RBA Annual Conference – 1992 The Objectives for, and Conduct of, Monetary Policy in the 1990s Charles Goodhart
1. Macro Policy
(a) A Consensus on Macro Policy
In contrast to the previous two decades, there is presently, at the outset of the 1990s, a considerable degree of consensus about the appropriate objectives and functions for a central bank in the conduct of macro monetary policy. During the late 1960s and 1970s there was a ferment of discussion about theoretical aspects of monetary policy, with conflict between the various schools of Monetarists and Keynesians. The decisive battle in that conflict was won by the Monetarists with the general acceptance that there was no medium- or longer-term trade-off between inflation and output growth, and that expectations, if not perhaps as fully informed as some theoreticians might propose, would at least not be subject to systematic error. Hence the medium- and longer-term Phillips curves were vertical, and the appropriate medium-term financial strategy for any monetary authority was to reduce inflation steadily towards the achievement of price stability; though the optimal dynamic path to price stability was, and remains, a contentious subject, to which I shall revert.
Consequently, by the end of the 1970s virtually all central banks had become converted to pragmatic monetarism, and the discussions in the first half of the 1980s revolved primarily around operational issues. Which monetary aggregate had the most predictable relationship with nominal income? Which aggregate(s) should be chosen as the target(s)? How should the target aggregate be operationally controlled; through the adjustment of interest rates or via some form of monetary base control? It was an era when econometric studies of demand for money functions multiplied like rabbits. Rarely have so many equations been claimed to have stable and satisfactory properties one moment, and have collapsed the next.
The sceptics who doubted the efficacy of rule-based monetary targetry won that particular battle, though the German rear-guard still holds fairly firm, despite being forced to shift ground from central bank money to M3 as their icon. Indeed, monetary targetry became so discredited, that many of the initial zealots, such as Nigel Lawson and the British Treasury, somersaulted from the claim that control of a broad monetary aggregate (£M3) was an essential and almost infallible route to price stability to the opposite position of asserting that there was no useful information at all to be gleaned from movements in broad money aggregates.
Yet the core of the monetarist position remains untouched, or has even strengthened. There has been no serious challenge to the claim that the medium- and longer-term Phillips curve is vertical, and hence that monetary policy should focus, primarily, if not solely, on controlling the level of some intermediate nominal variable, so as to anchor the rate of inflation at zero, or some very small positive number. Moreover, the ever increasing international mobility of money, combined with the patent greater success of free market economies (as compared with more centrally planned, controlled economies), has led to deregulation, and the abandonment of structures of direct controls, (exchange controls, quantitative controls over lending, etc.). Under these circumstances, it has become generally accepted that the central bank's primary policy instrument, (we shall discuss later whether secondary instruments do exist), lies in its ability to vary money market interest rates; it can do so via open market operations to alter the amount of base money available to the banking system, relative to a reasonably stable demand for such reserves by the banking system. I shall discuss later whether there are any potential threats to the central bank's control over short-term interest rates. (Note that the time frame here is relatively short, in terms of weeks, quarters or a few years. In the long term, nominal interest rates will be determined by real (international) forces and the expected rate of inflation.)
Thus the current consensus is that the central bank should use its (short-run) control over money market interest rates to achieve price stability. There remains some considerable discussion whether the central bank should use its powers of interest rate control directly to restrain inflation, or should still aim at some intermediate (nominal) target, such as some monetary aggregate or the exchange rate, (as in the Exchange Rate Mechanism in Europe). I shall discuss this shortly.
Nevertheless, most of the discussion has currently shifted to more institutional and constitutional issues, such as how best to design the constitution of the monetary authorities, and their incentives, so as to induce them to focus on the medium-term achievement of price stability and worry less about shorter-term concerns over employment, output and growth. Discussion over the optimal structure for the European System of Central Banks (ESCB) was central to the deliberations of the Intergovernmental Conference on European Monetary Union (EMU) leading up to the Maastricht Treaty last year, whereas questions on how the ESCB should actually operate in markets were put on one side and deferred for consideration by the European Monetary Institute (EMI) over the next four years.
All this has led to a delightful, but not notably conclusive, academic literature in which the statutes of the various central banks are reviewed in order to draw up an index of their constitutional characteristics, notably their independence from political control, and then such an index is regressed against the average national rates of inflation. Long studies by Cukierman et. al. (1992) and Cukierman (1992) are the most comprehensive to date; also see Alesina and Grilli (1991); and Grilli, Masciandaro and Tabellini (1991). Since this debate is familiar, I can perhaps, be allowed to leave it at that.
Be that as it may, central banks around the world are being increasingly put on their mettle to deliver price stability. If this outcome is not achieved, without some overriding and publicly-accepted rationale for failure in this respect, within the next couple of decades, academic commentators may increasingly question whether there is some inherent fault in the basic concept of entrusting sole control over base money to a single monopoly provider, subject to various institutional constraints and incentives. The alternative, which a number of radical economists have already put before us, is to limit the intervention of the state to me definition of the monetary unit, e.g. in terms of (some basket of) commodities or assets, and then to leave commercial banks completely free to issue their own notes and deposits defined in terms of, and ultimately convertible into, such real assets. While this may sound implausibly futuristic, should central banks fail to achieve price stability, despite being given detailed instructions to do so, the attractions of a system that is designed to achieve (an approximation to) price stability automatically might come to seem considerable.
