Job Market Paper

The Effect of Centrally Bargained Wages on Firm Growth


I study how firms adapt to exogenous changes in labor costs. To identify this effect, I exploit that industries in Sweden have a centrally bargained minimum increase in average wages each year. Since these are bargained on the industry level, they are detached from the growth prospects of individual firms. Data on centralized wage increases for the years from 2001 to 2006 are combined with administrative panel data on firms. I then compare the growth rates of firms in related industries that face different centralized wage increases, while controlling for differences in growth trends. Higher centralized wage increases cause the average firm to increase sales and employment growth. These effects come from increased investments and hiring. However, there is no effect on profitability. In addition, firms increase their share of skilled labor. Moreover, the effects on employment are larger for large firms. This suggests that higher centralized wage increases also cause a reallocation of labor.

Working Papers

Financial Constraints and Insurance Demand in Small and Medium-Sized Enterprises

(with Oliver Engist, Gustav Martinsson and Christian Thomann)

We study the impact of financing constraints on the insurance demand of small and medium sized firms. We build a panel data set that combines data on insurance purchases of 34,000 firms from a large Swedish insurer with administrative data. We measure financing constraints using credit scores commonly used by banks. We find that firms facing greater constraints purchase more insurance. The results are sustained when we control for net operating losses, collateralized loans and credit lines. A regression-discontinuity design supports that financing constraints have a causal effect on firms’ demand for insurance.

The Effect of Financial Constraints on Inventory Holdings


Standard finance theory predicts that more financially constrained firms hold larger inventories. I test this prediction using Swedish administrative data and data on credit scores. For identification, I exploit a discontinuity in the translation from continuous to discrete credit scores among Swedish firms. Firms that have a better credit score hold more inventories, both in the cross-section and around the cutoff. However, the regression discontinuity estimate is zero once we scale inventories by assets or sales. This suggests that inventories are determined by operational concerns rather than by financial concerns.

Non-Business Risk and Firm Growth

(with Oliver Engist and Christian Thomann)

The literature on recovery and creative destruction has analyzed the effects of large events like natural hazards or financial crises. However, large shocks induce significant general equilibrium effects that influence the recovery process of each individual firm. We complement the literature by studying how firms adjust to idiosyncratic shocks in an economy that is in equilibrium. We use a unique data set on the insurance purchases and insurance claims of 45,000 Swedish firms. Because insurance claims are unexpected and exogenous to productivity, we can identify the effect of shocks to physical assets. We find that firms, that experience a shock that leads to an insurance claim, quickly re-invest in their physical assets. Over a one-year time horizon this is accompanied by a temporary slowdown of the employment growth and a 50\% reduction in output growth. This effect is driven almost exclusively by firms with worse access to credit, whereas firms with the highest credit score experience no reduction in growth.

Spillover Effects of Insurance Claims: Evidence From Sweden

(with Oliver Engist)

Many of the consequential risks individuals or firms face have very low probabilities of occurring. It is therefore difficult to accurately judge the risks one is exposed to from experience alone. This leads agents to resort to simple heuristics, such as weighting recent experiences more heavily when assessing risk. Using a unique data set on insurance purchases and insurance claims of 40,000 Swedish firms, we analyze how firms react when a similar firm in their municipality experiences a damage to their property. Our results suggest that firms do not adjust their insurance when other firms have a loss. Moreover, highly salient events do not create significant insurance demand either. This stands in contrast to previous studies, which find that agents increase their insurance after observing natural hazards. Our results are consistent with a theory where firms adequately judge their risk exposure and purchase insurance accordingly.

Works in Progress

The Effects of Neighbourhoods on Income, Education, and Health

(with Viking Waldén)

Human Capital in Pre-Industrial Times: Evidence from Swedish Church Records

(with Erik Lindqvist and Björn Wallace)