What Should Central Banks Do? (1994) - Goodhart - PDF

What Should Central Banks Do? What Should Be Their Macroeconomic Objectives and Operations? Author(s): Charles A. E. Goodhart Source: The Economic Journal , Nov., 1994, Vol. 104, No. 427 (Nov., 1994), pp. 1424-1436 Published by: Oxford University Press on behalf of the Royal Economic Society Stable...

What Should Central Banks Do? What Should Be Their Macroeconomic Objectives and Operations? Author(s): Charles A. E. Goodhart Source: The Economic Journal , Nov., 1994, Vol. 104, No. 427 (Nov., 1994), pp. 1424-1436 Published by: Oxford University Press on behalf of the Royal Economic Society Stable URL: https://www.jstor.org/stable/2235461 REFERENCES Linked references are available on JSTOR for this article: https://www.jstor.org/stable/2235461?seq=1&cid=pdfreference#references_tab_contents You may need to log in to JSTOR to access the linked references. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at https://about.jstor.org/terms Royal Economic Society and Oxford University Press are collaborating with JSTOR to digitize, preserve and extend access to The Economic Journal This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:21:26 +00:00 All use subject to https://about.jstor.org/terms The Economic Journal, 104 (November), 1424-1436. ( Royal Economic Society 1994. Published by Blackwell Publishers, Io8 Cowley Road, Oxford OX4 iJF, UK and 238 Main Street, Cambridge, MA 02142, USA. WHAT SHOULD CENTRAL BANKS DO? WHAT SHOULD BE THEIR MACROECONOMIC OBJECTIVES AND OPERATIONS?* Charles A. E. Goodhart I. WHAT CAN CENTRAL BANKS DO? Before turning to the normative question of what Central Banks (hereafter CBs) should do, we need to deal with the contentious issue of what CBs actually can do in the field of monetary control. One might think that this should be a relatively straightforward matter of fact. Instead, there is a yawning chasm of mutual misunderstanding, which has persisted for decades, between economists and those working in CBs, which has bedevilled the subject. Virtually every monetary economist believes that the CB can control the monetary base (hereafter Mo), and, subject to errors in predicting the monetary multiplier, the broader monetary aggregates as well. After all, Mo (apart from some relatively unimportant qualifications about coins from the Mint), represents the liabilities of the CB, and the CB should be able to control its own liabilities by open market operations. Hence the normal assumption is that Mo is controllable within a narrow margin. If the Central Bank should fail to do so, it must be because it has chosen some alternative operational guide for its open market operations, e.g. holding interest rates constant at some level, which operational rule is frequently decried as sub-optimal. Assuming that CBs can, almost perfectly, control Mo, economists have constructed several simulated schemes of how an optimal rule for so doing might be set up; McCallum (I993a, b) provides good recent examples. Almost all those who have worked in a CB believe that this view is totally mistaken; in particular it ignores the implications of several of the crucial institutional features of a modern commercial banking system, notably the need for unchallengeable convertibility, at par, between currency and deposits, and secondly that commercial bank reserves at the CB receive a zero, or belowmarket, rate of interest. The first means that fluctuations in the public's demand for cash, which are both strongly seasonal and somewhat unpredictable, must be accommodated. The second means that commercial banks will not willingly hold free reserves at the end of each day (assuming for this purpose that a stated reserve ratio has to be held on each day, rather than averaged over, say, a couple of weeks) beyond that needed to meet late fluctuations in the demand for cash after the money market has closed, or become thin. Only if interest rates fall to the very low levels of the I930s, and/or risks of interest rate variability or late-in-the-day cash runs increase, would commercial banks increase their desired free reserves. Thus, given the * I am grateful to Norbert Schnadt and Dirk Schoenmaker for helpful suggestions. [ 1424 ] This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:21:26 +00:00 All use subject to https://about.jstor.org/terms [NOVEMBER 1994] THE ECONOMIC JOURNAL 1425 unpredictable fluctuations in cash flows, including late-in-the-day payments or receipts by government, any attempt by the CB on any day to achieve either some particular level, or rate of change, in Mo is bound to cause the residual available to commercial banks for their free reserves either to be excessive, or deficient. With no interest being paid on reserves at the CB, any alternative over-night yield is better than none at all. Hence whenever a Mo target led to excess bank reserves, overnight rates would be driven down approximately to zero. Per contra, a deficiency of reserves, given the low demand for free reserves, would force commercial banks below their (legal) requirements. Such a deficiency, at the end of any day, would drive interest rates up either to the (penal) rate at which the CB is prepared to fill the shortage, or the rate which makes the commercial banks willing to face such penalties as are imposed on ex post reserve deficiencies, whichever is less. Allowing reserve requirements to be averaged over some period, as in Germany, does reduce interest rate fluctuations consequent on a random