Introduction

Several former transition countries (FTEs) that joined the EU after 2004 have reversed economic reforms in recent years, sometimes central reforms. The reversals in Poland and Hungary, previously star performers of transition, have caught media attention, since they have touched the very foundations of a social market economy, such as the independence of the courts. These reversals triggered strong political reactions, also at the EU level. Reform reversals in other countries have been less in the media, but have in some cases been equally important.

Reversals are not unique to FTEs, not even within the EU, but their experiences in this regard may offer uniquely important lessons for several reasons. First, because these countries implemented reforms that were unprecedented in terms of their depth and scope within just one generation. This journey entailed a uniquely fast social learning process, one that stress-tested societies, but nevertheless might not have been fast enough in many areas and countries to make the reforms lasting. Second, external anchors--first the IMF and the World Bank, later European institutions--have played a uniquely strong role all the way through this process. Third, a relatively large group of countries travelled down this road together, countries that were geographically and culturally rather close to each other. Their initial social norms, albeit historically not necessarily homogenous, had been shaped by common factors prior to the start of this journey. Moreover, countries who turned their back on earlier reforms have seen others in their peer group continue on the path and pull ahead economically (Fig. 1). So not only the direct costs and benefits to the country (society) concerned, but also the opportunity costs of reforms and their reversals can be observed.

Many of the reform reversals took place during the recent crisis. In a crisis, decision horizons get shorter, not only for politicians, making reforms more vulnerable to reversal attempts.

While there is a huge literature on the political economy of reforms, and in particular on reforms in the transition economies (Rodrik 1995, 1996, 2006; Roland 2001, 2002), there is surprisingly little attention paid to reform reversals (Abiad and Mody 2005; Campos and Coricelli 2009). It seems that there are no general models of reform reversals or models that could capture many of the important aspects of the reversal episodes discussed below. Why did societies that had already undergone massive systemic reform, who had paid the associated political price (with politicians repeatedly facing the Juncker curse in the process, see Buti et al. 2008) who as a result had successfully entered the EU and fully integrated into the global economy, all of the sudden turn back and in a way, waste the previous efforts? The available models do not yet capture how internal and external social norms and social learning processes interact, how external anchors work, and what role European law and institutions have played in this process.

Episodes of reform reversals in these countries may offer many important lessons to help understand how the reform process evolves in a society, how domestic institutions work and interact with each other, how internal social norms evolve through social learning and interact with external social norms and how all these promote and anchor reforms or lead to their reversal. Many of the reform reversal episodes discussed here can help to understand how domestic and EU institutions, each of them backed by different social norms and evolving through social learning in different social spaces, interact in this process. This is perhaps the most unique contribution of this paper.

However, reversals deserve special attention not only because they can offer useful lessons for theorists, but also because they can have a major negative impact on the economies and societies concerned, which is perhaps larger than the original, positive impact of the reforms they reversed. Reversals can be particularly harmful to future development, since they can make it particularly difficult, if not impossible, for politicians to later (re)embark on reforms in the same areas. Moreover, concerted reversals in several policy areas can put a country on a lower development path for an extended period of time (path dependency).

Conceptual Framework for Reforms and Their Reversal

In what follows, we develop the 'evolutionary institutionalist perspective' that Roland (2001) offered in his seminal paper on the first 10 years of transition. We start from the viewpoint that the reform reversal episodes we discuss simply cannot be understood without carefully considering how developments in national institutions helped or hindered reform and also how this process was shaped by the interaction of national with European institutions. The very essence of EU membership is the common EU law (the acquis) and the institutions that are there to guard different parts of this law. Moreover, the EU also has rules and institutions to coordinate policies, particularly fiscal policy and structural reforms.

Roland (2001, p. 30) argues that "If anything, the experience of transition shows that policies of liberalization, stabilization and privatization that are not grounded in adequate institutions may not deliver successful outcomes." He also points to the importance of self-enforcing social norms, which drive society's acceptances and aversions and can help ensure that institutions gradually evolve toward more perfect institutions, in a sort of experimental way. Hence some flexibility in the national design of these institutions is important. Iancu and Ungureanu (2013) emphasize the role of 'social learning' and show how the lack of social learning leads to reform reversals in civil service reforms--an area where there is an absence of legal and institutional anchoring by the EU. The interaction between institutions on the one hand and social learning and norms that shape their legal forms (through laws) and behavior, on the other, plays an important role in the reform reversal episodes we shall discuss later in the paper.

Social norms are thought to evolve slowly, sometimes over centuries, through social learning (Young 2015; Rotter 1954). A key element of this process is reinforcement--some positive or negative reaction to an action by an individual or a group of people. The likelihood that the reinforcement happens and the value of the reinforcement action, be it positive (carrot) or negative (stick), to the individual or the group concerned determines the likelihood that a certain behavior will be performed. The same applies to politicians and political parties (movements), for whom elections (and polls between elections) deliver a very clear reinforcement. Through this process, a social norm emerges and effectively guides societal behavior. Norms therefore emerge bottom up, through learning in a society.

The countries considered here moved away from centrally planned, fundamentally autarchic economies with rather isolated societies. The social norms of this society had interacted with external social norms of a particular type--from the USSR and other similar countries. (1) They thus underwent a very different type of social learning to the other societies of a modern social market economy, developing social norms that in many aspects were fundamentally different from the ones that had previously acted as the dominant external anchors.

While norms emerge bottom up, laws and public institutions are set top-down, being created by elected national (and European) law makers (parliaments), typically but not always reflecting the evolution of a society's social norms. For the countries in the region, the process leading to EU accession entailed major improvements in institutions, perhaps most importantly in those areas falling under EU law, but also in other areas, in fact following a pattern close to the Washington Consensus idea of introducing institutions that had evolved as best-practice in highly developed countries (Roland 2001). This top-down process installed new laws and institutions reflecting social norms in developed countries which were in many cases distinctly different from the ones prevailing in FTEs. Moreover, as a reaction to the crisis, the EU embarked on a rapid change and improvement, of its own legal and institutional set up in central areas, such as policy coordination, banking and finance and fiscal systems. As several recent events suggest, social learning has not yet caught up with this process, even in many of the core euro area countries which are culturally, historically and geographically much closer to the center where these decisions were taken. From this perspective, it should perhaps not be surprising that FTEs have faced some major challenges in following the rapid evolution of the European institutional framework with their internal social learning processes.

The picture gets even more complex if we introduce the concept of parallel social norms and parallel social learning processes in these societies, which in our view are important to understanding what has been happening in Poland and Hungary (and perhaps also in Romania) lately. The notions of "Polska B" (Poland B, https ://en.wikipedia.org/wiki/Poland_A_and_B) or "videki Magyarorszag" (Hungarian countryside) are essential in this regard, as in all likelihood they reflect distinctly different social norms, social learning processes and reinforcements to those prevailing in other parts of these countries. As the social learning theory (SLT) emphasizes (Rotter 1954), the psychological situation is much more important than the objective situation.

