energy policies that promote equity, social rights and inclusion. Howbeit, these policies may vary across countries, i.e. uniform policies do not sufficiently fit each county’s needs. In other terms, the effects of energy poverty between the European households significantly vary, a fact that is potentially attributed to country-specific, political-specific or geographical-specific characteristics. Hence, the understanding of the transitional patterns of energy poverty is essential for policymaking and decisions.

In this study, we investigate the dynamic patterns of energy poverty among 27 European economies between 2005 and 2020. We follow the study of Phillips and Sul ( 2007 ) to examine whether countries converge to an equilibrium steady state or converge to different steady states. The results of the data-driven algorithm suggest the existence of three convergence groups of the population unable to keep home adequately warm showcase. We consider that the affordability of heating services is adequately explained by structural conditions of housing, weather conditions (i.e. temperature outcomes) and energy costs. Besides, we find a greater number of convergence clubs in terms of the arrears on utility bills. The adverse financial conditions, followed by the Great Recession, have caused severe effects on employment and disposable income; hence, arrears on utility bills across Croatian, Romanian, Spanish, Bulgarian, Cypriot, Irish and Slovakian households are triggered. Economies such as Belgium, Sweden, Finland, Czechia and German prosper in terms of energy poverty, due to effective disconnection protection and income schemes, as well as building renovation schemes. Interestingly, the results of material deprivation on the housing dimension suggest the formation of two convergence clubs. We consider that a large part of the European population suffers severely from structural problems, i.e. the deficiency of basic sanitary facilities and other issues on infrastructural terms. Ultimately, we infer that a significant proportion of households struggle with energy-related issues that affect their living conditions and well-being. Several disconnection and income burdens significantly affect the distribution of energy across households and the share of population at risk of poverty. We suggest that the analysis of convergence hypothesis is a precondition of shaping effective energy-related policies between specific groups of the European countries.

Policymakers and international organizations of energy should orient future strategies towards buildings renovation and income schemes for the reconstruction of dwelling that lack basic sanitary facilities. Precisely, economies like Portugal, Italy, Bulgaria, Romania, Latvia, Hungary, Lithuania and Greece are among the EU members with the highest rates of severe housing deprivation. To that end, policymakers should adopt long-term measures to protect vulnerable consumers and guarantee their access to basic energy resources. The renovation of old buildings which have poor thermal characteristics or suffer from leaking roofs, lack of bath or toilet and damaged walls is essential to improve the efficiency in dwellings. The investments in green energy and the effective exploitation of innovative technologies in buildings renovation are fundamental mechanisms that can improve the efficient use of energy within households. The construction of sanitary facilities entails the access to energy for all households, eliminates the energy inequalities and protects the people who are at risk of poverty. In view of the building renovation, low-income households who are in significant arrears on utility bills will lower the average cost of energy and increase their disposable income. Hence, these measures are not income related, i.e. they do not provide budget support in households. Instead, they are long-term schemes which will potentially improve the residential use of energy and the energy efficiency in the household sector. So, we suggest that the buildings renovation for the aforementioned economies is essential for governments to improve the living standards of the vulnerable population.

However, it is significant to highlight the necessity for short-term measures which are income related. To be more precise, people who reside in countries or regions that face extreme climatic conditions, i.e. heat or cold waves, are hit hard during summer or winter periods. Nevertheless, this is mainly evident to poor households who lack thermal or cooling services. Verily, these waves have detrimental effects on human health and life expectancy. Southeast and Eastern European countries face hot and dry summer, which implies that individuals who are unable to afford for cooling services are at risk of severe health problems. Cyprus, Spain, Portugal, Bulgaria and Lithuania are the most vulnerable members which are hit by continental warm summer and cold winter, implying the emergent need for the governments’ action. Governments should grant heating or cooling allowances to vulnerable consumers or social tariffs for electricity. The provision of income support schemes is essential for the replacement of ineffective appliances or the access to cooling or heating services. Besides, the use of renewables and effective allowances not only protects socially vulnerable consumers, but it also improves the social welfare and sustainable development and mitigates the greenhouse gas emissions.