cash flow on any day. But as the reserve calculation period nears its close, the CB still finds that surplus reserves would drive interest rates to zero, while deficient reserves have to be restored at an interest rate chosen by the CB, unless the commercial banks are to transgress their legal requirements. So reserve averaging allows the authorities to reduce the frequency of open market operations, (consistent with a desired level of interest rate volatility), but does not provide the means for tight control over Mo (Schnadt, I994). If the CB tried to run a system of monetary base control, it would fail. Much, perhaps most, of the time it would still be accommodating the day-to-day demand of the banking system for reserves at a penal interest rate of its own choice, whenever its Mo target was below the system's demand for reserves. Otherwise when its target was above the system's demand, overnight rates would fall to near zero. Some economists might prefer such a staccato pattern of interest rates, but it would not seem sensible to practitioners. There is, of course, nothing to stop CBs using the path of Mo as an information variable, as a possibly important input in the decision of whether, and how much, to vary interest rates. This could even be done by some rule, e.g. of the form di =f(Mo,_1-Mo*,-,) where the term in brackets is the deviation of the actual from the desired path of Mo. The operation of the non-borrowed reserve base system in the USA between October I979 and Summer I982 worked in much this way. But using the path of Mo as a guide to interest rate adjustments is not the same as monetary base control (though if such adjustments are required by rule to be large it may approximate to such control), and also immediately raises the question of why one should tie interest rates to movements in Mo rather than a (larger) set of information variables. One factor preventing CBs from being able to undertake monetary base control is the zero, or below market, interest rate on (free) reserves, since the desiredfree, close-of-day, reserves are low. While the payment of interest rat ( Royal Economic Society 1994 This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:21:26 +00:00 All use subject to https://about.jstor.org/terms 1426 THE ECONOMIC JOURNAL [NOVEMBER on (required) reserves would reduce the distortionary tax on banking, the payment of a (near) market rate of interest on free reserves would not eliminate the problems of trying to run a Mo control system. While such a payment might encourage the banks voluntarily to hold a larger average volume of reserves, the demand for reserves would then fluctuate depending on interest differentials, between that paid on reserves and in the money market, which the authorities cannot closely control. So the reserve/deposit ratio would become much more volatile than now, and greater control over the monetary base would not translate to a similarly improved control over the broader monetary aggregates. Moreover, the choice of the level, and changes in, the interest rate on reserves would become another important (and discretionary?) interest rate choice for the CB.' Consequently those in charge of CBs generally regard monetary base control as a non-starter. The instrument which they can, and do, control is the shortterm money market rate. They are as fully aware, as their critics among monetary economists, that holding nominal interest rates constant indefinitely would be a recipe for instability. Instead, they ask what should be their optimal reaction function, i.e. in response to what economic developments, how much, and how fast, should they adjust short-term interest rates, their main instrument. Critics, especially among Monetarists, often state that no-one, including those in CBs, has the information necessary to devise a satisfactory (forecasting) method for (discretionary) setting of interest rates (Brunner and Meltzer, I993). Instead, they reiterate their preference for some monetary rule. CBs respond that such a rule is operationally infeasible; the most that can be done, as in Germany, is to use some monetary aggregate(s) as an information variable in the process of deciding on interest rate adjustments. The two groups confront each other, each believing that the other has misguidedly failed to grasp a few simple issues. A nice example of this is given by McCallum (Ig 3b) and the rejoinder by Okina (I993). Moreover, the relationship between interest rate changes, and the various monetary aggregates is uncertain, and subject to quite long and variable lags; as is also, of course, the relationship between the monetary aggregates and nominal incomes and inflation. This makes the use of monetary aggregates as intermediate targets more problematic. Nevertheless, amidst all this uncertainty there is a temptation to err on the s of financial laxity. Raising interest rates is (politically) unpopular, and lowering them is popular. Even without political subservience, there will usually be a case for deferring interest rate increases, until more information on current developments becomes available. Politicians do not generally see themselves as springing surprise inflation on the electorate. Instead, they 1 Moreover, the institutional arrangements which shape commercial banks' demand for reserves are in the process of rapid change. These include the form of the payments system; the opening and closing hours of the money market and the times and conditions of potential access to CB