The Great Depression was a worldwide phenomenon. International goods and financial capital trade were as important as they are today, and so the effects of the Great Depression in the United States and Europe were rapidly transmitted to countries as far away as Australia.
The Wall Street stock-price crash of 1929 was mirrored in Australia, with stock prices halving between 1929 and 1930, and this created a very pessimistic environment for investment. Figure I presents the key macroeconomic variables for Australia from 1925 to 1945—the unemployment rate, real GDP growth and the rate of inflation of the GDP deflator.
In the late 1920s, Australia suffered two periods of minor recession. Yet, until 1930, unemployment remained stable at between 5 per cent and 7 per cent. However, by 1929 price deflation had already begun. This deflation, coupled with a collapse in export earnings from the drop in the terms of trade, set in play the vicious downward spiral that led to the Great Depression. Nominal wages were cut by up to 10 per cent through the award system in 1931, but relative to price deflation it was too little and too late—and so the real wage rose dramatically. Firms and farms found that their sales were collapsing, and at the same time the real value of their costs was increasing. Not surprisingly, many went bankrupt, putting great pressure on the balance sheets of the banks that had made loans to them. In addition, the banks were faced with unprecedented withdrawals. The money supply fell rapidly. In 1931 the economy suffered an exceptionally large collapse in output. Unemployment soared to almost 20 per cent in 1932, and remained stubbornly high for the next six years, returning to pre-depression levels only after World War II began. As in the United States, we see that output growth resumed in 1933, but unemployment took much longer to correct.
Why wasn't the deflation corrected earlier? The reason is the same as thc one we gave for the United States.The Australian monetary authority was the Commonwealth Bank (the precursor to the RBA. which was established in 1959), and it failed to stem the decline in the money stock—from 1929 to 1932 the nominal money stock was allowed to fall by almost 20 per cent. The raal interest rate went from about 5 per cent in 1929 to almost 15 per cent for the next three years, thus savaging aggregate demand. Eventually, the monetary authority realised the importance of maintaining liquidity in the financial system, but by then it was too late— the Great Depression was set. From 1933 to 1941, the real money stock increased by 54 per cent, which helped to gradually ease the situation.
Unemployment
Output growth
Figure I
Unemployment, output growth and inflation. Australia, 1925-45
SOURCE M. Butlin. RBA Discussion Paper. 1977.Table GIO.
The gradual recovery from 1933 was helped along by a belated cheap money policy, by an expansionary fiscal plan coordinated by the state premiers, by the real depreciation of the Australian currency, and by the recovery of the United States and Europe.The key lesson for Australia (as for all other countries) was to never allow the economy to get into a deflationary spiral, with unemployment rising too high for too long.
23.3 THE JAPANESE SLUMP
From thc end of World War II to the beginning of the 1990s. Japan's economic performance was spectacular. From 1950 to 1973 thc average growth rate was almost 8 per cent per year. As in other OECD countries, the average growth rate decreased after 1973. But from 197.3 to 1991 it was still a very respectable 4 per cent per year, a higher rate than in most other OECD countries. As a result of this growth, Japanese output per capita measured in 1'1'P terms), which was equal to only 20 per ccnt of thc LIS level in 1950, had climbed to 82 per cent of the US level in 1990.
This growth came to an abrupt end in the early 1990s, l igure 2.3.9 shows the evolution of the Japanese growth rate since 1990. From 1992 lo 2002, average annual growth was less than 1 per ccnt— far below what it had been earlier. This long period of low growth is called the Japanese slump. This slump was obviously noi as sharp and as deep as the Great Depression (recall from Table 23.1 that the average annual growth rate in the United States Irom 1929 to 1932 was -8.6 per ccnt), but it was still substantial.
Think of it this way: if average output growth between 1992 and 2002 had remained the same as it was during 1973 to 1991, output in Japan would have been roughly 30 per cent higher in 2002 than it actually was.
Since 2003, the growth rate has turned positive. But, unlike in the United States' recovery from 1933 on, the growth rate in lapan is still much lower than it was before thc slump. In the best ol eases, it will take many years lor the Japanese economy to return to normal. Unfortunately, in 2008-09, Japan laced the grim prospcct ol thc global economic slowdown brought about by thc financial crisis.
