Finance and Banking

1
Finance and Banking

• NAKE Course
• Lecture 6 Banking and Economic Growth

2
Financial development and growth (1)

• Well-functioning financial intermediaries and
  markets ameliorate information and transaction
  costs and hence induce faster long-run growth (we
  present a model by Bencivenga and Smith, 1991,
  hereafter that illustrates this point)
• But financial development can also hurt growth.
  It enhances resource allocation and hence returns
  to savings and so may lower saving rates
  (especially with externalities)

3
Financial development and growth (2)

• Empirics are plagued by a simultaneity bias is
  growth driving financial development or the other
  way round?
• Is it the banking system or is it the stock
  market that enhances growth? Do they both
  interact? Or is it the governance of financial
  systems that matters?
• Legal view financial development related to
  legal institutions is important for growth
  financial structure as such is rather unimportant

4
Financial development involves improvements in

• The production of information about investment
  projects
• The monitoring of firms and exerting corporate
  governance
• Trading, diversifying, and managing risk
• Mobilization and pooling of savings
• Easing the exchange of goods and services.

5
1 Producing information and allocating capital

• FIs may reduce information costs, and so
  improve resource allocation King and Levine
  (1993). There might be large costs associated
  with evaluating firms, managers, and market
  conditions
• Stock markets may stimulate the production of
  information about firms (Holmstrom and Tirole,
  1993). As markets become larger and more liquid,
  agents have greater incentives to spend resources
  in researching firms and markets

6
2 Monitoring firms and exerting Corporate
Governance

• Informational problems lead to adverse selection,
  moral hazard, and costly state verification
• FIs Diamond delegated monitoring model
  economizing on aggregate monitoring costs,
  eliminating the free rider problem as the monitor
  monitors on behalf of all investors (see
  hereafter)
• Debt contracts may lower the costs of monitoring
  firm insiders (Townsend, 1979)
• Equity holders can exert corporate governance if
  they have effective controls (Shleifer and
  Vishny, 1997)

7
3 Risk amelioration

• Cross-sectional risk diversification (Gurley and
  Shaw, 1955, Greenwood and Jovanovic, 1990)
• Intertemporal risk sharing long-lived FIs can
  facilitate intergenerational risk sharing
• Liquidity risk Diamond-Dybvig (1983),
  Bencivenga-Smith (1991). Liquidity risk arises
  due to uncertainties associated with converting
  assets into monies. DD savers allocate over
  illiquid and liquid assets. Random shocks force
  some savers to liquidate illiquid projects.
  Verification costs create incentives for
  financial markets to emerge

8
4 Mobilizing savings

• Overcome the transaction costs of collecting
  savings from multiple agents
• Overcoming the informational asymmetries
  individual savers should feel comfortable in
  handing over their savings
• Via financial markets or intermediaries (no big
  difference)

9
5 Facilitating exchange

• Technological innovation needs specialization
  that requires more transactions
• Financial arrangements that lower transaction
  costs facilitate greater specialization
• Feedback from productivity gains to financial
  market development (Greenwood and Smith, 1997)

10
Theory of financial intermediation

• Old view transactions costs
• New views
• Liquidity insurance (Diamond-Dybvig, 1983
• Information sharing coalitions
• Delegated monitoring (Diamond, 1984)

11
1 Diamond-Dybvig Liquidity insurance

• Idea consumers are uncertain about the timing of
  consumption and invest their funds at the bank in
  a deposit as an insurance
• FIs are a pool of liquidity
• If households vary in risk profiles FIs do not
  need to hold full cash for the deposits

12
Diamond-Dybvig (1)

• One good, three periods
• Continuum of ex-ante identical agents
• Each agent is endowed with one good at t 0
  this is the planning date
• A good can be consumed at t 1 or t 2
• There is a probability of a liquidity shock,
  through which a consumer wants to consume at t 1

13
Diamond-Dybvig (2)

• A good can be stored at a rate 1 or invested in a
  productive technology at a rate R gt 1 and this
  return will be paid out at t 2
• Early liquidation at t 1 gives a yield L lt 1
• Probability of being impatient is ?1, ?2 is the
  probability of consumption at t 2, ? discount
  rate
• Ex ante expected utility of the consumer
  U ?1u(C1) ??2u(C2)

