Financial Markets and Society

NOVEMBER 2002

Rebuilding the Transmission System for Monetary Policy BY JANE D’ARISTA

I. Introduction and Overview After enjoying a period of unprecedented adulation during the 1990s, the Federal Reserve and its policy judgments have come under increased criticism as the recovery of 2002 sputters and the economic outlook grows more uncertain. Some observers blame the central bank for incautiously cheering on the “new economy” of the late 1990s, while others fault it for needlessly letting the air out of stock prices. Or for failing to damp stock prices effectively. Or for mis-reading the evidence of the 2001 recession as it unfolded. Or for paying too little attention to the dangers posed by deflationary pressures.

Because they emanate from a broad spectrum of interests and perspectives, these objections naturally offer somewhat conflicting views of the Fed’s actions. Yet in the aggregate, they represent a healthy attempt to sort out the macroeconomic performance and policy lessons of the 1990s. What’s missing from virtually all the critiques, however, is an appreciation of how the Federal Reserve’s policy judgments are shaped and limited by the policy tools it uses. Since the late 1930s, relatively few changes have occurred in the way the Fed implements monetary policy. Open market operations – first used in the 1920s and formally acknowledged as a policy tool by the Banking Act of 1933 – have for many years remained the primary means for changing the price and supply of money and credit.

In the past, these transactions provided a highly effective transmission belt for monetary policy. But while the tools themselves remain in place, the institutional context in which they are employed has changed dramatically. These changes include a relative decline in the banking sector’s role in credit creation, reductions in the level of required reserves and the ascendancy of market forces – particularly unrestricted international capital flows – in determining the pace of credit expansion in a world of increased financial integration. This paper assesses the impacts of these far-reaching changes on the process of monetary policy implementation. It argues that the failure to modernize aging policy tools has deeply damaged the Fed’s ability to influence credit expansion – a task embedded in the Federal

INSIDE The Fed at a Crossroad ........................................................................................................ 11 Open Market Operations: A New Approach .............................................................. 22

Reserve’s mission and critical to the success of macroeconomic policy. Unrestricted credit expansion has, in turn, fueled unsupportable debt levels for households, businesses and financial institutions. And unchecked credit growth helped inflate asset-market bubbles. Moreover, weakened policy tools now hamper the Fed’s ability to implement effective countercyclical policies in the event of a significant downturn. In sum, the erosion of its monetary transmission mechanism has put the central bank in the position of inadvertently contributing to, rather than combating, economic imbalances and financial instability. The paper also contends that this underlying problem can be remedied by introducing a new system of reserve management that assesses reserves against assets rather than deposits. This new approach would enhance

monetary control by applying reserve requirements to all segments of the financial sector and increasing the Fed’s ability to respond effectively to credit expansions and contractions resulting from international capital flows.1

II. The Slipping Transmission Belt for Monetary Policy Institutional Change Remakes the Financial Sector Over the past quarter-century, the U.S. financial system has undergone a transformation, as household savings shifted from banks to pension funds and other institutional investment pools. Between 1981 and yearend 2001, the assets of all depository institutions plummeted from nearly half of total financial-sector assets to 24 percent.

OPEN MARKET OPERATIONS The Fed conducts open market operations by buying and selling U.S. Treasury securities in the secondary market through transactions with so-called “primary dealers.” These purchases and sales simultaneously change both the asset and liability sides of the Fed’s balance sheet, adding to or subtracting from both its holdings of assets (government securities) and its liabilities for bank reserves. Under a fractional reserve system, banks (and, since enactment of the 1980 Monetary Control Act, all other depository institutions) must hold a designated percentage of their deposits either as cash in their vaults or as non-interestbearing balances with their regional Federal Reserve Banks. The requirement to hold

reserves is the basic policy tool for any fractional reserve system and the fulcrum for open market operations. When the Fed changes the supply of reserves, it changes the volume of cash and deposits in the system (see Figure 3 on page 21). And since banks create deposits for their customers when they lend, constraining or augmenting their ability to add to their deposit base affects

their ability to expand credit. Open market operations also affect financial institutions that are not subject to reserve requirements. The Fed’s actions in supplying or acquiring cash in exchange for government securities serves to increase or reduce liquidity for the system as a whole, changing prices and conditions in U.S. money and credit markets and influencing economic activity.

DEFINITIONS Primary dealers: a group of commercial and investment banks that make markets in U.S. government securities and are designated by the Fed to act as its counterparts in open market operations. The Fed exercises a degree of supervision over their operations and acts as an ongoing lender in financing their inventories of securities through repurchase agreements. Fractional reserve system: any arrangement that requires depository institutions to hold assets equal to a given percentage of their deposits as reserves. The original intention of this system in the U.S. was to ensure the availability of an adequate amount of liquid assets in the event of deposit withdrawals.

1. For additional discussions of proposals for and experiences with asset-based reserve requirements, see U.S. House of Representatives 1972, Pollin 1993, and Palley 2000.

2

FINANCIAL MARKETS CENTER

Table 1: Shares of Financial Sector Assets ($ billion) 1951

1961

1971

1981

1991

2001

Depository Institutions $203.2 Insurance Companies 72.9 1 Pension Funds 20.4 Mutual Funds 5.7 2 Nonbank Lenders 10.1 GSEs & Federally Related Mortgage Pools 3.7 Security Brokers & Dealers 3.8 Others3 —

$378.7 131.9 91.9 29.0 30.9 13.0 7.6 0.2

$891.9 227.0 260.5 63.2 56.8 78.9 17.6 158.7

$2,518.3 495.0 951.1 253.2 237.8 361.6 59.8 263.9

$4,820.8 1,405.9 3,343.4 1,375.7 620.0 1,653.3 332.5 1,080.1

$8,610.3 2,392.7 7,096.1 6,510.0 1,192.9 5,130.9 1,465.7 3,370.5

Depository Institutions Insurance Companies Pension Funds1 Mutual Funds Nonbank Lenders2 GSEs & Federally Related Mortgage Pools Security Brokers & Dealers Others3

Percentage of Total Financial Sector Assets 55.4 50.8 49.0 32.9 19.3 12.9 9.6 9.6 13.5 14.8 18.5 22.9 4.2 3.6 4.9 9.4 4.5 3.2 4.6 4.2 1.9 4.5 7.0 11.3 1.1 1.0 1.2 2.3 — 9.0 5.1 7.4

63.5 22.8 6.4 1.8 3.2 1.2 1.2 —

24.1 6.7 19.8 18.2 3.3 14.3 4.1 9.4

1

Includes insured pension assets Includes finance companies and mortgage companies 3 Includes bank personal trusts and estates, asset-backed securities issuers, real estate investment trusts and funding corporations SOURCE: Federal Reserve System, Flow of Funds Accounts of the United States 2

Meanwhile, spurred in part by the funding requirements of the Employee Retirement Income Security Act (ERISA) of 1974, the assets of pension funds and mutual funds soared from 23 to 38 percent of financial sector assets. These institutional investment pools now provide the dominant channels for saving and investment flows. At yearend 2001, pension funds held $7.1 trillion of financial assets (including equities) and mutual funds’ holdings of money market instruments, stocks and bonds totaled $6.5 trillion. By contrast, the total assets of commercial banks, savings institutions and credit unions amounted to $8.6 trillion (Board of Governors, Flow of Funds). The shift in individual savings from banks to pension and mutual funds also produced a symmetrical increase in borrowing through capital markets, since securities constitute the primary assets held by institutional investors. As these investors increased their demand for

credit market instruments, corporations borrowed less from banks and issued substantially more bonds and commercial paper. Credit flows to individuals also moved into the capital markets as government-sponsored enterprises (GSEs) such as Fannie Mae and Freddie Mac and federally related mortgage pools securitized more mortgages and assetbacked securities (ABS) issuers used securitization techniques to fund car loans and other consumer receivables. Between 1991 and 2001, the liabilities of GSEs and mortgage pools – used mostly to finance single-family housing – rose by $3.6 trillion to $5.1 trillion, an increase of 240 percent. During the same period, assets of ABS issuers jumped 530 percent, soaring to $2.1 trillion (Board of Governors, Flow of Funds). Given these changes in the instruments and channels used for saving and investment, it is not surprising that the activities of institutions FINANCIAL MARKETS CENTER

3

have changed as well. Asset management has become the dominant function in U.S. financial markets and trading has become the principal activity. Bank lending remains important, particularly to small business borrowers that lack access to capital markets. But banks, too, manage mutual funds and offer asset-management services through their trust departments. And since passage of the 1999 Gramm-LeachBliley Act, larger banks have expanded their securities, asset-management and insurance operations through financial holding companies. The biggest U.S. banks have extended their activities beyond traditional lending in other ways as well. Since the 1970s these institutions have been the dominant foreign-exchange market makers. And more recently, they have developed a highly profitable niche – supported by their special relationship with the lender-oflast-resort2 – providing financial insurance as dealers in derivatives and sellers of committed lines of credit to back issues of commercial paper and other securities. Policymakers have long recognized the influence of these institutional changes on the conduct of monetary policy. In 1993, for example, Fed Chairman Alan Greenspan told attendees of the Kansas City Fed’s Jackson Hole conference, “the fairly direct effect that open market operations once had on the credit flows provided for businesses and home construction is largely dissipated” due to the diminished role of banks, the increase in savings channeled through institutional investors and the growth of securitization. Though Greenspan asserted that “the Federal Reserve can still affect short-term interest rates, and thus have an impact on the cost of borrowing from banks, from other intermediaries, and directly in the capital

markets,” he acknowledged that “this effect may be more indirect, take longer, and require larger movements in rates for a given effect on output” (Greenspan 1993).3 Subsequent events have underscored the accuracy of the chairman’s remarks. During the 1990s, the Fed kept real interest rates high in a noninflationary environment but its policy failed to moderate the rapid rise in stock prices, credit or GDP in the final years of the decade. In 2001, the central bank reversed course and undertook an aggressive easing campaign. Yet more than a year later, this effort had still not produced the intended impact on output, despite the injection of substantial fiscal stimulus.

The Erosion of Quantity Controls At the same time innovation and restructuring were transforming the financial industry, lawmakers and regulators were dismantling the quantity controls that once constituted a key feature of financial systems in the U.S. and other countries. Quantity controls include interest rate ceilings on deposits, international capital controls, liquidity and reserve requirements and direct limits on credit expansion. Historically, these “macroprudential” policy tools have contributed to financial stability by systemically restraining credit expansion and by promoting the soundness of individual institutions. From the perspective of financial firms, however, limiting the amount of credit they can extend to customers also restricts their opportunities for profit. During the 1970s and 1980s, offshore banking activity grew explosively as U.S. depository institutions sought the higher returns available in external markets lacking reserve requirements and other restraints on lending. The rise of the unregulated Eurodollar market gave the Fed and other central banks a

2. The size and importance of over-the-counter foreign exchange and derivatives markets and the dominant role of large U.S. banks in these markets ensure that the Federal Reserve will intervene to support the banks’ derivatives positions and financial guarantees. (The banks’ enormous off-balance sheet commitments relative to capital further lock the central bank into a supportive stance.) When the bond and derivatives markets seized up and the dollar fell 17 percent against the yen during the last quarter of 1998, the Fed fulfilled its guarantor-of-the-guarantors role by flooding the market with liquidity. 3. Taking a gloomier view of these developments, Bundesbank Vice President Hans Tietmeyer told the Jackson Hole conferees,“changes in the financial markets have generally made it more difficult for monetary policymakers to fulfill their stability mandate…In a number of countries, financial innovation and deregulation have distorted the intermediate targets used in the conduct of monetary policy and have altered the transmission mechanisms for monetary policy to the real economy”(Tietmeyer 1993).

