Notes from the Vault
Larry D. Wall
The asset side of the Federal Reserve's balance sheet has attracted considerable attention since the initiation of the Large Scale Asset Purchase program (more commonly called quantitative easing or QE) in 2008. The funding of those assets with liabilities and equity has received comparatively less attention until recently. However, interest in the fiscal implications of the funding side has grown with recent congressional action to fund an increase in the federal transportation budget in part by reducing the dividends paid on bank stock holdings and transferring part of the Federal Reserve Banks' surplus accounts to the Treasury. This raises the question, is this a transfer of wealth from the Fed to the U.S. taxpayer? Is there a free lunch to be had at the Fed?
This post is a primer on the implications for the Treasury's fiscal position of the Federal Reserve's liabilities and equity, including the surplus account, paid-in capital, outstanding Federal Reserve notes, and commercial bank deposits (reserve accounts).1 In order to simplify the analysis, and because it's a reasonable first approximation, I will assume that how the Federal Reserve funds its assets is independent of which assets it holds and independent of the Federal Reserve's revenue and expenses from operations. In other words, the analysis here is valid irrespective of the size and asset composition of the Federal Reserve's balance sheet and level of operating expenses.
Federal Reserve's balance sheet and income statement
To understand the significance of the Federal Reserve Banks' liability structure, I start with an overview of its financial statements. The Federal Reserve Board publishes audited financial statements that consolidate the 12 Reserve Banks into a single entity so it's possible to get a comprehensive view of the Federal Reserve's financial condition. Consolidated total assets of the Federal Reserve Banks were $4,497 billion as of December 31, 2014. Almost all the assets were acquired for monetary policy purposes and held in the System Open Market Account (SOMA), including $2,596 billion in Treasury securities and $1,789 billion of federal agency and government-sponsored enterprise mortgage-backed securities.
The primary liabilities of the Federal Reserve Banks are Federal Reserve notes outstanding ($1,299 billion), securities sold under agreements to repurchase ($510 billion), deposits by depository institutions ($2,378 billion), and Treasury general account deposits ($223,452 million). The Federal Reserve's equity consists of paid-in capital of $28.6 billion and surplus of $28.6 billion. For the purposes of this analysis, the securities sold under agreements to repurchase will be treated as essentially the equivalent of bank deposits.2
The Federal Reserve Banks earned $116 billion in interest last year, primarily interest payments on securities held in SOMA. The Federal Reserve also earned $1 billion from its provision of services to banks and government agencies. After interest payments to banks, operating expenses, and miscellaneous other items, the Federal Reserve Banks collectively earned $101 billion before remittances to the U.S. Treasury. Next on the Reserve Banks' income statement is remittances of $97 billion to the Treasury, resulting in net income of $4 billion. This net income was divided three ways: $1.6 billion in dividends to member banks (who provided paid-in capital of $28.6 billion in the form of stock ownership); $1 billion was retained in the Reserve Bank's surplus account; and the remainder was recorded in other comprehensive loss (in accord with conventional accounting procedures).
The ordering in the income statement may seem to imply that remittances are paid to the Treasury before net income is calculated; however, Carnegie Mellon Professor Marvin Goodfriend explains that in practice it is actually the reverse, the remittances are determined by the Federal Reserve's policy of maintaining a surplus account equal to the amount of paid-in capital. Thus, the remittance is the residual left over accounting for dividends, the increase (or decrease) from paid-in capital, and other comprehensive losses (or income). To verify this, one need only observe the variation over time in remittances, dividends, and changes to paid-in capital.
The following subsections analyze the implications of the equity and liabilities of the Federal Reserve Banks for the Treasury's financial position subject to two simplifications. First, it ignores the Treasury's deposits at the Federal Reserve. These deposits are roughly akin to the Treasury's checking account, which is to say the amount held in the account is determined by the Department of the Treasury based on its needs.3 Second, the interest paid both on securities sold under agreements to repurchase and on bank deposits will be combined for the purposes of this analysis, as both rates are set, in coordination, to achieve monetary policy objectives.
Paid-in capital and dividends
Nationally chartered banks and state chartered banks that become members of the Federal Reserve System must subscribe to stock equal to 6 percent of that bank's capital and surplus, and must pay in one-half of that amount (12 United States Code 209).4 The Federal Reserve Board observes that "The stock may not be sold, traded, or pledged as security for a loan." In return for this investment, the bank is entitled by law to dividends equal to 6 percent of its paid-in capital (12 USC 289), which are cumulative.
