6 The Framework of Monetary Policy (2024)

Objectives and Targets

Figure 6.1 is a flow diagram that illustrates the process of monetary policy implementation (see Friedman, 1975, and Madigan, 1994). Monetary policy objectives traditionally include economic growth, employment, price stability, and, by extension, nominal GDP.

Depending on the country, monetary policy may assign these objectives equal weights or, as many countries have done in recent years, place greater emphasis on the objective of price stability. Monetary policy may also seek other objectives—including the stability of long-term interest rates and financial markets, including foreign exchange markets—and may target economic activity in particular sectors of the economy, most notably agriculture. Although some objectives are consistent with each other, others are not; for example, the objective of price stability often conflicts with the objectives of interest rate stability and high short-run employment.

Monetary policy targets, as distinct from objectives, are proximate goals—goals that are not objectives in and of themselves but which, if attained, will work directly toward achieving the longer-term objectives of policy. Monetary policy targets are classified as either operating targets or intermediate targets. Intermediate targets are variables that, although thought to affect the ultimate objectives of monetary policy, are not controlled directly by the central bank. They include, for example, the money stock, outstanding credit, and long-term interest rates; these are variables that lie between the objectives and the instruments of monetary policy. In contrast, operating targets are tactical goals that the central bank tries to attain in the short run. Although central banks cannot use monetary policy instruments directly to affect intermediate targets, they can use them to affect operating targets, such as reserve money and short-term interest rates, which influence movements in intermediate variables.¹ Operating targets can also be altered in response to short-run economic conditions.

It is not always appropriate to use monetary policy instruments to achieve an intermediate target. In particular, if the intermediate variable does not show a stable and predictable relationship with the ultimate objective of policy, setting targets for intermediate variables will not promote overall monetary policy. Even in this situation, however, such intermediate variables as monetary or credit aggregates can continue to play an important role in the policy process—not as variables to be controlled, but rather as “indicator” (or information) variables.

These indicator variables may provide data on aggregate demand and inflationary pressures that serve as an information source in formulating monetary policy. Indicator variables provide signals about the stance of monetary policy, particularly changes in the degree of policy tightness. An example of an indicator variable is the term structure of interest rates. When monetary policy is tightened, short-term interest rates tend to increase; if, however, monetary policy lowers inflationary expectations, the inflation premium in some forward rates may be reduced. The impact of the tighter monetary policy, therefore, may be indicated by the effect on long-term rates, which is an average of the short-term rate effect and the effect on forward rates over the term of a particular financial instrument.

Three main criteria guide the selection of intermediate targets (and may also be relevant to choosing an operating target): an intermediate target must be measurable, it must be subject to some degree of central bank control (with an operating variable as the vehicle), and it must have a predictable effect on the ultimate policy objectives. Accurate and timely measurement of an intermediate variable is essential for judging the effectiveness of monetary policy actions. Central bank control over the intermediate variable is essential for moving the variable back on track when it diverges from its targeted path. But the most important criterion for an intermediate variable is the predictability of its impact: even if the variable can be controlled and measured accurately by the central bank, it has no application if its impact on the policy objective cannot be predicted. Suppose that the policy objective variable is GDP, and suppose that the central bank is considering targeting a particular intermediate variable. Then, even if a change in the target variable had only a small effect on GDP, it would be a suitable target if the central bank could predict the effect accurately.

One of the dilemmas in selecting intermediate targets may be illustrated by the comparative predictability of the impact of the money stock and interest rates on GDP. Suppose that the interest rate used as a target rises above its targeted level. The central bank could assume that the interest rate-GDP relationship is stable and take action to increase the money stock to reduce interest rates; alternatively, it could assume that the money-GDP relationship is stable and treat money as the target while also allowing interest rates to rise above the previously targeted level. The appropriate policy response would depend on why the interest rate increased. Given a fixed level or growth rate of the money supply, the interest rate increase would have been driven either by an unanticipated increase in the demand for money balances, owing to a fall in velocity, or by an unanticipated rise in nominal income. If a decline in velocity is responsible, the interest rate should be targeted, and to prevent a decline in GDP, action should be taken to increase the rate of money growth to move interest rates back to the targeted path. If, however, the rise in interest rates has been driven by an unanticipated increase in GDP, the money stock should be targeted, and the interest rate allowed to rise. In this case, targeting the interest rate would not allow its increase to act as an automatic stabilizer that moderates the rise in GDP.

The key advantage of using an intermediate target is that the central bank can move its instrument settings more quickly and accurately in response to a shock to the system than if it focused only on the current values of the ultimate target. However, if the central bank responds not to actual movements in the ultimate objective (for example, the inflation rate) but to projected movements, where the projection is based in part on movements in indicator variables, using an intermediate target may not speed up the policy response. Thus, whether or not an intermediate target improves how monetary policy functions is an empirical question, and the movement by central banks away from using monetary aggregates as intermediate variables has been caused by the weakened link that changes in the structure of financial systems have created between these aggregates and the ultimate objectives of monetary policy.

Operating targets must be chosen with similar criteria in mind: they must be able to be measured quickly and accurately; they must be under central bank control through its monetary policy instruments; and they must have a predictable effect on a selected intermediate variable. The choice of an operating variable will ordinarily depend on which intermediate variable has been chosen as a target, although in some countries intermediate targets have been deemphasized, and a direct link between the operating variable and the ultimate objective has been established, as, for example, in Canada (discussed later).

