1/27 Bankruptcy and Investment Evidence from Changes in Marital Property Laws in the U.S. South, 1840-1850. Peter Koudijs (Stanford & NBER) Laura Salisbury (York University & NBER) November 2015
2/27 Introduction How does personal bankruptcy affect household investment? Fresh start: stimulates entrepreneurship Credit market implications: financing more costly / more likely to be rejected Trade-off: protection vs credit constraints Evaluate trade-off empirically Exogenous variation in protection
3/27 This paper U.S. South: Married Women Property Laws in the 1840s Before: upon marriage, husband (virtually) unrestricted ownership of wife s assets, full-recourse loans After: wife s estate held in trust; could not be seized by husband s creditors; only used for necessities. Both before and after: married couple considered a single legal unit, governed by husband. Law changes not retroactive Consequences Inability to use wife s assets as collateral, wife unable to contract loans in her own name Provides downside protection, world without bankruptcy code Compare couples married before/after law change Net effect on household investment in 1850 Census
4/27 Related literature Macro Micro Livshits et al. (2007), Chatterjee et al. (2007) Gropp et al. (1997), Fan and White (2003), Berkowitz and White (2004), Berger et al. (2011) and Severino et al. (2013) (Homestead) exemptions in bankruptcy: differences across states Impact of exemptions on top of general bankruptcy protection General equilibrium effects Married women s property laws in the United States Impact on women s economic activity (variation at state level): Kahn (1996): patenting; Inwood and Van Sligtenhorst (2004): property holdings; Geddes et al (2012): school attendance. Decision to introduce laws: Geddes and Lueck (2002); Doepke and Tertilt (2009).
5/27 Outline 1 Historical context 2 Model 3 Data 4 Empirical results
6/27 American South 1840s Plantation economy 2/3 families in agriculture in 1850 Census Approx. 1/4 plantation owners (Wright 2006) Financial system Well-developed; slaves and plantations used as collateral (Kilbourne 1995, 2006) Access to loans North, trade-credit UK No dismissal of debt if insolvent, debtor s prison, all loans full recourse, (minimal) homestead exemptions Inheritance and dowries No primogeniture Normal to convey or will property to daughters; marriage market and grandchildren
7/27 Crisis of 1837 and its aftermath Crisis of 1837: sharp drop in cotton prices and land and slave values; foreclosures Common Law: husband s creditors could seize wife s property Widespread concern with position wives/daughters, family life in general The reverses of the last few years have shown so much devastation of married woman s property by the misfortunes of their husbands, that some new modification of the law seems the dictate of justice as well of prudence (Tennessee Observer, 1843) [There is no good reason]why property bequeathed to a daughter should go to pay debts of which she knew nothing, had no agency in creating, and the payment of which, with her means, would reduce her and her children to beggary. This has been done in hundreds of instances, and should no longer be tolerated by the laws of the land. (Georgia Journal, 1843)
8/27 Typical Married Women Property Law
9/27 Law changes per state
10/27 Setup Preferences: U M (c 0, c 1 ) = log c 0 + θ M E [log(c 1 )] Investment requires two inputs (Leontief): Fixed asset (land, slaves) - fraction α Variable inputs (wages, slave rents, seeds, etc.) - fraction 1 α Risky investment Failure (π = 1 2 ): αi Success (π = 1 2 ): [α + (1 α)r] I R > 2
11/27 Financial markets Lenders are risk neutral Collateral constraint in the spirit of Hart and Moore (1994) and Kiyotaki and Moore (1997) Creditors can always seize fixed assets αi In good state of the world can only seize [α + (1 α)βr] I Intuition: borrower can threaten to withdraw human capital; bargaining over surplus (1 α)ri. Incomplete markets: simple debt contracts