4

FINANCIAL MARKETS CENTER

powerful incentive to relax or remove quantity controls in order to moderate the shift in deposits and loans from domestic to external markets. In the U.S. and other industrialized countries, quantity controls also fell prey to an increasingly potent political movement favoring financial liberalization – and to the accompanying ideological argument that free markets allocate credit (and, in general, shape economic outcomes) more efficiently and productively than do governments. 4 After experimenting with direct limits on bank lending in an effort to manage capital outflows in the 1960s, the U.S. abolished capital controls in 1974. Six years later, the Monetary Control Act ended Regulation Q interest rate ceilings on depository institution accounts. Despite moving decisively in the direction of deregulation, the U.S. has not eliminated reserve requirements – the primary quantity control used by the Federal Reserve and the tool most critical to implementing monetary policy. However, reserve requirements have been seriously weakened. In response to international competitive pressures on U.S. banks during the late 1980s, the Fed supported a zeroing out of reserve requirements on time deposits in order to make the cost of domestic CDs comparable to Eurodollar sources of funds.5 The central bank also lowered reserve requirements on demand deposits from 12 percent to 10 percent, effective as of 1992. Simultaneously, the Fed played a key role in negotiating the Basle Accord, which set capital adequacy standards for all multinational banks located in the G-10 countries. By establishing

those standards at levels previously applied to U.S. banks, the Basle agreement eliminated a key competitive advantage enjoyed by non-U.S. depository institutions. Moreover, the Basle standards represented the kind of prudential requirement acceptable to liberalization proponents, since they ensure that market forces – i.e., the providers (or withholders) of capital – determine the amount of loans banks can extend. Over time, the Basle capital standards tended to displace rather than complement the role of reserve requirements. And this displacement created problems for monetary control. Since markets inevitably supply more capital during a boom and less during a downturn, capital requirements tend to impose a procyclical bias on bank lending. Thus, by exacerbating the effects of increased loan-loss provisions in a downturn without imposing any systemic constraints on a lending boom, the Basle regime compounds the challenges central banks face when they attempt to fashion effective countercyclical policy. In addition, the eclipse of reserve requirements by capital standards tends to weaken the link between central bank actions and financial institutions’ behavior. By reducing the leverage of the fractional reserve system, the shift from reserve to capital requirements – like the more overt removal of older-generation quantity controls – requires the Fed to engineer larger additions and contractions of liquidity to achieve its target short-term interest rate. These actions translate into greater volatility in the fed funds market and heighten risk by creating uncertainty about the availability and cost of funds.

4. Calls on developing countries to remove capital controls fostered controversy throughout the 1990s. However, many industrial nations maintained these controls throughout the Bretton Woods Era and abandoned them at a relatively late date. Moreover, other quantity controls long endured as standard tools of monetary management in many industrial countries. For example, direct limits on lending remained a primary monetary tool for the U.K. and other European economies into the 1970s (see footnote 29). European countries and Japan also extensively used regulations that required banks to match the maturity of assets and liabilities. Only by the end of the 1980s did all the EU nations undertake capital account liberalization to fulfill conditions for monetary union. 5. Ironically, the Fed’s attempt to improve monetary control in the 1980s contributed to the scaling back of requirements in the 1990s. In 1979, the central bank imposed reserve requirements on loans by U.S. banks’ foreign branches to their home offices. However, the Fed’s inability to establish parallel reserve requirements on non-U.S. banks operating offshore gave these foreign institutions a cost advantage that they used inter alia to gain a sizable share of the U.S. commercial and industrial loan market (McCauley and Seth 1992). These market-share gains occurred at a time when non-performing loans to developing countries were battering American banking industry profits, compounding the pressure on authorities to aid U.S. banks by easing their regulatory costs. FINANCIAL MARKETS CENTER

5

Reserve Balances Dwindle As the role of reserve requirements continued to weaken, banks’ reserve balances with the Fed fell from peak levels of $35-40 billion in 1986-1989 to $7 billion by August 2002. In large measure, this steady decline mirrors the fact that the Fed has been applying its lowered reserve requirements to a proportionately dwindling universe of liabilities. During the past half-century, depository institutions’ share of financial sector liabilities declined in tandem with their share of assets. And the relative diminution in deposit growth has been particularly acute. While financial sector liabilities expanded by more than 150 percent between 1991 and 2001, checkable deposits increased by a comparatively paltry 37 percent – with virtually all the growth occurring at savings institutions and credit unions (commercial banks’ checkable deposits

grew by less than one percent over the course of this period). Among other things, these trends reflect banks’ growing proclivity to sweep customers’ deposits from transaction accounts into time and money market accounts that are not subject to reserve requirements. From January 1994 to July 2002, sweep accounts rose from $5 billion to $501 billion (see Table 3) – a massive reduction in reservable liabilities that the Fed sanctioned after years of hearing bankers complain about the cost of holding non-interest-bearing reserve balances. The Fed also provided implicit approval of banks’ increased use of borrowed funds and repurchase agreements (repos), a practice that has further slowed deposit growth.6 By using existing assets as collateral for borrowed funds and to conduct repos, banks can leverage up their loans – and increase income – restrained

Table 2: Shares of Financial Sector Liabilities ($ billion) 1951

1961

1971

1981

1991

2001

Financial Sector Liabilities Depository Institution Liabilities Depository Institution Deposits Checkable Deposits Reserve Balances

$291.8 188.2 180.4 103.6 20.3

$621.1 349.3 328.6 131.0 17.3

$1,667.8 845.4 756.3 206.4 25.8

$5,017.7 2,439.0 1,955.8 373.2 26.3

$14,564.7 4,749.1 3,566.5 727.4 26.7

$36,993.4 8,491.5 5,247.0 998.3 9.1

MEMO ITEMS: Depository Institutions Liabilities as Pct. Financial Sector Liabilities

64.5%

56.2%

50.7%

48.6%

32.6%

23.0%

Depository Institution Deposits as Pct. Financial Sector Liabilities

61.8%

52.9%

45.3%

39.0%

24.5%

14.2%

Reserve Balances as Pct. of Depository Institution Deposits*

11.3%

5.3%

3.4%

1.3%

0.6%

0.2%

* Includes credit union and thrift deposits for years prior to Monetary Control Act (MCA) when reserve requirements were imposed only on banks. If reserve balances were calculated as a percentage of bank deposits only for 1951, 1961 and 1971, the results would be 14.3 percent, 8.1 percent and 5.3 percent respectively. SOURCE: Federal Reserve System, Flow of Funds Accounts of the United States

6. Between 1971 and 2001, deposits shrank from 90 percent to 62 percent as a share of total bank liabilities. The use of repos as “pseudo deposits” played a major role in this development: by yearend 2001, federal funds and security repurchase agreements of commercial banks in the U.S. stood at $788 billion, up 243 percent from the level ($230 billion) in 1991. These liabilities – which consist almost entirely of repos – outstripped checkable deposits ($627 billion) and equaled 20 percent of banks’ total deposits ($4.02 trillion) at yearend 2001.

6

FINANCIAL MARKETS CENTER

only by market forces, not monetary control. Like the movement to capital requirements, these techniques increase the financial system’s procyclical bias, providing greater access to funds at lower cost in a boom and less access at higher cost in a downturn. And, like other practices that rely on market forces to limit the supply of credit, they undermine the Fed’s capacity to implement anti-cyclical initiatives. As deposit growth declined and reserveevasion techniques became more prevalent, depository institutions have found it increasingly easy to meet their (more lenient) reserve requirements by holding vault cash rather than placing reserves with the Fed – the final contributing factor to the steep drop in reserve balances. Indeed, aided by the rise of ATMs, banks now satisfy more than four-fifths of their reserve requirements with holdings of vault cash, compared to 53 percent in 1991. With the shrinkage in reserve balances, Federal Reserve assets now provide coverage primarily for currency in circulation – the other major component of the central bank’s liabilities (see Figure 1(b) on page 19 for a simplified presentation of the Fed’s balance sheet).7 As banks’ reserve balances shrink, the Fed has tried to compensate by altering some of its operating procedures. Since it began announcing its target federal funds rate in 1994, the Fed has increasingly relied on the announcement itself – as opposed to actual open market operations – to implement policy decisions. In executing “open mouth operations,” the central bank essentially waits for market participants to adjust to the new rate level, based on their belief that the Fed can enforce the rate.8 In addition, the Fed has been forced to rely more on overnight repurchase agreements to implement policy. Between 1998 and 2000, overnight repos accounted for 58 percent of

Table 3: Composition of Depository Institution1 Liabilities

Type of Liability

Deposits Checkable Nontransaction Large time Other Sweep accounts Security RPs Credit market instruments Net interbank liabilities Other2 MEMO ITEMS: Sweep accounts as pct. total deposits Sweep accounts as pct. checkable deposits

Pct. all Depository Institution Liabilities 1994

Q2 2002

65.1 16.3 48.8 7.7 41.1 0.2 8.2 6.4 4.0 16.3

62.8 11.7 51.1 13.8 37.3 5.8 9.3 9.9 0.7 17.0

0.3

9.2

1.1

49.3

1

Includes U.S.-chartered commercial banks, foreign banking offices in U.S., bank holding companies, banks in U.S.affiliated areas, savings institutions and credit unions 2 Includes taxes payable, miscellaneous liabilities SOURCES: Federal Reserve System, Flow of Funds Accounts of the United States; Sweeps of Transaction Deposits into MMDAs

Fed transactions, up from 26 percent in 19941998 and ten percent before that (FRB New York 2001). These developments do not necessarily mean the Fed’s ability to enforce its target rate has atrophied.9 Indeed, at least one Federal Reserve economist argues that interest rates now respond more rapidly to policy shifts as a

7. At the end of August 2002, currency (including cash held outside U.S. borders) constituted 93 percent of total Federal Reserve liabilities (Board of Governors 2002). 8. Previously, the term “open mouth policy” was used to express the view that a central bank does not need to act; that it need only announce its intentions (Goodhart 2000). However, as Benjamin Friedman points out,“stated intentions matter only if there is something credible to back them up – and…whether the central bank can or cannot back up its intentions is a matter of institutional arrangements, subject to change”(Friedman 2000). 9. It could be argued that increased use of repos by the Fed strengthens its ability to influence short-term interest rates given the expanded use of repos for short-term borrowing and investment by virtually all segments of the financial industry. FINANCIAL MARKETS CENTER

