Project: FREE policy brief

The Expectation Boom: Evidence from the Kazakh Oil Sector

Posted on March 9, 2020 | Policy Brief

Image of Kazakhstan with skyscraper in the middle of the capital city representing oil price

This policy brief shows that an oil price boom may trigger dissatisfaction with one’s income and that this dissatisfaction is independent of the effect of the boom on real economic conditions. Unique data from Kazakhstan allows us to quantify the impact of the recent oil price boom on satisfaction with income. Compared to other households in the country, households related to the oil sector suffer a marked drop in their satisfaction with their income during the period of high oil prices. Based on our results, we argue that an oil price boom creates a gap between people’s expectations of the benefits from the boom and the observed economic conditions. Our results call for researchers, policy makers and companies to devote more attention to the dynamics of satisfaction, not only during resource busts but also during resource booms.

Local Impact of Natural Resources

Often, resource wealth is associated with a curse, slowing economic growth in resource-rich developing countries (Venables, 2016). While traditionally, this relationship has been explored across countries, more recently, the literature started exploiting plausibly exogenous spatial variation in resource wealth within countries (Cust and Polhekke, 2015). We now know that resources can generate local economic wealth (Aragon and Rud, 2013), while also attracting corrupted individuals to power (Asher and Novosad, 2018) and triggering local conflicts (Berman et al. 2017; Rigterink, 2018). But, up to now, we know very little about the impact of resource booms on individuals’ perceptions. Since perceptions and behavioral biases may also drive actions, understanding whether and how resources affect perceptions is key in understanding the local impact of natural resources (Collier, 2017).

In a new working paper (Girard, Kudebayeva and Toews, 2020) we use Kazakhstan as a case study to shed more light on the importance of such perceptions. We document the conditions that preceded and presumably contributed to the violent conflicts in the oil rich districts of Kazakhstan in 2011. We show that periods of high oil prices can actually lead to a drop in reported satisfaction with income. This implies that due to mere changes in perceptions, which are not reflected in economic conditions, a large number of people may experience a significant drop in satisfaction with income, creating a fertile ground for conflicts.

The Zhanaozen Conflict

Our attention to the case of Kazakhstan is driven by the extreme events that took place in 2011 in the city of Zhanaozen, a booming oil town in the west of the country’s desert. In May 2011, after several years of high oil prices, private sector workers in Zhanaozen demanded amendments to the pre-existing collective bargaining agreement asking in particular for a raise in wages. Difficulties in negotiating an agreement resulted in local oil companies dismissing more than 2000 employees in the summer of 2011 and oil production dropping by 7% in the first three quarters of 2011 relative to the same period in the previous year. At the conflict climax, the police tried to clear the central square of Zhanaozen for the upcoming preparations of the Independence Day, resulting in the killing of 17 and the injuring of over 100 people (Satpayev and Umbetaliyeva, 2015).

Oil Price Boom in Kazakhstan

Kazakhstan offers an ideal case study for our research question for two reasons. First, the government of Kazakhstan closely monitored citizens’ satisfaction with income throughout most of the 2000s using a representative household panel survey. Using this data allows us to link variation in the price of oil to within household variations in satisfaction with income – conditional on household income, thus, capturing the changing perceptions of household heads regarding their income. Secondly, Kazakhstan is a small open resource rich economy, with sparsely populated and remote districts, whose economic activity nearly exclusively depends on the extraction of oil and gas. The fact that Kazakhstan is a small open economy implies that changes in the oil price may be treated as exogenous to households located in Kazakhstan. The spatial isolation of the oil rich districts allows us to consider the group of household heads employed in the private sector in the oil rich districts as either directly or indirectly involved in the extraction of oil and gas.

Figure 1. Kazakhstan

Source: Resource rich districts are indicated as treated. The information on the spatial identification of oil and gas rich district is taken from the Petroleum Encyclopedia of Kazakhstan and captures more than 90% of total oil and gas production in Kazakhstan (Munayshy Public Foundation, 2005).

Satisfaction With Income

To identify the effect of oil price fluctuations on satisfaction with income, we exploit three sources of variation: location of the household, sectoral employment and time. The group affected by the price of oil is the group of oil-related households. Oil-related households consist of households whose head is employed in the private sector of the oil rich districts of Kazakhstan, and who are thus the closest to the oil sector (by nature of their activity and place of residence). The differential evolutions in satisfaction of household heads employed in other sectors and households located in other districts – in other words, households which are more remote from oil and gas extraction than the oil-related households provide a plausible counterfactual.

The main results are depicted in Figure 2 which represents the relationship between income and satisfaction for 8 groups based on the three sources of variation: oil price (which was low between the years 2001 and 2004, and high between 2005 and 2009), place of residence as indicated in Figure 1, and sector of activity.

First, we note that the relationship between income and satisfaction is upward sloping: reported satisfaction with household income increases with income. This is intuitive. Focusing on oil poor districts that appear in the bottom panel, we observe that the relation between satisfaction and income is virtually the same across sectors and time periods of low and high prices of oil. This is, however, not true for oil-rich districts, which are depicted in the top panel. Here, the relationship between income and satisfaction only remains unaffected across time for household heads who are not employed in the private sector. The picture changes if we turn to household heads employed in the private sector, who are the oil-related household heads. The satisfaction with the income of oil-related household heads shifts downwards, compared to other households, in the period of high oil prices (years 2005-2009). This downward shift is even more striking since oil-related household heads valued their income relatively higher than other households during the period of low oil prices (2001-2004).

Figure 2. Satisfaction with Income

Source: Authors’ calculations based on satisfaction with income and household income as reported in the Household Survey of Kazakhstan. The figures above depict the relationship between logged household income per family member and reported satisfaction with income by the head of the household on a scale from 1 to 5. The relationship is depicted conditional on household fixed effects and year fixed effects. The former account for time-invariant household specific characteristics such as individual biases. The latter account for Kazakhstan specific shocks affecting households in oil poor and oil rich districts simultaneously due to political and economic business cycles. As a result, the relationship between logged income per family member and satisfaction is normalized to zero in both dimensions.

Lastly, we document that the negative variation in satisfaction is related to the contemporaneous change in the price of oil. The satisfaction with income is not persistent, it is unrelated to past and future levels of the oil price.

Conclusion

Our results suggest that oil prices fluctuations can be linked to the individual’s perception of income. The fact that oil-related household heads express a strong dissatisfaction compared to other household heads may help to understand what made December 2011 possible, when 17 people were killed and over 100 people were wounded in Zhanaozen. If generalizable, such dynamics of perceived satisfaction with income should be kept in mind by both policy makers and extractive companies not only during resource busts but also during resource booms.

References

  • Aragon, Fernando M. and Juan Pablo Rud (2013). “Natural resources and local communities: evidence from a Peruvian gold mine.” American Economic Journal: Economic Policy 5(2):1–25.
  • Asher, Sam and Paul Novosad (2018). Rent-seeking and criminal politicians: Evidence from mining booms. Working Paper.
  • Berman, Nicolas, Mathieu Couttenier, Dominic Rohner and Mathias Thoenig (2017). “This Mine Is Mine! How Minerals Fuel Conflicts in Africa.” American Economic Review 107(6):1564–1610.
  • Collier, Paul (2017). “The institutional and psychological foundations of natural resource policies.” The Journal of Development Studies 53(2):217–228.
  • Cust, James and Steven Poelhekke (2015). “The local economic impacts of natural resource extraction.” Annu. Rev. Resour. Econ., 7(1):251–268.
  • Girard, Victoire, Alma Kudebayeva and Gerhard Toews (2020). “Inflated Expectations and Commodity Prices: Evidence from Kazakhstan.“ GLO Discussion Paper Series 469.
  • Rigterink, Anouk S. (2020). “Diamonds, rebel’s and farmer’s best friend: Impact of variation in the price of a lootable, labour-intensive natural resource on the intensity of violent conflict.” Journal of Conflict Resolution 64(1):90–126.
  • Munayshy Public Foundation (2005). “Petroleum Encyclopedia of Kazakhstan.”
  • Girard, Victoire, Alma Kudebayeva and Gerhard Toews (2020). “Inflated Expectations and Commodity Prices: Evidence from Kazakhstan” Working Paper.
  • Satpayev, Dossym and Umbetaliyeva, Òolganay (2015). “The protests in Zhanaozen and the Kazakh oil sector: Conflicting interests in a rentier state.” Journal of Eurasian Studies 6(2):122–129.
  • Venables, Anthony J. (2016): “Using Natural Resources for Development: Why Has It Proven So Difficult?” Journal of Economic Perspectives, 30, 161 – 84.

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.

Women at the Top of the Income Distribution: Are Transition Countries Different?

Posted on March 2, 2020 | Policy Brief

Shadows of women walking down the road in a sunset representing women at the top of income distribution

This policy brief reviews recent research on women at the top of the income distribution. The overall trend across a number of countries is that, while women are still a minority (and more so the closer to the top one moves), their share in top income groups has steadily increased since the 1970s. Detailed data from Sweden suggests that most of this rise is due to women increasingly earning high labor incomes (rather than capital becoming more important). It also shows that there are important differences between top income men and women, especially with respect to family circumstances. Comparing preliminary results from former Soviet and Eastern European countries indicates that there are, on average, more women at the top of the income distribution in these countries. On the other hand, the average time trend indicates that the share of women in top groups is falling. The preliminary results also indicate considerable heterogeneity across countries. These preliminary results require more detailed study, as does the question to which extent the relatively strong representation of women at the top of the income distribution reflects the “economic power” of women in the region.

The Gender Aspect of Rising Top Shares

Rising inequality has received a lot of attention in the policy debate as well as in the academic literature over the past decade. A particular feature of this discussion has been the increased concentration of both wealth and income in top groups. The summary of the World Inequality Report 2018 starts by stating that “The top 1% has captured twice as much of global income growth as the bottom 50% since 1980”. Such facts have, in turn, brought a lot of attention to the characteristics of top groups. What is driving their income growth? What is their income composition? Why have top shares increased so much in recent decades? (see, e.g., Roine and Waldenström, 2015, for an extensive overview, or Roine, 2016, for a brief summary).

However, one aspect which has received relatively little attention is that of gender. This may seem a little surprising. In a time when gender dimensions are often acknowledged as being important, one would expect that questions about the gender composition of top groups would also be of interest. If we know that top income shares are increasing, what is the gender composition of these groups? How has this changed over time?

This brief outlines some recent results on these questions and also points to some preliminary findings about a potential contrast between Western countries and (former) transition countries.

Evidence from Sweden, 1971-2017

Sweden is one of the few countries having had independent taxation of all taxpayers for a long period of time, allowing for a thorough analysis of the gender composition of top income groups. After having had joint taxation for married couples for most of the 20th century, and a short period of the option to be taxed independently even if married, Sweden switched to fully independent taxation in 1971. In a recent paper Boschini et al. (2020) study developments of men and women in top income groups in Sweden using detailed registry data on the full population for the almost 50-year period since.

The study finds a number of interesting results. First, it is evident that the share of women in top income groups has increased significantly, yet women remain clearly underrepresented, and more so the higher up in the distribution one moves. Figure 1 below shows the basic development over time for three top groups: the top 10 (P90-100), the top 1 (P99-100), and the top 0.1 group (P99.9-100) in the total income distribution and the labor income distribution respectively.

Figure 1. Share women in top groups in Sweden.

Source: Boschini et al. (2020)

Besides showing the general development comparing the two panels also reveals a subtler point: especially in the earlier decades and in the very top group (the top 0.1 group), there were substantially more women at the top of the total income distribution than at the top of the labor earnings distribution. In the 1970s and 1980s, the share of women in the top 0.1 group of the total income distribution is about two to three times as large as in the labor earnings distribution. Put differently, this means that in the past, to the extent that there were any women at the very top, they were mainly there thanks to capital incomes. Over time this changes and detailed analysis in the paper shows that the growth of the share of women in top groups is driven by an increasing share of high-income women in the labor income distribution.