Moreover, we recommend that information and awareness measures towards the effective use of energy will potentially contribute to the energy resources services. Instead, energy inequalities may be misleading due to the variation or even the lack in the definitions of vulnerable consumers and energy poverty. Inconsistent results and data collection on energy poverty indicators adversely affect policy measures and the initiatives, as well as the population that receive support by governments. Hence, we infer that the European Member States should adopt uniform and flexible measures of energy poverty, to protect vulnerable consumers by means of the access and affordability of energy services and well-being. Conclusively, energy poverty alleviation is a profound dimension for transition and welfare economies.

Finally, a major limitation of this paper is the lack of the investigation of potential factors that drive energy poverty outcomes. The examination of these factors across convergence clubs would provide sufficient information in terms of shaping country-specific policies. Therefore, future research on the determinants of energy poverty convergence clubs could prove beneficial to the energy poverty research.

Data availability

Our manuscript contains data, which will be made available on reasonable request.

Ethical approval and consent to participate

The research carried out in this work was taken into account with respect to the observance of all the rules of ethics that govern the conduct of such research.

Evaluation of the EU’s regulatory framework for electricity market design and consumer protection in the fields of electricity and gas ( https://eur-lex.europa.eu/resource.html?uri=cellar:20674470-b7b9-11e6-9e3c-01aa75ed71a1.0001.02/DOC_1&format=PDF ).

Akram, V., Rath, B. N., & Sahoo, P. K. (2020). Stochastic conditional convergence in per capita energy consumption in India. Economic Analysis and Policy, 65 , 224–240.

Article   Google Scholar  

Anastasiou, A. (2009). Central bank independence and economic performance. Cyprus Economic Policy Review, 3 (1), 123–156.

Google Scholar  

Arbolino, R., & Di Caro, P. (2021). Can the EU funds promote regional resilience at time of Covid-19? Insights from the Great Recession. Journal of Policy Modeling, 43 (1), 109–126.

Bajomi, A. Z., Feldmár, N., & Kőszeghy, L. (2020). Trapped in politics: Energy poverty in Hungary. In: Perspectives on energy poverty in post-communist Europe (pp. 25–54). Routledge

Baker, S. H., Carley, S., & Konisky, D. M. (2021). Energy insecurity and the urgent need for utility disconnection protections. Energy Policy, 159 , 112663.

Baldwin, R., & Di Mauro, B. W. (2020). Economics in the time of COVID-19: A new eBook. VOX CEPR Policy Portal, 2–3

Ballesteros-Arjona, V., Oliveras, L., Munoz, J. B., de Labry Lima, A. O., Carrere, J., Ruiz, E. M., ... & Mari-Dell'Olmo, M. (2022). What are the effects of energy poverty and interventions to ameliorate it on people’s health and well-being?: A scoping review with an equity lens. Energy Research & Social Science, 87, 102456.

Barnes, D. F., Khandker, S. R., & Samad, H. A. (2011). Energy poverty in rural Bangladesh. Energy Policy, 39 (2), 894–904.

Barro, R. J. (1991). Economic growth in a cross section of countries. The Quarterly Journal of Economics, 106 (2), 407–443.

Barro, R. J., & Sala-i-Martin, X. (1992). Convergence. Journal of Political Economy, 100 (2), 223–251.

Barro, R. J., Sala-i-Martin, X., Blanchard, O. J., & Hall, R. E. (1991). Convergence across states and regions. Brookings papers on economic activity, 107–182

Beg, N., Morlot, J. C., Davidson, O., Afrane-Okesse, Y., Tyani, L., Denton, F., & Rahman, A. A. (2002). Linkages between climate change and sustainable development. Climate Policy, 2 (2–3), 129–144.

Belaïd, F. (2022). Implications of poorly designed climate policy on energy poverty: Global reflections on the current surge in energy prices. Energy Research & Social Science, 92 , 102790.

Biernat-Jarka, A., Trębska, P., & Jarka, S. (2021). The role of renewable energy sources in alleviating energy poverty in households in Poland. Energies, 14 (10), 2957.

