Banks Are Not Intermediaries of Loanable Funds - Facts, Theory and - - PowerPoint PPT Presentation

SLIDE 1

Banks Are Not Intermediaries of Loanable Funds - Facts, Theory and Evidence

Zoltan Jakab, International Monetary Fund Michael Kumhof, Bank of England

Frankfurt, The Future of Money Conference, November 24, 2018

SLIDE 2

The views expressed herein are those of the authors and should not be attributed to the Bank of England or the International Monetary Fund.

SLIDE 3

1 Introduction: Banking Models in Economics

• Problem: Recent work uses intermediation of loanable funds (ILF) models.

— Banks are intermediaries between savers and borrowers of goods: ∗ Nonfinancial models. ∗ Banks = intertemporal commodity traders. ∗ Money = commodity money. — This theory misrepresents how credit is created in the real world.

• Solution: Use financing through money creation (FMC) models.

— Banks are intermediaries between spenders and spenders of money: ∗ Financial models. ∗ Banks = creators and intermediaries of money. ∗ Money = ledger entry money. — This theory is consistent with the actual credit creation process.

SLIDE 4

2 Understanding Banks: Key Insights

3.1 Banks are not Intermediaries of Loanable Funds 3.2 The “Deposit Multiplier” is a Myth

SLIDE 5

2.1 Banks are not Intermediaries of Loanable Funds

• The Loanable Funds Model - Postulated Credit Process

Intermediation = Physical Trading of Commodities — Banks collect a deposit of commodities or capital from a saver. — Banks lend those existing commodities to another agent, the borrower. — Deposits in this model are an input. — Money in this model is held as a store of value. — Rapid changes in credit: Switches between direct and indirect financing.

• The Financing Model — Actual Credit Process

Financing = Digital Creation of Monetary Purchasing Power — Banks make a loan of money to agent X. — Banks credit new money to the deposit account of the same agent X. — Deposits in this model are an output. — Money in this model is held as a medium of exchange. — Rapid changes in credit: Changes in gross balance sheet positions.

SLIDE 6

Intermediation of Loanable Funds Model Financing Through Money Creation Model

Bank Balance Sheet Saver

Deposit

• f

Goods

Bank Balance Sheet

Investor

Loan

• f

Goods Barter Loan

• f

Money Deposit

• f

Money

Investor

Monetary Exchange

Saver

SLIDE 7

• Why must ILF deposit-taking be a nonfinancial transaction?

— All financial transactions are variants of check or cash deposits.

• Check deposit:

— Households A and B bank with banks A and B. — B writes a check to A, A deposits in bank A. — Check only has value because the deposit already exists - in bank B. — This moves an existing deposit, it does not create a new one. — Also, bank A acquires reserves, not loanable funds. — The same logic applies to any deposits of private financial instruments.

• Central bank money is not loanable funds either:

— Central bank reserves cannot be lent to nonbanks, only to other banks. — Cash is never disbursed against new bank loans, only against existing deposits.

• New deposits in ILF models therefore do not represent financial transactions.
• Look at ILF budget constraints: They represent commodity accumulation.

SLIDE 8

• How is FMC deposit-creation a financial transaction?

— Loans are simultaneous ledger additions to assets and liabilities. — These ledger additions involve no intermediation. — Loan = right of bank to receive future installments from X. — Deposit = obligation of bank to deliver current funds to X. — Magic of banking: The obligation itself is current funds = money. — Banks are unique in their ability to do this. — Why? Because they are perceived to be safe. — Why? Mostly because of public support.

SLIDE 9

2.2 The “Deposit Multiplier” is a Myth

• Deposit Multiplier:

— Central bank fixes narrow money first. — Broad money is a function of narrow money.

• Kydland and Prescott (1990) showed that the actual monetary transmission

mechanism works in the opposite direction. — Broad money leads the cycle. — Narrow money (M0) lags the cycle.

• This is obvious under Inflation Targeting:

— If you control a price (the interest rate), ... — then you have to let quantities (reserves) adjust.

