The effect of employment protection on firm behavior and performance

Abstract

Stringent employment protection legislation makes existing employment contracts harder to terminate but its welfare effects are ambiguous if it simultaneously reduces productivity or induces firms to offer only fixed-term jobs. In this study, we examine how the reform in Cooperation Law in 2008 affected firm behavior and performance using Statistics Finland micro data and a novel administrative dataset on layoff negotiations. The Cooperation Law is composed of information-sharing mandates and a special employment protection institution that requires firms to hold layoff negotiations before they can furlough or dismiss workers. The theories of employment protection suggest that limiting the firm's choice set can never improve performance, whereas information-sharing theories predict that introducing employees into the firm’s decision-making process leads to more efficient use of information and increases productivity. The reform was applied to firms with at least 20 employees but left firms with less than 20 workers unaffected, creating an ideal setting for difference-in-differences design. Sectoral variation in collective bargaining agreements allows us to decompose the total effects of the reform to employment protection and information-sharing components.

We find no evidence that the reform affected employment flows or induced firms to substitute furloughs for dismissals but a significant reduction in the average tenure lengths suggests that some firms responded to the reform by increasing the use of fixed-term contracts. We find null effects on firm survival and profitability but our preferred estimates report that the reform increased labor productivity by 1.6%. We decompose the total productivity effect to individual effects of employment protection and information-sharing exploiting that fact that construction firms are partially exempt from the negotiation mandate. This decomposition suggests that the effect of employment protection on productivity may lie in the range from -4.8% to -1.0%, whereas the effect of information-sharing could be in the range from 2.0% to 5.2%. Both effects are even more pronounced after adjusting for take-up rates but the external validity of our findings remains uncertain as the reform overlaps with the Global Financial Crisis.