One important feature in the debate about the aftermath of the crisis is the imbalance between austerity and growth. Whether or not there has been too much austerity (Alesina and Giavazzi 2012) is obviously much more controversial than the need for an effective agenda for the resumption of economic growth. Such growth agenda is often presented as a set of structural reforms that, once implemented, will be able to increase competition and efficiency in labour, credit, and product markets.
Something not fully appreciated in this view is that there are various key aspects of reform that we still know little about, among them, their reversibility. Although a main objective of the theoretical literature is the design of reform packages that have a low probability of reversal (Fernandez and Rodrik 1991, Dewatripont and Roland 1995, Campos and Coricelli, forthcoming), there has been little effort so far to think about and empirically understand the main determinants of reform reversals. This column reports on recent research addressing this issue (Campos and Horváth, forthcoming).
What drives reversals?
What are the factors that may help us to better understand reform reversals? Unfavourable changes in economic conditions represent a possible reason (Dewatripont and Roland 1995). Sudden increases in unemployment or a slowdown in growth rates may cause reversals. Changes in political conditions generate another set of potential reasons (Gehlbach and Malesky 2010). For instance, turnovers in the major political party in government may make reversals more likely (consider the event of a left-wing succeeding a right-wing government or vice-versa). Trying to reflect these views, Campos and Horváth (forthcoming) estimate simple logit models in which the dependent variable is coded one if there was a reversal in a reform in a given year for a given country, and zero otherwise.
One important consideration is that although different reforms may be driven by a common set of factors, reversals in different reforms are better explained by different factors. Standard studies on reform dynamics explain both positive and negative deviations from the mean, while we expect in the case of reversals that the bottom of the distribution of changes in reform indicators (that is, reversals) is driven by different factors. It might also be noted that the question on the determinants of reform reversals has not been investigated as the (limited) focus has been mostly on the effects of reversals on economic performance (e.g., Merlevede 2003).
One might well ask: If the issue is so important, why so little research? Measurement of structural reforms provides one main reason for why little is known about reform reversals. The most widely used reform measures portray reforms as a smooth, uninterrupted process of continuous improvement where reversals occur only sparingly. This can be rationalised by policymakers being clever and well-informed, political pressure on the international organisations constructing such measures (to not lower scores), or national authorities mistakenly understanding survey questions as referring to cumulative (instead of current) efforts. We focus on improved measures of reform efforts (prices and wages liberalisation, external liberalisation and privatisation) in 25 central and eastern European countries (Campos and Horváth, 2012). According to these measures, reforms follow a much more turbulent process characterised by experimenting, learning by doing, sudden advances and indeed sharp reversals. Focusing on the EBRD reform indexes (which cover the same set of countries as we do), Merlevede (2003) reports that reversals are observed in about 9% of the cases. Our three indexes support more frequent reversals: 20.3% for external liberalisation, 14.2% for prices and wages liberalisation, and 18.1% for privatisation. Merlevede (2003) also reports that only half of the countries experienced reform reversals. According to our indexes, all countries experienced reform reversals. Figures 1 and 2 show the extensiveness and intensity of reversals using our indexes, respectively.
Figure 1. The share of countries experiencing reform reversals.
Figure 2. The intensity of reform reversals.
For the case of price and wage liberalisation, our econometric estimates suggest that reversals seem driven primarily by political factors, more specifically by various forms of direct political protest (for instance, general strikes). It is worth highlighting that factors that are often associated with the implementation of reforms (such as growth and democracy) seem to have little effect in explaining reform reversals (or indeed reversals duration). The importance of labour strikes also raises the issue of the timing of reforms: because price liberalisation was implemented to a large extent before external liberalisation and privatisation in this set of countries, the most effective method to revert prices liberalisation seem to have been direct mass political action.
The results for the two other reforms are also interesting and intuitive. In the case of reversals of privatisation efforts, the main explanatory factor we find is FDI inflows as a share of GDP. An increase in FDI inflows significantly lowers the probability of a reversal in privatisation efforts. The main reasons we associate with external liberalisation reversals are average growth rates of OECD economies (more favourable external conditions tempering the impetus to reform) and terms of trade (which have the opposite effect).
These results hold well against a series of robustness checks. For instance, they change little if instead the focus is on the persistence (or duration) of reversals. Moreover, none from a number of potentially important variables proved to be systematically related to reform reversals. Specifically, there seem to be no systematic effects from inflation, financial crises, fiscal deficits, EU negotiations, war, timing of elections, ideological alternation in government, number of leadership changes, distance from Brussels, and whether or not the country was previously part of the Soviet Union.
The evidence discussed above provides support for the interesting and unexplored hypothesis that the drivers of reform reversals differ both across specific reforms as well as in the determinants of overall reform dynamics. On the latter, although it is often claimed that factors such as democracy or macroeconomic stabilisation are key factors in the implementation of structural reforms over time and across countries, the above discussion suggests that such general factors are of limited assistance in understanding reform reversals. On the idea that reversals of different reforms are driven by different factors, the above discussion indicates that indeed this seems to be the case.
Specifically, the evidence shows that:
- Inflows of foreign direct investment diminish the likelihood of privatisation reversals,
- Worsened terms of trade boosts the probability of external liberalisation reversals, and
- Labour strikes increase the probability of price liberalisation reversals.
These results suggest that careful debate and reflection are still needed when designing and proposing a growth agenda based on structural reforms and aimed at taking us out of this crisis.
Alesina, A and F Giavazzi (2012), “The austerity question: “How” is as important as “how much”, VoxEU.org, 3 April.
Campos, NF and F Coricelli (forthcoming), “Financial liberalization and reversals: Political and economic determinants”, Economic Policy.
Campos, NF and R Horvath (forthcoming), “On the reversibility of structural reforms”, Economics Letters.
Campos, NF and R Horvath (2012), “Reform redux: Measurement, determinants and growth implications”, European Journal of Political Economy, 28:227-237.
Dewatripont, M and G Roland (1995), “The design of reform packages under uncertainty”, American Economic Review, 85:1207-1223.
Fernandez, R and D Rodrik (1991), “Resistance to reform: Status quo bias in the presence of individual-specific uncertainty”, American Economic Review, 81:1146-1155.
Gehlbach, S, EJ Malesky (2010), “The contribution of veto players to economic reform”, Journal of Politics, 72:957-975.
Merlevede, B (2003), “Reform reversals and output growth in transition economies”, Economics of Transition, 11:597-751.