SUMMARY
• Hyperinflations are periods of high inflation. The most extreme episodes took place after World Wars I and II in F.urope. But Latin America has had episodes of high inflation as recently as the early 1990s.
• High inflation comes from high nominal money growth. High nominal money growth comes from the combination ol large budget delicits and the inability to finance these budget delicits through borrowing, either from the public or Irom abroad.
• I he revenues from money creation are called 'seignorage'. Seignorage is equal to the product ol nominal money growth and real money balances. The smaller real money balances are. the higher the required rate ol nominal money growth, and therefore the higher the rate ol inflation required to generate a given amount of seignorage.
• Hyperinflations arc typically characterised by increasing inllation. There are two reasons lor this. One is that higher nominal money growth leads to higher inflation, inducing people to reduce real money balances, requiring even higher nominal money growth 'and thus leading to even higher inflation) to linance the same real delicit. The other reason is that higher inflation often increases the delicit, which requires higher nominal money growth and even higher inllation.
• Hyperinflations arc ended through stabilisation programs. To be succcsslul, stabilisation programs must include fiscal measures aimed at reducing the dclicit and monetary measures aimed at reducing or eliminating money creation as a source of financing tor the deficit. Some stabilisation plans also include wage and pricc guidelines or controls.
• A stabilisation program that imposes wage and pricc controls without changes in fiscal policy and monetary policy will tail. But even coherent and well-conceived programs don't always succeed. Anticipation of lailure may lead to the failure ol even a coherent plan.
KEYTERMS
hyperinflation, 554 • inflation tax, 561
debt monetisation, 556 • Tanzi-Olivera effect, 563
seignorage, 556 • stabilisation program, 563
barter, 558 • incomes policies, 564
dollarisation, 558 • orthodox stabilisation program, heterodox
1.after curve, 560 stabilisation program, 564
QUESTIONS AND PROBLEMS
Quick check
I. Using the information in this chaptcr, label each of the following statements 'true', 'false' or 'uncertain'. Explain briefly.
a. In thc short mn, governments can finance a deficit of any size through money growth, h. I he inflation tax is always equal to seignorage.
c. Hyperinflations may distort priccs but have no effect on real output.
d The solution to ending hyperinflations is simple: institute a wage and price freeze and inflation will stop.
e. As inflation is generally good lor those who borrow money, hyperinflations are the best times in which to lake out large loans.
I. Budget delicils usually shrink during hyperinflations.
2. Assume that money demand takes the following form:
M
— = Y[1 - (r + 7/)]
where Y = 1,000 and r = 0.1.
a. Assume that, in the short run тг ' is constant and equal to 25 per cent. Calculate the amount of seignorage il thc rate ol money growth, AM/M, equals:
1. 25 per cent
2. 50 per cent
3. 75 per cent
b. In the medium run. тг'' = тг \M/M. Calculate thc amount ol seignorage associated with the three rates ol money growth in question (a). F.xplain why the answers differ trom those in (a).
c. In the medium run, whai is ihe rate of money growth that maximises seignorage?
3. How would each of the following change the Tanzi-Olivera effect?
a. Requiring monthly instead ol yearly tax payments by households.
b. Assessing greater penalties for under-withholding ol taxes from monthly pay cheques.
c. Decreasing the income tax and increasing the goods and services tax.
Dig deeper
4. You are the economic adviser to a country suffering from a hyperinflation. Politicians debating the proper course of stabilisation have advocated various positions, listed in statements lai to fe). Discuss each statement in turn.
a. 'This crisis will not end until workers begin to pay their lair share ol taxes.
b. The central bank has demonstrated that it cannot responsibly wield its power ю create money, so we have no choice but to adopt a currency board.
c. I'ricc controls arc necessary to end this madness.
d. Stabilisation will be successful only il there is a large recession and il there is a substantial increase in unemployment.'
e. 'Let's not blame the central bank. Thc problem is liscal policy, not monetary policy.'
Explore further
5. High inflation around the world
Inflation rates are quite low in most of the advanced economies. However, you shouldn't think that high inflation is only a historical topic. Today, n number of countries are experiencing double-digit (or even triple-digit) inflation rates.
a. Co to the website ol the IMF (www.imf.org) and find ihe currcnt issue of the World Economic Outlook. In the Statistical Appendix, look at the table that lists inflation rates. Find countries that
have inflation rates of 10 percent or higher. Which country has thc highest inllation rate, and what is the rate?
b. Find Venezuela in the inflation table. How long has Venezuela had an inflation rate ol more than 10 per ccnt? Look at the projected inflation rates tor the current year and thc next year. Does inflation show any signs of slowing in Venezuela?
c. Venezuela is an oil producer, so its economy fluctuates with oil prices. Government lax revenues, in particular, depend heavily on the prosperity of thc oil industry. Willi oil prices rising, Venezuela has increased government spending dramatically in recent years. Suppose oil prices fall in thc future, hut Venezuela does not reduce government spending. How would a fall in oil prices affect ihe budget deficit in Venezuela? Given the effect on the budget deficit, and lollowing the logic of this chaptcr, how would a tall in oil prices make a hyperinflation possible in Venezuela?
We invite you to visit the Blanchard-Sheen page on the Pearson Australia website at
www.pearson.com.au/highered/blanchardsheen3e
for many World Wide Web exercises relating to issues similar to those in this chapter.
FURTHER READINGS
Back to Policy
Nearly every chapter of this book has looked at the role of policy. The next three chapters put it all together.
CHAPTER 25
Chapter 25 asks two questions: Given the uncertainty about the effects of macroeconomic policies, wouldn't it be better not to use policy at all? And, even if policy can in principle be useful, can we trust poiicy-makers to carry out the right policy? The bottom line is: Uncertainty limits the role of policy; policy-makers don't always do the right thing; but, with the right institutions, policy can help and should be used.
CHAPTER 26
Chapter 26 looks at monetary policy. It reviews what we have learned, chapter by chapter, and then focuses on two issues. The first is the optimal rate of inflation. High inflation is bad, but how low a rate of inflation should the central bank aim for? The second is the design of policy. Should the central bank target money growth or should it target inflation? What rule should the central bank use to adjust the interest rate?
CHAPTER 27
I
Chapter 27 looks at fiscal policy. It reviews what we have learned, and then looks more closely at the mechanics of debt, taxes and spending implied by the government budget constraint. It then considers several issues, from how wars should be financed to the dangers of accumulating too high a level of debt. It ends with a description of the current budget situation in Australia, and a discussion of the problems on the horizon.
CHAPTER ф
Should Policy-makers Be Restrained?
A
t many points in this book you saw how the right mix of fiscal policy and monetary policy could help a country out of a recession, improve its trade position without increasing activity and igniting inflation, slow down an overheating economy, stimulate investment and capital accumulation, and so on.
This conclusion, however, appears at odds with growing demands that policy-makers be tightly restrained. In the European Union, countries that have adopted the euro are required to keep their budget deficit under 3 per cent of GDP. In the United States, the first item in the 'Contract with America', the program drawn up by Republicans for the mid-term US elections in 1994, was the introduction of a balanced-budget amendment to the Constitution. With the elimination of the US budget deficit in the second half of the 1990s, the push for a balanced-budget amendment has weakened: but if deficits were to come back, the issue would surely return to centre stage. In Australia, the Howard-Costello government made fiscal surpluses into a political virtue. This was easy to promote because the growth of the Australian economy was remarkably good from 1996 to 2007. However, if Australia suffers a serious downturn in 2008-09 as expected, it may be politically difficult to counter it with necessary large fiscal deficits.
Monetary policy is also under fire. For example, the charter of the central bank of New Zealand, written in 1989, defines monetary policy's role as the maintenance of price stability, to the exclusion of any other macroeconomic goal. In Australia, the RBA has an agreement with the government that, in pursuit of its general objectives (stability of the currency, maintenance of full employment and economic prosperity and welfare), it will have only one quantified target, inflation between 2 and 3 per cent. This chapter looks at the case for such restraints on macroeconomic policy.
• Sections 25.1 and 25.2 look at one line of argument—namely, that policy-makers may have good intentions but they end up doing more harm than good.
• Section 25.3 looks at another, more cynical, line—that policy-makers do what is best for them, which isn't necessarily what is best for the country.
25.1 UNCERTAINTY AND POLICY
A blunt way ol slating the first argument in favour of policy restraints is that those who know little should do little. The argument has two parts: macroeconomists, and by implication the policy-makers who rely on their advice, know little,- and they should therefore do little. Let's look at each part separately.
How much do macroeconomists actually know?
Macroeconomists are like doctors treating cancer. They know a lot. but there is also a lot they don'l know.
Take an economy thai suffers irom high unemployment, and where thc central bank is considering the use ol monetary policy to increase economic activity. Think of the sequence of links between an increase in money and cn
increase in output—all the questions the ccntral bank laccs when deciding whether and by how much to decrease the interest rate:
• Is the current high rate of unemployment above the natural rate ol unemployment, or has the natural rate of unemployment itscll increased (Chapters 8 and 9)?
• It the unemployment rate is close to the natural rate of unemployment, isn't there a risk that a monetary expansion will lead to a decrease in unemployment below thc natural rate of unemploy¬ment and an increase in inflation Chapters 8 and 9)?
• What will be the ellect ol the decrease in the short-term interest rate on the long-term interest rate (Chapter t5)? By how much will stock prices increase (Chapter 15)? By how much will the currency depreciate (Chapters 20 and 21)?
• I low long will it take lor lower long-term interest rates and higher stock prices to at feet investment and consumption spending (Chapter 16)? How long will it take lor thc l-curvc effects to work themselves out and lor the trade balance to improve Chapter 19 -: What is the danger that the eflects come too late, when thc economy has already recovered?
When assessing these questions, central banks—or macroeconomic policy-makers in general—don't operate in a vacuum. Ihey rely in particular on macroeconometric models. The equations in these models give estimates ol how these individual links have looked in the past But dillerent models give different answers. This is because they have dillerent structures, different lists of equations and different lists of variables.
In the Australian context there arc iust a few operational macroeconometric models. Thc Reserve Bank of Australia (RBA i built a complex model in the late 1970s, which it used lor the next ten years to help it make forecasts. In thc 1990s it abandoned large-scale models, and began to use a variety of smaller models as part of the process. In common with many other central banks, it has developed in recent years its capacity to build dynamic stochastic general equilibrium I )SC.f:i models have a look at Chaptcr 27 for more on these). Thc Commonwealth Treasury lirst developed thc NIP model in the early 1970s, and substantially upgraded that with theTRYM model in thc early 1990s. There are other models in commercial use, such as the Access ACM model and the Murphy MM model. These Australian models have many differences, and their forecasts vary accordingly. In every country, this diversity of models and lorecasts is a fact of life.
figure 25.1 shows an example ol this diversity. Thc example comes Irom an American study commissioned in the late 1980s by the Brookings Institution a research institute in Washington DC—asking the builders of the twelve main macroeconometric models to answer a similar set ol questions. (The models arc described in the focus box Twelve macroeconometric models'.» The goal was to see how the answers would diller across models. One question was;
Consider a ease where thc US economy is growing at its normal growth rate and where unemployment is at its natural rate ot unemployment,- call this the baseline case. Suppose now that over the period of a year the l ed eases monetary policy compared with the baseline, so that, alter a year, nominal money is 4 per ccnt higher than it would have been in the baseline case. I rom then on nominal money grows at the same rate as in the baseline case, so the level of nominal money remains 4 percent higher than it would have been without the change in monetary policy Suppose further that interest rates in the rest of the world remain unchanged. What will happen to US output?
A description of the models and of the study is given in Ralph Bryanc et at.. Empirical * Macroeconomics for Interdependent Economies (Washington. DC: Brookings Institution. 1988). The study shows the effects not only of monetary policy, but also of fiscal policy (The simulation described in the text is simulation E in the supplemental volume.)
SHOULD POLICYMAKERS Br RfcSTRAINFD;
chapter 25
Figure 25.1 shows the deviation ol output from the baseline predicted by each ol the twelve models. All twelve models predict that output will increase for some time after the increase in money. Alter one year, thc average deviation of output from the baseline is positive. But the range ol answers is large. Irom nearly no change to close to an increase of 3 per ccnt,- even leaving out thc most extreme prediction, the range is still more than I per cent. Two years out. the average deviation is 1.2 per cent,- again leaving out the most extreme prediction, the range is still 2 per cent. And six years out the average deviation is 0.6 per cent, and the answers range from —0.3 per ccnt to 2.5 per cent. In short, if we measure uncertainty by the range of answers from this set of models, there is substantial uncertainty about the effects of policy.
Figure 25.1 The response of US output to a monetary expansion: predictions from twelve models
-0.5
-01 H 1 1 1 1 1 1 1
0 12 3 4 5 6 7
Years
MSG
VAR
- DRI
......UNK
COMPACT
MCM
WHARTON
- MINIMOD
INTERLINK
TAYLOR LIVERPOOL EPA
While all twelve models predict that US output will increase for some time in response to a monetary expansion, the range of answers regarding the size and the length of the output response is large.
