Political Selection of Firms into Privatization Programs. Evidence from Romanian Comprehensive Data

Exploiting a unique institutional feature of the early Romanian privatization setup, when a group of firms was explicitly barred from any privatization, we test how politicians select firms into privatization. Using comprehensive data that includes all firms inherited from socialism, we estimate the relation between pre-privatization firm characteristics - the information known to politicians at the time of decision making - and the effect of privatization on employment, efficiency and wages. With the estimated coefficients we simulate the effect of privatization on non-privatizable and privatizable firms separately, including in the latter group both actually privatized and not privatized enterprises. The simulations show that politicians expected privatization to increase the employment of the privatizable group by 7 - 10 percent, and to decrease it in the non-privatizable group by 10 - 30 percent, depending on the first-stage estimation method, OLS or matching combined with OLS. We do not find such discrepancies in the expected change in firm efficiency; the simulated efficiency effect of privatization is large and positive for both groups of firms, and it 52 - 65 percent for non-privatizable, and 41 - 43 percent for the privatizable firms. The analysis does not support the hypothesis that wages played an important role in privatization decisions. Our study suggests that employment concerns played the key role in selecting firms for privatization, even if efficiency gains had to be sacrificed.