After the housing price bubble burst, the Fed expanded its total asset portfolio five‐fold, reaching $4.5 trillion in October 2016 (and remaining there until September 2017) from a starting point of $900 billion in August 2008. With some small asset runoffs since October 2017, the assets currently stand at about $4 trillion. Ordinarily, monetary expansion is intended to increase the growth rate of the broader monetary aggregates to satiate the excess demand for money that arises during a liquidity crisis. But by paying interest on excess reserves (IOER) for the first time, the Fed deliberately prevented the expansion of its liabilities (mostly banks’ excess reserves), which financed the Fed’s asset purchases, from increasing the broader monetary aggregates like M2.
This dramatic change in the size of the Fed’s asset portfolio was accompanied by an equally dramatic change in its composition. Between October 2007 and November 2018, the Fed’s holdings of Treasury securities declined from 88 percent of its asset portfolio to 56 percent, while its holdings of mortgage‐backed securities (MBSs) went from zero to 40 percent. During the crisis, the Fed created lending facilities for nonbanks and began accepting riskier assets as collateral for loans. It also stood ready to lend to insolvent banks (e.g., standby credit lines for Citibank and Bank of America) and to purchase dodgy assets from nonbanks (e.g., Bear Stearns and AIG). The Fed discarded its longstanding practice of purchasing almost exclusively short‐term Treasury securities: it switched into longer‐term Treasuries (“Operation Twist 2”) and purchased $1.7 trillion worth of MBSs.
In all of its midcrisis and postcrisis improvisation, the Fed departed from a focus on overall market liquidity and stability of aggregate demand. It allocated credit to specific firms and sectors at the expense of the general market. It also greatly expanded its exercise of powers under Article 13 (3) of the Federal Reserve Act (Meltzer 2011). By lending on highly questionable collateral at subsidized rates, it departed dramatically from Walter Bagehot’s classical lender‐of‐last‐resort doctrine (Bagehot 1873; Hogan, Le, and Salter 2015). The Fed’s actions can be described as “preferential credit allocation” (White 2015), that is, as a move toward greater top‐down financial flows (Hummel 2011).
The Fed’s choices of unconventional monetary and credit allocation policies during and after the Great Recession have reopened a discussion of the political economy of Fed policymaking that had gone largely dormant during the Great Moderation. In this article, we offer a public‐choice account of the Fed’s unprecedented response to the Great Recession. We consider the Fed’s reluctance to pursue monetary policy “normalization” in this light. In theory, the Fed could readily shrink its bloated balance sheet and return it to normalcy. In practice, the Fed has dragged its feet under pressure from political, bureaucratic, and private interests. We conclude that the case for strict rules designed to limit the range of central bank actions, and the need to consider institutional arrangements that offer an alternative to central banking, are stronger than ever.
Our inquiry is rooted in a long‐standing literature that takes a “cynical” public‐choice approach to explaining Fed policymaking (see, e.g., Kane 1980 and Havrilesky 1990), in contrast to the “utopian” view that the Fed aims only at advancing the public interest. The renowned monetary historian Allan Meltzer concluded: “History does not offer evidence of [the Fed] seeking to optimize policy in the interests of consumer welfare” (Meltzer 2011: 47). Given how politics and powerful private interest groups shaped the legislation that created the Federal Reserve System (see Selgin 2016), it should not be surprising that elected officials, financial‐market actors, and its own bureaucratic imperatives have continued to shape Fed policies to the present day. Like any individual or firm or other agency, the Fed’s decisions can be explained as responses to the incentives and constraints it faces (Wagner 1986: 519).
Preferential Credit for Primary Dealers
In its initial response to the subprime mortgage crisis, the Fed redistributed liquidity from the general market toward financially suspect but “systemically important” financial institutions. It lent on highly questionable collateral and often at subsidized rates. It later made outright asset purchases designed to raise asset prices, specifically in the housing sector (Thornton 2015; Hummel 2011; Goodfriend 2014; White 2015).
With rising mortgage defaults, especially on adjustable‐rate mortgages, the prices of MBSs began to decline in late 2007 and early 2008. Many investment banks were exposed to substantial losses on their MBS holdings (Gorton 2010) and had difficulty rolling over short‐term funding in the form of overnight repos and commercial paper (Hummel 2011). Bear Stearns experienced a sudden stop in its funding in March 2008 when its short‐term funders suspected (rightly) that it was insolvent.