Monetary Policy Instruments

Monetary policy instruments are generally classified as either direct or indirect (Alexander, Baiiño, and Enoch, 1995). Direct instruments operate under the regulatory authority granted to the central bank; indirect instruments operate as a function of the central bank’s ability to issue reserve money and the consequent impact of changes in reserve money on money market conditions. Direct instruments function according to regulations that directly affect either interest rates or the volume of credit; indirect instruments function by influencing market demand or supply conditions. Indirect instruments are also termed “market-based instruments,” since their use affects the market-determined price of bank reserves as the central bank engages in transactions with both financial and nonfinancial institutions.

The most common types of direct instruments are administratively set interest rate ceilings, individual bank credit ceilings, and directed lending. There are three main types of indirect instruments: open market operations, central bank lending, and reserve requirements.

Other indirect instruments include central bank auctions of treasury bills or central bank obligations, and central bank auctions of central bank credit. The first is similar in some respects to open market operations, and the second is similar to central bank lending. Open market operations and central bank lending operate by affecting the level of reserve money, whereas reserve requirements are set and changed according to regulation and thus contain an element of direct control. However, since the effect of changes in reserve requirements is a function of the demand for reserve money, changes in reserve requirements are classified as an indirect instrument. Changes in reserve requirements generally will not change the aggregate level of reserve money, but since they affect the money multiplier, they will have an influence on the size of the money stock.

Although some countries have continued to use interest rate controls as direct instruments, they have become increasingly ineffective as markets and financial instruments develop. Typically, interest rate controls have been used when banking supervision is weak or when information on borrowers is not adequate for judging their creditworthiness. They have also been used when monetary authorities are unable to achieve a target interest rate through market mechanisms. Interest rate controls interfere with the market mechanism for credit allocation, and they promote disintermediation.

Individual bank credit ceilings are effective instruments for controlling total domestic credit growth. Like interest rate controls, however, they interfere with efficient credit allocation and lead to disintermediation.

Directed credits are restrictions placed on either central bank lending to banks or on commercial bank lending whereby particular sectors of the economy receive a predetermined proportion of the available bank credit. Directed credits are sometimes used to channel credit to public enterprises, thereby bypassing the fiscal accounts, or to promote certain “key” economic sectors. In either case, directed credits restrict market-determined credit allocation, thereby reducing the efficiency of intermediation, as well as creating incentives for disintermediation.

Open market operations consist of the central bank’s purchases or sales of financial instruments in the secondary market, although, when secondary markets are relatively undeveloped, they also may be carried out in the primary market, frequently involving transactions in central bank obligations. Instruments used for open market operations include treasury obligations, central bank obligations, and prime commercial paper. The transactions involved in open market operations may be outright purchases or sales, or they may take the form of a matched purchase and sale in which the purchase or the sale includes an accompanying obligation to reverse the transaction after a specified period. Outright purchases or sales are sometimes referred to as dynamic open market operations, because they seek to affect the growth of reserve money permanently. Matched sale-purchase transactions, also known as repurchase or reverse repurchase transactions, are referred to as defensive operations, because they are made with the intention of smoothing out the path of reserve money growth.

Central bank lending operations usually entail making loans or advances with short-term maturities. Depending on the country, they take the form of outright loans on a collateralized basis, rediscounts of high-quality financial assets (such as prime commercial paper or treasury obligations), or the extension of credit through auctions. By changing the lending or rediscount rate, central banks create an announcement effect that can enhance the transmission of monetary policy, and which is thought to be broader than the impact of an open market operation, which is limited to the securities dealers or banks in a few financial centers. However, changing the rediscount rate is of somewhat limited use in monetary targeting, since access to the rediscount window is at the initiative of the commercial banks. The central bank may also in fact be implementing a selective credit policy by setting the criteria for eligible paper for rediscounting. Some countries, notably the United States, Japan, and Germany, set the rediscount rate below market interest rates, but their central banks selectively determine which banks may use the facility. Most other countries set the rate above the market rate to discourage access.

Although changes in reserve requirements are available as an instrument of monetary policy, they are only infrequently used in practice because of their disruptive effects. They are not used for short-term liquidity management because frequent changes disrupt portfolio management by banks. Moreover, setting reserve requirements too high encourages financial disintermediation. Reserve requirements, to the extent they are used, help induce demand for reserves, thus enhancing the stability of reserve demand. Although an increase in required reserves helps to reduce liquidity, it represents a tax on financial institutions that is felt unevenly by commercial banks, depending on their balance sheet positions. And although the effective tax can be compensated by interest paid on reserves, frequent changes in reserve requirements, particularly in an upward direction, will disrupt a bank’s portfolio management.

Interaction Between the Exchange Rate and Monetary Policy

The most basic of all monetary policy choices is whether to adopt a fixed exchange rate (or some variation thereof, including such arrangements as an exchange rate band or a crawling peg) or a floating exchange rate. The conduct of monetary policy depends on this choice. Although several arguments can be made for either type of regime, or for some intermediate arrangement, some economists believe that decisions about this issue based on economic arguments can be reduced to one question: whether the resource allocation advantages of a fixed rate that arise from an extended currency area outweigh the macroeconomic or stabilization disadvantages of the inability of a country to conduct an independent monetary policy. Apart from the microeconomic advantages that might accrue to a country from choosing a fixed-rate regime, other arguments in support of a fixed rate seem to be largely political in nature (McCallum, 1999). Of course, an exchange rate band arrangement allows a country to maintain an exchange rate anchor as a way of gaining credibility while also giving the central bank some flexibility to conduct a more independent monetary policy.

