«Manuel Adelino, Duke University Antoinette Schoar, MIT and NBER Felipe Severino, MIT December, 2012 Abstract This paper explores the role of the ...»
In the last two columns we again split the sample by the amount of capital that is needed to start a business in a given industry as discussed above. We find that the effect of house prices on the net creation of sole proprietorships is stronger in industries with low start-up capital needs, which is in line with our findings for the other size categories. We should note, however, that the difference between the coefficients in the two specifications (below and above median capital needs) is not statistically significant.
3.4. Crisis Period (2007-2009)
One question that remains regarding the establishments that were created as a consequence of the increasing value of collateral during the run-up in house prices is whether these establishments were then eliminated once the housing bubble burst. The question is whether these were particularly 16 fragile firms that were disproportionately affected by the crisis or whether, on the other hand, they were not of different quality relative to the rest of the firms in the economy.
Given the data we currently have we cannot give a definite answer to this question. The problem is that we are not able to track individual establishments, which means we cannot know if the specific firms that were created in the 2002-2007 period survived the crisis or not. We can, however, test whether small establishments in general were more or less likely to downsize or disappear in the crisis. Put differently, we can assess whether employment loss was stronger at larger or smaller firms during the crisis in counties where the increase in house prices had been stronger in the pre-period (which are also the most levered counties as shown in Mian and Sufi, 2011a). We run those regressions in Table 6.
The results show that employment loss was similar across large and small establishments or, if anything, it seems to have been worse at large firms (in the specifications without industry fixed effects) in counties where house prices went up more. This suggests that, at least as a group, small firms were no more likely to destroy jobs as a consequence of the increased leverage accumulated during the pre-crisis period. This is consistent with the findings of Mian and Sufi (2011a) regarding the non-tradable industries for this period.
3.5. Total Employment and Migration
We finally want to consider the effect of house price changes on total employment as measured in the County Business Pattern (CBP). Columns (1) and (2) of Table 7 show county level regressions of change in Total Employment on house prices changes instrumented with the Saiz measure. Column (2) includes a number of county level controls such as population size, average educational attainment, and unemployment rate in the pre-period. We find that house price growth had no causal effect on total employment: the coefficient on house price changes is close to zero and insignificant in either of the specifications. In contrast, when we repeat the same regression set up using the level of unemployment as the dependent variable in Columns (3) and (4) of Table 6, we find a significant and negative relationship both with and without controls. Finally, in Columns (5) and (6) we show that house price changes also had a negative impact on the unemployment rate;, consistent with the results of Charles, Hurst and Notowidigdo (2012). How can the negative effect 17 on unemployment be reconciled with no changes in total employment? Our results suggest that the decrease in unemployment captures the transition of some agents in the labor force from being job seekers to a self-employment status. However, these people are not observed in the total employment measure, since the CBP data does not include non-employee firms (sole proprietorships).
Finally we also look at the net migration of people in and out of the county. We measure net migration as the difference between inflows and outflows at the county level. We repeat the same regression set up in Column (7) and (8) to estimate the effect of house appreciation on county to county migration and find that higher house prices caused a net out-migration from the counties with high house price appreciation. In unreported regressions we confirm that this was produced by larger outflows than inflows into those counties. This evidence is consistent with the idea that house prices affected the composition of households in each county and, therefore, indirectly affected the labor market dynamics.
Overall, the evidence we present in this paper identifies the causal effect house prices in the creation of new small firms. These results show that access to collateral allowed individuals to start small businesses or to become self-employed. We conjecture that without access to this collateral in the form of real estate assets, many individuals would not have made the transition from unemployment to starting a new business or self-employment.
We show that the effect of house prices is concentrated in small firms only and had no causal effect on employment at large firms. Importantly, our results also hold when we exclude industries that are most likely to be affected by local demand shocks and when we restrict our attention to manufacturing industries. The effect of house prices is also stronger in industries where the amount of capital needed to start a new firm is lower, consistent with the hypothesis that housing serves as collateral but is not sufficient to fund large capital needs.
Adelino, Manuel, Antoinette Schoar, and Felipe Severino. 2012. “Credit Supply and House Prices:
Evidence from Mortgage Market Segmentation.” NBER Working Paper n. 17832.
Black, Jane, David De Meza, and David Jeffreys. 1996. “House prices, the supply of collateral and the enterprise economy.” The Economic Journal (1996): 60-75.
Berger, Allen, and Gregory F Udell. 1998. “The economics of small business finance: The roles of private equity and debt markets in the financial growth cycle.” Journal of Banking and Finance, 22, 6: 613-673.
Bernanke, Ben S. and Mark Gertler. 1989. “Agency costs, net worth, and business fluctuations.” American Economic Review, 79: 14–31.
Cagetti, Marco and Mariacristina De Nardi. 2006. “Entrepreneurship, Frictions, and Wealth.” Journal of Political Economy, vol. 114, no. 5.
Chaney, Thomas, David Sraer, and David Thesmar. 2012. "The Collateral Channel: How Real Estate Shocks Affect Corporate Investment." American Economic Review, 102(6): 2381-2409.
Charles, Kerwin Kofi, Erik Hurst and Matthew J. Notowidigdo. 2012. “Manufacturing Busts, Housing Booms, and Declining Employment: A Structural Explanation.” Working Paper.
Evans, David S. and Boyan Jovanovic. 1989. “An Estimated Model of Entrepreneurial Choice under Liquidity Constraints.” Journal of Political Economy, 97:4, pages 808-27.