7

result of the central bank’s FEDERAL RESERVE announcement practices REPURCHASE AGREEMENTS (Sellon 2002). This economist also asserts that the Like commercial banks, the Fed engages in repurchase agreements Fed’s influence on spending (repos) and reverse repos (or matched sale-purchase transactions). It may have broadened due to does so when it needs to temporarily increase or decrease bank reserves. the effect of interest rate The central bank executes repos (or reverse repos) by buying (or selling) changes on the variable-rate securities that the seller (or buyer) agrees to repurchase (or resell) on a loans and mortgage specified date for a specified price. When the repo or reverse repo refinancings that have matures, both the asset and liability sides of the Fed’s balance sheet grown increasingly popular return to previous levels. with household borrowers. What this claim ignores, The decline in reserve balances has led to more extensive use of repos however, is the Fed’s by the central bank because diminished demand for the Fed’s liabilities evident inability to control (reserves) has forced it to rely increasingly on temporary injections and the quantity of credit. subtractions of liquidity to achieve its policy objectives. Despite the scale Traditional views of an interest-rate-based regime of recent increases, the use of repurchase agreements does not mark a assume that the central radical departure from earlier Fed practices. The central bank began bank’s ability to set the using repos and reverse repos to stabilize the government securities price of reserves will alter market in the 1920s and still uses them to provide short-term funding for the supply of money and primary dealers who are also its institutional counterparties when it bank credit, change the conducts open market operations. price of other short-term financial assets, affect demand for money and credit and thus influence the level of output in economic activity, the rapid growth of outstandthe economy. ing debt as a share of GDP clearly signals rising But sweep accounts, repurchase agreements risk since it indicates relatively slower growth and securitization have expanded the role of in the income needed to service borrowing. For non-reservable liabilities in bank credit creation that reason, both the Bank for International and thus weakened the ability of reserve Settlements (BIS) and the International Monrequirements to restrain bank credit growth. At etary Fund (IMF) have stressed concerns about the same time, nonbank financial firms have excessive credit expansion in their analyses of become much more prominent sources of U.S. credit market developments (BIS 2000; domestic credit. With traditional policy and IMF 2002). institutional restraints watered down or reAs Table 4 suggests, the dangers inherent in moved, debt levels and debt burdens have those developments may be heightened by grown explosively – and crimped the central changes in the composition as well as the bank’s influence over output. dimensions of U.S. borrowing. During the 1980s, the debt of all U.S. nonfinancial sectors The U.S. Debt Bubble surged from 139 percent of GDP to 191 percent. Over the course of the next decade, these At year-end 1971, outstanding U.S. credit sectors’ aggregate borrowing remained at market debt totaled $1.8 trillion, equaling 155 virtually the same level relative to GDP but the percent of GDP. By 2001, outstanding domesburden of debt dramatically shifted from the tic debt had climbed sixteen-fold to $29.5 federal government to households and nonfitrillion and reached 289 percent of GDP (Table nancial corporations. During the past 30 years, 4). While observers have long debated the household debt soared from 45 percent to 76 comparative influence of money and credit on 8

FINANCIAL MARKETS CENTER

Table 4: U.S. Credit Market Debt as a Percentage of GDP 1971

1981

1991

2001

Domestic Nonfinancial

138.0

139.4

191.2

190.3

Federal government

28.7

26.2

46.9

33.1

109.3

113.2

144.3

157.2

Households

44.6

48.4

64.5

75.7

Nonfinancial corporations

34.5

32.7

41.4

48.6

State & Local governments

14.8

11.9

18.2

13.6

12.3

21.8

47.1

91.8

5.0

7.0

5.1

6.9

155.3

168.3

243.4

289.0

27.7

32.2

45.8

53.2

4.9

5.4

26.4

48.4

Sectors

Non-Federal

Domestic Financial Foreign Total Credit Memo Items: Home Mortgages Debt of GSEs & Federally Related Mortgage Pools

SOURCE: Federal Reserve System, Flow of Funds Accounts of the United States

percent of GDP while nonfinancial corporate debt jumped from 35 percent to 49 percent. As debt loads increased, their constraints became more apparent. At yearend 2001, outstanding household debt ($7.7 trillion) equaled 104 percent of disposable income, up from 87 percent in 1990. Household debtservice burdens – the ratio of monthly debt payments to disposable personal income – matched the quarterly record set in 1986. And debt as a share of nonfinancial corporations’ net worth stood at 59 percent, up from 40 percent in 1990 (Board of Governors, Flow of Funds and Household Debt-Service Burden). Meanwhile, the financial sector itself experienced extraordinary increases in borrowing that reflected the dramatic shifts in its products, practices and structure. From 1990 through year-end 2001, financial institutions’ debt rose by $6.8 trillion to $9.4 trillion. Overall, the financial sector’s share of total annual borrowing jumped from about 20 percent in

1980 to about 50 percent at the end of the 1990s, peaking at an annualized rate of 55 percent in the third quarter of 1999. In other words, at the height of the 1990s expansion and during the downturn that followed, financial firms – paced by GSEs and federally related mortgage pools – routinely borrowed more than their domestic customers.10 In retrospect, it seems clear that this unrestrained credit boom played a crucial part in financing the U.S. stock market bubble of the late 1990s. Growing volumes of margin debt and home equity credit enabled households to bid up equity prices to unprecedented levels. In addition, executives and other employees borrowed heavily (often from their own companies) to exercise stock options – a practice that also put upward pressure on share prices. At the same time, debt-financed stock buybacks by corporations substantially reduced the supply – and thereby propped up the price – of their shares (Liang and Sharpe

10. Because GSEs rely on issues of securities rather than deposits to fund their lending, they played a major role in running up financial sector debt. Between 1971 and 2001, the debt of GSEs and federally related mortgage pools rose almost ten-fold as a share of GDP (from 5 to 48 percent). GSEs were also the driving force in the near-doubling of home mortgage debt as a share of GDP (from 28 to 53 percent) during the same period. FINANCIAL MARKETS CENTER

9

1999). More recently, relentless growth in mortgage borrowing has fueled fears of a bubble in U.S. housing markets. In addition, the unprecedented scale of financial-sector leverage carries troubling implications. Widespread borrowing among financial institutions exacerbates the possibility of problems in a few of those firms spreading to many others through their web of debt and derivative obligations. Finally, current U.S. debt levels constitute a substantial drag on the strength of demand and investment. For the economy merely to move forward, household and business borrowers must set aside large portions of their income just to service debt. And the financial sector must continue to raise the mountain of funding needed solely to refinance its own and its customers’ existing borrowing. Under these conditions, robust economic growth could well prove elusive. Slower economic growth would test the sustainability of debt levels for lenders as well as their borrowers. And a vital source of funding in the 1990s – inflows of foreign savings – may prove quite unsustainable in the years ahead.

The Effects of Capital Flows on U.S. Financial Markets During the past decade, the foreign sector became a large net supplier of credit to U.S. borrowers, strongly influencing both the amount and allocation of new borrowing in domestic credit markets. As net U.S. obligations to foreigners swelled during the course of the decade, the gap between U.S. ownership of foreign assets and foreign ownership of U.S. assets – the U.S. net international investment position – rose from -$165 billion, or -3 percent of GDP, in 1990 to -$2.3 trillion, or -23 percent

of GDP at yearend 2001 (Department of Commerce 2002). Foreign investors have provided more than ten percent of total funds loaned in U.S. credit markets in every year since 1992. Their share peaked in 1996 at 30.2 percent of total credit supplied to nonfinancial and financial borrowers – but it topped 20 percent in 1995, 1997, and 2001 as well. 11 Foreign lenders offered particularly strong support for the corporate bond market, becoming its dominant buyers during the second half of the 1990s. In the fourth quarter of 2001, they bought 34 percent ($206 billion) of net new issues and, by year-end 2001, owned 24 percent of outstanding corporate bonds, up from 13 percent in 1990. Foreign investors also own $1 trillion (about one-third) of outstanding Treasury securities and, as the supply of new Treasury debt dwindled in the late 1990s, they ramped up their purchases of agency securities. During the fourth quarter of 2001, for example, foreigners snapped up 32 percent of net new agency issues ($189 billion), lifting their share of outstanding agencies to 14 percent from four percent in 1990. Foreign purchases of U.S. credit market instruments played a prominent, albeit indirect, role in pumping up the equity bubble of the late 1990s. More than any other class of investors, foreign buyers purchased corporate bond issues supporting the stock buybacks that helped prop up unsustainable equity prices.12 And in the process of becoming the single biggest funder of GSE expansion, foreign lenders helped supercharge already buoyant mortgage activity, perhaps to bubble proportions. To a substantial degree, these huge inflows of foreign funds reflected monetary policy

11. Richard N. Cooper and Jane Sneddon Little (2000) offer another measure of the relative scale of capital flows: “U.S. international trade in securities has grown even faster and now looms larger, relative to GDP, than trade in goods and services. While nominal exports plus imports equaled 23 percent of GDP in 1999, gross international transactions in securities equaled 200 percent of GDP.” 12. While acquisitions of U.S. companies by foreign buyers contributed to net equity retirement, foreign investments in stocks lagged foreign bond purchases for most of the 1990s. At yearend 2001, foreign investors owned about 11 percent of outstanding U.S. equities, only slightly above their seven percent share in 1990. In 2000 and 2001, however, foreigners bought more U.S. stock on net than any other class of investor and thereby prevented the decline in equity prices from turning into a steeper plunge.

10

FINANCIAL MARKETS CENTER

decisions by the Federal Reserve – though not always in ways the central bank intended. In February 1994, for example, the central bank began a series of interest rate hikes intended to preempt inflation and prevent the economy from overheating. While many observers questioned the existence of these upside risks, rising rates undeniably helped attract inflows from foreign investors, encouraged U.S. investors to shift back into domestic assets, fueled domestic credit growth and interrupted what had been a consistent outflow of funds to Mexico (thereby precipitating the December 1994 peso crisis).13 Over the rest of the decade – with occasional relaxations – the Fed maintained interest-rate differentials between U.S. and other G7 currencies that favored dollar investments. As the U.S. current account deficit widened and inflows of foreign savings soared, domestic credit expansion blossomed into a full-fledged boom. While the strong dollar restrained inflation (aided by recurring financial crises), it did so at the expense of manufacturing and other export sectors while shifting a larger share of credit flows to households and to corporations financing equity buybacks. And at the peak of the 1990s expansion, the strong dollarsoaring inflows-expanding credit-strong consumption daisy chain exerted an increas-

ingly powerful pull on asset prices. 14 During 2001 and 2002, foreign investors’ appetite for U.S. financial assets remained strong, a skein of Federal Reserve rate cuts notwithstanding. However, should U.S. demand falter and import growth decline, new foreign investment in U.S. financial assets will diminish as the investors’ export income recedes. And a reduction in inflows will strain the ability of households and businesses to refinance extremely high levels of debt, thereby disrupting economic activity. In addition, a shock to the U.S. economy or a substantial loss of confidence by foreign investors in U.S. growth prospects could precipitate a dollar crisis. Should that occur, the Fed might feel forced to push up interest rates in order to halt capital outflows – a lingering prospect that underscores the degree to which monetary policy has become captive to high levels of foreign investment.

