Thoughts on "Firm Leverage and Unemployment During the Great Recession"

There has been several articles, papers and books I have been reading during the last week. The one I will write about today is an article written by Xavier Giroud and Holger Mueller entitled Firm Leverage and Unemployment During the Great Recession. (For those interested in reading the paper, I cannot seem to find the PDF link online, but you can perhaps contact the two professors for a copy of the paper.) Both are professors of Finance and affiliated with the NBER (National Bureau of Economic Research) and the CEPR (Center For Economic Policy Research) , one at the MIT Sloan School of Management and the other at the NYU Leonard N. Stern School of Business. The paper is interesting in that it listed an interesting conclusion in that there is a strong correlation between high-leverage firms and job losses in response to the household demand shocks during the Great Recession. This is an interesting observation that both professors were able to make, but with any article there are significant gaps in which the data was compiled. Due to certain lack of data, I think there needs to be significant improvements to potentially the next paper that they can gather. 

The 3 potential improvements that the two esteemed gentlemen can make are as follows:

1. Utilize more wage/labor data that is available, which could perhaps include statistics by the Bureau of Labor Statistics. This could be utilized perhaps in determining relevant wage levels of certain households within the two sets of data between the change in both high-leverage (ΔLev 02-06 > Median) and low-leverage (ΔLev 02-06 < Median) firms between the years 2002 and 2006. By utilizing wage and labor data, you can also factor the significance that the drop in employment that are caused by both high-leverage and low-leverage firms especially in the realm of (ΔLog Emp) 07-09, which could possibly confirm the conclusions, especially under the Alternative Hypothesis: Growth, Productivity, Wages (Tables 8, 9, 10) sections of the data tables that the paper had presented. With this wage and labor data, there could be potentially another set of information the user could interpret about the changes that are caused between the two time intervals of which the data is measuring.

2. Another striking improvement that I could potentially see is a change in the dependent variables on the regression model that the two professors have used to compute their data. One possible way is to shift the focus towards labor/wages versus unemployment. With this new information, they could generate another new set of information, which could possibly further validate their results that are written under Firm Leverage and Unemployment (Table 2), Instrumental Variable IV Estimation (Table 3) and Industry Sectors (Table 4). With the change in certain dependent variables, they could see a change in the data results in Establishment Closures (Table 5), in Firm-Level Analysis (Table 6), Within-Firm Spillovers (Table 7), Alternative Hypothesis: Growth (Table 8), Alternative Hypothesis: Productivity (Table 9), Alternative Hypothesis: Wages (Table 10) and in the County-Level Analysis (Table 11). The main piece of data that could be changed is how they measured the two corresponding variables: (ΔLog (Emp) 07-09) and (ΔLog (HP) 06-09). By introducing a dependent variable into the design, they can add another dimension in their measurements in order to incorporate better analysis at the end of the day.

3. The last improvement that they could incorporate is by organizing the time and the date each locale  had a significant drop in a better way that represents the drop in employment versus the drop in housing better. Perhaps by compiling a couple of new data tables with different time ranges than just the standard ranges that the data used (02-06, 07-09), we can perhaps see where the most striking changes are and what pertinent macroeconomic results are corresponded in the findings. 

Overall, the paper is a fantastic read for those who are just getting into reading published papers on financial economics, but also an interesting find in that unemployment is correlated heavily to firms that tightened their debt capacity in the run-up in response to household demand shocks than to certain firms that freed their debt capacity. Maybe the authors weren't exactly correct in their predictions and percolations on potential ramifications to macroeconomic theory, but they found an interesting correlation in the data from their current research.