If the levels and types of education are characteristically different in parallel communities, their assessment of the objective situation may be very different. If (as during the crisis) the uncertainty regarding the objective situation is also heightened, the psychological element becomes even more dominant in both communities. If the relative importance of psychological needs such as independence and protection dependency is significantly different in parallel communities, so can be the values of certain types of reinforcement actions, which may vary across parts of society. Most importantly, the reinforcement value (positive or negative) of the changes that a stronger and more centralized state can bring about may be very different in parallel communities. Therefore, it is probably not just by chance that a common central element of the Polish and Hungarian new approaches is a much stronger and much more centralized state.

As research on artificial intelligence shows, if the intensity of interaction between two communities declines below a critical level, parallel norms can emerge and become stable (Sen and Airiau 2007). In all likelihood, if such parallel norms exist in these countries, they emerged a long time ago. Why then did they not create the same outcome much earlier? Perhaps a few important recent developments help explain their recent emergence. FDI in the region tends to be heavily concentrated in certain areas and has little interaction with local firms (Bisztray 2016). Tourism is also heavily concentrated in a few areas, and so are foreign students (Erasmus). Generally, market forces have tended to produce strong agglomeration effects and particular migration routes (flows) within and across the FTEs. EU accession accelerated both elements, bringing global forces much closer to these countries than they were before. EU accession also made significant resources available to help FTEs promote regional development, and convergence more broadly, significantly boosting infrastructure investment and rural incomes. However, it seems this major financial support helped less with exposing isolated communities, where market forces were less at work, to external social norms and social learning processes.

Part of the puzzle of reform reversals may also be explained by social learning reversing its direction and, as a result, social norms changing (deteriorating). We tend to believe that reforms are, by nature, a good thing, at least for an economy or society as a whole. We assume that policy makers know which reforms are needed and how to introduce them; that reforms by nature increase growth potential, because there is a large body of literature that provides theoretical arguments and empirical evidence in this regard, and thus, long-term benefits outweigh any possible short-term cost (e.g., Varga and in't Veld 2014); and that reforms are sufficiently well designed and well implemented. Good design involves careful sequencing and flanking measures to mitigate short-term costs and/or the impact of reform on particular social groups (losers). Good implementation involves a continuous monitoring of the actual impact of implemented reforms so that unwanted or unforeseen negative impacts and genuine design problems can be identified and mitigated in a timely fashion. We tend to assume that governments have the capacity to pay for such mitigating measures and can maintain a stable and supportive macroeconomic environment in which households and firms can relatively easily judge the true impact of reforms and the true intentions of policy makers. Markets are also assumed to accurately judge the long-term (positive) impact of reforms.

In short, we tend to assume that policies are optimal and fully credible. So, once reforms are introduced, self-enforcing social norms and social learning should catch up with the new reality fast enough to create a domestic anchor for reforms that have also been introduced with external support or pressure, coming either from international organizations or financial markets.

But the reality in FTEs has been very different on almost all fronts. Apparently, in many cases, social learning has been slow and thus most institutions remained fragile, in many cases vulnerable to special interest group's attempts to (partly) capture them. For the same reason, laws and institutions created to promote good policies could fulfill this role only to varying degrees. Moreover, as we shall see below, the design of many reforms or institutions were less than perfect (e.g., pension reforms) or reforms may have caused unforeseen negative side effects that were not carefully monitored and addressed at the implementation stage.

The Role of External Anchors

Until the EU accession negotiations started, the IMF and the World Bank served as the main external anchors for economic policy and structural reforms (Roaf et al. 2014). Following EU accession, the IMF got involved in the region again in 2008 when the financial crisis started, first in Hungary, followed by Latvia and Romania (Szekely and Ward-Warmedinger 2018). The IMF has consistently identified the episodes of slow-down or reversal of reforms. However, its anchoring role has been weak at best.

While the first wave of reforms in early transition was driven by the desire to get away from a centrally planned economy toward a western-type market economy, the second wave was driven by the desire for EU membership. Membership offered an institutionalized anchor to the West. It secured access to the largest market in the world, the single market, the free movement of capital and labor and a massive, historically unprecedented, financial support to promote convergence (Keereman and Szekely 2010). It thus further increased and made tangible the perceived cost of a policy reversal, which carried with it the risk of being left out of the EU enlargement process. The public viewed such a possible outcome as very negative.

Once these countries entered the EU, most of the advantages became granted, while many of the reforms were not protected by EU legislation. Self-enforcing social norms and social learning were apparently not emerging fast enough to protect some of the previously undertaken reforms in those areas. In fact, reforms started to be reversed soon after EU accession and not only in combination with the negative impact of the crisis (Szekely and Ward-Warmedinger 2018).

EU membership, however, also comes with a number of obligations and mechanisms that monitor and enforce the fulfillment of these obligations. Traditionally, the macroeconomic surveillance of a member state concerned fiscal policy under the Stability and Growth Pact (European Commission 2017b), supported by a systematic monitoring of the long-run sustainability of public finances (European Commission 2015a, 2017a). Lessons from the recent crisis led to this framework being enhanced in 2012 by the addition of the surveillance of macroeconomic imbalances via the Macroeconomic Imbalances Procedure (MIP), including private sector imbalances in the financial sector. (2) Later on, the whole surveillance and policy coordination work was integrated into what is called the European Semester, with an increased emphasis on structural reforms. In this framework, every year, the European Commission issues Country Reports for all member states, based on which the European Council issues Country Specific Recommendations (CSRs) to the member states for reform measures. The country reports also assess the implementation of the previously issued CSRs. (3)

The European Commission is also the guardian of the EU treaty, that is, it has the obligation and the legal power to enforce it. To this end, it monitors the compliance of the member states with state-aid rules and takes action in case of unlawful stateaid. Other parts of the European Commission are guardians of regulations regarding fiscal systems, central bank independence, tax regulations, free movement of capital, and the banking union. Many of these areas are involved in the reform reversal episodes discussed in Szekely and Ward-Warmedinger (2018). In case a country does not comply with a given part of the treaty, and does not address problems when they are detected by the European Commission, it can be subject to an infringement procedure, which eventually may lead to the country being taken to the European Court of Justice to enforce the Treaty. Like other national legal enforcement measures, such procedure, however, typically requires a significant amount of time to reach this final stage.