III. The Fed at a Crossroads The confluence of these developments has placed the Federal Reserve in a situation uncomfortably reminiscent of the early years of the Great Depression, when the central bank’s passive reserve system proved inadequate to deal with bank runs and failures.15 With the

13. Given the limits of its policy tools, the Fed’s actions can also trigger large outflows inadvertently. At the beginning of the 1990s, the Fed lowered interest rates feverishly in hopes of jump-starting the economy. However, these rate cuts encouraged capital outflows by U.S. residents and foreign investors alike. The outflows absorbed a substantial share of the liquidity the Fed had provided, and was one ot the factors that forced macro policymakers to undertake a longer and larger effort to stimulate the economy. 14. During this period, the Fed made few public comments on the potential hazards of excessive debt growth and undue reliance on capital inflows. Instead, the central bank largely contented itself with lauding productivity gains. After the recession of 2001 began, New York Fed President William McDonough mused that “the major redirection of capital flows in the aftermath of the Asian and Russian crises – much of the redirection ultimately into the United States – has made us wonder whether we underestimate the impact of capital flows on large economies like our own. The surge in global flows of portfolio equity and direct investments in the 1990s also suggests that differentials across borders in the rates of return on financial equity and real capital are probably drivers of international capital flows that are as important as interest rates.” (McDonough 2001). 15. Some might view current conditions as analogous to the period that gave rise to the Federal Reserve System in 1913. At that time, the central bank’s architects aimed primarily to end the constraint of the gold standard on the money supply by creating a mechanism for note issue that would result in an elastic currency, expanding and contracting to meet the seasonably variable needs of trade and accommodate growth. Within this framework, the Fed discounted the eligible paper (trade bills) of member banks and issued Federal Reserve notes backed by a mixture of both eligible paper and the Fed’s holdings of gold. The 1913 Act also required all national banks and state member banks to hold reserves with their regional Federal Reserve Bank, entrusting System managers with the responsibility for investing these funds safely and lending them systematically across regions to prevent the bank runs and financial panics that had plagued the U.S. economy throughout the preceding six decades (D’Arista 1994). FINANCIAL MARKETS CENTER

11

1929 Crash, gold outflows limited the Fed’s ability to conduct the kind of open market operations pioneered by New York Fed chief Benjamin Strong.16 Moreover, slowing economic activity reduced the amount of paper eligible for rediscount and as collateral for note issues, compounding a dramatic contraction in the supply of money and credit. More significantly, the Fed’s maneuvering room was severely limited by its statutory inability to use the banking system’s large holdings of U.S. government securities as collateral for discounts and backing for note issues. 17 In 1932, emergency legislation authorized the use of Treasury securities as backing for Federal Reserve notes. This authorization, reaffirmed in the Banking Acts of 1933 and 1935 and made permanent in the 1940s, provided the framework for system operations still in use today (D’Arista 1994). As financial-sector change renders this framework increasingly ineffective, U.S. central bankers now find capital inflows limiting their actions in somewhat the same way gold outflows handcuffed their predecessors 70 years ago. In addition, the Fed’s contemporary balance sheet – now composed almost exclusively of Treasuries on the asset side – offers the central bank far too little operational flexibility, much as the balance sheet of 1930 stymied policymakers of that era. Hemmed in by these circumstances and hamstrung by other limitations, U.S. central bankers are confronted with four basic choices. They can rationalize the erosion of their primary policy mechanism

by changing objectives. They can attempt to muddle through. They can simply surrender to the inevitability of powerfully procyclical market forces. Or they can acknowledge the need for new policy mechanisms as decisively as their predecessors did in the 1930s.

Option 1: Change the Subject As the Fed’s policy transmission system has worn down, the central bank’s policy objectives have narrowed. Articulating a sentiment shared by many of his colleagues in the U.S. and abroad, St. Louis Fed President William Poole has asserted, “the primary goal of the central bank must be to control inflation. The reason is not that full employment is unimportant for society, but that the central bank does not have policy tools that enable it to reliably increase the level of employment in the long run” (Poole 2000). Many central bankers take this assertion a step further by declaring price stability the most important prerequisite for sustainable output and employment growth. Even though it is widely shared – and even though it offers a seemingly neat rationalization for weakened policy tools – this view has several major flaws. First, and most obvious, focusing on a single goal for policy would require a change in the Federal Reserve Act, which directs the central bank to pursue full employment and maximum output as well as stable prices. Second, tuning out other policy goals contradicts the Fed’s own acknowledgment that its mission includes influencing output levels (which, in turn, determine rates of

16. Governor Strong presided over the Federal Reserve Bank of New York from 1914 until his death in 1928. He understood that relying solely on discount window operations restricted the Fed’s ability to initiate needed policy actions – for if private banks did not actively seek central bank credit (as they did not during the post-World War I recession or the early years of the Depression), the economy might stagnate or fail to grow at full potential. Using the substantial gold inflows to the U.S. during and after World War I to raise the level of gold backing for Federal Reserve notes and reduce the need for discounted paper, Strong took the initiative by buying local government warrants and U.S. government securities in the open market. Buying and selling such assets backed solely by gold, he avoided the legal restrictions on holding assets other than trade bills and succeeded in using open market operations as a tool of stabilization policy until the final years of the 1920s. 17. The amount of U.S. government debt outstanding when the Federal Reserve began operations in 1913 was negligible ($2 billion) and most government securities were used to back national bank notes that were to be phased out and replaced by Federal Reserve notes. There was no provision in the Federal Reserve Act authorizing the use of government securities as collateral in the new system. By the end of World War I, outstanding Treasury debt had risen to $25 billion but could only be acquired by Federal Reserve Banks with 100 percent gold backing.

12

FINANCIAL MARKETS CENTER

employment).18 Third, the single-focus doctrine fails to account adequately for the stillemerging macroeconomic lessons of the past decade. During the 1990s, restrictive Fed policy indeed led to price stability, aided immeasurably by a strong dollar (which weakened other currencies and the prices of exports to the U.S) and globalizing production systems (which curbed producers’ pricing power and wage growth worldwide). While the achievement of price stability through these means coincided with robust output and employment growth in the U.S., it also permitted the build-up of deflationary pressures in product and labor markets that probably have not yet run their course. And, as previously noted, the Fed’s single-minded policy choices also failed to take account of growing debt burdens and a bubble in asset values that ultimately may prove as harmful as comparable levels of inflation in the prices of goods and services. To put it charitably, none of these outcomes validates the wisdom of monetary authorities focusing solely on price stability. Nevertheless, the concept has maintained its cachet in monetary policy circles. Indeed, a number of influential economists have urged the Fed to follow the example of other central banks and shift to an inflation-targeting regime. At first glance, inflation targeting – the notion of using price levels as the lodestar for policy to the exclusion or subordination of all other targeted macroeconomic variables – appears to offer central bankers a way around the problems caused by deteriorating transmission systems. But establishing an inflation targeting system does not automatically answer questions about how price stability targets can be met. If the Fed moved to an inflation-targeting regime and continued to implement policy by

changing the level of short-term interest rates, the effectiveness of the new regime would still depend on the existing transmission mechanism.19 But given its atrophied state, that mechanism likely could not target inflation any more effectively than other target variables – at least not without constant support from a strong dollar, the indefinite sustainability of which appears highly uncertain due to its distorting effects on the current account, capital flows and domestic credit expansion. In other words, narrowing the Fed’s policy goals is not a practical substitute for strengthening the basic tools it needs to implement policy. Equally important, the sustained pattern of divergence between credit and output growth over the last two decades points to a form of inflation that has not been curbed by the Fed’s interest rate policies. The persistence of credit inflation adds an important – and, to date, largely ignored – dimension to the longstanding debate over the relative influence of money and credit on macroeconomic outcomes. Moreover, the stubborn gap between credit expansion and economic growth calls into question the reliability of goods prices as a guide to aggregate demand. The argument for inflation targeting assumes this guide to be telling and predictable, if not inerrant. But a recent study published by the Bank for International Settlements argues that low and stable inflation actually “increases the likelihood that excess demand pressures show up first in credit aggregates and asset prices, rather than in goods and services prices” (Borio and Lowe 2002).20

Option 2: Muddle Through Rather than refashioning its objectives to suit an increasingly ineffective transmission system, the Fed could attempt to tinker with its balance sheet

18. Fed officials frequently note that monetary policy actions may be required to cushion the impact of price-distorting shocks on output and employment. The central bank’s Purposes and Functions publication – the Fed’s boilerplate explanation of itself to the public – also acknowledges this point and notes that the Fed contributes to economic performance by providing liquidity to limit the scope of financial disruptions or soften their effects. 19. By and large, the main advocates of inflation targeting by the Fed (including new Fed Governor Ben Bernanke) do not recommend changes in the conventional policy transmission process. 20. The failure of existing policy transmission tools to prevent or minimize the excess demand pressures that show up in credit and asset prices now poses a danger to stable prices – one that may result in persisting deflation on the downside as financial imbalances threaten to work their way through to the real economy. FINANCIAL MARKETS CENTER

13

or its policy mechanisms. In recent years, the central bank has made some tentative moves in this direction. To help compensate for a (previously problematic) runoff of Treasury debt, for example, the Fed authorized the use of longer-term securitized mortgage debt as collateral for repurchase agreements and actively explored the holding of other substitute instruments such as Ginnie Mae mortgage-backed securities (Board of Governors 2001). In April 2001, Congress approved Fed-requested legislation that authorized payment of interest on the Federal Reserve account balances banks use for clearing payments (Bennett and Peristiani 2001). In January 2002, growing concern over deflationary pressures and the difficulty of conducting stimulative monetary policy in a low-interest rate environment spurred the Federal Open Market Committee to discuss the possible need for “unconventional policy measures” if the economy flagged as short-term interest rates approach the lower bound of zero. Pressed by the Financial Times, a Fed official subsequently noted that these unconventional means could include purchasing corporate stocks, real estate or other assets that would add liquidity to the system (Despeignes 2002a).21 Finally, in May 2002, the Fed unveiled a plan to revamp its discount loan program and encourage depository institutions to borrow

more frequently at the discount window (Madigan and Nelson 2002). In theory, the additional borrowing would enable the Fed to regain some of the policy leverage lost to the decline in reserve balances.22 Like open market operations, lending through the discount window injects liquidity and adds to bank reserves. Building up levels of both discounts and reserves would tighten the slack in the Fed’s monetary policy ties to banks. And it could enhance the Fed’s ability to regulate the supply of money and bank credit.23 However, this kind of incremental change is unlikely to involve financial institutions other than banks and therefore is unlikely to have a substantial impact on policy transmission overall.24 As long as depository institutions hold less than one-quarter of total financial assets, discount-window reform and other bank-focused initiatives can exert a limited influence at best on general credit conditions. Such initiatives may very well allow the Fed to defer a day of reckoning with its underlying policy-implementation problems. But they will not cancel the event.

