Tag: Health

Alcohol-Related Costs and Potential Gains from Prevention Measures in Latvia

Posted on February 26, 2024 | Policy Brief

20240226 Alcohol Prevention Measures Image 01

Latvia has the highest per capita registered alcohol consumption rate among EU and OECD countries (OECD, 2024). In this brief, we show that the total budgetary (direct) and non-budgetary (indirect) costs associated with alcohol consumption in Latvia in 2021 amounted to 1.3–1.8 percent of the GDP. Non-financial costs from alcohol abuse amounted to a loss of nearly 90 thousand years spent in good health and with a good quality of life. We assess the potential effects of five alcohol misuse prevention measures, all recognized by the World Health Organization (WHO) as effective in reducing harmful alcohol consumption – especially when implemented together. Our analysis focuses on the individual effects of each measure and shows that raising the minimum legal age for alcohol purchases and enforcing restrictions on alcohol advertising and marketing are likely to yield the largest reductions in alcohol-related costs, although these effects will take time to fully materialize.

Introduction

Alcohol consumption is an important risk factor for morbidity and premature death worldwide. It is associated with over 200 diagnoses recorded in the International Statistical Classification of Diseases and Related Health Problems (CDC, 2021), including liver diseases, injuries, malignancies, and diseases of the heart and circulatory system (WHO, 2018). Alcohol consumption at any level is considered unsafe (Burton & Sheron, 2018).

Globally, an average of 3 million people die each year due to alcohol-related harm, accounting for 5.3 percent of all deaths (Shield et al., 2020). In 2019, alcohol consumption was the main risk factor for disease burden in people between 25 and 49 years of age and the second most important risk factor in people aged 10-24 years (GDB, 2019).

Alcohol use is associated not only with health problems but also with social issues, posing risks to people’s safety and well-being. It causes harm not only to the individual but also to family members and society at large (Rehm & Hingson, 2013). Various sectors, including health, justice, home affairs, and social care agencies, are involved in preventing the consequences of alcohol misuse and reducing the harm this causes. This demonstrates the multiple negative impacts of alcohol use on public health and well-being (Flynn & Wells, 2013).

Latvia has the highest per capita registered alcohol consumption rate among the EU and OECD countries (OECD, 2024), and no clear trend of declining levels has been observed in recent years. Moreover, the consumption of spirits, which can potentially cause more harm than other alcoholic beverages (Mäkelä et al., 2011), is steadily increasing. According to WHO data (WHO, 2024), the high per capita consumption of registered absolute alcohol in Latvia, compared to other countries, is largely due to the consumption of spirits. In Latvia, the share of spirits in total consumption is around 40 percent. By comparison, in the Czech Republic and Austria, where total per capita alcohol consumption is similar to Latvian levels, spirits account for only 25 and 16 percent of total consumption, respectively, while the proportions of beer and wine are higher.

This policy brief reports the estimated costs related to alcohol use in Latvia in 2021, based on the study Alcohol Use, its Consequences, and the Economic Benefits of Prevention Measures (Pļuta et al., 2023). It also provides an overview of the expected benefits from implementing preventive measures, such as raising the minimum legal age for buying alcohol and restricting alcohol advertisements.

Costs of Alcohol Use in Latvia

We estimate three types of costs associated with alcohol consumption:

  • Direct costs: These include budgetary costs related to alcohol consumption, such as healthcare, law enforcement and social assistance costs, as well as expenses for public education.
  • Indirect costs: These costs represent unproduced output in the economy and arise from the premature deaths of alcohol users, as well as their reduced employment or lower productivity.
  • Non-financial welfare costs: This type of cost arises from the compromised quality of life of alcohol users, their families, and friends.

We estimate direct costs by utilizing detailed disaggregated data on alcohol-related budget costs in the healthcare sector, law enforcement institutions (including police, courts, and prisons), costs of public education (e.g., educating schoolchildren about the consequences of alcohol consumption), costs of awareness-raising campaigns, and social assistance costs. For cost categories that are only partially attributable to alcohol consumption, we classify only a fraction of these costs as attributable to alcohol use (e.g., liver cirrhosis is attributable to alcohol usage in 69.8 percent of the cases, so only this fraction of the budget costs on compensated medicaments is attributable to alcohol use). To estimate social assistance costs, including public expenditure on social services, sobering-up facilities, social care centres, orphanages, and specialized care facilities for children and out-of-family care, we conduct a survey among social assistance providers.

