There are two reasons why the Fed, or any other central bank, should not act as a quasi-fiscal branch of the government, other than paying to the Treasury in taxes the profits it makes in the pursuit of its mandated macroeconomic stability objectives (maximum employment, stable prices and moderate long-term interest rates in the case of the Fed) and its appropriate financial stability objectives. The appropriate financial stability objectives of the central bank are those that involve providing liquidity, at a cost covering the central bank’s opportunity cost of non-monetary financing, to illiquid but solvent financial institutions.
Any action going beyond that, such as the recapitalisation of insolvent banks through quasi-fiscal subsidies, ought to be funded by the Treasury. The central bank should be involved only as an agent of the Treasury – an expert assistant. It should not put its own conventional or comprehensive balance sheet at risk.
The two arguments against the central bank acting as a quasi-fiscal agent are, first, that acting as a quasi-fiscal agent may impair the central bank’s ability to fulfil its macroeconomic stability mandate and, second, that it obscures responsibility and impedes accountability for what are in substance fiscal transfers. In the US such actions subvert the Constitution, which clearly states in Section 8, Clause 1, that the power to tax and spend rests with the Congress: “The Congress shall have Power to lay and collect Taxes, Duties, Imposts and Excises, to pay the Debts and provide for the common Defence and general Welfare of the United States; but all Duties, Imposts and Excises shall be uniform throughout the United States.”.
If, as happened in the USA on a vast scale, the central bank allows itself to be used as an off-budget and off-balance-sheet special purpose vehicle of the Treasury, and refuses to provide to the Congress some of the information essential for the quantification of the fiscal transfers it has made, the central bank not only subverts the constitution. By attempting to hide contingent commitments and to disguise de-facto subsidies by not divulging relevant information on the terms on which the central bank has offered financial assistance, it undermines its own independence and legitimacy and impairs political accountability for the use of public funds – ‘tax payers’ money’. It is surprising that a country whose creation folklore attributes considerable significance to the principle of ‘no taxation without representation’ would have condoned without much outcry such a blatant violation of the equally important principle of ‘no use of public funds without accountability’. This indeed amounts to a quiet usurpation of the power of the legislature by the central bank.
When the crisis started in August 2007, the Fed’s conventional balance sheet was just under $ 1 trillion – about seven percent of annual US GDP. At its peak, towards the end of 2008, the Fed’s conventional balance sheet was just over $2 trillion, about fifteen percent of annual US GDP. The Bank of England tripled the size of its balance sheet (as a share of GDP) over the same period. I see no problem at all with the size of the balance sheet per se. It is the logical consequence of the central bank, in a liquidity crisis, providing funding liquidity to systemically important financial entities (the lender-of-last-resort function) and market liquidity to markets for systemically important financial instruments (the market-maker-of-last-resort function).
The problem is not the size of the balance sheet but the size of the quasi-fiscal transfers the Fed has made to some of its private counterparties in its myriad interventions since the crisis started.
Let me start by restating that I believe there is no quasi-fiscal role for the Fed other than the one inherent in the pursuit of its macroeconomic objectives and of its legitimate financial stability objectives. The Fed should fund liquidity operations targeted at solvent counterparties. If it is required to deal (as agent of the Treasury) with potentially insolvent counterparties, the credit risk and counterparty risk should be assumed fully by the Treasury. This is not the practice of any of the leading central banks today; prior to the crisis, only the Bank of England came close.
I recognise the practical impossibility of determining whether an institution is illiquid but solvent or illiquid and insolvent. I also recognise that when a financial crisis threatens to engulf systemically important financial institutions, speed is of the essence and only the central bank can intervene with the necessary speed and on the necessary scale – assuming of course that it is domestic currency liquidity or default risk on domestic-currency-denominated securities that is at stake. But the central bank should, when it exposes itself to potential private sector credit risk, only do so as an agent of the Treasury, and not for its own account.
What this means is that whenever the central bank accepts private securities as collateral or purchases private securities outright, it always should do so with a full indemnity from the Treasury against losses due to credit risk.
The benchmark for the central bank should be a “Treasuries only” policy of balance sheet and liquidity management. Under a “Treasuries only” policy, the central bank only buys Treasury securities outright. In repos and other collateralised lending operations, it only accepts Treasury securities as collateral.
In an emergency, where the government requires the good offices of the central bank to stop systemically important institutions from collapsing, the form but not the substance of the ‘Treasuries only’ policy can be relaxed. The UK shows the way as regards outright purchases by the Bank of England of private securities. The UK Treasury and the Bank agree on an upper limit on the amount of private securities that can be purchased by the Bank (currently £ 50 bn) and on the nature of the private securities that can be bought outright. Then the Treasury provides the Bank of England with a full indemnity (guarantee) for any private securities purchased by the Bank up to that limit. That is the right way to separate fiscal policy from monetary and liquidity policy.
The ECB, which is committed to buy a very limited number of private securities outright (it has set itself a limit of €60 bn for covered bond purchases of which, as noted earlier, only €20 bn has been used thus far), does not have such an indemnity from the 16 national Euro area fiscal authorities. The ECB therefore takes credit risk on these outright purchases. Even if this credit risk is priced appropriately ex-ante, the realisation of the risk could blow a hole in the balance sheet of the ECB and reduce its capital. This is not a problem with the current scale of the outright purchase programme, but it puts the camel’s nose firmly in the tent.
