DEPOSITS and BANK CAPITAL STRUCTURE by Allen, Carletti, and Marquez DISCUSSION. By Paolo Fulghieri UNC, CEPR, ECGI

Transcription

1 DEPOSITS and BANK CAPITAL STRUCTURE by Allen, Carletti, and Marquez DISCUSSION By Paolo Fulghieri UNC, CEPR, ECGI Universita L. Bocconi CAREFIN October 24, 2014

2 1. INTRODUCTION This paper addresses an important problem: What are the determinants of banks capital structure? Common assumption is that equity is more expensive than deposits. Typical justifications is that deposits are a form of debt financing Deposits are cheaper because of tax deductions. This paper takes a new view: What about if the debt and equity markets are segmented? If deposit markets and equity markets are segmented, the standard MM argument of homemade leverage (arbitrage) does not apply.

3 2. THE MODEL Two types of investors: Capital investors: capital K, can be invested in either in equity or deposits; Depositors: capital D, can invest only in deposits. Markets are segmented. Technology: Risky: CRS with return r ~ U[0, R] and Er = R/2 > 1. Safe storage. Return on bank s equity: ρ R/2 (invest in risky technology). Return on deposits: u 1 (invest in safe technology). Deposits are cheaper, since they have a worse outside option! Financing with deposits is potentially costly since there are bankruptcy costs: => Recovery rate h B

4 THE MODEL (2) Banking system is competitive and earns zero profits. Paper examines several configurations and extensions: 1. No bankruptcy costs: capital structure is irrelevant (MM); Cost of equity = cost of deposits => ρ = u = R/2 Equity and debt are equivalent and bank competition raises the cost of debt to its marginal product, R/2. Depositors gain in that the return on deposits is above the reservation level, 1. Cost of equity is at the reservation level.

5 THE MODEL (3) 2. Bankruptcy costs: Deposit financing is potentially costly because of bankruptcy costs; Equity is valuable since it avoids these costs. In equilibrium ρ > R/2 u = 1 deposits are abundant N B (1- k B )<D u > 1 if deposits are scarce N B (1- k B )=D Interesting property: the private and social optimum coincide (the regulator would choose the same capital structure of banks as chosen individually by banks).

6 THE MODEL (4) 3. The role of deposit insurance With deposit insurance the government injects liquidity if the bank is not solvent. Deposit insurance is paid with a lump-sum taxation. Banks hold no equity capital. Capital providers and depositor fund banks as deposits with ρ =u=r. The government suffers bankruptcy costs. This generates the need of capital regulation, giving: Low R: ρ reg > R/2 > u reg = 1; Large R ρ reg > R/2 > u reg > 1. In addition, the regulatory solution provides social welfare and, interestingly, lower bank capital (since it reduces deposit rate and thus capital needs).

7 THE MODEL (6) Extensions: Banks do not hold assets directly, but lend to firms. Two cases: 1. Public firms (can raise external equity): banks act as conduits. Hold zero capital and default every time firms default. 2. Private firms (limited equity): banks hold positive capital for large R and large bankruptcy costs.

8 COMMENTS Interesting paper. Raises a lot of important issues. I like the idea that deposits are cheaper than equity because investors have different outside option than contributors of equity capital. I also like a lot the idea that equity capital is in short supply and competes with firms in seeking capital from investors. Nice paper!

9 COMMENTS (2) Possible future work: 1. CRS technology vs concave production function? 2. Surplus allocation: banking and corporate sectors are perfectly competitive and all surplus is allocated to investors (debt and equity). What if banks are oligopolistic? What if firms retain surplus? 3. Capital accumulation: relative supply of debt and equity is a key driver. But if deposits are scarce, they extract more surplus and become relatively more abundant. What is the steady state?