To estimate non-budgetary costs, we construct a counterfactual scenario where alcohol is not being overly consumed, ensuring higher productivity, a lower rate of unemployment, and lower mortality within the labour force. Finally, non-financial welfare costs are estimated by measuring the reduction in quality of life or QALYs lost (quality-adjusted-life-years) (for details, see the methodology section in Pļuta et al. (2023)).

The total direct and indirect costs of alcohol abuse in 2021 amounted to 1.3–1.8 percent of Latvia’s GDP. In comparison, revenues from the excise tax on alcoholic beverages in 2021 accounted for 0.7 percent of the GDP.

Direct costs, which entail expenses directly covered by the state budget, comprised 0.45 percent of the GDP. Among these costs, healthcare expenses were the largest component, constituting 37.8 percent  of total direct costs and 2.7 percent of general government spending on healthcare. Nearly half of these healthcare costs were attributed to the provision of inpatient hospital treatment for patients diagnosed with alcohol-related conditions. Another significant component of budgetary costs is associated with addressing alcohol abuse and combating illicit trade through law enforcement, accounting for 31.9 percent of total direct costs and 6.5 percent of general government spending on public order and safety.

Alcohol-related indirect costs amount to 0.9-1.3 percent of Latvia’s GDP. Despite not being directly covered by the state budget, they represent unproduced output and thus entail economic losses. The primary components of these indirect costs are linked to decreased output resulting from higher unemployment and reduced economic activity (0.6-0.8 percent of the GDP), as well as decreased output due to premature death among heavy drinkers (0.2-0.4 percent of the GDP). Notably, indirect costs attributed to alcohol misuse by males constitute almost two-thirds of the total indirect costs.

Finally, the non-financial costs from alcohol abuse in 2021 are estimated to reach 88 620 years spent in good health and with a good quality of life. These losses primarily stem from the distress experienced by household members from alcohol users, the decline in the quality of life among alcohol users themselves, and the premature mortality of such individuals.

The Effects of Preventive Measures

We consider five alcohol misuse preventive measures, all of which are included in the list of WHO “best buys” policies that effectively reduce alcohol consumption (WHO, 2017):

  • Reducing the availability of retail alcohol by tightening restrictions on on-site retail hours
  • Raising the minimum legal age for alcohol purchase from 18 to 20 years
  • Increasing excise tax on alcohol
  • Lowering the maximum allowed blood alcohol concentration limit for all drivers from 0.5 to 0.2 per mille (currently 0.2 for new drivers and 0.5 for all other drivers)
  • Restricting alcohol advertising and marketing

Our estimates of the expected reduction in alcohol-related costs resulting from these measures are based on two main components:

  • (1) our own estimates of alcohol-related costs in Latvia, as described above, and
  • (2) external estimates of the impact of the five misuse preventative measures on alcohol consumption derived from existing literature on other countries.

We then apply these external estimates to the calculated alcohol-related costs and Latvian data on alcohol consumption to determine the estimated impact for Latvia (for further details, see the methodology outlined in Pluta et al. (2023)).

Our findings indicate that the most substantial reduction in direct costs attributed to alcohol misuse is anticipated through raising the minimum alcohol purchase age to 20 years (yielding an 11.4-15.8 percent estimated cost reduction). Previous literature has shown that early initiation of alcohol use significantly increases the likelihood of risky drinking, and that risky drinking during adolescence significantly increases the risk of heavy drinking in adulthood (Betts et al., 2018; McCarty, 2004). Hence, raising the minimum legal age for alcohol purchase represents an effective tool to reduce alcohol consumption also among the adult population.

Another highly effective measure to reduce alcohol consumption is imposing restrictions on advertising, which results in a 5.0-8.0 percent estimated reduction of direct costs. There is a large body of literature indicating that alcohol advertising increases alcohol consumption among young people, as well as significantly increases the likelihood of alcohol initiation among adolescents and young adults (Noel, 2019; Jernigan et al., 2017). Also, among the adult population, alcohol consumption decreases with stricter advertising restrictions (see Casswell, 2022; Rossow, 2021).

However, it is important to emphasize that the full impact of both above discussed preventative measures will only manifest in the long run.