assistance; the fact that the shortest period for which money can now be lent or borrowed is overnight; and that the only time at which bank reserve positions are officially calculated and reviewed is at the close of business. All these institutional data are now likely to change, e.g. under the influence of modifications to the (international) payments systems. C) Royal Economic Society 1994 This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:21:26 +00:00 All use subject to https://about.jstor.org/terms 1994] THE ECONOMIC JOURNAL 1427 suggest that an electorally incon deferred, or a cut 'safely' accelerate end. This political manipulation of in aggregates (but indirectly, rather t and cynicism about whether the po believed. The disadvantages of tim understood, including among pol impetus for granting CBs more a rates as they, the independent CBs, objective, i.e. price stability, whic Fischer (I 994) describes as operational independence, but without goal independence. We turn to this next, and show how, in the process, it may help to finesse the disagreements between the Monetarists and CBs. II. WHAT MACRO VARIABLE SHOULD THE CENTRAL BANK SEEK TO CONTROL? It is easier for a principal to appoint an agent, if she can i achieve a single, quantifiable, easily recognised, measu outcome. This facilitates monitoring and accountability has multiple objectives for her agent, and when these conflict (e.g. there are trade-offs), then the principal delegate the choice among such objectives to the agent much closer control and supervision. Thus, with utilities t limitations on prices to consumers, an efficient provision services, and a rate of return sufficient to reward investor hence Ofwat, Ofgas. When an exploitable trade-off was perceived between inflation and unemployment, along the Phillips curve, the choicei of the optimal point was inherently, and rightly, a political decision. So, when monetary policy aims at several objectives simultaneously,2 with the need for choice, and balance, between them, policy will be subject to greater political oversight and the CB will be subservient. Greater autonomy is more likely when CBs are asked to achieve a single macro outcome, such as the maintenance of the Gold Standard up till I9I4, or price stability now. Thus current enthusiasm for independent CBs rests importantly on general acceptance of the vertical longer-term Phillips curve; that there is no medium or longer term trade-off to exploit; that the best sustainable outcome that th authorities can achieve through monetary policy is price stability. Those who do not accept this analysis, or who place great emphasis on th short-term, in which there remains a downward sloping Phillips curve, will fin such argumentation unappealing. For those who do accept the analysis, a shift 2 The European Central Bank (ECB) and the Bundesbank have as their overriding, primary objective t achievement of price stability. Both have a secondary objective, to assist the general policy of th Government, but this is conditional on this latter not being inconsistent with the attainment of price stabilit i.e. a lexicographic ordering. C) Royal Economic Society 1994 This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:21:26 +00:00 All use subject to https://about.jstor.org/terms 1428 THE ECONOMIC JOURNAL [NOVEMBER to CB 'independence', in circumstances where there is a single, quantifiable objective which the CB is mandated to achieve, makes the CB much more, rather than less, democratically accountable (Roll Report, I993). Both its objective, and success in achieving that target, can be made crystal clear. In those countries like Canada, New Zealand and the United Kingdom where the objective of price stability has been quantified, the target has been defined as a band for the rate of increase of the Retail (or Consumer) Price Index, e.g. 0-2 % in New Zealand, to be achieved between two and four years hence. There are many questions about the best way to quantify this general objective. One major issue is whether the objective should be price inflation alone, or for nominal incomes. There are several reasons for advocating a nominal income, rather than a price inflation target (Meade, I 994). The former gives some weight to deviations of output from its 'equilibrium'. Also, policy makers should not react to an adverse supply shock (e.g. the oil shocks of I973 and I979) by further deflating the economy. In practice, the latter point is largely met in the small print, or qualifications, to the inflation targets, whereby indirect tax increases, severe terms of trade shocks, energy and food price increases, and the own effects of interest rate increases on the RPI may be disregarded for the purpose of such targetry. Among the problems of using nominal GDP as a target are the delays in getting the data, and the errors and revisions in the series. Moreover, estimation of trend real output is difficult and contentious. Finally, the relatively long horizon for the RPI/CPI target provides some leeway for adjusting to cyclical fluctuations in the shorter term, while sticking to a single medium-term price inflation target. Consequently those countries adopting quantified final objectives have chosen a measure of inflation alone, not nominal incomes. The next question is whether that target should be set in terms of a desired price level, or a desired rate of change of prices. Ex ante, there is no real difference; a proposed future target level can be equivalently described in terms of a