The Fed has been taking massive credit risk in its outright purchase programmes. In the original TALF, for instance, up to $1 trillion could be guaranteed by the Fed, but the Treasury indemnity for the programme was capped at $100 bn, leaving the Fed with a potential credit risk exposure of $900 bn. Other Fed programmes too have involved actual or potential exposures to private credit risk that were not guaranteed by the Treasury.
As regards repos and collateralised loans, the most extreme departure from the ‘Treasuries only’ model has been the ECB. The Eurosystem accepts as collateral in repos and at the discount window an astonishingly wide range of private securities, including most asset-backed securities, as long as they have a rating of at least BBB-. This collateral policy has been implemented in such a loose and generous way, that international banks with subsidiaries in the Eurozone have packaged and wrapped securities they could not use as collateral anywhere else in formats that made them eligible collateral at the Eurosystem.
Only when the bank that borrowed from the Eurosystem has become insolvent, as in the case of Kaupthing’s Luxembourg subsidiary and Lehman Europe, has the ECB had to write down its risky exposure. But with many technically insolvent or near-insolvent banks as counterparties (including quite a few of the German Landesbanken and the Spanish Cajas), the true exposure of the ECB is bound to be higher than it is willing to own up to. The Eurosystem has on its books large amounts of loans to dodgy banks secured against poor collateral. The resulting credit risk falls entirely and entirely inappropriately on the ECB.
The Bank of England and the Fed now also accept private securities as collateral in repos at the discount window and at many of the special facilities that were created to resolve the crisis. None of these loans by the central banks to private entities and collateralised against private securities are guaranteed by the national Treasuries.
What makes the problem worse is that all the leading central banks are not just faced with the possibility that, having made a properly priced collateralised loan to a private counterparty, a bad state of the world is realised, the counterparty goes broke and the collateral turns out to be impaired also. When such a double default occurs, the central bank acts in an ex-post quasi-fiscal capacity if there is no full Treasury guarantee or indemnity.
There are good grounds for suspecting that many of these loans were not even priced properly ex ante to reflect the associated credit risk, but were instead handed out on terms that implied an ex-ante quasi-fiscal subsidy. None of the three central banks, the Fed, the ECB or the Bank of England have been willing to reveal how they value illiquid collateral. Requests to make public either the pricing models or the actual valuations of all illiquid private securities offered as collateral have been systematically stonewalled by the central banks. That makes it impossible for external assessors to determine whether an ex-ante subsidy was involved in the terms and conditions of the loans.
The Fed went well beyond even this. In its bail-out of AIG, it refused for a long time to reveal who the counterparties of AIG were that were made whole because of the Fed’s emergency loans to AIG. They were forced to reveal the information in the end, but this does not undo the earlier attempt to hide the identities of the beneficiaries of the Fed’s and the US Treasury’s largesse.
Exiting from unconventional monetary policy means revealing the true extent of the quasi-fiscal transfers handed out and quasi-fiscal taxes imposed by the central banks in their financial operations. At the moment, we really see not much more than the conventional balance sheet presented by the Fed. The Fed’s claims on the private sector are valued in ways that cannot be verified. We know from the reports on the former Bear Stearns assets tucked away in a Delaware-based special purpose vehicle (Maiden Lane) and from the AIG assets stowed in Maiden Lane II and III, that the Fed has got at least some rubbish in exchange for the loans it has provided. How much more write-downs and write-offs will we see?
It is true that central banks can be expected to make a profit on their lender-of-last-resort loans to solvent but illiquid counterparties. There is a wide gap between the liquidation value of the assets offered as collateral in times of stress and the present value of their held-to-maturity cash-flows. Central banks can and do exploit this situation to charge effective interest rates that are not just above the risk-free rate, but also at times above the default-risk adjusted opportunity cost of non-monetary funds to the central bank. The Fed just reported a provisional $14 bn profit from such lending activities.
It is good to know that when liquidity is scarce, the source of ultimate liquidity knows how to make a profit. This profit has no bearing on the question as to which counterparties benefited from quasi-fiscal subsidies from the Fed and from other central banks and in what amount. It is not just the aggregate or net quasi-fiscal subsidy of the central bank that matters. The redistributive quasi-fiscal activities of the central bank don’t necessarily require any net subsidies to the private sector.
Central banks don’t have a redistributive mandate. That job should be left to the Treasury and the legislature. This encroachment of unelected technocrats on the domain of distributive politics is simply not acceptable in an open democratic society. That much of the redistribution effected by the Fed and other leading central banks is consciously hidden by the agency and kept under a cloak of secrecy in the name of counterparty confidentiality, market sensitivity or stigma effects makes it worse.
Exiting from unconventional monetary policy will assist a fuller revelation of the exact nature of the quasi-fiscal actions of the Fed (and to a more limited extent the ECB and the Bank of England). The political benefits from the cleaning of the stables that will, I hope, result from this, will in my view dwarf the economic significance of a successful exit strategy.
 Some spokespersons for the central banks have stated that since they provide information on the haircuts applied to all collateral, including illiquid private assets, all relevant information is in the public domain. That is incorrect. The haircuts are supposed to apply to the price or valuation of the security, not to its notional or face value. Unless we know the valuation to which the haircut is applied, we know nothing.
 On August 25, 2009, Manhattan Chief U.S. District Judge Loretta Preska ruled against the Fed in a Freedom of Information Act lawsuit brought by Bloomberg News. She rejected the argument that loan records are not covered by the law because their disclosure would harm borrowers’ competitive positions. The Fed for the first time had to identify the companies in its emergency lending programs.
 See Fed makes $14bn profit on loans provided during financial turmoil, Financial Times, August 31, 2009.