The 6 percent dividend rate paid on Federal Reserve stock was set at, and has been constant since, the founding of the Federal Reserve. However, recently Congress passed the Fixing America’s Surface Transportation Act, which reduces the dividend for banks with consolidated assets exceeding $10 billion to the smaller of the 10-year Treasury note rate or 6 percent. At present, the 10-year Treasury rate is 2.2 percent, thus cutting the affected banks' dividends by nearly two-thirds. Although the cut in the dividends could be justified by the currently low market rates, the retention of the 6 percent cap raises the possibility that the dividend rate will fall well short of Treasury rates at some future date, such as was the case from the early 1970s until after 1991.
The Federal Reserve surplus account is similar to what most corporations called retained earnings, that is, it is the accumulation of comprehensive net income (after remittances to the Treasury) that has not been paid out as dividends. An obvious question, is the Treasury's fiscal position somehow weakened by the Federal Reserve retaining some of its earnings in its surplus account rather than paying higher remittances to the Treasury? The answer is no, as explained by Goodfriend. Indeed, from the perspective of the Treasury's financial condition, spending the Federal Reserve's surplus is identical to increasing the fiscal deficit.
First, suppose that the Federal Reserve had remitted the earnings to the Treasury. In this case, the Treasury would be able to issue less debt, reducing the Treasury's interest expense. Suppose, however, that the Federal Reserve retains the earnings in its surplus account and uses the extra funds to expand the Federal Reserve's Treasury portfolio by an equal amount. The Federal Reserve would then have higher earnings in the next period, which it would then remit to the Treasury. These two alternatives would have essentially the same implications for Treasury interest payments net of Federal Reserve remittances. In other words, the Fed's holding of a surplus account does not increase the burden of interest payments on U.S. taxpayers.
The surplus account has also been in the news, with the Fixing America’s Surface Transportation Act requirement that the Federal Reserve remit part of its surplus to the Treasury; the remitted surplus would then be counted as revenue by Congress and could be used to fund higher spending. Former Federal Reserve Chairman Ben Bernanke has characterized this use of the remittance as budgetary "sleight-of-hand." The Federal Reserve would need to sell Treasury securities to the public to obtain the funds for remittance, but the Treasury could not buy back the securities, as the funds would have gone to higher spending. Thus, the Treasury's interest payments net of Federal Reserve remittances would increase in future years.5
Federal Reserve notes outstanding
Federal Reserve notes are the currency being held by the U.S. public and by many foreign parties. These notes do not pay interest to their holders. Thus, Federal Reserve notes effectively provide an interest-free loan to the Federal Reserve, which, given Federal Reserve remittance policies, implies they represent an interest-free loan to the Treasury.
If Federal Reserve notes outstanding represent such a good deal for the Treasury, why don't we issue more of them? The answer is that the amount of currency relative to bank deposits is determined by the public. The Federal Reserve (with the help of the Bureau of Printing and Engraving) will supply as much currency to the public as it wishes to hold, but the Federal Reserve does not have the mechanism or desire to make people hold more currency than they want.
Deposits by depository institutions
When the Federal Reserve buys a Treasury security, its pays for it by crediting the deposit balance of a depository institution.6 The Federal Reserve pays interest on these deposits, currently at the rate of 25 basis points (0.25 percentage points per year). Further, the Federal Open Market Committee has indicated that it intends to control short-term interest rates (for the short and medium term) by raising and lowering the rate paid on deposits. The effect of a higher rate will be to reduce the amount of stimulation that monetary policy provides to aggregate demand and the real economy.
The interest rate the Federal Reserve currently earns on its Treasury securities portfolio is substantially higher than the rate it pays on deposits. So would the Treasury be in a better position if the Federal Reserve bought even more Treasury debt? The answer is no, if one believes in the expectations hypothesis of the term structure of interest rates. Under the expectations hypothesis, the rate paid on Treasury securities should be (approximately) equal to the average rate the Federal Reserve pays on bank deposits over the maturity of the securities. Thus, if the Federal Reserve's revenue from Treasury securities currently exceeds the rate it pays, the relationship will reverse in the future with the Federal Reserve paying more than it receives. The result is that the Federal Reserve will eventually take losses on the part of its portfolio funded by bank deposits that is approximately equal to what the Federal Reserve earns in the early years of this part of its portfolio. Thus, a long-run budget-neutral approach to remittances would actually have the Federal Reserve reduce its current remittances and increase its surplus to cover expected future losses.
