The Bank of England and the Sterling Money Markets: A Critique

Julian D. A. Wiseman

Abstract: The Bank of England does not implement its own monetary policy effectively. Its interventions in the money markets are at the wrong maturity, are in too small size, and the BoE accepts low-quality collateral. Yet monetary policy could be implemented simply and effectively.

Contents: Introduction, Tenure, Scale, Statistics, Collateral, What should be done?, Footnotes.

Publication history: only here. Usual disclaimer and copyright terms apply.

In each of the large currencies of the world, a short-term interest rate is determined by the central bank. The Federal Open Market Committee, the European Central Bank, the Bank of Japan, and the Bank of England each have a committee that meets a number of times each year to determine the official interest rate.

Once this committee has decided the level of rates, the central bank must implement it: the central bank must cause some market interest rate to trade near the chosen level. All central banks implement their monetary policy decisions by intervening in the money markets: either lending or borrowing.

The particular manner chosen by the Bank of England is ineffective, in that it does not cause any particular interest rate to equal the stated policy rate. Further, it need not be so, because there is a much simpler mechanism that would be more effective.

The Bank of England’s interventions in the money markets are flawed in three ways. The first two (tenure, scale) result in the bank’s failure to set any interest rate to the target; the third (collateral) results in the BoE taking huge credit risks.

1. Tenure

The Bank of England intervenes in two-week money. The BoE lends money for two weeks at the chosen level.

Imagine that tomorrow is an ‘MPC’ day: tomorrow the Monetary Policy Committee will announce the results of its meeting. This meeting is widely expected to result in a rate hike. Today, the BoE is lending two-week money at 6%. Tomorrow, the BoE is widely expected to be lending at 6.5%.

What should a commercial bank do? Of course, it should borrow all it can at 6% now, and deposit the unneeded part for 1 day. The result is two-fold. First, because everybody is depositing for one day, overnight money trades lower. Strange: because the BoE is expected to raise rates, overnight money is cheaper! This effect is known as ‘pivoting’. Second, two-week money will tend to trade somewhat high.

Of the big four central banks, two intervene in overnight money (the Fed and the BoJ), and two intervene in two-week money (ECB and BoE). So why is it acceptable for the ECB to intervene in two-week money but not for the BoE?

First, it must be emphasised that it would be better for the ECB to target overnight rates. The ECB ‘inherited’ the two-week target from the Bundesbank, which targeted two-week rates for the same reason that the BoE does: it always has. To the best knowledge of this author, neither has recently made a case for targeting two-week money, other than mere tradition.

Second, the ECB intervenes in money markets once each week, rather than every day. This reduces the scope for pivoting: the future higher rate will apply for at most 50% of the current intervention window (7 days out of 14) rather than 93% (13 days out of 14).

Thirdly, the ECB deals with many participants. In its money-market interventions it typically deals with 500 to 700 counterparties. The sterling markets are much more concentrated: the BoE typically deals with 20 participants during the day, and at the final intervention (after ‘the huddle’) with only one.

2. Scale

At the start of each day, the BoE chooses the amount of intervention that it is to provide, and the price at which it is to do so.

Wow! Does the BoE really believe that it can choose both price and quantity? Philip VI of France declared that something was worth what he said it was worth, on account of his being King*; does the Bank of England also claim divine right? Textbooks on economics deny that it is possible to choose both: either choose price and let the market choose quantity, or choose quantity and let the market choose price, perhaps by some form of auction.

A typical start-of-day money-market announcement from the Bank of England. Observe that the BoE attempts to set both price (“6.0%”) and quantity (“Stg 1600 mn”).

09:45 28FEB00           BANK OF ENGLAND                UK00181
BOE/MONEYOPS1

9.45 am Initial liquidity forecast Stg 1600 mn shortage
A round of fixed rate operations is invited. The Bank’s repo rate is 6.0%.
The operations will comprise repos to 13 & 14 March and outright offers of bills
maturing on or before 14 March
Principal factors in the forecast
Treasury bills & maturing outright purchases -106
Maturing bill/gilt repo -3271
Bank/Exchequer transactions +245
Fall in note circulation +1500
Bankers balances above target +25

Of course the staff at the BoE understand that even the BoE cannot determine both price and quantity. The truth is that the BoE’s preferred intervention mechanism (daily supply of limited-quantity two-week money) is ineffective at setting any interest rate. Instead, the staff in the BoE’s ‘boiler room’, where monetary policy is actually implemented, implicitly set their own monetary policy. By making slightly more or slightly less cash available, they choose the tightness or looseness of the sterling money-markets: they get to add or subtract a small amount from the committee’s official decision. The Monetary Policy Committee hiked rates in both September and November 1999, but neither of these rate hikes was ever visible in the market — in effect they were never implemented. Is this what Parliament intended when it passed the Bank of England Act 1998?

The Bank of England’s task is made harder by the concentration of power in the sterling markets. As already remarked, the ECB deals with several hundreds of counterparties. The Fed also deals with a large number of commercial banks. In contrast, the Bank of England must go toe-to-toe with the much smaller number of counterparties: only twenty or so commercial banks are active at the BoE’s monetary policy operations.

