Interactions between productivity and turnover at Ethiopian manufacturing firms

The industrial sector share of GDP almost remained stagnant in many less developed economies for many decades. The liberalization and structural adjustment reforms that were expected to create necessary preconditions for a substantial industrial development of these economies did not succeed in addressing this problem. The research that followed, on the manufacturing sector and its productivity, was very valuable in shedding light on the problems of the sector but also had data and model selection related problems. These problems include limited study of the exit behavior of firms.

The objective of this thesis is to examine the turnover behavior of firms and underlying factors that determine the turnover patterns. Our main findings include: First, there is a high turnover rate of firms (20.2 percent) in the manufacturing sector. This is fuelled primarily by the higher entry rate of firms. On average the firms that are the main actors in this turnover activity are the smaller firms in both the private and public sectors. Second, the mortality rate of firms in Ethiopian manufacturing is very high. About 60 percent of all entrant firms throughout the panel years have exited in the range of 5 years after they enter the market.  Third, the productivity of continuing firms on average is larger than the productivity of new entrants and old firms that exit. In addition, on average the entrant firms are more productive than firms that exit. Fourth, our main finding of exit decision shows that productivity strongly predicts exit. This shows the presence of the market selection mechanism, which differentiates productive firms that continue in the market from less productive firms that exit and works in both the private and public sectors. Finally, our analysis also show that turnover effect, in both its share effect and net entry forms, plays a dominant role as a source of aggregate productivity growth in Ethiopian manufacturing.   

This thesis has a modest contribution to the empirical literature on the topic by using panel data rather than survey data that most studies use, and by aiming to test the behavior of market in handling not only the manufacturing sector in aggregate but also the public sector. The study also uses the Stochastic frontier model, to estimate production functions and measure total factor productivity, that accounts for the technical inefficiency of firms contrary to averaging models other studies use that make behavioral assumptions which imply firms as efficient.