SLIDE 10

2.3 Understanding Banks: Conclusions

• Transmission starts with loan creation = deposit creation, and ends with

reserve creation.

• Alan Holmes, Vice President of the New York Federal Reserve, 1969:

In the real world, banks extend credit, creating deposits in the process, and look for the reserves later.

SLIDE 11

3 Key Features of Our Financing Model

• 1. Bank Assets: The Provision of Credit.
• Banks do not lend out pre-existing loanable funds.
• There are no loanable funds:

— Funds first exist in the mind of the banker. — They then materialize (digitally) along with the loan.

• 2. Bank Liabilities: Households Demand Bank Deposits.
• Bank deposits are not real savings.
• Banks do not collect deposits from non-banks:

— They create deposits for non-banks. — They collect deposits from each other.

• 3. Bank Equity: Subject to Basel regulation and aggregate risk.

SLIDE 12

4 The Models

• Two Models: One loanable funds and one financing model.
• Except for the loanable funds - financing difference, models are identical:

— New Keynesian monetary models. — Identical preferences, technologies, endowments. — Identical deterministic steady states. — Every single parameter (including adj. costs) is identical.

• We are therefore, as much as possible, comparing apples with apples.

SLIDE 13

Key Difference ILF-FMC: Budget Constraints

• Budget Constraints in ILF: Saver Household + Borrower Entrepreneur

— Saver Household ∆depositss

t = incomes t − spendings t

— Borrower Entrepreneur −∆loansb

t = incomeb t − spendingb t

• Budget Constraint in FMC: Representative Household only

∆depositsr

t − ∆loansr t = incomer t − spendingr t

Deposits and loans are predetermined variables Deposits and loans are jump variables

As we will see, this is highly favored by the data

SLIDE 14

5 Model Impulse Responses to Financial Shocks

SLIDE 15

Figure 3. Impulse Responses: Credit Crash due to Higher Borrower Riskiness

• 0.6
• 0.4
• 0.2

0.0

• 0.6
• 0.4
• 0.2

0.0 2 4 6 8 10 12 14 16

GDP

(% Difference)

• 0.5
• 0.4
• 0.3
• 0.2
• 0.1

0.0

• 0.5
• 0.4
• 0.3
• 0.2
• 0.1

0.0 2 4 6 8 10 12 14 16

Real Policy Rate

(pp Difference)

• 0.3
• 0.2
• 0.1
• 0.0

0.1

• 0.3
• 0.2
• 0.1
• 0.0

0.1 2 4 6 8 10 12 14 16

Consumption

(% Difference) 0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5 2 4 6 8 10 12 14 16

Real Retail Lending Spread

(pp Difference)

• 3
• 2
• 1
• 3
• 2
• 1

2 4 6 8 10 12 14 16

Investment

(% Difference)

• 4
• 3
• 2
• 1
• 4
• 3
• 2
• 1

2 4 6 8 10 12 14 16

Bank Loans

(% Difference)

• 0.3
• 0.2
• 0.1

0.0

• 0.3
• 0.2
• 0.1

0.0 2 4 6 8 10 12 14 16

Inflation

(pp Difference)

• 4
• 3
• 2
• 1
• 4
• 3
• 2
• 1

2 4 6 8 10 12 14 16

Bank Deposits

(% Difference) 0.0 0.2 0.4 0.6 0.0 0.2 0.4 0.6 2 4 6 8 10 12 14 16

Effective Price of Consumption

(% Difference)

• 4
• 3
• 2
• 1
• 4
• 3
• 2
• 1

2 4 6 8 10 12 14 16

Bank Net Worth

(% Difference) 0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5 2 4 6 8 10 12 14 16

Effective Price of Investment

(% Difference)

• 0.1

0.0 0.1 0.2 0.3 0.4

• 0.1

0.0 0.1 0.2 0.3 0.4 2 4 6 8 10 12 14 16

Bank Leverage Ratio

(Difference)

• - - = ILF Model, –— = FMC Model

Credit Crash due to Higher Borrower Riskiness

Financing Model: GDP drop is far larger Financing Model: Positive comovement

• f C and I

Financing Model: Far larger contraction in lending Financing Model: Much smaller increase in spreads Financing Model: Bank leverage is procyclical as lending contraction dominates net worth reduction

SLIDE 16

6 Stylized Facts and Related Empirical Literature

Simulations have generated three interrelated predictions for the financing model versus the loanable funds model:

• 1. Large and rapid changes in financial sector balance sheets.
• 2. Bank leverage is procyclical or acyclical.
• 3. Credit crashes have a large quantity rationing component.

