I

The Bank of England's ‘consultative document’ on Competition and Credit Control (C&CC) was published on 14 May 1971. It was a landmark occasion, representing a decisive break with the prior system of maintaining direct controls over bank lending to the private sector; the intention was now to achieve the monetary authorities’ objectives of policy via the operation of market mechanisms, notably adjustments in interest rates and open market operations. Although the ‘credit control’ aspect was, over the next few years, notably less successful than the encouragement of competition amongst the banks (where the London clearing banks previously had maintained a restrictive cartel with the support of the authorities), nevertheless the direction of travel towards a more liberal, market-based system, remained, despite a partial reversion towards a direct control system in the guise of the ‘corset’, introduced at the end of 1973, and finally laid to rest in June 1980.

This key event is now far enough removed in time to have become the subject matter of economic historians, notably Capie (Reference Capie2010) and Needham (Reference Needham2014), but with several other contributions, e.g. Moran (Reference Moran1984), Oliver (Reference Oliver, Needham and Hotson2014) and Sanderson (Reference Sanderson2009). But this note was neither stimulated by such histories,Footnote ¹ nor is it an attempt to provide either an alternative rounded, complete account of C&CC or a critique of such historical accounts. Instead, the purpose of this article is to recall a number of facets of this exercise which hold particular resonance for me personally as a participant in certain, but not all, parts of the formulation of C&CCFootnote ² and its aftermath, and which do not, perhaps, get quite enough emphasis in these current histories.Footnote ³

I shall start with the leading role played by Loan (overdraft) Facilities in generating problems for a system of direct credit controls, Section II. Then I shall discuss the pervasive intellectual error of placing too much weight on the efficacy of ratio controls, Section III. This will be followed by a description of what warnings were given within the Bank, and by the Bank to Her Majesty's Treasury (HMT), about the difficulties of monetary control, with comments about the possible utilisation of direct constraints on bank deposit rates, Section IV. Such warnings were not heeded. But as Needham (Reference Needham2014) emphasizes, a combination of the ‘dash for growth’ with Heath's refusal to countenance sufficiently upwards flexibility in interest rates would have scuppered C&CC in any case. In Section V I recall my own role in bringing this episode to an end with the introduction of the ‘corset’. Section VI concludes.

II

Capie (Reference Capie2010, p. 484) notes that there was dissatisfaction with ‘controls on lending, ceilings and quantitative controls in particular’. This is a considerable understatement. There were many reasons for such unhappiness, e.g. freezing the structure of the banking system, disintermediation into less efficient channels; many of these are repeated in the key paper on C&CC written by John Fforde at the end of 1970.Footnote ⁴ But a particularly acute concern was the incoherence of combining a system of providing (legally binding) loan facilities with a system of direct controls over, actually drawn, overdrafts and advances. The ability of a potential borrower to agree with a bank a facility which the borrower can then use at his or her discretion provides such borrowers with extremely valuable flexibility and added liquidity. It is one of the glories of an efficient modern banking system. But it means that, at any given moment, it is the borrowers, not the bank, which determine the growth in bank lending. Hence, so long as much, if not most, borrowing is done against previously arranged facilities, a bank cannot (legally) position itself so as to be certain of staying within a prescribed quantitative limit. The banks would have to make an informed guess about the future usage of undrawn, but available, facilities. Moreover, as macro-economic conditions worsened, and tougher credit conditions, possibly tighter controls, loomed larger, borrowers would bring forward their take-up of such facilities, to avoid having them subsequently withdrawn. There was, therefore, something of an adverse dynamic: the greater the pressure on the banks, the greater would be the incentive to crystallise and draw the available undrawn facilities. John Fforde on several occasions noted, ‘the occasional incompatibility of the overdraft system with the attainment of precise numerical targets set by authority’.Footnote ⁵

This problem had not gone unnoticed by the Treasury; thus an internal Treasury noteFootnote ⁶ raised the question, ‘whether a directive might be used to relieve the banks of legal obligations to fulfil a commitment, previously given, to provide overdraft facilities’. But the note continued by noting that the various legal authorities, e.g. the Treasury Solicitor, had taken ‘the view that the relief of banks from their contractual obligations is impractical’. Similarly modern monetary historians have been aware of this issue, but have not given it as much emphasis as, in my view, it deserves (Capie Reference Capie2010, p. 432; Needham Reference Needham2014, p. 30; Moran Reference Moran1984, p. 48).

