Chapter 12: Monetary policy

Aims of the chapter

This chapter discusses the design and the conduct of monetary policy, as well as its use as a tool to stabilise the economy over the business cycle. First, we revise the controversy between the Keynesian and the monetarist views on the best conduct (active versus passive) of macroeconomic policy, in particular highlighting the impact of macroeconomic uncertainty on the effectiveness of monetary policy actions. Then, we focus on three crucial
topics for monetary policy: the difference between money- and inflation targeting regimes, the debate between pre-committed and discretionary monetary policy, and the design of ad-hoc rules for the conduct of monetary policy.

• recognise the sources of business-cycle fluctuations and highlight the benefits of macroeconomic stability
• explain why the outcome of monetary policy actions is subject to a high degree of uncertainty
• describe how successfully monetary policy instruments can be used to achieve price stability
• explain the theoretical foundation of inflation targeting policies
• present the time-inconsistency problem and the debate between pre-commitment and discretion in the context of monetary policy
• critically discuss the implications of the main rules designed by macroeconomists for the conduct of monetary policy.

In this chapter of the subject guide we examine the design and the conduct of monetary policy, and its role in delivering economic stability. The first part of the chapter revises the concept of macroeconomic stability and discusses the controversy between the Keynesian view in favour of active macroeconomic policy, and the monetarist view, which supports a passive policy stance. We pay particular attention to the implications that this debate has for monetary policy. The rest of the chapter is dedicated to assessing the design and conduct of monetary policy. When intervening in the money market for the sake of stabilising macroeconomics fluctuations central bankers must choose precisely what instrument they wish to use: the supply of money or the nominal interest rate.

We explain why, since the late 1980s, most OECD central banks have adopted an inflation target rather than nominal money targets, and monetary policy has mainly been conducted by controlling the nominal interest rate, rather than the rate of nominal money growth. We give an overview of the theoretical underpinnings of so-called inflation targeting policies, and explain that the outcome of monetary policy depends on whether the central bank uses its full discretion or follows a pre-determined policy rule when setting its policy actions. We conclude the chapter with a brief review of the main rules proposed by macroeconomists for the conduct of monetary policy, paying particular attention to Taylor’s rule.

Macroeconomic stability

According to the traditional Keynesian view macroeconomic instability results from the combination of two fundamental factors: random disturbances or shocks, which continuously hit the economy over time, and nominal rigidities, which prevent output and inflation from adjusting instantaneously when a shock occurs. Under the theory the main sources of economic disturbances are demand, rather than supply, shocks. For this reason macroeconomic stabilisation policies are also referred to as demand-management policies. Since the aggregate demand curve is derived from the combination of the IS and the LM curves, demand shocks are ultimately IS shocks, such as unexpected changes in the confidence of consumers or businesses, and LM shocks, such as unexpected changes in the money multiplier that affect the overall money supply. In contrast, supply shocks derive mainly from unexpected changes in the costs of production, technological progress, and disturbances in the labour market.

Demand and supply shocks have different effects on output and inflation. Demand shocks cause these variables to co-move positively with one another. For example, a positive demand shock, such as an increase in consumer confidence, raises both inflation and output. In this case a monetary contraction would help to moderate both inflation and output. In contrast, supply shocks cause negative co-movement between inflation and output. For example, an adverse supply shock, such as a surge in the oil price, would increase inflation and reduce output. In this case, a monetary contraction would help to moderate inflation, but would further depress output. Conversely, a monetary expansion would increase output but contribute to fuel inflation even further.

The effectiveness of macroeconomic policy as a stabilisation tool not only depends on the source of the disturbances affecting the economy, but also on their duration. Macroeconomic shocks can be either permanent or temporary, and policymakers have the option of not reacting when facing a temporary shock as its effect will fade away over time.

It is important to bear in mind why macroeconomists and policymakers regard output and price stability as being beneficial for an economy. The main effect of economic stability is to reduce the uncertainty faced by firms, individuals and governments, since less uncertainty means that they can make their plans for the future with greater confidence.

Economic stability is particularly beneficial for growth, inflation, and unemployment. During a deep economic recession, for example, firms may quickly scrap physical capital, which could have taken a long time to build up in periods of growth. Uncertainty about future economic conditions harms investment spending, since firms will often find it convenient to delay, rather than embarking on, investment projects in the face of greater uncertainty about the future status of the economy. Inflation volatility can have several detrimental effects, including distorting investment decisions, and increasing the risk of making investment. It is also most often associated with high inflation. Unemployment volatility gives rise to a substantial social cost, in that high and persistent unemployment affects workers in terms of lost security and income, but it is also detrimental for the human capital and the overall productivity of a country, since people can find their skills depreciate very rapidly during spells in unemployment.

Active versus passive demand management policies

The AD–AS framework suggests that monetary policy can be employed as an effective stabilisation tool, at least in response to demand shocks: monetary expansions can successfully speed up the recovery of an economy from a recession, whereas monetary contractions can slow down output and inflation when the economy is growing too fast.

The main implication arising from this model is that monetary policy is an effective tool in delivering macroeconomic stabilisation, and central bankers should actively engage in fine-tune policies to limit business-cycle volatility due to fluctuations in aggregate demand.

This Keynesian view in favour of active macroeconomic policy has been challenged by the monetarist view that policymakers should, instead, be encouraged to refrain from using monetary policy as a stabilisation tool, on the grounds that the delays in both the implementation and the effects of monetary policy, as well as the uncertainty surrounding the structure and the behaviour of the economy, make it difficult for policymakers to
predict the exact response of the economy to policy changes. Supporters of the monetarist view argue that demand-management policies not only may be less successful than predicted by Keynesian theory, but may also lead to destabilising outcomes as their effects end up being felt in a phase of the business cycle different from that in which the shock occurred.