The inherent problems in less developed countries with the adoption of institutions that evolved in other social environments are well recognized in the literature on institutional monocropping (see, e.g., Evans 2004). The developments in the first phase of the transition process, in which the Washington institutions were the key external anchoring institutions in the region, are relatively well captured by the conceptual framework of this literature. The discussion of the introduction and reversal of pension reforms in the region given below provides a good example in this regard, showing some of the typical characteristics of institutional monocropping, such as technocratic rule and institution transfer, albeit in many cases strongly supported by political processes. In such episodes, the subsequent social learning about the way the new rules and institutions worked and their broader implications may have indeed been the very reason for reversal. However, the actual examples seem to suggest otherwise as apparently many of the problems originating from institutional monocropping could have been fixed without full reversals.

The second phase, in which the desire and later the formal preparations to join the EU was the key driver, however, raises major new issues in this regard. EU membership, by definition, involves the full adoption of the joint legal framework of the EU and the rules and institutions that are embedded in this framework. While many of the latter (such as central bank independence) were introduced prior to accession, and thus, there had been some social learning related to them, most of the social learning related to these new rules and institutions took place after EU accession. As we argued before, the political and public support to EU accession was extremely strong in the region, including referenda that showed in most cases overwhelming majority for EU membership.

Where the problems started to emerge, as we shall see in the discussions on the reform reversal episodes, was the lack of sufficient social learning following EU accession, and perhaps also the lack of a political process to promote and ease the adoption to the new rules and institutions, given that the formal reversal of individual elements protected by EU law is not a viable option. Apparently, in some cases, rules and institutions emanating from the common EU law have not yet been transformative enough and thus have not yet shaped the identities and mentalities of a sufficiently large part of society in these countries. As we shall see below, this may be a particularly important aspect regarding reform reversals in Poland and Hungary, where parallel social norms and social learning processes may exist and where the political balance of power might have shifted across these norms in a major way.

Episodes of Reform Reversals and Their Characteristics

Here we focus on a subset of the reform reversal episodes in the FTEs, by reform area (see Szekely and Ward-Warmedinger 2018, for a detailed account). We define a reversal either as a formal reversal of a previously taken reform measure or as a deterioration in the quality of an institution (behavioral reversal) which was central to one of the dimensions of the reform space. This definition reflects the fundamental point made in Roland (2001) that the transition to a modern market economy requires the necessary quality institutions that can make the economy work in the way we expect it to work. The first reversal type is relatively easy to identify, as one can refer to a formal/legal action. Changes in behavior are much more difficult to identify, usually being observed through a negative outcome captured in country surveillance (by the IMF or the European Commission) or through an episode of market turbulence or crisis (like, for example, in the episodes of critical weakening of banking supervisions in certain periods in Bulgaria and Slovenia). For reversals which were identified when problems reached a critical level, it is virtually impossible to pinpoint the time when the institutional behavior changed. Furthermore, reversals in the behavior of institutions, public or private, may also remain latent for a relatively long time. Consequently, there may have been other similar episodes of behavioral reversals that have not (yet) been identified.

Fiscal Policy and Pension System

Romania was among the first countries where the crisis led to a rapid deterioration of its fiscal situation and in its external financing conditions, forcing the country to ask for a joint EU-IMF financial assistance program in 2009 (see Section 4.1 in Szekely and Ward-Warmedinger 2018). This and the subsequent precautionary program helped the country to fully stabilize its fiscal position and more broadly rebalance its economy, while the economy grew above potential and growth and job creation accelerated. From 2016 onwards, however, fiscal policy embarked on a course of highly pro-cyclical loosening, moving from a close-to-balance structural position in 2015 to a large structural deficit by 2017. The only apparent limit to loosening was the 3% of GDP threshold specified within the Excessive Deficit Procedure (EDP) of the Stability and Growth Pact (SGP).

To show the full extent of the behavioral reversal behind this episode, in order to keep the galloping deficit stable at the 3% threshold, the technocratic Ciolos government started to opportunistically downsize the second pillar of the pension system in 2016, turning a behavioral reversal into a formal (legal) reform reversal in another area.

The lack of social norms (and social learning) to maintain sound fiscal policies is also revealed by the fact that the otherwise-international-best-practice domestic fiscal framework of Romania did very little to stop the fiscal behavioral reversal. The subsequent governments of very different types (political orientation, strength, etc.) simply disregarded their own domestic fiscal responsibility law and the stark warnings of their Fiscal Council (all created under an EU-IMF program). The public did not react negatively to any of this either. In fact, in the 2016 elections, it elected the central-left coalition with a program to continue the fiscal behavioral reversal. The lure of short-term gain in terms of increased fiscal space, short-term interest, proved much stronger than norms, the respect for newly created institutions and rules and the embodiment of longer-term gains from them (March and Olsen 2006). Certainly on the part of politicians, but perhaps also on the part of the wider public, as fiscal policies handed out direct gains to people (e.g., public sector wage increases). The EU anchor also did not work well in preventing the reversal. The European Commission triggered the Significant Deviation Procedure in 2017, but the Romanian authorities have not implemented the recommended fiscal adjustment. Instead they further loosened their fiscal stance.

In fact, the spillover into formal reform reversals in other areas intensified, as the government started preparations to create a 'sovereign fund' that would include state-owned enterprises (SOEs), which in turn would be exempted from the law that governed the selection of SOE managers, an important former reform (also introduced under the EU-IMF program). Moreover, they further downsized the second pillar of the pension system to keep the deficit at the 3% EDP threshold in 2018. Signs of further possible spillovers into other policy areas also emerged.

The pension reform reversals in FTEs (Section 4.2 of Szekely and Ward-Warmedinger 2018; Bielawska et al. 2017; Naczyk and Domonkos 2016) seem to have been rather opportunistic in nature. Short-term fiscal gains, deficit and debt reductions seem to have been the overwhelming motivation for these behavioral reversals, sometimes even during EU-IMF financial assistance programs. In Hungary, where reform reversals were motivated by a broader strategy, accumulated funds in the pension system were used to finance the renationalization of the energy sector. Political orientation seems to have played a minor role. Important design faults in the original reform made these reforms more prone to reversal, although they had been identified relatively early on (e.g., by the IMF) and could have been addressed inside the original system. In addition, the original systemic pension reforms were introduced with a very strong involvement of the World Bank (and sometimes during IMF programs), suggesting the existence of a rather weak domestic ownership of the reforms. Correspondingly, there was surprisingly little resistance from the public to such major reversals, despite their affecting the accumulated funds of contributors. The lack of sufficiently strong social norms (and social learning) to protect previous pension reforms, the apparently easy victory of short-term interests over weak norms (March and Olsen 2006), was also evidenced by the fact that these reversals came long after the initial reforms, in Hungary 12 years later. The difficulties with fully understanding the long-term implications of such a fundamental reversal of pension reforms also added to the problem. The governments involved certainly did very little to help people understand these consequences.