Nevertheless, the potential claim on a bank, and the associated monetary expansion, is generated when the facility is granted (with usage at the borrower's discretion). So, if you want to control bank lending, you really need to go back one step, towards getting a grip on the initial stage of providing facilities. But the authorities, the Bank and HMT, had no idea of what the volume of such facilities might be. Not surprisingly, this appreciation led HMT to press the Bank to collect data on undrawn facilities from the clearing banks. The Bank approached the Chairmen of the London Clearing Banks to ask what data might be made available. In his note, John Fforde comments that, ‘In response to the Treasury's request, we again raised the matter orally with the C.E.O.s. Their response was again discouraging. Besides telling us that the statistics would be valueless as “lead indicators”, the CEOs were extremely suspicious that statistics of facilities would in practice lead almost at once to direct official control over the total of facilities.’Footnote ⁷ In this, they were, of course, right. Nevertheless, so long as monetary management was not going to be undertaken by direct credit controls, Fforde believed in the ‘intellectual case for information about facilities’; but the clearing banks, fearful of the (indefinite) continuation of direct credit controls, remained obdurate and obstructive.

On a personal note, I sympathised with the position taken by the banks. In a note to Fforde, I wrote (p. 10) that, ‘Unfortunately a commitment is such an intangible, nebulous concept that the possibilities for the evasion/avoidance of a control related to commitments would be excessive.’Footnote ⁸ And earlier in this same paper, I noted that, one alternative would be:

III

This note of mine was one of three papers written at this time to consider alternative methods, alternative to direct ceilings on bank lending, for control over credit and the monetary aggregates in the autumn of 1969. The first was by Tony Coleby,Footnote ⁹ the second mine, and the third by Bill White.Footnote ¹⁰ These were mentioned by Capie (Reference Capie2010, p. 484). We were all young then, and Bill and I were relatively new at the Bank, but the senior officials, especially Fforde, were prepared to give us an opportunity to develop our analysis and to treat it seriously. As Capie notes (Reference Capie2010, p. 484), Fforde wanted to keep focused on the longer-range objectives and mechanisms of monetary management.

It is, perhaps, worth trying to recall the state of analysis on the determination of the supply of money at that juncture. Although monetary control was then actually achieved by direct controls, ceilings, on bank expansion, academic analysis had centred on the money multiplier, relating the high-powered monetary base, via the banks’ reserve/deposit ratio (their cash ratio) (and the public's currency/deposit ratio), to the broader monetary aggregates. However, Richard Sayers (Reference Sayers1951) had noted, partly by direct observation of money market operations, that, so long as the Central Bank wants to maintain some target level of short-term interest rates, it has got to relieve any pressures on banks’ desired cash ratios, up (or down), by withdrawing excess (injecting additional) cash. Otherwise short-term money market rates would diverge sharply from the desired official level of rates. So if the level of short-term interest rates is policy determined, then both the high-powered monetary base (H) and the broader monetary aggregates (M) become endogenous variables. Sayers, who also served on the Radcliffe Committee, was the dominant British monetary academic at this time, and his book Modern Banking, in all its various editions, was the standard textbook.

Sayers, however, never followed through to the above logical conclusion. Instead, he shifted the ‘fulcrum’ of the multiplier from the cash ratio to the wider liquid asset ratio, a ratio which the banks were then meant to maintain. The idea was that the authorities could squeeze the banks of liquid (primarily public sector) assets, by a combination of fiscal tightness and gilt sales to the non-bank private sector. Of course, that depended on being able to manage gilt-edged sales to the (non-bank) private sector to that end, and that issue, the management of the gilt market, was a perennial sore subject throughout these years.