While it seems that top income men and women have converged in terms of income composition and observable individual characteristics, the one area that still stands out as being markedly different is partner income. Figure 2 shows that top income women are much more likely to have partners who are also in the top of the income distribution. Even if the trend indicates convergence, large differences remain. Out of the top 1 women who are married, 70% have a partner who is at least in the top 10 (and about 30% are also in the top 1). For married top 1 men, only 30% have a partner who is in the top 10, and only a couple of percentage points are in the top 1. Part of this is, of course, a reflection of there being fewer women in top groups, but this is far from explaining all the difference (See Boschini et al., 2020 for more details).

Figure 2. Share of top income partners in Sweden.

Source: Boschini et al. (2020)

This is of course far from conclusive, but it points in the direction of family circumstances being a potential factor for explaining the relative absence of women in top income groups. Having a partner with a top (income) career is likely to be more demanding (for both parties) and such couples are much more common among top income women than men.

Several strands of research connect to this: for example, Fisman et al. (2006) find, among other things, that men are significantly “less likely to accept a woman who is more ambitious than he”. Also, work by Bertrand et al. (2015), on the impact of gender identity suggest that there is a social norm prescribing that men should earn more than women, which creates a discontinuity in the distribution of women’s contribution to total household income at 50 % (although Hederos Eriksson and Stenberg (2015) and Zinovyeva and Tverdostup (2018) find alternative explanations for this observation). Folke and Rickne (2020) find that women who are elected to high political office in Sweden face a higher probability of divorce (while this is not the case for men). Furthermore, according to the World Values Survey, close to 40% of Americans as well as Europeans agree with the statement “(i)f a woman earns more money than her husband, it’s almost certain to cause problems”. Taken together, findings like these suggest that, even in relatively progressive countries, social norms may contribute to women shying away from entering career paths leading to top incomes.

What About Other Countries?

Even though the Swedish data is unusually detailed, it is certainly not the only country where individual tax data exist. Atkinson et al. (2018) calculate the share of women in top groups for eight countries over time periods when individual tax data exist. Figure 3 puts their results next to those from Sweden. The resulting picture shows a remarkably similar development across countries and over time. The share of women in the top 10 has approximately tripled since the 1970s, from around 10% to around 30%. For the top 1 group, the level is slightly lower, but the relative increase is similarly large, from slightly above 5% to around 20%.

Figure 3. International comparison.

Source: Atkinson et al. (2018) and Boschini et al (2020).

Bobilev et al. (2019) explore the extent to which Luxemburg Income Study (LIS) data can be used to shed light on the presence of women at the top of the income distribution. Their findings point to a similar trend across a broader set of countries. Even though the main analysis has to be limited to the share of women at the top of the labor income distribution (since the possibilities to separate out individual capital incomes is limited), the picture in terms of the share of women in top groups is surprisingly similar across the 28 countries for which sufficient data exists from around 1980 until today. The overall finding is that the share of women in the top 10 group increases from about 10% around 1980 to just below 30% today.

To the extent that LIS data allows us to look at partners and family circumstances, the data shows a consistent pattern of asymmetries between top income men and women similar to that in Sweden found by Boschini et al. (2020). Having a partner and having children are positively associated with being in top income groups for men, but negatively associated for women (even though these differences have decreased over time). Also, top income men are likely to have partners who are not in the top of the income distribution, while this is not the case for top income women. Understanding patterns like these and the underlying channels is likely to contribute to our comprehension of the remaining differences in top income shares between men and women.

Are There Differences Between “East and West”?

A particularly interesting pattern in the LIS data is the difference that emerges when contrasting transition countries to Western countries.

As has often been pointed out, the Soviet Union and many of the countries in Eastern and Central Europe were, at least in some dimensions, forerunners in terms of promoting gender equality (e.g., Brainerd, 2000; Pollert, 2003; Campa and Serafinelli, 2019). This was mainly due to the high participation of women in the labor market as well as the (officially) universal access to basic health care and education.

However, some scholars have suggested that not all aspects of gender equality were as advanced in the countries in the Soviet Union and in Central and Eastern Europe (Einhorn, 1993; Wolchik and Meyer, 1985). Even though women were highly integrated in the labor market, they were still expected to take care of child rearing and housework at the same time (UNICEF, 1999). The gender pay gap and gender segregation in the labor market was also similar to levels found in OECD countries. In addition, despite the high number of women in representative positions in communist party politics, women were rarely found in positions of real power in the political sphere (Pollert, 2003).

Looking just at average values (in the labor income distributions), there are clear differences between East and West in top groups. The share of women among the top earning groups was considerably higher in some former Soviet countries during and after transition. However, the shares of women in top income groups have been converging in East and West.

Figure 4. Share of women in the top 10 / top 1 income groups, East vs. West.

Data source: Own calculations based on LIS data. West: unweighted average for Australia, Canada, Denmark, Italy, Norway, New Zealand, Spain, Great Britain. East: unweighted average for the Czech Republic, Estonia, Georgia, Hungary, Lithuania, Poland, Russia, Serbia, Slovenia and the Slovak Republic.

An analysis of the situation at the country level, provides a more complex picture. Figure 5 clearly indicates that the total representation of women in the top 10 income group has been higher in Eastern European countries than in the West (the pattern is similar for the top 1). However, while the share of women in top income groups has consistently increased in Western countries, the developments for women are much less homogenous in Eastern Europe (being below the diagonal indicates a higher share of women in the top 10 in 2005-2020 as compared to 1990-2005).

In Estonia, Slovakia and Poland, women are less likely to be part of the top income group in the period from 2005 to 2020 than they were in the years directly following transition. Considering that the most recent family policies in Poland have been shown to discourage female labor supply (Myck, Trzciński, 2019), this is maybe not so surprising.

Figure 5: Share of women in top 10 income group by country.

Data source: Own calculations based on LIS data. Eastern and Western countries defined as if Figure 4.

The share of women in the top 10 income group in Estonia declined from an astonishingly high 53% in 2000 to about 31% in 2013, which, admittedly, is still high compared to the corresponding average rate for Western countries (28%). Women in Russia, Hungary, Slovenia and the Czech Republic, by contrast, are more likely to be among the top earners in the period from 2005 to 2020 than they were between 1990 and 2005. Moreover, among all the countries in our sample, more recently, Slovenia is the country with the highest share of women in the top 10 of income earners (44% in 2007); Slovenian women seem to have gained grounds even after transition.

How come the representation of women in top income groups remains high (or even increases) in some transition countries but decreases in others? What is the role played by policy and regulation and what role can be attributed to social norms, family circumstances and institutions such as childcare? May economic growth have led to women dropping out of the labor force or never entering it to do care work, even when they had been or potentially could have been part of top income groups? What would be the impact of adding capital incomes to the picture?

Conclusion

Looking across a large number of countries, women seem to have increased their presence in top income groups since the 1970s. This has mostly been driven by women increasingly having high paying jobs. A preliminary look at LIS data indicates that former Soviet and Eastern European countries on average had higher shares of women in top groups around 1990, probably reflecting high labor market participation as well as relatively high education levels for women. But it also indicates that in some Eastern European countries, the share of women in top groups has dropped since the 1990s. As noted by Campa, Demirel, and Roine (2018) there seems to be an overall convergence in some dimensions of gender equality in transition countries, but there is also considerable variation across countries. More detailed studies of how men and women fare in terms of reaching top positions in incomes – but also in other areas like politics – are much needed and likely to be an interesting research area for years to come.

References

Removing Obstacles to Gender Equality and Women’s Economic Empowerment – What Can Policy Makers Learn from Global Research on Gender Economics?

Posted on February 24, 2020 | Policy Brief

20200224 Removing Obstacles to Gender Equality FREE Network Policy Brief Image 01

On November 15-16, 2019, the FREE Network and the ISET Policy Institute organized and conducted an international gender economics conference in Tbilisi, Georgia. The conference was organized as part of the FROGEE initiative – the Forum for Research on Gender Economics – supported by the Swedish International Development Agency (SIDA) and coordinated by the Stockholm Institute of Transition Economics (SITE). The conference brought together researchers, policymakers, and the broader development community to discuss obstacles to gender equality and women’s economic empowerment, as well as policies to remove existing constraints, with a particular focus on Eastern Europe and Emerging Economies. This policy brief provides an overview of the main takeaways from the presentations, with a special focus on policy-relevant lessons.

Introduction

In November 2019, Tbilisi welcomed its first international academic conference on gender economics, “Removing Obstacles to Gender Equality and Women’s Economic Empowerment”. The conference focused on the state of economic policy and gender issues around the world and more specifically in the ECA (Europe and Central Asia) region. The opening remarks were offered by two prominent keynote speakers – Dr. Caren Grown, Senior Director for Gender at the World Bank Group, Washington D.C, and Dr. Shahra Razavi, Chief of Research and Data at UN Women HQ in New York. The key addresses offered a global perspective on the current state of gender equality and progress made during the last 20 years. The global overview was followed by a policy panel discussion featuring prominent members of the policy-making community in Georgia. The panel participants reflected on how various policies have impacted gender (in)equality in the South Caucasus and in Georgia in particular. Later in the day, plenary presentations offered a preview of the South Caucasus Gender Equality Index, which is being developed by the ISET Policy Institute, and new research in gender economics done by academics in Georgia, Armenia, Belarus and Sweden.

The second day of the conference showcased research conducted by academics from over 15 countries covering 4 continents. It presented a range of diverse topics in gender economics, including, most prominently the links between childcare policies and labor supply decisions of women, female labor force participation (LFP) and happiness, evolving family structure and gender-selection preferences, the impact of economic, financial and public policies on women’s empowerment, the male-female earnings gap and gender aspects of international trade.

Below, we summarize the results and policy lessons that emerge from the body of work presented at the conference.

Gender Equality Progress in the ECA Region and Worldwide: Key Takeaways

First, as recent global data shows, the progress in women’s access to resources, in particular their access to the labor market, has on average stalled worldwide in the last 20 years. The labor market participation rate of women in 2018 stood at 63% globally, which is largely the same as in 1998, with some notable progress observed only in Latin America and the Caribbean (increase from 57% to 67% between 1998 and 2018), Australia and New Zealand (70 to 79%), as well as Northern Africa and West Asia (29 to 33%). The labor force participation gap between men and women is most pronounced for women who are married or in unions (44% gap, as opposed to 20% for single/never married or 17.9% for divorced/separated women).

Second, the ratio of time spent on unpaid care work by females was about 3-4 times that of males in most countries in the world, with some notable outliers: 11 times in Pakistan, 10 times in Cambodia and 9 times in Egypt. Only in Australia and New Zealand, the ratio of female to male time spent on unpaid work was slightly below 2. Thus, around the world, family responsibilities and unpaid work at home have clearly disproportionately burdened women, potentially preventing them from having an independent source of labor income, and generally weakening their financial position and bargaining power within the family unit. The recent UN Women report on Families in the Changing World (2019) argues for implementing a comprehensive package of family and women-friendly policy measures, which would include, among others, universal childhood education and care, universal healthcare coverage, long-term care for the elderly, etc. Such a comprehensive package would cost between 2-4% of GDP for most countries covered by the study. At the same time, the report argues that it would generate jobs, new investments and be a sizeable source of new tax revenue to the economies. Hence, the costs of such a program would be partially offset by the economic and tax benefits of formalizing the informal care economy. The study also details the ways in which countries could mobilize resources to pay for such packages, including improving tax collection, eliminating illicit financial flows, and leveraging aid and transfers.

For the South Caucasus in particular, the state of gender equality has not systematically been tracked until now. While there exists a number of thematic studies, surveys and narratives, as well as a more general Gender Inequality Index (GII) compiled by UNDP for all countries, a deeper systematic approach has recently been pioneered by the ISET Policy Institute, which started the ambitious project of developing a Gender Equality Index for the South Caucasus and, going forward, for the broader region of transition economies. The methodology behind the index is similar to the one adopted by the European Institute for Gender Equality, which tracks the Gender Equality Index for 28 European countries across a number of dimensions. Obviously, issues of data availability make it more challenging to build such an index in the context of transition economies. Thus, ISET-PI is working to construct some of the measures for the transition economies, using country-level data and household-level databases.

