How big should central bank balance sheets be?

In 2007, the Fed’s balance sheet was less than $1 trillion. Today, it is nearly $4.5 trillion. The U.S. experience is far from unique. Since 2007, global central bank balance sheets have nearly tripled to more than $22 trillion as of mid-2014. And, the increase is split evenly between advanced and emerging market economies (EMEs) (see chart).

Central Bank Assets (Trillions of U.S. Dollars)

Source: National central banks.

Source: National central banks.

So what’s the right size? The answer depends on the policy goals and the nature of the financial system. In the case of the Fed, we expect that it will be able to achieve its long-term objectives with fewer than half of its current assets.

The size of central bank balance sheets varies dramatically from country to country (see the two charts below). In advanced economies, the range today is from less than 5% of GDP in Canada to more than 80% in Switzerland. Central banks in EMEs have balance sheets that are nearly twice the advanced-economy norm, ranging from 12% of GDP in Colombia to nearly 140% of GDP in Hong Kong.

Advanced Economy Central Bank Assets (Trillions of U.S. Dollars)

Sources: National central banks and IMF.

Sources: National central banks and IMF.

Sources: National central banks and IMF.

Sources: National central banks and IMF.

We see five primary reasons for these differences:
  • Financial system structure
  • Regulation
  • Payments technology
  • Policy objectives
  • Degree of central bank independence

In financial systems where banks dominate, central bank balance sheets will be larger. Where reserve requirements are high, central banks need bigger balance sheets to satisfy commercial bank reserve needs. Where payments technology relies heavily on currency (rather than electronic transfers), the public holds more cash and commercial banks also seek larger reserve buffers to meet payments obligations. Policymakers focused on exchange rate stability typically need big foreign currency reserves. Finally, central banks that are not politically independent may hold a large volume of assets in their role as the government’s bank, allowing them to target favored assets. In contrast, independent central banks tend to hold small portfolios of relatively short-term government securities. Such “asset neutrality” establishes a bulwark against influence over the central bank by the fiscal authority.

These five factors go a long way toward explaining why emerging market central bank balance sheets tend to be larger than those in advanced economies. Emerging market financial systems are more likely to be bank-based, have high reserve requirements, and rely on currency as the premier means of payment. Their central banks also are more prone to target exchange rate stability and to act as the government’s agent in allocating credit to targeted assets.

The same factors also explain some of the within-category differences. Among advanced economies, for example, Canada has a zero reserve requirement while the Eurosystem is relatively bank-centric, helping to account for the latter’s bigger balance sheet. Among EMEs, China, Hong Kong, Saudi Arabia, and Singapore need a large stock of foreign currency to manage their exchange rates, while Colombia, the Czech Republic, and Poland let their currencies float.

So what factors drove the enormous and widespread balance sheet changes since 2007? Broadly speaking, there are four classes of motives associated with specific policy goals (see this recent paper):

  • Foreign exchange operations
  • Lender of last resort or market function operations
  • Stabilization of aggregate demand at the zero bound
  • Forced government financing (aka fiscal dominance)

Most of the crisis-driven balance sheet growth since 2007 was in the first and second categories, with some in the third. For example, the expanded balance sheets in China, Hong Kong and Switzerland reflect operations to steady exchange rates. Interventions in dysfunctional markets played a significant role in the United States and the United Kingdom. In many countries, including Japan, the United States, and the United Kingdom, demand stabilization at the zero bound eventually became the leading driver. [So far, no leading central bank has been compelled to finance its government.]

The vast balance sheet changes in recent years have been aimed at influencing far more than just the overnight interbank lending rate. U.S. monetary policy makers altered both the size and composition of their balance sheet, initially to counter dysfunctional markets, but eventually to influence both the term structure and credit risk structure of interest rates. By increasing their portfolio size and duration, the Fed compressed the term premium on U.S. government debt. And by purchasing government-insured mortgage-backed securities, the U.S. authorities were able to narrow the yield spread on mortgage loans. (You can find a discussion of various unconventional monetary policies, including balance sheet policies, here.)

But U.S. financial markets are now functioning well, and the aggregate demand rationale for altering the term and risk structure of interest rates is waning. So, what should the Federal Reserve aim to do about the extraordinary size of its balance sheet? The answer requires weighing the benefits and costs of maintaining a large balance sheet in normal times when monetary policy is set by altering the policy interest rate (see table).

 

Maintaining a Large Central Bank Balance Sheet in Normal Times
BenefitsCosts
Ability to influence a range of securities pricesPortfolio risk
Provision of high-quality liquid assets to economySubstitute for private intermediaries
Limit excessive liquidity/maturity transformationRisk to central bank independence

 

We see three main benefits. A central bank that holds a large, diverse portfolio of assets has the ability to influence a variety of securities prices and can provide a large volume of high-quality liquid assets to firms and households. In addition, by offering banks a large level of reserves at low cost (the leading central banks now all pay interest on reserves), central banks discourage bank risk-taking in the form of liquidity and maturity transformation. Put differently, paying interest on reserves can be viewed as a price-based version of a liquidity requirement (something we wrote about in a recent post).

We also see three main costs. First, a large balance sheet will mean portfolio risk, partly in the form of high leverage. For example, the Federal Reserve System currently has total assets of more than 150 times its capital. And, its holdings of long-term bonds expose it to substantial risk of loss if interest rates rise. Yet, quite a few central banks probably operate today without positive net worth – a state of affairs that poses risks only if their capital shortfall subjects the central bankers to political pressures. (You can find a general discussion of central bank finances here.)

Second, if central banks remain big players in a broad array of financial markets, they substitute their own balance sheet for that of private intermediaries. When markets were dysfunctional – as in the heat of the financial crisis – this was desirable. In normal times, private intermediaries have strong incentives to allocate savings to their most efficient uses, and are less prone than a central bank to political influence.

Finally, and most importantly, a government that comes to rely on a central bank to hold a large quantity of its debt poses a threat to central bank independence. History teaches us that when the central bank loses control of its balance sheet to the fiscal authority, the results can be catastrophic.

So, where does all of this leave us? In a recent statement of principle, the FOMC expressed the intention over the long run to “hold no more securities than necessary to implement monetary policy efficiently and effectively.” In addition, it will “hold primarily Treasury securities” to limit the balance sheet’s effect “on the allocation of credit across sectors of the economy.”

The stated goal of a minimal, asset-neutral balance sheet is well suited to a highly independent central bank in normal times. The Fed still retains the authority, in some future crisis, to intervene in dysfunctional markets. Its willingness to intervene provides the financial system valuable insurance against tail risk, but has negligible balance sheet impact in normal times.

So how much is the Fed’s balance sheet eventually likely to shrink? Plenty. There seems little desire to replace its $1.7 trillion of mortgage-backed securities as they mature. The remaining assets would be sufficient to support more than $1 trillion of bank reserves. But monetary policy can be implemented efficiently with few (or even no) reserves. Our view is that the most likely result is a long-run Fed portfolio at most half its current size, even if policymakers decide they wish to provide significant high-quality liquid assets, and substantially less if they do not.

It will take many years to get to this lower long-run level of assets, but making that transition ought not be a problem for setting monetary policy. With the ability to pay interest on reserves, combined with other new instruments like reverse repo, the Fed should be able to adjust its policy interest rate regardless of the size of its balance sheet.