Option 3: Give Up While most Fed officials acknowledge the importance of monetary policy in managing the economy, 25 proponents of a radically unregulated economic system might argue that the

21. In December 2000, a team of Fed economists released a working paper that examined emergency policy measures the central bank could use if short-term rates reached the zero bound (Clouse et al 2000). 22. Discount window operations provide short-term loans to banks and once served as the main channel for policy; over the past several decades, the Fed has turned to these operations primarily when it acts as lender of last resort. To lighten administrative costs and erase the perceived stigma of borrowing at the discount window, the Fed’s new plan converts the discount loan program from a belowmarket-rate to above-market-rate basis. In its explanation of the plan, the central bank stressed the new discount facility’s potential to compensate for ongoing volatility in the federal funds market and to release large volumes of liquidity in the event of a crisis. 23. Expanding discount window operations could also increase the number of counterparties used by the Fed in implementing policy. Currently, open market operations are conducted primarily through transactions with securities dealers. But wider use of the discount window could theoretically involve all banking institutions as counterparties. 24. Since the Fed’s proposal does not impose reserve requirements on nonbank financial institutions, they would remain ineligible for loans from the discount window other than for emergency reasons. Extending discount-window access to nonbanks without imposing reserve requirements on them would make bank credit more erratic by enabling nonbanks’ discounts to skew the level of bank reserves. This would erode any additional influence over the supply of money and credit the Fed might gain by reviving discount window operations. 25. This acknowledgement is not universal within the Federal Reserve System. In its 1994 annual report, for example, the Federal Reserve Bank of Cleveland asserted,“some people believe that when the economy slides into a recession, it has a natural tendency to stay there and so monetary and fiscal policy actions are needed… [But] such notions…are inconsistent with our view that a market economy is inherently resilient. That is, if an unexpected economic shock results in unemployment, the economy will naturally move back toward full employment without any policy stimulus.”

14

FINANCIAL MARKETS CENTER

difficulty of maintaining effective monetary policy tools makes the case for simply throwing in the towel and allowing market forces free rein to determine fluctuations in money, credit and macroeconomic performance. Such an approach would put laissez-faire philosophy to the test its purest advocates have long claimed they desire. But the growing incidence and painful consequences of financial disruption over the past two decades suggest the outcome wouldn’t be pretty.26 If shorn of all restraining influences, the procyclical bias of market forces would ensure that boom-bust cycles recurred with increasing frequency. And since markets, unlike central banks, do not have the power to create liquidity in a crisis, these violent swings would further destabilize and retard both economic and human development. 27

Option 4: Modernize the Transmission System Finally, the Fed could try to regain meaningful control over money and credit by modernizing – rather than fiddling with – its policy transmission system. Given the shortcomings inherent in its other options, given the gravity of the situation, and given the likelihood that the central bank’s underlying problem will not go away by itself, a thorough overhaul appears to be the most practical long-term remedy for what ails that system.

IV. Rebuilding Effective Transmission Mechanisms for Monetary Policy At this point in the evolution of financial markets, most traditional types of quantity

mechanisms – designed specifically for depository institutions, inapplicable to other segments of the financial industry and seemingly heavy-handed in a contemporary environment – would not provide the Fed an appropriate or practical means to restore monetary control. 28 However, reserve requirements offer a different and more promising set of possibilities. Reserve requirements can be and have been applied to all financial firms in asset-based systems. Such systems have been employed successfully in other countries. Asset-based reserves could be readily adapted to recent and future institutional changes in the financial sector. And among all the available alternatives, reserve-requirement reform appears the most direct and effective path to revitalizing the Fed’s policy transmission system. The remaining sections of this paper describe a comprehensive new system of reserve management that would accomplish this revitalization. The description begins with an outline of the system’s three key operational features: extending monetary control to the entire financial sector; applying reserve requirements to financial firms’ assets; and using repurchase agreements as the primary tool of open market operations while expanding the central bank’s eligible holdings. Next, the paper provides a detailed explanation of monetary policy implementation – and the attendant changes in balance sheets of financial institutions and the Fed – under the proposed system. The paper concludes with an overview of issues arising from the policy implementation process, including the impact of this proposed reserve-management regime on asset prices, capital flows and crisis management.

26. According to World Bank and International Monetary Fund officials, three-fourths of those institutions’ 180-plus member countries experienced a financial sector crisis during the 1980s or 1990s. (Conthe and Ingves 2001) 27. During congressional consideration of the Banking Act of 1935, Federal Reserve Chairman Marriner S. Eccles argued,“in the final analysis, in a depression, there is no liquidity, except that liquidity which can be created by the Federal Reserve or the central bank through its power of issue” (U.S. House of Representatives 1935). 28. One partial exception: mechanisms structured to limit the volume of capital inflows and outflows. FINANCIAL MARKETS CENTER

15

Key Ingredients of Policy Modernization

SECTION 19, FEDERAL RESERVE ACT BEFORE “Each depository institution shall maintain reserves against its transaction accounts as the Board may prescribe by regulation solely for the purpose of implementing monetary policy…Any reserve requirement imposed under this subsection shall be uniformly applied to all transaction accounts at all depository institutions. Reserve requirements imposed under this subsection shall be uniformly applied to nonpersonal time deposits at all depository institutions, except that such requirements may vary by the maturity of such deposits….. In order to prevent evasions of the reserve requirements imposed by this subsection, after consultation with the Board of Directors of the Federal Deposit Insurance Corporation, the Director of the Office of Thrift Supervision, and the National Credit Union Administration Board, the Board of Governors of the Federal Reserve System is authorized to determine, by regulation or order, that an account or deposit is a transaction account if such an account or deposit may be used to provide funds directly or indirectly for the purpose of making payments or transfers to third persons or others.”

1) EXTEND MONETARY CONTROL THROUGHOUT THE FINANCIAL SYSTEM None of the profound changes in saving, investment and financial-sector structure that have emerged over the past two decades evince any sign of abating or reversing. No plausible scenario suggests the likelihood of banks regaining their once-hegemonic role in credit creation. And no likely series of events promises to diminish substantially the influence of institutional investment AFTER pools and capital flows on “Each firm engaged in activities deemed to be financial in nature under credit expansion. rules issued pursuant to the Gramm-Leach-Bliley Act of 1999 shall As a result, any practimaintain reserves against its on- and off-balance sheet assets as the cal effort to rebuild effecBoard may prescribe by regulation solely for the purpose of implementtive transmission mechaing monetary policy…Any reserve requirement imposed under this nisms for monetary policy subsection shall be uniformly applied within each affected category of must establish new chanassets at all firms conducting activities that are financial in nature. For nels for exercising monfirms that are not a depository institution, bank holding company or etary control over all financial holding company, reservable assets shall consist only of those financial institutions. Simply put, banks can no holdings related to financial activities. In order to prevent evasions of the longer shoulder the transreserve requirements imposed by this subsection, after consultation with mission-belt function alone any financial regulatory agency, the Board of Governors of the Federal on either the up or down Reserve System is authorized to determine, by regulation or order, that an side of the business cycle – asset is a reservable asset if it is held incident to activities that are as amply demonstrated by financial in nature.” credit-market and realsector developments during the 1990s. And nonbank financial firms cannot participate meaningfully engaged in such activities. Requirements in the transmission-belt function unless they too should be imposed only on those portions of a meet reserve requirements. company engaged in financial activities (for How might the universe of reservable assets example, GMAC) but not those portions conbe defined? One sensible approach would be ducting nonfinancial operations (for example, to use the Gramm-Leach-Bliley Act’s definition GM’s auto making divisions). By drawing this of activities deemed financial in nature and distinction, the new system of reserve requireapply reserve requirements to companies ments would strengthen the crucial separation 16

FINANCIAL MARKETS CENTER

between banking and commerce, prevent commercial entities from making emergency liquidity claims on the lender of last resort, and give the Fed additional grounds for prudence in exercising its GLBA-granted authority to define financial activities. Section 19 of the Federal Reserve Act could be revised to codify this change (see box). 2) BASE THE NEW RESERVE SYSTEM ON ASSETS, NOT LIABILITIES To be sure, bringing non-depository institutions under the Fed’s monetary control demands significant adjustments to a reserve structure tailored to fit banks’ unique role in the financial system. Despite their growing dominance, nonbank financial intermediaries are not designed to engage in money creation. Unlike banks, they do not create new liabilities for customers when they add assets. Moreover, the liabilities of institutional investors such as pension funds and insurance companies are in longer-term contracts, rendering reserve requirements on those liabilities impracticable. In short, a liability-based system doesn’t permit central banks to create and extinguish reserves for nonbank financial firms. Efficiency and equity therefore require that reserves be held against assets. Though this notion may appear exotic, it actually embodies a range of real-world experiences, including the current model for U.S. insurance regulation. For years, states have required insurers to hold reserves against their assets and, in 1992, insurance commissioners instituted an asset valuation reserve (AVR) system that assigns risk weightings to various asset types (Palley 2000). Although these reserves were imposed for soundness purposes (as opposed to conducting

monetary policy) and are held by the firms themselves (rather than a public agency), they nonetheless illustrate the feasibility of systematically reserving and classifying institutional investors’ assets. The experience of European countries during the Bretton Woods era provides additional examples of asset-based reserve systems – some designed to control overall credit expansion and others to shield key sectors from cyclical excesses and droughts.29 And as recently as 1979, the Federal Reserve imposed reserve requirements on loans by U.S. banks’ foreign branches to their home offices. Industry resistance and pressures for deregulation doomed these earlier asset-based approaches to extinction. And the many changes that occurred in financial markets, institutions and transactions during the 1980s and 1990s make it highly unlikely that such systems could be revived in their past form. Nevertheless, historical evidence suggests that asset-based reserve models were often effective in accomplishing their primary goals (U.S. House of Representatives 1972). And no other models offer more pertinent lessons for modernizing the Fed’s policy tools today. Only by targeting financial firms’ assets can a reserve system hope to effectively influence a majority of total credit extended to nonfinancial and financial borrowers and to ensure greater balance in the distribution of resources across the business cycle. By shifting reserve requirements in this fashion, the Fed would be able to extend monetary control to an assortment of assets that, as of yearend 2001, was 36 times larger than the universe of reservable deposits ($35.8 trillion versus $998 billion: see Chart 1). Fifty years

29. As the foremost example of the latter objective, Sweden required all financial sectors to hold a given percentage of their total portfolio in housing-related assets, thereby erecting “a countercyclical shield against finance problems” for housing (U.S. House of Representatives 1972). Institutions that did not make real estate loans could meet the requirements by purchasing the liabilities of institutions that did. Financial firms that failed to meet the required percentage had to enter the shortfall on their balance sheet as reserves. In other words, Swedish financial intermediaries had to make an interest-earning loan for housing or an interest-free loan to the government. Meanwhile, credit ceilings were used by the Netherlands Bank before 1967 and by the Bank of England before 1971. Italy used such ceilings periodically to restrain inflation and Switzerland imposed credit limits in the years 1972-1975 to curb a housing boom. The reserve system established by France in 1967 targeted both deposits and assets, limiting additions to credit if assets increased more than a given amount over the previous year (U.S. House of Representatives 1976). FINANCIAL MARKETS CENTER