But, besides the question of gilt-edged market operations, there were two other major problems with control via a liquid asset ratio. The first was that not all liquid assets were from the public sector, notably commercial bills. The banking system, if it so wanted, could manufacture liquid assets by switching relative interest rates, in order to make it more attractive for the private sector to borrow on the basis of commercial bills rather than advances. The second, and more important, problem was that, if starved of liquidity, the banks could try to get hold of more by bidding for it, i.e. by raising the rate offered on (wholesale) deposits. But in the years up till 1969 they had not done so, because a combination of lending ceilings and the ‘cartel’ had meant that they had no cause to do so. But now the Bank was advocating a simultaneous abolition both of ceilings and of the cartel. Against that background ratio controls were unlikely to be effective. What would matter for control purposes would be the flexibility of the general level of interest rates, and, just as important, the relative pattern of interest rates. But these latter relativities had been fixed, by the cartel, for decades, so there was no experience of how shifts in relativities might influence monetary developments. There was some discussion of a possible need to retain a constraint on banks’ bidding for funds (a version of Reg Q as operated by the Fed in the USA), but this focused more on personal sector deposits, in order to protect building societies and National Savings, rather than as a means of monetary control.

In my own note (20 October 1969) I focused on the need to maintain a control on the maximum interest rate that banks could bid for funds, though the maximum could vary with the maturity, form and scale of the funding.

But, in order to prevent a competitive spiral of interest rates, especially if the political stomach is weak, the ability to constrain banks from bidding for funds would be essential.

In so far as there was a common theme to our three papers, it was that the joint abolition of ceiling controls and the cartel needed to have as a counterweight a continuing power to constrain banks from bidding for funds. Otherwise an interest rate spiral could ensue, which would place political sangfroid, let alone the real economy, under intense pressure. Meanwhile, a broadly based private-sector lending ratio could be a useful, but secondary and supporting, mechanism.

IV

The desire to get rid of direct quantitative credit ceilings, generally seen as distortionary and partially ineffective, was widely shared, not only by the banks and the Bank, but also in the Treasury and amongst politicians of all the main parties. The question was what alternative measures could be found to manage and control credit expansion and the monetary aggregates. Thus in October 1969, the Chancellor, Roy Jenkins, agreed with the Bank's Governor, L. K. O'Brien, that:

But as the three Bank papers, described in the previous section, had noted, it was not easy to find a half-way house between direct credit controls and pure reliance on flexible interest rates and open-market mechanisms. The latter was terra incognita, since although there had been a few, short periods when such controls had been relaxed, or even lifted, everyone knew, over the years from 1939 onwards, that, should there be a need to tighten monetary policy, the authorities would swiftly revert to the reimposition of such ceiling controls. The Treasury was scared of abandoning its ‘security blanket’, and doubtful whether any of the supporting measures, such as variable cash, or special deposits, interacting with one, or another, ratio would do much to provide a satisfactory alternative method of credit control.

Nor, at this earlier stage (i.e. pre-1971), did the Bank press for an immediate end to ceilings, and the cartel. So there was no immediate alternative strategy available to put before the Chancellor (Roy Jenkins). Instead of meeting the proposed two-week deadline for coming forward with an alternative policy, the Douglas Allen Working Party became transformed into a quasi-academic study group on methods of monetary control, going through (slightly revised versions) of the Bank's earlier internal papers; it met on numerous occasions, but got not much further forward. Nevertheless, a read of the papers/minutes, etc. of this group's deliberations in 1970 provides a good account of how the authorities, at this juncture, analysed the determinants, and potential methods of control, of the money supply.

One interesting, but hypothetical, question is whether a Labour government, if re-elected in June 1970, would have been as willing to introduce a reform along the lines of Competition and Credit Control, as were the Conservatives. As already noted, Jenkins was unhappy with the continued reliance on direct controls, and the Douglas Allen Working Party was set up under his direction. The change in government brought no noticeable alteration to the tone or direction of discussion, or to the membership of that working party. Moreover, so long as the outlook for the real economy appeared much the same – a Labour government might have been more relaxed about credit and monetary expansion than a Conservative government, some of whose members were by then beginning to follow monetarist analysis – C&CC was a relatively bipartisan measure. Indeed, the bulk of the criticism of the subsequent explosion of bank lending and broad money came from the Right, not the Left, of the political spectrum.