Childcare Policies and Labor Supply

One of the key messages emerging from the academic research in the area of childcare policies and labor supply was that gender-focused social policies need to be crafted carefully, with a focus on the binding constraints of the specific country context. A paper by Vardan Baghdasaryan and Gayane Barseghyan looked at how child-care service availability (affordability) affected the female labor force participation on the intensive and extensive margins in Armenia. The stage for a natural experiment in economic policy was set at the time when the Municipality of Yerevan unexpectedly decided to abolish childcare services fees (roughly 15% of average wage). The researchers hypothesized that such an intervention would have resulted in increased female LFP, as was the case in other (mostly developed) regions and countries around the world (e.g. Quebec in Canada). In the context of Armenia, however, the authors observe that there was no significant effect on female LFP rate on the extensive margin, meaning there was no evidence of inactive women entering the labor force. One possible explanation is that in the context of a developing country such as Armenia, the limiting factor to female participation in the labor force is the lack of market demand for the skills profile of non-active mothers. In such an environment, as the authors conclude, the monetary incentives do not suffice to lift the binding constraint on female LFP.

Yolanda Pena-Boquete presented a study on the case of Australia which analyzed how the labor hours and LFP of both women and men in the family are affected when either the mother’s or the father’s wages increase or when the price of childcare changes. The study finds that the mothers’ working hours respond positively and much stronger to a change in hourly wage than the fathers’. The policy implication is that an increase in mothers’ hourly wage would potentially result in a significant increase in their working hours and labor force participation. The wage effect on women’s working hours and LFP is much more pronounced even compared to the scenario when childcare prices decline.

Overall, the studies in this area demonstrated the need for a careful, multi-faceted approach in designing effective and cost-efficient labor market policies aimed at increasing labor force participation by married women with children.

Labor Force Participation and Happiness: Evidence from the South Caucasus

The paper by Norberto Pignatti and Karine Torosyan looked at the differences in the reported happiness levels between women of different labor market status in the three South Caucasus countries. The intriguing finding of the study is that while in Georgia, there is no difference in the reported happiness level between working women and housewives, in Armenia and Azerbaijan, working women with similar characteristics are much less likely to report being “very happy” than housewives. The interesting finding is that the overall results for Georgia also apply to the Armenian and Azerbaijani minority women in the country, implying that “cultural factors” may play a minor role in the reported differences between countries.

Family Structure and Gender-Selection Preferences

Gender-biased sex selection (GBSS) has been on the forefront of gender policy issues in the South Caucasus, as Armenia, Azerbaijan and, until recently, Georgia struggled with skewed sex ratios at birth (SRB). Understanding the driving forces behind GBSS, and in particular son-preference as a socio-economic phenomenon, is especially important. One of the recent studies on the issue was presented by Davit Keshelava of the ISET Policy Institute. The study “Social Economic Policy Analysis with Regard to Son Preference and Gender-biased Sex Selection” looked at the factors underlying GBSS rise and fall in Georgia over the last 15 years. The study also gleaned facts about the changing attitudes towards GBSS and son-preferences in different regions of Georgia. One of the study’s main findings is that the fall in the sex ratio at birth has been statistically significantly correlated with real income growth in the regions, reduction in poverty, and female employment. Among other factors significantly affecting the reduction in sex ratio at birth, was, surprisingly, the level of male education, while female education was statistically insignificant. The study documented a persisting son preference in Georgia, but also high awareness and strong negative attitudes towards gender biased sex selection in those regions that showed the sharpest improvement in sex ratio at birth over time.

Looking at the issue of gender preferences in the context of transition economies in Europe, Izabela Wowczko presented joint work with Michał Myck and Monika Oczkowska which investigated how preferences for the gender composition of children in the family might have changed in Central and Eastern European (CEE) countries after the fall of communism. The results showed that gender-neutrality was observed in almost all CEE countries before the transition. After the transition of the 1990s, many of the same forces which operated in the South Caucasus have affected the countries of Central and Eastern Europe – namely, decline in incomes, decimated traditional social safety nets and better access to ultrasound and family planning technologies. However, in the post-transition CEE countries, the authors observe a clear preference for a mix (boy/girl) or possibly boys at parity three (i.e. having two boys or a boy and a girl in the family reduced the likelihood of having a third child significantly, as opposed to having two girls). It was also observed that in most CEE countries (except Romania), there was an increased likelihood of having a second child if the first child is a boy – thus demonstrating a girl preference at parity two.

Policy Impact on Women’s Empowerment

A study from India by Mridula Goel and Nidhi Ravishankar looked at the impact of policy interventions on the long-term indicators of women empowerment. It shows that public policies were responsible for improving the so-called “power enablers”, such as literacy rates, financial access, property rights, political voice, etc. However, there is some evidence that not all traditional power enablers, e.g. having a bank account or working for money, are correlated with higher indicators of empowerment, measured by a woman’s autonomy in decision-making within the family. For example, working for money (receiving cash compensation) or having a bank account was found to be negatively correlated with a woman’s ability to decide how her own money is spent – possibly pointing to the existence of prejudice or negative attitudes within the household in such cases.

Another interesting study on this topic by Maria Perrotta Berlin, Evelina Bonnier and Anders Olofsgård looked at whether foreign aid projects foster female empowerment in the surrounding community using data from Malawi. It finds support for a small positive impact of aid on men’s and women’s attitudes related to domestic violence and sexual rights. There is, however, little systematic difference in the impact of gender-targeted aid versus general aid – with exceptions being the impacts on women’s experience of violence and women’s participation in decision-making.

Male-Female Earnings Gap and Gender Aspects of International Trade

The male-female earnings gap is a recurring topic in gender economics. Whether the gap is driven by differences in education and skills of men and women, labor market discrimination, choices of working hours, the “glass ceiling” or “sticky floor” phenomena, the gap is evident and persistent in both developed and developing countries. One of the papers presented by Dagmara Nikulin looked at the impact of trade liberalization on the gender wage gap in Europe. Generally, the economic literature does not provide conclusive evidence in this regard, and the link remains ambiguous. The paper, examining evidence from Europe, finds in particular that participation in global value chains (GVC), which the authors measure by foreign value added in exports, is correlated with reduced wages overall, but the negative effect on wage is lower for men than for women.

Echoing the results of the previous study, the paper by Marie-France Paquet and Georgina Wainwright-Kemdirim, “Since the effects of trade liberalization are not gender neutral, how can we improve its gender outcome? – Crafting Canada’s Gender Responsive Trade Policy” focuses on the problem of identifying and addressing potentially negative impacts of trade on female jobs. The study details a diagnostic modelling approach, which is to use CGE modeling combined with sectoral employment data (a labour module within CGE). The proposed model uses an overlapping generation framework and includes an occupational matrix to allow movements between occupations. This approach allows for specific potential impacts of generic FTAs by gender, age group and occupation.

Conclusion

To sum up, the first international academic conference on gender economics issues in Tbilisi highlighted the diversity and complexity of gender issues around the world and in the South Caucasus region in particular. It also became a powerful catalyst for new research and collaboration ideas among participating institutions and individual researchers. Finally, it demonstrated how policy-oriented research can help inform the policy-making community about the areas where intervention is most needed, design the most effective policies, and calculate the associated costs and benefits of interventions.

References to Selected Presentations

  1. Shahra Razavi “Policies for Gender Equality in an Unequal World: Challenges and Opportunities”, keynote presentation.
  2. Vardan Baghdasaryan and Gayane Barseghyan “Child Care Policy, Maternal Labor Supply and Household Welfare: Evidence From a Natural Experiment”.
  3. Michal Myck and Kajetan Trzcinski “From Partial to Full Universality: the Family 500+ Programme in Poland and its Labour Supply Implications”.
  4. Karen Mumford, Antonia Parera-Nicolau, Yolanda Pena-Boquete “Labour Supply and Childcare: Allowing Both Parents to Choose”.
  5. Norberto Pignatti, Karine Torosyan “Employment vs. Homestay and Happiness of Women in the South Caucasus”.
  6. Davit Keshelava et al. ISET Policy Institute Report “Social Economic Policy Analysis with Regard to Son Preference and Gender-biased Sex Selection”.
  7. Izabela Wowczko, Michał Myck and Monika Oczkowska “Gender Preferences in Central and Eastern Europe as Reflected in Family Structure”.
  8. Mridula Goel, Nidhi Ravishankar “Has Public Policy Succeeded in Enhancing Women Autonomy and Empowerment in India Over the Last Decade?”.
  9. Maria Perrotta Berlin, Evelina Bonnier and Anders Olofsgård “The Donor Footprint and Female Empowerment”.
  10. Dagmara Nikulin & Joanna Wolszczak-Derlacz “Gender Wage Gap and the International Trade Involvement. Evidence for European workers”.
  11. Marie-France Paquet, Georgina Wainwright-Kemdirim, “Since the Effects of Trade Liberalization are not Gender Neutral, How can we Improve its Gender Outcome? – Crafting Canada’s Gender Responsive Trade Policy”.

A Decade of Russian Cross-Domain Coercion Towards Ukraine: Letting the Data Speak

Posted on February 17, 2020 | Policy Brief

20200217 A Decade of Russian Cross-Domain Coercion FREE Network Policy Brief Image 01

Russia’s coercion towards Ukraine has been a topic of major international events, meetings and conferences. It regularly makes the headlines of influential news outlets. But the question remains open – do we really understand it? We diligently collect and analyze data to make informed decisions in practically all domestic issues but is the same done for international relations? This research paper introduces a number of tools and methods that could be used to study Russia’s coercion towards Ukraine beyond its most visible manifestation, looking into latent trends and relations that could reveal more.

Introduction

For the past decade, Ukraine has been in the headlines of the major world news outlets more frequently than ever before. Ukrainian-Russian relations have been and still remain the topic of international summits and other events. The occupation of a part of Ukraine’s territory has been denounced and Russia’s coercion towards Ukraine is now generally accepted as a fact. But what do we really know about the underlying empirics and dynamics and how can this multi-domain assertiveness be measured and tracked? This paper presents a number of data-driven approaches that allow looking beyond the headline stories to identify and track various dimensions of Russia’s coercion towards Ukraine and the dynamics of its development.

Academic Interest

Mapping the landscape of scholarly literature reveals a number of interesting results. First, the body of works studying Russia’s coercion towards Ukraine remains relatively modest. It quintupled in 2014 but afterwards the interest started tapering off. A search for papers on this topic in Scopus and Web of Science with a very precise query (to increase the accuracy of search) and publication time of 2009-2019 returned 155 papers most of which were published in or after 2014.

Figure 1. Scholarly publications on Russian-Ukrainian relations.

Source: WoS and Scopus, 2009-2019

A closer look at the content of these works with the use of a bibliometric software called CiteSpace shows that the majority of papers focus on Putin, once again emphasizing the significant role of his personality in Russia’s coercion towards Ukraine. The second largest cluster has the “great power identity” as its main theme and presumably looks beyond actions of Russia to identify its ideological grounds. Another group of publications is devoted to sanctions, pointing to their important role in studying Russian-Ukrainian relations.

Figure 2. The landscape of topics in scholarly publications on Russian-Ukrainian Relations.

Expressions of Coercion

The “practical” side of Russia’s coercion towards Ukraine is also frequently associated with the personality of Vladimir Putin and his attitudes towards Ukraine. To analyze this perception further, we created a corpus of speeches of Russian presidents published on the Kremlin website, filtered them to keep only those that mention Ukraine, divided them into pre-2014 and 2014 and after, and then analyzed them using an LDA topic modeling algorithm. This algorithm is based on the assumption that documents on similar topics use similar words. So, the latent topics that a certain document covers can be identified on the basis of probability distributions over words. Each document covers a number of topics that are derived by analyzing the words that are used in it. In simple terms, the model assigns each word from the document a probabilistic score of the most probable topic that this word could belong to and then groups the documents accordingly.

Figure 3. Speeches of Russian presidents before 2014, LDA topic modeling.

Figure 4. Speeches of Russian presidents in 2014 and after, LDA topic modeling.