17

earlier, the ratio stood at slightly more than two to one ($320 million versus $142 million). With authority over such a large universe of assets, the Fed could – and likely would – set reserve requirements at an extremely low level and still be able to conduct policy effectively. Thus, with the move to an asset-based system, reserve requirements would become far less onerous to depository institutions, represent a small burden (at most) to nonbank financial firms and remove a longstanding competitive inequality by leveling the field for the entire financial sector. 3) EMPLOY REPURCHASE AGREEMENTS AS THE CENTRAL BANK’S PRIMARY OPERATING TOOL AND EXPAND THE FED’S ELIGIBLE HOLDINGS As a proven tool of monetary policy, repurchase agreements provide the flexibility and adaptability needed to implement a reserve

system based on the broad universe of financial sector assets. Repurchase transactions are ideally structured to allow the Fed to interact with all financial firms on the asset side of their balance sheets. For example, the Fed can use a repo to buy government securities (or agencies, corporate bonds, loans, mortgages, commercial paper, etc.) from any of the many institutions that hold these assets – commercial and investment banks, mutual and pension funds, insurance and finance companies, or GSEs. Indeed, the Fed should be empowered to accept a wide variety of sound assets as backing for repurchase agreements. Broadening the holdings on its balance sheet would bring the Fed closer to the successful practices of other central banks.30 More importantly, authorizing the Fed to conduct repos with any sound financial asset – as long as its actions remain

$35,769.1

Chart 1: Financial Sector Assets vs. Estimated Reservable Deposits ($ billion)

40000 35000 30000 25000 $14,631.7

1961

1971

1981

1991

$998.3

$1,955.8

$5,140.7

$482.3

$683.2

$319.8

1951

$213.8

0

$142.3

10000

$1,754.6

15000

Reservable Deposits

$727.4

20000

5000

Financial Sector Assets

2001

NOTE: For years 1951, 1961 and 1971, estimated reservable deposits include transaction and nontransaction deposits for commercial banks, including non-member banks. For 1981, estimated reservable deposits include transaction and nontransaction deposits for all depository institutions. For 1991 and 2001, estimated reservable deposits include transaction deposits for all depository institutions. SOURCE: Federal Reserve Flow of Funds Accounts

30. Expanding the Fed’s eligible holdings would also confirm the wisdom of former Chairman Eccles, who argued that the Banking Act of 1935 should be amended to free the Fed to buy or discount “any sound asset” (U.S. House of Representatives 1935). At that time, the Federal Reserve Act permitted only trade bills to serve as backing for currency and bank reserves. The reluctance of a few powerful members of Congress to change this provision almost defeated the effort to extend emergency legislation that added government securities as eligible paper. Since then, government securities have attained the status trade bills once had as the enshrined central bank asset.

18

FINANCIAL MARKETS CENTER

consistent with policy objectives – would strengthen the central bank’s ability to halt runs, moderate crises and curb excessive investment across the entire financial system it oversees, from over-the-counter derivatives markets to the mutual fund industry. Extending the Fed’s range of eligible holdings would eliminate the central bank’s need to own a vast amount of Treasury securities – and its corresponding reliance on the federal government maintaining relatively high levels of indebtedness.31 The requirement that one government liability (government securities) be used to back another (outstanding currency) was redundant at the time of its adoption in 1932. While outstanding currency should of course remain a Federal Reserve liability, the central bank does not need to hold Treasury obligations to make good on that claim because it already wields a more powerful guarantee – the ability to create and extinguish Federal Reserve notes. By maintaining an enormous stockpile of Treasuries, the Fed has restricted the availability of this risk-free, highly liquid asset for use in private transactions, where it is needed to support market stability. In an asset-based system, the Fed could still acquire government securities as backing for repos. But most of its vast current holdings of Treasuries would be released for purchase by investors and financial institutions seeking the ultimate safe-haven asset.

Implementing an Asset-Based Operating System Moving to a system of reserve management that incorporates these three major features will necessarily involve balance-sheet changes for both financial firms and the Fed – and it will require changes in the conduct of policy. Figures 1(a), 1(b) and 3, along with the accompanying text, summarize balance-sheet categories and open market operations under the current liability-based reserve system. Figures 2(a), 2(b) and 4, also elaborated by text, ex-

Figure 1(a): Current Balance Sheet Structure Depository Institutions Assets

Liabilities

Reserves Loans Other

Deposits Capital Other

Figure 1(b): Current Balance Sheet Structure Federal Reserve System Assets

Liabilities

Government securities Repurchase agreements Discounts Foreign exchange reserves

Currency in circulation Bank reserves Government deposits Other

plain how an asset-based system would modernize the status quo. BALANCE SHEET STRUCTURES FOR AN ASSETBASED RESERVE SYSTEM Financial sector balance sheets. In a system of universally applied reserve requirements, financial institutions would book reserves on the liability side of their balance sheets rather than on the asset side. As Figure 1(a) illustrates, the current system treats bank reserves as a claim on the Fed, implying that depository institutions have loaned their funds to the central bank. Figure 2(a) shows where reserves would appear on the balance sheet of all financial intermediaries – banks, insurers, securities firms, pension funds, mutual funds, mortgage companies, and so forth – under a universal reserve-requirement regime (the Figure is a composite; not all the assets or liabilities listed in the composite appear on the balance sheets of all financial firms). Booking reserves on the

31. Maintaining some modicum of federal debt may well be necessary to achieve the optimal functioning of financial markets and to support solid macroeconomic performance. However, these considerations need not be entangled in efforts to improve the central bank’s policy operations. FINANCIAL MARKETS CENTER

19

liability side of financial firms’ balance sheets would be a practical consequence of moving to a reserve system that includes all financial sectors. In addition, it would more accurately reflect the role of both reserves and the central bank in America’s current monetary system.32 These balance sheet changes would also have two important implications for depository institutions as well as the broader financial industry. First, defining reserves as liabilities to the Fed would clarify and make explicit the fact that reserves represent the financial sector’s obligation to serve as a transmission belt for policy initiatives intended to affect economic activity. Second, recognizing reserves as liabilities to the Fed would moot the contentious issue of paying interest on reserves – removing a longstanding sore point for depository institutions and a potential expense for taxpayers. 33 The Federal Reserve’s balance sheet. Figure 1(b) shows the composition of the Fed’s current balance sheet. In this alignment, reserves are booked on the liability side, mirroring their placement as assets on depository institutions’ balance sheets and perpetuating the long-outmoded assumption that they are loans to the Fed.34 Figure 2(b) reverses that arrangement. In an asset-based reserve system, reserves would be recorded on the asset side of the Fed’s balance sheet, mirroring their entry as

Figure 2(a): Balance Sheet Structure Using Asset-Based Reserve Requirements Financial Institutions Assets Loans Bonds Shares Mortgages Treasuries Open market paper Other securities, advances & contracts Repos & Fed funds Cash

Liabilities Deposits Open market paper Loans Bonds Shares Mortgage securities Other securities, advances & contracts Repos & Fed funds Capital Reserves

Figure 2(b): Balance Sheet Structure Using Asset-Based Reserve Requirements Federal Reserve System Assets Financial sector reserves

Liabilities Currency in circulation Government deposits Repurchase agreements collateralized by loans, bonds, shares, mortgages, foreign exchange assets, Treasuries, open market paper and other securities Discounts

32. Booking bank reserves as a claim on the Fed has outlived its rationale and its usefulness. Because the central bank creates reserves and distributes them to deposit-taking institutions, those reserves already function as liabilities of private entities in the sense that their existence – both in terms of creation and extinction – depends on the actions of the central bank. 33. If a legislative change compelled the Fed to pay interest on reserves under the current system, taxpayers would effectively foot the bill because the central bank would rebate less money to the Treasury when it returns earnings in excess of expenses. Under an assetbased reserve system, it might be argued that financial institutions should pay interest on reserves to the Fed. However, the effect would be to impose a tax that encourages evasion and undermines the larger objective of maintaining a strong monetary transmission mechanism. This objective likely would be achieved more efficiently if financial firms simply held reserves as non-interest-bearing liabilities to the central bank. 34. Defining reserves as Federal Reserve assets would finally, if belatedly, achieve a fuller measure of consistency between the central bank’s balance sheet and its actual operations. During the drafting of the Federal Reserve Act, lawmakers forged a political compromise with the banking industry that made the new monetary authority appear to be nothing more than a bankers’ bank – a repository for the reserves banks would pay into the system as a safeguard in the event of future financial panics. In this conceptual framework, reserves could legitimately be viewed as a passive type of central bank liability. Soon after the Fed’s establishment, however, the invention of open market operations gave the System the ability to create reserves and exercise a level of influence on financial markets and economic activity not envisioned when the legislation was enacted. Later, the Banking Acts of 1933 and 1935, the Employment Act of 1946, and the Humphrey-Hawkins Full Employment and Balanced Growth Act of 1978 fully recognized and ratified this influence. Nevertheless, the Fed has maintained a set of bookkeeping arrangements that continue to treat its assets and liabilities like those of a mere bankers’ bank. Defining financial sector reserves as assets of the central bank would modernize these outdated arrangements by confirming that: a) the Fed’s major function is to create and extinguish liquidity; and b) it enjoys the unique ability to create the reserves that accomplish this function.

