1. How the Bank of England's independence has been compromised

In the depths of the recession which set in very quickly after the collapse of Lehman Brothers in September 2008, the Bank of England Monetary Policy Committee wanted to ease monetary policy further. Having already reduced short-term interest rates to historically low levels, it embarked on a programme of quantitative easing (QE), in which large quantities of gilts were purchased and lodged in a new Asset Purchase Facility (APF).^{Footnote 1} Initially, in March 2009, the MPC decided to buy £75 billion; by 2013, the QE programme amounted to £375 billion. In August 2016, after a long pause, the MPC decided to buy a further £60 billion of gilts and £10 billion of privately-issued bonds, and to provide £100 billion for a Term Funding Scheme to support bank credit, taking the QE total to £545 billion, of which £435 billion consisted of gilts.

The Bank of England was thought insufficiently capitalised to provide against the risk of such a large holding of assets, and in 2009 it sought and received an indemnity from the government against any losses that it sustained in the QE programme. The indemnity was extended in amount each time the Bank decided that it needed to enlarge the QE programme. At the end of February 2013, the Bank of England's leverage ratio was less than 1 per cent, far below any level thought adequate for a commercial bank.^{Footnote 2}

The Treasury welcomed the programme and agreed willingly to provide the indemnities. However, when the Bank decided in August 2016, after the Brexit referendum, that it needed to extend the programme further, the new Chancellor's letter agreeing to the Governor's request for an enlarged indemnity was noticeably cool. He did not say that he shared the MPC's view that further QE was necessary, and hinted at a degree of caution about the risks and the indemnity:

The Chancellor's caution is understandable. The QE programme is replacing large amounts of long-term government debt with overnight deposits in the Bank of England, and amounts to a very considerable shortening in the average maturity of public sector debt, which leaves the public finances much more exposed to variations in short-term interest rates than hitherto. And the assets against which he was being asked to indemnify the Bank included not only gilts but also corporate bonds and assets taken as collateral under the new Term Funding Scheme. Moreover, there is evidence that some purchases of gilts in the QE programme were conducted in a way which was exposed to gaming by the market, and unnecessarily increased the cost of the purchases.^{Footnote 4}

The Treasury could of course in principle refuse a request to extend the indemnity and, if it did so, the Bank would be unable to extend QE. Or the Treasury could insist on withdrawing the indemnity over a period, in which case the Bank would have to unwind QE. Any such action would have consequences, both for the economy and for the relationship between the government and the Bank of England. Nevertheless, the fact that the Bank of England depends on the Treasury's consent to deploy the main instrument of its monetary policy raises the question, to say the least, of whether its independence in conducting monetary policy has been compromised.

2. Is compromise of independence inevitable?

Central banks in many countries had to expand their balance sheets during the recession that followed the financial crisis, both to try to contain the damage to financial stability and to maintain an expansionary monetary policy at a time when short-term interest rates were already very low. By purchasing large amounts of assets, central banks have unavoidably trespassed on the territory either of the debt management authorities, if they bought government securities, or of the fiscal authorities, if they bought privately-issued securities.^{Footnote 5}

The control of public spending depends on the Treasury being in a position to authorise all spending decisions. Asset purchases by the Bank of England are, in a sense, public spending decisions, because the public finances ultimately bear the risks that they carry but, not being subject to Treasury authorisation, they are an exception to the general rule. The exception is tolerable if the Bank's balance sheet is small, but it becomes increasingly less so as the Bank's balance sheet grows. Issues of demarcation of power and responsibility are bound to arise.

The Bank of England's position is peculiar, however, because of its small capital base. In an obvious sense, the smallness of its capital base has made it weaker, because it has had to ask for indemnities for QE. In another sense, however, its small capital base has made it stronger. It is true that the Treasury could have refused a request for an indemnity, but to do so would have had political and possibly economic consequences which the Treasury might have wanted to avoid. If the Bank judges that the Treasury would shrink from the consequences of refusal, it can count on getting any indemnity it asks for, and can therefore pursue QE unrestrained by any concerns about the amount of risk that was being taken. The Bank of England's scope for action, therefore, depends on the Treasury's preferences and its perception of them. Its own financial situation is irrelevant. This is unsatisfactory. Any central bank must be subject to some limit on the financial risks it can take on its own initiative, but the limit should not depend on the current political situation.