Bigerna, S., Bollino, C. A., & Polinori, P. (2021). Convergence in renewable energy sources diffusion worldwide. Journal of Environmental Management, 292 , 112784.

Bollino, C. A., & Botti, F. (2017). Energy poverty in Europe: A multidimensional approach. PSL Quarterly Review, 70(283)

Bouzarovski, S., & Tirado Herrero, S. (2017). The energy divide: Integrating energy transitions, regional inequalities and poverty trends in the European Union. European Urban and Regional Studies, 24 (1), 69–86.

Bouzarovski, S., Thomson, H., Cornelis, M., Varo, A., & Guyet, R. (2020). Towards an inclusive energy transition in the European Union: Confronting energy poverty amidst a global crisis . European Commission: Brussels, Belgium.

Bouzarovski, S., Petrova, S., & Tirado-Herrero, S. (2014). From fuel poverty to energy vulnerability: The importance of services, needs and practices

Brucal, A. & McCoy, D., Social situation monitor, energy prices, energy policies and energy poverty: Exploring the connection. Publications Office of the European Union, Luxembourg, 2021, ISBN 978-92-76-28932-6, doi: https://doi.org/10.2767/85296 , KE-02-21-049-EN-N

Carfora, A., Scandurra, G., & Thomas, A. (2022). Forecasting the COVID-19 effects on energy poverty across EU member states. Energy Policy, 161 , 112597.

Carlino, G. A., & Mills, L. O. (1993). Are US regional incomes converging?: A time series analysis. Journal of Monetary Economics, 32 (2), 335–346.

Carlo Andrea Bollino & Fabrizio Botti. (2017). Energy poverty in Europe: A multidimensional approach. PSL Quarterly Review, Economia Civile, 70 (283), 449–472.

Carrion-i-Silvestre, J. L., & German-Soto, V. (2009). Panel data stochastic convergence analysis of the Mexican regions. Empirical Economics, 37 (2), 303–327.

Castaño-Rosa, R., Solís-Guzmán, J., Rubio-Bellido, C., & Marrero, M. (2019). Towards a multiple-indicator approach to energy poverty in the European Union: A review. Energy and Buildings, 193 , 36–48.

Charlier, D., & Legendre, B. (2019). A multidimensional approach to measuring fuel poverty. The Energy Journal, 40(2)

Cheng, Z., Liu, J., Li, L., & Gu, X. (2020). Research on meta-frontier total-factor energy efficiency and its spatial convergence in Chinese provinces. Energy Economics, 86 , 104702.

Clark, I. K. H., Chun, S., O’Sullivan, K. C., & Pierse, N. (2021). Energy poverty among tertiary students in Aotearoa New Zealand. Energies, 15 (1), 76.

Conte, A., Lecca, P., Salotti, S., & Sakkas, S. (2020). The territorial economic impact of COVID-19 in the EU. A RHOMOLO analysis. Territorial Development Insights Series, JRC121261, European Commission.

Costantini, M., & Lupi, C. (2005). Stochastic convergence among European economies. Economics Bulletin, 3 (38), 1–17.

Day, R., Walker, G., & Simcock, N. (2016). Conceptualising energy use and energy poverty using a capabilities framework. Energy Policy, 93 , 255–264.

Dayrit, M. M., & Mendoza, R. U. (2020). Social cohesion vs COVID-19. International Journal of Health Governance

Deller, D., Turner, G., & Price, C. W. (2021). Energy poverty indicators: Inconsistencies, implications and where next? Energy Economics, 103 , 105551.

Dobbins, A., Pye, S., Baffert, C., Brajković, J., Grgurev, I., De Miglio, R., & Deane, P. (2015). Energy poverty and vulnerable consumers in the energy sector across the EU: Analysis of policies and measures. Dobbins, S. Pye, C. Baffert, J. Brajković, I. Grgurev, R. De Miglio, P. Deane//Policy Report, 2

Dubois, U., & Meier, H. (2016). Energy affordability and energy inequality in Europe: Implications for policymaking. Energy Research & Social Science, 18 , 21–35.