20

FINANCIAL MARKETS CENTER

liabilities to the Fed on the balance sheets of banks, insurance companies, pension funds, mutual funds, GSEs and all other financial institutions. Meanwhile, repurchase agreements and discounts would move from the asset to the liability side of the Fed’s balance sheet to reflect the Fed’s liability for the private sector assets it acquires when it creates reserves. Foreign exchange assets (international reserves) also would become liabilities rather than assets since they too would be acquired through repurchase agreements. Outstanding currency would remain a liability, manifesting the Fed’s congressionally mandated authority to create money and manage its value. As Figure 2(b) shows, financial sector reserves would constitute the Fed’s only assets under the proposed system. The central bank would no longer hold a huge portfolio of government securities as backing for Federal Reserve notes, bank reserves and government deposits. Freed from its reliance on holding government debt, the Fed would be able to build on current practices by acquiring financial assets through repurchase agreements and discounts with financial institutions, allowing the volume of these liabilities to increase and

diminish to implement changes in policy. Periodic statements of condition would list the types and values of instruments backing the repos (Figure 2(b) lists examples of some types of instruments). Since the Fed would no longer earn interest on holdings of government securities – and since it would back repurchase agreements and discounts with non-interest-bearing reserves – the central bank would no longer have income to pay interest on its repos. But because financial institutions receive invaluable interestfree liabilities when they “loan” the central bank assets through repos, it seems eminently reasonable to compensate the central bank for its role in creating liquidity by allowing it to receive earnings on the collateral backing those repos.35 STEPS IN THE CONDUCT OF POLICY Under an asset-based reserve system, the Fed’s method of implementing expansionary and contractionary monetary policies would closely parallel its current implementation process (outlined in Figure 3) in three significant ways. The central bank would continue to buy and sell financial assets in transactions with private financial institutions. The Fed’s actions would

Figure 3: Current Open Market Operations Depository Institutions Assets

Federal Reserve System Liabilities

Assets

Liabilities

E X P A N S I O N ➊ + Government securities (or repos or discounts) ➋ + Reserves ➌ + Loans

➌ + Deposits C O N T R A C T I O N ➊ - Government securities (or repos or discounts)

➋ - Reserves ➌ - Loans

➊ + Bank reserves

➊ - Bank reserves

➌ - Deposits

35. If the Fed kept the earnings on financial assets held under repurchase agreements, the income – along with fees for check clearing and other services – should prove sufficient for it to continue operating at or near current levels. FINANCIAL MARKETS CENTER

21

still have the effect of simultaneously changing the amounts of its own assets and liabilities as well as those of private financial institutions. And policymakers would continue to rely on the federal funds market to distribute reserves and enforce the price effects of changes in reserves throughout the financial system. However, an asset-based system also would produce two major departures from the Fed’s current method of conducting policy. Rather than conducting transactions only with the primary dealers that today make markets in government securities, the central bank would deal with any financial institution. In addition, all private financial entities – not just banks – would be impacted when the Fed employed

open market operations to alter the volume of assets and liabilities on its balance sheet. To implement an expansionary policy under the proposed operating system, the Fed would add to reserves by engaging in a repurchase agreement with a financial institution. The expansion of reserves would occur in two steps, as illustrated in Figure 4. 1) The central bank buys an asset from a financial institution – for example, GE Capital, Fannie Mae, Prudential or JP Morgan Chase – agreeing to resell the asset in a designated period of time. The Fed pays for the asset by crediting the seller’s reserve account. In the example depicted in Figure 4, the Fed has added $1 of liabilities to its balance sheet (the

Figure 4: Open Market Operations Using Asset-Based Reserve Requirements Financial Institutions Assets Liabilities

Federal Reserve System Assets Liabilities

E X P A N S I O N Pre-Transaction Conditions

STEP ➊

1000

900 (to customers, investors, lenders) 100 reserves

1000

1000

- 1 asset + 1 repo 1000

STEP ➋

+ 10 assets

100 reserves

100 + 1 reserve

100 (cash, government deposits, repos & discounts) 100 + 1 repo

+ 1 reserve 1001

101

101

+ 9 liabilities (to customers, investors, lenders)

1010

1010

1000

900 (to customers, investors, lenders) 100 reserves

1000

1000

C O N T R A C T I O N Pre-Transaction Conditions

STEP ➊

- 1 repo + 1 asset 1000

STEP ➋

-10 assets 990

22

FINANCIAL MARKETS CENTER

100 reserves

100 - 1 reserve

100 (cash, government deposits, repos & discounts) 100 - 1 repo

- 1 reserve 999 - 9 customer liabilities 990

99

99

repo) and created $1 of assets (reserves).36 On the asset side of GE-Fannie-PruMorgan’s balance sheet, the transaction is a wash: the addition of a $1 repurchase agreement offsets the sale of $1 of assets to the Fed. However, the repo with the central bank (unlike the asset acquired by the Fed) does not bear interest for GE-Fannie-Pru-Morgan. Meanwhile, on the liability side of its balance sheet, GE-Fannie-Pru-Morgan has gained $1 of interest-free liabilities (the reserve deposit). 2) Using a fractional reserve requirement of ten percent for illustrative purposes – the actual percentage would be much smaller due to the large volume of reservable assets – the addition of $1 of reserve liabilities makes it possible for GE-Fannie-Pru-Morgan to support $10 of additional assets and to do so by acquiring only $9 of additional liabilities from customers.37 As is the case under current operating procedures, reserves would be distributed throughout the financial system by means of purchases and sales among the private institutions in the federal funds market.38 The system may not maximize the expansionary potential of the reserve increase due to the voluntary nature of this process. But the addition of a given amount of interest-free liabilities would provide powerful incentive – nothing to lose and more earnings to gain – for GE-Fannie-Pru-Morgan to acquire income-producing assets and thereby lead to a fairly predictable increase in credit. By providing this incentive, the asset-based reserve system would remedy a major flaw in the existing liability-based model. Under the prevailing system, the Fed can push on a string, creating excess reserves that aren’t used in the kind of credit crunch that developed during the 1990-1991 recession. Under the proposed system, string turns into stimulus. In a deflationary environment, this change

could prove the difference between recovery and prolonged recession. With the tools available in an asset-based system, the Fed could create reserves to encourage cancellation of non-performing loans and debt securities, allowing the financial sector to replace them with earning assets. This would channel liquidity directly to households and businesses, helping avoid the stagnation that develops when financial institutions resist issuing new credit and cannot cancel debt for troubled borrowers without jeopardizing their own survival. By strengthening private sector balance sheets in this fashion, monetary policy could powerfully reinforce fiscal initiatives designed to revive demand and investment. To implement a contractionary policy under the proposed system, the Fed would allow repurchase agreements to mature without renewal. The contraction of reserves also would take place through a two-step process illustrated in Figure 4. 1) The Fed extinguishes its liability to the seller of the repo (e.g. GE-Fannie-Pru-Morgan) by returning the collateral and debiting the seller’s reserve account. Thus the central bank reduces its balance sheet by $1 of liabilities and $1 of assets. GE-Fannie-Pru-Morgan has exchanged $1 of non-interest-bearing assets (the repo with the Fed) for $1 of interest-bearing assets (the backing for the repo). The book value of its assets is unchanged but GE-Fannie-Pru-Morgan has lost $1 of non-interest-bearing liabilities (the reserve deposit). 2) The loss of a $1 reserve deposit requires the institution to sell assets equal to a given multiple of the fractional reserve requirement. If the requirement is 10 percent, GE-FanniePru-Morgan must sell $10 of assets and reduce its liabilities to customers by $9.

36. Incidental to the core open-market transaction addition – and therefore not depicted in Figure 4 – the Fed receives interest or earnings on the asset it bought through the repurchase agreement and holds on its books. 37. Among other things, this feature ensures that asset-based reserves would impose less of a tax on the financial system than does the current requirement on banks to hold non-interest-bearing reserves against deposits. 38. Aside from broadening the fed funds market, an asset-based system is unlikely to change its structure. The Federal Reserve Bank of New York would continue implementing repo transactions for the Fed and would probably continue interacting with current primary dealers. And, as now, the primary institutions would sell or buy reserves to distribute changes throughout the system while individual institutions would use the market to adjust to changes in their own balance sheets. FINANCIAL MARKETS CENTER

23

Again, a change in the supply of reserves triggers adjustments that ripple throughout the financial system via the federal funds market. At the end of this process, contraction will have occurred in both the total supply of credit and the value of total credit market assets. Because its assets would consist entirely of the financial system’s reserves and because it would hold the backing for repos (e.g. bonds, notes, bills, commercial paper, etc.) only as liabilities under repurchase agreements, the Fed would not be able to engage in reverse repurchase agreements as it currently does. However, by holding liabilities for repos in a range of maturities, the central bank would be able to execute both rapid and gradual tightenings.39

Additional Considerations VARYING RESERVE REQUIREMENTS BY SECTOR By even-handedly requiring all institutions to hold reserves against assets, the operating system described in this report would reduce some of the distortions inherent in current monetary policy. Clearly, however, this new system would not eliminate fundamental differences between depository institutions and nonbank financial firms. While both banks and nonbanks extend credit, the most basic difference between them involves monetary creation. Only depository institutions create liabilities to customers when they increase loans. As a result, they make a uniquely active contribution to credit expansion and contraction under virtually all monetary operating procedures. In addition, bank liabilities still represent the main channel for private sector payments. Therefore, even in an asset-based reserve system, the Fed will tend to interact primarily with depository institutions in the day-to-day conduct of monetary policy. Even though changes in the reserves of insurance companies, pension funds and mutual funds would play a secondary role in

expanding or contracting money, those changes will still alter values on both sides of institutional investors’ balance sheets. For example, an increase in reserves would permit mutual funds or pension funds to expand their assets and consequently raise the value of existing liabilities to shareholders or beneficiaries. On the other hand, losses of reserves would compel these pools to sell assets, pay back the realized proceeds to owners and lower the value of existing mutual fund shares or beneficiaries’ retirement accounts. When stock prices are rising, all pooled investment funds that hold equities will need more reserves to cover the increased value of their assets. If the Fed decided that accommodating their demand for reserves would conflict with current policy objectives, it could leave reserve levels unchanged. At that point, investment funds would have the option of buying more reserves in the federal funds market to avoid selling stocks. However, if they did, their actions would drive up the fed funds rate, thereby discouraging further purchases of reserves in order to amass more equities. In addition, the Fed could set different reserve requirements for different types of assets and alter the requirements to address imbalances in credit flows (see below). Using these tools to alter institutional investors’ reserve liabilities, the central bank could thus influence changes in wealth that indirectly encourage or discourage new spending and saving. However, since pension funds, mutual funds, insurance companies and other nonbank sectors are passive intermediaries that do not create their own liabilities when they add assets, it would be appropriate for the Fed to set lower reserve requirements for their holdings than for bank loans. THE IMPACT OF RESERVE REQUIREMENTS ON ASSET PRICES The notion of reserve requirements directly affecting asset prices may seem radical to some.

39. The Fed will always have an adequate portfolio of repos to effectively moderate credit expansion. Since reserves are the assets that back repos, the central bank could only lose its ability to contract credit if it extinguished all reserves in the financial system.

24

FINANCIAL MARKETS CENTER

However, the Federal Reserve’s open market operations already impact asset prices through changes in interest rates and liquidity, both of which trigger portfolio shifts that disseminate the effects throughout asset markets.40 Though they do so indirectly, the Fed’s interest rate changes exert profound effects on the value of pension fund assets, mutual fund shares and housing, as recent experience has shown. In the growing debate over central banks targeting asset prices, some have sought to portray the technique as illegitimate or simply futile. In practice, however, all efforts to conduct monetary policy must take asset-price movements into consideration – at least at some level of the analytical or decisionmaking process. And, targeted or not, all efforts to conduct monetary policy must influence those price movements. As long as the Fed’s basic objectives – sustainable output, low unemployment, stable prices – remain constant, it makes scant philosophical difference whether policy transmits those influences indirectly (as in the current liability-based reserve system) or directly (as in an assetbased regime). In practical terms, the Fed’s influence on asset markets likely would function far more efficiently under an asset-based reserve system, even if did not become more potent in the aggregate. With all financial institutions participating in the federal funds market, volatility would decline as a result of those institutions making portfolio adjustments by purchasing and selling reserves rather than assets. This would be particularly important in the event of market disruptions, when forced sales of assets increase downward pressure on prices.