14
Diamond-Dybvig (3) cases

• Case I Autarky we will show that C1 lt 1 and C2
  lt R the worst solution
• Case II A bonds market is opened at t 1. This
  is a better solution, the feasible consumption
  set is extended C1 1 and C2 R
• Case III A depositing bank allocates liquidity
  and allows C1 gt 1 and C2 lt R

15
Diamond-Dybvig (4) autarky

• No trade between agents how big is individual
  investment I?
• C1 1 I LI 1 - I(1 - L) ? 1
• C2 1 I RI 1 (R - 1)I ? R
• So maximize utility given these values of C1 and
  C2 using the utility function U
  ?1u(C1) ??2u(C2)

16
Diamod-Dybvig (5) a market for bonds

• A bonds market opens at t 1. A bond pays one
  unit of the good at t 2. The bond price is p ?
  1 (if p gt 1 storage is more profitable)
• C1 is now 1 - I pRI early consumption leads to
  selling RI bonds at a price p. The consumer owns
  RI at date 2 and can sell these claims at t
  1 at pRI
• C2 (1 - I) / p RI buy (1 - I) / p of bonds.
  (1 - I) is savings that can be used to buy bonds

17
Bonds market (2)

• It is easy to prove that arbitrage will force
  p 1/R. A bond returns 1 unit at t 2
  at a price p at t 1 an investment returns R at
  t 2 at an investment of 1 at t 0. A high p
  will lead to infinite investment
• C1 1 and C2 R
• This result is better than autarky
• But still not Pareto optimal! C1 gt 1 and C2 lt R
  are likely to be better the bonds market does
  not supply full hedging

18
Pareto Optimality

• Find (C1,C2) that optimizes ex ante
  ?1u(C1) ??2u(C2) given ?1C1 ?2C2/R 1
• FOC u(C1) ?Ru(C2)
• Only in a special case u(1) ?R u(R) this is
  the case of the bond market.
• Assume risk aversion (so Cu(C) is decreasing)
  and we will have ?R u(R) lt ?u(1) lt u(1)
• So C1 gt 1 and C2 lt R are better in terms of
  utility the consumer wants to consume more at
  t 1, than the bond market allows to do so

19
What can a bank do?

• The consumer gives 1 unit in a deposit
• The bank pays the optimal C1 gt 1 to early
  consumers and C2 lt R in t 2 to patient
  consumers
• The bank stores ?1C1 and invests
  (1 - ?1C1) at a rate R and can pay
  precisely ?2C2 R (1 - ?1C1)
• Note that the budget constraint for our consumer
  is ?1C1 ?2C2/R 1

20
2 Information sharing coalitions

• Idea if a group of firms can credibly
  communicate the quality of investment projects,
  borrowing costs will decrease with the number of
  members of the group
• A FI is a coalition of firms
• Scale economies are generated in lender-borrower
  relationships if firms are better informed on the
  quality of the investment project

21
Leland and Pyle capital markets with adverse
selection

• Firms invest 1 unit. Net returns R(?) are
  normally distributed with unknown mean ?, and
  variance ?2 (?2 equal for all firms). The
  distribution of ? is known. Firms have initial
  wealth W0 gt 1
• Investors are risk neutral and have a costless
  storage technology

22
Leland-Pyle (2)

• Firms are risk averse and want to sell the
  investment project. They use a CARA utility
  function in final wealth w U(w) -exp-?w, ? gt
  0 is the CARA-index
• If ? was observable, the entrepreneur would sell
  the project at a price P(?) ER(?) ? and
  final wealth would be W0 ?
• Since ? is private information the price of
  equity will be the same for all firms

23
Leland-Pyle (3)

• Self financing the project will give
  Eu(w R(?)) u(W0 ? - ½??2)
• Selling the project to the market will yield u(W0
  P)
• So entrepreneurs with low expected return ? ?
  ?b P ½??2 will issue equity!
• Price of equity is P ? ? lt ?b

24
Leland-Pyle (4)

• Equilibrium of the equity market is characterized
  by the price P and the cut-off rate ?b. Suppose
  the distribution of ? is binomial P E?
  ?1?1 ?2?2
• Efficiency requires that all investors get
  outside finance ?b ? ?2
• So ?1(?2 - ?1) ? ½??2

25
Signaling through self-financing

• Idea high quality firms signal by investing part
  ? of their internal wealth
• As long as bad firms do not mimic the good firms,
  we get two equity prices (good and bad)
• Leland and Pyle prove that the informational cost
  of capital is ½??2?². This increases with the
  level of self-financing ?