Whereas the upward trend in prices, and indeed in inflation until recently, (the question of whether one needs to difference prices once, or twice, to achieve stationarity has been finely balanced), has been worrying, the volatility of short-term fluctuations in most economic variables, both nominal and real, around their longer-term growth rates has been quite low in recent decades, (apart from 1972–76). There is, in my mind at least, a question whether a free banking system, without central bank support, would exhibit a higher degree of systemic instability. It was, after all, micro-level concerns about such instability, and contagious bank failures, that led to widespread introduction of central banks in the latter part of the 19th and early 20th centuries, e.g. the Federal Reserve Act of 1913 was a direct consequence of the 1907 banking crisis in the United States. This was, it will be remembered, at a time when macro-policy was largely determined by the constraints of the gold standard. So, while I do not entirely dismiss the possibility of a reversion from a fiat monetary system to one in which the monetary unit is defined in terms of (some basket of) real assets, I doubt whether this would, ipso facto, spell the death-knell of the central bank. I shall revert to the topic of these micro-functions of the central bank later.
Another nightmare which occasionally troubles central bankers is that their ability to control, (over the short and medium term), the level of nominal interest rates might increasingly erode. Would the advent of the cashless society cause the basis for, and stability of, commercial banks' demand for the monetary liabilities (base money) of the central bank to disappear? Alternatively, how long can the central bank prevent commercial banks, or other financial institutions, from issuing interest-bearing notes (e.g. by offering lottery prizes based on their serial numbers)? In fact, such fears are largely chimerical. The cashless society is unlikely ever to appear, so long as we prefer, or find that it leads to cheaper transactions costs, to make certain payments on an anonymous basis. Even if there was a cashless society, the monetary authorities could still ensure a well-defined demand for their liabilities so long as some payments, e.g. taxes, had to be made in such form.
The old concept, enshrined in many text books, that money is, or should be, non-interest bearing has increasingly been eroded. Only cash is now seen to be such, and, as noted above, currency notes could easily be given an expected yield. The reason that this is not done is fiscal; seigniorage is a nice, steady source of revenue for the government. There is no monetary reason why the central bank need prevent the issue of (competitive) interest-bearing notes. Since it is non-profit maximising, unlike the commercial banks, it can always drive up interest rates, notably by offering higher interest rates, e.g. on its own notes, or drive down interest rates, by open market purchases of assets, making it unprofitable for commercial banks to offer higher interest rates. Neil Wallace (1983) claimed that restrictions on competitive commercial bank issues of notes were necessary to enable the central bank to control nominal interest rates. But this is so only when the central bank chooses to fix the yield on its own currency notes at zero. When the central bank is, itself, willing to offer a competitive rate of interest on its notes, then the necessary conditions for its maintenance of control over nominal interest rates are relaxed. These are that the central bank is prepared to undertake non-profit maximising operations, and that commercial banks have an incentive (e.g. caused by a penalty rate), or a direct constraining requirement, to limit their net indebtedness to the central bank beyond some given limit, see Goodhart (1992).
But this latter is, indeed, an academic digression. Let us return to present policy issues, and the choice of intermediate target for the central bank.
(b) The Choice of Intermediate Target
(i) A Monetary Aggregate
The failure of equations intended to model the relationship between a money stock and nominal income, usually in the form of demand for money functions, to predict the future path of velocity has been notorious, especially when the aggregate involved became the chosen target for the country' s monetary authority. Although there are a variety of reasons for such breakdowns, including some exaggeration of the reliability of the initial econometric findings, (and it is impossible to allocate the share of the blame with any accuracy), there is some general agreement that financial innovation has been mainly responsible.
In particular, deregulation and greater competition, both from non-banks and overseas banks, has led to the provision of higher, market-related interest rates on a wider range of bank deposits, which has led to large shifts in velocity. The only monetary aggregate not subject (yet) to such competitive pressures on yields has been the monetary base, where nominal interest rates have remained zero. As noted above, it is technically possible to pay interest on base money, obviously so for commercial bank deposits at the central bank, and the reasons for not doing so are primarily fiscal. I would note, en passant, the emergence of a minor dispute in the European Community (EC) about whether the seigniorage receipts of a future ESCB should be allocated to the member central banks on an (arbitrary) formula basis, or go to the EC central budget, as a Committee of Experts on which I was a member has advocated. Be that as it may, not only is the monetary base potentially subject to various other (technical) innovations, automatic teller machines, plastic cards, etc., but in certain countries, notably Germany, there have been sharp fluctuations in the demand for hard currency (high value) notes in substitution for the depreciating currency of collapsing centrally-planned economies. Surveys in the United States can only identify the resting place of a fraction of the outstanding stock of currency.
Nor are there good grounds for believing that such innovations were once-for-all, and that, having happened, we need only wait for a while until a stable relationship is re-established. On the contrary, the commercial banking sector has been having a difficult time. Greater competition, and bad debts, in part the result of the use of monetary policy, i.e. high nominal and real interest rates, to control inflation, have reduced the profitability, capital backing and credit rating of commercial banks in many countries around the world. This had led to some renewed concern whether it is appropriate to entrust the operation of the payments system entirely to such fragile financial intermediaries. On the one hand, this has resulted in attempts to boost capital adequacy, the Basle Agreement, and to constrain bank activity to be dependent on the relative level of such capital, as in the draft bill, ‘The Financial Institutions Safety and Consumer Choice Act of 1991’, proposed (but not in the event enacted) by the US Administration. Such measures are bound to affect the interest rates that banks can pay on deposits, their relative competitiveness, and the rate of growth of both assets and liabilities.
On the other hand, there is an increasing awareness that payments facilities, even if perhaps on a limited basis, e.g. plastic cards, electronic transfers, high value paper transfers, can be provided by a much wider range of intermediaries than commercial banks. Such payments services are now being routinely provided by the building society/saving and loan/mortgage bank system, and there is no reason why this should not spread further into the mutual fund/unit trust/undertakings for collective investments in transferable securities field, or beyond. As payments services become provided much more widely, the question of the definition of money, or of a bank, is likely to become increasingly blurred.