Flow equilibrium in the money market is reached when the rate of money expansion equals the rate of growth in the demand for money. But stabilization also requires equality between actual and desired stocks of money and foreign assets. In a closed economy, the only source of liquidity growth is the banking system’s purchases of domestic assets. An excess expansion of credit in terms of the demand for money will create an excess of expenditure over income; and if the expansion of credit continues to exceed the growth in money demand, inflation will tend to accelerate, despite the increased demand for cash balances as prices rise. This situation is similar to one that prevails in an open economy with a floating exchange rate, where money growth is determined by the central bank’s domestic asset purchases, given that the central bank does not intervene in the foreign exchange market and foreign exchange flows are not linked to the money supply process. In a floating exchange rate system, both the rate of inflation and its variability and the rate of currency depreciation and its variability are determined by imbalances in the money market. In an open economy with a fixed exchange rate or one of limited flexibility, a link is established between the balance of payments and the money supply. The central bank can set either the monetary base or the exchange rate, but not both. Choosing the monetary base as the exogenous variable corresponds to a flexible exchange rate regime; if the exchange rate is exogenously determined in a fixed-rate regime, the money supply becomes endogenous.

Explicit allowance for private capital flows may affect this analysis. Under fixed exchange rates, and thus in the absence of exchange rate uncertainty, foreign and domestic assets are perfect substitutes, and domestic and foreign interest rates must be the same, since any attempt by the central bank to lower domestic interest rates and increase the money stock, for example, would be met by capital outflows. The money stock and interest rates would remain unchanged; the only net effect would be a change in the asset composition of the central bank’s balance sheet, as foreign assets are reduced by an equivalent increase in the bank’s domestic assets. Sterilizing the capital flow is problematic, since sterilization does not alter the interest rate differential, and because the process would have to continue, resulting in unsustainable interest costs to the government.

Under a managed float or a fixed exchange rate that is not fully credible, the central bank is able theoretically to validate an interest rate differential between domestic and foreign rates that would be due, for example, to an expected change in the exchange rate plus a risk premium.² The differential between domestic and foreign interest rates can be written as

i - i* = δ + r,

where i is the domestic interest rate, i* is the foreign interest rate, δ is the expected devaluation, and r is the risk premium. The risk premium, which investors require as an incentive for shifting their portfolios away from the minimum variance portfolio, is a function of the comparative supply of domestic versus foreign assets. By changing the composition of the world stock of assets, sterilized intervention affects the interest rate differential and allows the central bank to pursue foreign exchange and money stock targets independently (see Dornbusch and Giovannini, 1990). The practical relevance of this proposition and of sterilized intervention has not been established, and some empirical studies have shown that sterilized intervention has very limited effects. Therefore, although the presence of capital mobility is believed to reinforce the conclusion that the money supply and the exchange rate are not independent variables and that only one of the two can be determined exogenously, the issue has not been settled.

Money Versus Credit and the Transmission Process

Analysis of the transmission process in monetary policy has traditionally focused on monetary aggregates, This focus has been termed the “money” view of the transmission mechanism, whereby short-term money market interest rates are determined by the supply of and demand for money, and changes in these rates affect longer-term rates, which affect spending and output.³ Monetary policy operates by affecting bank deposits and thus the liability side of commercial banks’ balance sheets. This mechanism can operate only in the absence of perfect substitutes for money, so that economic agents cannot completely offset the impact of monetary policy by substituting alternative financial assets. To the extent that money substitutes are available, the impact of monetary policy on short-term rates will be dampened in proportion to the degree of substitutability. In recent years, however, changing financial market structures and developments in the theory of imperfect information (as applied to financial markets) have shifted the focus to the role of credit markets in the transmission of monetary policy, although the view that the availability of credit can have significant effects on real economic activity has a long history.

The “credit” view of the transmission mechanism is based on the proposition that bank credit is not a perfect substitute for other forms of borrowing. At least for a large group of borrowers, financing by means other than bank credit either is impossible or comes at a higher cost than bank credit. Not only do banks intermediate between savers and investors, but they are also able to deal efficiently with asymmetric or incomplete information between borrowers and lenders, thereby bringing them together at lower cost than would be possible in direct transactions. The presence of a credit channel depends on two propositions: that loans and securities are imperfect substitutes as bank assets, which implies that banks will not respond to monetary tightening by reducing their securities and increasing their loans, and that bank loans and nonbank credit are also imperfect substitutes, which implies that borrowers cannot resort to other sources of finance to offset, without cost, a reduction in the supply of bank loans.

The credit view of monetary policy thus suggests that a tightening of monetary policy will force banks to reduce their loans and securities. To the extent that banks reduce their loans, spending by the recipients of bank loans will decline as borrowers become unable to substitute nonbank borrowing for the decline in bank credit, as is assumed in the money view. The resulting decline in aggregate demand will be greater than can be attributed to the money channel alone. Moreover, credit market frictions may magnify the impact of shocks, including monetary policy shocks, on borrowers’ spending. For example, a tightening of monetary policy will raise short-term interest rates, thereby lowering asset values and causing the balance sheets of firms to deteriorate. This deterioration, in turn, will reduce the availability of internal funds and raise the cost of external funds, since the external funds’ premium (the agency costs involved in optimal financial contracts under imperfect information) depends on a firm’s creditworthiness. Thus, credit market imperfections may have a “financial accelerator” effect on borrowers’ expenditure.