SLIDE 17

6.1 Large and Rapid Changes in Financial Sector Balance Sheets

SLIDE 18

Figure 7. Bank Balance Sheets: Time Series Evidence - US/EU/GER/FRA

• 8
• 6
• 4
• 2

2 4 6 8

• 8
• 6
• 4
• 2

2 4 6 8 dlog(Assets), % dlog(Equity), % dlog(Debt), % Slope -0.08 (0.42) Slope 1.11*** (0.00) United States (90Q1-16Q4)

• 4
• 2

2 4 6

• 4
• 2

2 4 6 dlog(Assets), % dlog(Equity), % dlog(Debt), % Slope 0.22 (0.08) Slope 1.01*** (0.00) Eurozone (97Q3-10Q3)

• 4
• 2

2 4

• 4
• 2

2 4 dlog(Assets), % dlog(Equity), % dlog(Debt), % Slope 0.42** (0.01) Slope 0.87*** (0.00) Germany (97Q3-10Q3)

• 4
• 2

2 4 6 8

• 4
• 2

2 4 6 8 dlog(Assets), % dlog(Equity), % dlog(Debt), % Slope 0.24** (0.02) Slope 1.18*** (0.00) France (97Q3-16Q4)

Aggregate banking system assets, debt and equity. Quarter-on-quarter % changes. Data: Flow-of-funds. Each point represents one quarter. Sample sizes shown in text. p-values of regression slopes in brackets.

53

Bank Balance Sheets: Time Series Evidence for 4 Regions

Bank assets and bank debt move virtually one-for-one The balance sheet changes are often extremely large

SLIDE 19

Figure 6: Physical Saving (SAV)

• 15
• 10
• 5

5 10 15 20 90 92 94 96 98 00 02 04 06 08 10 12 14 16 SAV/GDP, % Change in Credit to Nonfinancial Business/GDP, % d(LOANS&SEC)/GDP, % INV/GDP, % United States (90Q2-16Q4)

• 8
• 4

4 8 12 16 00 02 04 06 08 10 12 14 16 SAV/GDP, % Change in Credit to Nonfinancial Business/GDP, % d(LOANS&SEC)/GDP, % INV/GDP, % Eurozone (97Q4-16Q4)

• 6
• 4
• 2

2 4 6 8 10 12 03 04 05 06 07 08 09 10 11 12 13 14 15 16 SAV/GDP, % Change in Credit to Nonfinancial Business/GDP, % d(LOANS&SEC)/GDP, % INV/GDP, % Germany (03Q2-16Q4)

• 8
• 4

4 8 12 16 20 03 04 05 06 07 08 09 10 11 12 13 14 15 16 SAV/GDP, % Change in Credit to Nonfinancial Business/GDP, % d(LOANS&SEC)/GDP, % INV/GDP, % France (03Q2-16Q4)

Changes in Bank Balance Sheets versus Net Private Saving

Changes in bank balance sheets are extremely large and volatile Net private saving is very smooth by comparison Net private saving is also typically of very different size

SLIDE 20

(Data: Flow of funds. Quarterly. Based on stock data for d(LOA NS& SEC). All variables divided by the sam e quarter’s G D P.) (DFS/G DP for U S: Securities issued by Nonfinancial Business. DFS/G DP for EUR/G ER/FRA : Securities issued by O ther Residents.)