Should uncertainty lead policy-makers to do less?
Should uncertainty about the effects of policy lead policy-makers to do less? In general, the answer is yes. Consider the following example, loosely based on the situation in Australia in September 2008.
Suppose that the growth ol the Australian economy is just below normal and the unemployment rate is 4.3 per cent. The RBA is considering using monetary policy to expand output, and therefore il is not particularly concerned right now about the implications for inflation. To concentrate on uncertainty about the effects of policy on output, let's assume that the RBA knows everything else for sure. Based on its forecasts about thc world economy, it knows that, in the absence of changes in monetary policy, unemployment will go up to 5 per cent next year. It knows that the natural rate of unemployment is 5,5 per cent, and therelore the unemployment rale is below the natural rate. And it knows, from Okun's law, that I per cent more output growth for a year leads to a reduction in thc unemployment rate ot 0.5 percent.
Under these assumptions, the RBA knows that, if it could use monetary policy to achieve 1.4 per cent more output growth over the coming year, this would lower thc unemployment rate a year from now by 0.5 x 1 4% = 0.7%, cancelling out the effects on unemployment ol the sluggish world economy. By how much should the RBA lower the interest rate, which is currently at, say, 7 per cent?
What it could do is take all available Australian macroeconometric models and work out thc average ellect on output growth in a year's time of a I per cent cut in the short-run interest rate. Let's say the average effect is 1.4 percent.
Suppose the RBA takes this average relation as holding with certainty. What it should then do is straightforward. To keep the unemployment rate Irom rising to 5 per cent in one year requires 1.4 per cent more output growth. And 1.4 per cent output growth requires the RBA to decrease the interest rate by I per cent. The RBA should therefore decrease the cash rate from 7 per cent to 6 per cent today. If the economy's response is equal to the average response from the Australian models, this decrease in thc cash rate will maintain the unemployment rate at the end of the year at 4.3 percent, despite the sluggish world economy.
In the real world, of ► course, the RBA doesn't know any of these things with certainty. It can only make forecasts. It doesn't know the exact value of the natural rate of unemployment, or the exact coefficient n Okun's law. Introducing these sources of uncertainty wou'd reinforce our basic conclusion.
Suppose that the RBA actually decreases the interest rate by I per cent. But let's now take into account uncertainty, as measured by the range of responses of the dilferent models. Let's say that the
TWELVE MACROECONOMETRIC MODELS
Together, the set of models used in the Brookings project is representative of the different types of
macroeconomic models used for forecasting and policy in the world today:
• Two models, DRI (Data Resources Incorporated) and WHARTON, are commercial models.They are used regularly to generate and sell economic forecasts to firms and financial institutions.
• Five are used for forecasting and help in the design of policy. MCM (MultiCountry Model) is used by the Federal Reserve Board in Washington for the conduct of monetary policy; INTERLINK is used by the OECD in Paris: COMPACT is used by the Commission of the European Union in Brussels: EPA is used by the Japanese Plann ng Agency. Each of these four models was constructed by one team of researchers doing all the work—that is, building sub-models for countries or groups of countries and linking them through trade and financial flows. In contrast, the fifth. LINK, is composed of individual country models— models constructed in each country by researchers from that country and then linked together by trade and financial relations. The advantage of this approach is that researchers from a particular country are likely to understand that country very well; the disadvantage is that different country models may have quite different structures, and be hard to link to each other.
• Four models incorporate rational expectations explicitly: the LIVERPOOL model, based in England; MINIMOD, used at the International Monetary Fund; MSG, developed by Warwick McKibbin of the Australian National University and Jeffrey Sachs at Harvard University; and the TAYLOR model—which we saw in Section 7A—developed by John Taylor of Stanford University. Because it is technically difficult to solve for large models under rational expectations, these models are typically smaller models, with less detail than those listed above. But they are better at capturing the expectation effects of various policies. Thanks to more and more powerful computers, researchers are building larger and larger models with rational expectations. The modern versions of these models are called dynamic stochastic general equilibrium (DSGE) models, and are the subject of active research.
• The last one, VAR (for Vector Autoregression, the technique of estimation used to build the model), developed by Christopher Sims and Robert Litterman at Minnesota, is very different from the others.VAR models are not structural models; rather, they are statistical summaries of the relations between the different variables, without an explicit economic interpretation. Their strength is in their fit of the data, with a minimum of restrictions.Their weakness is that they are, essentially, a (very big) black box.
Note: A description of thc models and of the study is given in Ralph Bryant et ol. Empirical Macroeconomics for Interdependent Economies.
Brookings Institution, Washington, DC. 1988. The study shows the effects not only of monetary policy but also of fiscal policy.
range ol responses of output to a I per cent decrease in thc interest rate after one year varies Irom 0 per cent to .3.6 per cent. These output numbers imply in turn a decrease in unemployment from the monetary policy anywhere between 0 per cent and 0.5 x 3.6: = 1.8 per cent, or values ol the unemploy¬ment rate a year hence anywhere between 5 per cent and 3.2 per cent!
This range ol uncertainty is an issue that should constrain the central bank. But there is more. Wc assumed that the world economy was sure to deliver an increase ol unemployment by 0.7 per cent if the central bank did nothing. But this was just a forecast and no one knows for sure what will happen to the global economy and what impact it will have domestically—lets say. the range ol forecasts ol the effects on unemployment increase is between -0.3 per cent and 1.7 per cent. Adding this range of uncertainty to what we showed before suggests a range of unemployment between 6 per cent and 2.2 per cent:
FOCUS 'BOX
SHOiJI I.) POl ICY-MAKERS Rf RrSTRAINri>:' chaptc- 25
The conclusion is clear: given the range of uncertainty about the effects of monetary policy on output, decreasing the interest rate by 1 per cent would be inappropriate. If the ellects of the interest rate on output are as strong as suggested by one ol the models, and the global economy actually improves at the most optimistic forecast, unemployment by the end of the year could be 3.3 per cent below the natural rate ot unemployment, leading to very undesirable inflationary pressures relative to thc inflation target. Civen this uncertainty, thc RBA should decrease the interest rale by much less than
I per cent, l or example, decreasing the interest rale by 0.25 per ccnt leads to a range lor unemployment a year hence ol 6 per cent to 3.55 per cent, clearly a safer range of outcomes.
Thc RBA rarely changes the interest rate by more than 0.25 percent. I lowever, it did cut thc interest rate by I per cent in October 2008, citing its grave fears about the impact of the global financial crisis on output and unemployment in Australia.
Uncertainty and restraints on policy-makers
Lets summarise. There is substantial uncertainty about thc effects of macroeconomic policies. This uncertainty should lead policy-makers to be more cautious, to use less active policies. Policies should be broadly aimed at avoiding prolonged recessions, slowing down booms and avoiding inflationary pressure. The higher unemployment or inflation is, the more active thc policies should be. But they should stop well short ol fine-tuning, of trying to achieve constant unemployment, constant output growth or a constant rate of inflation.
These conclusions would have been controversial twenty years ago. Back then, there was a heated debate between two groups of economists. One group headed by Milton Friedman from Chicago, argued that because ol long and variable lags, activist policy is likely to do more harm than good. The other group headed by Franco Modigliani from the Massachusetts Institute ol Technology, had just built the first generation of large macroeconometric models and believed lhat the economists' knowledge was becoming good enough to allow lor increasing fine-tuning ol the economy. Today, most economists recognise that there is substantial uncertainty about the effects of policy. They also accept the implication thai ihis uncertainty should lead to less active policies.
Note that what we have developed so tar is an argument tor self-restraint by policy-makers, not tor restraints on policy-makers. II policy-makers understand the implications ol uncertainly—and there is no reason to think they don't—they will, on their own, follow less activc policies. There is no reason to impose lurther restraints, such as a requirement that they follow an interest rate rule rigorously, or thai money growth be constant or that thc budget be balanced. Let's now turn to arguments for restraints on policy-makers.
25.2 EXPECTATIONS AND POLICY
One reason the elfects ol macroeconomic policy are uncertain is the interaction of policy and expecta¬tions. I low a policy works, and sometimes whether il works at all, depends not only on how it allects current variables but also on how it affccts expectations about the future ! thc main theme of Chapter I7>. The importance ol expectations for policy goes however, beyond uncertainty about the effects
Warning: When economists say game they dont mean entertainment, ihey mean strategic inter¬actions between players In the context of macroeconomic policy, the players are the policy-makers and the economy people and firms . The strategic interactions arc dear: what people and firms do depends on what they expect policy-makers to do. In turn, what policy-makers do depends on whai is happening in ihe economy.
This example relies ► on the notion of multiplicative uncertainty—that because the effects of policy are uncertain, more active policies lead to more uncertainty. See William Brainard, 'Uncertainty and the effectiveness of policy', American Economic Review. May 1967. pp.411-25.
Friedman and Modigliani ► are the same two economists who independently developed the modern theory of consumption we saw in Chapter 16.
Even machines are becoming smarter: HAL (the robot in the 1968 movie 2001:A Space Odyssey) starts anticipating what humans in the spaceship will do.The result isn't a happy one. (See the movie.) >
Game theory is becoming an important tool in all branches of economics. Thc 1994 Nobel Prize in economics was awarded to three game theorists.
John Nash from Princeton. John Harsanyi from Berkeley (who did a Masters degree at the University of Sydney in die 1950s) and Reinha-d * Selten from Germany.
Game theory has given economists many insights, often explaining how some apparently strange behaviour makes sense when one understands thc nature of the game being played. One ol these
insights is particularly important for our discussion of restraints here. Sometimes you can do better in a game by giving up some of your options. To see why, lets start with an example Irom outside economics—governments' policies towards hijackers.
Hijackings and negotiations
Most governments have a stated policy that they won't negotiate with plane hiiackers. The reason lor this policy is clear: to deter hijacking by making it unattractive to hijack planes.
Suppose that, despite the staled policy, a hijacking takes place. Now that the hijacking has taken place anyway, why not negotiate? Whatever compensation the hijackers demand is likely to be less costly than the alternative—the likelihood thai lives will be lost il the plane has to be taken by force. So the best policy would appear to be: announce that you won't negotiate, but il a hijacking happens, negotiate.
()n reflection, it is clear thai this would in lact be a very bad policy. Hijackers' decisions don't depend on the stated policy hut on what they expect will actually happen il they hijack a plane. Il they know that negotiations will actually lake place, they will rightly consider the stated policy as irrelevant. And hijackings will take place.
So. what is the best policy; Despite the lact that once hijackings have taken place negotiations typically lead to a better outcome, the best policy is lor governments to commit not lo negotiate. By giving up the option to negotiate, they are likely to prevent hijackings in thc first place.
Let's now turn to a macroeconomic example, based on the relation between inflation and unemploy mem. As you will see, exactly the same logic is involved.
Inflation and unemployment revisited
77
« U
Recall the relation between inflation and unemployment derived in Chapter 8 i equation 8.91, with the time indexes omitted for simplicity):
• 25.1)
Inflation, тг. depends on expected inflation, 77', as embodied in wages sei in labour contracts, and on thc difference between the actual unemployment rale, v., and the natural rate ol unemployment. /<„. The coefficient tr captures the ellect ol unemployment on inflation, given expected inflation. When unemployment is above the natural rate, inllation is lower than expected- when unemployment is below the natural rate, inflation is higher than expected.
(25.2)
Suppose the RBA announces it will lollow a monetary policy consistent with zero inllation. On the assumption thai wage setters believe the announcement, expected inllation, 77' as embodied in wage contracts is equal to zero, and the RBA faces the following relation between inllation and unemployment:
77 - -at и - u„)
Il the RBA follows through on its announced policy ol zero inllation, expected inflation and actual inflation will both be equal to zero, and unemployment will be equal lo the natural rate ol unemployment.
Zero inflation and unemployment equal ю the natural rate ol unemployment is not a bad outcome. Bui it would seem thai lhe RBA can actually do even better:
• Recall from Chapter 8 that, in Australia, a is roughly equal to (1.3. So, equation 25.21 implies that, by accepting just I percent inflation, the RBA can achieve this year an unemployment rate ol 3 per cent below the natural rate ol unemployment.
This example was developed by Finn Kydland. from Carnegie Mellon, and Edward Prescott. from the University of Minnesota, in 'Rules rather than discretion: The inconsistency of optimal plans' .Journal of Political Economy, vol. 85, no. 3,
«June 1977.pp.473-92
A refresher: Given labour market conditions, and given their expectations of what prices will be. firms and workers set nominal wages. Given the nominal wages firms have to pay. the firms then set prices. So. prices depend on expected prices and labour market conditions. Equivalently, price inflation depends on expected price inflation and labour market conditions This is what is captured in
< equation (25.1).