Durlauf, S. N. (2003). The convergence hypothesis after 10 years . University of Wisconsin.

Evans, P. (1998). Using panel data to evaluate growth theories. International Economic Review, 295–306

González-Eguino, M. (2015). Energy poverty: An overview. Renewable and Sustainable Energy Reviews, 47 , 377–385.

Hajdinjak, M., & Asenova, D. (2019). Sustainable energy consumption and energy poverty: Challenges and trends in Bulgaria. In: Energy demand challenges in Europe: Implications for policy, planning and practice, 115–126.

Halkos, G. E., & Gkampoura, E. C. (2021a). Evaluating the effect of economic crisis on energy poverty in Europe. Renewable and Sustainable Energy Reviews, 144 , 110981.

Halkos, G. E., & Gkampoura, E. C. (2021b). Coping with energy poverty: Measurements, drivers, impacts, and solutions. Energies, 14 (10), 2807.

Hanke, F., Guyet, R., & Feenstra, M. (2021). Do renewable energy communities deliver energy justice? Exploring insights from 71 European cases. Energy Research & Social Science, 80 , 102244.

He, X., Khan, S., Ozturk, I., & Murshed, M. (2022). The role of renewable energy investment in tackling climate change concerns: Environmental policies for achieving SDG‐13. Sustainable Development.

Hesselman, M., Varo, A., Guyet, R., & Thomson, H. (2021). Energy poverty in the COVID-19 era: Mapping global responses in light of momentum for the right to energy. Energy Research & Social Science, 81 , 102246.

Horta, I. M., & Camanho, A. S. (2015). A nonparametric methodology for evaluating convergence in a multi-input multi-output setting. European Journal of Operational Research, 246 (2), 554–561.

Article   MATH   Google Scholar  

Huang, Z., Ming, X., & Duan, H. (2022). Heterogeneous effects of energy poverty convergence in Europe. Energy Strategy Reviews, 41 , 100822.

Karpinska, L., & Śmiech, S. (2020). Invisible energy poverty? Analysing housing costs in Central and Eastern Europe. Energy Research & Social Science, 70 , 101670.

Kim, Y. S. (2015). Electricity consumption and economic development: Are countries converging to a common trend? Energy Economics, 49 , 192–202.

Article   MathSciNet   Google Scholar  

Krugman, P. (1981). Trade, accumulation, and uneven development. Journal of Development Economics, 8 (2), 149–161.

Lee, K., Pesaran, M. H., & Smith, R. (1997). Growth and convergence in a multi-country empirical stochastic Solow model. Journal of Applied Econometrics, 12 (4), 357–392.

Liu, X., & Li, H. (2018). Regional differences and convergence analysis of green energy efficiency in China. Journal of Power and Energy Engineering, 6 (6), 48–58.

Lucas, Karen & Mattioli, Giulio & Verlinghieri, Ersilia & Guzman, Alvaro. (2016). Transport poverty and its adverse social consequences. Transport. 169. https://doi.org/10.1680/jtran.15.00073 .

Mafalda Matos, A., Delgado, J. M., & Guimarães, A. S. (2022). Linking energy poverty with thermal building regulations and energy efficiency policies in Portugal. Energies, 15 (1), 329.

Mankiw, N. G., Romer, D., & Weil, D. N. (1992). A contribution to the empirics of economic growth. The Quarterly Journal of Economics, 107 (2), 407–437.

Markandya, A., Pedroso-Galinato, S., & Streimikiene, D. (2006). Energy intensity in transition economies: Is there convergence towards the EU average? Energy Economics, 28 (1), 121–145.

Markandya, A., Pedroso, S., & Streimikiene, D. (2004). Energy efficiency in transition economies: Is there convergence towards the EU average?. Available at SSRN 556237.

Mishra, V., & Smyth, R. (2017). Conditional convergence in Australia’s energy consumption at the sector level. Energy Economics, 62 , 396–403.

