Working Paper Series no. 641: Adjustment Costs and Factor Demand: New Evidence From Firms’ Real Estate

Non-technical summary

In this study, we explore the effects of the costs associated with change in the volume of firms’ premises on factors demand, notably on the firm-level dynamics of the workforce size. More generally, this research contributes to the question regarding the role of adjustment costs on the allocation of production factors by documenting the effect of a quantifiable cost that varies across time and across firms.

As a firm grows, it will eventually be constrained by the size of its premises. Conversely, if it declines, the firm will operate in suboptimal oversized premises. When access to adjacent lands or buildings is constrained, or if sublease of unused premises is not possible, adjustments of the premises are likely to be realized through local relocations of activities. Those local relocations are costly and the associated cost varies across firms notably because of the tax on realized real estate capital gains. For real estate owners, these relocations indeed require the sale of previously occupied premises and the payment of the tax. It entails firm-level variations in the adjustment costs that are empirically exploited in this study. As a result, we expect to see more frequent relocations for firms that are more dynamic in terms of employment; this is indeed confirmed in Figure below.

To guide the interpretation of our results, we build a general equilibrium model with heterogeneous firms from which we derive predictions on the effect of the level of the fixed costs associated with real estate adjustment on the workforce growth at different level of productivity shocks. In this framework, profit-maximizing firms are heterogeneous with respect to their productivity level and make decisions on labor and real estate inputs in a context of adjustable real estate inputs conditional on paying a fixed cost. The level of these adjustment costs deter some firms from optimally adjusting their real estate inputs to the new productivity level and the complementarity between real estate and labor in the production process implies that it also affects firms' labor demand. We derive the existence of an interval of inaction for the difference in size between the optimal premises and the occupied ones in which firms do not adjust their real estate.

Other consequences of the model comes from the complementarity between real estate and labor which leads these firms to restrain employment growth as compared to their optimal employment growth they would observe had those firms adjusted their real estate. We show that the interval of inaction widens with the level of the adjustment costs and so do the number of firms affected. We also show that the non-relocating interval is not centered in zero and that a rise in the fixed adjustment costs has asymmetrical effects on the bounds of the interval. The effect of such a rise is larger in absolute value on the positive bound of the interval than the effect on the negative bound which implies that an increase in the fixed adjustment cost has an overall negative impact on mean firm-level employment growth. This asymmetry property strengthens with the level of the adjustment cost.

We test these predictions using microdata on French single-establishment firms from 1994 to 2013. We measure the level of the fixed costs associated with relocation by using the tax on capital gains. Different regression models confirm the theoretical predictions, namely that:

(i)        Relocation is associated with an increase (respectively decrease) in employment growth for firms that are growing (respectively declining).

(ii)       The larger the (latent) fixed costs, the less likely a firm is to relocate.

(iii)      The larger the (latent) fixed costs, the less dynamics the employment.

(iv)      Overall, employment is negatively affected by the size of the tax on capital gain.

Our baseline results suggest that a reduction of the relocation costs, through a decrease of the share of the real-estate market value that would be paid as a tax on capital gains of 1 standard deviation, would increase the propensity to relocate of affected firms by 10% and would raise the yearly employment growth rate of the growing firms by 5%. Besides, by reducing job reallocation from the least productive firms to the more productive ones, such fixed adjustment costs generate misallocation and affect global output.