Download to read offline
Uploaded byStructuralpolicyanalysis
Dynamism Diminished: The Role of Housing Markets and Credit Conditions
The document analyzes the effects of housing prices and credit supply on young firm activity using panel data at the metropolitan statistical area (MSA) level from 1981-2014. The key findings are: 1) Using an instrumental variables approach, the study finds large effects of local house price changes on local young firm employment growth and shares. 2) A separate, smaller role is found for locally exogenous shifts in bank lending supply on young firm activity. 3) Housing market fluctuations play a major role in driving medium-run fluctuations in young firm employment shares by acting as a transmission channel and driving force in recent decades according to the analysis.
More Related Content
Similar to Dynamism Diminished: The Role of Housing Markets and Credit Conditions
Dynamism Diminished: The Role of Housing Markets and Credit Conditions
- 1. Dynamism Diminished: TheRole of Housing Markets and Credit Conditions Steven J. Davis Research with John Haltiwanger OECD Global Productivity Forum Sydney, 20-21 June 2019
- 2. The Great Recessionand its aftermath saw the worst relative performance of young firms in at least 35 years. More broadly, as we show, young-firm activity shares move strongly with local economic conditions and local house price growth. In this light, we assess the effects of housing prices and credit supply on young-firm activity. Our panel IV estimation on MSA-level data yields large effects of local house price changes on local young-firm employment growth and employment shares and a separate, smaller role for locally exogenous shifts in bank lending supply. A novel test shows that house-price effects work through wealth, liquidity and collateral effects on the propensity to start new firms and expand young ones. Aggregating local effects to the national level, housing market ups and downs play a major role – as transmission channel and driving force – in medium-run fluctuations in young-firm employment shares in recent decades. The great housing bust after 2006 largely drove the cyclical collapse of young-firm activity during the Great Recession, reinforced by a contraction in bank loan supply. As we also show, when the young-firm activity share falls (rises), local employment shifts strongly away from (towards) younger and less-educated workers. Abstract
- 3.
- 4. 4 Share of Employeesin Young Firms, 1981-2014, U.S. Nonfarm Private Sector Source: Annual Rates, Business Dynamic Statistics (BDS) Employment in firms less than five years old fell from about 18% of private sector employment in 1981 and 1987-1988 to 9% in 2014. “Young” means < 60 months since the firm’s first paid employee as of March in the indicated calendar year. “Firm age” is set to age of its oldest establishment when the firm first becomes a legal entity, and increments by 1 each year thereafter. Establishment age is the number of years since operations began at the location in the same narrowly defined industry. 6 8 10 12 14 16 18 20 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Percent
- 5. 5 The Young-Firm Shareof Employment Exhibits Pronounced and Time Varying Cyclicality For each expansion and contraction episode, the chart shows annualized deviations from the overall mean, which equals -2.2 log points per year. (NOTE: Timing convention is that change in year t represents the change from March t-1 to March t). -5 -4 -3 -2 -1 0 1 2 3 LogPoints
- 6. Young Firm Outcomesat the State-Year Level, Using Data from 1981 to 2014 6
- 7. 7 Simple bivariate relationshipsshow great variation across states and time. Log Difference in Young Firm Employment Share, Unemployment Rates and Housing Prices in State-by-Year Data from 1981 to 2014 Employment share of young firms declines when local economic conditions deteriorate (as measured by local unemployment rate). Employment share of young firms rises when local housing prices increase -50 0 50 -4 -2 0 2 4 6 Change in Unemployment Fitted 1980-83 1983-90 1990-91 1991-01 2001-03 2003-08 2008-10 2010-14 Slope = -1.8153, SE = 0.1277 -50 0 50 LogDiffofYoungEmpShare -40 -20 0 20 40 Growth in Real Housing Price Fitted 1980-83 1983-90 1990-91 1991-01 2001-03 2003-08 2008-10 2010-14 Slope = 0.3021, SE = 0.0239
- 8. Our Main DataSources • Business Dynamic Statistics (BDS): Annual activity for firms with paid employees, with tabulations by firm age at State and MSA level. • Quarterly Workforce Indicators (QWI): Similar to BDS in key respects, but includes tabs by age-industry-MSA (including by gender-worker age-education). Covers shorter time period than BDS. • Local area unemployment rates (LAUS): This BLS program uses CPS data, UI claims, CES data and other sources to estimate local unemployment rates. • Local house price measures (FHFA): Federal Home Loan Finance Agency data, available monthly at MSA and State Level. • Saiz MSA housing supply elasticity: To instrument for house price changes • Local bank loan supply (CRA): Community Reinvestment Act data. Banks with assets >1 billion report # and volume of loans to businesses with <$1 million in gross revenue. We use to construct local “small business” loan supply shocks. • Quarterly Census of Employment and Wages (QCEW): Employment at MSA-industry level. To construct Bartik-like instruments for local cycle variables. 8
- 9.