The Effect of Illicit Markets

It is often argued that illicit alcohol markets, which provide access to cheaper alternative alcohol than registered commercial markets, can limit the effectiveness of preventive measures on overall alcohol consumption (Rehm et al., 2022).

To explore the interplay between illicit alcohol circulation and alcoholism prevention measures we conduct semi-structured interviews with experts regarding the prevalence of illicit alcohol circulation in Latvia and strategies to mitigate it.

While our main findings emphasize the inherent challenge of precisely quantifying the size of the illicit alcohol market, our analysis suggests that the share of illicit alcohol in total alcohol consumption in Latvia is relatively low. We also conclude that the size of the illicit alcohol market has been diminishing in recent years, and that public interest in engaging with illicit alcohol is declining. Given these findings, the current scope of the illicit market is unlikely to substantially undermine the efficacy of alcohol control measures. This is especially true as the consumers of illicit alcohol represent a specific group minimally affected by legal alcohol control measures in the country.

Conclusion

Our findings underscore the substantial costs associated with the large alcohol consumption in Latvia. In 2021, budgetary (direct) and non-budgetary (indirect) costs reached 1.3–1.8 percent of Latvia’s GDP. Furthermore, non-financial costs from alcohol abuse represent a loss of nearly 90 thousand years spent in good health and with a good quality of life.

Furthermore, non-financial costs from alcohol abuse represent a loss of nearly 90 thousand years spent in good health and with a good quality of life. This stems primarily from the distress experienced by alcohol users’ household members, and the decline in life quality and premature mortality among users themselves.

Latvia stands out as a country with exceptionally high levels of absolute alcohol consumption per capita compared to other countries. Policy makers should implement effective preventive measures against alcohol consumption to maintain the sustainability of a healthy and productive society in Latvia.

Acknowledgement

This brief is based on a study Alcohol Use, its Consequences, and the Economic Benefits of Prevention Measures completed by BICEPS researchers in 2023, commissioned by the Health Ministry of Latvia (Pļuta et al., 2023).

References

Disclaimer: Opinions expressed in policy briefs and other publications are those of the authors; they do not necessarily reflect those of the FREE Network and its research institutes.

Conflict, Minorities and Well-Being

Posted on June 17, 2018 | Policy Brief

20180618 Conflict, Minorities and Well-Being Image 01

We assess the effect of the Russo-Georgian conflict of 2008 and the Ukrainian-Russian conflict of 2014 on the well-being of minorities in Russia. Using the Russian Longitudinal Monitoring Survey (RLMS), we find that the well-being of Georgians in Russia suffered negatively from the 2008 Russo-Georgian conflict. In comparison, we find no general effect of the Ukrainian-Russian conflict of 2014 on the Ukrainian nationals’ happiness. However, the life satisfaction of Ukrainians who reside in the southern regions of Russia in close proximity to Ukraine is negatively affected. We also show that the negative effect of conflict is short-lived with no long-term legacy. Additionally, we analyze the spillover effect of conflict on other minorities in Russia. We find that while the well-being of non-Slavic and migrant minorities who have recently moved to Russia is negatively affected, there is no effect on local minorities who have been living in Russia for at least ten years.

Militarized conflict affects a myriad of socioeconomic outcomes, such as the level of GDP (Bove et al. 2016), household welfare (Justino 2011), generalized trust and trust in central institutions (Grosjean 2014), social capital (Guriev and Melnikov 2016), and election turnout (Coupe and Obrizan 2016). Importantly, conflict has also been found to directly affect individual well-being (Frey 2012, Welsch 2008).

However, previous research studying individual well-being in transition countries largely abstracts from heightened political instability and conflict proneness, while this has been particularly pertinent in transition countries. Examples of transition countries facing various types of conflicts are abound, such as Yugoslavia, Ukraine, Tajikistan, Russia, Armenia, Azerbaijan, Moldova, and so on. Therefore, it is imperative to explore how conflict shapes well-being in transition countries.

In a new paper (Gokmen and Yakovlev, forthcoming), we add to our understanding of well-being in transition in relation to conflict. We focus on the effect of Russo-Georgian conflict of 2008 and the Ukrainian-Russian conflict of 2014 on the well-being of minorities in Russia. The results suggest that the well-being of Georgians in Russia suffered negatively from the 2008 Russo-Georgian conflict. However, we find no general effect of the Ukrainian-Russian conflict of 2014 on the Ukrainian nationals’ happiness, while the life satisfaction of Ukrainians who reside in the southern regions of Russia in close proximity to Ukraine is negatively affected. Additionally, we analyze the spillover effect of conflict on other minorities in Russia. We find that while the well-being of non-slavic and migrant minorities who have recently moved to Russia is negatively affected, there is no effect on local minorities who have been living in Russia for at least ten years.