It is also important to mention that the original pension reforms were neither required nor protected by EU legislation, and thus, the European Commission had no mandate to formally intervene in the legislative process. Its broader fiscal surveillance work, while fully revealing the reversal, its nature and long-term consequences, was also apparently not a strong enough factor to stop any of the pension reform reversals.

Financial System

There were two major reform reversal episodes in the financial sector in Slovenia and Bulgaria. Regarding Slovenia, the analysis of the Bank of Slovenia (2015) and various Country Reports by the European Commission rather convincingly showed how the different factors played a role and how the reversals in the behavior (quality) of different institutions interacted to strengthen each other's negative impact, which eventually brought about a full-scale banking crisis. The strong capital inflow prior to the crisis, which accelerated with euro area membership, offered ample funding. But it was the weak corporate governance of state-owned banks and that of stateowned enterprises (SOEs) that turned this into a massive capital misallocation. Institutional weakness in banking supervision, in interacting with an apparent reluctance on the part of the state to address the rapidly accumulating loan quality problem in the state-owned banks, further distorted the behavior of these banks and amplified the problem. In fact, the state in general tried to postpone the rapidly accumulating problems in the SOEs, also by rapidly increasing state-aid.

By implementing the CSRs from the European Council (under the European Semester), the reversals were in large part reversed. The state-aid decision on the rescued state-owned banks also played a central role in reversing the behavioral reversals. However, such progress was not matched by social learning and by a sufficient weakening of vested interests. As a result, the vested interest groups fought back, via heavy criticism of the actions taken by the BoS in the AQR/ST and the subsequent recapitalization of the banks concerned. In some cases, formal reversals in other areas were also triggered, most importantly in CB independence. These reversal attempts were in turn stemmed by the ECB and the European Commission. Nevertheless, there are apparent signs of a significant behavioral reversal in this area, which in the future may eventually weaken the progress made in banking supervision. This latter development also suggests the limitations of the EU's role as an external anchor through formal (legal) instruments, in the absence of sufficiently strong social norms and sufficiently fast social learning. It is perhaps important to point out that, as mentioned earlier, central bank independence predated EU accession, so social learning had more time to take place. On the other hand, Slovenia was among the few countries where direct or indirect (through SOEs) state dominance in the banking sector prevailed even after EU and euro area accession.

Bulgaria also experienced major problems in its financial sector. The interaction of behavioral reversals in banking supervision and corporate governance in domestically owned banks and nonbank companies in the private sector led to a rapid deterioration in financial stability. This in turn triggered a strong external intervention from the EU (the finding of excessive imbalances under the MIP and the subsequent CSRs along with close monitoring of their implementation). This and earlier (summer 2014) events in the banking sector brought about action by the Bulgarian authorities to reveal the nature and full extent of the behavioral reversals and to correct the reversals by starting to clean up the banking system. This, however, triggered reactions affected by such action, who then successfully slowed down or even partially neutralized the corrective efforts by the authorities. The dissipating market pressure on the country was conducive to such counter efforts.

One of the large banks collapsed during this episode. Strongly prevailing social norms on fiscal discipline in Bulgaria ensured that the associated cost to the tax payer was minimized by a full bail-in of private claims, including a significant amount of unsecured corporate and retail deposits. This was a much stricter treatment of investors than in Slovenian in 2013 or what was typical in Europe and did not create a backlash. This shows that social norms can be resilient in one area, while massive reversals can take place in another. Interestingly, the Bulgarian currency board system was set up in 1997 followed a traumatic banking crisis. The social norms regarding the key prerequisites of a stable currency board system, well anchored fiscal policy, safe financial system and strong and independent supervision, emerged from this episode simultaneously.

The other large bank involved in the crisis still needs to strengthen its capital position. As the literature suggests (Peek and Rosengren 2005; Caballero et al. 2008), weak capital positions are conducive to a deterioration in corporate governance. Thus some vulnerability remains.

Regarding external anchors, the country surveillance of the European Commission (MIP) revealed the full extent and nature of the problems only in its 2015 report (European Commission 2015b). The massive behavioral reversals on several fronts thus remained latent in all likelihood for a rather long time. Apparently, the IMF surveillance did not fare any better.

Reform Reversals in Hungary and Poland

The nature of the reversal episodes in Hungary and Poland is rather different from those in other FTEs, as one could observe the emergence of a characteristically different overall approach to the whole functioning of the economy.

The Hungarian episode started at the deepest point of the crisis, when the country had an EU-IMF financial assistance program. Convergence had paused for a long time prior to the crisis (Fig. 2), and the socio-liberal coalition that governed the country previously had disintegrated. The country was led by a technocratic government prior to the 2010 elections.

Social inequality had increased, albeit not dramatically (Fig. 3). Large groups in society, mostly in the middle of the income distribution, were caught in a difficult situation with escalating monthly payments on foreign-denominated loans (typically mortgage and car loans). A partly overlapping group (toward the lower end of the income distribution) was struggling with paying their utility and energy bills. (4) In both cases, the industries involved (energy, utilities and banking) were showcases of the Hungarian transition reform drive and firms in these industries were almost fully privatized, predominantly to foreign strategic investors. The reform in the energy and utility sectors had included major regulatory reforms, creating an almost textbook-type clean system. In both cases though, unexpected negative side effects emerged in the form of strong pressures on family budgets.

Many of the first reform reversals in Hungary were justified by the need to address these problems under a rather tight fiscal constraint, imposed by the market (difficulties with, and a high cost, of market funding) and by external anchors (an EU-IMF program, and the EDP under the SGP). The previous model for the utility and energy sectors were largely dismantled, companies in this sector renationalized and retail price controls in the utilities and energy sector re-introduced. A special tax was levied on the banking sector and the cost of unwinding the foreign-denominated loans was largely pushed on the banks (most of which were foreign). The unwinding of these loans was key to removing the existential threat to the middle class, while the price control measures in the energy and utilities sectors were central to addressing the pressure on low-middle income families and thus to solidify the political support for the ruling coalition. (5) It was also considered essential to reestablish room for maneuver for the central bank, so that the weaker exchange rate of the domestic currency, an essential element of the new macroeconomic policy course, did not hit the families with foreign-denominated loans. This trade-off between the need for the economy to have a weaker exchange rate and the need to keep the exchange rate strong to avoid the escalation of payments by the loan holders made the necessary external adjustment politically painful and put a strait-jacket on monetary policy (previously) under the socialist-liberal government and under the EU-IMF financial assistance program. The funds transferred back to the government from the second pillar pension enhanced the government's capacity to move ahead with the re-nationalization of companies in the energy sector (possibly later in banking) in an environment where its borrowing capacity was rather limited.