Had the Fed followed the modern prescription for a lender of last resort, it would have let Bear Stearns fail while providing sufficient liquidity to the banking system as a whole. Instead, the New York Fed created “Maiden Lane LLC” to purchase nearly $30 billion of dubious mortgage‐related assets from Bear Stearns to sweeten JP Morgan Chase’s acquisition of the firm, thereby shielding Bear’s bondholders and other lenders from losses.
In the same month, the Fed also created a special emergency lending program for “primary dealers,” a group of broker‐dealers and investment houses (including Citigroup, Goldman Sachs, J. P. Morgan Securities, Merrill Lynch, Morgan Stanley, and Wells Fargo Securities) that serve as regular counterparties in the Fed’s bond purchases and sales. The Fed treated the primary dealers as “too big to fail” on the grounds that failures would disrupt the traditional money‐supply transmission mechanism (Fisher and Rosenblum 2009; Selgin 2012). Ironically, the Fed itself would soon disrupt the traditional money‐supply mechanism by paying IOER, which severed the linkage between open‐market operations and broader monetary aggregates. Under a more robust institutional arrangement for open‐market operations, as proposed by Selgin (2012), the Fed would not have a reason to rescue broker‐dealers.
The Fed began funneling credit to the primary dealers for the first time through the creation of two new lending facilities: (1) the Primary Dealer Credit Facility (PDCF), which extended collateralized loans to primary dealers; and (2) the Term Securities Lending Facility (TSLF), which allowed primary dealers to swap riskier assets on their balance sheets for Treasury securities to employ as collateral for PDCF loans. Acharya and Öncü (2010: 337) describe the creation of these special lending facilities to grant credit to well‐connected nonbank entities as the “most radical shift in monetary policy since the Great Depression.” But their creation is better described as a radical shift from monetary to credit policy (Goodfriend 2014).
The Fed held the monetary base constant for many months, by selling off $310 billion in Treasury securities to sterilize the special lending, which reallocated funds toward favored institutions and away from the rest of the banking sector.1 As Robert Eisenbeis (2010: 287) noted, the PDCF and TSLF served not to maintain liquidity in the market (as was warranted at the time) but to “reallocate to primary dealers reserves that would have otherwise been available to smaller banks” to support lending in the general economy.
Preferential Credit Allocation to the Housing Sector
In response to the developing recession, the Fed began to expand the monetary base dramatically, but it simultaneously adopted policies that sterilized the potential impact on broader monetary aggregates. Between September 3 and November 12, 2008 — as the Fed increased credit to banks, primary dealers, and foreign central banks — the monetary base doubled from $850 billion to $1.7 trillion. Simultaneously the Fed introduced two new policies to offset the impact on the broader monetary aggregates.
First, the Fed and the Treasury introduced the Supplemental Financing Account (SFA), whereby the Treasury’s account balance at the Fed grew by $560 billion to keep those funds from swelling bank reserves. The Federal Reserve Bank of New York (2008) stated: “Funds in this account serve to drain reserves from the banking system, and will therefore offset the reserve impact of recent Federal Reserve lending and liquidity initiatives.” As Alan Blinder (2010: 468) observed, this operation breached the traditional “wall between fiscal and monetary policy.”
Second, and more significant, the Fed began paying IOER in October 2008. As Ben Bernanke (2015: 325) noted, this policy was deliberately designed to reward banks for not lending their excess reserves — thereby preventing the broader money aggregates from swelling and preventing the federal funds rate from moving below the Fed’s target. Banks responded by dramatically increasing their holdings of excess reserves from less than $2 billion to $767 billion by December 2008. By paying generous interest rates on reserves (by comparison to Treasury yields at the time) the Fed in effect borrowed back the injected funds (Hummel 2011).
The same pattern prevailed in the Fed’s subsequent “quantitative easing” (QE) programs. The point of QE of the monetary base was evidently not to expand the money supply in the hands of the public. The growth of M2 hardly budged. The point was to finance bailouts (such as $85 billion to AIG), targeted lending facilities, and purchases of MBSs. When the Board of Governors (2008a) stated in its press release on October 6, 2008, that “the payment of interest on excess reserves will permit the Federal Reserve to expand its balance sheet as necessary,” they meant that it will allow the Fed to purchase MBSs and lend to favored recipients (including investment houses, money market mutual funds, and AIG) without creating a corresponding rise in M2 and the price level.2
In December 2008, the Fed announced that it would purchase $1.25 trillion in MBSs in addition to another $250 billion in agency debt and commercial paper. The policy, known as “QE1,” abandoned the Fed’s longstanding policy of buying predominately short‐term Treasuries. The Fed sought to drive housing prices back up and reduce risk premiums in the housing finance market. According to the Fed’s press release from November 25, 2008, “This action is being taken to reduce the cost and increase the availability of credit for the purchase of houses, which in turn should support housing markets and foster improved conditions in financial markets more generally” (Board of Governors 2008b). The Fed did not acknowledge that directing a larger share of credit toward housing markets meant a smaller share toward other market segments.