See the discussion of the post-1973 growth slowdown in major OECD countries in Chapter 12. ►
Forgotten the definition ► of GDP in PPP terms! See Chapter 10.
Figure 23.9 The Japanese slump: output growth since 1990 (%)
Figure 23.10 completes the macroeconomic picture by showing the behaviour ot the unemployment rate and inflation (using the CPI) since 1990. Looking at the unemployment numbers, you might conclude that Japan had in fact not done too badly True, thc unemployment rate increased from 2.1 per cent in 1990 to 5.4 per cent in 2002 (it has since declined to around 4 per cent). But even 5.4 per cent is lower than thc average unemployment rate in the Llnitcd States over thc past forty years, and it is a rate that many European countries can only dream of achieving.
6 -|
5 -
4 - GDP growth
3 -
2- 1 -
0 - ,/Mean growth \ j
-1 - 2 -
1 1 Г 1990 1992 1 1 1 1 1 1
1994 1996 1998 1 1 1 1
2000 2002 1 1 1 1 1 1
2004 2006 2008
From 1992 to 2002, average GDP growth in Japan was less than I per cent SOURCE.-Bank of japan.
Figure 23.10 The Japanese slump:
unemployment and inflation since
1990 (%)
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008
Low growth in output has led to an increase in unemployment Inflation turned into deflation from 1998 until 2007. SOURCE: Bank of Japan.
The conclusion would be wrong, however. An unemployment rale of 5 per cent in lapan is a sign ol a very depressed labour market. Го see why, go back to thc discussion in the focus box 'Okun's law across countries and time in Chapter 9. Wc saw there that Japanese firms offer substantial employment protection to their workers. So, when these firms experience a decrease in production, < See the focus box 'The they tend to keep their workers, leading to a small ellect ol the decrease in output on employment. Japanese unemployment As a result, low growth has not led to a very large increase in unemployment. But it does not mean rae ln Chapter 8. that Japan is doing well.
Turn finally to the inflation numbers. Low growth and high unemployment (by Japanese standards) have led to a steady fall in the inflation rate. Since 1995, lapan has had deflation—negative inflation— something that had not been observed in OECD countries since the Great Depression. These numbers raise an obvious set ol questions:
• What triggered Japan's slump?
• Why has it lasted so long?
• Were monetary and liscal policies misused, or did thev iail?
• What are the factors behind the current recovery? These are the questions we take up in the rest ol this section.
The rise and fall of the Nikkei
The 1980s were associated with a stock-market boom in Japan. The Nikkei Index, a broad index of Japanese stock prices, increased from 7000 in 1980 to 35,000 at the end of 1989—a fivefold increase. Then, within two years, the index fell sharply—down to 16,000 at the end of 1992. For the rest of the decade, the index remained low. By mid-2003 it stood slightly below 9900. almost a quarter of its value at the peak.
2-
-2
Why did the Nikkei rise so much in the 1980s and then fall so quickly in the early 1990s? Recall from Chaptcr I 5 that there can be two reasons for a stock pricc to increase:
• A change in the fundamental value of the stock price, coming, lor example, Irom an increase in currcnt or future expected dividends. Knowing that the stock will pay higher dividends either now or in the future, investors are willing to pay more lor the stock today. Its price goes up.
• A speculative bubble. Investors buy at a higher price simply because they expect thc price to go even higher in the luture.
figure 23.1 I gives the evolution ol dividends and stock prices in Japan Irom 1480 to 2008. The upper line shows the evolution ol the stock price index for the whole Japanese stock market; the lower line shows the evolution ol the corresponding index for dividends. For convenience, both variables arc normalised to be equal to one in 1980. A look at the ligure yields a simple conclusion. While the stock- price index increased in the 1980s, dividends remained llat. This isn't proof that the increase in thc stock-price index was a bubble. Investors may have expected large increases in future dividends, even il current dividends weren't increasing. Hut it strongly suggests that the increase in Japanese stock prices had a large bubble component, and that the later tall was largely a bursting ot that bubble. This conclusion is given more weight when we see how dividends and stock prices moved closely together from 2003 to mid-2008 (only to collapse in late 2008 due to the global linanciai crisis.