- 10. Identification • Instrument localhouse price changes to identify effects on local young-firm activity shares. Two IV approaches that rely on distinct variation: 1. IV(1) Exploit national housing boom & bust episodes that affect MSA-level house prices differently due to differences in local housing supply elasticities • Follows Mian and Sufi, except we stack boom and bust changes and consider different outcomes. Use Saiz housing supply elasticity instruments. • Stacking permits controlling for MSA specific trends. 2. IV(2) Exploit local area demand shifts interacted with local housing supply elasticities to instrument for local house price changes. 3. We also include additional local controls – omitted variable bias may yield violation of exclusion restriction even using Saiz instruments. • Supplement with local “small” business loan shocks by adapting approach of Greenstone, Mas and Ngyugen (2015). • Fits well with IV approach 2 above. • A shorter times-series dimension, because CRA data are only available from late 1990s.10
- 11. IV Approach (1):Stacked Boom/Bust Episodes 𝑌 𝑚𝑠 = σ 𝑠 𝜆 𝑠 𝐼𝑠 + σ 𝑚 𝜆 𝑚 𝐼 𝑚 + 𝛽𝐻𝑃𝑚𝑠 + 𝜀 𝑚𝑠 (1) Second Stage 𝐻𝑃𝑚𝑠 = σ 𝑚 𝛿 𝑚 𝐼 𝑚 + σ 𝑠 𝛿𝑠 𝐼𝑠 + σ 𝑠 𝑍 𝑚 𝐼𝑠 𝛾𝑠 + 𝜂 𝑚𝑠 (2) First Stage 𝑌 𝑚𝑠 = log change in MSA young-firm employment share 𝐻𝑃𝑚𝑠 = log change in MSA house price index 𝐼𝑠 is dummy for period s, and 𝐼 𝑚 is dummy for MSA m 𝑍 𝑚 is cubic in Saiz housing supply elasticity 𝜆𝑖 and 𝛿𝑖 are coefficients on dummy variables Exclusion restriction: 𝐶𝑜𝑣 𝑍 𝑚 𝐼𝑠, 𝜀 𝑚𝑠 = 0. That is, 𝑍 𝑚 𝐼𝑠 influences young-firm share only through house price growth, conditional on period and MSA effects in the first equation. 466 observations: 233 Boom changes (2002-2006) + 233 Bust changes (2007-2010). Stacking boom and bust episodes lets us control for arbitrary differences in MSA-level trends in the 2000s, addressing concerns that these trends are correlated with MSA-level housing supply elasticities, as argued by Davidoff (2015). 11
- 12. 12 Bivariate MSA- Level Relationship ofHouse Price Growth to Log Change in Young- Firm Share in Boom, Bust and Other Periods All three panels show annualized log changes
- 13. 13 (1) (2) (3)(4) OLS (Boom/Bust) IV (Boom/Bust) OLS (Boom/Bust) IV (Boom/Bust) Growth in real housing price 0.171*** (0.040) 0.190*** (0.070) 0.184*** (0.049) 0.194*** (0.057) F-Test for Excluded Instruments 31.4 35.3 Period Effects Yes Yes Yes Yes MSA Effects No No Yes Yes R2 0.247 0.247 0.515 0.515 Observations 466 466 466 466 Response of Log Difference in Young-Firm Share to Housing Price Growth, IV Approach (1): Stacked Boom/Bust Notes: Boom (2002-06), Bust (2007-10). Instruments are period effects interacted with cubic in (log) Saiz elasticity. Standard errors in parentheses clustered at MSA level. 466 observations in each specification. Controlling for Differential MSA- Level Trends
- 14. 14 The specification includes periodcontrols. See Column (2) in the table for standard errors that are adjusted for the two-stage nature of the estimation. Second-Stage Relationship between Log Change in Young-Firm Share and House Price Growth – Column 2 in the Previous Table Bust Boom