26
Proof

• Binomial no mimicking condition u(W0
  ?1) ? Eu(W0 (1 - ?)?2 ?R(?1)), or utility
  of a ?1 firm must be bigger than mimicking a ?2
  firm and self-financing ?
• Expected utility equivalent of the last term on
  the right-hand side ??1 - ½??2?2
• The last term is the informational cost of capital

27
Coalition of borrowers

• If high quality firms can credibly signal the
  quality of their projects, the variance can be
  reduced from ?2 to ?2/N (because of
  diversification).
• Diamond (1984) shows that the unit cost of
  capital decreases with the size of the coalition
  of borrowers

28
3 Diamond (1984) delegated monitoring

• Idea banks are good in monitoring projects
  increasing returns to scale
• n firms, 1 unit investment
• cash flow y uncertain, unobservable to FI
• K monitoring costs
• C costs of a debt contract K lt C if the firm
  had a unique financier, it would be optimal to
  choose the monitoring

29
Diamond (2)

• Per project m investors are needed
• Direct financing of n projects costs nmK in
  monitoring
• Let the FI do the job nK monitoring costs plus
  costs of monitoring the bank by the depositors
  Cn
• When does Cn nK lt nmK hold?

30
Delegated monitoring direct finance
Lender 1
Borrower 1
Lender m
Lender (n-1)m1
Borrower n
Lender m
Total cost nmK
31
Delegated monitoring Intermediate finance
Lender 1
Borrower 1
Bank
Lender m
Borrower n
Lender nm
Total cost nK Cn
32
Diamond (3)

• FI is cheaper if
• I Monitoring is efficient K lt C
• II Investors are relatively small m gt 1
• III Investment is profitable Ey gt K R

33
Proof of Diamond monitoring

• Definition of Cn. Bank offers a debt contract
  for a deposit 1 / m it offers a repayment Rd / m
• If the announce cash flow of the bank z lt
  nRd, the bank is liquidated z ?yi - nK
• Depositors are risk neutral and have an outside
  investment technology yielding a gross return R

34
Proof Diamond (2)

• What is the equilibrium repayment by the bank?
• The depositor must be indifferent between getting
  R on the outside technology or getting the
  minimum of the after cost result (?yi - nK)/n or
  Rd, the deposit rate.
• The cost of delegation are equal to the expected
  nonpecuniary penalty in case of bankruptcy Cn
  E max(nRd nK - ?yi, 0)

35
Costs of delegation

• Penalty is designed in such a way that the total
  return to the bank is equal to nRd, which is
  independent of the cash flow ?yi, so that the
  bank has no incentive to misreport earnings
• min(?yi - nK, nRd) Cn nRd, so
• Cn E max(nRd nK - ?yi,0)

36
Proof Diamond (3)

• When is nK Cnlt nmK, or K Cn/n lt mK? So it is
  sufficient to prove that Cn/n tends toward 0 if n
  goes to infinity.
• ?yi / n goes to Ey gt K R, so lim Rd R
  deposits are asymptotically riskless
• So Cn/n max(R K -Ey,0) 0

37
Theoretical example Bencivenga-Smith

• A general equilibrium model including a financial
  intermediary that influences the growth rate
• A 3-period OLG model with a liquid investment
  project (not directly productive) and an illiquid
  project (productive)
• Young agents, facing the likelihood of early
  liquidation, benefit from a Diamond-Dybvig
  (DD)-type of bank technology
• The DD-technology lowers liquid reserve holdings
  by the economy as a whole, and so also reduces
  the liquidation of productive capital
• Then, with production externalities, higher
  equilibrium growth rates will be observed in
  economies with a FI-sector

38
Diamond-Dybvig Optimal contract

• The optimal contract maximizes ex ante expected
  utility of the consumer
• Given (1) pay-outs are feasible, (2) the
  contract is Incentive Compatible
• Results (1) the optimal reserve is the amount to
  be paid out in period 1 to impatient clients, (2)
  it is never optimal to pay a consumer with a low
  return in period 1 or to make a payment to a
  high-return consumer in period 2