Whatever its origin thc rapid fall in stock prices had a major impact on spending and, in turn, a big impact on output. Table 23.-I shows thc evolution of GDP growth, consumption growth and investment growth from 1988 to 1993. Investment, which had been very strong during the rise ol the Nikkei, collapsed. In contrast to the Great Depression—where consumption had fallen sharply alter the stock- market crash—consumption was less affected. Hut the strength in consumption wasnt enough to avoid a sharp decline in total spending and GDP growth, Irom 6.5 per cent in 1989 to 0.4 per cent in 1993.
Recall from Chapter 15 ► that in the absence of a speculative bubble the price of a stock is equal to the expected present value of future dividends.
See the discussion of the ► effects of stock prices on consumption and investment in Chapters 16 and 17.
In short, there is no mystery about how the Japanese slump started. The more difficult question is why it continued tor so long. After all. perhaps the main lesson Irom the Great Depression was lhat macroeconomic policies could and should be used to help the economy recover. Were they used in Japan? I: so, why did they fail? These are the next two questions we take tip.
The failure of monetary and fiscal policy
Monday policy was used. Hut ii was probably used too late, and when it was used it was faced with the twin problems of the liquidity trap and deflation discussed in Section 23.1.
The increase in stock prices in the 1980s and die subsequent decrease in the early 1990s were not associated with a parallel movement in dividends. From 2003 to mid-2008 the correlation between them was strong. SOURCE: Datastrcam (TOTMKJP).
Table 23.4 GDP, consumption and investment growth, Japan, 1988-93
GDP Consumption Investment
Year (%) (%) (*)
1988 6.5 5.1 15.5
1989 5.3 4.7 15.0
1990 5.3 4.4 11.5
1991 3.1 2.1 4.4
1992 0.9 2.2 -7.3
1993 0.4 2.5 -11.6
SOURCE: OECD Economic Outlook, December 2001. Investment is private, fixed and non-residen'jal
The point is made in Figure 23.12 which shows ihe evolution ol the nominal interest rale ihe call money rate : and the real interest rate in Japan from 1990 to 2(108. (Because we don't observe expected inllation, wc construct thc real interest rate as the nominal interest rate minus actual—rather than expected—future inllation. •
Thc nominal interest rate was high in 1991 close to 8 per cent. This was in part because the Bank of Japan olten referred to as the BoJ », worried about thc rise of the Nikkei, had tried to decrease stock prices by increasing the interest rale. With inflation around 2 per cent, this nominal interest rate implied i Recall that the stock a real interest rate of about 6 per ccnt.
As growth slowed down, thc Bank of Japan cut ihe nominal interest rate. But il did so slowly, and by I 996 the nominal interest rate was down to less than I per cent. The cumulative effect ol low growth was such that inflation had turned to consistent deflation by 1998. As a result, the real interest rale was higher lhan thc nominal interest rate.
price depends positively on current and expected future dividends, ard negatively on current and future interest rates.
From the mid- 1990s until l'><)8 lapan was, in effect, in a liquidity trap. At times, the nominal short- term interest raic was equal lo zero. At the lime ol writing October 20081. it had climbed to about
Japan was in a liquidity trap until 2008: the nominal interest rate was close to zero. Deflation implies that even at a zero nominal interest rate the real interest rate is positive. SOURCE Bank of Japan.
0.5 per cent! At the same time, unemployment remained high though stabilised at around 4 per ccnt, suggesting that thc natural rate had been increasing towards that level, and thus the dangers of further deflation seemed to have been reduced. Inllation was actually positive at 2 percent in 2008, largely due to the cost pressures of higher global commodity prices. At the time of writing, the real interest rate was actually negative, which may help to stimulate demand and increase output. But it is not likely to be enough to avoid the slowdown arising from the global financial crisis in 2008-09.
Fiscal policy was used as well. Figure 23.1 3 shows what happened to tax revenues and to government spending as a proportion of GDP from 1990 to 2008. It shows thc dccrcasc in taxes at thc start ot thc slump in 1992, and thc steady increase in spending throughout the decade—an increase of close to three percentage points of GDP. Much of this increased spending has taken the form of public work projects, many of them of doubtful usefulness. But, from the point ol view ol increasing demand, one project is as good as another, and so this increase in government spending should have contributed to an overall increase in demand.
