This raises an identification issue for evaluating the effect of privatization on firm performance: if more profitable firms are more likely to be privatized, we may overstate the impact of privatization on profitability when we compare the performance of government-owned to that of privatized firms. The authors then proceed to use political variables as instruments for the privatization decision, adopting a two-stage least squares treatment effects regression.
After addressing the selection bias, they find that privatization still has a positive impact on performance in India.
This set of studies examines the effects of privatization on firm performance by comparing pre- and post-divestment data for companies privatized via public share offerings. Each firm is compared to itself a few years earlier using inflation-adjusted sales and income data. The first study using this methodology is by Megginson, Nash, and van Randenborgh As Megginson and Netter note, this methodology suffers from several drawbacks, among which selection bias is probably the greatest concern, since privatizations through share sales—Share Issue Privatization SIPs —represent the largest companies sold during a privatization program.
Another weakness is that the Megginson, Nash, and van Randenborgh methodology can only examine simple accounting variables assets, sales, etc. Most of the studies in this tradition also imperfectly account for macroeconomic or industry changes in the pre- and post-privatization window see Megginson and Netter , for a critique.
These studies also cannot account for the impact on privatized firms of regulatory or market-opening initiatives that are often launched in parallel with privatization programs. However, the Megginson, Nash, and van Randenborgh methodology allows the analysis of large samples of firms from different industries, countries, and time periods and, while carrying the risk of selection bias, SIP samples contain the largest and most politically important privatizations.
Most of these studies do identify a significant improvement in company performance, post-privatization, though methodological reservations remain. Research in this tradition has focused on specific industries banking [ Verbrugge, Owens, and Megginson ] and tele-communications [ D'Souza and Megginson ] ; has used data from a single country Chile [ Maquieira and Zurita ] and employed multi-industry, multinational samples.