In fact, most economists do not believe that the expectations hypothesis is literally true. They believe that investors demand higher returns for making longer-term, illiquid investments. Thus, the rate on the Federal Reserve's Treasury portfolio should be higher than the expected rate on bank deposits, which are essentially overnight deposits. However, risk premiums appear to explain only part of the upward slope in the term structure, as futures contracts on federal funds and eurodollar rates have both been predicting increases in short-term rates.7 If these predictions are realized, Federal Reserve earnings from that part of its portfolio funded by bank deposits will decrease.8
The relationship between the Federal Reserve Banks' balance sheets and the Treasury can be difficult to understand. This primer explains how the relationship depends to an important degree on the composition of the Federal Reserve's liabilities and equity.
Larry D. Wall is the executive director of the Center for Financial Innovation and Stability at the Atlanta Fed. The author thanks Paula Tkac for helpful comments on the paper. The views expressed here are the author's and not necessarily those of the Federal Reserve Bank of Atlanta or the Federal Reserve System. If you wish to comment on this post, please email firstname.lastname@example.org.
1 The Federal Reserve has also experimented with the use of reverse repurchase agreements to help control short-term interest rates, which could be a major source of funding of the Federal Reserve's securities holdings. For the purposes of this discussion, however, the cost of reverse repurchase agreements will be treated as essentially equivalent to the cost of bank reserve accounts. See my previous post for more of a discussion of why the Federal Reserve may use ON-RRP.
2 "Deposits by depository institutions" are the same as depository institutions' "Reserve balances with Federal Reserve Banks." These reserve balances are reported weekly on the Federal Reserve Statistical Release H.4.1.
3 The Treasury does not receive interest on its deposits. However, if the Federal Reserve did pay interest on these deposits that would result in a dollar-for-dollar reduction in the Federal Reserve earnings remittances, leaving the Treasury in no better a position.
4 Although the Federal Reserve member banks own "stock" in the Reserve Banks, this gives them only limited governance power. Banks have three of the nine members on each Bank's board of directors and these directors provide valuable input into the state of their respective District's economy. However, these directors are precluded by law from participating in the selection of the bank's president. Moreover, all Reserve Bank functions are overseen by the Board of Governors of the Federal Reserve System whose members are appointed by the president and confirmed by the Senate.
5 The Federal Reserve could repurchase these securities in its open market operations to expand its portfolio. However, doing so has two adverse implications. First, it would imply subordinating monetary policy to fiscal policy, something that most economists agree is generally a bad idea. Second, it would still not fully compensate the Treasury because the Federal Reserve's repurchase of the securities would increase banks' deposits at the Federal Reserve and these bank deposits (unlike the surplus account) are interest bearing.
6 Moreover, the deposits created by the Federal Reserve must be held within the banking system. Banks can transfer these deposits among themselves, but no bank can reduce aggregate bank deposits at the Federal Reserve. The two primary ways for reducing these deposits are outside the banks' control. First, the Federal Reserve can reduce deposits by selling some of its securities (for which it is paid by a reduction in bank deposits at the Federal Reserve). The second way to reduce bank deposits at the Federal Reserve occurs when the public converts part of its deposits in commercial banks into currency. However, the banks have very little influence over the public's preference for holding funds in bank deposits versus holding them in the form of currency.
7 Richard Crump and other Federal Reserve Bank of New York economists explain how various interest rate derivatives may be used to forecast future short-term rates, including the Federal Reserve's policy rate. The CME Group provides information on current quotes for its 30 Day Federal Funds Futures and for three-month Eurodollar Futures.
8 For an analyses of Federal Reserve earnings under various interest rate scenarios, see Federal Reserve Board economists Seth B. Carpenter, Jane E. Ihrig, Elizabeth C. Klee, Daniel W. Quinn, and Alexander H. Boote, Federal Reserve Bank of San Francisco economists Jens H.E. Christensen, Jose A. Lopez, and Glenn D. Rudebusch, and Federal Reserve Bank of New York economist Marco Del Negro and Princeton Professor Christopher A. Sims.