The existence of a credit channel in the monetary transmission mechanism would imply that a credit variable could be used as an intermediate target of monetary policy. To be useful, this variable would have to satisfy the standard criteria—it would have to show a reasonably strong relationship with both the objective variable and the operating target variable of monetary policy. In practice, most central banks that use indirect monetary instruments have been unable to exercise a high degree of control over credit aggregates in the short term, and monetary aggregates have been more popular as intermediate variables, although some central banks have decided that no financial aggregate can serve adequately as an intermediate target. Monetary policy, however, must still be conducted so that it anchors expectations and removes potential inflation bias. Therefore, some central banks have decided to base monetary policy decisions on a variety of information or indicator variables, among which are measures of credit. Although money market interest rates are usually considered an important indicator of monetary conditions, the transmission of monetary policy through both money and credit channels suggests that the spread between loan rates and money market rates is a better indicator of the impact of monetary policy actions on private sector spending. Several studies have found evidence for the credit channel view, in that the comparative availability of bank lending, as measured by interest rate spreads or by the ratio of bank credit to nonbank finance, is a predictor of future movements in investment and output. Other studies, however, have challenged this finding.⁴

Particularly because empirical evidence has indicated that changes in interest rates have a rather small impact on investment, several other forms of the monetary policy transmission mechanism have been suggested (see Mishkin, 1995). These fall into three categories, according to whether they operate through international trade, consumer spending, or investment spending.

The exchange rate channel is linked to the effect of monetary policy on interest rates. A tightening of monetary policy will increase domestic real interest rates, make domestic-currency deposits comparatively more attractive, and lead to an appreciation of the domestic currency, since the interest rate parity relationship requires that the domestic currency appreciate in order for expectations of a domestic-currency depreciation to equalize rates of return between domestic and world markets. This appreciation will cause a decline in net exports and thus in aggregate output.

The asset price channel of monetary policy operates through wealth effects on consumption and investment. As a result of a contractionary monetary policy, the public will find that its money balances have been reduced below the level demanded and will try to increase its money holdings by reducing spending. Consequently, the decreased demand for equities reduces equity prices and consumers’ wealth, leading to a fall in consumption. In addition, the drop in equity prices reduces the market value of firms in relation to the replacement cost of capital (Tobin’s q theory), and firms will reduce investment spending, increasing their capital assets instead by acquiring old capital through the purchase of other firms.

Dynamic Inconsistency and Inflationary Bias

Based on data from the United Kingdom, the original Phillips curve showed an inverse long-run relationship between wage inflation and unemployment. Although data from the United States through the 1960s seemed to confirm this relationship, both theoretical and empirical studies have demonstrated the fallacy of a long-run relationship of this type. The validity of a short-term relationship between wage inflation and unemployment, however, is generally accepted, despite early results based on rational expectations that showed that systematic monetary policy can affect only the inflation rate and not employment. The view that a short-run trade-off exists is supported by econometric evidence in the United States that links the Federal Reserve’s policy decisions to tighten monetary policy with subsequent recessions, as well as econometric evidence that predictable monetary policy affects output and not only prices. A widely accepted explanation for a short-run Phillips curve trade-off is based on the existence of sticky wages and prices. The possibility that an expansionary monetary policy may expand output and employment in the short run leads to one aspect of what has been termed the “problem of dynamic inconsistency.”

Dynamic inconsistency in the context of monetary policy refers to the difference between the optimal policies that a central bank would announce if it were considered credible by the public, and the policies that the central bank would carry out after the public had made decisions on the basis of its expectations. If the central bank announces that it will conduct monetary policy on the basis of a particular target rate of inflation, and the public engages in contracts based on that announcement, the central bank may attempt to adjust monetary policy to produce higher output by promoting surprise inflation. But the public will discount the announcements of the central bank, and the resulting inflation rate will be higher than it needs to be. Output may or may not rise above the full employment rate, depending on whether underlying wage rigidities prevent complete wage and price adjustment. Another way of putting this idea is to say that policymakers unconstrained by rules have an incentive to “cheat” the private sector in order to keep unemployment below its natural rate or to increase seigniorage revenue by raising inflation. However, since rational agents account for the incentive of policymakers to promote surprise inflation, they will adjust their behavior accordingly, creating an economy with an inflationary bias.

Rules Versus Discretion

Conducting monetary policy according to rules has a long history in economics (see, for example, Humphrey, 1992, and Prescott, 1977). The importance of the debate about rules versus discretion, which was reoriented in the mid-1980s, is related to the behavior of central banks in the conduct of monetary policy: rule-like behavior implies that the central bank will conduct policy systematically while refraining from exploiting existing expectations to achieve temporary gains in output. The debate about rules versus discretion is related to, and often confused with, the debate about activist versus nonactivist policy (McCallum, 1989). The distinction between rules and discretion is rather subtle. Whereas a monetary rule may be either nonactivist or activist, the distinction between rules and discretion depends on whether, in each period, the monetary authority implements a rule that it chooses to be applicable for a large number of periods, or a policy, even if determined systematically, that has no connection with the choices of different periods. In a rule-based policy, for example, the monetary authority seeks to maximize an objective function by designing an appropriate formula to be implemented over several periods. If a formula for growth in the money stock—Δm_t, where m_t is the logarithm of the money stock—is specified as

Δm_t = 0.01 + 0.05(U_t-1 - 0.05),

where U is the unemployment rate, then policy would be activist. If the formula were designed to apply to a large number of periods, however, it would still fit the definition of a rule. In contrast, discretionary policy entails making new decisions in each period. The precise analytical distinction between policy rules and discretion is made in the literature on time consistency (see Barro and Gordon, 1983). A policy rule refers to the optimal solution to a dynamic optimization problem, whereas discretionary policy refers to the inconsistent or shortsighted solution, even though it may be a “time-consistent” strategy.