Nevertheless the arrival of the new government was one of the factors leading Fforde to press ahead with the Bank's more radical approach, that of getting rid of ceilings and the cartel, without having any alternative ‘fool-proof’ control mechanism. Thus in his key paper (Appendix A), Fforde notes that the Treasury probably:

Later in this paper (p. 5), Fforde accepts that interest-rate and gilt-edged management could (would) not be used flexibly enough to achieve sufficient control. But when he comes to such controls (Section III), he focuses mainly on the difficulties of applying common ratios, and other conditions, both to the clearing banks, and to the heterogeneous group of other ‘outside’ banks. Nevertheless he did advocate the maintenance of restrictions on banks' bidding for small deposits (to protect National Savings, etc.), and for short-term large deposits (see proposal (ii), p. 7).

Fforde's impassioned démarche met no serious opposition in the Bank. With Fforde and Page having to turn to more immediate matters, Andrew Crockett was then put to work to transform JSFF's (and J. B. Page's) ideas into an internal document to be sent to HMT advocating the abolition of both ceiling controls and the cartel. A first draft of this was prepared by Andrew Crockett in January 1971, entitled ‘A new approach to credit control and the banking system’.Footnote ¹¹ This continued to propose restrictions on banks’ bidding for small-value accounts (paragraph 15) and for short-dated large deposits (paragraph 16), but mostly concentrated on how an appropriate ratio could be chosen and made applicable to the banking system as a whole.

Turning to the control of bank credit, Andrew noted (paragraph 43) that:

But the power with which any of these could be applied was limited. As earlier noted, so long as the authorities wanted to fix some policy-determined level of interest rates, then ratio control, even when varied by calls for Special Deposits, was a weak reed. Interest rate flexibility was subject to political override, and the then structure of the gilt market, largely based on a couple of weakly capitalised jobbers, was sufficiently fragile that the authorities could never be sure of making positive net sales of gilts, let alone a pre-arranged net volume of sales, in any quarter.

So what exercises were done in the Bank to assess what might happen to bank lending, once C&CC was adopted? This largely fell to me, and there are a series of papers on this subject written in February 1971.Footnote ¹² Two main lines of analysis were followed, first a forecast using what UK data were available, and second a study by Bill White of what had happened in Canada previously under roughly similar conditions.Footnote ¹³

My assessment was that:

From this work I drew two inferences:

These drafts were preparatory work for a Bank paper to be sent to HMT, under cover of a letter from C. W. McMahon, on ‘A new approach to credit control: some quantitative implications’. In this latter paper, the suggestion about putting greater weight on M1 survived, but in considerably muted form (p. 4);Footnote ¹⁴ the warning about the short-term interest inelasticity of both bank lending and M3, and hence the difficulty of controlling by this route, was, with the eye of the initiated, implicit, but never made explicit; and no mention at all was made of a back-stop need for controls over interest rate bidding by banks to prevent a disorderly competition for funds.

Instead the gist of the paper lies in the (correct) statement that, in our Demand for Money exercises, Crockett and I found no statistically significant effects of the imposition of ceilings on M3 (Goodhart and Crockett Reference Goodhart and Crockett1970). If there was no effect on M3 of putting on ceilings, why should there be an effect in removing them? The effect, it was suggested, would be more on the asset composition of bank portfolios, than on the overall aggregate. This might have been the case had the cartel not been abolished simultaneously with ceiling controls.

There was nothing in the paper sent to HMT that was not, on its own terms, accurate and believed to be true. Even so, it was drafted so as to allay HMT fears about the potential weakness of the credit control mechanism. It also held back the internal concerns that a few of us felt in the Bank. Be that as it may, C&CC was then officially adopted as government policy, and a series of negotiations with the relevant private-sector institutions begun. The public paper, unveiling C&CC to the wider world, was published on 14 May 1971.

V

As is well known, the Conservative government embarked on ‘a dash for growth’ in 1972/3. Under these conditions output, the money stock and inflation would in any case have risen quite sharply. But in this particular case, bank lending and broad money (M3) rose far faster than would have been expected, on the basis of previous (econometric) experience. When this was subsequently followed, in 1973–5, by an upsurge in inflation, to a peak of over 25 per cent, much of the blame was placed, by more monetarist commentators and politicians, on the prior failure of the credit control element within C&CC. How much of this surge in inflation was due to monetary stimulus (including low ‘real’ interest rates), and how much to a combination of rising input prices (especially the first oil shock in 1973) interacting with the (unnecessary and disastrous) policy of wage indexation, is too large a subject to be addressed here.