As far as monetary policy delays are concerned, economists distinguish between inside lags, namely the period of time it takes for a policy action to be implemented, and outside lags, namely the period that elapses between the actual implementation of a policy and the time it takes for that policy to exert its full effect on the economy. Inside lags are divided into recognition, decision and action lags. The recognition lag is the delay between the time a shock to the economy has occurred and the time that the central bank recognises that an action is required. The decision lag is the time between the recognition of the need for an action and the actual policy decision. The action lag is the time between the policy decision and its practical implementation.

The empirical evidence arising from OECD central banks suggests that inside lags are rather short. Recognition lags are on average about five months, whereas decisions lags are even shorter, since central banks discuss and decide on monetary policy very frequently over the year. Action lags are practically zero, since monetary policies (changes in the nominal interest rate) can be implemented almost immediately after a decision has been made. Outside lags are longer than inside lags, since once a monetary policy action has been implemented its effect on spending and output builds up only gradually over several quarters. This is because monetary policy changes affect, after some time, investment and consumption spending. The resulting changes in aggregate spending lead to a change in output, which may further affect investment and consumption. The delays caused by recognitions and outside lags imply that by the time the policy change begins to exert its effect on output the economy may be already going through a phase of recovery. In that case, the stabilisation policy may actually either accelerate the process of recovery already started in the economic system, or even have a destabilising effect, causing output to overshoot its potential level.

The success of demand management policy is also undermined by uncertainty surrounding the behaviour of the economy and the nature of economic shocks. Uncertainty about the structure of the economy means that policy-makers have to rely upon imperfect theoretical models of the economy and the quantitative predictions arising from their empirical estimation. There is ongoing disagreement among macroeconomists over the best model for this purpose, which ultimately implies that policymakers can never know the exact behaviour of the economy and cannot predict with precision how it will respond to a policy change.

Uncertainty about the nature of economic disturbances comes from the difficulty of distinguishing whether shocks are permanent or temporary, as well as understanding their source. The former issue is relevant since a temporary shock is likely to last only for a very short period, so that once the effect of the shock has faded away macroeconomic variables should revert to their initial levels. In this case a permanent policy change may have a destabilising effect on the economy, and the best policy action for the central bank may be either a short change in policy, or even not to respond to the shock at all. Similarly, it is important for the central bank to detect the exact source of economic shocks, since this is the starting point of any (possibly) successful policy action. Suppose that the central bank observes a sudden fall in output but it does not know whether it is the outcome of a demand or a supply shock. If the shock is demand-driven then a monetary expansion would be the right policy action. However, if the adverse shock comes from the supply side of the economy then a monetary expansion may not be, not least because of its undesirable effects on inflation.

To summarise, uncertainty about:

i. the duration of inside and outside lags
ii. the structure and behaviour of the economy
iii. the nature and source of economic shocks

implies that monetary policy must necessarily rely upon estimates and forecasts, so that the effects of stabilisation policies are subject to a high degree of uncertainty.

A third argument against demand management policies comes from the famous Lucas critique. The essence of the Lucas critique is that traditional models, when employed to assess the impact of macroeconomic policy changes on the economy, may yield misleading results if the parameters of the model are likely to change as agents revise their expectations when a new policy regime is implemented. As highlighted in Chapter 6, a famous application of the Lucas critique arises in the analysis of the inflation-unemployment trade-off, and estimates of the cost (sacrifice ratio) of alternative disinflation policies. Lucas’ argument is that it is incorrect to use the initial estimates of the AD–AS model to assess the likely response of the economic system to a changed inflation target since the original estimates of the model were consistent with the old policy and a change in policy will be associated with a change in the structure of the economy away from that which yielded these initial estimates.

A final element to consider when evaluating the likely effects of macroeconomic policy actions are the implications of the rational expectations hypothesis for policymaker decisions. The policy ineffectiveness proposition argues that, under rational expectations, an absence of price rigidities, and perfect information, anticipated demand management policies should be irrelevant for the economy, as rational wage- and price-setters will not be deceived by them into allowing output to depart from its natural level. Hence, only unanticipated policy changes can affect the short-run level of output. Moreover, disinflationary policy may be less costly than predicted by traditional Keynesian theory, since the cost of disinflation depends highly upon the way in which the private sector forms its expectations. Under adaptive expectations a reduction in inflation regimes leads to a, possibly large, short-run increase in unemployment. However, if agents are forward looking and have rational expectations, inflation can be reduced at almost zero cost, as long as the central bank’s policy is credible.

Monetary policy: targets and instruments

The debate about active and passive monetary policy is mainly a discussion of whether, and to what extent, central banks should respond to economic disturbances. A second important debate concerns the overall conduct and design of monetary policy. This debate is focused on two central questions:
What are the objectives that a central bank should try to pursue? And what instruments can it efficiently deploy to achieve these objectives?

The main goal of most modern central banks is to achieve price stability, by pursuing a policy that maintains a low inflation rate. Central banks have two main instruments at their disposal to carry out their policy actions, since they have the capacity to influence both the supply of
money and the short-term nominal interest rate. An important point to make is that a central bank cannot simultaneously set the interest rate
and the stock of money at arbitary target levels. As shown in Panel A of Figure 12.1, given a price level P, a central bank cannot, for instance,
simultaneously target an interest rate i * and a stock of money M * , since any combination is, ultimately, constrained by the slope of the money demand curve. Panel B in Figure 12.1 illustrates this issue from a different angle. Consider a central bank that targets an interest rate i * and a money supply M * , which are jointly consistent with the initial money demand curve M 1 D , and equilibrium at point E. Next, suppose that money demand rises from M 1 D to M 2 D . Targeting the interest rate i * clearly implies that the central bank must allow the money stock to deviate from M * as a result of the money demand increase. Conversely, targeting a money supply equal to M * implies that the central bank must let the interest rate rise above i * .