For simplicity, we assume that the RBA can and does choose the rate of inflation exactly. In doing so. we ignore uncertainty about the effects of policy (the topic of Section 25.1.
< but not central here).
Remember that the natural rate of unemployment is neither natural nor best in any sense (see Chapters 6 and 8). It may be reasonable for the RBA
^ and everyone else in the economy to prefer an unemployment rate lower than the natural rate of unemployment.
SI ЮШ D POLICY M/WRS BE RESTRAINED?
ch.ipte * 25
Suppose that the RBA—and everybody else in ihe economy—finds the trade-off attractive, and decides to decrease unemployment by 3 per ceni in exchange lor an inflation rate ol I per cent. This incentive to deviate Irom the announced policy once the other player has made his move—in this
case, once wage setters have set thc wage—is known in game theory as the time inconsistency of optimal policy. In our example, thc RBA can improve the outcome this period by deviating Irom its announced policy of zero inflation. By accepting some inflation, it can achieve a substantial reduction in unemployment.
• Unfortunately, this isn't thc end ot the story. II thc RBA has decreased the interest rate by more than it announced it would, wage setters arc likely to wise up and begin to expect positive inflation ol I per ccnt. II the RBA still wants to achieve an unemployment rate 3 per cent below thc natural rate, it will have to achieve 2 per cent inflation. However, if it docs achieve 2 per cent inflation, wage setters arc likely lo increase their expectations of inflation further, and so on.
• The eventual outcome is likely ю be high inflation. Bccausc wage sellers understand thc RB.Vs motives, expected inflation catches up with actual inflation, and the RBA must eventually be unsuccessful in its attempt to achieve unemployment below the natural rale ot unemployment. In short, attempts by the RBA lo make things better lead in the end to things being worse. Thc economy ends up with the same unemployment rale as would have prevailed ir the RBA had followed its announced policy, but with much higher inflation.
How relevant is this example? Very relevant. Reread Chapter 8. Wc can read the history of the Phillips curve and thc increase in inflation in thc 1970s as coming precisely from central banks' attempts to maintain unemployment below the natural rate of unemployment, leading to higher and higher cxpectcd inflation, and higher and higher actual inflation. In that light, the shitt of the original Phillips curve can be seen as thc adjustment of wage setters' expectations to the behaviour of the central bank So, what is the best policy for thc RBA to follow in this case? It is to make a credible commitment lhal it won't iry to decrease unemployment below the natural rale. By giving up the option of deviating Irom its announced policy, the RBA can achieve unemployment equal to thc natural rate ol unemploy¬ment and zero inflation. The hijacking analogy is clear. By credibly committing not to do something thai would appear desirable at the lime, policy-makers can achieve a better outcome: no hijackings in our earlier example, no inllation here.
Establishing credibility
How can a ccntral bank credibly commit not to deviate from its announced policy?
One way to establish its credibility is for the central bank to give up—or ю he stripped by law of— its policy-making power. For example, the mandate ol the bank can be defined by law in terms ol a simple rule, such as setting money growth ai 0 per cent forever. Or, it may involve requiring thc ccntral bank to follow an explicit intcrcst-ratc-setting rule, which wc analysed in Chapters 7, 9 and 21. An alternative, which we discussed in Chapter 2 I, is to adopt a hard peg, such as a currency board or even dollarisation. In that case, the ccntral bank gives up its ability to use thc exchange rate and the interest rate.)
Such a law surely takes care ol the problem ol time inconsistency. But such a tight restraint comes close to throwing thc baby out with the bathwater. We want to prevent the central bank from pursuing too low an interest rate (or loo high a rate ol money growth) in an attempt to lower unemployment below thc natural unemployment rate. But—subject to the restrictions discussed in Section 25.1—we still want the central bank lo be able lo lower the interest rale and raise the money supply when unemployment is far above the natural rate, and raise the interest rate and lower thc money supply when unemployment is tar below the natural rate. Such actions become impossible under a constant money growth rule, or when inllation and unemployment are very high at the same lime under an interest rate rule. There arc, indeed better ways to deal with time inconsistency. In the case ot monetary policy, our discussion suggests a way this can be done:
1. First, make thc ccntral bank independent. Appointing ccntral bankers for longer terms and making it harder to fire them will make them more likely to resist political pressure to decrease unemploy¬ment below thc natural rate ot unemployment.
2. Then, choose a conservative' central banker, somebody who dislikes inllation and is unwilling to accept more inflation in exchange lor less unemployment when unemployment is at the natural rate.
When the economy is at thc natural rate, such a ccntral hanker won't be tempted to embark on a
monetary expansion. Thus, the problem ol time inconsistency will disappear altogether.
Appointing as thc head of the central bank somebody who doesn't have thc same pretcrcnccs as the people as a whole might seem like a solution that only game theorists would concoct. But this is actually the way many countries have been responding to thc problem of time consistency in monetary policy. In many countries in the last two decades, central banks have been given more independence. Figure 25.2 suggests that this approach has been successful. The vertical axis gives the average annual inflation rate in eighteen OECD countries for the period 1960-90. Thc horizontal axis gives the value of an index of 'central bank independence', constructed by looking at several legal provisions in the bank's chartcr—for example, whether and how the government can remove the head of the bank. There is a striking inverse relation between thc two variables, as summarised by the regression line. More central bank independence appears to be systematically associated with lower inflation.
Governments typically have appointed central bankers who arc more 'conservative than the governments themselves—ccntral bankers who appear to carc more about inflation and less about unemployment than the government. 'Sec thc focus box 'Was Alan Blinder wrong in speaking thc truth?'.) This is what lies behind the growing number of central banks around the world adopting an explicit inflation targeting regime. By declaring a target and being seen to systematically act to achieve it, they assert their independence, gain credibility and reduce thc negative implications of time inconsistency.
Time consistency and restraints on policy-makers
Let's summarise what we have learned in this section.
We have examined arguments for putting restraints on policy-makcrs, based on thc issue of time inconsistency.
4 It is reasonable to argue that this doesn't prove that central bank independence leads to lower inflation. It may be that countries that dislike inflation tend to give more independence to their central bankers and have lower inflation, i Another example of the difference between correlation and causality is discussed in Appendix 3 at the end of the book.)
SHOULD POUCY-MAKLRS Bfc RtSIRAINU!
chapter 25
We have looked at the case of monetary policy. But similar issues arise in the context ol fiscal policy. In Chapter 26 we will discuss, for example, the issue of debt repudiation—thc option lor the
I Spain
France Ireland
United" Austria
Figure 25.2 Inflation and central bank independence
18 -.
I
16 - 14 - 12- 10- 8 6 - 4 - 2 -
Portugal
Greece
New Zealand .
Italy
United States
Australia
■
Belgium Netherlands
e
5
Kingdom Austra|lar-^^ Canada
Japan
Germany
Switzerland
Across OECD countries, the higher the degree of ccntral bank independence, the lower the rate of inflation.
SOURCE: Vittorio Grilli. Donate Mastiandaro and Guide Tabcllini. Political and monetary institutions and public financial policies in the industrial countries'. Economic Policy. October 1991. pp. 341 -92.
WAS ALAN BLINDER WRONG IN SPEAKING THE TRUTH?
In the summer of 1994. US President Bill Clinton appointed Alan Blinder, an economist from Princeton, vice- chairman (in effect, second-in-command) of the Federal Reserve Board. A few weeks later Blinder, speaking at an economic conference, indicated his belief that the Fed had both the responsibility and the ability, when unemployment was high, to use monetary policy to help the economy recover. This statement was badly received. Bond prices decreased, and most newspapers ran editorials critical of Blinder.
Why was the reaction of markets and newspapers so negative? It was surely not that Blinder was wrong. There is no doubt that monetary policy can and should help the economy out of a recession. Indeed, the Federal Reserve Bank Act of 1978 requires the Fed to pursue full employment as well as low inflation.
The reaction was negative because, in terms of the argument we developed in the text. Blinder revealed by his words that he wasn't a conservative central banker, that he cared about unemployment as well as inflation. With the unemployment rate at the time equal to 6.1 per cent, close to what was thought to be the natural ra:e of unemployment at the time, markets interpreted Blinders statements as suggesting that he might want to decrease unemployment below the natural rate. Interest rates increased because of higher expected inflation—bond prices decreased.
The moral of the story is that whatever views central bankers might hold they should try to look and sound conservative. This is why many heads of central banks are reluctant to admit, at least in public, the existence of any trade-off between unemployment and inflation, even in the short run.
government to cancel its debt obligations—and see lhat the conclusions are very similar to the case ol monetary policy.
The Reagan tax cuts both decreased tax rates and increased activity in the early 1980s (see the focus box in Chapter 20). But the cuts also led to a long sequence of deficits, which took nearly two decades to eliminate.We will look at the relation between current and future taxes
more closely when we ' examine the implications of the government budget constraint in Chapter 27.
When issues ol lime inconsistency are relevant, light restraints on policy-makers—such as a fixed money growth rule in the case of monetary policy—can provide a rough solution. Bui thc solution may have large costs il ii prevents the use of macroeconomic policy altogether. Better ways typically involve designing better institutions such as an independent central bank with an explicit inllation target) that can reduce the problem ol time inconsistency without eliminating monetary policy as a macroeconomic policy tool.
25.3 POLITICS AND POLICY
SHOULD POtlCY-MAKtKS BE RESTRAINED!
chapter 25
Now move from taxes to macroeconomic policy in general. Again, suppose that voters arc short¬sighted. II the politicians' main goal is ю please voters and get re-elected, what better policy than to expand aggregate demand before an election, leading to higher growth and lower unemployment? I rue growth in excess ol the normal growth rate cannot he sustained, and eventually the economy must return to the natural level ol output: higher growth must he lollowed later by lower growth. But with the right timing and short-sighted voters, higher growth can win thc elections. Thus, wc might expect a clear political business cycle, with higher growth oil average before elections than after elections.
The arguments wc have iust laid out are familiar,- in one form or another, you have heard them before. And their logic is convincing. So, it may come as a surprise that they don't lit the lacts veiy well.
Take lirst deficits and debt. The argument above would lead you to expect that budget deficits and high government debt have always been with us and always will he. l igure 25.3 which gives thc evolution ol thc ratio ol government debt to GDP in Australia since 1У00, shows that this isn't the case. Note how the lirst three build-ups in debt all happened in very special circumstances-. World War I the Great Depression and World War II —periods ol unusual declines in output or unusually high military spending. Note also how, Irom the end ol World War II to the present, the ratio ot debt to GDP has steadily decreased, from 1.87 to just under 0.1 now. The steady decrease in debt trom 1016 doesn't fit the argument ol short-sighted politicians very well. Or, in other words, short-sightedness doesn't explain much ol thc past evolution ot delicits and debt in Australia. I he data in Figure 25.3 exclude financial assets owned by thc government, and it you include these, net government debt is in fact negative 1 the OF.CD reported that net financial debt in Australia in 2008 was -7 per cent ot GDP). The negative net level ol debt in 2008 gave the Australian government signilicant Ireedom to run liscal deficits in the next few years, should the global financial crisis lead to a severe lall in output. Other countries like Japan and the United States have much higher government debt, and will be more constrained.
Return to the political business cycle argument in which policy-makers try to get high output growth bclore thc elections so that they will be re-elected. It the political business cycle was important, we would expect to see taster growth before elections than after. If it was rational tor politicians to exploit it then growth rales should on average be higher in the last year ot a term when an incumbent government wins re-election than when it loses.
From Okun's law, output growth in excess o: normal growth leads to a decline in the
: unemployment rate below the natural rate of unemployment, h the medium run. we know that the unemployment rate must increase back to the natural rate of unemployment. This, in turn, requires output growth below normal output growth for some time.
i The relation between the deficit, debt and GDP is explored in detail in Chapter 27.
Tabic 25.1 gives real GDP growth rates lor thc lirsi year, ihe middle yearl s > and thc last year of each Australian government since 1901, Oil average when a government won re-election, output growth was
The three major build-ups of government debt have been associated with World War I. the Great Depression and World War II.
Ever since, the ratio has been dropping.
SOURCE: M. Budin. RBA Discussion Paper. 1977. RBA Tables E09. G10
Figure 25.3
The evolution of the ratio of government debt to GDP. Australia. 1901-2008
Table 25.1 Output growth during the terms of office of Australian governments, 1901-2008
Government First year Year in office
Last year Middle year(s) (Win) Last year (Lose)
Right/conservative 3.5 3.7 4.0 1.9
Left/progressive 2.5 3.1 3.3 I.I
Average 3.2 3.5 3.8 1.5
1 1
SOURCE: RBA G10: M. Butlin. RBA Discussion Pnper. 1977.
higher in the last year ol a term compared with the middle year and the first year ot a term. This was true lor both right wing and left wing governments. This suggests that thc political business cycle argument worked and incumbent governments benefited from it. However, the average differences in the growth numbers across years are small <3.8 per cent in the last year ol a re-elected government versus 3.2 per cent in the lirst year). And so we cannot conclude that there is strong evidence of successlul manipulation. When we look at the governments that lost office, they had output growth rales in their last year on average quite significantly below growth in thc previous years of office (1.5 per cent in the last year versus .3.8 per cent in the first). This suggests lhat a poor economy is an important reason lor Australian governments losing office. When governments failed to stimulate the economy (perhaps because they were unable to), they lost office. When an incumbent government was re-elected, the growth rate in its last year was larger than in its earlier years. Therefore the data suggest lhai re-elected governments may have manipulated the economy to win re-election, while ihey were likely to have lost olfice il they allowed the economy to deteriorate.