Against that background, the statistical definition and calculation of monetary aggregates (other than Mo) on a 1/0 basis, (i.e. either all deposits in that category are included in the definition, or all are excluded), seems increasingly hard to defend. Quite why central bankers have been so adamant in their refusal to consider improved indices of moneyness, e.g. based on comparative interest rates (such as the Divisia index), or on comparative turnover, remains unclear to me. I do not suggest that such indices would, at a stroke, provide a stable relationship between money and nominal incomes, but, despite their complexity, I believe that they would add to our general information in a world where the relative competitiveness, interest rates paid and payments services offered by the various component elements of our shifting financial structure are continuously adjusting, and will go on doing so.
Against such a background of on-going structural change and financial innovation, there seems, therefore, little reason to expect (or hope) to find a monetary relationship (velocity) predictable enough to serve as a basis for a credible pre-commitment. Even in those cases where a modicum of econometric predictability may have remained, (Mo in the United Kingdom, M2 in the United States, M3 in Germany?), the authorities, well aware of prior problems with monetary targetry in other circumstances, have paid much more attention to a set of other indicators of inflationary pressure (e.g. exchange rates, wage increases, asset prices, retail prices), than to their continuing monetary target, (with Germany providing a partial exception). Indeed in the United Kingdom the continued role of Mo as a monetary target, (quintessentially so, now that we are in the ERM), has been an example of pure lip service to the concept of targetry. I cannot identify any monetary decision since 1985 in which the growth rate of Mo has played a significant part, and I can imagine the cynical laughter in the Chancellor's room were Mo to be mentioned during a discussion of interest rate policy.
What, however, does worry me is that the failure of the monetary aggregates to provide an accurate prediction of the movement of nominal income by themselves at all times has been taken by many authorities, notably in the United Kingdom, as a rationale for claiming that there is no useful information to be obtained from analysing the growth of the monetary aggregates at any time. In particular, the collapse in the growth rate of bank credit and in broad money in many countries in the last couple of years has provided a strong indication, to those who would look, that the depression would be longer, and the recovery more delayed and weaker, than those many official groups using standard Keynesian macro-models had expected. Until 10 April, many of us had expected that the UK Treasury's failure to factor such monetary conditions into their model would cost the Conservatives the election. Quite why that expectation, and the UK polls, were so wrong remains an interesting question for political scientists. Nevertheless, my message here remains. Despite the failure of monetary targetry, full analysis of monetary conditions should remain prominent, a constituent element in any central bank's set of information variables.
(ii) An Exchange Rate
The simultaneous failure, around the mid-1980s, of monetary targetry to provide predictable control over nominal incomes in many countries, and of floating exchange rates to provide stable and appropriate real exchange rates, enhanced the perceived advantages of pegging the exchange rate to whichever neighbouring major country could be expected to provide price stability. The mutual international links and surveillance, notably for example in the ERM, gave a degree of firmness and credibility to the exchange rate peg that no amount of domestic rhetoric about holding onto the monetary target could equal. Moreover, an exchange rate target is more tangible and understandable than a number (how obtained) for some (how and why chosen) abstract ‘M’.
There is, however, an important distinction between joining the ERM, which is but one (major) element within the wider context of the European Community, and a unilateral decision of a country to peg its exchange rate against another currency. The multilateralism of the ERM enhances its credibility well beyond that achievable by a unilateral, and hence more easily revoked, commitment. Australia, for example, cannot, if only for geographical reasons, join the European Community. If it was to commit to peg its exchange rate against the DM, or the ECU, it would not have the same effect as in the case of the United Kingdom joining the ERM.
Depending, therefore, in part on the degree of multilateralism involved, an exchange rate target can enhance the degree of commitment and credibility. Subject to differences in the rate of growth of productivity in tradeable goods and services, (a significant qualification emphasised in the Scandinavian theory of comparative inflation in fixed exchange rate systems, as Japan found out under Bretton Woods and Hong Kong is currently experiencing), the rate of growth of prices in the dependent country will approximate in the longer run to the rate of inflation in the anchor country.
The long-term trend outcome is fine so long as the anchor country does provide satisfactory price stability. But what happens if the anchor slips? Inflation in the United States in the 1960s was a major cause of the chaotic collapse of the Bretton Woods system. There are some worries that the disinclination of the German authorities to raise taxes to finance re-unification could threaten its record of price stability. If so, could the European Community switch, if necessary prior to Stage 3 of EMU, to an alternative anchor, e.g. the French franc, without massive disturbance? In the past the anchor country has generally been hegemonic, combining a good price stability record (if not necessarily the best in the system) with overall economic leadership, (the United Kingdom in the Gold Standard, the United States in Bretton Woods, Germany in the ERM). Does that have to be so, or can the anchor country be one of the smaller nations, e.g. could Germany ever anchor onto Switzerland, or does it always have to be vice-versa?
Even if the anchor country does continue to provide appropriate longer-term stability, pegging the exchange rate to another currency is likely to provide a bumpy ride when the (cyclical) shocks in the anchor country are not synchronised (are asymmetric) with those in the peripheral country (or countries). Let me take two current examples. German re-unification imparted an inflationary shock, exacerbated by their government's fiscal response, to Western Germany, requiring a strong hike in German interest rates. The resulting need to hold nominal, and real, interest rates high in the remainder of the European Community to hold the ERM peg intensified the recession there. Again, the low interest rates that have been appropriate in the depressed cyclical conditions pertaining in the United States have not been exactly right for the booming conditions in Hong Kong.
This asymmetry may result in a time path for real interest rates in the peripheral countries that is markedly inappropriate to their own short-term conjunctural needs, (otherwise known as the Walters critique of pegged exchange rates). Such potential asymmetries imply that, ceteris paribus, the road to price stability may be somewhat smoother when countries travel alone rather than by hitching their monetary policy to another's control and direction. But other things are not always equal. As in Europe, closer economic integration, in which exchange rate stability and subsequently currency union plays a major part, may be highly prized for its own sake. The credibility and security given by a currency link may, particularly for a small country facing a difficult future, vastly outweigh the inconvenience of temporary periods of inappropriate interest rates.