Data and Results

We employ the Russian Longitudinal Monitoring Survey (RLMS) which contains data on small neighborhoods where respondents live. Starting from 1992, the RLMS provides nationally-representative annual surveys that cover more than 4000 households with 10000 to 22000 individual respondents. The RLMS surveys comprise a broad set of questions, including a variety of individual demographic characteristics, health status, and well-being. Our study utilizes rounds 9 through 24 of the RLMS from 2000 to 2015.

In this survey, we identify minorities with the question of “What nationality do you consider yourself?” Accordingly, anybody who answers this question with a non-Russian nationality is assigned to that minority group.

We employ three measures of well-being. Our main outcome variable is “life satisfaction.” The life satisfaction question is as follows: “To what extent are you satisfied with your life in general at the present time?”, and evaluated on a 1-5 scale from not at all satisfied to fully satisfied. Additionally, we use “job satisfaction” and “health evaluation” as outcomes of well-being.

Our results suggest that our primary indicator of well-being, life satisfaction, for Georgian nationals has gone down in the Russo-Georgian conflict year of 2008 compared to the Russian majority (see Figure 1). The magnitude of the drop in life satisfaction is about 39 percent of the mean life satisfaction. Our estimates for the other two well-being indicators, job satisfaction and health evaluation, also indicate a dip in the conflict year of 2008. Lastly, our estimates show that the negative impact of the conflict does not last long. Although there is a reduction in the well-being of Georgians both on impact in 2008 and in the immediate aftermath in 2009, the rest of the period until 2015 is no different from the pre-2008 period.

Figure 1. Life Satisfaction of Georgian Nationals in Russia


Source: Authors’ own construction based on RLMS data and diff-in-diff estimates.

Furthermore, when we investigate the effect of the Ukrainian-Russian conflict of 2014, we find no negative effect on the life satisfaction of Ukrainians. One explanation for why the happiness of Ukrainians in Russia does not seem to be negatively affected in 2014 is that the degree of integration of Ukrainians into the Russian society is much stronger than the degree of integration of Georgians. On the other hand, our heterogeneity analysis reveals that in the southern parts of Russia closer to the Ukrainian border, where there are more Ukrainians who have ties to Ukraine, Ukrainian nationals are differentially more negatively affected by the 2014 conflict. The differential reduction in the happiness of Ukrainians is about 19 percent of the mean life satisfaction.

Moreover, we also look into whether there is any spillover effects of the Russo-Georgian and the Ukrainian-Russian conflicts on the well-being of other minorities. We first carry out a simple exercise on non-Slavic minorities of Russia. We pick the sample of non-Slavic ex-USSR nationals that are similar to Georgians in their somatic characteristics, such as hair color and complexion. This group of people include the nationals of Azerbaijan, Kazakhstan, Uzbekistan, Kyrgyzstan, Turkmenistan and Tajikistan. We treat this group as “the countries with predominantly non-Slavic population” as their predominant populations are somatically different from the majority Russians, and thus, might either have been subject to discrimination or might have feared a minority backlash to themselves during the times of conflict. This conjecture finds some support below in Figure 2 in terms of violence against minorities. We observe in Figure 2 that hate crimes and murders based on nationality and race peak in 2008.

Our estimates also support the above hypothesis and propose that there is some negative effect of the 2008 conflict on non-slavic minorities’ happiness as well as their job satisfaction, whereas 2014 conflict has no effect.

Figure 2. Hate Murders in Russia over Time

Source: Sova Center

Next, we investigate the spillover effects of conflict on Migrant Minorities. Migrant minorities are minorities who have been living in their residents in Russia for less than 10 years. We conjecture that these minorities, as opposed to the minorities who have been in place for a long time, could be more susceptible to any internal or external conflict between Russia and some other minority group for fear that they themselves could also be affected. Whereas other types of longer-term resident minorities, which we call Local Minorities, are probably less vulnerable since they have had more time to establish their networks, job security, and most likely also have Russian citizenship. Our estimates back up the above conjecture and demonstrate that migrant minorities suffer negatively from the spillover effects of the 2008 conflict onto their well-being captured by any of the three measures, and not from the 2014 conflict, whereas there is no negative impact on local minorities.