Games between policy-makers
Another line ot argument focuses not on games between politicians and voters but on games between political parties. For example, lake the issue ol budget delicit reduction in the United States. Despite the fact that by the mid-1980s large budget deficits were widely perceived to be one ol the main macroeconomic problems lacing the United States, it took another filteen years before the deficit was eliminated. Some ol the delays are part ol the normal democratic process: deficit reductions involve making paintul decisions, and lorging a consensus takes time. But other tactors seem to be at work as well While agreeing on the need for deficit reduction, the two political parlies—Democrats and Republicans—differ on how it should be done. Because they believe in a smaller role for government, Republicans locus on decreases in spending. In contrast. Democrats are more open to increases in taxes. Each side holds out, hoping the other side will give in.
Game theorists refer to these situations as wars of attrition. The hope that the other side will give- in leads to long and olten cosily delays. Such wars ol attrition happen often in the context of fiscal policy. Deficit reduction often occurs long alter ii should. This is particularly visible during episodes of hyperinflation. As you saw in Chapter 24, hyperinflations arise Irom the use of money creation to finance large budget delicits. While the need to reduce those delicits is usually recognised early on, support for stabilisation programs—which include the elimination of those delicits—typically comes about only when inflation has reached such high levels that economic activity is severely atlecied.
Wars of attrition are not limited to fiscal policy.Think of the pilots' strike in Australia in 1989-90. when the Labor government sided with the airlines to f.ght an expensive war against the pilots' union. Or think of the 2004-05 > National Hockey League lockout in the Un ted States, where the complete season was lost because owners and players couldn't reach an agreement.
ВЛСК ТО F*OUCY chapter 25
Another example of games between political parties is movements in economic activity brought about by the alternation ol parties in power. Many believe that the conservative Coalition governments in Australia typically worry more than Labor about inflation and worry less than l abor about unemployment. So, we would expect Labor administrations to show stronger growth—and thus less unemployment and more inflation—than Coalition administrations. This prediction doesn't appear to lit the facts very well. T.ook back at Table 25.1: on average, output growth was lower lor Labor govern-
SHOULD POUCYMAKERS Bt RESTRAINED?
chaptc- 25
ments, though the differences are small.) Hut the data are a little distorted h> the rapid successions of governments at the beginning of the twentieth century. Il we focus on the last hall of thai century, ihe evidence for a difference in the two types of administration is even weaker.
Politics and fiscal restraints
If politics sometimes leads lo long and lasting budget deficits, can rules be put in placc to limit these adverse effects?
A constitutional amendment to balance the budget would surely eliminate the problem ol dclicits. But. just like a constant money growth rule in the case ol monetary policy, il eliminates the use of fiscal policy as a macroeconomic instrument altogether. This is just too high a price to pay.
A better approach is to put in placc rules lhal put limits cither on dclicits or on debt. This is, however, harder than it sounds. Rules such as limits on the ratio of the delicit to C.DP or the ratio of debt to GDI' arc more flexible than a balanced budget requirement; bui they may still not be flexible enough il the cconomy is affected by particularly bad shocks. I his has been made clcar by the problems laced by the Growth and Stability Pact these problems are discussed at more length in the focus box The Growth and Stability Pact in Europe: A short history.' More flexible or more complex rules, like rules that allow lor special circumstances, or rules that take into account ihe slate ot the economy, arc harder to design and especially harder to enforce, hor example, allowing the deficit to be higher il thc unemployment rate is higher than thc natural rate requires having a simple and unambiguous way of calculating what the natural rate is, a nearly impossible task.
A complementary approach is to put in place mechanisms to reduce dclicits. were such deficits to arise. Consider lor example, a mechanism thai triggers automatic spending cuts when the deficit gets too large. Suppose that ihe budget delicit is too large and it is desirable to eul spending across the board by 5 per cent Parliamentarians will lind it dillicult to explain to their constituency why their favourite spending program was cut by 5 per ccnt. Now suppose lhal the delicit triggers automatic across-lhe- board spending cuts ol 5 per ccnt without any parliamentary action. Knowing that other programs will be cut, members ol parliament will accept cms in their favourite programs more easily. They will also be better able to deflect thc blame tor the cuts. Those who succeed in limiting the cuts to their favourite program to. say, 4 per cent (by convincing the House to make deeper cuts in some other programs, so as to maintain thc lower overall level ol spending) can then return lo their constituents and claim that they successfully prevented even larger cuts.
Phis was indeed the general approach used lo reduce deficits in ihe United States in thc 1090s. The Budget Enforcement Act passed in 1990, and extended hv new legislation in 1993 and 1947, introduced two main rules:
• It imposed constraints on spending. Spending was divided into two categories: discretionary spend¬ing i roughly, spending on goods and services, including defence! and mandatory spending (roughly, transfer payments to individuals). Constraints, callcd spending caps, were set on discretionary spending for the following five years. These caps were set in such a way as to require a small but steady decrease in discretionary spending (in real termsi. Hxplicii provisions were made tor emergencies. Por example, spending on Operation Desert Storm during thc Gull War in 1991 was not subject to the caps.
• It required that a new transfer program could be adopted only il ii could he shown not to increase deficits in thc future i either by raising new revenues or by decreasing spending on an existing program'. This rule is known as thc pay-as-you-go or PAYGO rule.
Thc focus on spending rather than on the deficit itscll had one important implication. Il there was a recession—hence, a decrease in revenues—the deficit could increase without triggering a decrease in spending. This happened in 1991 and 1992 when, because ol ihe US recession, the deficit increased— despite thc fact that spending satisfied the constraints imposed by the caps. This focus on spending had iwo desirable cffccts. It allowed lor a larger liscal deficit during a recession—a good ihing from the point of view of macroeconomic policy—and il decreased the pressure to break thc rules during a recession—a good thing from a political point of view.
FOCI JS THE GROWTH AND STABILITY PACT IN EUROPE: J A SHORT HISTORY
The Maastricht Treaty, negotiated by the countries of the European Union in 1991, set a number of convergence criteria that countries had to meet in order to qualify to join the euro area. (For more on the history of the euro, see the focus box 'The euro: A short history' in Chapter 21.) Among them were two restrictions on fiscal policy. First, the ratio of a country's budget deficit to GDP had to be below 3 per cent. Second, the ratio of its debt to GDP had to be below 60 per cent, or at least 'approaching this value at a satisfactory pace'.
In 1997. would-be members of the euro area agreed to make some of these criteria permanent. The Stability and Growth Pact (SGP). signed in 1997, required members of the euro area to follow the following fiscal rules:
• That countries commit to balance their budget in the medium run—that they present programs to the European authorities, specifying their objectives for the current and following three years in order to show how they are making progress towards their medium-run goal.
• That countries avoid excessive deficits, except under exceptional circumstances. Following the Maastricht Treaty criteria, excessive deficits were defined as deficits in excess of 3 per cent of GDP Exceptional circumstances were defined as declines of GDP larger than 0.75 per cent.
• That sanctions be imposed on countries that ran excessive deficits. These sanctions could range from 0.2 per cent to 0.5 per cent of GDP—so. for a country like France, up to roughly A$ 13 billion!
Figure I plots the evolution of budget deficits since 1990 for the euro area as a whole. Note how from 1993 to 2000 budget balances went from a deficit of 5.7 per cent of euro area GDP to a surplus of 0.01 per cent. The performance of some of the member countries was particularly impressive: Greece reduced its deficit from 13.4 per cent of GDP to 1.4 per cent of GDP (although it was revealed in 2004 that the Greek government had cheated in reporting its deficit numbers, and that the actual improvement, although impressive, was less than reported); Italy eliminated its deficit, going from a deficit of 10.3 per cent of GDP in 1993 to a surplus of 0.7 per cent in 2000.
Figure I
l 1 г
1998 2000
2002
2004
SOURCE. OECD Economic Outlook, December 2005.
Euro area budget deficits as a percentage of GDP since 1990
SHOULD POLICY-MAKERS BL RESTRAINLD/
chapte 25
Can all the improvement be attributed to the Maastricht criteria and the SGP rules? The answer is 'no'. The decrease in nominal interest rates, which decreased the interest payments on the debt, and the strong expansion of the late 1990s both played important roles. But the fiscal rules also played a significant role.The carrot—the right to become a member of the euro area—was attractive enough to lead a number of countries to take tough measures to reduce their deficits.
Things turned around, however, in 2000. Since then, deficits have increased.The ratio of the deficit to GDP for the euro area was back up to 2.9 per cent in 2005. The main reason is low output growth since 2001, which has led to low tax revenues. Net government debt to GDP in the euro area has reached just below 60 per cent.
Although the deficit for the euro area as a whole is just below the 3 per cent limit, this is not the case for a number of individual countries. The first country to break the limit was Portugal, in 2001. with a deficit of 4.4 per cent.The next two were France and Germany, both with deficits in excess of 3 per cent of GDP since 2002. In each case, the government of the country decided it was more important to avoid a fiscal contraction that could lead to even slower output growth than to satisfy the rules of the SGP.
Faced with clear'excessive deficits' (and without the excuse of exceptional circumstances because output growth in each of these countries was low but positive), European authorities found themselves in a quandary. Starting the excessive deficit procedure against Portugal, a small country, might have been politically feasible, although it is doubtful that Portugal would have ever been willing to pay the fine. Starting the same procedure against the two largest members of the euro area, France and Germany, proved politically impossible.
After an internal fight between the two main European authorities, the European Commission and the European Council—the European Commission wanted to proceed with the excessive deficit procedure, while the European Council, which represents the states, did not—the procedure was suspended.
Since 2003, the legal status of the SGP has been in limbo, and the credibility of the Pact has been severely affected.This crisis has made clear that the rules were too inflexible. Romano Prodi, the head of the European Commission, admitted as much. In an interview in October 2002. he stated,'I know very well that the Stability Pact is stupid, like all decisions that are rigid'. And the attitudes of both France and Germany have shown that the threat to impose large fines on countries with excessive deficits was simply not credible.
The European Commission has therefore explored ways to improve the rules so as to make them more flexible and. by implication, more credible.The current proposals are to keep the 3 per cent deficit and 60 per cent debt numbers as desirable goals, but to focus less on the numbers for a particular year and more on the path of debt forecast for the medium run. Fines are not viewed as credible, so the plan is to rely on early and very public warnings, as well as on peer pressure from other euro area countries.
We saw the potential problems with such proposals earlier in the chapter. Flexible rules are harder to interpret and more prone to disagreements of interpretation. And public warnings and peer pressure only go so far. Finding the right rules is hard, and it is not clear whether or how the SGP will survive.
By 1998, liscal delicits in the Llnited States were gone and lor thc lirst time in twenty years the federal budget was in surplus Not all ol the deficit reduction was due to the Budget Enforcement Act rules. A decrease in defence spending due to the end ol the Cold War and a large increase in tax revenues due to the strong expansion ol the second hall ol the 1990s were important lactors. But there is wide agreement that the ailes played an important role in making sure that decreases in delence spending and increases in tax revenues were used for deficit reduction, rather than lor increases in other spending programs
Once budget surpluses appeared, however Congress became increasingly willing to break its own rules. Spending caps were systematically broken and the PAYGO rule was allowed to expire in 2002. In 2005 the deficit was again large and was predicted to remain so for many years to come. It is clear lhat, although spending caps and PAYGO niles were essential in reducing delicits in the 1990s, they have not prevented large delicits Irom reappearing in the 2000s. This is leading some economists and policy-makers to conclude that in the end nothing short ol a constitutional amendment can do the job ol avoiding delicits. The issue is likely to rise to the forefront of discussions ol LIS fiscal policy in thc near luture.
In Australia, while there is no one pushing lor a balanced-budget amendment or spending caps, the Coalition government Irom 1996 to 2007 built a political platform based on budget surpluses. For electoral reasons, the treasurer in that government. Peter Costello, created an artificial constraint on fiscal policy. From a macroeconomic perspective, this constraint is undesirable—a more flexible approach is needed. In 2008-09, all countries have bccomc more flexible in their fiscal policies as a necessary response to the global financial crisis lhal has threatened to bankrupt many financial institutions and push the economy towards recession.