- 15. IV Approach (1):On Identification 1. Measurement Error: IV addresses concerns that OLS yields a (downwardly) biased estimate of 𝛽 due to measurement error in HP – a serious concern, in our view, given the difficulties of constructing good house price indices. 2. Reverse causality: Not a serious concern in our view, given that our lhs variable is log change in the young-firm employment share. In other words, we do not think exogenous shocks to the local young-firm share drive changes in local house price growth. The small size of the young-firm share also limits concerns about reverse causality. 3. Omitted Variables: Our specifications might not adequately control for local cycle conditions that affect local house price growth and 𝑌 𝑚𝑠 -- a serious concern in our view. 15
- 16. IV Approach (1):On Identification Omitted variables can cause the 𝐶𝑜𝑣 𝑍 𝑚 𝐼𝑠, 𝜀 𝑚𝑠 = 0 assumption to fail. To see how, suppose the true specification is given by: 𝑌 𝑚𝑠 = σ 𝑠 𝜆 𝑠 𝐼𝑠 + σ 𝑚 𝜆 𝑚 𝐼 𝑚 + 𝛽𝐻𝑃𝑚𝑠 + 𝛼𝑋 𝑚𝑠 +𝜀 𝑚𝑠 (1)’ 𝐻𝑃𝑚𝑠 = σ 𝑚 𝛿 𝑚 𝐼 𝑚 + σ 𝑠 𝛿𝑠 𝐼𝑠 + σ 𝑠 𝑍 𝑚 𝐼𝑠 𝛾𝑠 + 𝜃𝑋 𝑚𝑠 + 𝜂 𝑚𝑠 (2)’ Where 𝑋 𝑚𝑠 is a local shock that affects the young-firm activity share in (1)’. Such local shocks may also affect local house prices in (2)’. Suppose further that 𝐶𝑜𝑣 𝑍 𝑚 𝐼𝑠, 𝑋 𝑚𝑠 ≠ 0, i.e., that the local shock 𝑋 𝑚𝑠 is also correlated with our instrument. In this case, estimating (1) rather than (1)’ will violate the assumption that 𝐶𝑜𝑣 𝑍 𝑚 𝐼𝑠, 𝜀 𝑚𝑠 = 0. This limitation can be overcome by estimating (1)’ rather than (1) along with (2)’ rather than (2). 16
- 17. IV Approach (1):On Identification The next slide considers the sensitivity of the key coefficient of interest to the inclusion of several local shock control variables: • Local cycle control: The average annualized change in the MSA-level unemployment rate during the period. • Bartik-type local demand shifter: The average annualized employment growth rate implied by (lagged MSA-level industry employment share) X (national industry employment growth) summed over all industries at the 2- digit NAICS level. Results are similar using 4-digit NAICS data, but there is much suppression of cell-level data at 4-digit NAICS by MSA level. • Local population growth: The average annualized change in the MSA-level population during the period. 17
- 18. 18 (1) (2) (3)(4) IV IV IV IV Growth in real housing price 0.194*** (0.057) 0.174** (0.075) 0.173** (0.081) 0.161** (0.079) F-Test for Excluded Instruments 26.9 26.7 26.0 23.4 Period & MSA Effects Yes Yes Yes Yes MSA Unemp. Rate Change No Yes Yes Yes MSA Bartik Shock No No Yes Yes MSA Population Growth No No No Yes R2 0.515 0.519 0.520 0.522 Response of Log Difference in Young-Firm Share to Housing Price Growth, IV Approach (1): Stacked Boom/Bust Episodes with Additional Controls Notes: Boom (2002-06), Bust (2007-10). Instruments are period effects interacted with cubic in (log) Saiz elasticity. Standard errors in parentheses clustered at MSA level. 466 observations in each specification. Same as column (4) in previous table