Until the 1990s many central banks in the OECD area had medium-term nominal money growth targets, which they tried to hit by determining the appropriate short-run rate of money growth. This policy was mainly dictated by the fact that under the classical quantity theory of money the inflation rate is entirely determined by the money growth rate in the medium run. Therefore, close control of nominal money growth ought to – according to this theory – deliver close control over inflation. But recall that the link between inflation and money growth under the quantity theory of money only holds as long as money demand – as measured by velocity – is constant over time. Indeed, as already discussed in the previous chapter of the subject guide, the velocity of money has been increasingly volatile over the last 20 years as a result of continuous changes in money demand, often induced by technological innovations that have affected the costs of holding or obtaining money. This has made the relationship between money growth and inflation less and less tight over time, so that inflation has become more difficult to control through the money supply. Therefore since the beginning of the 1990s, most OECD central banks have switched their focus to direct inflation rate targets, rather than using nominal money growth rate targets as an indirect means of controlling the price level. To achieve their targets these central banks have switched to using the short-run nominal interest rate as the main policy instrument.

The source of macroeconomic shocks is also a fundamental element when evaluating whether monetary policy should be carried out through the control of either the interest rate or the money stock. If macroeconomic disturbances are mainly due to IS shocks, then it is generally best for the central bank to keep the money supply constant. This is illustrated in Panel C of Figure 12.1. The economy is initially in equilibrium at point
E, where output equals Y * and the interest rate equals i * . An adverse IS shock reduces output and the interest rate to Y 1 and i 1 respectively, as the IS curve shifts from IS 1 to IS 2 but the LM curve is unaffected so long as the central bank keeps the money supply constant. Conversely, if the central bank keeps the interest rate constant at i * then it has to contract the money supply which further decreases output to Y 2 . If macroeconomic shocks are, however, mainly the result of LM disturbances then it is best for the central bank to conduct monetary policy by keeping the interest rate constant. This is shown in Panel D of Figure 12.1. The economy is initially in equilibrium at point E, where output equals Y * and the interest rate equals i * . An adverse money demand shock shifts the LM curve upwards, reducing output to Y 1 and increasing the interest rate to i 1 . The central bank will not respond to this shock if it keeps the money supply constant, but if it keeps the interest rate constant at i * then it has to expand the money supply restoring output to the initial level Y * .

Inflation targeting

As mentioned in the previous section, since the early 1990s many OECD central banks have adopted inflation targeting regimes for the conduct of their monetary policy. Broadly speaking, inflation targeting is defined as the official commitment of the central bank to the objective of achieving and maintaining the stability of the general price level. In practice, inflation targeting regimes are characterised by either explicit or implicit quantitative inflation targets, either intervals or point targets, where the centre of the interval or the point target currently varies across countries from 1 to 3 per cent per year. Under an inflation targeting regime monetary policy is carried out through the appropriate manipulation of the short-term nominal interest rate in an attempt to ensure that the desired inflation rate is realised.

The analytical underpinning of inflation target policies can be assessed by considering the expectations-augmented Phillips curve, which is reproduced below for convenience:
pi t = pi t − α ( u t − u n ) (12.1)

The Phillips curve shows that unemployment exceeds its natural rate if the expected inflation rate is higher than the current rate. If the central bank’s inflation target is indicated with bar pi, and expectations are anchored to this target so that pi_t^e = bar pi , then the Phillips curve yields:

pi_t = bar pi - alpha (u_1 - u_n) (12.2)

which shows that if the central bank hits the inflation target in any period, unemployment will be at its natural level. Under this framework stabilising fluctuations in output caused by demand shocks will also mean stabilising deviations in inflation from its target level. To see this, recall that Okun’s law shows that deviations of output from its potential level are inversely related to deviations of unemployment from its natural level. Therefore, the Phillips curve in equation (12.1) can also be written as:

Y_t = Y_n + gamma (pi_t - pi_t^e) (12.3)

for some parameter γ > 0, which shows that if the central bank hits the inflation target then output equals its potential value. This implies that while manipulating the short-term nominal interest rate to achieve its inflation target, the central bank will also indirectly stabilise real output. For instance, equation (12.3) suggests that an adverse demand shock will reduce inflation below target while triggering an economic recession. In this circumstance, a nominal interest rate cut contributes to restoring the inflation to target, but will also stimulate the real economy. The coherence between the needs of inflation stabilisation and output stabilisation in such circumstance is sometimes known as the ‘divine coincidence’.

Rules versus discretion

An important issue concerning the conduct of inflation targeting monetary policy is whether a central bank should use full discretion or instead commit to a pre-determined policy rule when setting the nominal interest rate.

Many economists argue that pre-commitment to a policy rule is superior to discretion because of the so-called time-inconsistency issue, which arises whenever policymakers have an incentive to deviate from a pre-announced policy once the private sector has formed its expectations based upon that initial announcement.

To illustrate the time-inconsistency argument consider the expectation- augmented Phillips curve obtained from rearranging equation (12.1) as:

u_t = u_n − β ( pi_t − pi_t^e ) (12.4)

where u t is the current unemployment rate, u n is the natural unemployment rate, pi_t is the current inflation rate, pi_t^e is the expected inflation rate, and the parameter β = 1/α measures the response of unemployment to surprise inflation. Suppose that the central bank’s preferences are represented by a loss function L, given by:

L ( u , pi ) = γ u ( u t − u n ) + γ_pi pi_t^2 (12.5)

which shows that the central bank dislikes deviations of unemployment from its natural level, u_t – u_n , and inflation volatility, pi_t^2 . The parameters γ_u and γ_pi are preference weights that measure how much the central bank dislikes unemployment and inflation volatility, respectively.