The effects ol macroeconomic policies arc always uncertain. This uncertainty should lead policy¬makers ю be more cautious, to use less active policies. Policies must be broadly aimed at avoiding prolonged recessions, slowing down booms and avoiding inflationary pressure. The higher the level of unemployment or inflation, the more active the policies should be. But they should stop short of fine-tuning, ol trying to maintain constant unemployment, constant output growth or a constant rate of inflation. By targeting a range tor inflation the RBA has acknowledged thc need for caution. Using macroeconomic policy to control the economy is fundamentally different from controlling a machine. Unlike a machine, ihe economy is composed of people and firms who try to anticipate what policy-makers will do, who react not only to current policy but also to expectations of future policy. In this sense, macroeconomic policy can be thought of as a game between policy-makers and thc economy. When playing a game, it is sometimes better for a player to give up some options. For example, when a hijacking occurs, il is best to negotiate with hijackers. But a government lhat credibly commits not to negotiate with hijackers—that gives up the option ot negotiation—is actually more likely to deier hijackings from occurring.
The same argument applies to various aspects of macroeconomic policy. By credibly committing noi to use monetary policy to decrease unemployment below the natural rale of unemployment, a ccntral bank can alleviate tears that money growth will be high and decrease both expected and aciual inllation. When issues ol lime inconsistency arc relevant light restraints on policy-makers— such as a rigid interest-rate-setting rule in ihe case ol monetary policy—can provide a rough solution. But ihe solution may have large costs il it prevents the use ol macroeconomic policy in particular acute situations. Better methods typically involve designing better institutions such as an independent ccntral bank) that can reduce thc problem of time inconsistency without eliminating monetary policy as a llcxiblc macrocconomic policy tool.
Another argument for putting restraints on policy-makers is that policy-makers may play games either with the public or among themselves, and these games may lead to undesirable outcomes. Politicians may try to fool a short-sighted electorate by choosing policies with short-run benefits but large long-term costs—for example, large budget deficits. Political parties may delay painful decisions, hoping lhat the other party will make ihe adjustment and take the blame Belter ways typically involve better institutions and better ways of designing the process through which policy and decisions are made.
dynamic stochastic general equilibrium. 573 fine-tuning, 576 optimal control, 576 game, 576
optimal control theory. 576 game theory, 576 strategic interactions, 576
players, 576 time inconsistency, 578 political business cycle, 581 wars of attrition, 582 spending caps. 583 PAYCO rule, 583
5HOULD POLICY-MAKERS Bt RESTRAINED?
chapte'- 25
QUESTIONS AND PROBLEMS
Quick check
1. Using the information in this chaptcr, label each of the following statements 'true', 'false' or 'uncertain'. Explain briefly.
a. There is so much uncertainty about thc effects ot monetary policy that wc would be better oil not using it.
b. Elect a Labor government if you want low unemployment.
c. There is clear evidence of political business cycles in Australia: high growth and low unemployment around elections, lower growth and higher unemployment thc rest of the time
d. Rules arc ineffective in reducing budget deficits.
c. Governments would be wise to announce a no-negotiation policy with hostage takers.
I. If hostages arc taken it is clearly wise lor governments to negotiate with the hostage takers, even
if the government has announced a no-negotiation policv. g. When a ccntral bank announces a target inflation rate, it has no incentive to deviate trom thc target.
2. Implementing a political business cycle
You are the economic adviser to a newly elected government. In four years' time, the government will face another election. Voters want a low unemployment rate and a low inflation rate. However, you believe that voting decisions are influenced heavily by the values of unemployment and inflation in the last year before the election, and that the economy's performance in thc first three years of a government's term of office has little effect on voting behaviour.
Assume that inflation last year was .1 per cent and that the unemployment rate was equal to the natural rate. The Phillips curve is given by
IT, 77j_ , - «(», - u„)
Assume that you can use fiscal and monetary policy to achieve any unemployment rate you want for each of the next four years. Your task is to help the government achieve low unemployment and low inflation in the last year of its administration.
a. Suppose you want to achieve a low unemployment rate that is, an unemployment rate below the natural ratel in the year before the next election (four years Irom today). What will happen to inflation in the lourth year?
b. Given the effect on inflation you identified in pari (a), what would you advise ihe government to do in the early years of its administration to achieve low inflation in thc fourth year?
c. Now suppose that the Phillips curvc is given by
77, = 7rf - ot(u, - иn)
In addition, assume that people form inflation cxpcctations7r! based on consideration ol thc luture (as opposed lo looking only ai inflation last year), and are aware lhat thc government has an incentive to carry out the policies you identified in parts 'ai and 'hi. Arc the policies you described in parts 'a) and (b) likely lo be successful? Why or why not?
3. Suppose that the government amends the constitution to prevent government officials from negotiating with terrorists.
Whai arc the advantages of such a policy? What are the disadvantages?
4. New Zealand rewrote the charter of its central bank in the early 1990s to make steady, low inflation its only goal.
Why would New Zealand want to do ihis?
Dig deeper
5. Political expectations, inflation and unemployment
There are two parties: Labor, which cares a lot more about unemployment than inflation, and the Coalition, which cares a lot more about inflation titan unemployment. When Labor is in power, il chooses an inllation rate of — |, and when the Coalition is in power, il chooses an inflation rate of TTi l-Ve assume that ТГ[ > ■irt. The Phillips curve is given by
ir, = rf - a(u, - u„)
An election is about lo be held. Assume that expectations about inflation in the coming year тг\ are formed before the election in l. (Essentially, this assumption means that wages for the coming year are set before the election.I Moreover. Labor and the Coalition have an equal chance of winning the election.
a. Solve for expected inflation, in terms ol 77; and 77(.
b. Suppose that l abor wins the election and implements its target inllation rale. 77,. Given your solution lor expected inflation in part (a), how will the unemployment rate compare with the natural rate ol unemployment?
c. Suppose that the Coalition wins the election and implements its target inllation rate. 7tl-. Given your solution lor expected inflation in part a , how will the unemployment rale compare with the natural rate ot unemployment?
d. Do these results tit the evidence in Table 25.1? Why or why not?
e. Now suppose that everybody expects Labor to win the election and Labor indeed wins. Il Labor implements its target inflation rale, how will the unemployment rate compare with the natural rate?
6. Deficit reduction as a prisoner's dilemma game
Suppose there is n budget deficit. It can be reduced by cutting military spending by cutting welfare programs, or by cutting both. Labor has lo decide whether to support cuts in social security programs. The Coalition has lo decide whether to support cuts in military spending. The possible outcomes are represented in the following table:
Defence cuts Social security cuts
Yes No
Yes (C= 1,1. = —2) (C - 2. L 3)
No (C ■ 3, 1. -2) (C. = -l, L = -1)
The table presents payoffs to each party under the various outcomes. Think of a payoff as a measure of happiness for a given parly under a given outcome. If Labor votes for welfare cuts, and the Coalition vole against cuts iu military spending, the Coalition receives a payoff of Л and Labor receives a payoff of -2.
a. II the Coalition decides to cut military spending, what is the best response ol Labor? Given this response, what is the payoll lor thc Coalition?
b. It the Coalition decides not to cut military spending, what is the best response ol Labor? Given this response, what is the payoll lor the Coalition?
c. What will the Coalition do? What will Labor do? Will the budget delicit be reduced? Why or why not?
(A game with a payotl structure like the one in this problem, and which produces the outcome you have just described, is known as a prisoner's dilemma. Is there a way to improve the outcome?
SHOUI D POIICYMAKFRS Bf- RESTRAIN! D!
chapter 25
Explore further
7. Games, precommitment, and lime inconsistency in the news
Current events offer abundant examples of disputes in which the parties are involved in a game, try to commit themselves t.i lines of action in advance, and face issues of time inconsistency. Examples arise in the domestic political process, international affairs and labour-management relations.
a. Choose a current dispute 1 or one resolved recently) to investigate. Do an Internet search to learn thc issues involved in the dispute, the actions taken by the parties to date, and the current state of play.
b. In what ways have the parties tried to precommil to certain actions in the luture? Do they face issues of time inconsistency? Have the parties failed to carry out any of their threatened actions?
c. Does the dispute resemble a prisoner's dilemma game a game with a payoff structure like the one described in problem 6)? In other words, does it seem likely (or did it actually happen) that thc individual incentives of thc parties will lead them to an unfavourable outcome—one that could be improved lor both parties through cooperation? Is there a deal to be made? What attempts have the parlies made lo negotiate?
d. How do you think the dispute will be resolved (or how has it been resolved)?
We invite you to visit the Blanchard-Sheen page on the Pearson Australia website at
www.pearson.com.au/highered/blanchardsheen3e
for many World Wide Web exercises relating to issues similar to those in this chapter.
FURTHER READINGS
CHAPTER ф
Monetary Policy: A Summing Up
N
early every chapter of this book has touched on an aspect of monetary policy. This chapter puts it all together and ties up the remaining loose ends.
Let's first briefly review what you have learned (the focus box 'Monetary policy: What you have learned and where', gives a more detailed summary):
• In the short run. monetary policy affects the level of output as well as its composition:
- A decrease in interest rates involves an increase in money and a depreciation of the currency.
- These lead to an increase in the demand for goods and an increase in output.
• In the medium run and the long run. monetary policy is neutral:
- Changes in either the level or the rate of growth of money have no effect on output or unemployment.
- Changes in the level of money are associated with proportional increases in prices.
- Changes in the inflation rate are associated with corresponding changes in the rate of nominal money growth. With these conclusions in mind, this chapter explores three issues.
• Section 26.1 discusses the inflation rate that central banks should try to achieve in the medium and long run.
• Section 26.2 discusses how monetary policy should be designed to achieve this inflation rate in the medium and long run. as well as to reduce output fluctuations in the short run.
• Section 26.3 describes how monetary policy is actually carried out in Australia today. 26.1 THE OPTIMAL INFLATION RATE
Tabic 26.1 shows how inflation has steadily gone down in rich countries since the early 1980s. In 1981 average inllation in the OFCD was 10.5 per cent,- in 2005 it was down to 3.5 per cent. In 1981 two countries (out of thirty had an inflation rate below 5 percent.- by 2008 the number of countries had increased to twenty-two.
Does this imply that most central banks have now achieved their goal? Or should they aim lor an even lower inflation rate, perhaps 0 per cent? Thc answer depends on the costs and benefits of inflation.
The costs of inflation
We saw in Chapter 24 how very high inflation—say, 30 per cent a month or more—can disrupt economic activity. The debate in OFCD countries today, however, isn't about the costs ol inflation rates of 30 percent a month or more. Rather, it centres on the advantages of, say, 0 per cent versus 4 per cent inflation a year. Within that range, economists identify
In Chapter 4 we looked at the determination of money demand and money supply, and the effects of monetary policy on the interest rate and the money supply.
You saw how a decrease in the interest rate, achieved through an open-market operation, leads to an increase in the money supply (or vice versa).
In Chapter 5 we looked at the short-run effects of monetary policy on output.
If the central bank fixed the money supply and then increased it, you saw how that led to a decrease in the interest rate, to an increase in spending and an increase in output. Equivalently, if the central bank fixed the interest rate and then lowered it. the money supply would have to be increased, which would lead to an increase in spending and output.
In Chapter 7 we looked at the effects of monetary policy on output and prices, not only in the short run but also in the medium run.
Whether the central bank fixed the money supply or the interest rate, you saw that in the medium run money is neutral: changes in money are fully reflected in changes in the price level. In the short run, tighter monetary policy reduced output and prices.
In Chapter 9 we looked at the relation between inflation, unemployment and output growth.
You saw that in the medium run a tighter monetary policy (either thought of as a reduction in the rate of growth of the money supply or a lower inflation target) leads to a one-for-one lower inflation rate and nominal money growth rate, leaving the unemployment rate unaffected.You saw that in the short run, however, tighter monetary policy leads to higher interest rates, and thus to lower output and higher unemployment for some time.
In Chapter 14 we introduced a distinction between the nominal interest rate and the real interest rate.
You saw how easier monetary policy leads to a lower nominal interest rate and a lower real interest rate in the short run. but to a higher nominal interest rate and an unchanged real interest rate in the medium run.
In Chapter 17 we returned to the short-run effects of monetary policy on output, taking into account the effects of monetary policy on expectations.
You saw that monetary policy affects the short-term nominal interest rate (the cash rate in Australia), but that spending depends primarily on both current and expected future short-term real interest rates. You saw how the effects of monetary policy on output depend on how expectations respond to monetary policy.
In Chapter 20 we looked at the effects of monetary policy in an economy open both in goods markets and financial markets.
You saw how. in ar open economy, monetary policy affects spending and output not only through the interest rate but also through the exchange rate. A decrease in the interest rate leads to a depreciation, both of which increase spending and output.
In Chapter 21 you saw how monetary policy leads to exchange rate overshooting, amplifying the real effects in the short run. However, monetary policy was neutral in the medium run. and the equilibrium was the same whether exchange rates were fixed or flexible.