- 19. IV(2) Approach: OnIdentification 𝑌 𝑚𝑡 = 𝑡 𝜆 𝑡 𝐼𝑡 + 𝑚 𝜆 𝑚 𝐼 𝑚 + 𝛽𝐻𝑃 𝑚𝑡 + 𝛼𝐶𝑌𝐶 𝑚𝑡 + 𝑋 𝑚𝑡 ′ Α +𝜀 𝑚𝑡 𝐻𝑃 𝑚𝑡 = σ 𝑚 𝛿 𝑚 𝐼 𝑚 + σ 𝑠 𝛿𝑡 𝐼𝑡 + 𝐶𝑌𝐶 𝑚𝑡 𝑍 𝑚 ′ Γ + 𝛼𝐶𝑌𝐶 𝑚𝑡 + 𝑋 𝑚𝑡 ′ Β + 𝜂 𝑚𝑡 𝑌 𝑚𝑡 = log change in MSA young-firm employment share 𝐻𝑃 𝑚𝑡 = log change in MSA house price index 𝐼𝑡 is dummy for period t, and 𝐼 𝑚 is dummy for MSA m 𝑍 𝑚 is cubic in Saiz housing supply elasticity 𝑋 𝑚𝑡 is a vector of local controls (e.g., Bartik local demand shifter and population controls). 𝐶𝑜𝑣(𝐶𝑌𝐶 𝑚𝑡 𝑍 𝑚, 𝜀 𝑚𝑡) = 0); i.e., the interaction between local cycle and local supply elasticity affects 𝑌 𝑚𝑡 only through its effect on local house price growth, 𝐻𝑃 𝑚𝑡, conditional on controls. IV addresses measurement error and endogeneity in the house price index. MSA effects and local controls address unobserved MSA trends and omitted variable bias (related to concerns by Davidoff (2015). 19
- 20. 20 IV Estimates for1992-2014 Sample, IV Approach (2) Dependent Variable: Log Change in Young-Firm Employment Share, MSA by Year Data Notes: Standard errors in parentheses clustered at MSA level. All specs include the change in the MSA unemployment rate. Specs without year effects include a quadratic in National GDP Growth. For IV estimates, overidentification tests show we cannot reject the null of instrument validity. 5322 observations. * p < 0.1, ** p < 0.05, *** p < 0.01. OLS IV2 OLS IV2 IV2 Growth in real housing price 0.181*** 0.384*** 0.092*** 0.285** 0.300** (0.022) (0.127) (0.027) (0.132) (0.149) F-test for Excl. Instruments 45.3 47.1 41.4 MSA Effects Yes Yes Yes Yes Yes Year Effects No No Yes Yes Yes MSA Bartik Shock No No No No Yes MSA Population Growth No No No No Yes
- 21. Extend Specification (from1999 on) using Local “Small” Business Bank Loan Supply Shocks (GMN (2015) 21
- 22. 22 ESTIMATES FOR 1999-2014 Sample: IVestimates yield positive and statistically significant impact of housing prices and small business loan supply shocks on young-firm activity shares. Using small business activity share as outcome yields much weaker effects -- e.g., no effect of Small Business Loan Supply shock. OLS OLS IV2 IV2 IV2 Growth in real housing price 0.178*** (0.022) 0.163*** (0.023) 0.297*** (0.090) 0.289*** (0.091) 0.322*** (0.102) Local Small Business Loan Supply Shock 0.030*** (0.010) 0.024** (0.011) 0.020* (0.012) F-test for Excluded Instruments 43.0 43.7 38.8 MSA Effects Bartik Controls Population Growth Yes No No Yes No No Yes No No Yes No No Yes Yes Yes Dependent Variable: Log Difference of Young Employment Share, Using MSA by Year Data Notes: Standard errors in parentheses clustered at MSA level. All specs include the Change in MSA level Unemployment Rate and a quadratic in National GDP Growth. For IV estimates, overidentification tests show we cannot reject the null of instrument validity. 3728 Observations. * p < 0.1, ** p < 0.05, *** p < 0.01.
