The Impact of Owner Characteristics & Policy on Small Firms: Using the SSBF

Transcription

1 The Impact of Owner Characteristics & Policy on Small Firms: Using the SSBF National Accounts & Economic Performance Conference May 14, 2008 Neus Herranz, Stefan Krasa and Anne Villamil University of Illinois

2 Small Firms In the US, small firms Produce more than 50% of non-farm output Employ 50% of workers Pay 45% of total private payroll They are important to the macroeconomy. There is risk but returns can be great.

3 Debate: Nature vs. Environment Do entrepreneurs have different innate characteristics? More willing to bear risk More optimistic How important is the business environment? Bankruptcy & other institutions Access to credit (liquidity constraints) Return distribution

4 Research Agenda Examine data from the Survey of Small Business Finance (SSBF) 1993, 1998, 2003 Construct a model to organize and explain the data Use HPC to solve the model and conduct policy experiments

5 Goal: Understand Small Firms & Policies that Promote Success Why do people become entrepreneurs? What policies promote or hinder entrepreneurship? Why do we observe differences in entrepreneurship across countries?

6 Survey Small Business Finance Conducted in 1987, 1993, 1998, 2003 CS sample of non-farm, non-financial, non-real estate small businesses Represent about 5 million firms: about 4000 observations Data on firms & primary owners: age gender, industry, firm type, financial info on firm & owner 6

7 SSBF Facts Small firm returns are risky All debt-equity ratios are equally likely Owners: invest significant personal net-worth More than 90% work at the firm 18-26% firms have negative equity Annual default rate is %

8 Real Firm Returns in Density SSBF 1993 µ=1.300, σ=1.575 Normal density: µ=1.300, σ= Density S&P µ=1.212, σ=0.651 Normal density: µ=1.212, σ= moment median mean standard dev. skewness kurtosis 1993 SSBF % conf. [1.08, 1.11] [1.22, 1.38] [0.95, 2.13] [2.3, 17.3] [29, 488] 1993 S&P % conf. [1.07, 1.10] [1.16, 1.28] [0.28, 1.02] [3.1, 14.6] [20, 277]

9 Net-worth Invested Entrepreneurs with positive equity in their firm: Table 2: Net-Worth Invested, 1998 SSBF % net-worth invested 20% 40% 60% 80% mean median % of entrepreneurs 52% 25% 11% 3% 27% 21% Negative Equity for Incorporated firms: * 18.4% in 1993 * 26.3% in 1998 * 23.4% in 2003 Unincorporated firms: 12.2%, 18.6%, 13.8%

10 Capital Structure: 93 & 98 Inc Cumulative probability and1998data Uniform distribution equity/assets

11 Change over time for Unincorp Cummulative Probability Incorporated Unincorporated Equity/Assets

12 Dynamic Model Describe an entrepreneur by Preferences (heterogeneous) Resources Production Technology Information Lender: resources & opportunities

13 Heterogeneous Entrepreneurs: ρ t t+1 A (scale of production, assets) ε (% self finance) 1-ε (% debt finance) c (consumption) v (repayment) realization x Assets: Ax Equity: A(x-v) Net worth outside firm: (1+r)(1-c-εA) Debt: Av

14 Problem 3 v S = max c,a,ɛ, v u(c) + βv B x Subject to: x [ ( 1 ρ (1 + r) 1 ɛa c)] df (x) ( )] 1 ρ ] x +βv S [A(x v) + (1 + r) 1 ɛa c x df (x) 0 x x df (x) + x 0 (1 δ)x df (x) + x x v df (x) = (1 ɛ)(1 + r B ) x = max { v [ 1 ( vb v S ) ] 1 1 ρ } (1 + r)(1 ɛa c), x A c + ɛa 1 (1 ɛ)a b c 0, A 0, 0 ɛ 1.

15 HPC Methods We use HPC methods to solve the model: Compute distributions: firm returns, net-worth, debt-equity, firm size. Choose model parameters that fit the data Construct the distribution of risk aversion Choose decision rules: c, A, ε, v Need HPC to compute distributions

16 Quantitative Analysis Parameter Value Comment/ Observations β 0.97 determined from r and r B T 11 U.S. credit record δ 0.10 Boyd-Smith (1994) r B 1.2% real rate, 6 mo T-Bill, r 4.5% real rate, 30 year mortgage, f(x) SSBF 1993 (Appendix D) Remaining parameters: µ and σ: N(µ,σ) distribution of risk aversion ρ b: borrowing constraint

17 Quantitative Analysis Parameter Value Comment/ Observations β 0.97 determined from r and r B T 11 U.S. credit record δ 0.10 Boyd-Smith (1994) r B 1.2% real rate, 6 mo T-Bill, r 4.5% real rate, 30 year mortgage, f(x) SSBF 1993 (Appendix D) Remaining parameters: µ: 1.55 σ: 0.83 b: 21.5% 75% of entrepreneurs have risk aversion between 1 and 3.

18 Model Predictions Cumulative probability data model % net worth invested Cumulative probability data model Assets as % of net worth outside firm

19 Model Predictions Cumulative probability model and1998data equity/assets Cumulative probability model equity/assets

20 Model Predictions Parameter Interpretation Value Data median A% median firm assets default % firm default rate cons. % c as % of net worth neg. Eq. % neg. equity in firm

21 Policy Experiment: T Welfare Effects risk aversion ρ T = T = T = T = T = T = 11 T = T = T = T = T = T =

22 Policy Experiments: T Change in Loan Interest Rate risk aversion ρ T = T = T = T = T = T = T = T = T = T = T = T =

23 Experiment: µ µ fit median A % default % cons. % neg Eq. % % women owned businesses: 16% (1993), 24% (1998) Median Asset level normalized by net worth outside firm: 53% (men) 39% (women) Negative equity (1993) 14.8%(men) 19.5% (women) (1998) 19.4%(men) 26.1% (women)

24 Optimism 10% additional return T µ σ b % fit median A % default % cons. % neg Eq. %

25 Conclusion Model & data help us understand owner behavior, financial & legal structure, default, output, welfare SSBF data are essential for questions & measurement (firms & owners) Institutions matter & interact with modest heterogeneity in owner characteristics Bankruptcy makes entrepreneurs more willing to bear risk (80% with ρ between 0.74 and 3) Negative equity: Option value of continuing to operate the firm. Big welfare effects with production