Figure 9: VAL+DFS for Nonfinancial Business Issuers Only

• 20
• 15
• 10
• 5

5 10 15 20 90 92 94 96 98 00 02 04 06 08 10 12 14 16 d(LOANS&SEC)/GDP, with corporate bonds, % (DFS+VAL)/GDP, with corporate bonds, % d(LOANS&SEC)/GDP, with equities and corporate bonds, % (DFS+VAL)/GDP, with equities and corporate bonds, % United States (90Q2-16Q4)

• 4
• 2

2 4 6 8 10 12 98 00 02 04 06 08 10 12 14 16 d(LOANS&SEC)/GDP, with securities other than shares, % (DFS+VAL)/GDP, with securities other than shares, % Eurozone (97Q4-16Q4)

• 4
• 3
• 2
• 1

1 2 3 4 03 04 05 06 07 08 09 10 11 12 13 14 15 16 d(LOANS&SEC)/GDP, with securities other than shares, % (DFS+VAL)/GDP, with securities other than shares, % Germany (03Q2-16Q4)

• 4

4 8 12 03 04 05 06 07 08 09 10 11 12 13 14 15 16 d(LOANS&SEC)/GDP, with securities other than shares, % (DFS+VAL)/GDP, with securities other than shares, % France (03Q2-16Q4)

(Data: Flow of funds. Quarterly. B ased on stock data for all series. A ll variables divided by the sam e quarter’s G D P.) ((VAL+D FS)/G D P: Securities issued by N onfinancial Business for all countries in the sam ple.)

Valuation effects and direct financing substitution are very small compared to total balance sheet changes

Changes in Bank Balance Sheets versus Valuation Effects and Direct Financing Substitution

SLIDE 21

6.2 Procyclical Bank Leverage

• Nu˜

no and Thomas (2012): — Study comovement between cyclical components of U.S. bank leverage and aggregate output. — Commercial banks are acyclical, shadow banks are strongly procyclical.

• Our Analysis:

— Follows Nu˜ no and Thomas (2012). — In addition: ∗ Takes account of lags of output. ∗ Studies several other countries.

SLIDE 22

Table 2: Correlation of Financial Sector Leverage and GDP in Four Economies

Cross-correlation between cyclical components of logarithm of lagged GDP and leverage ratio (with assets = cumulated flows)

US Regulated US Shadow US Regulated + Shadow EUR GER FRA Lags 90:1 - 16:4 90:1 - 16:4 90:1 - 16:4 97:3 10:3 97:3 10:3 97:3 16:4 0.18* 0.66*** 0.53*** 0.40** 0.24 0.40*** 1 0.20* 0.65*** 0.50*** 0.56*** 0.33** 0.54*** 2 0.19* 0.50*** 0.38*** 0.62*** 0.33** 0.60*** 3 0.16 0.29*** 0.23** 0.56*** 0.25 0.58*** 4 0.15 0.08 0.10 0.41*** 0.14 0.48*** 5 0.13

• 0.07

0.00 0.21 0.03 0.33***

(D ata: Flow of funds. Q uarterly.) (* = Significant at 10% confidence level, ** = Significant at 5% confidence level, *** = Significant at 1% confidence level)

Strongly procyclical credit

SLIDE 23

7 Conclusions

• Key Contributions of This Paper:
• 1. Theory: Loanable funds models of banking are not a correct represen-

tation of the real-world credit/money creation process. The objective of financing models is to fix that.

• 2. DSGE Model Comparison: Financing models have very different simu-

lation properties. — Far larger and far faster changes in bank lending. — Much smaller changes in spreads. — Much larger effects on the real economy.

• 3. Stylized Facts: Financing models are consistent with key stylized facts.

— Large discontinuous jumps in credit and money. — Procyclical bank leverage. — Credit rationing during downturns.

SLIDE 24

8 Looking Ahead

• The topic of this conference is “The Future of Money”.
• To understand various monetary reform proposals one has to first understand

“The Presence of Money”.

• And that means the present modus operandi of banks and the central bank.
• The ILF model offers a misleading starting point.
• The FMC model aims to fix that.
• Many important implications.
• Example: Ability of banks to engage in procyclical lending:

— Very high in the FMC model. — Much more modest in the ILF model.

SLIDE 25

Thank you!