If γ_u = 0, the central bank is said to be strict inflation targeting, since it cares only about inflation volatility. If γ_pi = 0, the central bank is said to be strict unemployment targeting, since it cares only about deviations of unemployment from its natural level. In our example we do not impose any restriction on these two parameters, which is equivalent to assuming that the central bank cares about both unemployment and
inflation volatility.

Consider what happens under commitment. For simplicity, assume that the central bank controls directly the inflation rate and announces that it is committed to a specific inflation target bar pi . If the commitment of the central bank is credible, the public forms its expectations on the basis that inflation will be equal to the announced target, bar pi = pi_t^e. The Phillips curve in equation (12.4) implies that if the central bank then goes ahead with its policy commitment there will be no surprise inflation and unemployment will be at its natural rate. Note that the best rate of inflation for the central bank to commit to, given the loss function that we have assumed, is zero.

Now consider what happens when the central bank has discretion to deviate from its pre-announced policy. It still is assumed to have direct control over the inflation rate, but does not have to pre-commit to any specific target.

We assume that the public rationally understands the policy choice facing the central bank, and uses this understanding when setting pi_t^e. Next, the central bank sets the current inflation rate pi_t , The difference between the expected and the current inflation (i.e. surprise inflation) determines the unemployment rate and the actual level of output. So the key point is that, under discretion, the central bank takes inflation expectations as given when deciding about the actual rate of inflation. In this context, the optimal inflation rate chosen by the central bank can be calculated by first substituting the Phillips curve in equation (12.4) into the loss function of
the central bank described by equation (12.5), to obtain:

L (ui Pi t ) = - yu β ( Pi − Pi ) + γ Pi 2 .

Next the central bank chooses the inflation rate that minimises this loss function. This can be calculated by differentiating L with respect to the inflation rate as follows:

d L ( u t , pi t ) / d pi_t = = − γ_u β + 2γ_pi pi_t .

Setting the above first-order derivative equal to zero and solving for pi_t yields:

pi_t = γ_u β / 2_γpi (12.6)

which shows that the inflation rate chosen by the central bank under discretion is generally different from zero. Therefore, the public will rationally form its expectations by setting

pi_t^e = γ_u β / 2_γpi

so that inflation is positive when unemployment is at its natural level.

It is important to compare the outcome of the two solutions. The unemployment rate equals the natural rate under both pre-commitment and discretion. However, optimal pre-commitment yields zero inflation whereas discretion leads to positive inflation. For this reason pre- commitment is preferred to discretion, since it yields a lower valve for pi_t^2

Note that under discretionary monetary policy the central bank has an incentive to announce, for instance, a zero inflation target but then to choose the optimal positive inflation rate in equation (12.6), as this would result in surprise inflation and reduce unemployment. Conversely, a zero inflation rate policy is not credible under discretion: the public knows that the central bank has an incentive to renege on its announcement in order to reduce unemployment (inflation bias), and will not believe the announcement in the first place.

The result in equation (12.6), however, shows that there is a case in which discretionary monetary policy yields a zero inflation rate as under pre-commitment. This occurs when the central bank dislikes inflation much more than unemployment. Analytically, this case occurs when the parameter γ_pi is much larger than γ_u . This result suggests that zero or very low inflation can be achieved under discretion if the head of the central bank has a reputation for strongly disliking inflation. This has been used as an argument for appointing ‘conservative’ central bankers when delegating monetary policy. More generally, the main solutions that have been proposed for overcoming the time-inconsistency problem are:

• Constitutional rules, which consist in writing down the commitment to a low inflation rate in the constitution of the central bank.
• Reputation, which means that central banks have an incentive to stick to their announced policies if they are sufficiently concerned about their future credibility.
• Central bank independence, which consists in delegating the conduct of monetary policy to an independent central bank, which should be solely focused on achieving price stability.

Figure 12.2 gives a graphical illustration of the time-inconsistency problem. The long-run equilibrium of the economy must lie on the LRPC. Of all such possible equilibria, both the public and the central bank prefer point A 1 since it delivers full employment at zero inflation. But suppose that the central bank had announced a zero inflation target, consistent with point A 1 . Then the economy will be operating on SRPC 1 , and the central bank has an incentive to inflate the economy to point B, at which point the marginal loss from higher inflation equals the marginal benefit from lower unemployment. At point B inflation is higher than anticipated, so the Phillips curve shifts from SRPC 1 to SRPC * . The new long-run equilibrium is at point A 2 , at which output is at the natural level but inflation is positive and equal to pi * , even though everyone prefers zero inflation. If the central bank announces that it will return to zero inflation, this promise is not credible as everyone knows that once on SRPC 1 the central bank has an incentive to renege on its initial commitment and to choose point B.

Monetary policy rules

The design of a monetary policy rule requires the preliminary definition of the objectives that the central bank aims to achieve by committing to that rule. According to the traditional monetarist view, if the central bank wants to achieve a constant rate of inflation, then it has to follow a
constant nominal money growth-rate rule described as:

∆ M t +1 / M t = abs g M

which implies that the central bank controls the growth rate of the money supply, ∆ M t + 1 / M t , so that it is always in line with the targeted growth rate g M . Note that under the assumptions of the quantity theory of money, inflation is entirely determined by money growth, so g M represents the implicit inflation target of the central bank. 1 As discussed earlier in the chapter, the success of money growth-targeting rules in achieving stable inflation is seriously undermined by the instability of money demand and the objective difficulty faced by central banks in controlling the overall money supply in the short run. For this reason, money supply rules are mainly set to establish medium and long-term targets in the conduct of monetary policy, but have limited use in the short run.