In other words, if the monetary authority is a rational policymaker free to choose the best monetary policy, the time-consistent monetary policy is the one that the monetary authority selects to optimize each time it selects a policy, even though the optimal policy would be to select a plan or rule at the beginning and then adhere to it over time. The latter procedure is referred to as “time-inconsistent,” Although the time-consistent policy may yield significant short-run social benefits, economic agents will learn to anticipate the period-by-period optimization, and the policymaker will lose credibility.⁵

Policy rules may not necessarily be determined by a mechanical formula, although the term “policy rules” connotes either a fixed setting for policy instruments or a simplistic mechanical procedure (Taylor, 1993). A policy rule might more appropriately be renamed “systematic monetary policy program,” to connote the important properties of a policy rule more effectively without also implying fixed settings or mechanical formulas.

The distinction between rules and discretion can be illustrated with a comparison between a particular objective function for the monetary authority and the money growth rates that would be established alternatively under rule-based policy and discretion-based policy (McCallum, 1989). Assume that the central bank wants to minimize the following objective function:

$O_t=a\mathrm{\Delta }{m}_t^2-b(\mathrm{\Delta }{m}_t-\mathrm{\Delta }{m}_t^e),$

where O_t is the value of the function and a and b are positive constants, and expectations are formed rationally so that the expected value for money growth in period t, $\mathrm{\Delta }{m}_t^e$, will be, on average, equal to Δm_t. Because rational agents will on average neither overpredict nor underpredict money growth, ${\mathrm{\Delta }m}_t-\mathrm{\Delta }{m}_t^e$ will equal zero, and the value of Δm_t that minimizes O_t will be zero for all t. Thus, in this example, a rule-based policy would lead to a money growth rate of zero in each period.

In contrast, for a particular period, the central bank will minimize O_t by recognizing that $\mathrm{\Delta }{m}_t^e$ at the end of the previous period has already been determined, and that the minimization of O_t may not imply making Δm_t equal to $\mathrm{\Delta }{m}_t^e$. Thus, minimization of the objective function requires setting its derivative equal to zero and solving for Δm_t = b/2a for any given period. The optimal growth rate of money under a discretionary policy would be a positive value.

In summary, although a discretionary policy would choose for a particular period a value for Δm_t that will differ from $\mathrm{\Delta }{m}_t^e$ Δm_t will, on average, equal $\mathrm{\Delta }{m}_t^e$ because of rational expectations under both types of policy. Furthermore, since the average value of the objective function is zero when Δm_t = 0 under a policy rule, and is b²/4a when Δm_t = b/2a under discretion, the value of O_t which is to be minimized, is higher under discretionary policy. Discretionary policy is thus less desirable. In other words, neither type of monetary policy will have an effect on unemployment, on average, but discretionary policy will lead to positive inflation. For any particular time period, faster monetary growth will lower unemployment and raise inflation, and the trade-off can be evaluated. Nevertheless, the effect of this policy on expectations is not captured in decisions made period by period, and the central bank does not account for the fact that it cannot reduce unemployment, on average. Decision making by rules also can indicate the authorities’ choice of an ongoing process with desirable properties (McCallum, 1989, p. 243). Although monetary authorities are unlikely to agree to surrender their power to use discretionary monetary policy, they may still find it helpful to use a policy rule as a guide for monetary policy decision making.

Nominal feedback rules are a form of policy rule whereby the central bank need not rely on a specific model of the economy in order to implement them. The nominal feedback rule includes a feedback mechanism that indicates exactly what change in the policy instrument is required when the nominal target variable deviates from its target path. Nominal feedback rules were developed in response to shortcomings in constant money-growth rules and to the inflationary bias in purely discretionary policy, among other problems (Dueker and Fischer, 1998).

Two monetary rules, in particular, illustrate the rule approach to monetary policy: the McCallum rule and the Taylor rule. Each contains a feedback mechanism for adjusting the operating variable of monetary policy in response to observed deviations of policy objectives from target or trend. The operating variable in the McCallum rule is base money; in the Taylor rule it is short-term interest rates. The McCallum rule feeds back from deviations in nominal GDP from an assumed target path, whereas the Taylor rule feeds back from deviations in inflation from the target and in real GDP from the trend. When, for example, output is above the trend and inflation is above the target, the feedback rule indicates that monetary policy should be tighter than a neutral stance.

The McCallum rule is illustrated by the following:

m = k* - v_t-1 + λ(x* - x)_t-1,

where $x_t^{*}=x_{t-1}+k^{*}$ is the nominal income growth target, m is the growth of the monetary base, x is the logarithm of nominal GDP, and an asterisk denotes a target value. The constant term k* sets the path for steady-state nominal income growth, υ is an adjustment for changes in the velocity trend of base money, and λ (x* - x)_t-1 is the feedback term that allows policy to be changed according to the deviation of nominal income from the assumed target.

