TRADE CREDITORS BEHAVIOR IN BANKRUPTCY BEN IVERSON WITH VICTORIA IVASHINA SEPTEMBER 2014

Size: px

Start display at page:

Download "TRADE CREDITORS BEHAVIOR IN BANKRUPTCY BEN IVERSON WITH VICTORIA IVASHINA SEPTEMBER 2014"

Grant Burns
5 years ago
Views:

1 TRADE CREDITORS BEHAVIOR IN BANKRUPTCY BEN IVERSON WITH VICTORIA IVASHINA SEPTEMBER 2014

2 MOTIVATION Firms obtain large amounts of financing from their suppliers. Trade credit represents 23% of total liabiliqes for average U.S. firm on par with the size of bank loans. RelaQvely few studies on trade credit because hard to get data. We use detailed informaqon on trade creditors from 132 bankrupt firms to examine their behavior during bankruptcy.

3 MOTIVATION Our focus: Do suppliers have private informaqon about firms that borrow from them? Can we detect that via their trading behavior during bankruptcy? Bankruptcy is a nice laboratory that helps to focus on this parqcular quesqon. Our findings help understand why trade credit exists and is so prevalent.

4 PREVIEW OF RESULTS Private and smaller firms rely more heavily on trade credit as a source of financing Sales by informed suppliers predict low recovery rates Result is strongest for firms that rely more heavily on trade credit as compared to bank debt Informed suppliers lead the market when recovery rates are lower

5 OUTLINE 1. Theories of trade credit 2. Data 3. Trade credit sales and recovery rates 4. Trade credit sales and other bankruptcy outcomes

6 WHY DOES TRADE CREDIT EXIST? (1) Trade credit is prevalent: Share of cases with type of debt Average share of total liabilites Bank debt Trade credit Bonds Asset managers, HF & PE Individuals Government Other Total: Source: Ivashina, Iverson & Smith (2014) Why do firms borrow from their suppliers instead of financial intermediaries?

7 WHY DOES TRADE CREDIT EXIST? (2) Main theories: a. Suppliers have advantage in obtaining informaqon about borrower s future prospects b. Suppliers have role in verifying product quality c. Suppliers are more willing to provide liquidity because strategically Qed d. Trade credit reduces transacqon costs e. Trade credit is legal means of price discriminaqon Not mutually exclusive! All could be reasons trade credit exists

8 WHY DOES TRADE CREDIT EXIST? (3) Main theories: a. Suppliers have advantage in obtaining informaqon about borrower s future prospects b. Suppliers have role in verifying product quality c. Suppliers are more willing to provide liquidity because strategically Qed d. Trade credit reduces transacqon costs e. Trade credit is legal means of price discriminaqon Once firm is in bankruptcy, we can observe trading behavior of trade creditors è tells us about information Ignore provision of new trade credit, which might be affected by other motives.

9 WHY DOES TRADE CREDIT EXIST (4) Previous research has focused on trade credit as a subsqtute for (informed) bank debt: Firms with weaker banking relaqonships have higher accounts payable Petersen & Rajan (1997); Giannef, Burkart & Ellingen (2008) Trade credit increases when bank credit contracts Nilsen (2002); Garcia- Appendini & Montoriol- Garriga (2013); Carbo- Valverde, Rodriguez- Fernandez & Udell (2014) Use of trade credit is higher in countries with less- developed banking systems Demirgüç- Kunt & Maksimovic (2001); Fisman & Love (2003) ContracQng terms respond to repeated transacqons and new informaqon Antrás & Foley (2014)

10 DATA Main obstacle in studying trade credit is lack of data. When firm files for Ch. 11, must create Schedule of LiabiliQes, containing nearly complete informaqon on all creditors. Data source: leading claims administrators - BMC Group, Epiq Bankruptcy SoluQons, Kurtzman Carson Consultants, Donlin Recano & Company Total of 132 bankruptcies between private firms, 64 public

11 BANKRUPT FIRMS Table 1 Obs. Mean Std. dev. Median Total assets at filing (million $US) 129 1,967 4, Total liabilities at filing (million $US) 129 1,850 4, Total liabilities/total assets (no outliers) By industry: Mining & Construction 5.5% Manufacturing 38.3% Transportation, Communications, and Utilities 15.6% Wholesale & Retail Trade 21.1% Finance, Insurance, and Real Estate 8.6% Services 10.9% Restructuring outcome: Reorganized 44.7% Sold to a financial buyer 9.9% Sold to a strategic buyer 12.1% Liquidated Piecemeal 33.3%