The central bank could alternatively commit to a real GDP targeting rule described by:

∆ Y t +1 / Y t = bar g y

where abs g y is the real GDP target. Such a rule assumes that the central bank adjusts monetary policy so that the growth rate of real GDP, ∆ Y t + 1 / Y t , in every period t is equal to the real GDP target g y . If the real GDP target is, however, higher than the natural growth rate of output, then commitment to a real GDP rule must lead to an ever-increasing inflation rate, and it is, therefore, an unsustainable policy in the long run.
If the central bank has both inflation and real GDP targets, it could commit to a nominal GDP-targeting rule described by:

( ∆ Y t + 1 / Y t ) + pi t = abs g

where the left-hand side of the policy rule indicates nominal GDP growth (real GDP growth plus inflation) and the term g specifies the constant nominal GDP target. Under a nominal GDP-targeting rule the central bank chooses the growth rate of nominal GDP, which means that it will respond
to fluctuations in both real GDP and inflation.

As discussed earlier in the chapter, the main policy instrument employed by modern central banks for the conduct of monetary policy is the short-
term interest rate. Monetary policy rules that focus directly on this instrument variable, rather than broader targets, are often referred to as
interest rate rules. The best known example of such rules is the Taylor rule, which describes how central banks could manipulate the short-term
nominal interest rate in order to achieve a desired inflation target and keep real GDP at its natural level. Analytically, the Taylor rule is described by the following equation:

i_t = r^* + pi_t + f_pi ( pi_t − bar pi ) + f_y ( y_t − y_n )

where i t represents the central bank policy rate, r * the equilibrium real interest rate, pi_t the inflation rate, bar pi the inflation target and y t – y n the output gap, measured as the deviation of real GDP from its potential level. Taylor found that in the USA a rule with parameters set arbitrarily to r * = 2, pi_t = 2, f_pi = 0.5 and f y = 0.5 tracked the actual federal funds rate fairly well between 1987 and 1992, so that this rule seems to give a good description of how policy was actually set in the USA over this period. In this case, the Taylor rule is numerically written as:

i_t = 2 + pi_t + 0 . 5 ( pi_t − 2 ) + 0.5 ( y_t − y_n ),

or equivalently as:

i_t = 1 + 1.5 pi_t + 0.5 ( y_t − y_n ). (12.8)

Equation (12.8) shows that under the Taylor rule the interest rate responds more than one-for-one to changes in inflation, whereas it responds less than one-for-one to changes in the output gap. The former feature is known as the Taylor principle. Consider what happens if inflation increases by one per cent. Under the Taylor rule the nominal interest rate i t , which equals the real interest rate plus inflation, increases by more than the increase in inflation. Consequently, the Taylor rule ensures that the real interest rate, which is the variable determining investment spending and thus affecting aggregate demand, will rise when inflation increases, hence helping the economy to slow down.

Blanchard, O., Johnson, D.R., (2013) Macroeconomics (sixth edition), Pearson

Chapter 17: Expectations, Output, and Policy

17-1 Expectations and Decisions: Taking Stock

The theme of Chapter 16 was that both consumption and investment decisions depend very much on expectations of future income and interest rates.

Define aggregate private spending as the sum of consumption and investment spending:

A (Y, T, r) := C (Y - T) + I (Y, r)

where A stands for aggregate private spending, or, simply, private spending. With this notation we can rewrite the IS relation as

Y = A (Y_+ , T_- , r_- ) + G

The second step is to extend equation to take into account the role of expectations. The natural extension is to allow spending to depend not only on current variables but also on their expected values in the future period:

Y = A (Y_+ , T_- , r_- , Y_+^e , T_-^e , r_-^e ) + G

Increases in either current or expected future income increase private spending.
Increases in either current or expected future taxes decrease private spending.
Increases in either the current or expected future real interest rate decrease private spending.

A decrease in the current real interest rate, given unchanged expectations of the future real interest rate, does not have much effect on spending.

The multiplier is likely to be small. Recall that the size of the multiplier depends on the size of the effect a change in current income (output) has on spending. But a change in current income, given unchanged expectations of future income, is unlikely to have a large effect on spending.

The LM relation we derived in Chapter 4 and have used until now was given by M/P = YL(i) where M>P is the supply of money and YL1i2 is the demand for money. Equilibrium in financial markets requires that the supply of money be equal to the demand for money. The demand for money depends on real income and on the short-term nominal interest rate—the opportunity cost of holding money.

We have seen that expectations about the future play a major role in spending decisions. This implies that expectations enter the IS relation: Private spending depends not only on current output and the current real interest rate, but also on expected future output and the expected future real interest rate.

In contrast, the decision about how much money to hold is largely myopic: The two variables entering the LM relation are still current income and the current nominal interest rate.

17-2 Monetary Policy, Expectations, and Output

The interest rate that enters the LM relation, which is the interest rate that the Fed affects directly, is the current nominal interest rate. In contrast, spending in the IS relation depends on both current and expected future real interest rates. Economists some-times state this distinction even more starkly by saying that, while the Fed controls the short-term nominal interest rate, what matters for spending and output is the long-term
real interest rate.

The real interest rate is approximately equal to the nominal interest rate minus expected current inflation:
r = i - p^e

When the Fed increases the money supply—decreasing the current nominal interest rate i—the effects on the current and the expected future real interest rates depend, therefore, on two factors:

Whether the increase in the money supply leads financial markets to revise their expectations of the future nominal interest rate, i^e .
Whether the increase in the money supply leads financial markets to revise their expectations of both current and future inflation, p e and p^e . If, for example, the change in money leads financial markets to expect more inflation in the future— so p^e increases—the expected future real interest rate, r^e , will decrease by more than the expected future nominal interest rate, i^e .

The IS curve is steeply downward sloping. We saw the reason why earlier: For given expectations, a change in the current interest rate has a limited effect on spending, and the multiplier is small. The LM curve is upward sloping. An increase in income leads to an increase in the demand for money; given the supply of money, the result is an increase in the interest rate. Equilibrium in goods and financial markets implies that
the economy is at point A, on both the IS and the LM curves.