MONETARY POl ICY- A SUMMING LP chapter 26
us
T
(BOX
We discussed the pros and cons of different monetary policy regimes, flexible exchange rates versus fixed exchange rates. We discussed the pros and cons of adopting a common currency such as the euro, or even giving up monetary policy altogether, through the adoption of a currency board or dollarisation. In Chapters 22 and 23 we looked at the implications of a financial crisis that may lead into a liquidity trap, the fact that monetary policy cannot decrease the nominal interest rate below zero. A financial crisis is typically associated with falls in stock prices, and an increase in the cost of borrowing, which reduces aggregate investment and thus output. Monetary policy can be used to try to moderate a recession, but you saw how the liquidity trap and deflation can combine to turn a recession into a slump or a depression.
• In Chapter 24 we studied hyperinflations and looked at the conditions under which such episodes arise and eventually end.
We focused on the relation between the budget deficit, nominal money growth and inflation.You saw how a large budget deficit can lead to high nominal money growth and, in turn, to hyperinflation.
• In Chapter 25 we looked at the problems facing macroeconomic policy in general, and monetary policy in particular.
You saw that uncertainty about the effects of policy should lead to more cautious policies. You saw that even well-intentioned policy-makers may sometimes not do what is best, and that there is a case for restraints on policy-makers. We also looked at the case for giving independence to the central bank and appointing a conservative central banker.
From Chapter 14: In the medium run. the real interest rate isn't affected by inflaton. Thus, an increase in * inflation is reflected one- for-one in an increase in the nominal interest rate. (This is called the Fisher effect.)
Some economists argue that the costs of bracket creep were much larger. As tax revenues steadily increased, there was itde pressure on the government to control spending. The result, they argue, was an increase in the size of the government in the * 1960s and 1970s far beyond what would have been desirable.
• In this chapter we discuss the issues of the optimal inflation rate, the design of monetary policy, and how the RBA actually conducts monetary policy in Australia today.
Table 26.1 Inflation rates in the OECD, 1981-2008
Year 1981 1985 1990 1995 2000 2008
OECD average (%) 10.5 6.6 6.2 9.0 5.3 3.5
Number of countries with
inflation below 5% 2 10 15 21 22 22
The number of OECD countries with inflation rotes of the GDP deflator less than 5 per cent has more than doubled since I99S.
SOURCE: OECD Economic Outlook. Annex Table 16.
five main costs of inflation.- i I ! shoe-leather costs, '2) tax distortions, (3? money illusion, 4 inflation variability and 5' redistribution costs trom creditors to debtors.
Shoe-leather costs
In the medium run, a higher inllation rate leads to a higher nominal interest rate, and so to a higher opportunity cost ol holding money. As a result, people decrease their money balances by making more trips to thc bank—thus the expression shoe-leather costs. These trips would be avoided if inllation were lower, and people could be doing other things instead, such as working more or enjoying their leisure time.
During hyperinflations, shoe-leather costs can become quite large. But their importance in times of moderate inflation is limited. It an inflation rare of 4 per cent leads people to go the bank say, one more time every month, or to do one more transaction between their term deposit funds and their currcnt account every month this hardly qualifies as a major cost of inflation.
Tax distortions
1 he second cost ol inflation comes from the interaction between thc tax system and inflation.
BACK TO POLICY
chaple" 26
Consider, lor example, income taxes. Thc income levels corresponding to dillerent income-tax rates aren't usually increased automatically with inflation. As a result, many people arc pushed into higher tax brackets as their nominal income—but not necessarily their real income- increases over time, an effect known as bracket creep. Increases in these income thresholds arc often manipulated by governments in their annual budgets to make them look like tax cuts, which arc alwavs popular with voters. The truth is that often these tax cuts are simply cancelling out the automatic lax increase implied by bracket creep from inllation.
MONETARY POLICY: Л SUMMING UP chapter 26
Another related issue is taxes on asset income such as interest or rent or dividends. Although we know that the real rate ol return on an asset is the real interest rate, not thc nominal interest rate, income tor the purpose of income taxation typically includes nominal interest payments, not real interest payments. This would make little difference il:
• all forms ol capital income were taxed in the same way,-
• any interest paid was fully tax-deductible,- and
• income-tax brackets were fully adjusted for inflation.
In many countries, the nominal interest paid on borrowings tor investment in assets such as shares or investment property is tax-deductible. In Australia, this feature is associated with what is called negative gearing—when you borrow to buy a property or shares, you can pay more nominal interest than you get as rent or dividend on the asset which is where thc negative comes from), and then claim the loss as a tax deduction on your other taxable income sources. Since rental rates and dividends should also have an equivalent compensation lor the inflation component, there is no inflation distortion when thc diffcrcncc between the rent or dividend and the interest rate is subject to tax. Thus, contrary to popular belief, negative gearing doesn't introduce a distortion related to inflation. Why do people accept the tax-subsidised loss in their negatively geared investments? Because, in the long term, they are hoping for substantial capital gains. Ideally, any realised real capital gains should be taxed in the same way as normal real income.
So, how are capital gains taxed? Taxes on capital gains in many countries are based on the change in the local currency price ol the asset between the time it was purchased and the time it is sold. This implies that the higher the rate ot inflation, the higher the tax. An example will make this clear:
• Suppose that inflation nas been running at тг percent a year lor the last ten years.
• Suppose you bought shares in one ot these countries tor $50,000 ten years ago, and you are selling them today for $50,000 x (1 i тг%)м'—so their real value is unchanged.
• II the capital gains lax is 30 per cent, the effective tax rate on the sale ol the shares—defined as the
ratio ol thc lax you pay to the price for which you sell the shares—is equal to: 4 The numerator от the
50,0001 I + 77% j'"-50,000 50,000(1 + 77%i'°
• Because you are selling your shares lor the same real price lor which you bought them, your real capital gain is zero and you shouldn't be paying any tax. Indeed, if тт - 0—il there has been no inflation then the effective tax rate is 0. But if тт = 4 per cent then the effective lax rate is 9.7 per cent. Despite thc fact that your real capital gain is zero, you end up paying a high tax.
In Australia over the last twenty years, policy on capital gains taxation has varied considerably. Prior to 1985, there was no capital gains taxation, which meant a distortion in favour of capital earnings relative to labour income. When capital gains taxation was introduced in 1985. only realised real capital gains were taxed at a person's marginal income-tax rate (that is, the tax rate a person pays on any additional income). So, no inllation distortion existed. But perversely in 1999. the government decided to change the Riles again, so thai only realised nominal capital gains were to be taxed, but ai a conces¬sional rate ol hall of the person's marginal income-tax rate. Thus capital gains taxation in Ausiralia now suffers from two distortions: the marginal tax rate is half that on income,- and capital gains due to inflation are taxed.
But there is more. Since capital gains tax was introduced, it has applied to all asset ownership except owner-occupied housing. The 'laniilv home' has never been subject to capital gains tax, and this represents a major tax distortion. It is a distortion because:
• Only those with owner-occupied homes benefit.
• It pushes tip house prices, making it harder (or young people to enter ihe housing market.
fraction equals die sale price minus the purchase price.The denominator is the sale price.
(30%)
• It shilts financial investments towards housing and away Irom other torms ol asset accumulation that are subject to capital gains tax.
SACK TO POLICY chapter 26
Economists have estimated thai average house prices in Australia in 2008 were up to 30 per cent over-priced. Also, you saw in Chapter 15, Figure 15.0, what appeared to be housing bubbles in Australia Irom 2001 to 2004 and again in 2006 to 2007. Could these capital gains tax distortions have led to over¬priced housing and amplified the housing bubble? The answer is almost certainly 'yes'.
Ii is possible to argue that these tax costs aren't examples of a cost of inflation per se, but rather the result ol a badly designed tax and regulation system. In the income-tax threshold example we discussed, thc government could eliminate the problem il it committed itself to automatic indexation of the income thresholds to the price level. This isn't impossible—the United States did it in the early 1980s. The Australian government could, and it should, lax all capital gains at the lull marginal income-tax rate. But because governments arc often reluctant to introduce such systematic adjustment, ihe inflation rate matters and leads to distortions.
Money illusion
The lliird cost comes from money illusion the notion lhat people appear ю make systematic mistakes in assessing nominal changes versus real changes. A number ol calculations lhat would he simple under price stability become more complicated when there is inllation. In comparing their income this year with their income in ihe past, people have to keep track ol ihe hisiory ol inflation. In choosing between different assets or deciding how much to consume or save, they have to keep track of the difference between the real interest rate and the nominal interest rate. Casual evidence suggests that many people find these calculations difficult and often fail lo make ihe relevant distinctions. Economists and psychologists have gathered more formal evidence, and it suggests lhal inflation olten leads people and lirms lo lake incorrect decisions isee the locus box 'Money illusion ). II this is the case then a simple solution is to have no inllation.
Inflation variability
This cost comes from thc fact that higher inflation is typically associated with more variable inflation. And more variable inflation means that financial assets such as bonds, which promise fixed nominal payments in the future, become riskier.
Take a bond that pays $ I 000 in ten years. With constant inllation over the next ten years, not only the nominal value bin also the real value ol the bond in ten years is known with certainty—we can calculate exactly how much a dollar will be worth in ten years. But with variable inflation, the real value ol $1,000 in ten years becomes uncertain. Thc more variability, the more uncertainty. Saving for retirement becomes more difficult. For those who have invested in bonds, lower inflation than expected means a better retirement; but higher inflation may mean poverty. This is one ol the reasons that retirees, for whom part ot income is fixed in dollar terms, typically worry more about inflation than A good but sad movie ► other groups in the population, about surviving on a
PWorld War II in Italy is
Umberto D byVittorio You may argue, as in the case of laxes, lhat these costs aren't due to inllation per se, but rather to thc de Sica. made in 1952. financial markets' inability to provide assets that protect their holders against inflation. Rather than issuing only nominal bonds (bonds that promise a fixed nominal amount in the future), governments or firms could also issue indexed bonds bonds that promise a nominal amount adjusted for inflation, so people con't have to worry about the real value of thc bond when they retire. Indeed, as we saw in C hapter 15, a number ol countries, including Australia and the United States, have introduced such bonds, so that people can better protect themselves against movements in inflation. However, in Australia there has been little demand tor indexed bonds.
The conclusion remains: anyone who owns an asset whose value isn't indexed lo inllation will suffer when inflation increases. This is because the real value ol that asset must lall. Anyone who has a debt whose value isn't indexed lo inflation will be better olt when inflation increases. The real value ol that debt decreases. In every country this cost is a tact ol lile and therefore a good reason for keeping inflation low.
MONEY ILLUSION
There is a lot of anecdotal evidence that many people fail to adjust properly for inflation in financial calculations. Recently, economists and psychologists have started looking at money illusion more closely. In a recent study in the United States, two psychologists, Eldar Shafir from Princeton University and Amos Tversky from Stanford University, and one economist, Peter Diamond from the Massachusetts Institute of Technology, designed a survey aimed at finding the presence and the determinants of money illusion. Among the many questions they asked of people in various groups (people at Newark International Airport, people at two New Jersey shopping malls, and a group of Princeton undergraduates) is the following.
Suppose that Adam. Ben and Carl each received an inheritance of $200,000 and each used it immediately
to purchase a house. Suppose that each sold his house one year after buying it. Economic conditions were.
however, different in each case:
• During the time that Adam owned the house there was a 25 per cent deflation—the prices of all goods and services decreased by approximately 25 per cent. A year after Adam bought the house, he sold it for $ 154,000 (23 per cent less than he had paid for it).
• During the time Ben owned the house, there was no inflation or deflation—the prices of all goods and services didn't change significantly during the year. A year after Ben bought the house, he sold it for $ 198,000 (I per cent less than he had paid for it).
• During the time Carl owned the house, there was a 25 per cent inflation—the prices of all goods and services increased by approximately 25 per cent. A year after Carl bought the house, he sold it for $246,000 (23 per cent more than he had paid for it).
Please rank Adam, Ben and Carl in terms of the success of their house transactions. Assign I to the
person who made the best deal and 3 to the person who made the worst deal.
In nominal terms. Carl clearly made the best deal, followed by Ben. followed by Adam. But what is relevant is how they did in real terms—adjusting for inflation. In real terms, the ranking is reversed: Adam, with a 2 per cent real gain, made the best deal, followed by Ben (with a I per cent loss), followed by Carl (with a 2 per cent loss).
The survey's answers are summarised below.
Rank Adam (%) Ben <%) Carl (%)
1st 37 15 48
2nd 10 74 16
3rd 53 II 36
1 1
Carl was ranked first by 48 per cent of the respondents, and Adam was ranked third by 53 per cent of the respondents.These answers are very suggestive of money illusion. In other words, people (even Princeton undergraduates) have a hard time adjusting for inflation.
SOURCE: E. Shafir, R Diamond and A.Trersky, 'Money illusion'. Quarterly Journal of Economics, vol. 112. no. 2. May 1997. pp. 341-74. ©The President and Fellows of Harvard College and the Massachusetts Institute ofTechnology.
