Sergio Godoy Saturday, June 12th

1. Columbia Business School Growth, Initial Conditions, Law and Speed of Privatization in Transition Countries: 11 Years Laterpresented at Conference“Comparative Transitions: A Critical Review”London Business School Sergio Godoy Saturday, June 12th

2. Main Themes • Motivation • Contributions • Literature Review • Data • Results • Conclusions • Future Line of Research

3. Motivation • The record of transition countries have been at best mixed and several countries has shown a dismal growth performance in the last decade (see Figure 1). • We would like to check whether privatization policies, institutions and /or initial conditions can explain these huge cross sectional growth differences among transition countries in the last eleven years. • Since the mid-90s a growing empirical literature appeared to look at these issues one or two at the time. • In addition, for the most part, this literature failed to recognize the possibility that policy measures can be endogenous to the institutional setup and/or to the legacy left by the Communist era. • Moreover, this literature failed to measure correctly the concept of privatization speed except for one paper.

4. Figure 1

5. Contributions (I) • We consider simultaneously which of the these three dimensions are more important in explaining cross sectional growth among transition countries. • The liberalization and privatization indexes can be endogenous to the economic success of these countries. Thus, we make policy indexes endogenous. • Liberalization indexes are really a measure of the level of liberalization that has been achieved by a country and not how fast it has liberalized.

6. Contributions (II) • Thus, for assessing policy speed it is necessary to use some measure of change in these indexes. In our paper we consider both policy level and policy speed. • Another contribution of our paper is that for gauging initial conditions we include direct measures of institutional conditions. • We believe that this is a more important part of the legacy left by the Communist era than the traditional economic initial conditions.

7. Literature Review • Importance of policies and, more specifically, liberalization and privatization (De Melo, Denizer and Gelb (1996); Aslund, Boone and Johnson (1996)). • Role of initial conditions as a “burden” left by the Communist era and use of instrumental variables in order to correct for endogeneity of policy indexes (De Melo et al. (1997) and Heybey and Murrel (1999)). • Role of institutions and, in particular, legal institutions (Havrylyshyn and Rooden (2000), Stiglitz (2001)).

8. Data • The dependent variable is the total growth rate from 1990 to 2001 of 23 transition countries. • For measuring strength of legal institutions we use data from a survey conducted by the World Bank and ERBD. Specifically, we use the proportion of firms in the countries of the sample that in 1997 agree to: To what degree do you agree that the legal system will uphold contract and property rights? • We call this variableEnf.

9. Data: Policy Variables • The effect of policy is quantified by two distinctive variables. First, for policy speed, we use (one tenth) of the difference between the 2000 average of the ERBD indexes of Small-scale and Large-scale privatization and the 1991 average of the same indexes (Dp00p91). • Second, the effect of policy level is captured by the 2000 observation of the same index (Privi00). • We tried other measures of liberalization and of macroeconomic stability but they did not much add much explanatory power.

10. Data: Instruments (I) • We use as instruments the following variable: • Years under communist rule • Late 1980s defense expenditure as % of GDP • 1990 black market premium • Late 1980s sector distortion • Late 1980s trade distortions as % of GDP • For details see De Melo, Denizer, Gelb and Tenev (1997) and Popov (2000).

11. Data: Instruments (II) • These instruments are chosen based on the two distinctive criteria used for choosing instrumental variables: • First, we believe that our instruments are clearly correlated with our measures of policy since speed and final level of privatization are certainly related to pre-existing level of reform. This pre-existing level is captured with our instruments because they reflect accurately the amount of distortions left by the communist policies. • Second, like the dependent variables, they should be exogenous in the growth regressions. To be sure, we ran the test of overidentifying restrictions and we found that effectively they are uncorrelated with the errors of the growth regressions.

12. Data: Initial Conditions • For measuring the institutional initial conditions we use the first principal component of the 1989 observation of the institutional variables from Campos (1999) (PCinst1). • These are an indicator that measures both quality of law enforcement and "substance" of law itself, an indicator that captures quality of local bureaucracy, an indicator that measures civil liberties, political rights and influence of civic organizations, and an indicator that captures both transparency of policy-making and accountability of the executive. • This component is strongly correlated with each of these variables and explained most of the variability of these variables.

13. Results: Ordinary Least Squares

14. Results:Two-Stage Least Squares

15. Conclusions (I) • The most striking result is that our measure of institutional strength is positively associated with growth, a result that is preserved in the two sets of estimations. This result is not surprising and is similar to what other author have found. • However, we are really the first to consider simultaneously the three standard explanations offered in the literature because we are considering policy speed and not only policy level. • We also found that initial conditions have an insignificant effect on cross sectional growth. A possible explanation is that we are using a longer time series. A second possible explanation is that the earlier literature did not control much for institutional features.

16. Conclusions (II) • The privatization speed has a negative effect on cross-country growth and the level of privatization has a marginal positive effect on growth. • When we consider a possible endogeneity of these policy variables, their sign and size of their coefficients are preserved but they loose some significance. • An explanation for these distinctive results is the earlier literature did not really consider the issue of policy speed and only looked at policy level (the only exception is Heybey and Murrell (1997)).

17. Future Line of Research (I) • First, it is important to disentangle the effect of privatization itself from related policy decisions that also influence long-term growth. • Second, privatization can influence growth not only through expected efficiency gains of private companies but also through other indirect channels such as encouraging financial market development. • The next step is to relate these micro-financial proxies with some macro measures such as cross sectional growth rate across transition countries.

18. Future Line of Research (II) • The main challenge is to find good proxies for micro-financial variables that go beyond the traditional indexes used in the literature on country cross sectional growth. • Micro-financial variables that capture, for instance, privatization method, privatized company performance, ownership structure, corporate governance measures and quality of regulation and competition policies.