Now suppose the economy is in recession, and the Fed decides to increase the money supply.

Assume first that this expansionary monetary policy does not change expectations of either the future interest rate or future output. In Figure 17-4, the LM shifts down, from LM to LM'. (Because I have already used primes to denote future values of the variables, we shall use double primes, such as in LM', to denote shifts in curves in this chapter.) The equilibrium moves from point A to point B, with higher output and a supply has only a small effect on output: Changes in the current interest rate, unaccompanied by changes in expectations, have only a small effect on spending, and in turn a small effect on output.

The effects of monetary policy—the effects of any type of macroeconomic policy for that matter—depend crucially on its effect on expectations:
* If a monetary expansion leads financial investors, firms, and consumers to revise their expectations of future interest rates and output, then the effects of the monetary expansion on output may be very large.
* But if expectations remain unchanged, the effects of the monetary expansion on output will be small.

The role of expectations is even more central in a case we have left aside, namely the case where the economy is in the liquidity trap and the short-term rate is already equal to zero.

Saying that the effect of a particular policy depends on its effect on expectations is not the same as saying that anything can happen. Expectations are not arbitrary. .. Economists refer to expectations formed in this forward-looking manner as rational expectations.

Chapter 22: Should Policy Makers Be Restrained?

22-1 Uncertainty and Policy

Take an economy with high unemployment, where the central bank is considering the use of monetary policy to increase economic activity. Assume that it has room to decrease the interest rate; in other words, leave aside the even more difficult issue of what to do if the economy is in the liquidity trap. Think of the sequence of links between a reduction in the interest rate that the central bank controls and an increase in output—all the questions the central bank faces when deciding whether, and by how much, to reduce the interest rate:

* Is the current high rate of unemployment above the natural rate of unemployment, or has the natural rate of unemployment itself increased?

* If the unemployment rate is close to the natural rate of unemployment, is there a significant risk that an interest rate reduction through monetary expansion will lead to a decrease in unemployment below the natural rate of unemployment and cause an increase in inflation?

* What will be the effect of the decrease in the short-term interest rate on the long-term interest rate? By how much will stock prices increase? By
how much will the currency depreciate?

* How long will it take for lower long-term interest rates and higher stock prices to affect investment and consumption spending (Chapter 16)? How long will it take for the J-curve effects to work themselves out and for the trade balance to improve? What is the danger that the effects come too late, when the economy has already recovered?

Should uncertainty about the effects of policy lead policy makers to do less? In general, the answer is: yes.

There is substantial uncertainty about the effects of macroeconomic policies. This uncertainty should lead policy makers to be more cautious and to use less active policies. Policies should be broadly aimed at avoiding large prolonged recessions, slowing down booms, and avoiding inflationary pressure. The higher unemployment or the higher inflation, the more active the policies should be. But they should stop well short of fine tuning, of trying to achieve constant unemployment or constant output growth.

Until 30 years ago, macroeconomic policy was seen in the same way as the control of a complicated machine. Methods of optimal control, developed initially to control and guide rockets, were being increasingly used to design macroeconomic policy. Economists no longer think this way. It has become clear that the economy is fundamentally different from a machine, even from a very complicated one. Unlike a machine, the economy is composed of people and firms who try to anticipate what policy makers will do, and who react not only to current policy but also to expectations of future policy. Hence, macroeconomic policy must be thought of as a game between the policy makers and “the economy”—more concretely, the people and the firms in the economy. So, when thinking about policy, what we need is not optimal control theory but rather game theory.

A first step is to make the central bank independent. By an independent central bank, we mean a central bank where interest rate and money supply decisions are made independent of the influence of the currently elected politicians. Politicians, who face frequent reelections, are likely to want lower unemployment now, even if it leads to inflation later. Making the central bank independent, and making it difficult for politicians to fire the central banker, makes it easier for the central bank to resist the political pressure to decrease unemployment below the natural rate of unemployment.

22-3 Politics and Policy

By definition, tax cuts lead to lower taxes today. They are also likely to lead to an increase in demand, and therefore to an increase in output for some time. But unless they are matched by equal decreases in government spending, they lead to a larger budget deficit and to the need for an increase in taxes in the future. If voters are shortsighted, the temptation for politicians to cut taxes may prove irresistible. Politics may lead to systematic deficits, at least until the level of government debt has become so high that politicians are scared into action.

Look first at the evolution of the ratio of debt to GDP from 1900 to 1980. Note that each of the three buildups in debt (represented by the shaded areas in the figure) was associated with special circumstances: World War I for the first buildup, the Great Depression for the second, and World War II for the third. These were times of unusually high military spending or unusual declines in output. Adverse circumstances—not pandering to voters were clearly behind the large deficits and the resulting increase in debt during each of these three episodes. Note also how, in each of these three cases, the buildup was followed by a steady decrease in debt. In particular, note how the ratio of debt to GDP, which was as high as 130% in 1946, was steadily reduced to a postwar low of 33% in 1979. The more recent evidence, however, fits the argument of shortsighted voters and pandering politicians much better. Clearly, the large increase since 2007 is due to the crisis. But, leaving it aside, note how the debt-to-GDP ratio increased from 33% in 1980 to 63% in 2007. This increase in debt can be largely traced back to two rounds of tax cuts, the first under the Reagan administration in the early 1980s, the second under the Bush administration in the early 2000s.

Dornbusch, R., S. Fischer and R. Startz Macroeconomics (2011)

CHAPTER 16: The Fed, Money, and Credit

• The Federal Reserve provides the monetary base (bank reserves and currency) upon which the money supply (currency and deposits) is built.
• The primary tool for controlling the money supply is open market purchases, purchases of bonds paid for with newly printed money.
• The Federal Reserve chooses both intermediate and ultimate targets. The key consideration in choosing targets is uncertainty about different kinds of economic shocks.