Some economists claim that the Bank of England should not be over-concerned about its own solvency. They point out that the value of the monopoly over note issue that the Bank enjoys is a kind of additional capital; and that other central banks, such as that of Chile, have continued to operate effectively even though they have been insolvent. However, the income from the note issue is transferred, after deduction of expenses, directly to the government, and any change to that arrangement would require legislation. And, more generally, any loss made by the central bank is a kind of public spending, which ought to be authorised by the Treasury.

It follows from all this that the pursuit of QE by the Bank of England had unavoidable consequences for the Bank's financial relationship with the government, which have compromised its independence. The consequences would have been smaller had the Bank had a higher ratio of capital to assets.

3. Is central bank independence important any longer?

It is now widely acknowledged that, with short-term interest rates close to the lowest possible level, the once-clear demarcation between monetary policy, public debt management and fiscal policy has become blurred.

Nevertheless, the arguments for some degree of central bank independence that were carefully elaborated in the 1990s and after still have some force. Above all, there is a risk that governments, if in a position to do so, will use monetary policy to pursue short-term growth by demand expansion at the expense of more longer-term inflation and probably less longer-term growth. Such was Britain's experience in the 1950s, 1960s and 1970s. Some degree of central bank independence is worth protecting.

These arguments have nothing to do with financial stability, and the Bank of England Act 1998, which made the Bank independent in monetary policy, did not make it independent in financial stability. And it is worth recalling that in managing the multiple financial crises of the 19th century, the Bank worked closely with the government.

4. What could be done to protect the Bank of England's independence now?

The Bank's vastly increased assets have been financed by vastly increased liabilities, in the form of bankers' deposits with the Bank of England. My proposal is for a drastic shrinkage of the Bank's balance sheet, together with some increase in the Bank's capital so as to give it the scope for some further QE if it were to see the need for it. Of course, if the Bank saw the need for further QE on a larger scale than its increased capital allowed, it could apply to the Treasury for yet more capital. That would be a limitation on the Bank's independence, but a necessary one.

It would not be wise, or possible, to sell all or most of the bonds in the Bank's portfolio over a short period. Gross gilt sales undertaken by the Debt Management Office currently run at around £130 billion a year, while the total of gilts in the APF is £435 billion, and there are an additional £10 billion of corporate bonds. However, it would be possible for the gilts in the APF to be exchanged with the Treasury for newly-issued Treasury bills. The market risk exposure of the Bank's assets would be massively reduced, and moreover the Bank could sell the newly-acquired Treasury bills fairly quickly, within a period of months. The counterpart to the reduction in the Bank's assets would be a reduction in the commercial banks' swollen deposits with the Bank of England. The Bank of England would have a much smaller asset total and much healthier-looking capital ratios. The indemnity from the Treasury could be allowed to lapse (compare table 1 and table 2).

A large increase in the Treasury bill issue matched by a reduction in deposits with the Bank of England would of course mean a big change in the composition of the commercial banks' liquid asset portfolios. Yet it is not plausible that a switch in the form of liquid assets from deposits in the Bank of England to Treasury bills, in which there would be a liquid market, would affect the availability of bank credit, or any other aspects of the banks' behaviour towards their customers. The implementation of short-term interest rate policy might need to change. At present, if the Bank wants to change the level of short-term interest rates, it simply changes the interest rate it pays on deposits held with it by commercial banks. If its balance sheet were to be shrunk as suggested, it might need to revert to open-market operations in Treasury bills, gilt repos, and perhaps other liquid assets, on the pre-crisis pattern.