Finally, if monetary and exchange rate policy is to be predicated to the maintenance of an exchange rate peg, (and thereby to the same long-term rate of inflation as in the anchor country), can fiscal policy be used more aggressively to stabilise the peripheral countries against asymmetric shocks, e.g. those coming from the anchor country. To revert to my previous two examples, could the rest of the European Community have expanded their fiscal deficits to offset German-led higher real interest rates, and could Hong Kong have raised its budget surplus (by more than it actually did)? These examples show quite clearly, that whatever may be possible in theory, in practice much more aggressive use of fiscal policy, (and larger year-on-year fluctuations in the size and sign of the Public Sector Borrowing Requirement (PSBR)), would be fraught with serious political and economic problems.
In my view there needs to be a major politico-economic argument in favour of exchange rate pegging, e.g. to achieve wider economic integration or to maintain credibility under especially difficult circumstances, to justify the transfer of monetary policy control to another country, whose own needs and objectives will, in the shorter run at least, at times be out of line with your own. In the examples quoted, the EMS and Hong Kong, such justification clearly exists, but it will not hold as a generality.
(iii) The Direct Targeting of Inflation
Monetary targetry has failed. Exchange rate pegging is only to be advised in special circumstances. Yet central banks are being charged with particular responsibility for achieving price stability. How then should they achieve this? The most general answer is that they should use their main instrument, the control of short-term interest rates, directly to target inflation.
This involves a number of problems. Chief among these is the presence of (long and variable) lags, slow response/adjustment processes, within the economic system. In a world without lags, where nominal incomes and expenditures reacted instantaneously to movements in (real) interest rates, and where inflation reacted instantaneously to the pressure of demand, monetary policy would be vastly easier. Even in the face of stochastic shifts in economic relationships, the instantaneous nature of the relationships would enable the authorities to adjust their policy response quite closely to immediate needs.
The main problems are caused by the interaction of uncertainty about, and stochastic shifts in, the underlying economic relationships and the long and variable lags. Our best estimates suggest that domestic expenditures react quite slowly, building up over a number of quarters, to changes in interest rates. Meanwhile inflation, perhaps particularly asset price inflation, (notably housing and property prices), may also have a dynamic of its own.
Let me illustrate with a heuristic model of the inter-relationship between lending to finance real property, and property prices:
where LD, LS are the demand and supply of loans for real property, P, the price level of property, EṖ, its expected rate of appreciation, i is the nominal rate of interest, and the signs of the partial derivatives are as shown. Thus, the demand for loans is a negative function of the own real interest rate; the supply of loans is a positive function of the price level, (since this raises the available collateral), and the yield on such loans, and the price level responds to credit expansion. Finally, we assume a less than perfectly rational, backwards-looking expectations function.
Obviously if an equilibrium, Pt=Pt−1, is disturbed, instability results, since both P and EṖ enter positively into both LS and LD, and L adjusts to reinforce the disturbance. In this kind of model, if the authorities are slow to spot what is happening, i may have to be varied very sharply indeed, in order to offset increasingly strong expectations of future price changes. But once the authorities do manage to brake the expansion, then a given nominal level of interest becomes an increasingly severe real level of interest rates, as buoyant expectations first falter and then reverse.
Not only in the United Kingdom, but in quite a number of industrial countries, the last cycle of boom and subsequent recession has been much influenced by the path of real property prices, (especially housing in the United Kingdom), and the associated time path of bank lending connected with that. Housing and property prices rose steadily in the late 1980s fuelled by buoyant bank lending (with associated rapid growth in broad money). Asset price inflation preceded retail price inflation, but the growth in personal wealth in the United Kingdom surely had some considerable responsibility for the decline in the saving ratio, which drove up the pressure of demand and inflationary pressure in 1988–90. Checking this boom took a long time. Interest rates went to 14 per cent in May 1989 and 15 per cent in October 1989, but the economy did not decisively turn till Q3, 1990. Subsequently, of course, the reversal of the previous dynamic in the property and banking sectors, and the need to keep nominal interest rates high to support sterling within the ERM, induced a deep and lengthy recession, from which an anaemic recovery is only just beginning.
While much of this (unhappy) story is particular either to the United Kingdom, and/or to the special characteristics of the end of the 1980s, there are some more general issues. First, what exactly does one mean by inflation? This is usually measured by the Retail or Consumer Price Index (RPI, CPI), or, sometimes, by the estimated GDP price deflator. But should the authorities also concern themselves with asset price inflation/deflation, either for its own sake or because it may be a leading indicator of the CPI? Say that the authorities were in a position, much like that in Japan, with the CPI rising at 4 per cent per annum, but equity and property prices collapsing. What is the true rate of inflation in such circumstances, and how should monetary policy be adjusted then? In part, this question relates to the accompanying paper, by Blundell-Wignall et al., which asks whether asset-price and commodity-price inflation, (among others), may be leading indicators of RPI inflation. But even were there no such interconnecting links, should the authorities be concerned for asset-price inflation, (and deflation), as well as for the RPI?
Given that monetary policy's primary function is to achieve price stability, then interest rates are likely to (need to) be raised as, and when, inflation worsens. But the rise in inflation is almost bound to be accompanied by a worsening of inflationary expectations. Hence, in order to raise real interest rates, nominal interest rates must be raised significantly more than the prior increase in the annualised rate of growth of the RPI. In order to nip an inflationary spiral in the bud, the initial increase in interest rates needs to be rather dramatic, and once the down-turn comes the subsequent dismantling should also be rapid.
But the need for a sharp interest rate response to the early stages of inflationary/deflationary spirals goes right against a central banker's innate caution, reinforced by uncertainty not only about the future, but also about the present state of the economy. Moreover, whatever the constitutional position of the central bank, there will be a wider political need to justify interest rate changes to the general public. This need also tends to make the authorities hesitate too long.