The money supply consists mostly of deposits at banks, 2 which the Fed does not control directly. The key concept to understand is fractional reserve banking.. . Rather, the coins would form a “base” supporting deposits avail-able through the banking system. The real money supply is determined by a blend of these two fictional systems.

High-powered money (or the monetary base) consists of currency (notes and coins) and banks’ deposits at the Fed. The part of the currency held by the public forms part of the money supply. The currency in bank vaults and the banks’ deposits at the Fed are used as reserves backing individual and business deposits at banks. The money multiplier is the ratio of the stock of money to the stock of high-powered money.

Bank reserves consist of deposits banks hold at the Fed and “vault cash,” notes and coins held by banks. In the absence of regulation, banks would hold reserves to meet (1) the demands of their customers for cash and (2) payments their customers make by checks that are deposited in other banks.


The Federal Reserve has three instruments for controlling the money supply: open market operations, the discount rate, and the required-reserve ratio.

The striking result is that the Fed can create high-powered money at will merely by buying assets, such as government bonds, and paying for them with its own liabilities.

The discount rate is the interest rate charged by the Fed to banks that borrow from it to meet temporary needs for reserves

Federal funds are reserves that some banks have in excess and others need.


When the bank makes a loan, the person receiving the loan gets a bank deposit. At this stage, when the bank makes a loan, the money supply has risen by more than the amount of the open market operation.


We make a simple but important point in this section: The Fed cannot simultaneously set both the interest rate and the stock of money at any given target levels that it may choose.

When the Fed decides to set the interest rate at some given level and keep it fixed—a policy known as pegging the interest rate —it loses control over the money supply.


If output deviates from its equilibrium level mainly because the IS curve shifts about, output is stabilized by keeping the money stock constant. The Fed should, in this case, have a money stock target.

If output deviates from its equilibrium level mainly because the demand-for-money function shifts about, the Fed should operate monetary policy by fixing the interest rate.


Proponents of the central role of credit also argue that credit rationing makes interest rates an unreliable indicator of monetary policy. Credit is rationed when individuals cannot borrow as much as they want at the going interest rate.


Rather, in choosing intermediate targets, the Fed has to trade off between those targets it can control exactly and those targets that are most closely related to its ultimate targets.

CHAPTER 17 Policy

• Uncertainty about the economy places limits on the reach of successful policy.
• Our imperfect knowledge of the economy sometimes argues for a go-slow approach in the application of economic policy.
• Choice of policy targets should be influenced by the limits of our knowledge as well as by the extent of our knowledge.
• Democracies face the difficult problem of structuring policymaking bodies to avoid temptation toward an inflationary bias.

Acknowledging the limits of policy is very different from attempting to avoid policy altogether. A large country does not have the option of not having a macroeconomic policy. The choices of government spending, of taxation, and of the money supply will affect the economy. So in deciding on their budgets and on monetary policy, governments need to consider how best to affect the economy—or at least, how to avoid some common mistakes.


There are delays, or lags, at every stage, and these can be divided into two stages: an inside lag, which is the time period it takes to undertake a policy action— such as a tax cut or an increase in the money supply—and an outside lag, which describes the timing of the effects of the policy action on the economy.

The recognition lag is the period that elapses between the time a disturbance occurs and the time the policymakers recognize that action is required.

The decision lag —the delay between the recognition of the need for action and the policy decision—differs between monetary and fiscal policy. 3 The Federal Reserve System’s Open Market Committee meets frequently to discuss and decide on policy. Thus, once the need for a policy action has been recognized, the decision lag for mon-etary policy is short. Further, the action lag —the lag between the policy decision and its implementation —for monetary policy is also short. For fiscal policy it can be considerable.

The existence of the inside lag in policymaking focuses attention on the use of automatic stabilizers. An automatic stabilizer is any mechanism in the economy that automatically—that is, without case-by-case government intervention—reduces the amount by which output changes in response to a shock to the economy. One of the major benefits of automatic stabilizers is that their inside lag is zero. The most important automatic stabilizer is the income tax. It stabilizes the economy by reducing the multiplier effects of any disturbance to aggregate demand.

The inside lag of policy is a discrete lag —so many months—from recognition to decision and implementation. The outside lag is generally a distributed lag: Once the policy action has been taken, its effects on the economy are spread over time. There may be a small immediate effect of a policy action, but other effects occur later.



Uncertainty about the size of the effects that will result from any particular policy action—whether because of uncertainty about expectations or about the structure of the economy—is known as multiplier uncertainty.

Multipliers measure the quantitative effect of policy. The argument that the less certain we are about the size of a multiplier, the more cautious we should be in application of the associated policy instrument is intuitively plausible.


Economic variables play a variety of roles in policy discussions. It is useful to divide variables into targets , instruments , and indicators .


We see no case for arguing that monetary and fiscal policy should not be used actively in the face of major disturbances to the economy. Fine tuning presents more complicated issues. In the case of fiscal policy, the long inside lags make discretionary fine tuning virtually impossible, though automatic stabilizers are in fact fine tuning all the time. But with monetary policy decisions being made frequently, fine tuning of monetary policy is indeed possible.

The case for fine tuning is a controversial one. The major argument against it is that in practice policymakers do not in fact behave as suggested—making only small adjustments in response to small disturbances. If allowed to do anything, they may do too much.

The major lesson is not that policy is impossible but that overly ambitious policy is risky. The lesson is to proceed with extreme caution, always bearing in mind the possibility that policy itself may be destabilizing.



The case for modest, activist, discretionary policy seems clear. Why then do countries follow procedures sometimes that seem to have a bias toward too much inflation? The answer to these questions is found in an examination of the idea of dynamic inconsistency. Essentially, the argument is that policymakers who have discretion will be tempted to take short-run actions that are inconsistent with the economy’s best long-run interests.