In September 2011, the Fed introduced “Operation Twist 2,” whereby it sold short‐term Treasuries and bought long‐term Treasuries in an effort to lower long‐term interest rates relative to short‐term rates. The express purpose was to benefit housing finance firms, including the government‐sponsored enterprise (GSE) guarantors and private holders of MBSs, by lowering rates on 30‐year fixed‐rate mortgages and, correspondingly, raising house prices. The operation was enlarged in June 2012. In September 2012, the Fed introduced QE3, a plan to buy $40 billion in MBSs per month, plus some Treasuries, until further notice.
To be clear, the Fed’s preferential redirection of credit was not necessary for the Fed to run an expansionary monetary policy or to fight recession. Open‐market purchases of Treasuries could have accomplished any desired expansion of monetary aggregates.
In fact, the Fed did not run an expansionary monetary policy (relative to trend) during the recession, or even an anticontractionary monetary policy. To Chairman Bernanke’s chagrin, nominal GDP actually shrank during 2009. Both the price level and real output fell as the Fed allowed an unsatisfied excess demand for money to develop. In other words, the Fed failed to offset a sharp decline in the velocity (turnover) of M2.
The primary beneficiaries of the Fed’s credit policies were financial firms heavily invested in housing. Through the final quarter of 2008, the Fed made clear its intention “to help reduce the cost and increase the availability of residential mortgage credit” (Board of Governors 2008b). The reallocation of credit to housing finance came at a substantial cost to the rest of the real economy. Several classes of credit demanders — such as small businesses, college students, and auto buyers — complained about “tight credit conditions,” while the Fed was busy diverting credit to housing (White 2015: 20). The U.S. flow of funds accounts show that despite the decline of housing construction, the growth of mortgage credit remained slightly positive throughout late 2008 and 2009, while credit to other parts of the private sector, such as nonmortgage consumer credit and total business credit, contracted (Board of Governors 2014: 3). Credit allocation policies therefore delayed the necessary adjustment in financial intermediation away from overinvestment in housing and toward more sustainable lending.
The Fed was not forthcoming about the details of the recipients and the terms of its lending. Only a freedom of information request by Bloomberg News revealed the extent to which the Fed was lending at below market interest rates. Reporters estimated that the total subsidy to the borrowers was roughly $13 billion (Ivry, Keoun, and Kuntz 2011).
The Fed’s Move into Fiscal Policy
Borrowing money and spending it on bailouts, or relending it to subsidize special interests, is traditionally the role of Congress rather than the Fed. It is a fiscal policy, not a monetary policy — as it neither aims to provide liquidity to the financial system as a whole, nor to alter the trajectory of the money supply to achieve macroeconomic objectives (see Buiter 2009; Blinder 2010: 476; Meltzer 2011: 47; Cargill and O’Driscoll 2013: 428).
With its $4 trillion balance sheet, the Fed now allocates a larger proportion of the nation’s credit than at any time in its history. Its evident willingness to bailout failing financial firms invites moral hazard. Because the Fed allocates funds with discretion, it invites rent seeking — that is, socially unproductive lobbying efforts from would‐be recipients of preferential credit.3 Although the postcrisis Dodd‐Frank Act limits Fed rescues to “broad‐based lending programs,” this appears merely to ban single‐firm rescues like the AIG or Bear Stearns interventions. It does not limit preferential allocation of funds to the firms in a particular segment of the financial industry, say housing finance. As Lee Hoskins and Walker Todd (2018) stated, “The Dodd‐Frank Act of 2010 imposed new conditions on but did not contract the greatly expanded federal safety net and failed to reduce the substantial increase in moral hazard.”
When the Fed redirects funds that would otherwise flow to more productive sectors, it stunts long‐run overall economic growth. Thus, the Fed’s preferential credit programs likely contributed to the sluggishness of the economic recovery after 2009.