Quite surprisingly, we discovered that the overall rhetoric of speeches is very similar for the two periods. Although some speeches do differ and the later corpus includes new vocabulary to reflect some changes (i.e “Crimea”, “war”) the most common words remain practically the same. Thus, regardless of the apparent shift in relations between the two countries, Russian leadership still relies on the same notions of collaboration, interaction, joint activities, etc. The narrative of “brotherhood” between the nations persists despite and beyond the obvious narrative of conflict.

To include a broader circle of Russia’s leadership we also looked at the surveys of the Russian elite conducted regularly by a group of researchers led by William Zimmerman and supported by various funders over the years (in 2016 – the National Science Foundation and the Arthur Levitt Public Affairs Center at Hamilton College). Seven waves of the survey already took place; the most recent one in 2016. The respondents were the representatives of several elite groups (government, including executive and legislative branches, security institutions, such as federal security service, army, militia, private business and state-owned enterprises, media, science and education; for practical reasons from Moscow only).

The survey revealed a number of interesting observations. For instance, while the prevailing Russian opinion on Russia’s occupation of Crimea had been that it was not a violation of international law, a closer look at the demographic characteristics of respondents shows that they were not as coherent as it might seem from the outside. While the “green” answers from respondents with backgrounds such as media or private business may have been anticipated, the number of members of the legislative and especially executive branch and the military that had at least some doubt on the legality was surprisingly quite sizable, and they even demonstrated some support of the “violation of law” interpretation.

Figure 5. Elite and public opinion on Russia’s annexation of Crimea.

Comparing these elite opinions to the public opinion poll by Levada Center conducted in the same year shows that even the general public is slightly more likely to choose the most extreme “full legality” option than the respondents from the executive branch.

Beyond the elite or general opinion polls, we tried to develop a metric that might allow us to track Russian sensitivities towards Ukraine. For that, we examined two different ways of expressing “in Ukraine” in Russian language: ‘на Украине’ (the ‘official’ Russian expression) vs. ‘в Украине’ (the version preferred by Ukrainians). [In English, this can be compared so saying ‘Ukraine’ vs ‘the Ukraine’.]

Our first visual plots how many search queries were done on Google Search with both versions over the last decade.

Figure 6. Search queries for “в Украине” (green) versus “на Украине” (red), Google Trends, 2009-2019.

We can clearly observe that during less turbulent times the more politically sensitive version is much more common. This however drastically changes during the peaks of Russia’s coercion towards Ukraine when the number of searches with the less politically correct term increases significantly.

A different trend can be observed if we look at official media publications stored in the Factiva database. We estimated the ratio of search volumes for each term and observed that until the beginning of 2013, about a third of articles and news reports used “in Ukraine”. This changed around January 2013 when the ratio starts to decrease for “in Ukraine” searches and plummets to a mere 10% of outlets still preferring this term.

Figure 7. The ratio of “в Украине” to “на Украине” occurrences in large Russia media (2009 – 2019), Factiva.

Tracking Coercion Itself

What is the track record of Russia’s actual coercion over this decade? For this, we turn to a few recent datasets that try to systematically capture verbal and material actions (words and deeds): the automated event datasets. The largest one of those, called GDELT (Global Database of Events, Language, and Tone), covers the period from 1979 to the present, and contains over three quarters of a billion events. It is updated every fifteen minutes to include all “events” reported in the world’s various news outlets. To exclude multiple mentions of the same event by one newswire, the events are “internally” deduplicated. The events are not compared across newswires.

An event consists of a “triple” coded automatically to represent the actor (who?), the action (what?) and the target (to whom?) as well as a number of other parameters such as type (verbal or material; conflict or cooperation; diplomatic, informational, security, military, economic), degree of conflict vs cooperation etc. Other similar datasets include ICEWS (Integrated Crisis Early Warning System) and TERRIER (Temporally Extended, Regular, Reproducible International Events Records). For this analysis, we filtered out only those events in which Russia was the source actor and Ukraine was the target country. We present two metrics: (1) the percentage of all world events that this subset of events represents and (2) the monthly averages of the Goldstein score, which captures the degree of cooperation or conflict of an event and can take a value from -10 (most conflict) to +10 (most cooperation). Also, to add a broader temporal perspective, we looked beyond the last decade. It can be clearly seen that the number of events before 2013 was significantly lower, especially in “material” domains. Some verbal assertions from Russia towards Ukraine happened during the Orange Revolution and so-called “gas wars”.

The situation changes radically starting from 2013. The proportion of events increases with some especially evident peaks (i.e. during the occupation of Crimea). The verbal events remain quite neutral while the actions towards Ukraine move from some fluctuations to steadily conflictual.

Figure 8. Russia’s negative assertiveness towards Ukraine, 2000-2019.

Measuring Influence

We have seen that the past decade was exceptional in the scale of Russian assertiveness towards Ukraine. But what do we know about Russia’s influence on Ukraine and Ukraine’s dependence on Russia? Influence measures the capacity of one actor to change the behavior of the other actor in a desired direction. In an international context this often concerns the relations between countries. Influence can be achieved by various means, one of which is to increase the dependence of the target country upon the coercive one. This strategy is frequently employed by Russia willing to regain and/or increase control over the former post-Soviet countries. The Formal Bilateral Influence Capacity (FBIC) Index developed by Frederick S. Pardee (Center for International Future) looks at several diplomatic (i.e. intergovernmental membership), economic (trade, aid) and security (military alliances, arms import) indicators allowing to identify the level of dependence of one country upon another. This is especially interesting from a comparative perspective. Figure 9 shows that countries such as Armenia and Belarus remain highly dependent on Russia. For half of the decade, Ukraine was number three on this list. Today the situation has changed. Ukraine’s dependence on Russia has gradually decreased and has become even smaller than Moldova’s, moving closer to the steadily low level of dependence of Georgia. This may signify a positive trend and a break of a decade-long relationship of dependence.

Figure 9. Dependence of post-Soviet countries on Russia, FBIC.

Conclusion

Consequently, Russia and Ukraine have become much more visible in the international academic and policy research efforts. This can be measured through a number of instruments, including a comprehensive mapping of the academic landscape itself with regard to salience and topics that are being studied, analysis of the word choice (that could be represented by the use of the terms to describe events in Ukraine by the government media and Google search users (“на Украине” versus “в Украине”); speeches of Russian presidents that use the same rhetoric of collaboration when talking about Ukraine despite the obvious change in relationships) and material coercion (significant increase in number of assertive conflictual Russia’s actions towards Ukraine). Some findings do give hope for change: the opinions of the Russian elite on recent Russian actions towards Ukraine while remaining generally unfavourable are not as cohesive as it might appear and Ukraine’s dependence on Russia has decreased significantly.

Disclaimer

This research is a part of a larger research effort titled RuBase funded by the Carnegie Foundation of New York and implemented jointly by The Hague Centre for Strategic Studies and Georgia Tech with the help of the Kyiv School of Economics StratBase team in Ukraine. The ‘Ru’ part of the title stands for Russia; and ‘base’ has a double meaning – both the knowledge base built during the project, and the (aspirationally) foundational nature of this effort. The project intends to look beyond the often-shallow traditional understanding of coercion and apply innovative tools and instruments to study coercion in its multifaceted form. This is only a small selection of the tools that have been successfully tested in the course of this (ongoing) research project and applied to the study of Russia’s coercion in different domains. The prospects of any progress in resolving the Russian-Ukrainian conflict are currently slim, thus further work that would allow identifying patterns and trends that the human eye may oversee to understand Russia better and develop an informed foreign policy strategy both for Ukraine and the West is crucially important.

References

Indexation Formula for Natural Gas Procurement in Ukraine

Posted on February 10, 2020 | Policy Brief

Natural Gas power plant

Due to the high volatility of natural gas market prices, it is almost impossible to adequately plan the purchases for the year ahead, so contract prices need to be regularly updated. This fact creates uncertainty for the contracting authorities, as well as room for unfair competition and corruption. We offer an indexation formula which uses the European gas prices as a benchmark for procurement prices and calculate the potential economic effect of this formula on the Ukrainian gas procurement market.

Problems with the Public Procurement of Natural Gas

Natural gas procurement poses a number of challenges for the contracting authorities (CAs), suppliers and controllers. Due to price volatility it is almost impossible to adequately plan the purchases for the year ahead, so prices need to be regularly adjusted. After the heating season starts, CAs find themselves in a weak position in price negotiations since they almost never have storage for accumulating stocks, and if the contract is cancelled the new procurement would take at least one month due to the existing public procurement regulations. The new version of the Law on Public Procurement, which was recently adopted by the Ukrainian parliament, addresses this problem by allowing CAs to have a new contract fast in case the previous contact was cancelled because of the supplier.

CAs often lack reliable data on market dynamics. There are cases when unreliable price references are provided by specialized agencies to support higher price claims of suppliers. As a result, CAs bear administrative responsibility if they do not have proper justification for changing contract prices when controlling agencies initiate an audit.

Natural gas suppliers may also find themselves in a situation of unfair competition. Since it is possible to win an open auction (i.e. by quoting a considerably lower than market price) and later raise the price to the market level with an additional contract, honest businesses might feel demotivated to participate in the procurement process. They cannot be sure if the contract price can be changed later because there is no proper legal mechanism to assess the need of such an adjustment.

Previous research shows that every third contract of natural gas purchase was amended with an additional contract at least once, usually raising the price for the customer (Shapoval, Memetova, 2017). Additional contracts are indeed used 1) as a tool for price overstatement by the supplier, 2) as a loophole for corruption, and 3) as a way to get a market price for a supplier who used dumping to win the auction (Gribanovsky, Memetova, 2017).

International Drivers of Gas Prices in Ukraine

Since 2016, the EU has been the only official exporter of natural gas to Ukraine. According to PwC, Ukraine imported 14.1 billion m3 in 2017, which is 44% of total gas consumption – the remaining 56% are extracted locally. In Ukraine, the prices for industrial consumers are not regulated, while the household prices are set by the government. Today, the average price on the unregulated gas market is in line with the prices in neighbouring countries – the Baltic states, Poland, Slovakia and Hungary (PricewaterhouseCoopers Advisory LLC, 2018).

European prices are formed on the large marketplaces. The two biggest hubs, Dutch TTF and British NBP, by far outweigh their competitors (ACER Market Monitoring Report, 2017). However, the third-biggest hub, German NCG, is the closest to the Ukrainian border, so its prices often become the benchmark for private traders. In some cases, NCG is the official benchmark for gas price – for instance, the purchase parity import price in Ukraine in 2017-2018 was based on this hub’s index.

In order to assess the impact of the European natural gas prices on procurement prices in Ukraine, we used the Month+1 futures hub prices from TTF, NCG and Austrian VTP CEGH. Procurement prices were extracted from the analytical module of ProZorro (the Ukrainian e-procurement system which CAs are obliged to use at all levels). We excluded irrelevant procurements and selected the contracts which had information on the volume procured. We calculated the average daily prices weighted by volume. Our dataset covers the time period from January 1, 2017 to December 31, 2018.

Figure 1: Natural gas prices at ProZorro and European hubs

Source: ProZorro data, hubs data

As one can see in Figure 1, hubs prices are highly correlated, so they cannot be used as independent variables within a single model. Thus, we decided to take the NCG Month+1 price as a benchmark for explaining the relation between internal procurement prices and international market prices.

NCG Impact on Procurement Prices in ProZorro

In the period of low business activity on natural gas markets, especially in summer, few contracts are awarded. One might have noticed from Figure 1 that this leads to higher variance in daily prices caused by random factors. Therefore, in our model we decided to use the weighted average of weekly prices instead.

Figure 2: Weekly gas price fluctuations in ProZorro and NCG

Source: ProZorro data, Pegas (https://www.powernext.com/futures-market-data)

Our econometric estimation shows that the NCG Month+1 price influences procurement prices with a lag of 7 weeks. In other words, the price at the German hub becomes relevant for the Ukrainian procurement market after almost 2 months on average.