12 CAVEAT Our data is condiqonal on bankruptcy might not be representaqve. Reassuring facts: Aggregate reliance on trade credit matches other studies Rajan and Zingales (1995): 22.8% of total liab. Compustat firms from : 23.4% Our sample: 22.5% The pre- bankruptcy trend is minor (next slide) There is no formal market for trade claims outside of bankruptcy But more can and should be done: Verify individual suppliers using segment data

13 TRADE CREDIT USAGE BY FINANCIALLY DISTRESSED FIRMS Figure 2 6% Accounts payable / Total liabilities 4% 2% 0% -2% -4% -6% -8% Quarters before bankruptcy

14 DATA At the claim- level, we observe the owner of the claim, the total amount, and amount that is secured at the bankruptcy filing IdenQfy claimholders by matching name to parent insqtuqon, focusing only on claims > $50K. 14,870 unique trade creditors holding 28,072 claims Selling of trade claims subject to disclosure under Rule 3001(e). All trades of instruments NOT registered with SEC or a loan syndicate must be disclosed - excludes public bonds and syndicated loans 2,176 trade claims sold (7.8% turnover) For a subset of trades, we also see the date of the trade

15 WHO RELIES ON TRADE CREDIT? Table 3 Hypothesis: private and smaller firms are more opaque. Therefore they should use more trade credit if their suppliers have advantage in obtaining information. Dependent variable: Trade credit share of total claims (1) (2) (3) (4) Private 0.260*** 0.273*** 0.370*** 0.366*** (0.030) (0.020) (0.025) (0.028) Public (marginal effect) ** * (0.038) (0.040) (0.059) Ln(Assets) *** * (0.004) (0.009) Industry fixed effect No Yes Yes Yes Observations R- squared Result is intuitive, but contrary to Rajan and Petersen (1997)

16 IDENTIFYING INFORMED SUPPLIERS To get cross- secqonal variaqon, we idenqfy suppliers who are most likely to be informed: We are able to match 19% of our sample of trade creditors to Capital IQ to get balance sheet informaqon For these firms, a supplier is informed if it has an above- median payables- to- total- assets raqo. Idea: idenqfy firms that have lent a lot relaqve to their own capacity, and assume they pay the most auenqon to these relaqonships For firms without balance sheet data, we assume they are likely to be small suppliers, and define them as informed if their claim amount is above the 75 th percenqle in size. Robustness: get same results with subsample of firms with balance sheets, or using different thresholds

17 SALES BY INFORMED SUPPLIERS (1) Use claim- level regressions (trade credit only) of the form: Recov ic =α+ β 1 Informed ic + β 2 Sold ic + β 3 Informed ic Sold ic +γ X ic +δ Z c + ε i InteracQon term Informed ic Sold ic idenqfies whether sales by informed suppliers predict low recovery rates. Z c includes firm size, profitability, bankruptcy outcome dummies, Qme in bankruptcy, and industry fixed effects.

18 SALES BY INFORMED SUPPLIERS (2) Table 4 Dependent variable: Claim- level recovery rate (1) (2) (3) Claim sold 0.136*** 0.084*** 0.019** (0.039) (0.030) (0.009) Informed supplier (0.015) (0.014) (0.011) Informed supplier * Claim sold *** *** * (0.019) (0.017) (0.011) Share of claim that is secured 0.612*** 0.588*** 0.610*** (0.070) (0.068) (0.061) Case- level recovery rate *** - - (0.103) Other case- level controls Yes Yes - - Industry fixed effect Yes Yes - - Bankruptcy case fixed effect Yes Obs erva ti ons 23,125 22,814 23,987 R - squared

19 SALES BY INFORMED SUPPLIERS: FURTHER EVIDENCE Hypothesis: When firms rely more heavily on trade credit, suppliers are most informed If informed debt is the cheapest, firm will want to rely most heavily on informed debt à main source of debt should be the most informed source We split the sample (above- and below- median) by reliance on trade credit relaqve to: Total liabiliqes Bank debt Test if sales are most predicqve for heavy trade credit users