Fan, Wei and Michelle J. White. 2003. “Personal Bankruptcy and the Level of Entrepreneurial Activity.” Journal of Law and Economics, Vol. 46.
Fairlie, Robert W., and Harry A. Krashinsky. 2012. “Liquidity constraints, Household Wealth, and Entrepreneurship Revisited.” Review of Income and Wealth. 59: 279-306.
Gentry W.M, and R. Glenn Hubbard. 2005. “Success Taxes, Entrepreneurial Entry, and Innovation.” Innovation Policy and the Economy, Volume 5, 87-108. The MIT Press.
Gertler, Mark and Simon Gilchrist. 1994. “Monetary Policy, Business Cycles, and the Behavior of Small Manufacturing Firms.” The Quarterly Journal of Economics, MIT Press, vol. 109(2), pages 309-40.
Holtz-Eakin, Douglas, David Joulfaian and Harvey S. Rosen. 1994. “Entrepreneurial Decisions and Liquidity Constraints.” NBER Working Paper No. 4526.
Hurst, Erik and Annamaria Lusardi. 2004. “Liquidity Constraints, Household Wealth, and Entrepreneurship.” Journal of Political Economy, vol. 112, 2: 201-248.
Kashyap, Anil K., Jeremy C. Stein, and David W. Wilcox. 1993. “Monetary policy and credit conditions: Evidence from the composition of external finance.” American Economic Review, Volume 83, Issue 1, 1993, p. 78-98 Kiyotaki, Nobuhiro, and John Moore. 1997. “Credit cycles.” Journal of Political Economy 105, no.
Kocherlakota, Narayana. 2010. “Inside the FOMC.” President’s Speech. Marquette, MI, August 17, 2010.
Mian, Atif and Amir Sufi. 2011a. “What Explains High Unemployment? The Aggregate Demand Channel” NBER Working Paper.
Mian, Atif and Amir Sufi. 2011b. “House Prices, Home Equity Based Borrowing, and the U.S.
Household Leverage Crisis.” American Economic Review 101: 2132-2156.
Mulligan, Casey. 2011. “The expanding social safety net.” NBER Working Paper n. 17654.
Petersen, Mitchell A., and Raghuram G. Rajan. 1994. “The Benefits of Lending Relationships:
Evidence from Small Business Data.” The Journal of Finance, Vol. 49, No. 1, 3-37.
Robb, Alicia M. and David T. Robinson. 2012. “The Capital Structure Decisions of New Firms.” Review of Financial Studies, forthcoming.
Romer, Christina D. 2011. “Dear Ben: It’s Time for Your Volcker Moment.” New York Times, October 29, 2011.
Saiz, Albert. 2010. “The Geographic Determinants of Housing Supply,” Quarterly Journal of Economics, vol. 125(3), pages 1253-1296, August.
20 Table 1. Summary Statistics Panel A reports summary statistics for all counties in the sample in Column 1, and Columns 2 and 3 show the summary statistics for counties above and below the median elasticity of housing supply in the sample. For each variable we show the pooled average, median (italicized) and standard deviation (in parenthesis). Total Employment refers to the total number of employees in a county in thousands across all establishment sizes and industries using the County Business Patterns data as of 2002. Unemployment Rate is shown in percentage and comes from the Bureau of Labor Statistics Local Area statistics in 2002. Number of Households, in thousands, comes from the Census Bureau in 2000. Growth of Total Employment is the percentage change of Total Employment between 2002 and 2007. Growth in DTI is the percentage change in debt to income ratio between 2002 and 2007 and Growth in Income is the percentage change in income in a county during the same period. The debt to income ratio is estimated using county level household debt data from the New York Fed-Equifax and income is computed using IRS county-level information. Growth in House Prices is the percentage change in house prices between 2002 and 2007at the MSA level from the Federal Housing Finance Agency. Finally, Change in Unemployment Rate, is the change in the rate between 2002 and 2007. Panel B shows the Total Employment in 2002 in thousands, Employment Growth between 2002 and 2007 in percentage points, and the percentage of Total Employment for each establishment size for all firms, as well as split by the start-up amount of capital needed to start a firm.
21 Table 2. Employment Growth, Firm Size and House Price Appreciation The table shows two-stage least squares regressions of employment growth on house price growth instrumented with the elasticity of housing supply, indicator variables for each establishment size (not shown in the table) and interactions of house price growth with the size of establishments. All regressions are weighted by the number of households in a county as of 2000. Employment growth is the percentage change in employment between 2002 and 2007 estimated using County Business Patterns (CBP) data. Growth in House prices is the percentage change between 2002 and 2007, and each interaction is with a dummy indicator for the size of the establishment. Column 1 shows the first stage regression of the change in house prices between 2002 and 2007 on the Saiz elasticity measure. Columns 2 through 4 “All Industries” shows the results for the whole sample of firms, first the weighted least squares results, then the IV at a county level and, finally, the IV results at a county and industry level. Columns 4 through 8 show the coefficients split by the start-up capital amount (above and below the median) also at the county and at the county and industry levels. The omitted category refers to establishments with 50 or more employees. All regressions control for the natural logarithm of population, the percentage of the population with a college degree, the percentage of the labor force that is employed, the share of the population in the workforce, and the percentage of homes that are owner-occupied. Columns 4, 6 and 8 include 4-digit NAICS fixed effects. Controls are at a county level for the year 2000 and are obtained using Census Bureau Data Summary Files. Standard errors are in parenthesis and are clustered by MSA. *, **, *** indicate statistical significance at 10, 5, and 1% levels, respectively.