- 23. Local Employment GrowthRate Effects 23
- 24. Exploring the ChannelsThrough Which House Price Changes Affect Young- Firm Activity Shares 24
- 25. (Local) Housing Pricesand (Local)Young-Firm Activity: Potential Transmission Channels 1. Wealth and Risk Tolerance: Home equity up greater willingness to take on risks of new/young business. (Khilstrom-Laffont, 1979, Guiso-Paiella, 2008) 2. Wealth Effect on Desire to Be Own Boss: Demand for being own boss is a normal good and increases in wealth. (Hurst and Pugsley, 2015) 3. Liquidity and Collateral Effects: Households tap home equity to relax liquidity constraints, increasing their ability to finance new/young businesses. Higher house prices greater collateral value. (Evans and Jovanovic, 1989) 4. Local Credit Supply: Local housing conditions affect local banks’ lending capacity + young firms are relatively dependent on local bank credit. 5. Local Outlook and Credit Supply: Banks see local housing prices as indicators of (future) local business conditions, affecting their willingness to lend; and new and young firms are relatively dependent on bank credit. 6. Non-uniform Consumption Expenditure Responses: Young firms supply goods and services whose demand is relatively sensitive to local income/wealth.25
- 26. Exploring Channels Novel testto distinguish between contribution of consumption demand channels and other channels (all of which operate on financial conditions for the young firms and their owners) Idea: Test whether the local industry growth rate response to local house price changes depends on the local industry’s firm-age structure of employment. If house prices work entirely through consumption demand channels, then we expect the local industry response to be invariant to its firm-age structure. The alternative view says the local industry response rises with its young-firm activity share due to wealth, collateral, and liquidity effects of house prices on the relative propensity to start a new business or expand a young business. Implement using annual QWI employment data at 2-digit NAICS by MSA level for 1999-2015. 26
- 27. Specifications for Implementingthe Invariance Test 𝐺𝑅𝑗𝑚𝑡 = 𝑎 + 𝑏1 𝐶𝑌𝐶 𝑚𝑡 + 𝑏2 𝐻𝑃 𝑚𝑡 + 𝑏3 𝑒𝑚𝑝𝑠ℎ𝑎𝑟𝑒𝑗𝑚,𝑡−1 +𝑐 ∙ 𝐻𝑃 𝑚𝑡× 𝑒𝑚𝑝𝑠ℎ𝑎𝑟𝑒𝑗𝑚,𝑡−1 + 𝑓𝑡 + 𝑓𝑚 + 𝑓𝑗 + 𝜀 𝑗𝑚𝑡 (1) 𝐺𝑅𝑗𝑚𝑡 = 𝑎 + 𝑏3 𝑒𝑚𝑝𝑠ℎ𝑎𝑟𝑒𝑗𝑚,𝑡−1 +𝑐 ∙ 𝐻𝑃 𝑚𝑡× 𝑒𝑚𝑝𝑠ℎ𝑎𝑟𝑒𝑗𝑚,𝑡−1 +𝑓 𝑚𝑡 + 𝑓𝑗 + 𝜀 𝑗𝑚𝑡 (2) 𝐺𝑅𝑗𝑚𝑡 = 𝑎 + 𝑏3 𝑒𝑚𝑝𝑠ℎ𝑎𝑟𝑒𝑗𝑚,𝑡−1 +𝑐 ∙ 𝐻𝑃 𝑚𝑡× 𝑒𝑚𝑝𝑠ℎ𝑎𝑟𝑒𝑗𝑚,𝑡−1 +𝑓 𝑚𝑡 + 𝑓𝑗𝑡 + 𝜀 𝑗𝑚𝑡 (3) where j is industry, m is MSA, and t is time. Industry classifications based on 2-digit NAICS codes. 𝐺𝑅𝑗𝑚𝑡 is log employment change from t-1 to t for industry j in MSA m, and 𝑒𝑚𝑝𝑠ℎ𝑎𝑟𝑒𝑗𝑚,𝑡−1 is the lagged young-firm share.27