Fundamental changes in economic policy did not stop there. Many of the next steps entailed reversals of previous reforms, sometimes central ones (Bokros 2014). Energy and utility companies were repurchased, made easier by regulatory changes. Two large banks (MKB and Budapest Bank) were repurchased, in both cases in market transactions initiated by their foreign owners. Regulatory changes in the food retail-trade sector targeted foreign-owned chains. Tobacco trade was restricted to state-licensed shops. Barriers to entry were introduced in the pharmacy market. Education and health-care service provision was centralized, and more generally, many of the previously main functions of local governments (e.g., education) were removed and centralized, and the entire approach to social benefits was overhauled. Major labor market reforms which were directed at returning older workers, people with partial disability and previously long-term unemployed to the labor market, were reversed. A main element of the labor market reforms was a major expansion of the public work scheme, which supported by reforms of the social benefit system, made this the principal form of income support for many people. As a result of the labor market reforms, labor force participation increased rapidly.

While in the non-tradable sectors measures targeted foreign-owned firms, in the tradable sectors, support to FDI increased and the government signed strategic collaboration agreements with large foreign companies operating in Hungary. The exchange rate policy also strongly supported this sector. The role of the state increased on many fronts, and many of its functions became more centralized. The economic situation stabilized and the economy returned to relatively strong growth by the next elections in 2014. With growth kicking in, and employment picking up, real disposable family income started to grow, poverty indicators improved and inequality started to subside, reversing the previous trends (Figs. 4, 5).

Turning to Poland, in sharp contrast the PiS-led coalition took over in late 2015 in a mainly strong economic environment following an enviable economic performance during the crisis under the previous PO-led central-right coalition governments. Convergence had progressed at a rapid pace (Fig. 2) during the two terms of the PO-led coalition, following a period of previously rather slow convergence (including during a short-lived previous PiS-led coalition that had not been very successful). Unlike in many other countries during the crisis, social inequality in Poland had continuously declined from an already modest level (Fig. 3). The Polish middle class was perhaps the biggest winner during this period, gradually improving its position inside the country and enjoying enviable income growth, also relative to its peer groups in Europe before and during the crisis (Figs. 4, 5).

Unlike in Hungary in 2010, there was no need for an immediate fiscal adjustment. The government's plan to increase child support put some upward pressure on the budget deficit, which stood rather close to the 3% of GDP EDP threshold since Poland had just exited the EDP prior to the 2015 elections. Moreover, there was no sign of any pressing issues which might pose a (real or perceived) economic (existential) threat to any particular segment of society.

While in the political arena a conflict with the European institutions emerged rather soon after the new PiS-coalition took power, on the economic front both the attitude and communication was demonstratively more in line with EU requirements. The diagnosis underlying the strategy of the government (Responsible Development Plan, Ministry of Development 2017) was very close to the one the European Commission offered in its European Semester Country Reports for Poland. In many areas, it built on recent initiatives to address similar issues across Europe.

A main element of this plan is the intention to rebalance the roles of foreign and domestic firms in the Polish economy, by increasing the share of the latter. The state is envisaged a major role in this process, including in channeling capital into export and innovation-oriented projects, as well as into infrastructure. The role of SOEs is also envisaged to increase, and their corporate governance and control to be changed. Previous privatization plans (by the PO-led coalition) have been put on hold, and with the repurchase of PKO SA and its transfer to the state-owned insurance (financial) group (PZU), reversal in a central area has started (albeit initiated by the foreign parent bank group, Unicredit, and carried out on market terms). So, as in Hungary, a stronger role for the state and a centralization of functions and consolidation of institutions is a key characteristic of the developments in Poland.

The formal (legal, institutional) changes in this plan, some of which are clear reversals relative to the transition and EU accession reforms, are in themselves not necessarily creating major deviations from a well-functioning and well-governed social market economy. In fact, their declared goal is to achieve just that, as interpreted by the PiS-led coalition. Moreover, they do not create unresolvable conflict with EU law. They may well test the limits of the existing capacities of the Polish state, so capacity building is essential (as realized by the strategy itself), but the largest source of potential risk, as we have seen in many of the episodes described above, is that of behavioral reversals inside institutions, and the oft-problematic interactions and spillovers of these reversals. As in Slovenia and Bulgaria, the financial (banking) sector is a vulnerable part of the economy in this regard, and one which can propagate the behavioral reversals in a quick and for long undetected manner. SOEs are another similarly vulnerable part of the system.

The Plan also pays close attention to the underdeveloped parts of Poland (Polska B), which is a traditionally strong voter base of the incumbent coalition (Fig. 6). The simultaneous emphasis on a stronger role of the state, on a more top-down approach, and on the development of these parts of Poland is not a coincidence, but a key characteristic of the overall approach.

While the Hungarian and Polish reform reversal episodes show some similarities, they are fundamentally different in nature. The Hungarian reform reversal episode seems to be a classical case in political economy and fits well also into our overall framework. It starts with an election crush of a disintegrating coalition which had significantly, and for a long time, underperformed on the economic front in absolute and relative (to other countries in the region) terms. It had left the major existential problems of the middle class, lower-middle class and poor families unaddressed. The value of previous reforms was not tangible for many in a society where protection and dependency was traditionally strong. Furthermore, a long period of uncertainty had made the benefits of previous reforms look smaller, even for those who had perceived them.

The Polish reform reversal episode followed a period of enviable economic and social performance, in which not only social inequality had been continuously reduced, but the Polish middle class enjoyed an unparalleled absolute and relative (to other countries) income gain and Polish private business thrived. There were no apparent signs of any existential threat to any group in society. All this appears to be in sharp contrast with the prediction of the evolutionary-institutional school, which expected the middle class and the emerging new private sector to support and protect reforms (Roland 2001). In many aspects, the economic (development) strategy of the PiS-led coalition demonstratively builds and relies on the EU approach. Nevertheless, similar to the Hungarian approach, it entails a much stronger role for the state, a much more centralized and consolidated state, and aims to increase the role of Polish capital (firms) in the economy. Unlike the Hungarian approach, the focus of the latter is the tradable sector, with an emphasis on the move toward an innovation-based economy. Similarly to Hungary, the underdeveloped parts of the country (Polska B) are given particular attention.

Lessons from the Reversal Episodes

The fast journey from central planning to EU (euro area) membership stress-tested the social learning processes of FTEs. The desire for a higher standard of living and freer life, to be anchored to the West and to enter the EU, spurred major reform waves and led to the rapid introduction of institutions that evolved as bestpractice institutions in highly developed countries. This process most likely accelerated social learning, but apparently in many FTEs the pace of such learning was not fast enough to keep up with the pace of the reforms, leaving these best-practice institutions with social norms in the countries concerned that were not sufficiently strong to maintain them. Perhaps not surprisingly, widespread reversals emerged in the region, especially when the crisis hit these countries. This seems to suggest that reversals are an inherent characteristic of the FTEs' journey toward a modern social market economy.