39
BS-observations

• Banks face a predictable withdrawal demand
• Banks hold liquid reserves against withdrawal
  demand
• Banks issue liabilities that are more liquid than
  their assets
• Banks reduce the need for self-financing of
  investment
• Banks economize on liquid reserve holdings

40
BS Development literature

• Legislation and government regulation determine
  exogenously the state of development of financial
  markets (McKinnon-Shaw financial repression)
• In underdeveloped markets we see only banks
• Long delays between receipt of profits and
  investment expenditure
• In absence of banks too much investment is
  self-financed
• In absence of banks agents self-insure against
  liquidity needs
• Role of banks is to optimize the composition of
  savings
• Economies with well-developed financial systems
  grow faster

41
BS (1)

• Two goods consumption and capital
• At t 0 we have an initial old generation,
  endowed with a per firm capital stock k0 at t 0
  and an initial middle-aged generation with k1 at
  t 1
• Capital is owned by a subset of old agents
  (entrepreneurs). There is no rental market for
  capital. kbt average capital stock per
  entrepreneur at t
• Consumption good is produced by a technology
  kbt?kt?Lt1-?. 0 lt ? 1 - ? lt 1. Lt is labor, ?
  denotes the external effect. Capital depreciates
  fully in one period

42
BS (2)

• All young generations are identical (no
  population growth) with a continuum of agents.
  Each young agent supplies a single unit of labor
  when young (no supply at age 2 and 3). ci age i
  consumption
• U(c1,c2,c3,?) -(c2 ?c3)-?/?, where ? 0 with
  probability 1 - ? or 1 with probability ?. All
  young income is saved. DD-technology is assumed

43
BS (3)

• Two assets liquid investment inventory of
  consumption good return n gt 0 at t 1 or t 2
• Illiquid investment good invest at t and get a
  return R at t 2 the slow cycle of production.
  Early liquidation yields a scrap value 0 ? x lt n
• All capital is in the hands of age-3
  entrepreneurs
• Maximization of kbt?kt?Lt1-? wtLt gives the
  demand for labor Lt kt (1 - ?)kbt?/wt1/?

44
BS Financial Intermediation (4)

• Not all old agents are entrepreneurs. 1 - ?
  liquidate assets at date 2 and have no capital. ?
  each hire Lt labor. Lt 1/?, so the
  equilibrium wage rate wt kbt(1 - ?)??
• Return to capital (share ?) is ???-1kt
• Banks invest deposits from young savers by a
  fraction zt ? 0 in liquid and qt ? 0 in
  illiquid investment zt qt 1

45
BS (5)

• r1t return (in consumption goods) to depositor
  for a withdrawal after 1 period
• r2t (r2t) return in capital (consumption)
  goods to depositor for a withdrawal after period
  2 (both capital and consumption goods are
  returned!). Let ?it be the liquidation fractions
  (i 1,2). The resource constraints are
• (1 - ?)r1t ?1tztn ?2tqtx
• ?r2t (1 - ?2t)Rqt
• ?r2t (1 - ?1t)ztn

46
BS (6)

• Bank is seen as a coalition of young agents at t.
  Young agents deposit entire labor income wt. At
  time t 1 a fraction 1 - ? liquidates and
  consumes r1t wt. A fraction ? gets r2t units of
  capital and r2t of consumption goods. Capital
  is used for production kt r2twt and earns
  profits ???-1kt
• Expected utility of a representative depositor
  is -(1 - ?)/? (r1twt)-? - (?/?)
  ???-1r2twt r2twt- ?
• Proposition If ???-1R gt n, then none of the
  capital invested is liquidated early and reserves
  are entirely liquidated ?1t 1 and ?2t 0.
  This is a famous DD-result! So r2t 0

47
BS (7)

• This result simplifies the problem of the bank.
  It can simply optimize qt, the fraction invested
  in the illiquid technology. See BS, p. 201
  qt f(?, ?, ?, n, R)
• From the simplified resource constraints we get
  (1 - ?)r1t ztn (1 - qt)n, so t 1
  consumption is r1twt (1 - qt)nwt /(1 - ?) and
  ?r2t Rqt, so period 2 consumption is ???-1r2twt
  ???-1Rqtwt /?. Patient agents will value
  consumption at time 2 higher than in period 1
  this leads to the assumption in the proposition
  ???-1R gt n