- 28. Detail on IndustryClassifications for This Test 23 Construction 31-33 Manufacturing 42 Wholesale Trade 44-45 Retail Trade 48-49 Transportation and Warehousing 51 Information 52 Finance and Insurance 53 Real Estate and Rental and Leasing 54 Professional, Scientific, and Technical Services 55 Management of Companies and Enterprises 56 Administrative and Support and Waste Management and Remediation Services 62 Health Care and Social Assistance 71 Arts, Entertainment, and Recreation 72 Accommodation and Food Services 28 We omit the following industries because few MSAs have positive employment and/or QWI coverage limitations: 11 -- Agricultural Services 21 -- Mining 22 -- Utilities 61 -- Educational Services (Mostly non-profits in QWI) 81 -- Other Services (Many religious organizations and other non-profits in QWI). The parts of 61 and 81 included in the QWI also have weak relationships to cyclical variables, including housing prices.
- 29. Dependent Variable: AnnualLog Employment Change at Industry-MSA level OLS OLS OLS IV2 Chang in -0.939*** -0.750*** Unemp. Rate (0.149) (0.106) Housing Price 0.088*** 0.175*** Log Chang (0.011) (0.042) Young-Firm 0.029*** 0.031*** 0.037*** 0.031*** Emp. Share (0.010) (0.010) (0.010) (0.010) HP x Young_Sh 0.813*** 0.780*** 0.588*** 0.672*** (0.059) (0.075) (0.091) (0.118) MSA FE Yes No No Yes Year FE Yes No No Yes Industry FE Yes Yes No Yes MSA-by-year FE No Yes Yes No Ind-by-year FE No No Yes No N 39627 39627 39627 39627 R2 0.140 0.267 0.317 0.137
- 30. Quantifying Departures fromAge Invariance • Are the departures from age invariance large in magnitude? • To address this question, compute the regression-implied response differential between an MSA-Industry at the 90 percentile for the young-firm employment share and one at the 10 percentile. Evaluate at the local house price log change. 𝑅𝑒𝑠𝑝_𝐷𝑖𝑓𝑓 = Ƹ𝑐 𝑌𝑜𝑢𝑛𝑔_𝑠ℎ 𝑡 90−10 𝐻𝑃𝑡(𝑝), where • Ƹ𝑐 = coefficient on interaction term in the regression. In practice, we use Ƹ𝑐 = 0.672, the estimate from the rightmost column • 𝑌𝑜𝑢𝑛𝑔_𝑠ℎ 𝑡 90−10 = the 90-10 differential in the young-firm employment across local industries at time t • 𝐻𝑃𝑡 𝑝 = pth percentile of log change from t-1 to t in MSA-level housing prices • The lower panel on the next slide implements this calculation. We report the annual average response differential during boom and bust periods and the corresponding cumulative response differentials. 30
- 31. Dispersion in Young-FirmEmployment Shares Industry-MSA Young-Firm Share 1999- 2015 Boom Period Bust Period 90th Percentile 0.262 0.274 0.255 10th Percentile 0.049 0.063 0.048 Std Deviation 0.086 0.086 0.083 90-10 0.213 0.211 0.207 31 Dispersion in Local Log House Price Changes Log MSA House Price Change 1999- 2015 Boom Period Bust Period 90th Percentile 0.078 0.128 0.035 10th Percentile -0.062 0.005 -0.138 Std Deviation 0.066 0.053 0.082 90-10 0.134 0.123 0.173 Use Ƹ𝑐 = 0.672 Boom Period Bust Period P90 P10 P90 P10 𝐻𝑃𝑡 𝑝 (average annual log change) 0.128 0.005 0.035 -0.138 𝑌𝑜𝑢𝑛𝑔_𝑠ℎ 𝑡 90−10 0.211 0.207 𝑅𝑒𝑠𝑝_𝐷𝑖𝑓𝑓 Annual, Percentage Points 1.8 0.01 0.5 -1.9 Cumulative, Percentage Points 7.3 0.3 1.5 -5.8 Boom = 2002-2006 and Bust =2007-2010 These results show that departures from age invariance are large for SMSAs that had especially big house price gains (losses) during the national house price boom (bust). However, the magnitude of departures from age Invariance are modest for most SMSAs most of the time.
