Banks and the Broader

Economy
The Long-run and Short-run
Relationships
Dr. Tony Hu

1/45
 Lecture overview

• Banks and economic growth

◦ the long-run impact
◦ the debate on whether finance matters
• Credit cycles and the amplification of macroeconomic shocks
◦ the short-run impact
◦ why small shocks can have large impacts given financial friction
• Next week
◦ Banking crisis
◦ Regulations for financial stability

2/45
Finance and Economic Growth

3/45
The Empirical Correlation

4/45
 Do financial systems matter?

• A debate that lasted for decades:

◦ Schumpeter (1911): financial service essential for economic development
◦ Robinson (1952): “where enterprises leads, finance follows”
◦ Goldsmith (1969): “a rough parallelism can be observed between economic and
financial development”
◦ Lucas (1988): “importance overstated”
• ⇒ Answer the question with real-world data
• Empirical challenges: reverse causality and simultaneity
◦ finance causes economic development
◦ or, finance only reflects economic development
• Three milestone papers that conclude finance does promote development

5/45
 King and Levine (1993)
Finance and Growth: Schumpeter Might be Right

• How do pre-determined financial development relate to long-term growth?

• Reasoning: causality implies sequences (a necessary condition)
• Measuring financial development
◦ traditional measurement ("financial depth") = size of financial intermediary sector
◦ liquid liabilities of the financial system to GDP
◦ other measures of financial services: deposit money banks, credit to private sector
◦ pre-determinacy: value taken from year 1960
• Measuring long-term economic growth
◦ average real per capita GDP growth over 1960-1989

6/45
Correlation Between Finance and GDP Growth

7/45
8/45
 Jayaratne and Strahan (1996)
The Finance-Growth Nexus: Evidence from Bank Branch
Deregulation

• Explore the variation in financial development due to reasons other than economic
development
• A natural experiment in the U.S.
• Policy shock that lifted branching restrictions
◦ inter-states M&A were allowed
◦ banks are allowed to open new branch in other states
• Leading to more efficiency in the financial sector
• The change of policy was not synchronised across different states
• Compare the economic performance of states that lifted the restrictions with that of
those that did not

9/45
Treatment vs Control States

10/45
11/45
 Channel of Influence

• Growth increase following intrastate branching is long-lasting

• No significant increase in lending quantity
• Lending quality improves: decrease in non-performing loans, charge-offs, insider
loans
◦ Better competition
◦ Economies of scale

12/45
 Rajan and Zingales (1998)

• Financial development should affect different industries differently

• Industries relying more on external financing should benefit more
• Industry characteristics determined by technologies
◦ being exogenous to the relationship between finance and development
◦ diff-in-diff estimation
• Measuring external dependence
◦ fraction of capital expenditure not financed by internal funds
• Measuring financial development
◦ total capitalization
◦ private credit to GDP ratio
◦ accounting standard

13/45
14/45
Industry Categorization

15/45
16/45
17/45
 Can it go too far?

• Frictions in the financial sector and possibility of banking crises

• Rent extraction in banking and allocation of social resources
• Kneer (2013): A study highlighting the possibility of human resource allocation
◦ US banking deregulation increases skilled labor in banking
◦ at the expense of the non-financial sector
◦ especially industries that are skilled labor and R&D intensive

18/45
Regression results

19/45
20/45
Credit Cycles

21/45
 Credit Cycles: An Overview of Theories

• What is a credit cycle, and why is it a problem?

• The quantity aspect
◦ excessive (and risky) credit expansion in boom
◦ excessive credit contraction in recession
• The quality aspect
◦ Greenspan: “the worst loans are made at the top of the business cycle”
• Causes of credit cycles
◦ Endogenous credit cycles due to market imperfections
— economic shocks amplified by financial friction
— greater friction during economic downturns ⇒ amplification
◦ behavioral exposition: institutional memory hypothesis
◦ regulatory reason: time-invariant capital requirement in Basel II

22/45
Loan Growth and Provisioning: Allied Irish Bank

23/45
 Endogenous Credit Cycles

• Amplification based on borrowers balance sheet

• First observed by Irving Fisher (1933): “During a debt-deflation, because of an
unanticipated fall in the price level (or, alternatively, a fall in the relative price of
borrowers’ collateral, for example, farmland), there is a decline in borrower net
worth. This has the effect of making those individuals in the economy with the most
direct access to investment projects suddenly un-creditworthy (i.e., the agency costs
associated with lending to-them are high). The resulting fall in investment has
negative effects on both aggregate demand and aggregate supply.”