So early choices by the policymaker must anticipate later stages, which themselves depend on the choices made earlier. Decision makers work out their choices by starting at the end and working backward. This choice method is a simple example of dynamic programming.

1. The potential need for stabilizing policy actions arises from economic disturbances.
2. Some of these disturbances, such as changes in money demand, consumption spending, or investment demand, arise from within the private sector. Others, such as wars, may arise for noneconomic reasons.
3. Wise policymakers work with what we know about the economy while also recognizing the limits of our knowledge. Good policy design includes an assessment of the risks associated with unforeseen errors.
4.The three key difficulties of stabilization policy are that ( a ) policy works with lags; ( b ) the outcome of policy depends very much on private sector expectations, which are difficult to predict and may react to policy; and ( c ) there is uncertainty about both the structure of the economy and the shocks that hit the economy.
5. When forming economic policy, policymakers must choose between sudden policy changes and gradual changes. Sudden policy changes may enhance the policymakers’ credibility but are based on limited information. Gradual changes allow policy- makers to incorporate new information as the economy moves toward its target.
6. For the purposes of policy, economic variables can be classified as targets (identified goals of policy), instruments (the tools of policy), and indicators (economic variables that signal whether we are getting close to our policy targets). There are clearly occasions on which active monetary and fiscal policy actions should be taken to stabilize the economy. These are situations in which the economy
has been affected by major disturbances.
7. Fine tuning—continuous attempts to stabilize the economy in the face of small disturbances—is more controversial. If fine tuning is undertaken, it calls for small policy responses in an attempt to moderate the economy’s fluctuations, rather than to remove them entirely. A very active policy in response to small disturbances is likely to destabilize the economy.
8. In the rules-versus-discretion debate, it is important to recognize that activist rules are possible. The two important issues in the debate are how difficult it should be to change policy and whether policy should be announced as far ahead as possible. There is a tradeoff between the certainty about future policy that comes from rules and the flexibility of the policymakers in responding to shocks.
9. Central bank independence is one avenue democracies use to add to the credibility of policy and to help mitigate the problem of dynamic inconsistency.

Mankiw, N. G. Macroeconomics. (Worth, 2009)

If you are using Makiw, you should read all the material in Chapter 18, including the appendix on the time-inconsistency problem.

Chapter 19: Money Supply, Money Demand, and the Banking System

19-1 Money Supply

We begin by recalling that the money supply includes both currency in the hands of the public and deposits at banks that households can use on demand for transactions, such as checking account deposits. That is, letting M denote the money supply, C currency, and D demand deposits, we can write
Money Supply = Currency + Demand Deposits

M = C + D.

If banks hold 100 percent of deposits in reserve, the banking system does not affect the supply of money.

Now imagine that banks start to use some of their deposits to make loans - when they do so, we have fractional-reserve banking, a system under which banks keep only a fraction of their deposits in reserve. Thus, in a system of fractional-reserve banking, banks create money.

The process of transferring funds from savers to borrowers is called financial intermediation.

The monetary base B is the total number of dollars held by the public as currency C and by the banks as reserves R. It is directly controlled by
the Federal Reserve.

The reserve–deposit ratio rr is the fraction of deposits that banks hold in reserve. It is determined by the business policies of banks and the laws
regulating banks.

The currency–deposit ratio cr is the amount of currency C people hold as a fraction of their holdings of demand deposits D. It reflects the
preferences of households about the form of money they wish to hold.

We begin with the definitions of the money supply and the monetary base:
M = C + D,
B = C + R.

The first equation states that the money supply is the sum of currency and demand deposits. The second equation states that the monetary base is the sum of currency and bank reserves. To solve for the money supply as a function of the three exogenous variables (B, rr, and cr), we first divide the first equation by the second to obtain

M/B = (C+D) / (C+R)

19-2 Money Demand

In previous chapters, we used simple money demand functions. We started with the quantity theory, which assumes that the demand for real balances is proportional to income. That is, the quantity theory assumes

(M/P)^d = kY,

We then considered a more general and realistic money demand function that assumes the demand for real money balances depends on
both the interest rate and income:

(M/P)^d = L(i, Y ).

Recall that money serves three functions: it is a unit of account, a store of value, and a medium of exchange. The first function—money as a unit of account—does not by itself generate any demand for money, because one can quote prices in dollars without holding any. By contrast, money can serve its other two functions only if people hold it. Theories of money demand emphaize the role of money either as a store of value or as a medium of exchange.

1. The system of fractional-reserve banking creates money, because each dollar of reserves generates many dollars of demand deposits.
2. The supply of money depends on the monetary base, the reserve–deposit ratio, and the currency–deposit ratio. An increase in the monetary base
leads to a proportionate increase in the money supply. A decrease in the reserve–deposit ratio or in the currency–deposit ratio increases the money
multiplier and thus the money supply.
3. The Federal Reserve changes the money supply using three policy instruments. It can increase the monetary base by making an open-market
purchase of bonds or by lowering the discount rate. It can reduce the reserve–deposit ratio by relaxing reserve requirements.
4. To start a bank, the owners must contribute some of their own financial resources, which become the bank’s capital. Because banks are highly lever-
aged, however, a small decline in the value of their assets can potentially have a major impact on the value of bank capital. Bank regulators require
that banks hold sufficient capital to ensure that depositors can be repaid.
5. Portfolio theories of money demand stress the role of money as a store of value. They predict that the demand for money depends on the risk and
return on money and alternative assets.
6. Transactions theories of money demand, such as the Baumol–Tobin model, stress the role of money as a medium of exchange. They predict that the
demand for money depends positively on expenditure and negatively on the interest rate.
7. Financial innovation has led to the creation of assets with many of the attributes of money. These near moneys make the demand for money less
stable, which complicates the conduct of monetary policy.