Nine years since the official end of the Great Recession, why has the Fed not yet normalized the composition and size of its balance sheet? The special loan facilities have been wound up, but other programs have not. Only in October 2017 did the Fed start, very slowly, the process of not replacing some Treasury and MBSs as they matured. If the aim of QE, including purchases of trillions in MBSs, and of lengthening of the Fed’s securities portfolio, were only to serve the public interest by promoting macroeconomic recovery, it stands to reason that the programs would have ended by now. The measured output gap, which was −6.5 percent in 2009, became slightly positive in 2018: actual output exceeds estimated full‐employment output (Quandl 2018). The current unemployment rate, at 3.7 percent, is well below the estimated natural or full‐employment rate.
The Fed would serve the public interest by minimizing its intervention into the allocation of credit and returning financial intermediation to private firms subject to market discipline. Delay in doing so suggests that the Fed places a higher value on special interests — particularly, the Treasury, the Fed bureaucracy, and those of rent‐seeking financial institutions that have benefited from the programs in place — rather than on the welfare of society as a whole.
The U.S. Treasury has received larger income transfers from the Fed with the expansion of the Fed’s balance sheet and the change in its composition from short‐term Treasuries to higher yielding (riskier) MBSs and longer‐term Treasuries. In the five years before the financial crisis and QE, the Fed remitted an average of $29 billion per year to the Treasury. From 2010 through 2017, the Fed remitted an average of $86.1 billion per year.4 Total remittances for the 2010–17 period total $689 billion, which is $457 billion more than if remittances had remained at $29 billion per year. Given that the Fed’s interest income on Treasury securities comes from the Treasury to begin with, we can equally think of the Fed’s QE2 and QE3 purchases of Treasury securities as simply relieving the Treasury of its debt service burden on those securities.
The Fed’s own budget has expanded with its balance sheet. According to the Fed’s 2017 Annual Report, “From the actual 2008 level to the budgeted 2018 amount, the total operating expenses of the Federal Reserve System have increased an average of 4.7 percent per year. Over the same period, nondefense discretionary spending by the federal government has increased an average of 1.4 percent per year.”5 This trend may reflect a variety of factors, but it is consistent with Mark Toma’s (1982) hypothesis that the Fed retains some share of additional seigniorage. The Fed cannot use its retained seigniorage to pay its employees bonuses, or to pay higher dividends to its member‐bank shareholders (dividends are capped by statute), but it can increase expenditures on its own operations, especially by hiring more staff.6
As the logic of collective action (Olson 1965) implies, the benefits of the Fed’s current policies are concentrated on the aforementioned interest groups and the costs are dispersed throughout society. The alignment of these interests favors a continuation rather than a reversal of the postcrisis status quo. Fed officials have accordingly shown very little concern for ending their credit allocation policies promptly or for shrinking the size of their balance sheet.7 Reducing excess reserves to a more normal level means selling off MBSs and long‐term Treasuries, and thus depressing their prices — which is contrary to the interests of housing finance firms. On October 25, 2018, Richard Clarida, vice chairman of the Fed’s Board of Governors, gave a talk at the Peterson Institute on the “Outlook for the U.S. Economy and Monetary Policy.” Although he used the term “normalization,” he said nothing about shrinking the balance sheet, ending Operation Twist 2, or IOER (see Clarida 2018).
Conclusion: Toward a Nondiscretionary Future
The highly discretionary nature of the Federal Reserve’s credit policies means that a great deal rides on the theoretical views and tactical concerns of the Fed’s leadership. Under a purely discretionary regime, future Fed policy is highly uncertain. That is why Milton Friedman (1962: 50) argued that “any system which gives so much power and so much discretion to a few men … is a bad system.”
One way to limit discretion is through institutional reform that would replace the present‐day Fed with an arrangement bound by a strict constitution. Many authors have written about desirable features and characteristics of a monetary constitution (see, e.g., Yeager 1962; White, Vanberg, and Köhler 2015; Buchanan 2010; Horwitz 2011; and Salter 2014). Regardless of the specifics, under a constitutional approach either no monetary authority exists, or it is subject to a binding rule specifying a policy target and delineating allowable actions in pursuit of that goal.