Instead, the focus here is to ask what caused this surge in bank lending, and why it came as such an (unwelcome) surprise to the monetary authorities. As reported in the study on ‘Monetary developments in the first six banking months of 1973/74’,Footnote ¹⁵ M3 rose by 3700 in this half-year, an annualised rate of increase of almost 30 per cent.

Why had this happened? Prior to C&CC, both bank lending and deposit rates offered by the clearing banks were kept pegged by the cartel. Consequently the spread, between lending and deposit rates, had been constant over the whole data period. In the various studies of demand for bank lending and for the monetary aggregates, M1 and M3, interest rates had been included, often, e.g. in the bank lending equations mentioned earlier, with negligible econometric significance, but the spread, having been constant, was not.

One of the insights of monetarism is that there is likely to be a fairly stable relationship between agents’ access to funds that can be used for payment and their total expenditures. However, the question of what funds can be used for payment depends on financial structure. In an overdraft system, as in the UK, unused loan facilities are as much a source of funds available for making payments as are bank deposits. Most UK companies, certainly all big ones, will simultaneously have unused loan facilities alongside deposits. When an agent has a choice whether to pay a bill by drawing on its unused facilities or by running down its deposits, that choice is determined by the spread between lending and deposit rates. Widen the spread and there is an incentive to pay bills out of deposits (not by increasing loans), so both bank lending and monetary growth are (somewhat artificially) reduced. Narrow such spreads and the reverse occurs. When spreads become negative, as happened from time to time in 1972/3, there is even an incentive to borrow just to swell interest-bearing deposit accounts. This latter was termed the ‘merry-go-round’, also known as ‘round-tripping’, reported above (NB the merry-go-round can involve borrowing from a bank different from that in which the deposit is made).

What we had not foreseen, when introducing C&CC, was that, in the pursuit of expansion and a competitive increase in each of their market shares, the clearing banks would narrow their spreads, i.e. bid up for deposit funds, while keeping lending rates down to attract business. What was even worse was that a major control method incorporated in C&CC, i.e. calls for Special Deposits, intended to ‘starve’ banks of funds, actually made the situation worse by narrowing the spread further, thereby giving a further twist to the ‘merry-go-round’. In my view the changing pattern of relative rates of interest was the main culprit for the froth in the banking system and of the need to end this by driving ‘a wedge between borrowing and lending rates’.Footnote ¹⁶ Much of the increase then in both bank lending and M3 was artificial, and of little significance for the ‘real’ economy. That said, there was an unsustainable boom in 1972/3, and monetary factors aided, rather than counteracted it. By the end of 1973 real interest rates were negative, credit was plentiful and the government was still running a significant deficit.

In some large part this inappropriate stance of monetary policy was due to political unwillingness to raise interest rates, especially at a time when prices and incomes controls were being, once again, attempted. Businessmen complained to the Prime Minster that they could hardly be expected to keep prices down if their interest rates were driven up. Accordingly, on 15 November 1973, the Chancellor, Barber, wrote to the Governor, effectively to pass on Heath's demand that some other way had to be found to control monetary growth that did not involve further increases in interest rates (Capie Reference Capie2010, pp. 520–1). Richardson spoke to Fforde, who asked me to come up with an answer, which I did a few days later with my paper on ‘Alternative methods of credit control’.Footnote ¹⁷

As an economist, I was aware that a balance sheet had two sides, and that one could focus on average, or marginal, conditions. Hence there were four possibilities, to impose an average, or marginal, control on bank lending, or on bank bidding for funds. My strong conclusion was that the best buy was to penalise bank bidding for additional interest-bearing funds, partly because that was what had got the system into such difficulties, i.e. the punishment fitted the crime. This paper was the genesis of the ‘corset’ control (the term ‘corset’ was invented by Gilbert Wood, in the Bank, to indicate an external constraint to disguise and conceal internal flab), in operation most but not all of the time from then until summer 1980. No doubt in my absence others might have gone down this same road. Moreover, as will be obvious from reading this paper, it was written in a deeply cynical spirit. Nevertheless it was, I would contend, the best possible answer to a tricky, and unavoidable, problem. This story is not fully set out by Capie (Reference Capie2010, p. 521), so I am glad to have had the chance of adding to the record.