- 32. Additional Results froma Dynamic Specification Extending the specification to include the lagged main effect for local housing price changes and its interaction with the lagged young-firm employment share in the local industry: 1. The local industry response to higher local house prices rises even more steeply with the local industry’s young-firm share. 2. The effect of a local housing price increase on local industry employment growth rises in period t with the local industry’s young-firm share, and it rises even further in period t+1. 3. In terms of local industry employment levels, these results imply powerful hysteresis effects of local housing price changes that vary with the firm-age structure of employment in the local industry. 32
- 33. How Important AreHouse Price Movements and Bank Loan Supply Shifts for National Changes in Young-Firm Activity Shares? 33
- 34. Our Quantification Method Weuse IV coefficient estimates and actual state-level house price changes from 1981 to 2014 to quantify national effects of housing market developments. By using all house price changes, we capture the effects of exogenous house price changes and the role of house prices in transmitting shocks that originate elsewhere. We aggregate state-level changes to the national level using state-level employment shares. Given correctly identified causal effects of local house price changes, our quantification exercise may overstate or understate the role of housing market developments in national young-firm activity shares: • Overstate? Spatial equilibration of young-firm activity across local areas may attenuate national responses relative to the aggregated local responses. • Understate? (1) Positive spillovers of young-firm activity across local areas. (2) Entrepreneurs may own houses outside the area where they operate young firms, an effect not captured by our regression model or aggregation. 34
- 35. 35 Contribution of HousingPrice Changes to Log Changes in Young-Firm Employment Shares By Cycle Episode Based on IV2 Estimate of Coefficient on HP Solid Bar is Actual. Striped and Dotted Bars are counterfactuals implied by IV (2) approach with, respectively, controls for MSA, Year Effects and Local controls. Counterfactuals use actual state-level house price changes. Annualized deviations from overall means depicted. The mean decline is -2.2 log points per year. 61 percent of the (trend-deviated) decline in young-firm activity share in Great Recession is due to decline in housing prices, according to this exercise. -5 -4 -3 -2 -1 0 1 2 3 LogPoints
- 36. 36 Year-By-Year results showthat the housing boom attenuated the secular decline in young-firm employment shares from from 1998-2007 and accelerated the decline after 2007. IV(2) estimates. Recall that the mean change in the young firm employment share is -2.2 log points per year. -10 -8 -6 -4 -2 0 2 4 6 8 10 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 LogPoints Actual Housing Prices (IV Results) Cumulative increase from 1997-2007 from Housing Prices = 11 log points Cumulative decrease from 2008-2013 From Housing Prices = 10 log points
- 37. 37 Contribution of HousingPrice Changes and “Small” Business Bank Loan Supply Shocks to Log Changes in Young-Firm Employment Share by Cycle Episode Sold Bar is Actual, Diagonal Striped Bar is Counterfactual (Housing Prices only), Dotted Bar is Counterfactual (Loan Supply only), Horizontal Striped Bar is (Housing Prices + Loan Supply). Using IV2 estimates from column 4 of previous table. Annualized deviations from overall means depicted. The mean decline is 2.4 log points per year from 1999-2014. Taken together, the decline in housing prices and bank loan supply shocks account for 85 percent of trend-deviated decline in Young-Firm activity share in Great Recession During 2001-08 period these effects tended to boost Young-firm activity shares – working against forces leading to diminished dynamism over this period of time. -4 -3 -2 -1 0 1 2 3 LogPoints
- 38. How Do theFortunes of Young Firms Affect the Labor Market Opportunities of Younger and Less-Educated Workers? 38
- 39.