It was quite largely to avoid such pressures for the interest rate response to be too little/too late that the authorities embraced monetary targetry. This gave an additional rationale for interest rate adjustments (we had to raise them as M was growing too fast), hopefully at an early stage (i.e. was M growth a leading indicator of inflation?): some even used it as a smokescreen, (i.e. ‘we try to control M, or non-borrowed reserves, and it is the market, not us, that varies interest rates’). The smokescreen was of great use in the early 1980s, notably to Volcker to enable him to raise interest rates sufficiently violently to check the worsening inflationary trend in the United States.
The record in the United Kingdom, (prior to its entry into the ERM, which changed the rules of the game), suggests that some continuing strengthening of the authorities' backbone against ‘too little, too late’, would still be desirable. But how can one factor that into a system in which the authorities act judgmentally, against a set of information variables, in order to achieve price stability? One route has been to consider changing the constitutional position, and incentive structure, of the central bank as a backbone brace. Might there be other routes? I have sometimes wondered whether, starting from a presumed equilibrium with zero inflation and 3 per cent nominal interest rates, there should not be a presumption that such interest rates should rise by 1½ per cent for each 1 per cent that inflation rises above zero, and that the Governor should be asked, say twice a year, to account for any divergence from that ‘rule’. But I fear that any such rule would be too mechanical; and hence it would be too easy for a Governor to justify, and commentators, politicians and the public to accept, lower interest rates.
Back in the early 1980s, when inflation threatened to stay in double figures, there was much discussion about the optimal speed of adjustment of price inflation, (gradual or as quick as possible). Now, with inflation commonly under 5 per cent in many industrialised countries, the more immediate question is exactly how do you define price stability? From its peak in 1979–80 many industrialised countries managed to reduce inflation to around 2–4 per cent by 1986–87, (aided by falls in oil/commodity prices). Then many finance ministers proclaimed victory over inflation, (prematurely, as evidenced in 1989–90), and turned to other matters. But that experience suggests that the public stomach for continuing (monetary) restraint may weaken below some low positive number. And is the calculation of inflation sufficiently reliable, (e.g. taking proper account of quality changes, etc.) to make it worthwhile to get the rate down to a somewhat arbitrary 0 per cent, especially if wringing out the final few percentage points involves a high cost in lost output? Exactly what do we mean, how should we define, ‘price stability’? It is an important, and quite difficult, question; and a question which the drafters of the Maastricht Treaty comprehensively burked.
Given the sources of dynamic instability, and the long and variable lags, within the economy, I have argued that the monetary authorities need to be brave in the face of uncertainty, and be prepared to vary interest rates earlier and more violently than their natural caution would normally entertain. If they are to follow this advice, it will have implications for interest rate risk. Acting early, rather than late, will make it harder for the private sector to predict such actions, and a larger movement will have a greater impact on those in an exposed position. Of course, there should be a quid pro quo, in that decisive early action should prevent the need for much larger, longer cycles; but that thought may be little solace to those on the receiving end.
These latter are not just randomly distributed among all economic agents, but the adverse effects are likely to be concentrated amongst certain sectors, among possibly fragile financial institutions and certain classes of borrowers. The sudden explosion of house repossessions from mortgage defaulters was a particularly painful aspect of the recent recession in the United Kingdom. Insofar as the continuation, and perhaps intensification, of interest rate risk will be a feature of a monetary system dedicated to the achievement of price stability, do the monetary authorities have a need, possibly a responsibility, to ensure structural adjustments to the financial system, (to its markets, instruments, conventions, e.g. on collateral, and institutions), that will enable it to absorb such interest risk satisfactorily? I shall try to address such micro-level structural concerns subsequently.
(c) Secondary Instruments and Objectives
In prior decades, the objectives of a central bank, when spelt out, were usually held to be several, including protection of the external and internal value of the currency, high employment, financial stability, etc., with the trade-offs between them left unspecified. As noted in (a) above, there is now a consensus that absolute primacy must be given to the achievement of domestic price stability. But this does not mean that central bankers are urged to concentrate entirely on only one variable, price stability. Rather their objective function is to become lexicographic; i.e. they can consider all these other subsidiary issues, e.g. external balance, growth, financial stability, etc., so long as it is consistent with the prior achievement of price stability.
This is made plain in Article 2 of the Statute of the ESCB (1992). ‘The primary objective of the ESCB shall be to maintain price stability. Without prejudice to the objective of price stability, it shall support the general economic policies in the Community with a view to contributing to the achievement of the objectives of the Community …’. So there are to be subsidiary objectives, given the primary objective. If so, are there secondary instruments of monetary policy usable for that purpose, given that the main instrument will have to be predicated to controlling domestic inflation (or, pro tem, within the ERM to holding the exchange rate peg)? What are these instruments?
I shall discuss two sets of such subsidiary instruments. The first of these is debt management; the second, direct (prudential) controls. Let us start with debt management. In order to control short-term money market interest rates, sufficient open market operations must be undertaken to operate on and determine the reserve base of the banking system. This, however, leaves considerable freedom to use debt operations to influence the maturity structure of the outstanding debt and to attempt to sell more, or less, debt to the various counterparty sectors, the banking sector, non-bank private sector and the external sector. Such techniques can be used either to try to twist the yield curve, ‘Operation Twist’, or to influence monetary growth, given the level of short-term interest rates. Exactly this latter was attempted in the ‘overfunding’ exercise in the United Kingdom, 1981–85. More gilts were sold to the non-bank public than necessary to fund the PSBR, so reducing bank deposits (and £M3); and the resulting cash stringency and upwards pressure on short-term interest rates were then relaxed by the Bank of England buying short-term (commercial) bills from the banks.