desired rate of change, and vice versa. The difference occurs ex post, when a miss in period t- i in achieving a level has to be rectified in period t, whereas for targets expressed in terms of rates of change, bygones are bygones, thereby causing the price level in the latter case to trace out a random walk. Consequently, the variance and uncertainty about future price levels is m less under a levels target. Moreover, when the regime has become credibl target expressed in levels should be more self-stabilising; unexpected deflatio (inflation) should lead to an expectation of future price inflation (deflation) restore the desired level; this will drive ex ante real interest rates in an equilibrating direction. The theoretical advantages of a target expressed in terms of levels have been thoroughly rehearsed by Scarth (I 994), by Fillion and Tetlow (I994), and by Duguay (I994). Nevertheless all quantified targets so far adopted have been in terms of the rate of increase of the RPI/CPI. Moreover, descriptions given by senior CB officials, in countries where no quantified target has been set, of their definition of price stability are often in terms of a rate of increase low enough not to enter into people's consciousness and calculations (Greenspan, I988). One reason ( Royal Economic Society 1994 This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:21:26 +00:00 All use subject to https://about.jstor.org/terms I994] THE ECONOMIC JOURNAL I429 may be that the desired optima a (very) low positive number, perhaps because of concern about an upwards bias, e.g. relating to quality changes, in the RPI, though this has been calculated as only about 05 % (Crawford, I 993, also see Fischer, I 994). The is also fear of downwards nominal wage rigidity and some residual feeling tha a little inflation may serve as a lubricant to the economic system, though the is little hard evidence of that, see Crawford and Dupasquier (I 994). Even s this would not of itself necessarily settle the issue of levels vs rates of chang because the ex ante target could still be set in terms of a slowly rising level. Perhaps, stating that the price level in three years' time, (starting from i oo) should be i o6- i, within a band of I05 to I07, sounds less impressive than that inflation should be held between i and 3 % throughout the period. Of course, a levels target is more demanding, since misses then have to be rectified. Given that this whole approach is in its initial phase; that quantified price objectives and 'independent' CBs are only now being introduced; that sustained low inflation, let alone zero inflation, has yet to be achieved; that political and public support for this regime have in many instances yet to become solid; for all these transitional reasons there is a need not to over-egg the pudding in the transitional phase. If this new regime is shown to have worked, and sustained (very) low or zero inflation is achieved, we may, perhaps, then switch to the theoretically preferable basis of a target defined in levels. The next question is which index should be chosen. We have already considered some reasons why the RPI or CPI is preferred to the GDP deflator. The latter is only provided quarterly, with considerable delay; often subject t considerable revision; the public do not find it easy to understand. The RPI (o CPI), however, only measures changes in present goods and services prices, an excludes current changes in the prices of future goods, i.e. omitting adjustmen in asset prices, houses, land, equity, etc. Some economists have argued that such asset price changes should, in theory, be included in the cost of living index (Alchian and Klein, I973, Goodhart, I993). But the theoretical basis for including asset prices-in a price index remains contentious, and asset prices are erratic, subject to fads and fashions. The question is posed, 'If there was a (speculative) boom in the equity market, would you want, just for that reason, to deflate the economy?' Per contra, asset prices, flexibly set in efficient markets, may often reflect the thrust of monetary policy before stickier current goods and service flow prices react. So, the present consensus is that asset prices are among the set of useful information variables, but should not be included in the index measuring inflation itself. Perhaps the most important asset price, at least for smaller, open economies, is the exchange rate. The exchange rate represents the most commonly used intermediate target. Many smaller countries have seen advantages in pegging their currency to that of a large neighbour, especially one with a good counterinflation reputation, notably within the ERM; but there are many other examples. This paper, however, is not the place to discuss exchange rate regimes. The point here is that both final and intermediate targets are ( Royal Economic Society I994 ECS 104 This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:21:26 +00:00 All use subject to https://about.jstor.org/terms 1430 THE ECONOMIC JOURNAL [NOVEMBER frequently used in tandem, but one of th target, the main information variable to he Thus countries within the ERM saw this as achieving price stability; but this latter wa bands were, and were the main influence contra, in New Zealand the RPI target is but in working practice the RBNZ has es bands for the exchange rate, which, if br response. Thus the exchange rate acts as a future progress to the final objective. Such an intermediate exchange rate tar mechanism, as in