Figure 3: Correlations between procurement prices and NCG Month+1 with different lags

According to the model, the weighted average gas price in ProZorro is more dependent on the NCG Month+1 gas price than on the reservation price in ProZorro. Thus, a UAH 1 increase in the reservation price adds UAH 0.41 to the final price, while each additional hryvnia of the NCG price leads raises the final price by UAH 0.63 in 7 weeks if the price growth trend is not taken into account.

Potential Cost-Saving Using the Price Indexation Formula

The Price Indexation Formula

As European gas prices strongly influence prices on the internal Ukrainian market, it is obvious that they should be included into the indexation formula, as well as exchange rate fluctuations. After consultations with stakeholders, the Ministry of Economic Development and Trade of Ukraine (MEDT) decided to adjust the initial formula proposed by the KSE and included price fluctuations on the Ukrainian Energy Exchange (UEEx) with a small weight into the formula in order to stimulate UEEx development.

The final formula was officially published in December 2018. This formula is not compulsory for any contract authorities, though it is recommended for use by the smaller public entities who do not have the in-house analytical capacity to make a realistic price assessment during negotiations with the suppliers.

where:

  • CP – new price in UAH for 1000 m3 of natural gas (including value-added tax, VAT)
  • PCP – current price in UAH for 1000 m3 of natural gas (including VAT) before adjustment
  • K(cur) – average National Bank of Ukraine (NBU) UAH/EUR exchange rate for 5 days before the price change
  • K(base) – average NBU UAH/EUR exchange rate on the day of the previous price adjustment (contract signed)
  • NCG(avg) – average of daily NCG Month+1 index during 20 previous trading days before the day of price amendment, EUR per MW-hour
  • NCG(base) – NCG Month+1 index on the day of the previous price amendment (contract signed), EUR per MW-hour
  • VAT – rate of value-added tax, which is currently 20% in Ukraine
  • CV – heating value of natural gas in MW-hour/1000 m3 on the date of the price adjustment
  • UEEx(avg) – weighted monthly average natural gas price of UEEx (including VAT) on the day of price amendment
  • UEEx(base) – weighted monthly average natural gas price on the UEEx (including VAT) on the day of the previous price amendment (contract was signed)

Thus, the formula includes current gas price, exchange rate changes, changes in NCG index and UEEx index.

Estimation of Potential Cost-Saving for Contract Authorities

The simplest yet time-efficient way to empirically verify the hypothesis of potential cost-saving after the introduction of the price indexation formula in the gas market is a retrospective analysis of the contracts which had already been signed.

The basic principle of estimation is comparing actual prices with the potential prices calculated based on the price indexation formula. For this, we collected a dataset of natural gas procurement contracts covering the time period from August 2017 to the end of August 2018. This period includes both short-term contracts signed for the heating season or its part (usually signed in August-September, sometimes in January-February) and middle-term contracts which are active for at least one year (usually signed in December-March). We took into account all the additional contracts to these contracts signed before January 1, 2019.

Supply schedules and prices of additional contracts are not readily available in a machine-readable format, so we kept only contracts with the total value higher than UAH 1 million. These are 27.5% of all contracts but they cover 79.3% of the total value of natural gas procurement in Ukraine. The final dataset contains prices of additional contracts and monthly supply schedules.

Our earlier analysis of all the contracts on the shorter time scale showed no correlation between prices and volumes in gas procurement contracts (Shapoval, Memetova et al., 2017), therefore our results can be extrapolated to all the gas contracts.

The biggest gap between actual and indexation prices would be in November 2017, averaging UAH 623. However, until the end of the year the gap reduced threefold to UAH 170.

Figure 4: Monthly increase of gas procurement prices

Source: bipro.ProZorro, ProZorro API

We combined the supply schedules with the prices found in the additional contracts in order to estimate potential savings. Obviously, the highest savings were observed during the heating season. However, in September they were negative (see Figure 6). Thus, while the market prices of natural gas started rising in August, actual procurement prices lagged behind until the end of September-October.

Figure 5: Monthly cost savings in case of applying price indexation formula

Source: bipro.ProZorro, ProZorro API

In total, for contracts of over UAH 1 million, potential cost savings from applying the price indexation formula would have been equal to UAH 120.25 million. If these estimations are extrapolated to all the contracts, this figure would reach UAH151.6 million. This is a rather modest sum in relative terms – only 2.7% of the total contract value. However, using the formula is expected to assist smaller CAs who often lack the knowledge of market dynamics to negotiate the optimal price more effectively and limit their dependence on the suppliers’ estimates.

Besides, the parties concluded the contracts without taking into account the opportunity of using the indexation formula. Therefore, actual cost savings might be lower, first of all because the suppliers’ auction strategy would be different. In particular, the dumping strategy with subsequent price increase through additional contracts would become useless. If the formula is used, a lower starting price would mean a lower increase in absolute terms (UAH per 1000 m3), because the formula calculates the change in relative terms (in per cent). For example, if the market price grows by 15% during the indexation period, the starting price can also be raised by only 15%.

Conclusion

The application of the price indexation formula for natural gas procurement may have a positive impact on the public procurement market. We recommend taking into account the prices of the European hubs adjusted by exchange rate fluctuations.

Had the price indexation formula been used for additional contracts in gas procurement in 2017-2018, the average price would have declined by UAH 623, potentially allowing CAs to save UAH 151.6 million.

Formula pricing would raise the negotiation power of customers (CAs) before the start of the heating season. This is especially true for the smaller ones which are not able to professionalize procurement processes. Natural gas price indexation within clearly defined boundaries will create more favourable conditions for fair competition by eliminating the stimuli for dumping at the auction stage.

References

Data Sources

Does Gender Diversity Actually Matter?

Posted on February 3, 2020 | Policy Brief

20200203 Does Gender Diversity Actually Matter FREE Network Image 03

Measuring the effects of gender diversity on performance is important to understand the impact of gender quotas. However, the effects of gender diversity remain understudied. We need data with a reliable assessment of team member quality to disentangle the effects of diversity from compositional effects (when higher-quality women replace mediocre men). We use the unique database of the trivia game “What? Where? When?” which has information on both the performance and gender composition of the team and allows to track each player individually. We find that the gender diversity of the team has no statistically significant effect once we control for the quality of each player. In this particular environment, with little evidence of gender discrimination, instruments like gender quotas have no merit. This result does not apply to discriminatory environments where gender quotas could bring benefits through compositional effects.

Introduction

As gender quotas have been widely introduced in politics and in the corporate world, the effects of gender diversity have become the center of attention of many economists. Many observational studies find positive effects of gender diversity on corporate boards’ performance (Desvaux, Devillard, & Sancier-Sultan, 2010). Other studies, using the introduction of gender quotas in boards as a natural experiment, find negative effects on stock valuation, which disappear in the longer run (Ahern and Dittmar, 2012; Matsa and Miller, 2013; Eckbo et al., 2019).

The effect of gender diversity on team performance may run through two different mechanisms. One mechanism is compositional effects due to discrimination: if women face a glass ceiling, only the best women get into teams/boards, and they are on average of higher quality than men. Hence, boards with female representatives perform better. The discrimination mechanism has been shown to be at work in the political setting, for example: gender quotas in parties lead to higher-quality women replacing mediocre men (Besley et al., 2017). The other mechanism is the true effect of gender diversity through complementarity between men and women: if they differ substantially in some dimensions, these differences might become the source of better team decisions, or, on the contrary, inefficiencies in decision-making.

To separate between the two mechanisms – compositional effects and diversity effects – we need data with reliable quality measurement for each team member. Controlling for team member quality would take care of the compositional effect, and the gender composition would be significant only if there is a true gender diversity effect.

We use the What? Where? When? trivia game dataset to measure the effects of gender diversity on team performance with and without control for a player’s quality.

The What? Where? When? Game

What? Where? When? (WWW) is a team-played trivia game popular in post-Soviet countries. Teams of six players are asked questions and have one minute to come up with an answer. Typically, in order to find the correct answer, a team needs to combine both logical thinking and knowledge. A tournament usually consists of 36-90 questions. The team with the most correct answers wins the first place. In 2003, a unified database of the game was created. This database contains records of more than 218,000 individuals who have played in at least one of the 6,000 recorded tournaments.

The What? Where? When? Dataset

A unit of observation in our dataset is one game played by a team. It contains the unique ID of the team, the ID of each player, information about the number of games played by the team and by each player, the tournament date, the difficulty of the tournament and the number of teams. We identify the gender of the players through their names and patronymic names. Overall, we use 74,475 team-game observations which were played by 2,854 teams (23,000 single players) from 2013 to 2018.

Performance Measure

The measure of a team’s performance in a tournament is the percentage of correct answers normalized by the average percentage of correct answers in this tournament. We use player’s individual fixed effects as a measure of their quality in our regression analysis.

Gender Aspects in What? Where? When?

Only 31.5% of the players in the sample are female, however, other than that, we fail to find any significant evidence indicating gender discrimination or segregation. Table 1 presents the actual shares of team-game observations by gender composition as well as the predicted shares if assignment to teams was random. The difference between the actual shares and predicted shares does not appear to be economically significant.

Table 1: The actual distribution of women across teams is not different from random

Source: Authors’ calculations based on the What? Where? When? dataset. Random assignment assumes that the share of women across all teams is equal to 31.5% as in the actual data.

Results

The basic model of our analysis, Model 1 examines the association between the performance of a team, normalized by tournament difficulty, with dummy variables for gender diversity (defined as the number of minority gender players in the team, i.e. diversity_1 is true if there is only one woman or only one man on the team). We also include the individual fixed effects of each player in the second specification (Model 2), to control for the quality of players and rule out possible composition effects.

Table 2. Effect of diversity on performance with and without the individual quality controls

Source: Authors’ calculations based on What? Where? When? dataset. Individual fixed effects are included in the specification with the quality control. Only players who played at least a median number of games (62) are included.

The coefficients of Model 1 and 2 are shown in Table 2. While diversity is significant in the first specification, after accounting for the individual quality of players, we cannot reject the hypothesis of insignificance of gender diversity. These results hold under different specifications: with controls for player experience, with different player experience cutoffs, or including the neural network-generated predictions of performance.

Figure 1. The distribution of individual coefficients (proxy for player quality) for female and male players

Source: Authors’ calculations based on the What? Where? When? dataset. Each individual coefficient is a proxy to the player’s quality estimated in the regression from Table 2. Only players who played at least a median number of games (62) are included.

Figure 1 presents the distributions of individual coefficients of female and male players. In our sample, the female distribution centers slightly to the left of the male one. It explains the negative diversity coefficients in the specification without the individual fixed effects – in this case, the diversity dummies capture the lower average quality of female players.

Conclusion

Our study aimed at disentangling compositional and pure effects of gender diversity by using a novel dataset of a team played trivia game. Our main finding is that after accounting for the individual quality of team members, the gender composition of a team does not appear to be significant for a team’s performance.

Although it is always dangerous to extrapolate findings obtained in specific settings, we believe that the positive gender diversity effects found in other studies are often manifestations of the change in the average quality of team/board members i.e. compositional effects rather than gender diversity effects per se. From a policy point of view, this means that while we need gender quotas in areas suffering from gender discrimination, once we reach equal opportunities such instruments may no longer have any positive effects.

References

  • Ahern, Kenneth R., and Amy K. Dittmar, 2012. “The changing of the boards: The impact on firm valuation of mandated female board representation.” The Quarterly Journal of Economics 127.1: 137-197.
  • Besley, Timothy, Olle Folke, Torsten Persson, and Johanna Rickne, 2017. “Gender quotas and the crisis of the mediocre man: Theory and evidence from Sweden.” American Economic Review 107, no. 8 : 2204-42.
  • Desvaux, Georges, Sandrine Devillard, and Sandra Sancier-Sultan, 2010. “Women at the top of corporations: Making it happen.” McKinsey & company : 7-8.
  • Eckbo, B. Espen, Knut Nygaard, and Karin S. Thorburn, 2019. “Board Gender-Balancing and Firm Value.” Dartmouth College working paper.
  • Matsa, David A., and Amalia R. Miller, 2013. “A female style in corporate leadership? Evidence from quotas.” American Economic Journal: Applied Economics 5.3: 136-69.