Neacsa, A., Panait, M., Muresan, J. D., & Voica, M. C. (2020). Energy poverty in European Union: Assessment difficulties, effects on the quality of life, mitigation measures. Some Evidences from Romania. Sustainability, 12 (10), 4036.

Nussbaumer, P., Bazilian, M., & Modi, V. (2012). Measuring energy poverty: Focusing on what matters. Renewable and Sustainable Energy Reviews, 16 (1), 231–243.

Phebe Asantewaa Owusu & Samuel Asumadu-Sarkodie | Shashi Dubey (Reviewing Editor) (2016) A review of renewable energy sources, sustainability issues and climate change mitigation. Cogent Engineering, 3:1.  https://doi.org/10.1080/23311916.2016.1167990

Pachauri, S., & Rao, N. D. (2020). Advancing energy poverty measurement for SDG7. Progress in Energy, 2 (4), 043001.

Pan, L., & Maslyuk-Escobedo, S. (2019). Stochastic convergence in per capita energy consumption and its catch-up rate: Evidence from 26 African countries. Applied Economics, 51 (24), 2566–2590.

Pan, X., Liu, Q., & Peng, X. (2015). Spatial club convergence of regional energy efficiency in China. Ecological Indicators, 51 , 25–30.

Papada, L., & Kaliampakos, D. (2018). A stochastic model for energy poverty analysis. Energy Policy, 116 , 153–164.

Pesaran, M. H. (2007). A pair-wise approach to testing for output and growth convergence. Journal of Econometrics, 138 (1), 312–355.

Article   MathSciNet   MATH   Google Scholar  

Phillips, P. C., & Sul, D. (2003). The elusive empirical shadow of growth convergence. Available at SSRN 384923.

Phillips, P. C., & Sul, D. (2007). Transition modeling and econometric convergence tests. Econometrica, 75 (6), 1771–1855.

Phillips, P. C., & Sul, D. (2009). Economic transition and growth. Journal of Applied Econometrics, 24 (7), 1153–1185.

Primc, K., Slabe-Erker, R., & Majcen, B. (2019). Energy poverty: A macrolevel perspective. Sustainable Development, 27 (5), 982–989.

Quah, D. T. (1996). Twin peaks: Growth and convergence in models of distribution dynamics. The Economic Journal, 106 (437), 1045–1055.

Rademaekers, K., Yearwood, J., Ferreira, A., Pye, S. T., Hamilton, I., Agnolucci, P., ... & Anisimova, N. (2016). Selecting indicators to measure energy poverty.

Romero, J. C., Linares, P., & López, X. (2018). The policy implications of energy poverty indicators. Energy Policy, 115 , 98–108.

Saba, C. S., & Ngepah, N. (2022a). Convergence in renewable energy consumption and their influencing factors across regions: Evidence from convergence algorithm approach. Environmental Science and Pollution Research, 29 (40), 61412–61445.

Saba, C. S., & Ngepah, N. (2022b). Convergence in renewable energy sources and the dynamics of their determinants: An insight from a club clustering algorithm. Energy Reports, 8 , 3483–3506.

Salman, M., Zha, D., & Wang, G. (2022). Assessment of energy poverty convergence: A global analysis. Energy, 255 , 124579.

Samarakoon, S. (2019). A justice and wellbeing centered framework for analysing energy poverty in the Global South. Ecological Economics, 165 , 106385.

Santamouris, M., & Kolokotsa, D. (2015). On the impact of urban overheating and extreme climatic conditions on housing, energy, comfort and environmental quality of vulnerable population in Europe. Energy and Buildings, 98 , 125–133.

Sareen, S., Thomson, H., Herrero, S. T., Gouveia, J. P., Lippert, I., & Lis, A. (2020). European energy poverty metrics: Scales, prospects and limits. Global Transitions, 2 , 26–36.

Siksnelyte-Butkiene, I. (2022). Combating energy poverty in the face of the COVID-19 pandemic and the global economic uncertainty. Energies, 15 (10), 3649.