Similar to the varied factors and motivations that can promote reforms, the reversal episodes suggest that there are a number of rather different forces at work in creating reversals. Some factors are common in certain areas of reform, such as in the reversals of the pension system, and appear to hold across countries, political families and types of government. Some factors appear to be present when certain types of governments are in place, and work across many areas, as appears to be the case in Hungary and Poland at present.

Reform reversals can be formal (which challenge legislation or rules) or behavioral (which erode the quality of an institution by materially changing the way it works). Spillovers, from one type of reversal to another, from one area to another, or from one institution to another can play an important role in influencing the nature and dynamics of reform reversals. In many cases, it is the interaction of reversals in different sectors that has created a full-blown reform reversal episode. Partial and opportunistic reversals can happen, particularly in areas where the implications of such reversals are not immediate and not easily visible (such as pension reforms). These types of reversals have tended to occur more frequently when governments are weak. In contrast, major and multifaceted reversals seem to require strong and stable governments.

While the crisis undoubtedly made countries more prone to reversals, even fundamental reform reversals have occurred when a country had weathered the crisis well and where inequalities had declined. For example, although the Polish middle class enjoyed an unparalleled income gain and Polish private business thrived during the crisis, contrary to previous expectations (Roland 2001), this did apparently little to protect the previously introduced reforms.

While reversals seem to have occurred when institutions and social norms were not sufficiently strong and social norms tend to impact different sectors similarly, there are sometimes interesting asymmetries. For example, in Bulgaria, the social norms which supported well-anchored fiscal policy and strong and independent banking supervision (following the creation of the currency board in the 1990s) changed asymmetrically. While fiscal discipline remained strong, the quality of banking supervision was allowed to deteriorate significantly. Interestingly, Romania showed an opposite pattern of asymmetric reversals in which banking supervision remained strong and independent but fiscal discipline eroded. The fiscal policy regime remained sufficiently strong in Bulgaria to allow and encourage a reduction in the fiscal cost of bank failure through a full bail-in of creditors in a failed bank. Such strict bail-in did not happen in Slovenia. Even the bail-in of a relatively small amount of subordinated debt triggered a behavioral reversal in another area (central bank independence).

The banking (and more broadly, the financial) sector seems particularly prone to behavioral reversals, both in public and private institutions. Like in the other cases, it is the confluence and interplay of behavioral reversals in different areas and institutions (banking supervision, private and state-owned banks, nonbank corporations) that has led to full-blown reversal episodes. The rather low level of transparency and public scrutiny in this sector seems to be particularly conducive to reversals. For example, the very restricted access to decision making in banking supervision, or central banking, may have provided a convenient veil in some countries for forces seeking to bring about behavioral reversals in financial institutions. State control over state-owned banks and nonbank companies are also frequently not subject to strong public scrutiny and the level of transparency tends to be low. Finally, public scrutiny of private banks and companies is perhaps even weaker and the degree of transparency even lower. Since reform reversals are dominantly behavioral in nature in this area, and thus by nature more difficult to detect, reversals may tend to remain latent for a long time and are revealed only when they lead to a crisis or a major event (such as the collapse of a large bank). This lack of transparency may also explain why social norms, even if they happened to be not sufficiently strong, could not prevent more efficiently reversals that turned out to be so costly for society and why vested interest, for which the sizable direct gain could overrule norms, could drive reform reversals so far.

Well-designed fiscal systems, national or European, are thought to be the best anchor for fiscal policy and an efficient way to limit myopic political intentions to opportunistically loosen fiscal policy. This is the most traditional area of EU policy coordination and rules (SGP), and reforms at the European level following the crisis also focused on further strengthening national fiscal systems and rules. However, as the Romanian example shows, if social norms are not sufficiently strong, such national fiscal systems (or rules), even if they are close to best practice, will have little impact. The preventive arm of the SGP (the SDP) in this case, which is the EU anchor, does not seem to have been strong enough either. When fiscal rules and systems are sufficiently strong (such as the 3% of GDP EDP deficit limit or the Polish national debt rule), but social norms are not, the intention to create more fiscal space may generate spillovers that lead to reform reversals in other areas. This has occurred most frequently in pension systems, but also in tax policy and in corporate governance and the state control of SOEs.

While the opportunistic fiscal behavior of governments is a phenomenon well described by the traditional political economy models, spillovers from the fiscal side into reform reversals in other areas seem a unique characteristic of countries with weak social norms, such as the FTEs. Perversely, a perhaps premature pension reform introduced with strong external involvement, and with some characteristics of institutional monocropping, might have been an easy target for opportunistic politicians in such an environment. So in a way, the major pension reform reversals in Hungary and Poland were accidents waiting to happen. The design problems and other circumstances might have only added to the temptation and made it easier for reversals to be accepted by the public. The political orientation or other characteristics of the government do not appear to have mattered much. The ultimate determining factor has clearly been the weakness of social norms. The fact that such opportunistic reversals in the Baltic countries were later stopped or even reversed further supports this conclusion.

Finally, poor design of the original reforms and a lack of systematic assessment of reforms after their introduction may have made certain reforms more prone to reversal--particularly if the political system allowed such problems to accumulate, and the negative consequences impacted on distinct groups in society. However, in almost all cases discussed here, such design problems could have been addressed without a reform reversal.

The reform reversal episodes in Hungary and Poland entail special characteristics, raising the question of whether these episodes represent more than just the usual manifestations inherent to the transition process. In fact, it is the nature of developments in some areas that are not covered by this paper, such as the independence of courts and media, which raises this question more forcefully than the reform reversals described above (Mizsei 2018). While in the past, the Hungarian government used more confrontational rhetoric about the issues discussed in this paper, the Polish government economic program in many aspects is very close to the mainstream in Europe. In both countries, standard legal processes could in most cases resolve the conflict with EU law. A key problem here is path dependence. Such reversals can put these economies on a lower growth path for a long time, partially taking away the benefits of EU membership.

External Anchors

The Washington institutions played a dominant role in shaping the transition process from the very beginning until the EU accession process started, and played a key role thereafter. Following the start of the accession process, the EU gradually took over as the dominant external anchor. The IMF (joint EU-IMF) financial assistance programs--which played an important role in some countries during the crisis--were strong forces that promoted reforms and discouraged reform reversal intentions. But it seems they could not resist opportunistic reform reversals, such as a temporary reversal of pension reforms. The surveillance work of the IMF, while highly effective in detecting formal reform reversals, was less effective in detecting behavioral reversals and overall could do little to prevent or re-reverse such reversals.