The Taylor rule may be expressed as

i = p_t-1 + w₁[(Y - Y*)/Y*]_t-1 + w₂(p - p*)_t-1 + r*,

where i is the nominal interest rate, p is the inflation rate, p* is the target inflation rate, r* is the equilibrium real interest rate, (Y - Y*)/Y* is the output gap, and the w’s are weights. The performance of rules has been assessed on the basis of counterfactual simulations in which monetary policy determined according to a rule has been simulated and compared with the actual results of the stance of monetary policy (see, for example, Haldane, McCallum, and Salmon, 1996). An alternative approach is taken by Stuart (1996), who assessed whether the McCallum and Taylor rules would have been able to provide useful information about the stance of monetary policy during particular episodes. Since, in practice, central banks are unlikely to entrust monetary policy to a nominal feedback rule, such rules may have their most promising use as a reference guide for analyses of current monetary conditions.

Central Bank Independence

Regardless of the rules or the absence of rules to guide the conduct of monetary policy, it is generally recognized that a time-inconsistency problem arises when a central bank is concerned with both inflation and employment and possibly other objectives as well. For example, a central bank forced to finance the fiscal deficit by creating money may have to promise low inflation to maximize the demand for money, which forms the base of the inflation tax, and then to break that promise to increase inflationary revenue. This example suggests that time inconsistency may be a pervasive feature of monetary policy. It has been argued, more generally, that a credible framework for monetary policy is necessary to convince the private sector that the central bank will not exploit its ability to engineer short-run welfare gains, seigniorage revenue, or even electoral gains. The case for central bank independence has been argued in part on the basis of the proposal that conservative central bankers who are more averse to inflation than is society as a whole be appointed central bank governors (see Rogoff, 1985). The appointment of conservative central bankers would mitigate the inflationary bias of monetary policy, because the public would know that they would refrain from using unexpected inflation to expand employment. The conservative central banker would not attempt to stimulate employment but would be able to promise and deliver low inflation.

The issue of a central bank’s reputation and thus its credibility has been studied extensively at a theoretical level. One of the difficulties of testing the credibility problem empirically, however, is that credibility cannot be measured directly. In the theoretical discussions, credibility is linked to the central bank’s ability to make binding promises, but it is difficult to imagine how its ability to do so can be observed and measured. Therefore, empirical studies have focused on central bank independence and the relationship between independence and inflation. Since central bank independence may emerge as part of society’s attempt to eliminate the inflation bias caused by the lack of policy credibility, this approach is an indirect test of the hypothesis that more credible central banks will deliver lower inflation.

Empirical evidence on the relationship between inflation and central bank independence has been mixed. For the 1960-92 period, a measure of legal central bank independence, calculated for 18 industrial countries as a sum of 15 different legal provisions, including whether the government appoints the central bank governor and the length of the governor’s term, showed a strong negative relationship with inflation (Fischer, 1995). In a comprehensive attempt to quantify central bank independence and explore the relationship between independence and inflation, Cukierman, Webb, and Neyapti (1992) developed measures of central bank independence that were defined not as unconditional independence from government, but rather as independence to pursue the objective of price stability, even at the cost of other objectives that might be more important to political officials. They developed different rankings of central bank independence: by legal independence, by governors’ turnover rates, by the responses of specialists to a questionnaire on central bank independence, and by an aggregation of the first two rankings. The relationship observed for industrial countries, in which legal independence is an important determinant of inflation, did not extend to developing countries, where legal independence is not a statistically significant determinant of price stability. However, developing countries showed wide discrepancies between actual and legal independence, and combining information on legal independence and information on the turnover rate of governors contributed significantly to explaining inflation, since turnover is strongly and positively associated with inflation.

Other evidence bearing on the case for central bank independence includes a study of the impact of central bank independence on real economic performance (Alesina and Summers, 1993). Whereas some evidence suggests that independent central banks may be able to achieve lower levels of inflation, this line of research investigates whether a correlation exists between central bank independence and the level and variability of growth, employment, and real interest rates. A central bank free from political pressure may behave more predictably, thus promoting economic stability and reducing risk premiums in real interest rates. In addition, to the extent that high inflation affects economic performance adversely, central bank independence may improve economic performance. Conversely, to the extent that monetary policy can achieve real objectives, independent central banks may achieve lower rates of inflation at the cost of lower real economic performance. The study compared several measures of economic performance in 16 industrial countries with measures of central bank independence. The results confirmed that, although central bank independence reduces the level and variability of inflation, it does not have large benefits for (or costs to) real macroeconomic performance.

One of the most striking instìtutional developments in monetary reform in recent years has been the changing role of central banks, Between 1989 and 1995, the legal framework of the central banks in 7 industrial countries and 10 developing countries was modified in order to effect changes in how the central bank conducted monetary policy. Three objectives underscored the changes: to emphasize price stability to a greater degree; to increase the degree of central bank independence from political pressure (the central banks whose status was modified were generally freed from any obligation to finance either the government or government-controlled agencies and were given sole responsibility for setting policy instruments); and to place greater emphasis on the accountability of central banks for their actions. The reform has resulted in an increase in the operational discretion of central banks to attain their goals (Cottarelli and Giannini, 1997), As described earlier, the case for central bank independence has been based on the negative relationship between independence and inflation and the absence of an observed growth trade-off. Support for central bank independence is reinforced by the case of Germany, which has a highly independent central bank and has achieved comparatively rapid economic growth with outstanding inflation performance.