Conclusion

In this paper, instead of focusing on the direct impact of conflict on happiness in war-torn areas, we contribute to the discussion on conflict and well-being by scrutinizing the well-being of people whose country of origin experiences conflict, but they themselves are not in the war zone. Additionally, we show that some other minority groups also suffer from such negative spillovers of conflict. Being aware of such negative indirect effects of conflict on well-being is essential for policy makers, politicians and researchers. Most policy analyses ignore such indirect costs of conflict, and this study highlights the bleak fact that the cost of conflict on well-being is probably larger than it has been previously estimated.

References

  • Bove, V.; L. Elia; and R. P. Smith, 2016. “On the heterogeneous consequences of civil war,” Oxford Economic Papers.
  • Coupe, T.; and M. Obrizan, 2016. “Violence and political outcomes in Ukraine: Evidence from Sloviansk and Kramatorsk”, Journal of Comparative Economics, 44, 201-212.
  • Frey, B. S., 2012. “Well-being and war”, International Review of Economics, 59, 363-375.
  • Gokmen, Gunes; and Evgeny Yakovlev, forthcoming. “War and Well-Being in Transition: Evidence from Two Natural Experiments”, Journal of Comparative Economics.
  • Grosjean, P., 2014. “Conflict and social and political preferences: Evidence from World War II and civil conflict in 35 European countries” Comparative Economic Studies, 56, 424-451.
  • Guriev, S.; and N. Melnikov, 2016. “War, inflation, and social capital,” American Economic Review: Papers & Proceedings, 106, 230-35.
  • Justino, P., 2011. “The impact of armed civil conflict on household welfare and policy,” IDS Working Papers.
  • Welsch, H., 2008. “The social costs of civil conflict: Evidence from surveys of happiness” Kyklos, 61, 320-340.

Disclaimer: Opinions expressed in policy briefs and other publications are those of the authors; they do not necessarily reflect those of the FREE Network and its research institutes.

Old-Age Poverty and Health – How Much Does Income Matter?

Posted on September 30, 2013 | Policy Brief

20130930 Old-Age Poverty and Health Image 01

The question concerning the material situation of older people and its consequences for their wellbeing seems to be more important than ever. This is especially true given rapid demographic changes in the Western World and economic pressures on governments to reduce public spending.  We use data from the Survey of Health, Ageing and Retirement in Europe (SHARE) to examine different aspects of old-age poverty and its possible effects on deterioration in health. The data contains information on representative samples from 12 European countries including the Czech Republic and Poland. We use the longitudinal dimension of the data to go beyond cross sectional associations and analyze transitions in health status controlling for health in the initial period and material conditions. We find that poverty matters for health outcomes in later life. Wealth-defined and subjective poverty correlates much more strongly with health outcomes than income-defined measure. Importantly subjective poverty significantly increases mortality by 58.3% for those aged 50–64 (for details see Adena and Myck, 2013a and 2013b). 

Measuring Poverty

When measuring poverty, the standard approach is to define the poverty threshold at 60% of median equalized income. This standardized measure offers some advantages, such as simplicity and comparability with already existing studies. However, there are valid arguments against its use when analyzing old-age poverty. The permanent-income theory provides arguments against current income as a major determinant of quality of life of older people. Moreover, poverty defined with respect to current income while taking account of household size through equalization, ignores other important aspects of living costs such as disability or health expenditures. Additionally, most analysis using income-poverty measures ignore such aspects as housing ownership and housing costs.

Our analysis examines different aspects of poor material conditions of the elderly. The first poverty definition refers to respondents’ wealth as an alternative to income-defined poverty. Poor households, defined with reference to wealth (“wealth poverty” – WEALTH), are those that belong to the bottom third of the wealth distribution of the sample in each country. For this purpose, household wealth is the sum of household real assets (net of any debts) and household gross financial assets. Secondly, we compare the above poverty measures to a subjective measure of material well-being. This measure is based on subjective declarations by respondents, in which case (“subjective poverty” – SUB) individuals are identified as poor on the basis of a question of how easily they can make ends meet. If the answer is “with some” or “with great” difficulty, individuals in the household are classified as “poor”.