Argentina, Hong Kong peg, as in the ERM, leaves the ultimate deci price stability to the hegemonic, central cou lies between a final objective for the rat intermediate target for some monetary agg Opinion, I believe, is moving in favour predominant as the quantified target. Gi intermediate) can be clearly and precisely being equal, a natural advantage in going by definition, what people care about; it understood; restrictive policies may be mor to control inflation, rather than to control of times the links between the intermediat appear complex and uncertain to the lay perhaps break-down of, the stability of functions, uncertainty about their use has Tabellini (I994: I4-I5), 'it is clear that the and simpler to enforce.... Hence, a contract based on an intermediate monetary target is much more demanding on -the principal's information compared to an inflation target.... Generally, the principal finds it easier to monitor the outcome rather than the policy instrument, because the optimal instrument choice depends on detailed information which may not be available to the principal.' But, ceteris paribus does not hold. The key argument for an intermediate target is that this may be quicker and easier to achieve. Thus Persson and Tabellini ask (p. I 5), 'Why do we see exchange rate targets or monetary targets often imposed (or self imposed) on central bankers, but rarely see centra bankers accountable for the rate of inflation? One reason may have to do with the commitment technology available to the principal. The effect of policy actions on asset prices or the money supply is readily [sic] observable.3 The 3 Like many other monetary economists, Persson and Tabellini exaggerate the speed and ease of achieving monetary targets. As already explained, CB officials see no practicable and/or sensible alternative to the us of short term interest rates as their main instrument. The effect of such interest rate adjustments on th monetary aggregates is neither 'readily observable', nor necessarily quick acting, nor of well understood and predictable magnitude. ( Royal Economic Society 1994 This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:21:26 +00:00 All use subject to https://about.jstor.org/terms 1994] THE ECONOMIC JOURNAL 1431 effect on prices is observable only with substantial delay. It may thus be harder for society to commit to 'punishing' a central bank for actions undertaken six months or a year ago. If the central bank deviates from a financial target the penalty is more immediately related to the policy actions. It may therefore be easier to sustain such penalties than in the case of inflation targets.' In contrast to their optimism about the ease and speed of achieving intermediate monetary targets, monetarists emphasise the difficulty of forecasting inflation (on a policy unchanged basis), and hence of knowing how to set interest rates to achieve a given rate of inflation. Indeed Brunner and Meltzer (I993) reckon that this is so difficult that trying to set interest rates order to achieve price stability directly is more likely than not to be destabilising. No doubt such forecasting is difficult and subject to considerable error. The role of chief forecaster in a CB committed to an inflation target is both difficult and exposed. Yet most officials, practitioners and even economists who have worked inside CBs would, I believe, not put the difficulty or speed of achieving final RPI objectives as much worse than the task of attaining intermediate monetary targets. During the period of the Medium Term Financial Strategy the intended downwards path for inflation was almost exactly achieved (I980-5), but the monetary aggregate targets were not. So far the targets for inflation in Canada and New Zealand have either been met, or exceeded (i.e. inflation below the lower band). This has also been true for the United Kingdom since Autumn I 992, but it remains to be seen how much autonomy the Bank of England may have to vary interest rates if the desired inflation bands (1-4 %) look like being broken. Indeed, in some respects experience with final inflation targets, at least in New Zealand, have been too good in that success so far in holding inflation within a band-width of 2% may lead the public to believe that this standard can regularly be met, whereas unforeseen shocks may cause misses quite often when the bands are so narrow. The question of optimal band-width involves issues of credibility, as well as of the size of shocks and ability to forecast. With CB officials themselves often happier to go for final price objectives than for intermediate monetary targets, why not let them try? Monetarists may predict failure, but as long as the cost of failure is internalised in CBs by an appropriate incentive structure, there would not seem a disadvantage for society in this strategy. Perhaps the recent relative success with hitting inflation targets in Canada and New Zealand has been due to the greater incentives on CB officials so to do. This introduces the final question here, of the appropriate incentive structure for CBs in order to give society the best chance of price stability under this regime. At present, the main incentive is that the Governor may not be reappointed if the target is missed. The unwillingness of the incoming Liberal government in Canada to reappoint John Crow may have been due to their concern about the prior undershoot, and the extent of deflation in keeping near to, or below, the bottom band. When the target is expressed