Ethnic Geography: Measurement and Evidence

Posted on January 27, 2020 | Policy Brief

Crowd of people crossing street representing Ethnic Geography - Measurement and Evidence

The effects of ethnic geography, i.e. the distribution of ethnic groups across space, on economic, political and social outcomes, are not well understood. We develop a novel index of ethnic segregation that takes both ethnic and spatial distances between individuals into account. Importantly, we can decompose this index into indices of spatial dispersion, generalized ethnic fractionalization, and the alignment of spatial and ethnic distances. We use ethnographic maps, spatially disaggregated population data, and language trees to compute these four indices for 161 countries. We apply these indices to study the relation between ethnic geography and current economic, political and social outcomes. We document that country level quality of government, income and trust increase with the alignment component of segregation.

Ethnic Geography: Key Idea

There is a vast literature on how a country’s ethnic diversity affects economic, political and social outcomes. This literature provides evidence for negative effects of ethnic diversity on e.g. peace, public goods provision, redistribution, the quality of government, and economic development in general. In these studies, ethnic diversity is typically quantified by indices based on the different ethnic groups’ country-wide population shares. By definition, these indices ignore ethnic geography, i.e. the distribution of ethnic groups across space.

Alesina and Zhuravskaya (2011) make an important first step towards taking ethnic geography into account. They construct an index of ethnic segregation that is based on the various ethnic groups’ population shares in different subnational units such as regions or provinces. We contribute to the literature on ethnic diversity by proposing a set of indices that capture important aspects of ethnic geography.

Theory

We derive a new segregation index that is based on both spatial and ethnic distances between pairs of individuals. Starting from a general class of indices that are expressions of the relation between a randomly selected pair of individuals, we uniquely characterize an index via a set of axioms. Our index avoids the standard problems of the so-called a-spatial segregation measures (based on population shares in administrative units), in particular the border dependence mentioned by Alesina and Zhuaravskya (2011) and the checkerboard problem (White 1983, Reardon and O’Sullivan 2004). Both problems are potentially very severe and are illustrated in detail in the paper. Importantly, our index can be decomposed into three (sub)indices: an index of spatial dispersion, a well-known index of generalized ethnic fractionalization (see below), and a measure of the alignment of spatial and ethnic distances between individuals (i.e. ethno-spatial alignment or, simply, alignment hereinafter). Interested readers can also find a stylized illustration of what each component stands for in the paper.

Data and Illustration

We compute these four indices of ethnic geography for 161 countries from all over the world using a combination of digital maps showing the distribution of ethno-linguistic groups all over the world and the current population at a very high resolution. As robustness check, we also compute our measure using historical population maps as well as a simpler map based on global land cover data that should proxy for the exogenous component of the spatial distribution of a country’s population.

This provides us with real world examples of countries that differ in, for example, alignment, but are otherwise similar. For example, Togo and Benin are neighboring countries located in West Africa with comparable climatic, geographic and demographic characteristics. Moreover, they were both French colonies after WWI, became independent in 1960, and started their post-colonial history in tumultuous ways that culminated in coups. The ruling autocrats both managed to stay in power for many years.

Benin and Togo are also comparable in terms of generalized ethnic fractionalization (between the median and the third quartile of our sample) and spatial dispersion (above the third quartile). Ethno-spatial alignment is however considerably higher in Benin (1.35, which is above the third quartile) than in Togo (1.11, which is below the median). Figure 1 shows the different ethnic homelands and the main language groups which these ethnic homelands belong to.

Ethno-spatial alignment is relatively high in Benin as there is a relatively clear divide between Kwa speaking groups in the South, Defoid speaking groups in the Center, Gur speaking groups in the North, and some smaller groups speaking very different languages in the North East. As a result of this divide, linguistically distant individuals tended to live far apart from one another. In contrast, ethno-spatial alignment is relatively low in Togo, mainly because there are Gur and Kwa speaking groups in the country’s South, its Center and its North. As a result of these large and widespread language groups, linguistically distant individuals often lived relatively close to one another.

Figure 1:

Source: Maps of Togo (left) and Benin (right) showing the traditional homelands of language groups according to WLMS and our grid cells. Each grid cell constitutes a different location in the computation of our indices, each color indicates that the corresponding grid cell belongs to the traditional homeland of a certain language group (with the relevant language groups given in the legend), and the brightness of this color indicates the current population (also given in the legend). The legend entries Gur/Kwa and Gur/Defoid indicate the traditional homelands of multiple language groups, some speaking a Gur language and some a Kwa or Defoid language. WLMS indicates no traditional homelands in the white areas.

Empirical Analysis

We use our indices in cross-country regressions to improve our understanding of the role that ethnic geography plays in economic, political and social outcomes around the globe. We first focus on the associations between our index of ethnic segregation on the one hand, and the quality of government, incomes and generalized trust on the other hand. We find a negative relation between ethnic segregation and the quality of government, similar to Alesina and Zhuravskaya (2011) with their index of a-spatial segregation in their sample of 97 countries. We further find that our index of ethnic segregation tends to be negatively associated with incomes, too, but unrelated to generalized trust. See also Ejdemyr et al. (2018) and Tajima et al. (2018) for recent contributions on Malawi and Indonesia.

More importantly, we study the relation between the three components of our index of ethnic segregation – ethnic fractionalization, spatial dispersion and ethno-spatial alignment – and these outcome variables. Strikingly, we find a positive and statistically significant association between the alignment of ethnic and spatial distances between individuals, and the quality of government, incomes and trust. Hence, societies perform better when ethnically diverse people live far apart, relative to what they would have, had all ethnic groups been represented in all locations with population shares equal to their country shares.

Conclusion

To better understand the role of ethnic geography and to mitigate well-known problems of a-spatial segregation measures, we have developed a new segregation index that is based on ethnic distances between groups and spatial distances between locations rather than categorical data on ethnic groups and administrative units.

The decomposition of our segregation index reveals that it corresponds to the product of generalized ethnic fractionalization, spatial dispersion, and the alignment between ethnic and spatial distances. This ethno-spatial alignment is a novel concept that captures, broadly speaking, whether ethnically different individuals tend to live far from each other, relative to the situation where all groups appeared in each location with population shares equal to their country ones.

Using these indices in cross-country regressions suggests, among other things, that countries with higher ethno-spatial alignment tend to be better governed, richer, and more trusting.

Of course, the indices we have developed can also be applied to measure the ethnic geography of cities. For example, one could use our segregation index instead of a-spatial measures to compare segregation across Russian metropolitan areas or within metropolitan areas over time. Finally, we would like to stress that our theoretical framework is not specific to the ethnic dimension. Instead of categorizing individuals by ethnic groups and measuring linguistic distances, future research could focus on other social or socio-economic cleavages that are believed to be salient in a particular setting.

References

  • Alesina, Alberto, and Ekaterina Zhuravskaya, “Segregation and the Quality of Government in a Cross Section of Countries,” American Economic Review, 101 (2011), 1872–1911.
  • Ejdemyr, Simon, Eric Kramon, and Amanda Lea Robinson, “Segregation, Ethnic favoritism, and the Strategic Targeting of Local Public Goods,” Comparative Political Studies, 51 (2018), 1111–1143.
  • Hodler, Roland, Michele Valsecchi and Alberto Vesperoni, “Ethnic Geography: Measurement and Evidence,” CEFIR Working Paper No. 253, June 2019. http://cefir.ru/download.php?id=4681
  • Reardon, Sean F., and David O’Sullivan, “Measures of Spatial Segregation,” Sociological Methodology, 34 (2004), 121–162.
  • Tajima, Yuhki, Krislert Samphantharak, and Kai Ostwald, “Ethnic Segregation and Public Goods: Evidence from Indonesia,” American Political Science Review, 112 (2018), 637–653.
  • White, Michael J., “The Measurement of Spatial Segregation,” American Journal of Sociology, 88 (1983), 1008–1018.

The Political Economics of Long Run Development in Eastern Europe: Insights from the 2019 SITE Academic Conference

Posted on January 20, 2020 | Policy Brief

Roller coaster photographed from below symbolising Eastern Europe transition

Thirty years after the fall of communism, many assume that the economic transition of Eastern Europe and the former Soviet States towards a system of market economy is complete. But the region faces new challenges, of both economic and political kind, which renders a thorough understanding the past even more important. This policy brief is based on the scientific contributions presented at the 7th SITE Academic Conference held at the Stockholm School of Economics from December 16th to December 17th, 2019. Organized by the Stockholm Institute of Transition Economics (SITE), the conference brought together academics from all over Europe and the United States to share and discuss their research on economic and political development in Eastern Europe.

The Imperial and Soviet Periods

In the first section of the conference, papers with a focus on the long-term history of Eastern Europe and its implications for more recent events were presented. Marvin Suesse presented his research on how the Russian State Bank financed Tsarist Russia´s belated industrialization, a question that had been discussed by historians, but never thoroughly analyzed quantitatively. By geo-coding historical manufacturing censuses around the turn of the century and using distance between bank branches and factory location, the causal impact of the expansion of the State Bank is estimated, revealing large effects on firm revenues and productivity. These effects are largest in areas where alternative means of financing were least available and where human capital was more abundant.

Natalya Naumenko presented her findings on the economic consequences of the 1933 Soviet famine, which in terms of casualties was extremely devastating. She uses the meteorological conditions a year earlier as an instrumental variable and finds that the famine, which was mostly a rural phenomenon, had a persistent negative effect on the urban population while the rural population recovered relatively quickly.

Gerhard Toews discussed the long-term consequences on regional development of the displacement of an estimated 3 million “enemies of the people”, political prisoners typically belonging to the elite of the society, into the gulags in the early years of the Soviet Union. Using archival data, he has constructed a large database describing the gulag population in terms of the shares of “enemies” relative to other prisoners and taking into account their socio-economic characteristics i.e. the much higher levels of education of the former group. Exploiting variation within gulags, the results suggest that a historically higher density of “enemies” means higher economic prosperity today as measured by nightlight intensity.

Taking another angle, Christian Ochsner investigated the effects of the Red Army´s occupation on post-war Europe, using the demarcation line crossing the Austrian state of Styria as a natural experiment. His conclusion is that even the temporary occupation affected the region’s long-term development, the main channel being age-specific migration.

Finally, Andreas Stegman offered an analysis of the effects of the 1972 East German Extended Visitors Program. The program reduced travel restrictions for West German visitors traveling to certain districts of East Germany. Using a geographic regression discontinuity design comparing similar districts with and without the program, he shows that included districts indeed received much more visits from West Germany and that their citizens were more likely to protest against the Communist government and less likely to vote for the ruling party. This suggests that face-to-face interaction can influence beliefs and attitudes in non-democratic regimes, in turn influencing individual behavior and societal outcomes during transition.

Corruption, Conflict and Public Institutions

Another topic of the conference was the current role of corruption, conflict, electoral fraud and public sector effectiveness for the region. Scott Gehlbach presented his most recent research on the ownership patterns and strategies of Ukrainian oligarchs before and after the Orange revolution. By mapping oligarchs to changing political leadership, he shows how firm owners in Ukraine take actions to protect their property depending on their connections with the current government. He finds that obfuscation of ownership behind holding companies and complicated structures is a potentially valuable strategy in this environment in general but becomes particularly important when an oligarch loses direct connections to the ruling regime.

Likewise, Timothy Frye analyzed election subversion by employers in Russia, Argentina, Venezuela, Turkey and Nigeria. He finds that in Russia, public sector employers and especially state-owned firms are more likely to influence their employees’ decision to vote than private companies. Furthermore, work place mobilization by employers in Russia is clearly negatively associated with the freedom of the press. Election subversion is more likely to be successful when the degree of dependence of the employee is high and the employer’s potential threats are credible. Among Russian firm officials, the most frequently named motivations for them to practice election subversion are the desire to improve their relationship with the authority and the intention to help their party.