24/45
 Amplification Mechanism: An illustration with H-T (1997)
• Net worth: an amplification mechanism in credit cycles
• Model re-interpretation and modification
◦ re-interpreting the borrower as a bank, with equity A (the net worth)
◦ variable investment size: the loan portfolio size, I, is the bank’s choice
◦ constant returns to scale (CRS): R being the per unit return
◦ the moral hazard gain B also a per unit
◦ pL and pH the same as before
• Equity multiplier due to moral hazard
◦ when pH (R − B/∆p) < 1, the project is not self-financing
◦ the (maximum) size of the investment
A
I∗ = = kA
1 − pH (R − B/∆p)

• with k > 1 when the project is not self-financing

• £1 reduction in equity ⇒ £k reduction in investment

25/45
 An Numerical Illustration

• Return of per unit of investment: R = 1.2

• Moral hazard problem
◦ probability of success when the entrepreneur behaves: pH = 1
◦ probability of success when the entrepreneur shirks: pL = 0.8
◦ private benefit from shirking: B = 0.08 per unit of investment
• Pledgeable income per unit of investment: 0.8
• 1 unit of investment must be at least financed by 0.2 of entrepreneur’s own funding
• Maximum leverage ratio equals 5.
• £1 reduction in the entrepreneur’s own funds ⇒ £5 reduction in investment

26/45
 Endogenous Credit Cycles (Cont’d)

• Formalising the idea by introducing market imperfections in dynamic models

◦ Bernarnke & Gertler (1989): costly state verification
◦ the importance of borrowers’ net worth
◦ Kiyotaki & Moore (1997): inalienable human capital
◦ the importance of collateral value and asset price
• The mechanics of endogenous credit cycles
◦ borrowers are financially constrained because of market frictions
◦ net worth and collateral help to relax credit constraints
◦ financial shock ⇒ firm net worth or collateral value goes down
◦ agency problem intensifies, and the borrowers can borrow less
◦ lower production, lower profit, and lower net worth/collateral value
◦ initial shock amplified by financial friction ⇒ a credit spiral

27/45
Amplification in Kiyotaki & Moore (1997)

28/45
 Kiyotaki-Moore (1997): Model Setup
• An economy of infinite horizon, t = 1, 2, 3, ...
• Economy populated by ‘farmers’ & ‘gatherers’, both groups of mass 1
• Asset: ‘land’ in fixed supply K, consumption good: ‘fruit’ produced by ‘land’
• Farmers (borrowers in the economy)
◦ produce with a CRS P technology: yt+1 = (a + c)kt
∞
◦ are impatient: U = t=0 β t xt
◦ a is tradable (can be used to purchase land), but c is not
• Gatherers (lenders in the economy)
◦ DRS technology: ẏt+1
P∞ = G(k̇t ), with G0 (0) = ∞ and G0 (K) < a + c
◦ more patient: U̇ = t=0 β̇ t ẋt , with β̇ > β
• There exists K0 s.t. G0 (K0 ) = a + c: K0 being the efficient land allocation.
• Asset market and (short-term) debt market
◦ spot price for 1 unit of ‘land’ at t: qt
◦ interest rate for debt R

29/45
 Financial Friction: Inalienable Human Capital
• Borrowing capacity determined by financial frictions
• Hart and Moore (1994): inalienable human capital
◦ borrowers’ human capital indispensable to the project
◦ output not pledgeable at all
◦ lenders require collateral to protect their investment
◦ seizing the collateral in the case of borrower renegotiation
• Total debt liability cannot exceed the value of collateral.

R · bt = qt+1 · kt

• This financial constraint is the driving force of the amplification mechanism.

• To get a flavor of the model, we will
◦ focus on the steady-state equilibrium (convergence),
◦ and focus on binding borrowing constraint (all tradable goods reinvested).
◦ both can be established in formal proofs but left out for simplicity.

30/45
 Solving for the Steady State
• Linearity of the model facilitates aggregation (holy grail in Macroeconomics)
• Farmers: financially constrained investment
◦ (binding) period borrowing constraint: RBt = qt+1 Kt
◦ the evolution of net worth: At = (a + qt )Kt−1 − RBt−1
◦ (binding) period budget constraint: qt Kt = At + Bt (similar to H-T)
• Borrowing & budget constraint ⇒ an investment multiplier
At At
Kt = = q (1)
ut qt − t+1
R

• ut captures the required downpayment, and 1/ut the leverage ratio.