20 SALES BY INFORMED SUPPLIERS: TRADE CREDIT VS. TOTAL LIABILITIES Table 5 Panel A Dependent variable: Claim- level recovery rate Heavy trade credit users Light trade credit users (1) (2) (3) (4) Claim sold 0.103** 0.077** (0.044) (0.034) (0.040) (0.029) Informed supplier (0.009) (0.007) (0.020) (0.020) Informed supplier * Claim sold *** *** (0.015) (0.016) (0.023) (0.022) Share of claim that is secured 0.676*** 0.660*** 0.581*** 0.580*** (0.055) (0.047) (0.115) (0.115) Ca s e- l evel recovery ra te *** *** (0.071) (0.188) Other case- level controls Yes Yes Yes Yes Industry fixed effect Yes Yes Yes Yes Obs erva ti ons 8,976 8,896 14,149 13,918 R - squared

21 SALES BY INFORMED SUPPLIERS: TRADE CREDIT VS. BANK DEBT Table 5 Panel B Dependent variable: Claim- level recovery rate Heavy trade credit users Light trade credit users (1) (2) (3) (4) Cl a i m s ol d 0.094** 0.048* (0.039) (0.026) (0.041) (0.031) Informed supplier (0.008) (0.006) (0.021) (0.021) Informed supplier * Claim sold *** ** (0.019) (0.014) (0.021) (0.023) Share of claim that is secured 0.665*** 0.648*** 0.598*** 0.591*** (0.096) (0.094) (0.080) (0.081) Ca s e- l evel recovery ra te *** * (0.093) (0.195) Other case- level controls Yes Yes Yes Yes Industry fixed effect Yes Yes Yes Yes Obs erva ti ons 10,949 10,722 12,176 12,092 R - squared

22 TIMING OF SUPPLIERS EXIT (1) For 282 trades, corresponding to 20 different bankruptcy cases, date of the sale is recorded For this sample, we test whether informed suppliers lead the market by selling their stakes first when recovery rates are low. Dependent variable: Days unql sale Split 20 cases into above- and below- median recovery rates Main explanatory variable: Informed supplier*low recovery

23 TIMING OF SUPPLIERS EXIT (2) Table 6 Dependent variable: Days until sale (1) (2) (3) Informed supplier (25.250) (9.892) (6.316) Informed s uppl i er * l ow recovery * *** ** (62.653) (55.559) (18.438) Low recovery ( ) (92.331) Other case- level controls - - Yes - - Industry fixed effects - - Yes - - Bankruptcy case fixed effects Yes Find similar results with fraction of bankruptcy until sale as dep. variable Obs erva ti ons R - squared

24 INTERACTION OF SUPPLIERS EXIT AND BANKRUPTCY OUTCOME Suppliers hold implicit stake in reorganized company Wilner (2000) hypothesizes that in such cases trade creditors might grant concessions to: Allow quicker exit Preserve firm as going concern (to preserve relaqonship) Want to test whether sales by informed suppliers are correlated with these outcomes. Important caveat: relaqonships not necessarily causal!

25 SUPPLIERS TRADING DECISION AND BANKRUPTCY OUTCOMES Table 7 Dependent variable: Months in ba nkruptcy Ba nkruptcy outcome: Reorganization Sale (1) (2) (3) (4) Share of informed claims that are sold * 0.330*** Liquidation (19.065) (0.149) (0.073) (0.186) Share of total trade claims that are sold (25.430) (0.516) (0.589) (0.528) No informed claims dummy (4.211) (0.336) (0.340) (0.092) Ca s e- l evel recovery ra te ** (6.217) (0.084) (0.155) (0.154) Prepa cka ged ba nkruptcy *** 0.272** *** (3.800) (0.089) (0.137) (0.087) Other case- level controls Yes Yes Yes Yes Industry fixed effects Yes Yes Yes Yes Obs erva ti ons R - squared

26 CONCLUSIONS Our findings point towards suppliers holding informaqonal advantage in providing credit: Trade credit is larger source of financing for small, private firms Informed suppliers more likely to sell when recovery rates are low Especially for firms that rely heavily on trade credit Informed suppliers also lead the market when recovery rates are low Sales by informed suppliers correlate with lower propensity to survive as going concern

27 Thank you!

28 SIZE DISTRIBUTION OF INFORMED AND UNINFORMED SUPPLIERS Figure 1 Distribution of supplier size by informed measure Density Ln(trade creditor assets) kernel = epanechnikov, bandwidth = Informed Not informed

Trade Credit, the Financial Crisis, and Firm Access to Finance

Trade Credit, the Financial Crisis, and Firm Access to Finance Santiago Carbó-Valverde Francisco Rodríguez-Fernández Gregory F. Udell Presented at the BdE-CNMV Workshop on SME Finance Broad topic: THE