Solarin, S. A. (2019). Parametric and non-parametric convergence analysis of electricity intensity in developed and developing countries. Environmental Science and Pollution Research, 26 (9), 8552–8574.

Solow, R. M. (1956). A contribution to the theory of economic growth. The Quarterly Journal of Economics, 70 (1), 65–94.

Sy, S. A., & Mokaddem, L. (2022). Energy poverty in developing countries: A review of the concept and its measurements. Energy Research & Social Science, 89 , 102562.

Szép, T. S. (2016). Energy convergence of the European Union toward 2020. The Central European Journal of Regional Development and Tourism, 88.

Thomson, H., & Snell, C. (2013). Quantifying the prevalence of fuel poverty across the European Union. Energy Policy, 52 , 563–572.

Thomson, H., Simcock, N., Bouzarovski, S., & Petrova, S. (2019). Energy poverty and indoor cooling: An overlooked issue in Europe. Energy and Buildings, 196 , 21–29.

Thomson, H., & Bouzarovski, S. (2018). Addressing energy poverty in the European Union: State of play and action. EU Energy Poverty Observatory, Manchester.

Thomson, H. & Bouzarovski, S., Transforming the measurement and mitigation of energy poverty. Executive summary of the European Union Energy Poverty Observatory. Publications Office of the European Union, Luxembourg, 2020, ISBN 978–92–76–20942–3, doi: https://doi.org/10.2833/758612 , MJ-02–20–601-EN-N.

Trinomics. Selecting indicators to measure energy poverty. Available online: https://ec.europa.eu/energy/sites/ener/files/documents/Selecting%20Indicators%20to%20Measure%20Energy%20Poverty.pdf (accessed on 22 October 2020).

Turai, E., Schmatzberger, S., & Broer, R. (2021). Overview report on the energy poverty concept. Energy poverty in the privately-owned, multi-family environment.

Zervoyianni, A., & Anastasiou, A. (2007). Symmetry of shocks and trade integration in the enlarged European Union. Global Business & Economics Anthology, I , 47–61.

Zervoyianni, A., & Anastasiou, A. (2009). Convergence of shocks and trade in the enlarged European Union. Journal of International Trade and Economic Development, 18 (1), 79–114.

Zervoyianni, A., Anastasiou, A., & Anastasiou, G. A. (2014). Does central bank independence really matter? Re-assessing the role of the independence of monetary policy-makers in macroeconomic outcomes. International Journal of Economics and Business Research, 8 (4), 427–473.

Zhang, W., Pan, X., Yan, Y., & Pan, X. (2017). Convergence analysis of regional energy efficiency in china based on large-dimensional panel data model. Journal of Cleaner Production, 142 , 801–808.

Zhang, D., & Broadstock, D. C. (2016). Club convergence in the energy intensity of China. The Energy Journal, 37(3)

Download references

Open access funding provided by HEAL-Link Greece.

Author information

Authors and affiliations.

Laboratory of Data Science and Digital Transformation, Department of Management Science and Technology, University of Peloponnese, Tripoli, Greece

Athanasios Anastasiou & Eftychia Zaroutieri

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Athanasios Anastasiou .

Ethics declarations

Conflict of interest.

The authors declare no competing interests.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Anastasiou, A., Zaroutieri, E. Energy poverty and the convergence hypothesis across EU member states. Energy Efficiency 16 , 38 (2023). https://doi.org/10.1007/s12053-023-10113-9

Download citation

Received : 26 October 2022

Accepted : 09 April 2023

Published : 05 May 2023

DOI : https://doi.org/10.1007/s12053-023-10113-9

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

  • Log- t regression test

JEL Classification

  • Find a journal
  • Publish with us
  • Track your research

IMAGES

  1. nuLL modeL. dePendent vAriAbLe: Poverty At time t

    null hypothesis about poverty

  2. Null hypothesis for spearmans rho

    null hypothesis about poverty

  3. Difference Between Absolute and Relative Poverty(With Table)

    null hypothesis about poverty

  4. The Null Hypothesis and Charter Schools

    null hypothesis about poverty

  5. 15 Null Hypothesis Examples (2024)