as an inflation rate in some terminal year, CB governors with a strong incentive to deliver will ? Royal Economic Society 1994 This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:21:26 +00:00 All use subject to https://about.jstor.org/terms 1432 THE ECONOMIC JOURNAL [NOVEMBER want to get inflation to, or below, the requi give themselves the best chance of hitting t Nevertheless, the idea that the incentive too potent would seem fanciful to many. A before the I 989 Act, I advocated relating bo the outcome, relative to the target. Although had been introduced, it was actually reject that it would evoke headlines such as 'Go action to throw 500,000 out of work'. Anyhow, CB officials are (curiously) reluc their own success in achieving their targets, even though this should considerably raise their expected average salaries. Perhaps they find payment by results demeaning; perhaps, as Persson and Tabellini (I994) suggest, they are very risk averse. One fallacious argument is that the outcome will be affected by unforeseeable shocks. So, of course, are business profits, but few would argue that businessmen's remuneration should not be partially profit related. It still strikes me as sensible to relate the remuneration of senior CB officials to their success. I ended Section I by suggesting that the adoption of price stability targets for CBs might reduce the intensity of dispute between Monetarists and CB officials over operational techniques. If CBs are mandated to achieve price stability, and the incentive structure properly rewards them for success (penalises failure), then CB officials have every incentive to find the best way to achieve that objective. If they continue to believe that this can only, or best, be done by interest rate adjustments, that is their responsibility, and it is they who would suffer if they got it wrong. III. WHY USE CENTRAL BANKS TO TRY -TO ACHIEVE PRICE STABILITY? CB officials have expressed their determination to defe Yet, for a variety of reasons, inflation accelerated in now more subdued, is hardly eliminated. The publi political and economic events that could reignite inflat institutional innovation of giving the CB a formal, price stability, and both the autonomy and incentives t improve on the past record, and may succeed in objective. Nevertheless it remains to be seen whether this institu If the CB cannot develop and maintain a supportive c the general public, and the main political parties, th sustainable, irrespective of the CB's technical experti Some monetary economists, mostly, but not only, of (for a definition see Laidler (I992), Schuler and W (i 984)), argue that this route, via Central Bank 'indepe ( Royal Economic Society 1994 This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:21:26 +00:00 All use subject to https://about.jstor.org/terms I994] way THE to ECONOMIC achieve JOURNAL price I433 stability, and with a fiat currency, which the operations, why not have the go numeraire, in terms of a basket commercial banks to maintain con bank notes, which they could free Achieve price stability by defin away with the CB altogether; an in the provision of both bank n revised still basket-commodity-based keeping the CB: but m advoca governments would be once again to revert to an increasing fiduciar Banking set of institutional chan abolition of the CB and 'free In their view the main problem with the Gold Standard was that the ba currency base was defined in terms of convertibility into a single commodity, so that world monetary conditions were subject to idiosyncratic shocks affecting it. Broaden the base to a representative basket of goods (and services), and the desired stable price level may be achieved, so it is claimed, almost by definition. There are, of course, different views about what the components of the basket should be, or whether the standard should be defined in terms of (a basket of) goods and services, or over labour (Thompson, i986), but these are second order.4 The first serious problem is how convertibility in such a system would be enforced. Someone doubting the solvency of one particular bank, or perhaps of banks overall, can hardly go into a bank and withdraw a basket of goods, let alone services, or labour time. One suggestion (Greenfield and Yeager, I989) is that convertibility be maintained indirectly, by having the banks hold reserves in gold and paying out as much gold as necessary to allow the creditor to exchange her notes/deposits so as always to buy the guaranteed amount of basket goods/services/labour. Obviously the amount of gold needed to pay the given sum to be withdrawn would vary as gold prices shifted relative to the basket. This idea is close t Irving Fisher's compensated dollar. A problem, however, is that it would seem to involve a bank, in the process of meeting its convertibility guarantee, often having to set a price of gold in terms of the basket different from that set in th market. This might lead to massive arbitrage flows, and hence to great instability in banks' gold reserves, and in interest rates (Schnadt and Whittaker, I993). Should Central Bank 'independence' fail to deliver price stability, this issue (currently, perhaps, regarded as arcane and academic), might suddenly become policy relevant. Assuming that a resolution is found, and indirect convertibility becomes 4 Hayek has advocated allowing banks to compete by offering convertibility into whichever objects of value they individually chose. Network economies of scale, however, make