THE IMPACT OF RESERVE REQUIREMENTS ON FOREIGN RESERVES AND CAPITAL FLOWS Even though the Fed does not currently engage in repurchase agreements backed by foreign securities there are no technical reasons why it shouldn’t under an asset-based reserve system. Such transactions would eliminate the central bank’s need to hold international reserves as precautionary investments. And they would underscore the Fed’s commitment to preserving stable conditions in domestic financial markets.41 In addition, asset-based reserve requirements would provide a more effective tool for sterilizing capital inflows than does the prevailing liability-based regime. Under the current system, the central bank’s reserve adjustments only affect depository institutions – not intermediaries that hold corporate bonds and other assets preferred by foreign investors. Thus while the Fed can influence systemic liquidity by buying or selling government securities in open market operations, it cannot offset the supply or distribution of foreign investment in assets other than bank deposits or Treasuries. Because of this handicap, the central bank cannot play an effective restraining role when foreign inflows or outflows cause the issuance volume or price level of such assets to veer worrisomely from overall economic growth patterns. Under an asset-based reserve system, the Fed could respond to excessive investment of foreign funds in one or more U.S. asset markets by allowing repos backed by those holdings to run off and replacing them with repos in foreign assets.42 This would effectively mop up the inflow, leaving reserves, interest rate

40. Currently, for example, a hike in the federal funds rate raises interest rates (and lowers prices) on other short-term instruments such as Treasury bills and commercial paper. Investors respond to these changes by adjusting holdings of various assets to maximize returns. In the process, their purchases and sales tend to alter prices for and returns on medium- and long-term assets and even equities. Though market developments (including cross-border capital flows) may inhibit or amplify the results, these price effects constitute a primary element of the current monetary transmission system. 41. International reserve holdings are central banks’ investments in financial assets issued by the government or private residents of other countries. By holding and managing these reserves, central banks can inadvertently have destabilizing effects on financial markets in the nation that issued the assets. 42. This would constitute a form of exchange market intervention that would influence market conditions and exchange rates for other countries as well. However, it would be more effective than current intervention strategies because, by targeting portfolio shifts, it would neutralize the allocative effects of capital flows.

FINANCIAL MARKETS CENTER

25

levels and credit conditions largely unchanged. Alternatively, the Fed would conduct repos in assets sold by foreign investors, increasing reserves to counter the contractionary effects of an outflow. RESERVE REQUIREMENTS, MACROECONOMIC CONTROL AND CRISIS MANAGEMENT At the end of the day, the main purpose of reinstating quantitative policy tools is to improve monetary control and overall macroeconomic performance. However, because they effectively constrain or stimulate specific asset types or sectors, asset-based reserve requirements could be further refined to deal with sectoral bubbles or credit crunches. For example, an increase in reserve requirements on banks’ commercial real estate loans would have defused a bubble in the late 1980s when such loans were rising by over 20 percent a year at New England banks. The recent bubble in high tech stocks also could have been deflated more effectively by raising reserve requirements on those equities instead of hiking interest rates or margin requirements. On the other hand, a fall in stock prices would produce excess reserves that allows intermediaries to buy stocks and moderate the drop more efficiently than a cut in interest rates or capital gains taxes. Last, but certainly not least, an asset-based reserve system would give all financial institutions direct access to the Fed’s discount window and provide the financial sector more meaningful lender-of-last-resort coverage. For example, if mutual funds faced runs by share-

26

FINANCIAL MARKETS CENTER

holders, they could avoid selling assets (and thus prevent a free-fall in prices) by discounting assets with the Fed and acquiring reserves needed to offset customers’ withdrawals. While current law authorizes the Fed to provide emergency liquidity to nondepository financial firms, the authority remains untested in practice. If it were bundled with complementary reforms in prudential supervision and financial sector guarantees, the comprehensive lender-oflast-resort facilities achievable through assetbased reserves would make the Fed’s crisis interventions more coherent, less costly and, hopefully, less necessary. Like the other reforms proposed in this paper, this improvement in crisis-management technique and strategy would begin forging a policy framework appropriate for the 21st century’s financial system. ■ Jane D’Arista is director of programs at the Financial Markets Center, where she conducts educational seminars and authors Capital Flows Monitor, a quarterly analysis of international financial developments, and Flow of Funds Review & Analysis, a quarterly commentary on the Federal Reserve’s main compilation of domestic financial statistics. Previously, she served as an international economist at the Congressional Budget Office, a staff member of the House Banking Committee and chief finance economist for the Subcommittee on Telecommunications and Finance of the House Energy and Commerce Committee. Her publications include The Evolution of U.S. Finance (M.E. Sharpe).

References Bank for International Settlements. 2000. BIS Quarterly Review: International Banking and Financial Market Developments. Basle:BIS (November). Bennett, Paul and Stavros Peristiani. 2001. “Are U.S. Reserve Requirements Still Effective?” Financial Innovation and Monetary Transmission. Conference sponsored by the Federal Reserve Bank of New York (April).

Feinman, Joshua N. 1993. “Reserve Requirements: History, Current Practice, and Potential Reform.” Federal Reserve Bulletin (June). Washington, DC: FRS. Fieleke, Norman S. 2000. “U.S. Monetary Policy in an Integrating World: Discussion.” The Evolution of Monetary Policy and the Federal Reserve System Over the Past Thirty Years. Conference Proceedings, Federal Reserve Bank of Boston (October).

Board of Governors of the Federal Reserve System (FRS). 2002. “Factors Affecting Reserve Balances of Depository Institutions and Condition Statement of Federal Reserve Banks” (Release H.4.1). August 29. Washington, DC: FRS.

Friedman, Benjamin M. 2000. “The Role of Interest Rates in Federal Reserve Policymaking.” The Evolution of Monetary Policy and the Federal Reserve System Over the Past Thirty Years. Conference Proceedings, Federal Reserve Bank of Boston (October).

…………2001. Minutes of the Meeting of the Federal Open Market Committee Held on January 30-31, 2001. Washington, DC: FRS.

Goodhart, C.A.E. 2000. “Can Central Banking Survive the IT Revolution?” International Finance 3 (July).

…………1994. The Federal Reserve System: Purposes and Functions. Washington, DC: FRS. …………Flow of Funds Accounts of the United States. Washington, DC: FRS (various issues). …………Household Debt Service Burden. Washington, DC: FRS (various issues). …………Sweeps of Transaction Deposits into MMDAs. Washington, DC: FRS (various issues). Borio, Claudio and Philip Lowe, 2002. “Asset prices, financial and monetary stability: exploring the nexus.” BIS Working Papers, No. 114. Bank for International Settlements. Basle: BIS (July). Clouse, James, Dale Henderson, Athanasios Orphanides, David Small and Peter Tinsley. 2000. “Monetary Policy When the Nominal Short-Term Interest Rate is Zero,” Finance and Economics Discussion Series (December). Washington, DC: FRS. Cooper, Richard N. and Jane Sneddon Little. 2000. “U.S. Monetary Policy in an Integrating World: 1960 to 2000”. The Evolution of Monetary Policy and the Federal Reserve System Over the Past Thirty Years. Conference proceedings, Federal Reserve Bank of Boston (October). Conthe, Manuel and Stefan Ingves, 2001, “Financial Sector Assessment Program: A Commentary.” International Monetary Fund Views and Commentaries (at www.imf.org). Washington, DC (March). D’Arista, Jane. 1994. The Evolution of U.S. Finance; Vol. I; Federal Reserve MonetaryPolicy: 1915-1935. Armonk, NY: M.E. Sharpe, Inc. Despeignes, Peronet. 2002a. “Fed considered emergency measures to save economy.” The Financial Times. March 25. …………2002b. “Fed proposes change to its role in emergency lending to banks.” The Financial Times. May 18.

Greenspan, Alan. 1993. “Opening Remarks.” Changing Capital Markets: Implications for Monetary Policy. Symposium sponsored by the Federal Reserve Bank of Kansas City (August 19-21). International Monetary Fund. 2002. Global Financial Stability Report: Market Developments and Issues. Washington, DC: IMF (March). Liang, J. Nellie and Steven A. Sharpe. 1999. “Share Repurchases and Employee Stock Options and their Implications for S&P 500 Share Retirements and Expected Returns.” Finance and Economics Discussion Series (December). Washington, DC: FRS Madigan, Brian F. and William R. Nelson. 2002. “Proposed Revision to the Federal Reserve’s Discount Window Lending Programs.” Federal Reserve Bulletin (July). Washington, DC: FRS McCauley, Robert N. and Rama Seth. 1992. “Foreign Bank Credit to U.S. Corporations: The Implications of Offshore Loans.” Federal Reserve Bank of New York, Quarterly Review (Spring). McDonough, William J. 2001. “Opening Remarks.” Financial Innovation and Monetary Transmission. Conference sponsored by the Federal Reserve Bank of New York (April). Palley, Thomas. 2000. “Stabilizing Finance: The Case for Asset-Based Reserve Requirements.” Financial Markets & Society. Philomont, VA: Financial Markets Center. Pollin, Robert. 1993. “Public Credit Allocation Through the Federal Reserve: Why It Is Needed; How It Should Be Done,” in Gary A. Dymski, Gerald Epstein, and Robert Pollin, eds., Transforming the U.S. Financial System. Economic Policy Institute. Armonk, NY: M.E. Sharpe, Inc. Poole, William. 2000. “Monetary Aggregates and Monetary Policy in the Twenty-First Century.” The Evolution of Monetary Policy and the Federal Reserve System Over the Past Thirty Years. Conference Proceedings, Federal Reserve Bank of Boston (October).

FINANCIAL MARKETS CENTER

27

Sellon, Gordon H., Jr. 2002. “The Changing U.S. Financial System: Some Implications for the Monetary Transmission Mechanism.” Federal Reserve Bank of Kansas City Economic Review (First Quarter).

U.S. House of Representatives. 1935. Banking Act of 1935: Hearings on H.R. 5357. Committee on Banking and Currency. Washington, DC: Government Printing Office. …………1972. Foreign Experience with Monetary Policies to Promote Economic and Social Priority Programs. Staff Report of the Committee On Banking and Currency. Washington, DC: Government Printing Office.

Tietmeyer, Hans. 1993. “Overview.” Changing Capital Markets: Implications for Monetary Policy. Symposium sponsored by the Federal Reserve Bank of Kansas City (August 19-21). U.S. Department of Commerce. 2002. “The International Investment Position of the United States at Yearend 2001.” Survey of Current Business, Vol.82, No.7 (July). Washington, DC: Government Printing Office.

……….1976. International Banking: A Supplement to a Compendium of Papers Prepared for the FINE Study. Staff Report of the Committee on Banking, Currency and Housing. Washington, DC: Government Printing Office.

FINANCIAL MARKETS CENTER

PO Box 334 Philomont, VA, USA 20131 Phone: 540 . 338 . 7754 Fax: 540 . 338 . 7757 E-mail: [email protected] Web address: www.fmcenter.org

Additional copies of this report are available from the Center at no charge.