The benefits of inflation
FOCUS 'BOX
MONFTARY TO! ICY: A SUMMING LP chapter 26
Inflation is actually not all bad. One can identify three benefits of inflation: I) seignorage, (2) thc option of negative real interest rates tor macroeconomic policy, and (3) (somewhat paradoxically) the use of the interaction between money illusion and inflation in facilitating real wage adjustments.
Seignorage
Money creation—tine ultimate source of inllation—is one of the ways in which the government can linance its spending. Put another way, money creation is an alternative to borrowing from the public or raising taxes.
As you saw in Chapter 24, the government typically does not create' money to pay lor its spending. Rather, the government issues and sells bonds and spends the proceeds. But it the bonds are bought by the central bank, which then creates money to pay for them, the result is the same: other things being equal, thc revenues trom money creation—that is, seignorage—allow the government to borrow less from thc public or to lower taxes.
How large is seignorage in practice? When looking at hyperinflations in Chaptcr 24, you saw that seignorage is often an important source of government finance in countries with very high inflation rates. But its importance in OECD economies today, and for thc range of inflation rates we are considering, is much more limited. Take the case of Australia. In 20(18 the ratio ol thc monetary base— the money issued by the RBA • see Chapter 4)—to GDP was 4.2 percent. An increase in nominal money base growth of 10 per cent per year (which should eventually raise inflation would lead therefore to an increase in seignorage ol 10% x 4.2% or 0.42 per cent ol GDP. This is a small amount ol revenues to get in exchange lor up to 10 per cent more inflation. Compared with the 37 per cent ol GDP raised by other taxes it is small indeed:
Therclore while the seignorage argument is sometimes relevant tor example, in economics that don't yet have a good liscal system in place), it seems hardly relevant in thc discussion of whether OF.CI) countries today should have. say. 0 per ccnt versus 10 per ccnt inflation.
The option of negative real interest rates
This argument lollows from our discussion ol thc liquidity trap and its macroeconomic implications in Chapter 23. A numerical example will help here.
• Consider two economies, both with a real interest rate equal to 2 per cent. In the lirst economy, the central bank maintains an average inflation rate of I per ccnt, so the nominal interest rate is on average equal to 2% + 4% - 6%. In thc second economy, the ccntral bank maintains an average inflation rate of 0 per ccnt, so the nominal interest rate is on average equal to 2% + 0% = 2%.
• Suppose that both economics are hit by a similar adverse shock, which leads, at a given interest rate, to a decrease in spending and a decrease in output in the short run. In the lirst economy, the central bank can decrease the nominal interest rate Irom 6 per ccnt to 0 per cent, a decrease of 6 per cent. Under the assumption that cxpectcd inflation doesn't change immediately, and remains equal to 4 per cent, the real interest rate decreases from 2 per cent to -4 per ccnt. This is likely to have a strong positive cllcct on spending and help thc cconomy to recover. In the second economy, the central bank can only decrease thc nominal interest rate from 2 per ccnt to 0 per ccnt. a decrease of 2 per ccnt. Under thc assumption that cxpectcd inllation doesn't changc right away, and remains equal lo 0 per cent, the real interest rale decreases by only 2 per cent, from 2 per cent to 0 per cent. This small decrease in the real interest rate may not increase spending by very much.
In short an economy with a higher average inllation rate has more scope to use monetary policy to fight a recession. An economy with a low average inllation rate may find itsell unable to use monetary policy to return output to the natural level of output. As you saw in Chaptcr 2.3, this is lar Irom being iust a theoretical possibility. Japan laced precisely such a limit on monetary policy, and its recession turned into a long slump. Some economists worry lhal other countries may also be at risk. Many countries, including Australia, have low inflation and low nominal interest rates. If, for any reason, some of these countries were to be laced with further adverse shocks to spending, thc room tor monetary policy to help avoid a decline in output would dearly be limited.
Let H denote the ► monetary base— the money issued by the central bark.
Then seignorage/ У = \HIPY = (ЛН/Н) (HIPY) where ЛН/Н is the rate of growth of the monetary base, and HIPY is the ratio of the monetary base to nominal GDP
At the time of writing.
the three-month nominal interest rate in Australia is 6.3 per cent and the inflation rate is 4.S per cent. In the United States, these are 2.1 per cent and 5.4 per cent, and for the euro area 5.4 per cent and ► 3.6 per cent In the early 1990s. thc US Fed decreased the nominal interest rate by 7 per cent. This was still not enough to avoid a
This is ihen a major reason why ccntral banks need lo choose inflation targets well above zero. It is possible that the targets lhat have been chosen could be a little low. Examples of central banks lhat have explicit inflation targets are the Reserve Bank of Australia with a range of 2-3 per cent, ihe Bank ol Canada 1-3 per ccnt, thc Bank ol England 2.5 per cent, the Reserve Bank ot New Zealand 0-3 per cent,
and the Swedish central hank I 3 per cent. Though there are other central hanks that are less explicit about inflation targets, inflation is still important in their monetary policy design. For example the European Central Hank aims for price stability . defined as inflation below or close to 2 per cent in the medium term. Even at the top end of the range of these inflation targets, it is quite possible that they may have been set too low—an inllation target ol 3 per cent and the implied medium-run nominal interest rate ol about 6 per cent probably gives a central bank too little room to manoeuvre it future negative shocks require large decreases in interest rates. In other words il the current deflation risks are realised in the luture, central banks may come to regret having set their inflation targets so low.
Money illusion revisited
Paradoxically, the presence ol money illusion provides at least one argument lor having a positive- inflation rate.
Го see why. consider two scenarios. In the lirst inflation is 4 per cent and your wage goes up by I per cent in nominal terms—in dollars. In the second, inflation is 0 per cent and your wage goes down by 3 per cent in nominal terms. Both lead to the same 3 per cent decrease in your real wage, so you should be indifferent. The evidence however, is that many people will accept thc real wage cut more easily in the lirst case than in the second case.
Why is this example relevant to our discussion? Because, as you saw in Chapter 13, the constant process ol change that characterises modern economics means that some workers must sometimes take a real pay cut. Thus, the argument goes, the presence ol inllation allows for these downward real-wage adjustments more easily than when there is no inflation. This argument is plausible. Economists haven't established its importance hut because so many economies now have very low inllation we may soon be in a position to test it.
The optimal inflation rate: The current debate
At this stage, the debate in OF.С I) countries is between those who think some inllation say, 3 percent is line and those who want to achieve price stability—that is 0 percent inflation.
Those who want an inllation rate around 3 per cent emphasise lhat the costs ol 3 per cent versus
0 per cent inflation are small and that the benelits of inflation are worth keeping. They argue that some
1 the costs ol inflation could be avoided by indexing thc tax system and issuing more indexed bonds. They alst) say that going Irom current rales ol inflation to 0 percent would require some increase in unemployment lor some time, and that this transition cosi may well exceed thc eventual benefits.
See. for example, the results of a survey of managers by Alan
* Blinder and Don Choi, in 'A shred of evidence on theories of wage rigidity'. Quarterly Journal of Economics. 1990. pp. 1003-16.
4 A conflict of metaphors: because inflation makes these real-wage adjustments easier to achieve, some economists say that inflation 'greases the wheels' of the economy. Others, emphasising the adverse effects of inflation on relative prices, say that inflation 'puts sand' in the economy.
MONITARY POLICY: A SUMMING U:>
chapter 26
Those who want to aim lor 0 per cent make lhe point that 0 per cent is a very different target rale from any other— it corresponds to price stability This is desirable in itself. Knowing that the price level will be roughly the same in ten or twenty years as it is today simplifies many decisions, and eliminates the scope for money illusion. Also, given the time consistency problem facing central banks (discussed in Chapter 25) credibility and simplicity ot the target inflation rate are important. Price stability may achieve these goals better than a target inflation rate ol 3 per cent.
The debate isn't settled. For the lime being, most central banks appear to be aiming for low but positive inflation that is, inflation rates ol between 1 per cent and 3 per cent, but some argue that this may not be a sufficiently cautious butler against deflation risks.
26.2 THE DESIGN OF MONETARY POLICY
Until the beginning ol thc lc)<)0s the design ol monetary policy typically centred around nominal money growth. Central banks chose a nominal money growth target tor the medium run. and they thought about short-run monetary policy in terms ot deviations ol nominal money growth from thai target. In the past decade, however, this design has evolved. Most central banks have adopted an inflation rale target rather than a nominal money growth rate target. And they think about short-run monetary policy in terms ol movements in the nominal interest rate, rather than in terms ol movements in the rale ol nominal money growth. This is why we placed so much emphasis in earlier chapters on
monetary policy as setting the interest rate on the hasis ot an inflation target. Let's lirst look at what central hanks did earlier, before turning to what they do now.
Money growth targets and target ranges
Llntil the 1990s, monetary policy in thc Llnitcd States and in other OEC.D countries was typically conducted as follows:
• Thc central bank chose a target rate for nominal money growth corresponding to the inflation rate it wanted to achieve in the medium run. Il, tor example, it wanted to achieve an inflation rate ol 4 per ccnt and thc normal rate ol growth of output (the rate ot growth implied by the rate ot technological progress and the rate ol population growth) was 3 per cent, thc central bank chose a target rate of nominal money growth of 7 per cent.
• In thc short run, thc central bank allowed for deviations of nominal money growth Irom the target. II. lor example, the economy was in a recession, the central bank increased nominal money growth above the target value to allow lor a dccrcasc in the interest rate and a laster recovery ol output. In an expansion, it might do thc reverse lo slow down output growth.
• Го communicate to the public both what it wanted to achieve in the medium mn and what it intended to do in the short run, the central bank announced a range lor the rate of nominal money growth it intended to achieve. Sometimes this range was presented as a commitment from the ccntral bank,- sometimes it was presented simply as a forecast rather than a commitment.
Recall how inflation ► and nominal money growth move together during episodes of hyperinflation (Chapter 24).
From Chapter 4: MI ► measures the amount of money in the economy, and is constructed as the sum of currency and current deposit accounts.The RBA doesn't dirccdy control Ml What it can control is H. the monetary base: but it can choose H and through that try to achieve any value of MI it wants. So, n is reasonable to think of the RBA as controlling MI.
Over time, central hanks became disenchanted with this way ol conducting monetary policy. Let's now look at why.
Money growth and inflation revisited
The design of monetary policy around nominal money growth is based on thc assumption of a closc relation between inflation and nominal money growth in the medium mil. The problem is lhat ihis relation is in tact nol very light. II nominal money growth is high, inflation will also be high: and il nominal money growth is low. inflation will be low. But the relation isn't tight enough that, by choosing a rate of nominal money growth, the central bank can achievc precisely its desired rate ol inflation, not even in thc medium run.
Mt growth and inflation
The relation between inllation and nominal money growth is shown in Figure 26.1. which plots ten- year averages ol the inflation rate using the CPI as the price indexi against ten-year averages of the growth rate of two measures of money stock 'Ml and M3) Irom 191 I to 2008. Thc reason for using ten-year averages is this: in thc short run, changes in nominal money growth (via the interest rate affect mostly output, not inllation. It is only in the medium run lhat a relation between nominal money growth and inflation should emerge. Taking ten-year averages ol both nominal money growth and inflation is a way ol detecting such a medium-run relation.
Figure 26.1 shows that for Australia sincc 1912 the relation between Ml growth and inflation hasn't remained very light. There have been some extended periods when they did move closely together— from 1920 until World War II, and Irom ihe mid-1950s lo 1986. Bui note how inflation started declining in thc early 1980s, reaching 2.4 per ccnt in 2005, while nominal money growth has remained high at around I I per cent. Though the gap between the two series may in small part be due to thc growth in real GDP, the general conclusion is clcan the link between ten-year averages of growth in Ml and inflation isn't tight enough lor the central bank to use Ml to achieve its desired rate of inflation.
From Ml to M3, and other monetary aggregates
Why is the relation between Ml growth and inflation not tighter? Here arc two reasons:
ВАОСЮ TOUCY chapLer 26
• Because of changes in the competitiveness of the financial system. Consider the increasing gap from 1986. In the mid-1980s thc financial system was deregulated, which resulted in many new bank
MONE1ARY POLICY: A SUMMING UP chapter 26
Figure 26.1 Ml and M3 growth and inflation:
ten-year averages.
Australia.
20
1910
—I—
1920
—I—
1930
1940
1950
I960
1970
1980
1990
2000
1911-2008
The relation between MI or M3 growth and inflation isn't tight, not even in the medium run, and definitely not since the 1980s.
entrants Irom within Australia and Irom abroad. Banks were allowed to pay interest on their deposits at any rate they liked As old and new banks competed lor market share, they also introduced innovations that made it cheaper and easier for customers to access their deposits electronically. I hus, the money stock grew considerably, even though the central bank decided to reduce inllation by raising interest rates Irom 1989. Increased linanciai system competitiveness can also help to explain the gap that emerged irom 1945 to 1956—the banks expanded considerably at a time when wartime wage and price controls were only slowly removed. • Because of shifts in the dew and for money. An example will help. Suppose that as thc result of thc introduction ol credit cards people decide to hold only hall the amount of money they held before,- in other words, the real demand tor money decreases by hall. In the medium ain. the real money stock must also decrease by hall. Fora given nominal money slock, the price level must double, liven il the nominal money stock is constant, there will be a period ol inflation as the price level doubles. During this period there will be no tight relation between nominal money growth (which is zero) and inllation (which is positive .
