Unconventional monetary policies: an appraisal by Claudio Borio* and Piti Disyatat* Bank for International Settlements Keynote lecture at the Money Macro.

Unconventional monetary policies: an appraisal

byClaudio Borio*

andPiti Disyatat*

Bank for International Settlements

Keynote lecture at the Money Macro and Finance Research Group 41st Annual Conference

Bradford University School of Management, 7-9 September 2009

* Claudio Borio and Piti Disyatat are, respectively, Head of Research and Policy Analysis and Senior Economist at the BIS. The views expressed are those of the authors and not necessarily those of the BIS.

2

Background and objective

The unprecedented monetary policy response to the financial crisis …

… has triggered a heated and at times confusing debate about “unconventional monetary policy” (UMP)

Objective

• provide a general framework to think about UMP

• highlight key issues concerning the transmission mechanism

• briefly consider key policy challenges

2

3

Classification framework

Distinguish

• interest rate policy (IRP) & balance sheet policy (BSP = UMP)

BSP = active use of central bank (CB) balance sheet to affect directly market prices & funding conditions beyond a short-term (overnight) interest rate

Classify BSPs in terms of two criteria:

• how they alter the structure of the private sector balance sheet

• the market segment explicitly targeted

3

4

Five substantive points BSPs are not unconventional in essence, but only because of the market

targeted • exchange rate policy !

BSPs can be decoupled completely from IRP (“decoupling principle”) • no such thing as a well behaved demand for monetary base (bank

reserves) as basis for setting interest rates !

Strong emphasis on bank reserve balances (monetary base) in the context of BSP is misplaced

• bank reserves are not special !

BSPs need to be viewed as part of consolidated public sector balance sheet

• loss of monopoly over MP (= IRP) and tricky issues about coordination and operational independence

BSPs can have a major impact on the financial risks incurred by CB • tricky issues about independence and credibility

4

5

Roadmap

Framework

• start with IRP (market for bank reserves & decoupling principle)

• lay out taxonomy for BSP

• apply the taxonomy

Transmission mechanism

• general channels

• focus on role of bank reserves

Policy challenges

• calibration and communication

• link with debt management

• financial risks absorbed by the central bank

• exit

5

6

IRP & the market for bank reserves

Until the crisis, monetary policy was largely synonymous with IRP

• announce policy rate and control tightly a market reference rate (overnight)

Done through two complementary mechanisms

• Signalling the desired interest rate

• Provide the corresponding amount of central bank funds (“liquidity management operations”)

Core of implementation of IRP is the market for bank reserves

• CB’s monopoly over supply underlies credibility of the signal

7

The decoupling principle

Principle: amount of bank reserves and interest rates can be set independently • same amount of bank reserves can coexist with many different levels of

interest rates (Scheme 1)• different amounts of reserves can coexist with the same level of interest rate

(Scheme 2) Scheme 1: penalty (or no) remuneration on excess reserves

• very inelastic demand curve; not well behaved• signalling coordinates expectations on target

• policy rate = market (overnight) rate > remuneration on excess reserves Scheme 2: remuneration at the policy rate sets opportunity cost to zero

• effectively a horizontal demand curve• policy rate = market (overnight) rate = remuneration on excess reserves

Reserve requirements do not change the picture

IRP often implemented without changing the size of the CB balance sheet

8

9

BSP: definition and key implications

BSP = actively use the CB balance sheet to affect directly broader financial conditions (beyond very short-term interest rate)

• eg, purchase of private sector assets

Core implications

• can be implemented regardless of level of interest rate (decoupling principle)

• need to insulate market for bank reserves• sterilise (Scheme 1)• or pay interest at policy rate (Scheme 2)

• not that unconventional ! • very common: exchange rate policy (Graph 1)• less conventional: market segment targeted

9

10

Graph 1: Central bank assets and liabilities In trillions of respective currency units

11

BSP: a taxonomy

Two criteria (Table 1) • change in composition of private sector balance sheets • market segment explicitly targeted

Exchange rate policy• change net exposure of private sector to FX risk• target exchange rate (level, volatility, etc.)

Quasi debt-management policy• change in the composition of public sector claims held by the private

sector• target yields on government securities

Credit policy• change in profile of private sector claims and/or composition of public

vs private sector claims• target financing conditions

Bank reserves policy• target private sector holdings of bank reserves, regardless of how

injected

11

12

Table 1: Typology of balance sheet policies

13

BSP: core features and implications Analytical distinctions rely on economic substance of the transactions

• transfer of ownership of cash flows vs. funding (eg repos)

Impact on the CB balance sheet depends on „financing“ • running down other claims vs. issuing liabilities• given limited control over broader financial conditions, may need to

transact in large quantities

May involve increasing the CB‘s intermediation role vis-à-vis the private sector

BSP needs to be considered through the lens of the consolidated public sector balance sheet

• largely mirror image of private sector’s

BSP can have a significant impact on the CB‘s absorption of financial risks

• FX, interest rate and credit risk

13

14

BSP: Applying the taxonomy

Credit easing• mix of quasi-debt management and credit policy

Quantitative easing• bank reserves policy plus (possibly) quasi-debt management and

credit policies

Evolution of BSP during the crisis (Table 2 and Graph 2))• phase 1 (pre-Lehman): credit policy in interbank market (mainly) • phase 2 (post-Lehman):

• broader credit policy (general) • quasi-debt management policy (Fed, BoE) • exchange rate policy (SNB, EMEs)• bank reserves policy (BoE)

• phase 2 saw major expansion of CBs’ balance sheet• and greater intermediation role for private sector

14

15

Table 2: Balance sheet policy during the crisis

16

Graph 2: Central bank assets and liabilities In billions of respective currency units

17

BSP: Transmission mechanism

Two channels

• signalling: announcement of operations• via expectations and beliefs (exclusively)

• broad portfolio: altering the structure of private sector balance sheets • via imperfect substitutability of different assets and liabilities

• portfolio balance effect• credit channel• risk-taking channel

Various types of policy rely differentially on these mechanisms (in addition to signalling)

• eg: quasi-debt management portfolio balance effect• monetary policy as debt management (Tobin/Friedman) !