The general idea is that, when you want to ease monetary conditions (prices and quantities) relative to your, otherwise determined, short rates, you underfund, and vice versa. Indeed, with short rates in the United Kingdom now largely determined by German rates (within ERM), a number of economists, especially those focusing on broad money aggregates, such as T. Congdon, are imploring the Chancellor consciously to underfund, in order to relieve the severe monetary squeeze, see, for example, Anthony Harris' column in the Financial Times, on Monday, June 20th, 1992.
While I support this latter proposal to some considerable extent, there is obviously a major question whether such debt management is at all a powerful secondary weapon. ‘Operation Twist’ in the United States was not found to be notably successful, and studies of the effect of relative debt supplies (at the long, medium and short end) on the yield curve have not found any strong, significant effect. There is better evidence that overfunding in the United Kingdom did hold down the growth of £M3 by several percentage points per annum, but it did not lower the growth rate of bank lending to the private sector, or narrow money. Moreover, its (beneficial) effect on £M3 was achieved at the cost of causing certain tensions, and potential sources of arbitrage, in money markets. So the demise of this policy in 1985 was then regretted by only a few.
Not only are there doubts about the effectiveness of such debt management as a monetary instrument, there are also queries whether it should be operated by the monetary authorities at all, but rather should be controlled by the fiscal authorities. It is so in most countries, and to use debt management for monetary purposes would then require some co-ordinating accord. Those charged with formulating the Maastricht Treaty had such a terror of the ESCB being forced to provide monetary finance to (any sector of) government, that they insisted that the ESCB not participate in any primary public sector debt markets. Although the ESCB could still advise on debt management (and operate in secondary markets), the effect of Maastricht will be to reinforce the allocation of debt management responsibility to the Ministry of Finance.
Be that as it may, there will be continuing interest over the potential effects of debt management (given the level of short-term interest rates) on monetary conditions. From time to time some commentators and economists will advocate its use as a monetary instrument.
The use of my next set of instruments as secondary mechanisms for achieving macro monetary ends is even more contentious. These are direct controls over banks. In general, in this age of the triumph of free markets and of internationally mobile factor resources, such controls are eschewed. They are interventionist and anti-competitive. They lead to boundary problems, disintermediation, avoidance and evasion, distortion, moral hazard and to all the curses of the Pharaohs. Nevertheless, we have them in banking. They are usually imposed for prudential reasons, e.g. capital adequacy requirements, minimum collateral requirements, liquidity requirements, etc.. In one case, that of required reserve ratios, the advocates of such ratios also point to their fiscal results and their purported effect in facilitating monetary control. This latter is going to prove another contentious issue in EMU, where the Maastricht Treaty gives the Governing Council of the ESCB the right to set common reserve ratio requirements, Article 19 of the Statute of the ESCB, but the initial national levels (and use of) such ratios varies from none in the United Kingdom to high in certain Southern European countries.
The application of the Basle (and EC) capital adequacy requirements in recent years has revealed that from a conjunctural, cyclical standpoint, they are counterproductive. During recessions, bad debts will be high, profits low. Capital adequacy requirements will bite, prolonging the downturn, weakening the recovery. Per contra, they will be slack in booms, and have no constraining effect. The same feature will tend to hold more, or less, for most other prudential requirements, e.g. collateral margins, liquidity requirements, etc.
Suggestions have been made, e.g. by the Bush Administration, that the severity of the application of the capital adequacy requirement be relaxed in this recession. But it is possible to go much further, as the Labor Party was indicating, before the last general election in the United Kingdom, and make variations in direct controls, (e.g. in margin requirements for house mortgages), which have a primarily prudential purpose, into a secondary monetary instrument. It is quite possible that a future ESCB might vary reserve ratios for a similar purpose.
There are, of course, problems with this tactic. Would it work? While one may hope to obtain world-wide agreement to a common prudential level, e.g. of capital, getting such agreement to co-ordinated contra-cyclical variations in the level seems most improbable. Hence, there are bound to be considerable problems of disintermediation across boundaries, avoidance, non-level-playing fields, etc.. The Channel Islands must be looking forward to high EC reserve ratio requirements with great relish.
Although cyclical variations in such requirements can be defended on prudential grounds, e.g. it makes some sense to raise margins when housing (stock exchange) prices are cyclically high, there are nevertheless many problems. In an efficient market, how can the authorities infer when prices are (cyclically) (too) high? Could one deal with this by adjusting such controls in response, not to the level, but to the rate of change of prices, (but what is the average trend rate of increase?). Moreover, once attempts are made to adjust the level of such prudential requirements for cyclical reasons, might not there be pressures that could drive the average, over the cycle, away from its appropriate level. For example, an individual country might use such an excuse to gain a quasi-permanent regulatory arbitrage advantage. Having lowered, say, capital adequacy requirements on (valid) cyclical grounds, lobbying from its banks might prevent any return to the prior (agreed) level, whatever the circumstances. Alternatively, once having imposed tougher requirements, the authorities might be afraid to relax them, perhaps because of a fear that the resulting expansion could be unduly large, or merely because, in a world where inflation is the regular enemy, it is difficult ever to find a good time for an explicitly expansionary signal. For such reasons the monetary authorities in the United Kingdom persisted with direct controls over bank lending far too long in the 1950s and 1960s.
For all such reasons, the attempt to piggyback on prudential controls, to use them as the basis of a secondary conjunctural monetary instrument, is quite likely to pervert the initial prudential intention. Yet, as already noted, constant prudential controls are likely to exacerbate cyclical fluctuations. What conclusions does one draw from all this? One conclusion that I draw is that fashions change, partly in response to historical events. Booms and busts in asset prices (equity, housing) tend to make variations in collateral margins (equities, post-1929; housing, post-1990?) appear more sensible, whereas in other circumstances they are seen as, at best, otiose. Some decades ago, (variable) required cash/liquidity ratios were the dernier cri; now it is capital adequacy ratios. Even the most free-market, anti-intervention economists (e.g. the Shadow Open Market Committee in the United States) are advocating a direct constraint over the permissible business activities of banks in accordance with a (somewhat arbitrarily defined) ladder of capital availability.