    null hypothesis about poverty

  6. Solved Suppose Jamal is studying poverty in his economics

    null hypothesis about poverty

VIDEO

  1. What's a Null Hypothesis, an Alternative Hypothesis, and Statistical Power

  2. How Poverty Occurs

  3. Null vs Alternative Hypothesis

COMMENTS

  1. Theories of the Causes of Poverty

    Thirty years later, poverty research continues to lack clear theory. There are few explicitly named theories and little explicit theoretical debate. If a theory of the causes of poverty is articulated, it is usually only compared against the null hypothesis of no effect. Studies rarely compare two or more theories against each other.

  2. 9.1: Null and Alternative Hypotheses

    Review. In a hypothesis test, sample data is evaluated in order to arrive at a decision about some type of claim.If certain conditions about the sample are satisfied, then the claim can be evaluated for a population. In a hypothesis test, we: Evaluate the null hypothesis, typically denoted with \(H_{0}\).The null is not rejected unless the hypothesis test shows otherwise.

  3. The Social Consequences of Poverty: An Empirical Test on Longitudinal

    Abstract. Poverty is commonly defined as a lack of economic resources that has negative social consequences, but surprisingly little is known about the importance of economic hardship for social outcomes. This article offers an empirical investigation into this issue. We apply panel data methods on longitudinal data from the Swedish Level-of ...

  4. Full article: The impact of poverty cycles on economic research

    Poverty is a global problem, and fighting it is a historical task faced by humanity. The outbreak of poverty crises has a certain cyclicality, and the development of economic research may react to poverty cycles. ... If the null hypothesis is rejected, it indicates that there exists a Granger causality between the economic variables and the ...

  5. Spatial Regression Analysis of Poverty in R

    Poverty has been studied across many social science disciplines, resulting in a large body of literature. Scholars of poverty research have long recognized that the poor are not uniformly distributed across space. ... The p-value is 1 out of 1000, which is 0.001 for a null hypothesis that there is no spatial dependence versus a one-sided ...

  6. Inequality, poverty and economic growth

    López (2006) endorses the hypothesis that poverty retards growth through various channels including education, institutions, health and physical capital accumulation. ... or instrument proliferation. The null hypothesis of the test is that instruments are exogenous. When the p-value is small, we reject the null hypothesis and conclude that the ...

  7. The Impact of Education and Culture on Poverty Reduction: Evidence from

    The study also tests which is better between the OLS and RE model using the Breusch-Pagan (BP)-Langragian Multiplier (LM) test. The null hypothesis of the BP-LM test is that there is no substantial variance between regions. A probability value of less than 5% for the BP-LM test indicates that the RE model is appropriate and the OLS pool is not.

  8. Alleviating time poverty among the working poor: a pre-registered

    Poverty entails more than a scarcity of material resources—it also involves a shortage of time. ... we found strong evidence in support of the null hypothesis that there were no differences in ...

  9. Null & Alternative Hypotheses

    The null and alternative hypotheses offer competing answers to your research question. When the research question asks "Does the independent variable affect the dependent variable?": The null hypothesis ( H0) answers "No, there's no effect in the population.". The alternative hypothesis ( Ha) answers "Yes, there is an effect in the ...

  10. The Effects of Poverty on Child Health and Development

    Poverty has been shown to negatively influence child health and development along a number of dimensions. For example, poverty-net of a variety of potentially confounding factors—is associated with increased neonatal and post-neonatal mortality rates, greater risk of injuries resulting from accidents or physical abuse/neglect, higher risk for asthma, and lower developmental scores in a ...

  11. Testing an hypothesis about differences in poverty estimated from

    Get rights and content. Abstract. This paper develops procedures for testing an hypothesis about differences in poverty estimated from grouped data. The methodology developed in the paper is applied to analyze changes in poverty in India between 1983 and 1986-1987. Previousarticle in issue.