12/27 Impact Married Women Property Laws Total investment, I = w M + w F c 0 + l Before the law, creditors can seize: [α + (1 α)βr] (w M + w F c 0 + l) (good state) α (w M + w F c 0 + l) (bad state) After the law, creditors can only seize: [α + (1 α)βr] (w M c 0 + l) (good state) α (w M c 0 + l) (bad state)
13/27 Consumption / borrowing decision - before law Household will always pick a riskfree loan, never at constraint max log c 0 + θ M c 0,l 2 log {[α + (1 α)r] (w M + w F c 0 + l) l} + θ M 2 log {α (w M + w F c 0 + l) l} l = R(2α 1) 2α 2(1 α)(r 1) (w M + w F c 0 ) c 0 = w M + w F 1 + θ M We assume that R(2α 1) > 2α: always borrow
14/27 Consumption / borrowing decision - after law Risky loan: (1 + ρ)l = (2 α)l α(w M c 0 ) max log c 0 + θ { M [α + (1 α)r] c 0,l 2 log (wm + w F c 0 + l) (2 α)l + α(w M c 0 ) } s.t. l l = c 0 w M 2α + (1 α)βr (1 α)(2 βr) (w M c 0 ) with α + (1 α)r > 2 α (2 βr) > 0 (assumption)
15/27 Three different cases 1 w M /w F < φ 1 c 0 = w M ; l = 0 2 φ 1 w M /w F φ 2 l = l; c 0 = 2 2 + θ M w M + (2 βr) [α + (1 α)r] w F (2 + θ M ) (1 β)r 3 w M /w F > φ 2 c 0 = c 0 ; l = l
16/27 Main results φ 1 w M /w F φ 2 Lemma c 0 < c 0 Lemma Define y and y to be optimal total investment before and after the law (w m + w F c 0 + l) 1 l l is strictly increasing in w M /w F 2 y y is strictly increasing in w M /w F Lemma y > y for at least part of the w M /w F distribution Note: no general equilibrium effects
Introduction Historical context Model Data Empirical results Conclusions 17/27
18/27 Data sources 1 Records of individual marriages contracted between 1840 and 1850 in US South (familysearch.org) 2 Complete count census of 1850 (NBER) Information about gross value real estate & slave holdings individual couples - proxy for total investment 3 Complete count census of 1840 (NBER) Information about slave holdings families - proxy for wife s and husband s wealth at marriage
19/27 Construction dataset First step: match (1) to (2) ID couples living in South Search 1850 census for couples whose names match those of marriage records (NYIIS code + test of string similarity) Unique matches: construct distribution of probable ages Refine multiple matches, drop observation if multiplicity remains Overall match rate around 20% Second step: construct pre-marital wealth from (3) Little information: surname of household head and number of household members only Compute average slaveholdings by surname-state Match to husband s surname / wife s maiden name and state of birth Southern born only: 88% of people in 1850 census, 76% of these can be linked to 1840 census through surname-state
20/27 Summary statistics
21/27 Investment decision log I i,j,s,t = βlaw s,t + with: γ [log(w i,1840 /W j,1840 ) median] LAW s,t + δ 1 log W i,1840 + δ 2 log W j,1840 + δ 3 X i + δ 4 X j + τ t + σ s + u i,j,s,t I i,j,s,t gross (unemcumbered) value of real estate and slaves i = husband, j = wife, s = state of marriage, t = year of marriage X i /j : additional controls including age, state of birth, literacy, state-specific linear time trend. τ t : year of marriage fixed effects, σ s : state fixed effects Estimated by OLS and Tobit Notes: in state marriages only
Investment - OLS log(w i,1840 /W j,1840 ) : p(75) p(25) = 1.3. 22/27
Investment - Tobit log(w i,1840 /W j,1840 ) : p(75) p(25) = 1.3. 23/27
24/27 Investment Mix
Introduction Historical context Model Data Empirical results Conclusions Investment vs Endowments 25/27
26/27 Results driven by changing marriage matches?
27/27 What do we learn? Exogenous variation in bankruptcy protection Comparable with a (hypothetical) introduction of a modern bankruptcy code No general equilibrium effects Outcomes Bankruptcy protection matters for investment decision Net effect of bankruptcy protection depends on share assets protected Increase (decrease) in investment for the upper (lower) part of the w M /w F distribution