Determinants and Effect of Firm-level Adjustment on Productivity

Abstract

This thesis examines the motivations for and impacts of different channels of adjustment on firm-level productivity. It specifically focuses on how firms systematically choose between different paths of adjustment and the impact of these choices on firm-level productivity.

The first empirical chapter of this thesis examines the determinants of firms’ choice of adjustment. Using a multinomial logit model, it considers the role of the following 4 characteristics: firm size, adjustment size, firm-level variables (R&D, age, multi-plant and foreign ownership) and other factors. The chapter shows that large firms tend to rely more on external forms of adjustment – greenfield investment and mergers and acquisition for expanding firms and; plant closure and plant sale for contracting firms - than small firms. It also shows that firms tend to rely more (less) on external forms of expansion (contraction) when the desired size of adjustment is large. With regards to the firm-level variables considered, this chapter shows that R&D is negatively related to greenfield investment with no/negligible effect on mergers and acquisition, plant closure and plant sale. Age has a negative (no) impact on greenfield investment (mergers and acquisition) and plant closure (plant sale). Multi-plant firms tend to rely more on external forms of adjustment. Lastly, we find that foreign-owned firms are more likely to acquire and close existing plants.

The second empirical chapter studies the impact of alternative forms of adjustment on firm-level productivity. This chapter uses the system GMM approach to tackle two sources of bias: simultaneity of adjustment paths-productivity relationship and endogeneity of factor inputs (and self-selection of firms in and out of an industry) in the production function. This chapter shows that there is no statistical relationship between adjustment paths and the long-run productivity of firms. However, given our choice of appropriate control groups and the fact that we use the system GMM approach to alleviate endogeneity concerns, we view our finding of no long-run adjustment effect as novel.