  12. Null Hypothesis: Definition, Rejecting & Examples

    When your sample contains sufficient evidence, you can reject the null and conclude that the effect is statistically significant. Statisticians often denote the null hypothesis as H 0 or H A.. Null Hypothesis H 0: No effect exists in the population.; Alternative Hypothesis H A: The effect exists in the population.; In every study or experiment, researchers assess an effect or relationship.

  13. Recent origin and evolution of obesity-income correlation across the

    If the null hypothesis of non-linearity could not be rejected, by implication an S-curve would not be present in the data. ... (2017) Investigating the poverty-obesity paradox in Europe. Econ ...

  14. PDF Mexico: Poverty estimates 2000-2002 Test of hypothesis

    Summary. Hypothesis testing was made on the poverty estimates developed for the period 2000-2002 by the Mexico's Technical Committee on Poverty Measurement, for three defined concepts of poverty: food, capacities and patrimony poverty, including the three geographic levels: national, urban and rural. For this purpose, the National Income and ...

  15. Inequalities in Poverty and Income between Single Mothers and Fathers

    Thus, the null hypothesis can often be rejected in cases with very large sample sizes . ... The poverty threshold criterion was from the US Census Bureau definition based on family size and the number of individuals in the family under the age of 18 . Furthermore, similar to Kramer et al. ...

  16. PDF Financial Development, Economic Growth and Poverty Reduction in India

    The above equation indicates that its past values influence the level of poverty. The null hypothesis is that there is no co-integration vector in the model that is H 0: ¥ 1 =0, ¥ 2 =0, ¥ 3 =0 whereas, the alternative hypothesis that a long-run relationship exists H 1: ¥ 1 ≠0, ¥ 2 ≠0, ¥ 3 ≠0.

  17. How to Write a Null Hypothesis (5 Examples)

    Whenever we perform a hypothesis test, we always write a null hypothesis and an alternative hypothesis, which take the following forms: H0 (Null Hypothesis): Population parameter =, ≤, ≥ some value. HA (Alternative Hypothesis): Population parameter <, >, ≠ some value. Note that the null hypothesis always contains the equal sign.

  18. What Is The Null Hypothesis & When To Reject It

    A null hypothesis is rejected if the measured data is significantly unlikely to have occurred and a null hypothesis is accepted if the observed outcome is consistent with the position held by the null hypothesis. Rejecting the null hypothesis sets the stage for further experimentation to see if a relationship between two variables exists.

  19. Contributions of Community Education in the Eradication of Poverty

    This is a descriptive survey research study that sets to examine the extent community education has contributed to the eradication of poverty among communities in Rivers State of Nigeria. Two research questions and one null hypothesis were developed to guide the study. The population of the study was 1,672 respondents made up of 270 married men, 517 married women and 885 youths in there ...

  20. Out-of-Pocket Health Expenditure and Poverty: Evidence from a Dynamic

    When the same poverty models were estimated for the Eastern Mediterranean region, it was observed that the effect of out-of-pocket health spending on poverty was statistically insignificant for both β ^ 1 and β ^ 2, although the three models rejected the null hypothesis of no threshold effect. Moreover, most of the explanatory variables were ...

  21. Childhood Poverty, Adverse Childhood Experiences, and Adult ...

    This study aimed to consider childhood poverty in relation to a count measure of adverse childhood experiences (ACEs) as a predictor of adult health outcomes and to determine whether associations are sensitive to how childhood poverty is operationalized. A sample of 10,784 adult residents was derived using data 2014-2015 Wisconsin annual ...

  22. Energy poverty and the convergence hypothesis across EU ...

    Energy poverty is an emerging issue towards global affairs. Currently, the development of energy-related policies is becoming essential, with regard to new societies, social inclusion and social rights. In this paper, we examine the dynamic patterns of energy poverty among 27 EU member states between 2005 and 2020. We use the log-t regression test to investigate the convergence hypothesis, and ...

  23. The Impact of Microfinance Institutions on Poverty Alleviation

    The coefficient of poverty in model 2 is positive and statistically significant, indicating that an increase in poverty leads to a rise in microfinancing as people seek ways to survive scarceness. The more people become poor, the more they cannot access the banking channels and resort to microfinance institutions.