Inflation Targeting Versus Price-Level Targeting

The practice of targeting inflation attracted substantial attention in the 1990s, and several industrial countries and Israel now conduct their monetary policy on this basis (see Chapter 7 of this volume). Inflation targeting implies not just the government’s announcement of a short-term inflation target—a practice that has been followed by many countries for some time—but the announcement of a target path for inflation over the next few years, as well as the implementation of a procedure for monitoring how well the monetary authorities achieve the target path. Since inflation targets in this framework are set for the medium term, the conduct of monetary policy is not discretionary over the medium term; however, because countries that have adopted inflation targeting typically maintain a flexible exchange rate policy, domestic monetary conditions are independent of external monetary conditions.

In the short term, however, inflation targeting gives monetary policy some discretion. The monetary authorities in inflation-targeting countries use a feedback mechanism from expected, rather than actual, inflation. In the United Kingdom, the Bank of England projects inflation two years into the future and forms policy according to the extent to which the forecast and target inflation differ. Canada, New Zealand, and other countries follow a similar procedure. This method for determining monetary policy can be illustrated by the following monetary rule:

ΔI_t = γ(E_t Π_t+j - Π*),

where I_t (is the policy instrument, Π_t is inflation, Π* is the inflation target, γ is a positive feedback parameter, j is the number of time periods it takes for policy to have its maximum effect on inflation, and E_tΠ_t+j denotes the expected inflation outcome j periods ahead based on information known through time t.

An alternative approach to inflation targeting is to target the price level. The difference between these approaches is that, under inflation targeting, the rate of inflation actually experienced in any period may differ from the target because of factors beyond the control of the authorities, such as supply shocks, but the inflation target for the next period will remain unchanged; thus the impact of the missed inflation target on the price level is not compensated for. When, alternatively, a price-level target is missed, policy is adjusted over the next several periods to bring the price level back in line with its targeted path. Thus, inflation targeting has the potential to allow a drift in the price base (usually upward), creating uncertainty about the price level in the future. A price-level target, however, will create less uncertainty about the future price level, since higher-than-targeted inflation outcomes will he followed by a lower target for inflation so that the price level can move back to its target path. The possibility that negative inflation may be necessary to move the price level to its target path is cited as a reason against targeting the price level, but whether these measures are necessary would depend on the steepness of the targeted path of the price level.

Types of Frameworks

In general, the monetary framework of a country can be classified according to the existence of institutions, announced rules, or other policy guidelines that affect how the monetary authorities conduct monetary policy. Cottarelli and Giannini (1997) identified nine basic monetary frameworks, based on announced policy rules, from a longitudinal sample of the principal monetary frameworks of 100 countries. The nine frameworks are shown in Table 6.1, listed in order of increasing discretion, defined according to two considerations: whether the monetary authority is able to set short-term interest rates independently of foreign interest rates, and whether the monetary authority can surprise the private sector by promoting unanticipated inflation, without repudiating its policy announcements.

The first four frameworks are completely nondiscretionary because domestic interest rates are pegged to foreign rates through arbitrage. The fifth and sixth frameworks—exchange rate pegs with capital controls (detailed in Table 6.1)—allow some short-run discretion in monetary policy; however, for countries adopting this framework and a short-term intermediate target, the announcement of this target eliminates the scope for discretionary policy. The seventh framework, inflation targeting, does not allow medium-term discretion because the central bank must announce the targeted future inflation path and allow the public to monitor its performance in attaining the target. Since countries adopting this framework typically have flexible exchange rate systems, short-term domestic monetary policy is independent of external monetary conditions.⁶ The eighth framework includes countries with a short-term intermediate target, such as a monetary aggregate, and includes the United States until 1993, when the Federal Reserve officially announced that it was giving less weight to monetary aggregates. Countries in this category are constrained in the short term but not the medium term, since the targets are periodically revised. The last framework is one in which the monetary authority has complete discretion. Alternative classification systems are of course possible, and it may be the case that a central bank operating under a floating exchange rate system may nevertheless be more committed to maintaining an inflation or monetary target than a central bank that operates under a fixed rate system but devalues frequently.

The data presented by Cottarelli and Giannini (1997) show that, prior to the breakdown of the Bretton Woods system, only 10 percent of the sample countries used a completely discretionary framework; 56 percent followed an exchange rate peg with some capital controls, so that an element of discretionary policy was available. From 1970 to 1994, however, countries moved increasingly toward a more discretionary framework, and by 1994, 36 percent of the countries had adopted a completely discretionary framework. An index of monetary discretion shows that, on average, the degree of discretion increased particularly strongly until the late 1980s and then stabilized. Another index calculated in the same study shows that, even in countries following disinflation programs, the trend has clearly been toward discretion.

Recent Experience Under Alternative Monetary Regimes

As described in the previous section, both industrial and other countries have operated under several types of monetary regimes, but a few stand out, in particular: exchange rate targeting, monetary targeting, inflation targeting, and discretionary policy with an implicit nominal anchor.⁷ Both industrial and developing countries have long used monetary regimes based on exchange rate targeting in the form of a fixed rate; more recently, a number of developing countries have adopted exchange rate targeting in the form of a crawling peg as the basis for their monetary regimes. Exchange rate targeting has proved to be an effective regime for reducing inflation. For example, Argentina, which established a currency board in 1990 because it needed a strong disciplinary mechanism, pegged its currency to the dollar and was able to lower its annual inflation rate from more than 1000 percent in 1989 to 3.3 percent by 1995. In 1990 the United Kingdom pegged its currency to the deutsche mark and was able to reduce its inflation rate from 9.5 percent in that year to 3.7 percent in 1992.