A movement to tie the Fed’s hands cannot be expected to arise from the Fed’s current beneficiaries — namely, Congress, the Fed’s leadership, or the recipients of Fed largesse in the financial industry and academia. For this reason Wagner (1986: 532) argued that “monetary reform without political reform to redress the rent‐seeking excesses of prevailing political institutions seems likely to be a short‐lived aberration.” Thus, the goal of constitutional monetary reform needs to be linked to the broader goal of constitutional fiscal reform to adopt what Buchanan (1962) called a “generality norm,” requiring that federal programs be beneficial to citizens in general, rather than providing rents to some at the expense of others. Concerned scholars and citizens will have to propose appropriate rules for fiscal as well as monetary policies, and offer guidance on how to move toward a rules‐based monetary and fiscal regime.
Acharya, V. V., and Öncü, T. S. (2010) “The Repurchase Agreement (Repo) Market.” In V. Acharya, T. F. Cooley, M. P. Richardson, and I. Walter (eds.) Regulating Wall Street, 319–50. Hoboken, N.J.: Wiley, for the New York University Stern School of Business.
__________ (2013) “Transcript of Chairman Bernanke’s Press Conference” (June 19). Available at www.federalreserve.gov/mediacenter/files/FOMCpresconf20130619.pdf.
Board of Governors of the Federal Reserve System. (2008a) “Press Release” (October 6). Available at www.federalreserve.gov/monetarypolicy/20081006a.htm.
__________ (2008b) “Minutes of the Federal Open Market Committee.” (November 25). Available at www.federalreserve.gov/monetarypolicy/files/fomcminutes20181108.pdf.
__________ (2011) “Minutes of the Federal Open Market Committee” (June 21–22). Available at www.federalreserve.gov/monetarypolicy/files/fomcminutes20110622.pdf.
__________ (2014) “Minutes of the Federal Open Market Committee” (July 29–30). Available at www.federalreserve.gov/monetarypolicy/files/fomcminutes20140730.pdf.
__________ (2016) “Press Release” (November 23). Available at www.federalreserve.gov/newsevents/pressreleases/bcreg20161123a.htm.
Clarida, R. (2018) “Outlook for the U.S. Economy and Monetary Policy.” Speech delivered at the Peterson Institute for International Economics, Washington D.C. (October 25). Available at www.federalreserve.gov/newsevents/speech/files/clarida20181025a.pdf.
Federal Reserve Bank of New York (2008) “Statement Regarding Supplementary Financing Program” (September 17). Available at www.newyorkfed.org/markets/statement_091708.html.
Goodfriend, M. (2014) “The Case for a Treasury‐Federal Reserve Accord for Credit Policy.” Testimony before the Subcommittee on Monetary Policy and Trade of the Committee on Financial Services, U.S. House of Representatives (March 14).
Hoskins, W. L., and Todd, W. F. (2018) “Twenty Years after the Fall of the Berlin Wall: Rethinking the Role of Money and Markets in the Global Economy.” Levy Economics Institute, Working Paper No. 908.
Quandl (2018) “United States Output Gap, % of Potential GDP.” Retrieved from Quandl: www.quandl.com/data/ODA/USA_NGAP_NPGDP-United-States-Output-Gap-of-potential-GDP.
__________ (2016) “‘Boss’ Aldrich and the Founding of the Fed.” Cato at Liberty (July 8). Available at www.cato.org/blog/boss-aldrich-founding-fed.
1 Between July 2007 and September 2008, the Fed’s holding of Treasuries fell to $480 billion from $790.6 billion. For this reason, the monetary base only increased by 2.24 percent between August 2007 and August 2008, well below its 7.5 percent annual average over the preceding two decades.
2 Some of the Fed’s targeted credit programs included the Commercial Paper Funding Facility, the Term Asset‐Backed Securities Loan Facility, the Revolving AIG Credit Facility, and Maiden Lane II and III, which benefitted major AIG counterparties — particularly Goldman Sachs (see GAO 2011).
3 For instance, lobbyists from the Commonwealth of Puerto Rico sought the Fed’s aid to avoid default on Puerto Rican bonds in 2015.
4 Data from the Fed’s January 10, 2018, press release, www.federalreserve.gov/newsevents/pressreleases/other20180110a.htm. These remittances exclude the one‐time $19.3 billion transfer from the Fed’s capital in 2015.
6 “The FAST Act reduced the dividend rate applicable to Reserve Bank depository institution stockholders with total assets of more than $10 billion (large member banks) to the lesser of 6 percent or the most recent 10‐year Treasury auction rate prior to the dividend payment” (Board of Governors 2016).