48
BS (8)

• Equilibrium kbt2 r2twt Rqtwt /? kt2
• kbt2 / kbt ? R(1 - ?)??-1qt drives the
  growth rate this is the basic result
• A higher labor share (1 - ?) increases the growth
  rate
• If capital becomes easier to produce (higher
  values of R), growth increases
• If qt increases we get a higher growth rate

49
BS no intermediation (1)

• Young agents save their entire income wt and
  choose the fraction qt to invest in the illiquid
  capital
• The yields are x on liquidated capital in period
  1, n on inventories at time 1 and 2 and ???-1R on
  illiquid capital at 2
• We can now compare qt and qt to get to the
  differences in growth rates

50
BS no intermediation (2)

• The equilibrium growth rate is higher if young
  agents are sufficiently risk averse
• Financial autarky is likely to result in
  relatively large holdings of liquid assets by
  individual agents
• Intermediaries reduce the reliance on
  self-finance this leads to higher growth!

51
Empirics

• Three issues
• What are the key indicators? How do these develop
  over time?
• What is the existing evidence?
• Methodology econometrics of cross sections and
  dynamic panel data estimators

52
Measuring financial development

• Indicators of the size of the financial sector
• Indicators for the development of banks
• Indicators for the development of the equity
  market

53
The size of the financial sector

• M3/GDP is a popular measure
• Disadvantages
• how is credit used?
• who is the lender?
• growth enhancing services of financial
  intermediaries are not picked up

54
Development of banks

• Commercial bank credit to the private sector over
  GDP
• Bank loans to private sector/ total loans
  financial sector
• Total assets commercial banks/total assets
  central bank
• These proxies indicate how credit is distributed

55
Development of the equity market

• Capitalization market value all listed shares
  over GDP
• Total traded shares/ GDP
• Turnover ratio total value of traded
  shares/total value of listed shares
• Standard deviation of market returns volatility
  measure

56
Empirics cross-country-banks

• Goldsmith (1969) 35 countries, 1860-1963
  financial intermediary size relative to size of
  the economy rises as countries develop
  graphical positive correlation
• King and Levine (1993) 77 countries, 1960-1989.
  Condition for other growth factors. DEPTH
  liquid liabilities/GDP, BANK bank credit to
  total credit, PRIVY credit to private
  firms/GDP. Finding strong positive correlation.
  Disadvantages focus only on banks, andproblems
  of causality

57
Empirics cross country-stock markets

• Levine-Zervos (1998) 42 countries, 1976-1993.
  Various measures market capitalization,
  turnover, volatility. Initial values of financial
  development influence growth no tension between
  stock market and bank-based results. Main
  channel productivity growth

58
Empirics cross-country causality

• Econometrics of Instrumental Variables
• Example Levine, Loayza, Beck (2000) use legal
  origin measures as instruments. 71 countries,
  1960-1995. Strong link between financial
  development and growth is not due to simultaneity
  bias (see hereafter for a more detailed
  discussion)

59
Industry/Firm level analyses

• Rajan and Zingales (1998) 36 industries, 42
  countries, 1980-1990. Some industries are
  strongly dependent on external finance and
  benefit strongly from financial development
• Wurgler (2000) industry-level, 65 countries,
  1963-1995. Countries with highly developed
  financial systems increase investment more in
  growing industries
• Firm-level Demirguc-Kunt and Maximovic (1998)
  26 countries, 1980-1991. Both bank development
  and stock market development are positively
  associated with excess growth of firms. Beck.
  Demirguc-Kunt, Levine, and Maksimovic (2001)
  confirm the findings in a larger sample

60
The methodology of intermediation-growth
regressions

• Main problems (1) simultaneity bias growth
  determines financial development, so we want the
  exogenous component of financial development in
  growth regressions, (2) omitted variables, (3)
  unobserved country-specific effects
• Solution see Levine, Loayza, Beck, LLB,
  (2000) (A) instrumental cross-sections, and (B)
  Dynamic Panel data (GMM) estimators