The problem is further exacerbated by ‘the huddle’. This semi-secret arrangement is unique to the sterling markets. At the end of the day the commercial banks ‘huddle’ together, and one of them (chosen by rota) takes on the position of all. Assume that bank A is at the centre of the huddle, bank B has £100 million too much, bank C £200 million too little, and bank D is about flat. B lends A £100 million and A lends C £200 million, at a rate calculated using a formula. Bank A then turns to the BoE to borrow £100 million, plus its needs. Thus, at the end of the day, the BoE must go toe-to-toe with a single counterparty. No other central bank faces such a maximally concentrated counterparty list.

The following table is presented as evidence of the BoE’s failure to implement its own monetary policy. It shows, for both SONIA (Sterling OverNight Inter-bank Average) and closing two-week repo rates, the average spread to the official policy rate (in 0.01s of a %), the number that were more than 0.125% under, were within 0.125% of, and were more than 0.125% over the BoE’s policy rate. Observe that two-week repo was ‘near’ policy only 56% of the time since independence, and the average cost of overnight money (SONIA) was within 1/8 % of official policy on only 31.7% of days (this latter statistic ignores days on which rates changed).

3. Collateral

The third flaw in the BoE’s monetary operations is the eligible collateral. Central banks do not wish to take credit risk. They properly believe that the taking of credit risk is for a commercial bank. Hence, when a central bank lends money, it takes collateral against it. The collateral acts as a credit enhancement: BoE lends bank A £100 million, and bank A lends the BoE slightly more than £100 million worth of government bonds. If bank A should become bankrupt, the BoE can sell the government debt to recover the value of its loan. This operation is known as a repo, and is covered by legal documentation deliberately designed to be proof against default.

So far, so good. But what collateral is eligible?

It used to be that the BoE only accepted gilts, and certain types of multiply-guaranteed trade bills. This list proved too small, so in 1998 the BoE widened out the list of eligible collateral, to include all EU government debt.

This should be acceptable in the event of a default of one of the BoE’s counterparties: the failure of a large UK clearing bank is unlikely to cause the failure of (say) the Italian Republic. But the reverse would be disastrous. If Italy defaulted on its debt (of most of a trillion pounds), it would most likely take down several commercial banks. (This is not unknown: when Charles II defaulted on 2nd January 1672, the six largest holders of the debt failed; when Russia defaulted in Autumn 1998, the world’s largest hedge fund failed.)

The problem is worse than it seems. The BoE’s counterparties can choose what collateral to give the BoE. They choose, unsurprisingly, the collateral that it is cheapest for them to obtain. This collateral is usually the weakest eligible credit: the BoE will typically be given ‘rubbish’, not ‘quality’.

Of course, many would argue that an Italian default is unlikely. But the leading rating agencies all give Italy a less than top rating. The UK and a small number of other countries receive the top rating of Aaa/AAA/AAA; the various rating agencies mark Italy either two or three notches beneath this. The BoE’s current strategy seems to be that of prayer: let us pray that Italy’s weaker rating proves unjustified. This author believes this ‘strategy’ to be imprudent. A private-sector bank adopting this ‘prayer’ approach would have some awkward meetings with its regulators: so should the Bank of England.

What should be done?

All of the above can be fixed easily.

The Bank of England should intervene at a maturity of 1 day: overnight money. That eliminates pivoting.

The Bank of England should have a borrowing and a lending rate, of 0.99 and 1.01 times the official policy rates. The BoE should accept unlimited deposits at the lower rate, and should make unlimited loans (against collateral) at the higher rate. Thus, if the official policy rate is 6%, overnight collateralised money would trade in the 5.94% to 6.06% range.

Eligible collateral should include only the ‘maximally-top rated’ paper. This means UK government debt, and other government debt issued by any OECD government that has been Aaa/AAA/AAA for at least five years, and the rating of which is not under review by any of the main three rating agencies. This is deliberately a high standard. Currently this criterion is met by debt issued by the UK, as well as by debt issued by the US, Germany, France, and few others. Against deposits, the BoE should issue 1-day Bank-of-England Bills.

The BoE should enforce a steep ‘haircut’ on loans. That is, it should lend only £100 against £105-worth of securities (the rules will need to be more complicated, so that debt with a more variable sterling price has a steeper haircut). The steep haircut ensures that better-credit entities (who can borrow the extra £5 most cheaply) will tend to deal with the Bank of England, and that worse-credit entities will tend to borrow from other private-sector banks.

To all this, the Bank of England, when consulted informally, has one major objection: “it would nationalise the money market”. Well, yes and no. The Bank of England would indeed be the dominant player in the market in overnight collateralised sterling. But what of it? The usual objection to nationalisation is that it distorts price signals. But the whole purpose of the BoE’s money-market interventions is for the state to set a price, not the free market. And it would not ‘nationalise’ any part of the money market other than that in overnight collateralised money; longer-maturity loans, and unsecured loans, would trade at a free-market rate.

Perhaps, when the UK Debt Management Office finally takes over the government’s cash management, the BoE will take the opportunity to improve its monetary-policy operations.

Footnotes

* This footnote is copied from footnote 3 of chapter 1 of Frozen Desire (The Meaning of Money), James Buchan.
Aucun puisse ne doit faire doute, que à Nous et à Nostre Majesté royal n’appartiengne seulement & pour le tout … de faire monnoier teles monnoyes, & donner tel cours, pour tel prix comme il Nous plaist … en usant de nostre droit. Letres adressés au Seneschal de Beaucaire … touchant le cours des Monoyes, 16 Janvier 1346 [i.e., 1347] in Ordonnances des Roys de France, ed. E.-J. de Laurière, Paris, 1729, Vol II, p. 254.

Julian D. A. Wiseman, February 2000