“…So, while labor market outcomes are measured in 1996, 2000, 2004, and 2008, data on institutions and policies come from the years 1995, 1999, 2003, and 2007. Similar to most other studies, we control for omitted factors (including unobserved characteristics of countries) by using random-or fixed-effects specifications of our regression 15 Change in inflation is the key control variable introduced in Nickell (1997) to account for the deviation of the unemployment rate from its natural level and is used in most subsequent studies. As there are concerns about the appropriateness of this measure in the transition context (Cazes 2002), our baseline specification includes a measure of a recent change in GDP, which aims to better account for macroeconomic shocks to which transition economies were still prone to even after the initial recession of the late 1980s -early 1990s. 16 We also have considered several additional control variables, such as proxies for the enforcement of institutions, which is likely to be sub-optimal in the countries studied.…”