Discussion of "Market Structure, Credit Expansion and Mortgage Default Risks" Liu, Bo; Shilling, James; and Sing, Tien Foo

Transcription

1 Discussion of "Market Structure, Credit Expansion and Mortgage Default Risks" Liu, Bo; Shilling, James; and Sing, Tien Foo Discussed by Yao-Min Chiang, Department of Finance National Chengchi University, Taiwan 1

2 Securitization and deregulation change market structure of the banking industry. Market structure will affect loan demand, and market structure will affect loan supply, and then consequently affects the loan stock in market and the risk of loans. Question: The paper describes the motivation like a time-series story. However, the paper focuses on cross-sectional analysis. This paper can provide implication for the subprime crisis. 2

3 Before regression analysis, run portfolio analysis. Form quintile portfolios based on market structure, concentration and contestability, respectively. At the beginning of year t, zip-code counties are ranked in ascending order based on their concentration and on their contestability and are assigned to a corresponding quintile. The credit supply and credit risk on these quintile portfolios are calculated for the same period and are rebalanced in each year. Then calculate the supply difference and the default risk difference between the low and the high concentration / contestability. This will help to have a clear view on the dynamic impact of market structure on credit supply and credit risk. 3

4 Mt thinking on the research issue: Market structure affects loan demand. 4

5 Market structure affects loan supply. 5

6 Question: Two opposite views on how competition affects credit risk. (1) A positive relationship between banking market concentration and bank risk-taking. Boyd, De Nicoló and Al Jalal (2006), DeNicoló and Loukoianova (2007) (2) As the number of banks increases, the probability of bank default first declines but increases beyond a certain point. Martínez-Miera and Repullo (2007) So, (1) Will concentration cause risk premium to increase or to decrease? (2) Will concentration cause credit quality to deteriorate? 6

7 Question: When banks decide to originate higher risk loans, they bear higher risk. Do they also make higher return? When we discuss increasing credit risk, should we also discuss whether banks make higher return? As Martín, Salas and Saurina (2006) argue that failure to take the risk premium into account would result in significant biases in measuring bank market power 7

8 This paper uses a 4-quadrant Model for the Credit Market 8

9 (a) Securitization, changing market structure, affects borrower s demand for loans. (b) Securitization, changing market structure, affects lender s supply of loans. (a) and (b) together decide mortgage stock and mortgage quality. This paper considers demand and supply of loans are endogenously decided by market structure, and then consequently affect mortgage stock and mortgage quality. (c) Most literatures discuss directly how market structure affects mortgage stock and mortgage quality. 9

10 Question: However, this paper focuses only on supplydriven story. This paper only treats F, credit supply as an endogenous variable, but does not incorporate endogenous demand in the empirical model. Mortgage stock and mortgage quality, especially, mortgage quality will be affected by the demand side, the borrowers side. 10

11 Gross and Souleles (2002) estimate a duration model for bankruptcy using a panel of credit card accounts from a single credit card issuer for They put forth two leading explanations for the increasing trend in bankruptcy throughout the period: a deterioration in the risk composition of borrowers (supply or risk effect) versus an increase in borrowers willingness to default due to declines in default costs (demand effect). Although this paper incorporates several variables on the demand side, there are still some important variables missing. Like income, unemployment, etc. Again, there is no endogenous model for mortgage demand. 11

12 Question: Securitization affects loan supply and loan demand. Should the important variable, WAC, be incorporated in to the supply function, and the demand function, if exists. 12

13 This paper uses the Gini coefficient and the H-statistic of Panzar- Rosse (1982 and 1987) as the measures the concentration and competition in the banking markets, and test how these two indicators affect credit supply and credit quality using samples of non-agency mortgages in the US for the periods Question: A matching problem. Gini coefficient and the H-statistic: based on those non-agency mortgages, or on Wharton s banking database of the University of Pennsylvania as inputs for the computations of the market structure variables. (covering agency and non-agency loans) 13

14 (1) Other than the mortgage suppliers concentration, the concentration of high-risk loans in at the zip-code and neighborhood levels was also found to have significant effects on mortgage default risks. (2) Mian and Sufi (2009) found that mortgage increases in subprime zip-code counties were twice as much as the increases in prime zip-code mortgages during the periods from 2002 to They credit growth coincided with increases in securitization activities contributed to the sharp rises in defaults rate in the subprime zip-code counties in

15 (3) Agarwal, Ambrose, Shomsisengphet and Sanders (2009) when testing mortgage default in Phoenix, Arizona found high concentration of subprime mortgages in selected zip-codes, especially in the lower income neighborhoods and the Central Business District. Their results, however, showed no significant relationship between concentration of the subprime mortgages and default probability. However, this paper shows there is a strong relation between mortgage suppliers concentration and default probability. (4) Also, Gini coefficients based on real estate loan value and total assets moved in opposite direction during some periods. It means that concentration measured based on different level (type) of data may affect research results. 15

16 In the appendix, it says there are three market structure variables: number of banks, concentration (Gini coefficient) and contestability (H-statistics). When/where is number of banks used? As the number of banks increases, the probability of bank default first declines but increases beyond a certain point.( Martínez-Miera and Repullo, 2007) Treating number of banks as a controlling variable? 16

17 During the period of concentration decreased and contestability increased. In terms of impact on credit quality, which dominates which? 17

18 The erosion of competitiveness in the markets could be due to entry of hit-and-run players into the markets. After the subprime crisis in 2007, some inefficient hitand-run players may have exited the markets, and the market competitiveness increases in Question: The increase of credit supply and of credit risk is purely brought by these hit-and-run players??? Will those non-hit-and-run lenders also increase credit supply and get higher credit risk after the entry of those hit-and-run players? 18

19 Zip-code county year observations are used in the regression estimate Two-way clustering problem? Petersen (2009) Fama-Macbeth standard errors Newey-West standard errors 19

20 10%, 50%. This paper form sub-samples based on concentration and on contestability. Sometimes, the cut point is 10%, and sometimes it is 50%. It is confusion. 20

21 The coefficient on the non-linear contestability term is positive and significant, which shows a downward convex relationship of the market contestability on mortgage supply. The coefficients on the concentration variable and the squared term are both significant, but the signs are different. The mortgage supply is a positive concave function of the concentration of banking market. Find out the optimal levels of concentration and of contestability to be cutting of points (concentration: 4.91; contestability: 0.048) 21

22 The HPI difference between strong concentration market and weak concentration market becomes larger in recently years. The HPI difference between strong contestable market and weak contestable market becomes larger in recently years. Is it due to low interest rate? If so, interest may affect the relationship between market structure and credit supply and credit risk. Interest (WAC) should be incorporated into the model. If not, what behind this? 22

23 Some minors: 1. Some minors: 2. No page number. 3. Alphabetic order in the reference. 4. Table format: highlight, (coefficient, stand error, p-value) (coefficient, t-value, ***) 5. On page 2, Section 2 6. On page 13, when describing Figure 7, three contestable markets, three? Page 14, three submarkets? 7. On page 19, However, the the negative 8. Figure 3: Residential Mortgage default rate) at zip-code country level (Originated at year Figure 5 Note This figure shows a 10. The pooled OLS regressions are estimated using maximum likelihood method. 11. On page 50, Table : Test 23