it efficient for a government to define the monetary standard in its own area. ( Royal Economic Society 1994 This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:21:26 +00:00 All use subject to https://about.jstor.org/terms 1434 THE ECONOMIC JOURNAL [NOVEMBER possible, what then are the comparative advantages and disadvantages of such a system; but compared with what? If Central Bank 'independence' should deliver (acceptable) price stability, the case for a massive jump to a new, and therefore uncertain, institutional framework would not attract much following. The case for switching to a completely different monetary system is conditional on how current institutional reforms work, and we do not yet know this. To continue the discussion, assume that CBs fail to deliver the desired outcome. What are the remaining pros and cons? There are some minor issues. The resource costs of each bank holding its own reserves of gold would be greater. There might be some greater informational problems, and resource costs, with each (major) bank issuing its own notes. Against this, competition in note issue might lead to an improved product, possibly cracking the problem of an effective means of paying interest on currency. Somewhat more important, governments would lose command over seigniorage. In countries without high reserve requirements and with low inflation, seignorage represents a small proportion of government revenue (Rovelli, I994), but every little helps. This source of government revenue is relatively painless, provides a tax on the black economy (large cash users), and helps to reduce other distorting taxes. Although the purpose of this exercise is to eliminate the inflation tax, that is the final source of revenue to a government in desperation when tax collection breaks down and the bond market dries up. Countries in the throes of war, revolution or collapsing civil order will, inevitably, turn to the printing press. Glasner (1 989) argues that this is why the creation of money has been so closely tied with sovereignty. The main issue is, however, whether a free banking system, without a CB, would suffer banking crises and panics, or would be inherently stable. This micro-level question is the chief bone of contention between those in favour of free banking, and those against. [This is a large question, which has been addressed at some length by George Selgin in his accompanying paper.] Bank runs rarely arise from a stochastic cumulation of ordinary, personal depositors seeking early withdrawal of deposits into cash (Diamond and Dybvig, I 983). Instead they arise, in the context of informational asymmetries when large commercial and interbank depositors fear, but do not exactly know that the bank involved has made serious losses on its asset portfolio (Jacklin and Bhattacharya, I988). If bank assets had an immediately calculable market value (not necessarily constant), this main source of runs should be halted, either by increasing restrictions on the bank's actions, ultimately closure, as its capital became impaired [FDICIA goes in this direction], or by moving towards mutual fund banking, in which the value of bank liabilities varies with that of bank assets. Such mutual fund banking is growing rapidly in the United States. With an improved payment system, mutual fund banking, and mark-tomarket valuation of bank assets, it is possible to envisage a future in which the CB's role in ensuring systemic stability, as Lender of Last Resort, supervisor, etc., could be phased out. But that time has not yet come, and there are good grounds for believing that banking crises and panics can still occur. Even ? Royal Economic Society I994 This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:21:26 +00:00 All use subject to https://about.jstor.org/terms I994] THE ECONOMIC JOURNAL I435 thereafter, parts of the wider financial system might still be subject to (liquidity) crises, and a need for a LOLR may well remain. The sizeable number of bank failures in several countries, e.g. Scandinavia and New England, and weakness elsewhere (e.g. Japan and Victoria in Australia), recently is evidence that systemic instability can still occur. Some free banking advocates claim that such problems are largely caused by the moral hazard associated with (improperly priced) CB insurance. If the CB was not there to protect them, banks would, they suggest, hold larger reserves and capital, so systemic instability would naturally decline. There is evidence of a downwards trend in bank capital ratios, prior to the Basle Capital Adequacy Requirements in I988, which trend may possibly be ascribed to the protection felt to be given to the system by CBs. Otherwise I have seen little persuasive evidence that the presence of CBs (as contrasted with complete, ioo% deposit insurance, e.g. as provided by FDIC and FSLIC, which did play a large role in the S & L debacle), was responsible for the banking crises of recent decades, e.g. the UK fringe bank crisis of I973 [those banks had not been under the wing of the Bank of England], the LDC crisis in I982, and the property-related crises in I990-2. Moreover there were plenty of banking crises and panics prior to the formation of CBs (Sprague, 19Io; the I893 crisis in Australia; Dowd, I992). Indeed, to lessen the intensity and incidence of such banking crises was one of the main reasons for the establishment of CBs. In some cases, as in the United States in I929-33, the CB's own policies may well have aggravated the disaster. But so long as panics and crises continue to be possible, and feared, the public, and politicians, will want to maintain some institutional mechanism for intervening in order to prevent such systemic instability, and that mechanism is, and will remain, the Central Bank. REFERENCES Alchian, A. A. and Klein, B. (I 973). 