One reflection of potential problems with the standard income poverty measure becomes visible when it is compared with the subjective measure. The graph below shows the differences in country rankings when using one or the other poverty measure.  The country with the greatest disproportion is Czech Republic. While being ranked as second according to the income measure, it is ninth according to the subjective measure.

Figure 1. Country Ranks in Old-Age Poverty According to an Income versus a Subjective Measure

Slide1

Source: Authors’ calculations using SHARE data (Wave 2, release 2.5.0).

Even more striking is the fact that the differences between ranks are not because of over or under classification of individuals as poor, but rather because of misclassification. Figure 2 shows that there is little overlap between different poverty measures. The share of individuals classified as poor according to all three measures is only 7.95%, whereas it is 60% according to at least one of the measures.

Figure 2. Poverty Measure Overlap

Slide1

Notes: Data weighted using Wave 2 sample weights. Source: Authors’ calculations using SHARE data (Wave 2, release 2.5.0).
 

Measuring Well-Being

We examine three binary outcomes measuring the well-being of the respondents – two reflecting physical health, and one measuring individuals’ subjective health. The two measures of physical health are generated with reference to the list of twelve symptoms of bad health and the list of twenty-three limitations in activities of daily living (ADLs). In both cases, we define someone to be in a bad state if they have three or more symptoms or limitations. The two definitions are labelled as: “3+SMT” (three or more symptoms) and “3+ADL” (three or more limitations in ADLs). Subjective health “SUBJ” is defined to be bad if the subjective health assessment is “fair” or “poor”. Finally, we also analyze mortality as an “objective” health outcome.

Poverty and Transitions in Well-Being and Health

There is some established evidence in the literature that poverty negatively affects health and other outcomes at different stages of life.[1] At the same time, there is little evidence on how the choice of the poverty measure might result in under- or over-estimation of the effects of poverty. We address this question by examining different poverty measures as potential determinants of transitions from good to bad states of health.

The results confirm that living in poverty increases an individual’s probability of deterioration of health. In a compact form, Figure 3 presents our results from 12 separate regressions (4 outcomes, three poverty measures). Here we report the odds ratios related to the respective estimated poverty dummies. Individuals classified as poor according to the income measure are 37.7% more likely to report bad subjective health in a later wave of the survey than their richer counterparts; they are 4.5% more likely to suffer from 3 or more symptoms; 18.7% more likely to suffer from 3 or more limitations; and 5% more likely to die. The last three effects, however, are not statistically significant.

In contrast, the effects of wealth-defined poverty and subjectively assessed poverty are 2-8 times stronger than those of income poverty, and they are also significant for all outcomes but death. Overall, wealth-defined poverty and subjective assessment of material well-being strongly correlate with deterioration in physical health (exactly the same goes for improvements in health, see Adena and Myck 2013b).

Figure 3. Poverty and Transitions from Good to Bad States Overlap

Slide2

Notes: Data weighted using Wave 2 sample weights. Source: Authors’ calculations using SHARE data (Wave 2, release 2.5.0, Wave 3, release 1, Wave 4, release 1).
 

Poverty and Mortality in the Age Group 50-64

Our analysis reveals differences between age groups and confirms the decreasing importance of income (and thus income defined poverty) with age. As compared to the average effects presented in Figure 3, for the younger age group 50–64 income poverty proves more important as a determinant of bad outcomes, with transition probabilities between 20 and 40% for all outcomes (see Figure 4). The magnitudes are closer to those of other poverty measures, but still lower in all cases. Importantly, we find that wealth-defined and subjective poverty is an important determinant of death in the age group 50–64.

Figure 4. Poverty and Transitions from Good to Bad States 50-64 Slide3
Notes: Data weighted using Wave 2 sample weights. Source: Authors’ calculations using SHARE data (Wave 2, release 2.5.0, Wave 3, release 1, Wave 4, release 1).
 

Conclusions

The role of financial conditions for the development of health of older people significantly depends on the measure of material well-being used. In this policy brief, we defined poverty with respect to income, subjective assessment, and relative wealth. Of these three, wealth-defined poverty and subjective assessment of material well-being strongly and consistently correlate with deterioration and improvements in physical and subjective health. We found little evidence that relative income poverty plays a role in changes in physical health of older people. This suggests that the traditional income measure of household material situation may not be appropriate as a proxy for the welfare of older populations, and may perform badly as a measure of improvements in their quality of life or as a target for old-age policies. To be valid, such measures should cover broader aspects of financial well-being than income poverty. They could incorporate aspects of wealth and the subjective assessment of material situations as well as indicators more specifically focused on the consumption baskets of the older population.