Michal Myck studied the impact of the transition experience on economic development around the Polish-German border. Polish communities close to the border were economically backward at the beginning of the transition but could potentially benefit from trade opportunities with an opening towards the West. Using similar methods to those of Stegman above, and nightlight intensity as a measure of economic activity as for instance Toews, Myck finds significant evidence for economic convergence both between Germany and Poland, and between Polish border regions and the rest of Poland.

Vasily Korovkin presented his research on the impact of the conflict in Eastern Ukraine on trade in non-conflict areas in Ukraine, hypothesizing that the conflict may cause a trade diversion away from Russia, particularly so in areas with many ethnic Ukrainians. Using variation in the share of the Russian speaking population at the county level as well as detailed firm level export and import data, he finds that the decrease in trade with Russia is negatively correlated with the share of the Russian speaking population. Potential mechanisms include a decline in trust at the firm level and changes in local attitudes including consumer boycotts.

Finally, Tetyana Tyshchuk analyzed the effects of a Ukrainian public sector reform on civil servants’ capacity and autonomy. The reform created public policy directorates parallel to the regular bureaucracy in 10 ministries. Members of the directorates were hired based on a different procedure and different merits relative to regular public servants and received significantly higher salaries.  Tyshchuk finds that the better paid civil servants indeed score higher on many, though not all, indicators of capacity and autonomy.

Information, Populism and Authoritarianism Today

The final important theme of the conference was the role of information and media, old and new, in today’s politics. In the event´s first keynote speech, Ruben Enikolopov analyzed the political effects of the Internet and social media whose low entry barriers and reliance on user-generated content make them decisively different from traditional media channels. On the one hand, this represents a chance for opposition leaders and whistleblowers to make their voice heard and may improve government accountability. On the other, these media may also become a platform for extremists. Enikolopov presented some of his work analyzing to what extent social media has contributed to fighting corruption in Russia. Using the timings of blog posts by the famous Russian opposition leader Alexei Navalny on corporate governance violations in state-owned companies, he shows that revelations resulted in an immediate drop in the price of the traded shares of the respective companies. He also finds evidence suggesting that Navalny´s blog posts resulted in management changes in these companies. In related papers, he exploits the spread of VKontakte (VK), the Russian version of Facebook, to better understand the influence of social networks on political activism, voting and the occurrence of hate crime. He finds that the spread of VK is indeed causally related to political protests, though not because it nurtures opposition to the government, but rather because it facilitates protest co-ordination. With respect to hate crime, he finds that social media only has an effect in areas where it falls on fertile grounds and where there already are high levels of nationalism. The tentative conclusion is that in Russia – as in Western countries – social media seems to have increased political polarization.

On a similar topic but taking a more theoretical approach, Galina Zudenkova investigated the link between information and communication technologies (ICT), regime contestation and censorship. In a game theoretical framework, where citizens use ICT both to learn about the  competency of the government and to coordinate protests, governments can use different tools to censor information to increase their chances of survival. Zudenkova finds that less competent regimes are more likely to censor coordination, whereas intermediate regimes are more likely to focus on censoring content. These theoretical predictions are then tested using country level data.

The targeted use of information has also played a key role in Putin’s Russia according to Daniel Treisman. In his keynote speech, he argued that while the 20th century dictatorships were mainly based on violence and ideology, the 21st century has been characterized by a sizeable shift towards what he calls “informational autocracy”. Constructing a dataset on the methods used by authoritarian regimes to maintain power between 1946 and 2015, he shows that the use of torture and violence peaked among those dictators who took power in the 1980s and has declined since. Furthermore, he highlights a remarkable shift from topics of violence towards topics of economic competency in dictators’ speeches. However, Treisman finds that by instrumentalizing information, dictators fool the public “but not the elite”. In democratic regimes, those with tertiary education tend to rate their political leaders higher than people without tertiary education. In the new informational authoritarian regime, the opposite seems to be the case. According to Treisman, this is because the “informed elite” has a better understanding of the political reality in places where the media is censored, Putin’s Russia being a good example. Treisman concluded that this new model of authoritarianism has become the prevalent model outside of Europe and today also has its advocates inside the European Union.

The conference ended with a final keynote speech by Sergei Guriev on the political economy of populism. Using existing definitions, he first confirmed that Europe has seen a rise in right-wing populism in the last 20 years. Secular trends, such as globalization and new communication technology, but also the recent global financial crisis, are driving factors behind the rise of populist parties. For instance, analyzing regional variation in voting patterns suggests that the Brexit vote was primarily driven by economic motives rather than by anti-immigrant sentiments. Ironically, though, most evidence suggests that populist governments have a below-average economic performance once in office, the US and Poland being notable exceptions. A key point of Guriev’s presentation was that populism seems to be a good method to obtain power, but, once in power, populists tend to be less successful in promoting citizen welfare. These findings seem to be of high importance given the increasing public support for populist parties around the world and in parts of Eastern Europe

The conference was very well received and on behalf of SITE, the authors would like to express their appreciation to all speakers and participants for sharing their knowledge and to Riksbankens Jubileumsfond for financial support. For those interested to learn more about the papers summarized very briefly above, please visit the conference website and the presenters’ websites as indicated in the text and here below.

Speakers at the Conference

Andreas Stegman, briq – Institute on Behavior and Inequality

Christian Ochsner, CERGE-EI and University of Zurich

Daniel Treisman, University of California, Los Angeles

Galina Zudenkova, TU Dortmund University

Gerhard Toews, New Economic School Moscow

Marvin Suesse, Trinity College

Michal Myck, CenEA

Natalya Naumenko, George Mason University

Ruben Enikolopov, New Economic School Moscow

Scott Gehlbach, University of Chicago

Sergei Guriev, Sciences Po Paris

Tetyana Tyshchuk, Kyiv School of Economics

Timothy Frye, Columbia University

Vasily Korovkin, CERGE-EI

Income Inequality in Transition. New Results for Poland Combining Survey and Tax Return Data

Posted on January 13, 2020 | Policy Brief

20200113 Income Inequality in Transition FREE Policy Brief Image 01

We re-examine the evolution of income inequality in Poland in the process of post-socialist transition focusing on the previously neglected problem of under-coverage of top incomes in household survey data. Multiple statistical techniques (Pareto imputation, survey reweighting, and microsimulation methods) are applied to combined household survey and tax return data in order to obtain top-corrected inequality estimates. We find that the top-corrected Gini coefficient grew in Poland by 14-26% more compared to the unadjusted survey-based estimates. This implies that over the last three decades Poland has become one of the most unequal European countries among those for which top-corrected inequality estimates exist. The highest-income earners benefited the most during the post-socialist transformation: the annual rate of income growth for the top 5% of the population exceeded 3.5%, while the median income grew on average by about 2.5% per year. This brief summarizes the results presented in Brzezinski et al. (2019).

Introduction

There is a large economic literature documenting income inequality changes experienced by former communist countries during their post-1989 transformations. While in Russia and in many post-Soviet economies, inequality exploded during the transition, Poland is often perceived as a country where inequality grew rather moderately. However, these conclusions may be unreliable as they are based on inequality measures estimated using income data only from household surveys.

Many recent studies show that surveys are plagued by significant under-coverage of top incomes, which leads to a severe downward bias of the inequality estimates. Several approaches have been proposed to correct for this problem. One of them is to combine survey data with income information taken from administrative sources such as tax returns. While top-corrected inequality estimates have been produced for many advanced economies, transition countries received little attention in this context so far.

For Poland, Bukowski and Novokmet (2019) provided series of top income shares estimated using tax data. However, their estimates are constructed for gross (pre-tax) income distributed among tax units. This kind of income concept deviates considerably from the primary measure of the standard of living analysed in income distribution literature, namely disposable equivalized household income defined for the entire population. Estimates based only on tax data are not directly comparable to standard survey-based measures, which makes it difficult to decide which of the two kinds of results are closer to the underlying inequality trends and levels.

In a recent paper (Brzezinski, Myck, Najsztub 2019), we provide the first estimates of top-corrected inequality trends for real equivalized disposable incomes in Poland over the years 1994-2015. These estimates can be readily compared with standard survey-based estimates available from Statistics Poland or from Eurostat. Our analysis re-evaluates distributional consequences of post-socialist transition in Poland.

According to the standard view, the Polish transition to a market economy was an almost unqualified success story. Poland managed to achieve fast and stable economic growth (around 4.3% per year since 1994) that was at the same time broadly inclusive and shared rather equally by various social classes and segments of the income distribution. Survey-based estimates suggest that the Gini index for Poland has not increased significantly since 1989 and reached the average level among the EU countries in 2015. In contrast to the standard view, our top-corrected results show that the inequality of living standards in Poland grew sharply over 1989-2015. The adjusted Gini index grew by 4-8 p.p. to a level that ranks Poland among the most unequal European countries for which comparable estimates exist.

Data and Methods

We use data from two sources. Our survey income data comes from the representative Polish Household Survey (PHBS) conducted annually by Statistics Poland since 1957. We use the PHBS data for 1994-2015 as the pre-1994 surveys do not contain data on individual incomes (required for our microsimulation modelling) and 2015 is the last year for which estimates of tax-based inequality measures are available. We adjust the baseline PHBS survey weights to match the census-based number of males and females by age groups (population weights). We also create a further adjusted set of weights to match the number of PIT payers in each tax bracket according to the Polish tax scale (tax weights).

Our main income variable is real equivalent household disposable (post tax and transfer) income. We obtain it from the Polish microsimulation model SIMPL applied to the PHBS data. The microsimulation model allows us to construct a gross (before PIT and employee SSCs) income distribution among the tax units, which is unavailable in the raw PHBS data. This is crucial as it is the gross income distribution between tax units to which we impute top incomes estimated using tax-based statistics.

Our second data source is the series of tax-based top income shares for Poland taken from Bukowski and Novokmet (2019). To construct top-corrected inequality estimates, we follow the methodological approach of Bartels and Metzing (2019). Using the microsimulation model applied to the PHBS data we obtain the distribution of gross income among tax units (individuals). In the next step, we use data on top income shares to estimate the parameters of a Pareto distribution for gross income distribution in terms of tax units. Then, we replace the top 1% (or 5%) of tax units’ incomes with the incomes implied by the estimated Pareto distribution. The resulting imputed gross distribution is subsequently reweighted using either population or tax weights. After imputing top incomes, we again use the microsimulation approach to compute top-corrected real equivalized household net incomes.

Corrected Income Inequality Trends

Figure 1: The Gini index for Poland, 1994-2015: unadjusted vs top-corrected estimates

Note: Vertical lines show 95% confidence

Figure 1 presents our income inequality estimates in terms of the Gini coefficient. For the period 1994-2005, we present two top-corrected series, which can be considered as lower and upper bound estimates of the “true” Gini. The results for this period are more uncertain as they are affected by the 2004 tax reform in Poland that introduced an optional flat tax for non-agricultural business income, which reduced the marginal tax rate for the highest income taxpayers from 40% to 19%. Research suggests that before the reform the problems of tax evasion and avoidance could have been more pronounced in Poland and some of the top incomes were unreported or under-reported. The upper bound series on Figure 1 corrects for the possible higher tax evasion and avoidance before 2005.

The unadjusted Gini series suggests that income inequality in Poland was rather stable over 1994-2015.  On the other hand, our top-corrected series point to a very different story.  Until 2005, our two correction procedures show similar inequality trends, but somewhat different levels. After 2005, our corrected series shows systematic and high divergence between unadjusted and top-corrected Ginis ranging from 4 to 8 p.p. The top-corrected Ginis increase in the range from 14 to 26% over 1994-2015.  While according to the unadjusted data Poland is only moderately unequal, the comparison of top-corrected estimates shows that in 2015 Poland has higher level of income inequality than even high-inequality EU countries such as Germany, Spain or UK.

We also show that each percentile of the disposable income distribution in Poland saw income increases in absolute terms between 1994 and 2015. This implies that on average the incomes of all social groups increased during the transition to market economy. However, these gains were shared unequally. According to our adjusted estimates, the cumulative growth in real income over 1994-2015 for the top 1% of Poles reached 122-167%, while for the bottom 10% the corresponding number is at most 57%.