Such problems, and debates, which will persist, arise originally because some prudential intervention is seen as required. If intervention is necessary, then its conjunctural effects will have to be assessed, and perhaps harnessed and utilised. But is such intervention necessary in the first place? This is the subject of the final section.
2. Micro Policy
Whereas there is now something of a consensus on the main lines of macro policy, there is less agreement on micro policy, by which I mean the central bank's concerns with the structure and stability of the banking system. Not only are there growing doubts whether there are clear dividing lines between those parts of the system (e.g. banks), for which the central bank should have the lead regulatory responsibility, and those for which it should not (e.g. capital market institutions), but also there are those who would transfer (all) such regulatory, structural micro functions to another body.
Indeed in the Maastricht Treaty sizeable barriers, e.g. in the shape of a unanimous agreement in the Council (Article 105 Paragraph 6 of the Treaty), must be overcome to allocate any regulatory responsibility, beyond that of advising, to the European Central Bank (ECB). Meanwhile, the individual national European central banks have quite widely differing formal constitutional responsibilities for supervision and regulation in their own countries. There is, however, more difference between the formal positions than in reality. Even when a separate body is constitutionally responsible for banking supervision, insurance, closure, authorisation, etc. as in Germany, it must, and does, work in practice closely in concert with the central bank.
The desire to separate the macro from the micro functions of central banks springs from several sources. Given that the ECB was to be independent of political control, there may have been some reluctance to give the ECB micro-structural powers as well, lest it seem too powerful relative to the democratic, political process. Probably more important, there may have been (in my view groundless) fears that concern with micro-level banking stability might divert the central bank's focus from its primary responsibility of achieving price stability, and that micro-level action, notably Lender of Last Resort lending (LOLR), could infect monetary policy in an inflationary fashion. (Exactly why such targeted LOLR support could not be offset in its overall market effect by generalised open market operations has never been clear to me.) Finally, there may have been a view that regulation and supervision could be done at the national, rather than the European, level and hence, under the principle of subsidiarity, should so be done.
This concentration, in the Maastricht Treaty, on the macro-functions of central banks, and the implicit willingness not only to downgrade their micro-functions, but perhaps to hive off such functions to some extent to some other, constitutionally separate, body runs counter to the historical record. The establishment of the central banks before 1914 was done in a context in which the price level, and the main elements of monetary macro policy, were set by the operation of the quasi-automatic gold standard. The function of the central bank was to prevent micro-level banking disturbances, banking crises and panics, from disturbing the macro-economy. Moreover in the then developing countries, e.g. Germany, Russia, Japan, the central bank was perceived as having an extremely important role in developing the structure of the financial system to underpin industrial and economic growth. The idea that the central banks should concentrate on the aggregates, and ignore the development of the constituent parts, and the structure of the financial system, is alien to the historical record.
Moreover, it is doubtful whether such a division of responsibilities, (macro to the central bank, micro to some other bodies), is feasible or viable. Banking crises have usually involved sizeable deposit withdrawals, or runs, from banks whose solvency was subject to doubt. In those circumstances other banks would not lend to such risky brethren. The decision, whether or not to provide last resort loans in such cases, is traditionally that of the central bank. Who could replace it? The argument is sometimes made that no such lending is desirable, particularly since such runs are often (well justified) symptoms of insolvency. So long as the central bank looks after the aggregates, let the chips fall where they may for the individual institutions, or so it may be argued. That proposal strikes me as extremely dangerous, given the contagious nature of confidence and panic, and the effect on asset prices of such changes in mood. Moreover, once panic should take hold, it could radically alter the structure of financial institutions, with a flight to those institutions perceived as relatively safer.
The central bank will have concern for systemic stability. One key feature of the financial system is the continuing smooth working of the clearing, payments and settlements systems. I find it difficult to envisage how such systems can remain secure under all circumstances without the involvement, support and guarantee of the central bank. If the central bank is to provide its guarantee, e.g. by providing day-light overdrafts to banks with net debt positions, it is going to want to ensure that their financial conditions remain satisfactory. Hence, concern with the central payments mechanism would seem to imply that the central bank retain a direct supervisory, and regulatory, oversight over the main monetary institutions in the country.
Once, however, we establish that a central bank's concern with systemic stability will make it want to maintain its regulatory oversight (and LOLR functions) with respect to some banks, is there any clear, or convenient, outer limit, or boundary, to the central bank's field of authority. Most financial boundaries have become fuzzy. Many of the large banks are now international, and the BCCI affair has underlined both the importance, and difficulties, of international regulatory (and accounting) co-ordination. The acceptance of the universal banking principle in Europe, the erosion (and possible future repeal) of Glass-Steagall in the United States and Article 65 in Japan, is blurring the distinction between banks and non-banks. As earlier noted, not only are banks entering capital market activities, but non-banks may increasingly compete by offering certain (perhaps limited) payments services.
It is no longer as clear, as it used to be, which are the set of (banking) institutions which should come under the central bank's protective wing, and which should be the responsibility of other regulators. That raises the question of which parts of the system really need such support, and why, and whether those parts within the safety net can, or should, be effectively separated from other associated activities. There have been suggestions to restrict deposit insurance and protection to those banks whose permissible asset holdings would be narrowly defined, e.g. Litan's narrow banks (Litan 1987). But deposits with such ‘narrow’ banks would presumably yield less than deposits with other financial intermediaries. Would it be acceptable to prevent these latter from offering payments' services on their uninsured deposits? Moreover, such ‘narrow’ banks would presumably only be able to lend money to the private sector, if at all, through purchases of high-grade marketable commercial paper. Ordinary loans and overdrafts to the rest of the private sector, (apart from fully collaterised mortgages), would have to be made on the basis of uninsured deposits. Could, should, a central bank stand quietly by while any significant proportion of such credit creating capacity collapsed?