▪

References

  • Adena, Maja and Michal Myck (2013a): “Poverty and transitions in key areas of quality of life”, in: Börsch-Supan, Axel,  Brandt, Martina , Litwin, Howard and Guglielmo Weber (eds.) “Active Ageing and Solidarity between Generations in Europe – First Results from SHARE after the Economic Crisis.”
  • Adena, Maja and Michal Myck (2013b) Poverty and Transitions in Health, IZA Discussion Paper 7532, IZA-Bonn.

 

[1] For a literature review, see our publications.

Can Anti-Smoking Campaigns Increase Obesity? Evidence from Belarus

Posted on May 20, 2013 | Policy Brief

20150518-Expected-Effects-of-Tobacco-1

Authors: Aliaksandr Amialchuk, University of Toledo, and Kateryna Bornukova, BEROC.

In this brief, we discuss the possible effects of an anti-tobacco campaign on obesity levels in Belarus based on results of Amialchuk et al (2012). Both smoking and obesity are among the main health concerns in Belarus. Negative correlation between smoking and body weight is well documented, but can anti-tobacco campaign cause an increase in obesity rates? Results of studies from developed countries provide mixed evidence. In Amialchuk et al (2012), we use household survey data from Belarus to establish the link between smoking and body mass index (BMI). We use cigarette prices and regional smoking prevalence as instruments for smoking, and find a negative effect of smoking on BMI. Moreover, using the quantile regression approach, we find that smoking has different effects on body weight for different BMI quantiles, with the largest negative effect in the upper part of the conditional BMI distribution. These findings suggest that anti-tobacco campaigns may slightly increase obesity rates, and campaigns should therefore ideally also include measures to promote a healthy lifestyle. On the other hand, the potentially modest weight gain from an anti-tobacco campaign is likely to be more than offset by the general improvements in health.

Smoking and Obesity in Belarus

Smoking prevalence in Belarus, like in many other transitional countries, is quite high. According to the Belarusian Household Survey of Income and Expenditure from 2010, the smoking rate was 26%, with a much higher prevalence of among men (49.3%) compared to women (9.5%).[1]

Despite the troubling levels of smoking prevalence, little has been done to combat smoking in Belarus. While most of the post-Soviet economies liberalized the tobacco industry, it remains under government control in Belarus. The profits of the state-owned cigarette producers, along with tobacco taxes, constitute an important part of Belarusian budget revenues. This might explain why the Belarusian government has not engaged in anti-tobacco campaigns in the past. However, Belarus is currently implementing Anti-Tobacco Plan for 2011-2015 in cooperation with the World Health Organization.

The Anti-Tobacco Plan includes a variety of anti-tobacco actions and measures. In particular, the government has plans to gradually increase tobacco taxes, introduce smoking-free zones and restrict smoking in public places, along with a massive informational campaign about the dangers of smoking and ways to quit. These measures have the potential to lead to a significant decrease in smoking prevalence. However, an unintended consequence of these policies might be an increase in overweight and obesity rates.

In fact, obesity is another important health problem of Belarus. In 1996-2008, (the period of analysis in Amialchuk et al (2012)), the mean BMI among adults was 26, which suggests that an average Belarusian adult is just on the borderline between healthy weight and overweight. In particular, 34% of adults are overweight, while approximately 15% of adults are obese. Moreover, the distribution of weight status has undergone substantial changes over time: the percentage of individuals in the right tail of the BMI distribution has increased over time, with the percentage of obese increasing faster than the percentage of overweight individuals.

The Link between Smoking and Obesity

The negative relationship between smoking and body weight is well-documented in the medical literature. This inverse relationship is mostly attributed to how smoking affects body weight by boosting metabolism and suppressing appetite.  However, causality is usually difficult to establish: for example, a smoking person may also be more likely to eat unhealthy foods and care less about their health in general. Nevertheless, most of the previous studies have found a significant negative effect of smoking on body weight.

Since in many developed countries, the decrease in smoking prevalence coincided in time with the surge in both overweight and obesity rates, the question arises whether anti-smoking campaigns are in part responsible for the increase in obesity rates. However, the evidence on the effects of anti-tobacco campaigns on overweight/obesity rates in developed countries is mixed. Some studies do not find any significant effect on obesity (Nonnemaker et al, 2009).