The strong desire to join the EU has been a major force promoting reforms in the region, also in areas beyond what is required by EU legislation. The capacity of the EU to promote reforms or deter reversals following accession has, however, been markedly weaker. Euro area membership seems to have worked in a similar way. The political ambition to join the euro area, even as an exit route out of a program (as in Latvia), not only prevented reversals but also promoted a new wave of reforms. On the other hand, as the example of Slovenia shows, euro area membership in itself cannot arrest reform reversals, including in critical areas, if that area is not covered by EU Law and regulation, nor overseen by a euro area institution.

The European Commission has strong legal instruments to act against reform reversals in areas which are protected by EU legislation, such as competition, stateaid, or central bank independence. It has therefore been a major force in stemming the tide of formal reform reversal in almost all episodes discussed here. The anchoring role of the EU was however much weaker in the case of behavioral reversals, even in areas where EU law had some relevance to the case and/or EU institutions had a mandate to intervene (such as the MIP). Generally, it seems that institutions in FTEs are still rather fragile, their quality is at risk from erosion, and tend to be prone to partial capture by interest groups. These forms of reform reversal are much more difficult to address with formal instruments. In fact, strongly protected (also by EU legislation) independent institutions in such an environment can be abused as platforms for reform reversal and can further weaken the already weak anchoring power of social norms, as vested interest with direct gain from reform reversals can hide behind this veil.

Besides its classical role as the guardian of the SGP, that is, the set of rules and procedures that were introduced to anchor the fiscal policies of its member states, the European Commission was given new mandates to promote reforms via the MIP and the European Semester. These new mechanisms, particularly the MIP, have the power to reverse reform reversals. But, as the reform reversals in Slovenia and Bulgaria show, it has only been when these reversals have led to excessive macroeconomic imbalances that effective action has been taken. The preventive power of these new instruments appears to be much weaker.

The fact that several reforms had not been developed "in house" by the FTEs themselves, but rather imported from these external anchors, may have weakened the social learning process. In many cases, FTEs took the best practice from other countries, with a lack of thorough conviction and indigenous development, which left the best-practice institutions with social norms that were not sufficiently strong to maintain them.

How to Make These Countries More Resilient to Reform Reversals

Our analysis naturally leads to the conclusion that the ultimate solution to prevent reform reversals is to accelerate social learning processes, particularly among parallel communities. This could include increasing the exposure of people in the FTEs to communities with social norms that correspond more closely to the existing rules and institutions, both inside their own countries and outside. Experiences suggest that both the introduction of the initial reform and also the duration of the reform process would benefit from stronger national debate to develop a sovereign national understanding and collective memory of the reasons for the optimal design chosen.

It is also important to focus on the quality and internal coherence of reforms and newly created institutions and to carefully monitor their functioning to detect behavioral reversals as early as possible. In addition to and interacting with domestic anchors, external anchors, particularly EU institutions, can help a lot in this regard by deepening their surveillance work and also focusing on behavioral reversals and the typical spillovers patterns found in this paper. The risk that a successful formal (legal) step to stem a reform reversal can trigger a behavioral reversal in the same or another area deserves particular attention.

The ambition to join the euro area could boost reform efforts in FTEs that are not yet a member of the euro area. Creation of the Banking Union has significantly deepened integration in the euro area in the area of banking supervision via the creation of the Single Supervisory Mechanism (SSM). There is now also a formal way for non-euro area members to establish a close cooperation with the SSM, which may offer a route for aspiring euro area Member States, such as Bulgaria, to anchor the necessary improvements in banking supervision early on in the process toward membership. Close cooperation with the SSM may also help to avoid the kind of banking reform reversal episodes that are discussed in the paper in other countries where there are new developments (such as a major increase in the market share of domestically owned, private or public, banks) which suggest increased vulnerabilities in this regard. Given the central role of the banking/financial sector in propagating (behavioral) reversal in the system, it seems important to look for ways to strengthen the external anchors in this area.

The recent initiative of the European Commission to set up a Structural Reform Support Service (SRSS) and the new plans to significantly increase its size can help to strengthen the anchoring role of the EU. The SRSS can help to promote quality reforms and coherence among reforms and perhaps can also look into how to create stronger domestic anchors for reforms. More generally, European institutions should develop a better understanding of the social learning processes in these countries and finds ways to help strengthen them.

References

Abiad, A., and A. Mody. 2005. Financial reform: What shakes it? What shapes it? The American Economic Review 95: 66-85.

Bank of Slovenia. 2015. Report of the Bank of Slovenia on the causes of the capital shortfalls of banks and the role of the Bank of Slovenia as the banking regulator. Ljubljana, March 2015.

Bielawska, K., A. Chlori-Domiriczak, and D. Startko. 2017. Retreat from mandatory pension funds in countries of the Eastern and Central Europe in result of financial and fiscal crisis: Causes, effects and recommendations for fiscal rules. Paper presented at the 2018 ASSA meetings.

Bisztray, M. 2016. The effect of FDI on local suppliers: Evidence from Audi in Hungary. Discussion papers MT-DP--2016/22, Center for Economic and Regional Studies, Hungarian Academy of Sciences.

Bokros, L. 2014. Reform reversal in Hungary after a promising start. In The great rebirth: Lessons from the victory of capitalism over communism, ed. A. Aslund and S. Djankov, 39-51. Washington, D.C.: Peterson Institute for International Economics.

Bocskei, B. 2015. Rezsicsokkentes: a kozpolitikai valtozas mint politikai innovacio. Politikatudomanyi Szemle, XXIV/4,94-114.

Buti, M., A. Turrini, P. Van den Noord, and P. Biroli. 2008. Defying the 'Juncker curse': Can reformist governments be re-elected? European Commission, Economic Papers 324.

Caballero, R., T. Hoshi, and A. Kashyap. 2008. Zombie lending and depressed restructuring in Japan. American Economic Review 98: 1943-1977.

Campos, Nauro F., and Fabrizio Coricelli. 2009. Financial liberalization and democracy: The role of reform reversals. IZA discussion paper no. 4338.

Enyedi, Z., Z. Fabian, and R. Tardos. 2015. Partok es szavazok, 2002-2014. TARKI.

European Commission. 2015a. The 2015 aging report. European Economy 3/2015.

European Commission. 2015b. Macroeconomic imbalances. Country report: Bulgaria 2015, occasional papers 213, June 2015. European Commission, DG ECFIN.

European Commission. 2017a. Debt sustainability monitor 2016. Institutional Paper 047, January 2017.

European Commission. 2017b. Vade Mecum on the stability and growth pact, 2017 Edition. Institutional Paper 052, March 2017.

Evans, P. 2004. Development as institutional change: The pitfalls of monocropping and the potentials of deliberation. Studies in Comparative International Development 38 (4): 30-52.