• exchange rate policy portfolio balance effect

• credit policy credit and risk-taking channels

17

18

Are bank reserves special?

Thesis: too much attention placed on bank reserves in BSP• do not add potency to BSP

Uniqueness of bank reserves• means of final settlement (plus exogenous remuneration)• but BSP relies on other features

• attractive liquidity services (ease to obtain finance)• attractive yield

Short-term government paper is very close substitute in these respects (Table 3)

In other words: • bank reserves play special role in IRP because of settlement

services plus below market/penalty remuneration rate • CB needs to meet amount very precisely to control the overnight

rate (Scheme 1)

• to induce an expansion, need to make them perfect substitutes with other short-term default-free assets (Scheme 2)

• they are no longer special ! 18

19

Table 3: A comparison of bank reserves and other forms of sovereign claims

20

Bank reserves and credit creation Thesis: bank reserves in and of themselves are unlikely to boost credit

creation much

Why?

• notion of a stable money multiplier is misconceived• reserves never constrain lending

• amount is demand determined (Scheme 1)• additional reserves do not loosen lending constraints (Scheme

2)• except for change in liquidity composition of banks’ balance

sheet • but then similar to issuing more short-term paper !

• eg BoJ’s quantitative easing (Graph 3)

Ironically, if they have any effect, it is more likely when the transmission mechanism is broken

20

21

Graph 3: The Bank of Japan‘s quantitative easing

22

Money creation via direct purchases

Thesis: seeking to boost the money stock through direct purchases from the non-bank public is unlikely to have a significant effect on aggregate demand

Why?

• identity of initial acquirer makes little difference to final (equilibrium) stock of deposits

• amount depends on endogenous choices of banks and the public

• CB controls only what assets it buys and counterparts on its balance sheet

• final change in deposits is a poor signal of policy effectiveness (for given income, etc)

• impact on relative yields is largest for purchases of assets that are least substitutable for bank deposits

• but this minimises the impact on equilibrium deposits !

22

23

Is financing with bank reserves inflationary?

Thesis: there is nothing uniquely inflationary about bank reserves financing of BSP

Why?

• small extra effect over financing via short-term paper

Revisiting the notion of “monetisation” of deficits

• should distinguish

a) financing at a low interest rate

b) financing instrument (quasi-debt management operation)

• critical effect is via (a) not (b) • Scheme 1: would drive interest rates to zero !• Scheme 2: similar to short-term debt financing at the given

interest rate (decoupling principle)

confusion: prevailing view sees ↑ in bank reserves (monetary base) and ↓ interest rates as the dual of each other !

23

24

BSP challenge 1: calibration and communication

Issue

• little precedent (effectiveness and side effects)

• less control over relevant asset prices/yields

• disagreement over transmission mechanism• role of beliefs and expectations

Quasi-debt -management policy

• diminishing returns beyond „surprise effect“? (Graph 4)

Credit policy

• long-term effectiveness depends on CB’s catalyst role

• concerns with collateral damage• level playing field; market functioning

24

25

Graph 4: Central bank bond purchases

26

BSP challenge 2: government debt management policy

Issue

• all that the CB does can be, and is, done by the government

• balance sheet of the government is much larger

• objectives of debt management are different from those of MP

• eg minimise the cost of debt lengthen maturity in bad times to lock-in unusually low rates

CB options: degree of coordination

• none – full – in between

• CBs have made different choices

Effectiveness of policy depends on degree of subordination of debt management objectives

26

27

BSP challenge 3: financial risks for CB

Issue

• BSP can raise substantial financial risks for the CB

• could influence credibility and operational independence • via reputation• via loss of budgetary independence

Risks are country-specific

• accounting (market-to-market vs accrual)

• rules for profit distribution & recapitalisation

• broader institutional and political factors

28

BSP challenge 4: exit strategy

Issue

• decoupling principle allows CB to choose IRP from BSP independently

• main difficulty is choice of timing

Risks on both sides …

• too early? too late?

… but experience suggests that exiting too late is the bigger risk

• political economy pressures are overwhelmingly in this direction

29

Conclusion Monetary policy will probably never be the same again

• major experiment under way (UMP) Paper has

• provided a framework to classify and understand UMP

• explored analytical issues in the transmission mechanism

• highlighted key policy challenges Main takeaways

• BSP can be entirely decoupled from IRP• no such thing as a well behaved downward sloping demand for bank

reserves/monetary base !

• Bank reserves are not special in the context of BSP

• BSP can be properly understood only through lens of the consolidated public sector balance sheet

• UMP is only unconventional in terms of the markets targeted• exchange rate policy is quintessential BSP !

Effectiveness? Jury is still out.