Thc reason why the demand lor money shilts over time goes beyond the introduction ol credit cards. To understand why, we must challenge an assumption we have maintained until now— namely, that there is a sharp distinction between money and other assets. In fact there arc many financial assets that are close to money. They cannot be used for transactions at least not without restrictions—but they can be exchanged for money at little cost. In other words, they arc very liquid this makes them attractive substitutes lor money. Shilts between money and these assets are the main factor behind shifts in the demand for money.
Thc presence of shifts between money and other liquid assets has led central banks to construct and report measures that include not only money hut also other liquid assets. These measures arc called monetary aggregates, and typically come under the names of M3, broad money and so on. In Australia. M3 is Ml (currency and current account deposits) plus all other deposits with banks by the non-bank public (such as term deposits that have an explicit maturity ol a few months to a few years, and a penally for early withdrawal . Broad money includes M.3 plus deposits hy thc non-bank public in non-bank financial intermediaries (such as building societies, cash management trusts, finance companies and other general financiers).
From equation (5.3) (the LM equation): the real money supply (the left side) must be equal to the real demand for money (the right side) M/P=YL(i). If, as a result
4 of the introduction of credit cards, the real demand for money halves, then MIP = YL(i)f2. For a given level of output and a given merest rate, M/P must also halve. Given M. this mplies that P must double.
4 In 2008. MI was equal to $233 billion, compared with $1035 billion for M3. and $1118 billion for broad money. (Note: There is no such thing as M2 in Australia.)
The construction ol these wider monetary aggregates would appear lo oiler a solution to our earlier problem. Il mosl ol the shifts in the demand tor money are between Ml and other assets within M.3, the demand for M3 should be more stable than the demand for Ml, and so there should be a tighter relation between M3 growth and inllation than between M I growth and inflation. II so, the central bank
could choose targets for M.3 growth rather than lor Ml growth. This is indeed the solution that many central hanks adopted. But il hasn't worked well cither, for two reasons:
• While the relation between M3 growth and inllation is tighter than the relation between MI growth and inllation in the two periods when M I growth predicted badly (1040 56 and 1986 onwards), it did worse when Ml growth performed well. This is seen in Figure 26.1. With neither Ml nor M3 growth dominating in all circumstances, ccntral hanks were never comfortable about the use ol growth targets lor monetary aggregates.
• More importantly while the ccntral bank can control M I it doesn t control M3. II people shilt Irom Treasury bonds lo term deposits, this will increase M3—which includes term deposits but doesn't include Treasury bonds. There is little thc central bank can do about this increase in M3. Thus, N13 is a strange target: it is not under the direct control of the central bank nor is it what the central bank ultimately cares about.
In Australia from 1977 to 1985 the RBA set itsell target ranges for conditional projections lor thc growth of M3, and it was marginally successful in achieving these. But once the impact of thc deregulation measures of the early 1980s began to take ellect, the RBA quickly realised that it would certainly fail il it continued to target any monetary aggregate. Instead, in 1985 thc RBA developed a checklist approach to monetary policy. This list contained everything lhat could conceivably matter lor Iraming monetary policy. The problem with this approach was that, although everything now mattered, in laci nothing really mattered in particular. This allowed too much discretion, and thc private- sector couldn t easily determine what to expect ot monetary policy. In thc next ten years thc RBA began to appreciate that its monetary policy would be more successful il it refined and communicated its medium-run objectives. By thc early 1940s it had adopted inflation targets which is the topic that we now take up.
Inflation targeting
In many countries, ccntral banks have dclincd as their primary goal the achievement ol a low inllation rate, both in the short run and in the medium run, I his is known as inflation targeting. The lirst country to adopt an explicit inllation target was New Zealand, in 1989. Canada followed in 1991, thc United Kingdom in 1992, Sweden and Australia in 1993, Finland in 1994, Spain in 1995, thc Czech Republic in 1997, Poland, Israel and Brazil in 199") Switzerland, South Africa Thailand, I lungary, Colombia and Korea in 2000, and Iceland, Norway and Mexico in 2001. Inflation targeting has to be dclincd in some time frame:
• Trying to achieve a given inllation target in the medium run would seem, and indeed is a clear improvement over trying to achieve a nominal money growth target. After all, in the medium run the primary goal ol monetary policy is to achieve a given rate of inflation. It is better to have an inflation rate as the target than a nominal money growth target, which, as we have seen, may not lead to the desired rate ol inflation.
• Trying to achieve a given inflation target in the short run would appear to be much more controversial. Focusing exclusively on inflation would seem to eliminate any role that monetary policy could play in reducing output fluctuations. But in lact this isn't the case. To see why, return to the Phillips curve relation between inflation, тт., lagged inflation, and the deviation ol the unemployment rate, it,, from the natural rate of unemployment u„ (equation 18.10]),-
IT, = 7Г, , - a[u, - u„)
Let thc inflation rate target be тг:. Suppose that the central bank could achieve its inflation target cxactly in every period. Then the relation would become
77' = 77; - Ot(u, - U„)
0 = -
lead the central hank to act in such a way as to eliminate all deviations of output from its natural level ol output.
The intuition: Il the central bank saw that an adverse demand shock was going to lead to a recession, it would know that, in the absence ol a monetary expansion, the economy would experience a decline in inllation below the target rate ol inllation. To maintain stable inllation, the central bank would then rely on a monetary expansion to avoid the recession. The same would apply to a favourable demand shock. Fearing an increase in inflation above the target rate the central bank would rely on a monetary contraction to slow down the economy and keep output at the natural level ol output. In short, as a result ol this active monetary policy, output would remain at the natural level of output all the time.
I he result we have just derived—that inflation targeting eliminates deviations of output from its natural level—is striking. But it is too strong, for two reasons:
■"TONFTARY POLICY: A SUMMING UP chapter' 26
• I he central bank cannot always achieve the rate ol inflation it wants in the short run. So, suppose- that, for example, the central bank wasn't able to achieve its desired rale ol inflation last year, so 77;_, is higher than — . Then it isn't clear lhat the central bank should try to hit its target this year, and
achieve
TT, - ll
The Phillips curve relation implies that such a decrease in inflation would require
a potentially large increase in unemployment. We return to this issue below. • Like all other macroeconomic relations, the Phillips curve relation above doesn't hold exactly. It will happen lhat lor example, inflation increases even when the unemployment is at its natural rale. In this case, the central bank will lace a more difficult choice: whether to keep unemployment at thc natural rate and allow inllation to increase, or to increase unemployment above the natural rate, to keep inllation in check
These qualifications are important, and even if valid, il remains true that an explicit inflation target is a clear signal ot the central bank's intention lor inflation, which helps people lo lorm their expectations about prices in the future. Thus, it helps to lock in expectations of inllation at that level, which then reduces the inflationary impact of any macroeconomic shocks.
Ihe general point remains: inllation targeting makes good sense in the medium run, and allows lor monetary policy to stabilise output around its natural level in ihe short run.
Interest rate rules
I "he next question is how to achieve the central bank's targets. Inllation is clearly not under the direct control ol the central bank. In answer to this question, in the early 1990s, lohn Taylor, in the Llnited States, argued that, since it is the interest rate thai directly affects spending, central banks should think in terms ol setting an interest rate rather than the money supply or its rate ol growth. He then suggested a rule that the central bank might want to follow. Taylor's rule goes as follows:
From Chapter 14: In the medium run. the real interest rate is given, equal to r„, so the nominal interest rate moves one-for-one with the inflation rate. If r„ is 2 per cent and the target inflation rate тг' is 2 per cent, then the neutral nominal interest rate i„ = 2% + 2% = 4%. If the target inflation rate is 0 per cent,
< then i„ = 2% + 0% = 2%.
4 Recall, from Chapter 4. that, through open- market operations, the central bank can achieve any short-term nominal interest rate it wants. Also recall how we applied various interest rate rules in Chapters 5. 7.9.14.15,17.20 and 21.
Let 77. be the rate of inflation, and тг1 he the target rate ol inflation.
(26.1)
Let i| be the nominal interest rate, and i„ be the medium-run nominal interest rate—thc nominal interest rate associated with the target rate ol inllation 77' in the medium run. Think of the central bank as choosing the nominal interest rate. 1. Let u, be the unemployment rate and 11,, be the natural unemployment rate. Then, Taylor argued, the central bank should follow a rule such as:
(ii77. - 77') - bin,
where a and l> are positive coefficients, and / is the neutral or medium-run value ol the nominal rate ol interest (which, recalling what you saw in Chapter 14, must equal r„ + 77r). Let's look ai what the rule says:
• It inflation is equal to target inflation ! 77. = 77 , and the unemployment rate is equal lo the natural rate of unemployment in, = u,. then the central bank should set the nominal interest rate, ;',, equal to its medium-run value:
" !„ = r„
This way, thc cconomy can stay on the same path, with inflation equal to thc target inflation rate and unemployment equal to the natural rate of unemployment. Obviously it is important that the central bank chooses /,, appropriately so that the economy can reach a medium-run equilibrium—for example, you saw how this worked out when we discussed liscal consolidation in Chapter 7.
• Il inflation becomes higher than the target (тг, > тг' \ thc central bank should increase the nomina interest rate, /.. above i,,. This higher interest rate will rcduce investment and thus aggregate demand and output, which, according to equation (7.2), will reduce inflationary pressures.
Thc coefficient a should therefore rcllcct how much thc ccntral bank cares about unemployment versus inflation. Thc higher a is the more the central bank will increase the interest rate in response to inliation the more the economy will slow down the more unemployment will increase, and thc laster inflation will return to the target inflation rate.
In any ease, as Taylor pointed out, a should be larger than one. Why? Because what really matters for spending is thc current real interest rate not the nominal interest rate. When inflation increases, the central bank will want to decrease spending and output, and so must increase the real interest rate. In other words, it must increase the nominal interest rate more than one-for-one with inflation.
• If unemployment is higher than thc natural rate of unemployment и > »,,), the ccntral bank should decrease the nominal interest rate. The lower nominal interest rate will increase output, leading to a decrease in unemployment. Like the coefficient a, thc coefficient l> should reflect how much thc ccntral bank cares about unemployment relative to inflation. The higher b is, the more thc central bank will be willing to deviate Irom target inflation lo keep unemployment close to the natural rate ot unemployment.
In stating this rule, Taylor didn t argue thai it should he followed blindly. Many other events, such as an exchange rate or credit crisis or the need to change the composition of spending on goods, and thus the mix between monetary policy and fiscal policy, justify changing the nominal interest rate for other reasons than those included in thc Rile. But. he argued, thc rule provided a useful way ot thinking about monetary policy. Once the central bank has chosen a target rate ol inflation, it should try to achieve it by adjusting the nominal interest rate. Thc rule it should follow should take into account not only currcnt inflation but also current unemployment.
Since it was first introduced, the Taylor rule has generated a lot ol interest, both from researchers and from central banks.
• Interestingly, researchers looking al the behaviour ol both the Fed in thc United States and thc Bundesbank in Germany «prior to joining the euro have found that, although these two central banks certainly didn t think of themselves as following a Taylor rule, this rule actually described their behaviour in the 1980s and 1990s quite well. Thc evidence suggests thai the l ed seems to have an implicit inflation target of about 3 per cent. Thc RBAs behaviour since 1991 is also captured well by a Taylor rule.
• Other researchers have explored whether it is possible to improve on this simple rule: for example, whether the nominal interest rate should be allowed to respond noi only to current inflation but also to expected future inflation.
• Yet other researchers have discussed whether central hanks should adopi an explicit interest rale rule and follow it closely, or whether they should use thc rule more informally and feci free lo deviate from the aile when appropriate
In general, most central banks have now shitted from thinking in terms ol nominal money growth lo thinking in terms of an interest rate rule. Whatever thc implications lor nominal money growth ol following such a nominal interest rate rule, these are increasingly seen as unimportant, both by the central bank and by financial markets. To take an example, which you saw in Figure 26.1, thc large rates of growth ol Ml and M3 observed in Australia since the mid-1990s didn't seem to trouble either thc RBA or the financial markets. Fhcv are seen as shifts in the demand for money, shilts that the RBA can accommodate without running the risk ol higher inflation. However there are some economists who suspect that the large uncontrolled growth rates ot monetary aggregates arc implicated in thc serious asset price inflation over the last twenty years.
Комментариев нет:
Отправить комментарий