Redistribution and Progressivity of the Tax System

We also analyse how our correction procedures affect measures of redistribution and progressivity of direct taxation (income taxes, employees’ mandatory social security contributions, and health insurance). The top-corrected estimates show that the percentage reduction in the Gini index due to social insurance contributions and PIT has fallen from 19.2% in 1999 to 11.6% in 2015.

While the unadjusted series suggests that the progressivity of the Polish system of PIT and social insurance contributions has decreased only mildly over time, the top-corrected series points to a much steeper fall, especially during 2005-2009. Without the top-correction, the progressivity in 2015 is overestimated by 2.3 p.p. (or by 40%). Much of the decline in tax progressivity over 2005-2009 is due to the reduction from three PIT brackets and marginal tax rates to just two brackets and rates (18% and 32%) in 2009. Even in terms of the unadjusted data, Poland ranks in the recent years as the country with the lowest PIT and SICs progressivity in the EU.

Conclusion

Our recent paper on estimating the top-corrected measures of income inequality shows that while Poland was already a relatively unequal country in the early 1990s, it has become one of the most unequal European countries (not including Russia) among those for which comparable estimates exist. The results have important implications for the assessment of the distributional consequences of post-socialist transformations or modernization processes in emerging countries. They indicate that using income tax data and imputation or reweighting techniques to account for the problem of missing top incomes in survey data can significantly alter the conclusions about income inequality levels and trends. More reliable inequality estimates would contribute not only to a better understanding of economic transformation and modernization processes but could also shed some light on recent political turmoil in many transition and emerging countries (such as Turkey, Hungary or Poland). As suggested by some recent research, the growing distributional tensions in emerging countries of Eastern Europe and Central Asia may be associated with more distrust in governments and an increased propensity to vote for radical political parties.

Acknowledgements

The authors gratefully acknowledge the support of the Polish National Science Centre (NCN) through project number: UMO-2017/25/B/HS4/01360. For the full list of acknowledgements see Brzezinski et al. (2019).

References

  • Bartels, C., Metzing, M. (2019). An integrated approach for a top-corrected income distribution. The Journal of Economic Inequality, 17(2), 125-143.
  • Brzezinski M., Myck M., Najsztub M. (2019), Reevaluating Distributional Consequences of the Transition to Market Economy in Poland: New Results from Combined Household Survey and Tax Return Data. IZA DP No. 12734.
  • Bukowski P., Novokmet F. (2019), Between Communism and Capitalism: Long-Term Inequality in Poland, 1892-2015. CEP Discussion Paper No 1628 June 2019.

From Partial to Full Universality: The Family 500+ Programme in Poland and Its Labour Supply Implications

Posted on December 16, 2019 | Policy Brief

20191216 From Partial to Full Universality FREE Network Policy Brief Image 01

The implementation of the ‘Family 500+’ programme in April 2016 represented a significant shift in public support for families with children in Poland. The programme guaranteed 500 PLN/month (approx. 120 euros) for each second and subsequent child in the family and the same amount for the first child in families with incomes below a specified threshold. As of July 2019, the benefit has been made fully universal for all children aged 0-17, an extension which nearly doubled its total cost and benefited primarily middle and higher income households. We examine the labour market implications of both the initial design and its recent fully universal version. Using the discrete choice labour supply model, we show that the initial Family 500+ benefits generated strong labour supply disincentives and were expected to result in the withdrawal of between 160-200 thousand women from the labour market. The recent removal of the means test is likely to nullify this negative effect, leading to an approximately neutral impact on labour supply. We argue that when spending over 4% of GDP on families with children, it should be possible to design a more comprehensive system of support, which would be more effective in reaching the joint objectives of low child poverty and high female employment combined with higher fertility rates.

Introduction

Following the 2015 parliamentary elections in Poland the ruling Law and Justice Party was quick to fulfill their campaign promise of implementing a generous quasi-universal family support programme. In April 2016, all families began receiving PLN 500 (approx. 120 euros) per month for each second and subsequent child, while households that passed an income means test were granted the same amount for their first or only child.  At a cost of nearly PLN 22 billion (5.2 billion euros, approx. 1.1% of GDP) per year, the Family 500+ benefit became the flagship reform of the Law and Justice government’s first term.

With new elections approaching in October this year, the government announced a significant expansion of the programme in May, which made it fully universal. The extended programme is nearly twice as expensive with an additional cost of PLN 18.3 billion (4.3 billion euros) per year, valuing the whole package at over 2% of GDP. This takes the total value of financial support for families with children, including family benefits and child-related tax breaks, to 4% of GDP and it means that as far as family support is concerned, the ruling party has brought Poland from one of the lowest-spending countries in the EU to one of the highest over the course of 4 years.

The initial design of the benefit had a significant impact on childhood poverty in Poland, with an absolute and relative decrease from 9.0 to 4.7 percent and 20.6 to 15.3 percent respectively between 2015 and 2017 (GUS, 2017). While a more targeted design could have made a far greater impact, these changes still reflect a significant improvement in the material situation of families with children. The policy may have also had a modest upward effect on fertility rates in the first years following its implementation, although this is difficult to assess given the parallel roll out of several other fertility-oriented policies and other changes which could have played a role in family decisions. Simultaneously, as argued in the ex-ante analysis by Myck (2016) and ex-post analysis by Magda et al. (2018), these positive outcomes came at the cost of reduced female labour market participation. This reduction primarily affected women with both lower levels of education and living outside of large urban areas (Myck and Trzciński, 2019).

The Family 500+ Reform: Design and Distributional Implications

The initial Family 500+ programme directed funds to 2.7 million families in addition to any already existing financial support and has been excluded from other means-tested support instruments.  Since families that had a net income of less than PLN 800 per month per person could receive the benefit for the first or only child, the policy had a distinct redistributive element and meant that  the bottom half of the income distribution received nearly 60% of the funds. However, the design was characterised by clear labour market disincentive effects, which were particularly strong for second earners and single parents.

In a one-child household (53.3 percent of families with children, GUS, 2016) with the first earner bringing in an income equivalent to 125% of the national minimum wage, the second earner needed only to earn PLN 940 per month in order for the family to cross the means test threshold and stop receiving the Family 500+ benefits. The benefit design is presented in Figure 1 in the form of budget constraints for the first earner (Case A) and the second earner (Case B) in a couple with one child. In the latter case the first earner is assumed to receive earnings equivalent to 125% of the minimum wage. The disincentive effects of the means test are clear in both cases and we can see that for the second earner, the benefit withdrawal comes at a very low income level – far below the national minimum wage of PLN 2100 per month. The “point withdrawal” of the benefit implied that it was enough for the family to marginally exceed the means test threshold for it to completely lose eligibility for the Family 500+ support for the first child.

The expansion of the Family 500+ programme, which came into effect in July 2019, eliminated the means-tested threshold thus making the policy fully universal. It came, however, at the cost of the redistributive character of the programme. Over 32% of the additional expenditure resulting from the universal character of the policy has been passed on to the top quintile of the income distribution and in its new version, the bottom half of households only receive 45 percent of all spending. The expansion of the programme is thus unlikely to further reduce child poverty significantly and – since its beneficiaries are mainly families with middle and high incomes – it is not expected to bring noticeable changes in fertility levels.

Figure 1: Family budget constraints for the first and second earner

Source: Authors’ calculations using the SIMPL microsimulation model.

Partial and Full Universality of the Family 500+ Programme and the Implications on Female Labour Supply

With the use of modelling tools to simulate the labour market response to changes in financial incentives to work, we have updated the initial simulations of Myck (2016) using the latest pre-reform data and examined the simulated labour supply decisions to the expanded fully universal programme, as if it were implemented instead of the initial version of the benefit. The analysis was conducted with data from the 2015 Polish Household Budget Survey, a detailed incomes and expenditure survey conducted annually by the Polish Central Statistical Office.

Table 1: Effects of the initial and the expanded Family 500+ programme on female labour supply

Results of the simulations are presented in Table 1. Simulations were conducted separately for single women, and under two scenarios for women in couples assuming that both partners adjust their behaviour (Model A) and that the labour market position of the male partner is unchanged (Model B).  The simulated labour supply response to the initial reform confirms the magnitude of earlier results and suggests an equilibrium effect of 160-200 thousand women leaving the labour force. This is also consistent with results presented by Magda et al. (2018), who found that female labour market participation decreased by approx. 100 thousand women after the policy had been in place for one year.

However, as we can see in the right-hand part of Table 1, the response to a fully universal design – modelled as if it was introduced in 2016 instead of the means-tested version – is essentially neutral. For single mothers the reduction is only about 3000, while for women in couples, the model suggests a small positive reaction under the Model A specification and a small negative one under Model B. In total, the universal design of Family 500+ benefits can be described as labour supply neutral. Since the reaction has been modelled on pre-reform data, and because some women have already withdrawn from the labour market after the introduction of the initial benefit design in 2016, the remaining uncertainty is whether the new set of incentives will motivate these mothers sufficiently to return to work.

Conclusion

The introduction and subsequent expansion, of the Family 500+ programme has substantially increased financial resources of families with children in Poland. The policy rollout of the initial, partially universal programme has seen substantial changes in the level of child poverty in Poland and may have contributed to a modest increase in fertility in the initial years following the introduction of the reform. The means-tested design of the benefit, however, incentivised a significant number of women to leave the labour market. One year after the introduction of the policy approximately 100,000 women were estimated to have left the labour market (Magda et al. 2018), while the equilibrium effect of the policy suggested long-run implications of over 200,000 (Myck, 2016). The updated simulation results using the latest available data suggest slightly lower, though still substantial equilibrium implications of the initial partially universal design of the Family 500+ programme in the range of between 160,000-200,000. However, as we show in our latest analysis, these labour market consequences could be reversed after the expansion of the programme to a fully universal set-up. The simulated effects of the universal design of the programme, which has been in place in Poland since July 2019, modelled as if it was implemented instead of the initial means-tested version, are broadly neutral for female labour supply. The only question is how likely the mothers who left employment in response to the initial policy will return to work given the new set of financial incentives. Considering these positive implications of the fully universal programme, one has to bear in mind that the extended programme, which will cost over PLN 40 bn per year (approx. 2% of GDP), is unlikely to contribute to the other key objectives set by the government, namely reducing child poverty and increasing fertility. Including the Family 500+ programme, the Polish government currently spends about 4% of GDP on direct financial support for families with children. Given the design of the policies which make up this family package, it seems that the joint objectives of higher fertility, reduced poverty and higher female employment could be achieved more effectively under a reformed structure of support that would be better targeted at poorer households, include specific employment incentives, and incorporate support for childcare, early education and long-term care.

Acknowledgements

This brief summarizes the results presented in Myck and Trzciński (2019). The authors gratefully acknowledge the support of the Swedish International Development Cooperation Agency, Sida, through the FROGEE project. For the full list of acknowledgements see Myck and Trzciński (2019).

References

  • Goraus, K. and G. Inchauste (2016), “The Distributional Impact of Taxes and Transfers in Poland”, Policy Research Working Paper 7787, World Bank.
  • GUS (2016), “Działania Prorodzinne w Latach 2010-2015”, Główny Urząd Statystyczny – Polish Central Statistical Office, Warsaw.
  • GUS (2017), “Zasięg ubóstwa ekonomicznego w Polsce w 2017r.”, Główny Urząd Statystyczny – Polish Central Statistical Office, Warsaw.
  • Magda, I., A. Kiełczewska, and N. Brandt (2018), “The Effects of Large Universal Child Benefits on Female Labour Supply”, IZA Discussion Paper No. 11652, IZA-Bonn.
  • Myck, M. (2016), “Estimating Labour Supply Response to the Introduction of the Family 500+ Programme”, Working Paper 1/2016, CenEA. Jacobson, L., LaLonde, R. and Sullivan, D. (1993). “Earnings losses of displaced workers”, American Economic Review, 83, pp. 685–709.
  • Myck, M. and Trzciński, K. (2019) “From Partial to Full Universality: The Family 500+ Programme in Poland and its Labor Supply Implications”, Ifo DICE report 3 / 2019.