Evidence from Belarus

As mentioned above, smoking behavior and BMI may be jointly determined, and to deal with the challenge of establishing causality, we utilize the method of instrumental variables analysis. We employ two instrumental variables in our estimation: (i) the mean number of cigarettes smoked per day in the same year-region-gender- and education group as the respondent, and (ii) the average yearly price per pack of cigarettes in the region where the respondent lives. Gilmore et al. (2001) identify important demographic and socio-economic differences in smoking rates, which dictates our use of gender and education categories (below secondary, secondary, university degree) to construct groups of observations that will be followed over time. The use of region as a grouping variable allows us to capture the social norm associated with smoking at the regional level. We exclude the individual’s own cigarette smoking when we create group-level means. Group-specific smoking prevalence is likely to be predictive of the individual’s own smoking preferences, but is unlikely to have a direct effect on individual’s weight status other than through the effect on individual’s smoking. After accounting for the fixed differences in average smoking among regions, gender, and education groups within each year, the source of variation that is available to identify the effect of the instrument on individual’s smoking is the differences in smoking prevalence among various interactions of year, region, gender and education categories.

We use lagged prices as instrument for current year cigarette consumption of the individuals in order to account for the addictive and inelastic nature of demand for smoking and the inability to quickly change smoking behavior after a price change. Furthermore, we use natural log of cigarette prices in order to account for the potentially non-linear effect on the number of cigarettes smoked. Cigarette prices are likely to influence an individual’s BMI only through its effect on smoking.

Other controls in our regressions include total personal income; household size; age; gender; single vs. married indicator; indicators of self-reported health status (good health, fair health, and poor health indicators); number of medical visits in the last 3 months; indicator for having been hospitalized in the last 12 months; indicator for whether health affects ability to work; sports practicing indicator; indicators for the educational attainment (university diploma, secondary education); and indicators for being currently employed, having ever worked, and being a student.

Our endogeneity-corrected estimates suggest that one additional cigarette per day would decrease BMI by roughly 0.23 units, and would reduce the probability of being overweight by approximately 2.5%. Furthermore, there is a small but significant effect on the likelihood of being obese: an additional cigarette smoked per day decreases the probability of being obese by 1.3%. Our results suggest an important implication that smoking is inversely related to body weight, and has some effect on obesity rates.

We also explore the difference in the effect of smoking on body weight across different quantiles of conditional BMI distribution. The largest effect is obtained for the 75th and 90th percentiles, and the smallest effects for the 10th and 25th percentiles. Smoking has a large effect on the body weight of individuals who are at the upper tail of the BMI distribution. These findings suggest that a reduction in smoking rate may lead to an increase in obesity rates by inducing weight gain among the population near the top end of the conditional BMI distribution.

While we found evidence of a possible increase in obesity rates resulting from the anti-tobacco campaign, it is important to remember that adverse health effects of smoking are numerous and the health benefits of smoking cessation are far in excess of the risk of weight gain. The current high prevalence of smoking and number of overweight individuals in Belarus constitute a major public health concern. Our results suggest that the prevalence of overweight and obesity might be exacerbated by the anti-tobacco campaign. From a policy perspective, an increase in obesity rates among the general population may be a reasonable concern for policy instruments targeted at reducing the overall smoking rates. It would therefore be wise to promote healthy eating habits and sports together with the anti-smoking campaign. However, the potentially modest weight gain from anti-tobacco campaign only is likely to be more than offset by the general health improvements associated with a decline in smoking rates.

▪

References

  • Amialchuk, A., K. Bornukova, M. Ali, 2012. Smoking and Obesity Revisited: Evidence from Belarus. BEROC Working Paper Series, WP no. 19
  • Gilmore, A.B., McKee, M., Rose, R., 2001. Prevalence and determinants of smoking in Belarus: A national household survey, 2000. European Journal of Epidemiology 17: 245-253
  • Nonnemaker, J., Finkelstein, E., Engelen, M., Hoerger, T., Farrelly, M., 2009. Have efforts to reduce smoking really contributed to the obesity epidemic? Economic Inquiry 47, 366–376

 

[1] The social norms explain difference in smoking rates of men and women. In younger population, however, gender differences in smoking rates are less pronounced.