- 40. 40 1. Concentration of seculardecline in Great Recession is apparent for all groups. 2. Younger workers and less educated workers have larger declines than national average in Great Recession (2008-10). In turn, older workers and more educated workers have smaller declines in Great Recession than national average.
- 41. 41 Regressions at theMSA Level Regressions at the MSA-Industry Level Men Women Men Women Demographic Group OLS (1) Two-Stage (2) OLS (3) Two-Stage (4) Two-Stage (5) Two-Stage (6) 19-24 Years of Age 0.017 (0.010) 0.258 (0.046) 0.01 (0.009) 0.239 (0.043) 0.259 (0.052) 0.384 (0.062) 25-44 0.028 (0.010) 0.359 (0.074) 0.032 (0.009) 0.550 (0.108) 0.360 (0.076) 0.633 (0.119) 45-54 -0.028 (0.011) -0.540 (0.093) -0.031 (0.008) -0.588 (0.099) -0.539 (0.096) -0.786 (0.134) 55-64 -0.017 (0.005) -0.082 (0.027) -0.012 (0.005) -0.201 (0.036) -0.077 (0.025) -0.230 (0.041) < High School 0.020 (0.007) 0.283 (0.053) 0.021 (0.007) 0.295 (0.053) 0.289 (0.048) 0.313 (0.051) High School 0.01 (0.007) 0.110 (0.024) 0.016 (0.007) 0.151 (0.027) 0.062 (0.019) 0.089 (0.023) Some College -0.010 (0.002) -0.105 (0.021) -0.017 (0.005) -0.195 (0.031) -0.126 (0.024) -0.166 (0.026) Undergrad or More -0.022 (0.010) -0.294 (0.055) -0.020 (0.009) -0.257 (0.051) -0.225 (0.042) -0.238 (0.045) Dependent variable: One-Year Change in the group-level share of employment at the MSA Level or the MSA-Industry Level Reported coefficients are effect of change in young firm employment share. Two-stage approach isolates systematic variation in young firm employment share projected from first stage on local Cyclical shocks, housing price Growth and small business Lending shocks. This addresses measurement error from noise infusion for the QWI. All specifications include MSA Effects and aggregate cyclical Controls. Industry specs control For industry effects and industry Interacted with national cyclical controls
- 42. Summary of MainResults 1. Over the past 35 years, there has been a large secular decline in the U.S. young-firm employment share. That decline accelerates during recessions and attenuates during booms. 2. Local house price changes have large causal effects on the local employment shares of young firms. 3. These causal effects work partly through wealth, liquidity, collateral, and credit supply effects on the propensity to start a new business or expand a young one. 4. House price changes are a major driver of medium-run fluctuations in national young-firm employment shares. They account for more than half of the trend-deviated drop in young-firm activity shares during the Great Recession. 42
- 43. Summary of MainResults 5. Locally exogenous shifts in bank loan supply are also a driver of local and national changes in the young-firm employment share in certain episodes, e.g., the Great Recession. 6. When the young-firm share of local employment rises, employment shifts from older to younger workers and from more- to less-educated ones. 7. Together with our other results, Result 6 says that housing busts and credit crunches hurt younger and less-educated workers through their particular effects on the fortunes of younger firms in addition to their broader effects on the economy. 43