Iancu, D.C., and M. Ungureanu. 2013. Conditionality and reform reversal: The case of the Romanian civil service. Draft paper presented at the NISPAcee ..., 2013. nispa.org.

Keereman, F., and I P. Szekely (eds.). 2010. Five years of an enlarged EU: A positive sum game, 288. Berlin: Springer.

March, J.G., and J.P. Olsen. 2006. The logic of appropriateness. In The oxford handbook of public policy, ed. R.E. Goodin, M. Michael Moran, and M. Rein, 689-708. Oxford: Oxford University Press.

Ministry of Development, Poland. 2017. Responsible development plan, https://www.mr.gov.pl/media /14873/Responsible_Development_Plan. pd f.

Mizsei, K. 2018. The role of the state in transition: Reversals of reforms. Paper presented at the 2018 ASSA meetings.

Naczyk, M., and S. Domonkos. 2016. The financial crisis and varieties of pension privatization reversals in Eastern Europe. Governance 29: 167-184.

Peek, J., and E. Rosengren. 2005. Unnatural selection: Perverse incentives and the misallocation of credit in Japan. American Economic Review 95: 1144-1166.

Roaf, J., R. Atoyan, B. Joshi, K. Krogulski, and an IMF Staf Team. 2014. 25 years of transition, postcommunist Europe and the IMF, regional economic issues special report, October 2014. Washington, D.C.: The IMF.

Rodrik, D. 1995. The dynamics of political support for reform in economies in transition. Journal of the Japanese and International Economies 9: 403-425.

Rodrik, D. 1996. Understanding economic policy reform. Journal of economic Literature 34: 9.

Rodrik, D. 2006. Goodbye Washington consensus, hello Washington confusion? A review of the World Bank's economic growth in the 1990s: Learning from a decade of reform. Journal of Economic literature 44: 973.

Roland, G. 2001. Ten years after...transition and economics. IMF Economic Review 48: 29-51.

Roland, G. 2002. The political economy of transition. The Journal of Economic Perspectives 16: 29.

Rotter, J.B. 1954. Social learning and clinical psychology. New York: Prentice-Hall.

Sen, S., and S. Airiau. 2007. Emergence of norms through social learning. In IJCAI'07 Proceedings of the 20th international joint conference on artificial intelligence, 1507-1512.

Szekely, I.P., and M. Ward-Warmedinger. 2018. Reform reversal in former transition economies (FTEs) of the European Union: Areas, circumstances and motivations. IZA Working Paper.

Varga, J., and J. in't Veld. 2014. The potential growth impact of structural reforms in the EU. A Benchmarking Exercise. Economic Papers, European Commission 541. http://ec.europa.eu/economy_ finance/publications/economic_paper/2014/pdf/ecp541_en.pdf.

Young, H.P. 2015. The evolution of social norms. Annual Review of Economics 7: 359-387.

(1) A beautiful memory of this period has been given recently by President Tusk in his acceptance speech at the University of Pecs where he received an honorary doctorate. He said: "To put it simply, Hungary was, for my generation, a reflection of the West, dreamed of and unreachable for decades". http://www. consilium.europa.eu/en/press/press-releases/2017/12/08/acceptance-speech-by-president-donald-tuskupon-receiving- honorary-doctorate-from-the-university-of-pecs/#.

(2) For more information on the Macroeconomic Imbalances Procedure, see https://ec.europa.eu/info/ business-economy-euro/economic-and-fiscal-policy-coordination/eu-economic-governance-monitoring -prevention-correction/macroeconomic-imbalance-procedure_en.

(3) For more information on the European Semester, see https://ec.europa.eu/info/business-economy-euro/ economic-and-fiscal-policy-coordination/eu-economic-governance-monitoring-prevention-correction/ european-semester_en.

(4) For the strong impact of these two problems on the voting patterns, see Enyedi et al. 2015). FIDESZ benefitted most from the dissatisfaction of these people, especially among those that were in the middle of the income distribution. The parties of the previous ruling coalition suffered significant vote losses because of these problems.

(5) For an analysis of how FIDESZ utilized these problems to gather strong public support, see Bocskei (2015).

https://doi.org/10.1057/s41294-018-0077-1

Istvan P. Szekely [1,2] (iD) * Melanie Ward-Warmedinger [1,3]

Published online: 23 October 2018

Views expressed in this paper are strictly personal and do not necessarily represent the views of any of the institutions the authors are affiliated with. An earlier version of this paper was presented at the 2018 ASSA Meetings in Philadelphia, PA, January 4-6. We are grateful for the comments of our discussants, John Bonin, Mitchell Ornstein, and Sergei Guriev, and the session participants. We are also grateful for comments and/or assistance to an anonymous referee and to Francisco Barros Castro, Lajos Bokros, Agnieszka Chlon-Dominczak, Stefan Ciobanu, Ivan Csaba, Patrick D'Souza, Per Eckefeldt, Stanislaw Gomulka, Bostjan Jazbec, Vasileios Karantounias, Julia Kiraly, Gilles Mourre, Wojciech Paczynski, Marc Puig, Ralph Schmitt-Nilson, Andras Simor, Attila Karoly Soos, Klara Stovicek, Gyorgy Suranyi, Marton Szili, Chin Tong, Maarten Verwey, Liviu Voinea, Bartlomiej Wiczewski, Rafai Wieladek and Tomasz Zdrodowski. We remain responsible for all remaining errors.

[mail] Istvan P. Szekely

Istvan-Pal.SZEKELY@ec.europa.eu

[1] Directorate-General for Economic and Financial Affairs (DG ECFIN), European Commission, Rue de la Loi 170, 1049 Brussels, Belgium

[2] Budapest Corvinus University, Budapest, Hungary

[3] IZA, Bonn, Germany

Caption: Fig. 1 Per capita GDP relative to frontier (in PPS, group of high-income countries = 100). Note The frontier (= 100) is defined as the average of high-income countries (Sweden, Denmark, Netherlands and Austria). Baltics is a simple average of the series for Estonia, Latvia and Lithuania. Source: Eurostat

Caption: Fig. 2 Per capita GDP relative to frontier (in PPS, group of high-income countries= 100). Note The frontier (= 100) is defined as the average of high-income countries (Sweden, Denmark, Netherlands and Austria). Source: Eurostat

Caption: Fig. 3 Social inequality in Hungary and Poland. Source: Eurostat

Caption: Fig. 4 Income share of the middle income quintiles in Hungary, Poland and the euro area. Source: Eurostat

Caption: Fig. 5 Household income developments in Hungary, Poland and the euro area

Caption: Fig. 6 Regional income differences (2013) and voting patterns (2015) in Poland. Source: Ministry of Development Poland (2017) and Wikipedia