Thus, the appropriate boundary for central banks' micro-level responsibilities has become unclear and fuzzy. By the same token, the extent of the central bank's responsibility to support the system is, and should remain, ambiguous. One hundred per cent insurance, whether explicit as in the United States, or implicit, causes severe moral hazard problems. One of the major problems facing central bank regulators is that the ‘too big to fail’ syndrome is commonly believed to hold in all major countries. Not only does this involve some moral hazard for the large banks, who are perceived as invulnerable, but it also provides a marketing advantage to them relative to their smaller competitors.
It is, perhaps, the ‘too big to fail’ syndrome that gives the main edge to the proposals that banks should be required to value their capital on a ‘mark-to-market’ basis, so that increasing controls can be placed on their operations, including, if need be, enforced sale, or merger, while their residual capital value is still positive. But can this really be done?
While I have clearly only been able to sketch in a few of the micro issues facing central banks, I hope to have demonstrated that, unlike macro policy where something of a consensus now holds, there is great uncertainty about the appropriate ambit and functions of a central bank at the structural, regulatory level. Such problems and issues will continue to confront them through the 1990s, whatever degree of success they may achieve in taming inflation.
3. Conclusions
There is a general acceptance among economists that the medium, and longer, term Phillips curve is vertical. Hence, there is no trade-off in the longer run between growth and inflation. Consequently, there is now also a consensus that the primary macro-policy objective of a central bank should be price stability. Given the consensus on the objective, much current discussion has shifted to structural questions on how best to design the constitution of, and incentives for, the monetary authorities so as to induce them to focus on the achievement of price stability.
In part owing to the collapse of the stability of demand-for-money functions, there is also now a general acceptance that the monetary authorities' main instrument lies in their control of short-term interest rates, rather than monetary base control for example. Yet given the demise of monetary targetry, on what basis should the authorities set such interest rates? The option of pegging the exchange rate, and hence interest rates, to that of another (presumably more counter-inflationary) country has the disadvantage that, when the cycles in the centre and periphery are unsynchronized, the time path of the real interest rate in the peripheral country will become inappropriate. Except where there are supplementary advantages from exchange rate pegging, the best option would seem to be for the central bank to target inflation, or the growth of nominal income, directly. But the long and variable lags of the operation of monetary policy, and uncertainty about, and changes in, the structure of the economic system make that difficult. The authorities have to use a set of information variables, including the growth rate of the monetary aggregates themselves, as their (partial) guide.
An upsurge in inflation is quite likely to cause a worsening in inflationary expectations. This is, perhaps, especially so in the case of asset prices; the links between asset price inflation and general inflation deserve careful consideration. Consequently, following such an upsurge, nominal interest rates need to be raised sharply, if there is to be an increase in real interest rates. But, once inflationary expectations are dampened, that same level of nominal interest rate could imply a sharply rising level of real rates. Central bankers need to brave their innate caution and be prepared to vary nominal interest rates sharply, both up and down, over the cycle. But such interest rate volatility impinges primarily upon a relatively small proportion of the economy, which may require structural reform in order to absorb such volatility.
With interest rates, the central bank's primary instrument, being predicated to the control of inflation (or of exchange rates), its primary objective, are there any secondary instruments that could be employed in the pursuit of secondary objectives, e.g. encouraging investment? Two such potential instruments are considered; the first is debt management, mainly in the form of over-, or under-funding, whose effect on the economy is assessed as rather weak; and the second is the application of prudential controls to macro-policy, cyclical objectives, which may pervert their primary prudential purpose and whose value in this role is contentious.
Indeed, unlike monetary macro-policy, where there is now a general consensus, the whole issue of the role of the central bank in prudential micro-policy supervision has become a matter of debate. There is no agreement, or common view, on the extent and nature of a central bank's appropriate responsibilities in this field, and which financial institutions, if any, the central bank should supervise. The costs of compliance, the moral hazard arising from (implicit) insurance, particularly the ‘too big to fail’ syndrome, regulatory failures (BCCI), are all deplored, but there is no agreement on a better way forward.
References
Alesina, A. and V. Grilli (1991), ‘The European Central Bank: Reshaping Monetary Politics in Europe’, CEPR Discussion Paper No. 563.
Blundell-Wignall, A., P. Lowe and A. Tarditi (1992), ‘Inflation, Indicators and Monetary Policy’, this volume.
Council of the European Communities (1992), Treaty on European Union, (commonly known as the Maastricht Treaty), Office for Official Publications of the European Communities, Luxembourg, signed on February 7th.
Cukierman, A. (1992), Central Bank Strategy, Credibility and Independence: Theory and Evidence, M.I.T. Press, Cambridge, MA, forthcoming.
Cukierman, A., S.B. Webb and B. Neyapti (1992), ‘Measuring the Independence of Central Banks and its Effects on Policy Outcomes’, The World Bank Economic Review, 6, pp. 353–398.
Goodhart, C.A.E. (1992), ‘Can We Improve the Structure of Financial Systems?’, Schumpeter Lecture at the European Economic Association, 7th Annual Congress, 30th August.
Grilli, V., D. Masciandaro and G. Tabellini (1991), ‘Political and Monetary Institutions and Public Financial Policies in the Industrial Countries’, Economic Policy, 13, pp. 341–392.
Harris, A. (1992), ‘A New Twist to Ease the Slump’, Financial Times, Monday, July 20th, p. 17.
Litan, R. (1987), What Should Banks Do?, Brookings Institution, Washington, DC.
Wallace, N. (1983), ‘A Legal Restriction Theory of the Demand for ‘Money’ and the Role of Monetary Policy’, Federal Reserve Bank of Minneapolis, Quarterly Review, 3, pp. 1–7.