'On a correct measure of inflation.' Journal of Money, Cr vol. 5, pp. I83-I9I. Brunner, K. and Meltzer, A. H. (1993). Money and the Economy: Issues in Monetary Analysis Cambridge University Press. Crawford, A. (I 993). 'Measurement biases in the Canadian CPI.' Bank of Canada, Technical Report No 64. Crawford, A. and Dupasquier, C. (1994). 'Can inflation serve as a lubricant for market equilibrium?.' In Economic Behaviour and Policy Choice under Price Stability, pp. 49-80. Ottawa: Bank of Canada. Diamond, D. W. and Dybvig, P. H. (I983). 'Bank runs, deposit insurance and liquidity.' Journal of Political Economy, vol. 9I, pp. 40I-4I9. Dowd, K. (ed.) (I992). The Experience of Free Banking. London and New York: Routledge. Duguay, P. (I 994). 'Some thoughts on price stability versus zero inflation.' Paper presented at 'Paolo Baffi' Centre for Monetary and Financial Economics, Conference, 4 March, on Central Bank Independence and Accountability. Milan: Universith Commerciale Luigi Bocconi (forthcoming). Fillion, J.-F. and Tetlow, R. (I 994). 'Zero inflation or price stability'. In Economic Behaviour and Policy Choice under Price Stability, pp. I29-I66. Ottawa: Bank of Canada. Fischer, S. (I 994). 'Modern central banking.' Paper prepared for the tercentenary of the Bank of England Central Banking Symposium, 9 June. London: Bank of England (forthcoming). Glasner, D. (I989). Free Banking and Monetary Reform. Cambridge: Cambridge University Press. Goodhart, C. A. E. (I993). 'Price stability and financial fragility.' Paper presented at Bank of Japan Conference, Tokyo. 28/29 October, on Financial Stability in a Changing Environment. London: Macmillan (conference proceedings forthcoming). C Royal Economic Society I994 This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:21:26 +00:00 All use subject to https://about.jstor.org/terms I436 THE ECONOMIC JOURNAL [NOVEMBER I994] Greenfield, R. L. and Yeager, L. B. (i 989). 'Can monetary d 9, PP- 405-2 I. Jacklin, C. J. and Bhattacharya, S. (i988). 'Distinguishing panics and information-based bank runs: Welfare and policy implications.' Journal of Political Economy, vol. 96, pp. 568-592. Laidler, D. E. W. (I992). 'Free banking theory.' In The New Palgrave Dictionary of Money and Finance (ed. P. Newman, M. Milgate and J. Eatwell), vol. 2, pp. i96-i97. London: Macmillan. McCallum, B. T. (1993 a). 'Specification and analysis of a monetary policy rule for Japan.' Bank of Japan Monetary and Economic Studies, vol. I I, pp. I-45. McCallum, B. T. (I 993 b). 'Monetary policy rules and financial stability.' Paper presented at Bank ofJapan Conference, Tokyo, 28/29 October, on Financial Stability in a Changing Environment. London: Macmillan (conference proceedings forthcoming). Meade, J. E. (i994). Full Employment without Inflation. London: Social Market Foundation. Okina, K. (I 993). 'Comments on " Specification and analysis of monetary policy rule for Japan ": a central banker's view.' Bank of Japan Monetary and Economic Studies, vol. i i, pp. 47-54. Persson, T. and Tabellini, G. (I 994). ' Credibility and accountability in monetary policy.' Paper presented at 'Poalo Baffi' Centre for Monetary and Financial Economics, Conference, 4 March, on Central Bank Independence and Accountability. Milan: Universita Commerciale Luigi Bocconi (forthcoming). Roll Report, a report of an independent panel chaired by Eric Roll. (I 993). Independent and Accountable: A New Mandate for the Bank of England. London: The Centre for Economic Policy Research. Rovelli, R. (I 994). 'Central banking, seignorage and the financing of the government.' Paper presented at the 'Paolo Baffi' Centre for Monetary and Financial Economics, Conference, 4 March, on Central Bank Independence and Accountability. Milan: Universitk Commerciale Luigi Bocconi (forthcoming). Scarth, W. (I994). 'Zero inflation vs. price stability.' In Economic Behaviour and Policy Choice under Price Stability, pp. 89- I I9. Ottawa: Bank of Canada. Schnadt, N. (1994). The Domestic Money Markets of the UK, France, Germany and the US. Subject Report VII, Paper I, City Research Project. London: London Business School. Schnadt, N. and Whittaker, J. (I993). 'Inflation-proof currency? The feasibility of variable commodity standards.' Journal of Money Credit and Banking, vol. 25, pp. 2I4-22I. Schuler, K. and White, L. H. (I992). 'Free banking history.' In The New Palgrave Dictionary of Money and Finance (ed. P. Newman, M. Milgate and J. Eatwell), vol. 2, pp. 198-I99. London: Macmillan. Sprague, 0. M. W. (I977 [191O]). History of Crises under the National Banking System. Fairfield, N.J.: Augustus M. Kelley. Original edn., US National Monetary Commission (6Ist Congress, 2nd session, Senate doc. no. 538), Washington, DC: Government Printing Office. Thompson, E. A. (i986). 'A perfect monetary system.' Paper presented at the Liberty Fund/Manhattan Institute Conference on Competitive Monetary Regimes, New York. White, L. H. (1 984). Free Banking in Britain: Theory, Experience and Debate (I8oo-i845). Cambridge: Cambridge University Press. X Royal Economic Society I994 This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:21:26 +00:00 All use subject to https://about.jstor.org/terms