• (Holmstrom-Tirole (97) model with variable investment has a similar feature.)
• Borrowing constraint & net worth evolution ⇒ At = aKt−1
• By equation (1), we know
aKt−1
Kt = q (2)
qt − t+1R

31/45
 Steady State of the Equilibrium
• Gatherers: maximizing the NPV by deciding on the capital holding
• Implication: indifferent between investment and asset sale
G0 (K − Kt ) qt+1
= qt − = ut
R R
• A steady-state equilibrium characterized by {q∗ , u∗ , A∗ , K ∗ , B∗ }.
• In a steady state, the endogenous variables do not change over time.
• Equilibrium asset price in the steady state
◦ Examining equation (2) in the steady state, we obtain
aR
q∗ = and u∗ = a.
R−1
◦ Asset pricing perspective: asset price = discounted present value
∞
X 1 G0 (K − K ∗ ) aR
q∗ = s
· G0
(K − K ∗
) = =
R R−1 R−1
s=1
• Farmer’s investment pinned down by G0 (K − K ∗ ) = aR
• Farmer’s net worth pinned down by A∗ = aK ∗
32/45
 Amplification in Kiyotaki & Moore (1997)
• Consider a negative productivity shock to a
◦ asset price decreases, because q∗ = R−1
R
a
◦ farmer’s investment decreases, because G0 (K − K ∗ ) = Ra
◦ farmer’s net worth decreases, because A∗ = aK ∗
◦ farmer’s debt capacity decreases, because B∗ = q∗ K ∗ /R
• Static amplification: Farmers’ investment drops by
1
>1
q − q/R
per unit of capital lost
• Persistence
◦ Constrained farmers do not make profitable investments
◦ Depressed future returns so At+1 &, At+2 &...
• Dynamic amplification through prices
◦ Bad news for productivity: qt &, qt+1 &...
◦ This further reduces capital At &, At+1 &...

33/45
 A Behavioural Perspective: Institutional Memory

• Berger and Udell (2004): Institutional Memory Hypothesis

• Loan officers’ abilities to screen/monitor loans can deteriorate as the painful
memory of the last crisis fades.
◦ ⇒ lower lending standard
◦ ⇒ fast loan growth
• The standard will be restored when the next crisis hits ⇒ credit cycles
• One possible driving force: the number of experienced loan officers
• Empirical test: regressing loan growth and spread on variable ‘TIME-ALLL’
◦ (TIME-ALLL: # years since the ratio of the allowance for loan and lease losses to total
loans (ALLL) was at its maximum over the prior 10 years. )
◦ a greater value = weaker institutional memory
◦ correlated with great loan growth
◦ and lower spread when loans were underwritten

34/45
Institutional Memory: Loan Growth

35/45
Institutional Memory: Loan Spread

36/45
 Procyclical Capital Regulations (Basel II)

• Procyclicality of capital regulation

• Under Basel 2, capital requirement vary with value of asset
◦ in boom, default risk decreases, lower risk weight, and less capital
◦ in bust, the opposite happens
• In an economic downturn
◦ in order to meet capital requirement
◦ banks substitute risky corporate lending by safe government bonds
◦ because it is expensive to raise new equity in recession (dilution)
◦ cut credit to real sector when they are most in need
• In economic boom
◦ capital requirement non-binding
◦ credit expansion, and accumulating risk (bubble or ineffective screening)
• ⇒ the need of counter-cyclical capital requirement

37/45
 Counter-cyclical Capital Requirement

• Addressing the pro-cyclicality of Basel II

• Basel III
◦ capital surcharge during credit booms
◦ counter-cyclical buffer within a range of 0-2.5%
◦ comprising common equity
◦ imposed when credit growth resulting in a build-up of systematic risk
• Rule-based vs. discretionary
◦ Basel III suggested a common reference point
◦ i.e., the deviation of credit-to-GDP ration from its trend
◦ acknowledging the possibility that the variable may give misleading signal

38/45
The financial crisis

39/45
 Bernanke (2018)
The Real Effects of Disrupted Credit

• Channels of effect
◦ “financial fragility”: panic in wholesale funding markets, credit supply
◦ “household leverage”: household deleveraging, credit demand
• External finance premium: cost of borrowing - return to safe asset
◦ Borrower’s net worth, collateral
◦ The financial health of banks

40/45
External finance premium

41/45
 Stages of financial crisis

• Stage 1: deflation of housing bubble, mortgage trouble

• Stage 2: liquidity pressure on financial institutions
• Stage 3: contagion to non-mortgage credit, wholesale funding run
• Stage 4: capital losses at banks and other lenders

42/45
Stages of financial crisis

43/45
Summary

44/45
 Take-home Messages

• The long-term impact of the banking sector on the macroeconomy

◦ the debate on banking/finance and economic growth
• The short-term impact of the banking sector on the macroeconomy
◦ amplification of business cycles
◦ small shock, large and long-lasting outcomes

45/45