As table 2 shows, the proposed shrinkage of the Bank's balance sheet would get the leverage ratio up to 3.5 per cent, which is still barely respectable. Some recapitalisation would be needed so as to give the Bank sufficient capital for the assets it would still have after the balance sheet shrinkage, and some more might be desirable to allow the Bank some further scope to use its own balance sheet as a monetary policy tool.

The recapitalisation of the Bank and the withdrawal of the indemnity would internalise the problems of risk management within the Bank. The management of the Bank would not be expected to maximise profits, but they would be answerable for the financial condition of the Bank, and the Treasury would no longer need to interest itself so closely in the Bank's risk management techniques. The Bank should be allowed to retain profits and build up its capital gradually so as to maintain the utility of its balance sheet as a policy tool.

5. Unwinding QE

If the Bank were to exchange the gilts and corporate bonds in the APF for Treasury bills as suggested, then the eventual task of unwinding QE by selling off the gilts would fall to the DMO. Because the amounts are so large, it would not be a simple task.

Should it be done at all? Should not the gilts simply be cancelled? Cancelling the gilts would amount to replacing them permanently by the Treasury bills which the DMO would have transferred to the Bank, and which the Bank would have sold to the market. It would mean that the shortening of the maturity of the national debt which QE brought about would be made permanent. The public finances would be permanently more vulnerable to fluctuations in short-term interest rates – at present, a 1 per cent increase in short-term interest rates would lead automatically to an increase of 0.2 per cent in the ratio of the budget deficit to GDP. Tension between monetary policy and fiscal policy would be permanently heightened. The gilts should not be cancelled. There would however be a case for reconsidering the maturity structure of the debt, in the light of the Liquidity Coverage Ratio requirement which has caused the commercial banks to increase their gilt portfolios massively and, specifically, for issuing more short-medium gilts of the kind that the banks are likely to want.

Unwinding QE will put additional strain on the liquidity of the gilt-edged market, which is already under stress. The capacity of the market makers, which are predominantly banks, to provide a warehouse for gilts between the time of the auctions and the appearance of investor demand, has been curtailed by the post-crisis intensification of bank regulation, including the introduction of a mandatory minimum ratio of capital to total assets – the leverage ratio. Liquidity is already strained by the regular issuance programme and the DMO has had to introduce special inducements for market-makers, such as syndicated offerings and the post-auction facility to acquire additional gilts at the auction price. In addition, it has increased the number of separate auctions and reduced the average size in order to reduce the burden on the market makers.

It is unlikely that the current market structure would be able to absorb an auction programme which included the unwinding of QE over any period of a few years without the risk of wild price fluctuations, loss of market liquidity, and the accompanying risks to financial stability. Such a programme would require a change of technique, including perhaps a reversion to the pre-1986 practice of minimum price auctions, in which the government in effect acts as its own underwriter and sets its own underwriting price. Gilts that were left unsold at the tender were sold later in the secondary market. Minimum price auctions entail a degree of price-setting by the government.

Before 1986, such operations were conducted by the Bank of England; the terms of the issue and the minimum price were agreed with the Treasury, and the secondary market operations were conducted by the Bank subject to general guidance from the Treasury. If the proposed transfer of the gilts in the APF to the Treasury were to take place, the unwinding of QE would be managed by the DMO, not the Bank of England. The timing of the unwinding, and the way in which it was managed, would probably affect the economy and the economic outlook, and the Bank would need to be informed about what the DMO planned to do, and to have the right to express an opinion about it, but the Treasury would have to have the final say.^{Footnote 6} Thus the arrangements for debt management would be the same as they are for fiscal policy.

If, instead, the gilts in the APF were to be retained by the Bank and sold by the Bank at a time of the Bank's choosing, there would be two official bodies selling gilts into an illiquid market at the same time. Co-ordination would obviously be needed, and the co-ordination would have to be so close that there was in effect only one seller. The only realistic alternative to transferring the gilts in the APF to the Treasury would be to return the function of government debt management to the Bank of England, and abolish the DMO.