Although exchange rate targeting has yielded benefits to the countries that have adopted this monetary regime, monetary policy under exchange rate targeting is not able to respond to domestic shocks; moreover, countries operating under this regime are subject to shocks that, acting through the interest rate link, affect the anchor country. When reunification with eastern Germany led to a fiscal expansion and higher interest rates in Germany, the shock was transmitted to the other countries of the European exchange rate mechanism (ERM) because their interest rates rose along with those in Germany; in particular, rates in France and the United Kingdom were higher than they would have been had monetary policy been focused on domestic conditions. The result was that growth slowed and unemployment rose in France, and the United Kingdom dropped out of the ERM, primarily to be able to adapt monetary policy to its domestic circ*mstances.

Several countries adopted monetary targeting in the 1970s. The United States, the United Kingdom, and Canada all targeted monetary aggregates, but all were unsuccessful at using this method to control inflation, although it has been argued that none of them pursued monetary targeting seriously (Mishkin, 1998). Not until the early 1990s did the three countries formally abandon monetary targeting, although all three had substantially reduced their reliance on monetary targets since the early 1980s, recognizing that the relationship between monetary aggregates and inflation and nominal income had become extremely unstable.

The success of monetary policy in both Germany and Switzerland, however, has given strong support to advocates of monetary targeting. Both Germany and Switzerland use a flexible approach to monetary targeting. The German Bundesbank, in particular, has allowed its medium-term inflation goal to vary in response to shocks, but since 1984 it has set a “normative rate of price increases” of 2 percent a year. On the basis of estimates of potential output growth and velocity trends, a target growth rate for the monetary aggregate is then set. The target ranges for money growth, however, have frequently been missed because both the Bundesbank and the Swiss National Bank have been concerned with output and exchange rates in addition to inflation. An important aspect of the monetary regime in both countries has been a strong commitment to communicating monetary strategy to the public, thereby enhancing transparency. Despite the inability of the monetary regimes in both countries to attain their intermediate targets consistently, both Germany and Switzerland have achieved good inflation outcomes, and it is apparent that monetary targeting can work successfully toward the ultimate objectives of monetary policy.

Largely in an effort to reduce inflation, and in response to the instability of velocity, several countries have adopted inflation targeting as their monetary regime. In 1990 New Zealand was the first country to adopt inflation targeting formally, with several other industrial countries following;⁸ Israel and Chile also adopted a form of inflation targeting. Inflation targeting requires that medium-term numerical targets for inflation be announced to the public, and that the paramount goal of monetary policy be a commitment to price-level stability. All of the inflation-targetìng countries have chosen to target the inflation rate rather than the price level; target rates of inflation currently range from a midpoint of 1.5 percent a year in New Zealand to a midpoint of 2.5 percent in Australia, Spain, and the United Kingdom, with somewhat higher levels for other countries.

Inflation-targeting countries do not ignore fluctuations in output and employment, and they have shown flexibility in response to supply shocks. The official price index used to assess whether the inflation target is met frequently excludes items particularly likely to reflect the effects of shocks, such as food and energy prices, indirect tax rate changes, and the direct effects of interest rate changes on the index. The targeting regime of New Zealand’s Reserve Bank contains an explicit clause that allows the bank to deviate from its planned policies.

Finally, an important aspect of inflation targeting is the increased accountability of the central bank for the inflation outcome. In particular, New Zealand’s government has the right to dismiss the governor of the Reserve Bank if inflation targets are breached, even for one quarter.

In recent years, several countries have operated within a monetary framework that has no explicit nominal anchor, whether the exchange rate, a monetary aggregate, or inflation. In 1979 the Federal Reserve System in the United States began to announce annual targets for a narrow-money (Ml) aggregate as its intermediate variable. But because Ml became extremely interest-elastic and its velocity more variable, it became less useful as an intermediate target, and the Federal Reserve ceased setting target ranges for Ml growth in 1987. Broader monetary aggregates replaced Ml, but only as indicator variables, since no single variable or set of variables has been considered a reliable intermediate target (Madigan, 1994). Therefore, the policy regime in the United States does not target a nominal variable explicitly, although it has been characterized as having an implicit nominal anchor in the sense of the Federal Reserve’s overriding concern to control inflation over the long run (Mishkin, 1998). The discretionary strategy pursued by the United States has had demonstrable success in addressing inflation and employment, but this policy is not transparent and is particularly prone to the problem of time inconsistency. Moreover, the success of monetary policy in the United States seems to depend heavily on the public’s confidence in the person occupying the position of Chairman of the Board of Governors of the Federal Reserve System.

1

Reserve money, also known as the monetary base or high-powered money, includes currency issued plus bank and nonbank deposits with the central bank.

2

This result is based on the mean variance approach to portfolio choice. See Dornbusch (1983).

3

The impact of changes in nominal money balances on real economic activity depends, however, on the existence of sticky prices arising from contractual rigidities or imperfect information, or expectations, as described later.

4

For a review of this literature, see Kashyap and Stein (1993).

5

For examples to illustrate this point, see Sachs and Larrain (1993) and McCafferty (1990).

6

For example, see the last section of this chapter for a discussion of Canada’s monetary policy framework, which was changed to one of inflation targeting in 1991.

7

This section draws on Mishkin (1998).

8

These included Australia, Canada, Finland, Spain, Sweden, and the United Kingdom.