Has it? The economists who have looked at this question have concluded that it has but that it was just not enough to increase spending and output. Put another way, in the abscnce of increased government spending, output would have declined even further. Could the Japanese government have done more? Probably not. Fligh government spending and low taxes have led to a long string of budget deficits and a steady accumulation of government debt. Thc ratio of net government debt to CI)P increased Irom I 3 per cent of GDP in 1991 to almost 100 per cent in 2008. With a current interest rate on government bonds very close to zero, interest payments on the debt arc small. But il the interest rate were to increase in the future, interest payments might represent a very heavy burden on the government budget. For this reason, the Japanese government is understandably reluctant to continue to use fiscal policy.
The recovery
A joke circulating in Japan is that, as a result of public works projects, by the end of the slump
the entire shoreline of ► the Japanese archipelago will be covered in concrete.
At an interest rate of I per cent per year, a 100 per cent debt-to- GDP ratio impiies interest payments equal to I per cent of GDP. At an interest rate of. say. 6 per cent per year. ► the same debt-to-GDP ratio implies interest payments equal to 6 per cent of GDP. a much higher interest charge for the government.
Output growth has been higher since 2003. and most economists cautiously predicted that the recovery will continue, notwithstanding the global slowdown in 2008-09 that is bound lo affect Japan. This raises the last question. What arc thc factors behind the recent recovery? There appear to be two main factors.
Figure 23.13
Japanese
government
spending and
revenues,
1990-2008
The budget deficit steadily increased in the 1990s, mostly as a result of increased government spending.
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008
SOURCE: Bank of Japan, ф
A regime change in monetary policy
In the strange world ol the liquidity nap. higher expected inllation is good. At a zero nominal interest rate, higher expected inflation implies a lower real interest rate. A lower real interest rate stimulates spending. Higher spending leads to higher output and lower unemployment. Lower unemployment through the Phillips curve raises actual inllation.
This suggests that, even il the nominal interest rate is close to zero and thus can hardly be reduced further, the central hank might still be able to lower thc real interest rate by affecting inflation expectations. This may not be easy to do. Suppose that the central bank announces an inflation target, a rate of inflation it will try to achieve over the next lew years. II people believe thc announcement, then expected inflation will indeed increase, helping the economy get out of the slump. But, it people don't believe the announcement and continue to expect deflation, then deflation will continue. i Note the symmetry
Therefore, the advice given to the Bank of Japan by many economists during the second halt of thc 1990s was that it should try to influence and increase inflation expectations. At worst, this wouldn't work; at best, this might gel thc economy out ot the slump. In 2003 the new chairman ot the Bank of Japan decided to follow this advice. He announced thai lhe bank was now committed to keeping lhe nominal interest rate equal to zero until there was strong evidence of sustained inflation. At the same time, the bank acted to flood financial markets with liquidity, lust like in 1933 in the United Slates, this statement and the action was perceived as a signal of regime change in monetary policy, and it appears lo have changed inflation expectations. Although the current inflation rate is still negative, inflation is now expected to become positive in the future, and the long-term real interest rate has fallen. This appears to be one ol thc lactors behind the strong increase in investment spending since 2003.
The clean-up of the banking system
It became clear in the 1990s thai the banking system in Japan was in trouble. Largely as a result of thc slump in output, many firms were doing poorly, and banks carried on their books many bad loans, loans the borrowers weren't able to repay. (Why this was. and how it happened, is discussed in the focus box Thc Japanese banking problem . •
Many bad firms—firms thai were incurring losses and should have closed -continued to be financed by the hanks, and so continued to operate. At the same time, as a large proportion ot bank financing continued lo go to the firms with bad loans, good firms'—firms with good prospects and good investment projects—couldn't lind financing and thus couldn't invest. In short, bad loans further depressed investment spending.
In such a case, lhe appropriate policy is clear. Banks thai have made too many bad loans should be forced to either close or restructure. Firms that cannot pav their loans should be forced to do thc same. These measures achieve two goals.
1 They eliminate the had lirms. leading eventually—as these firms are replaced by more productive
ones—to higher productivity and lo a higher natural level of output. 2. They allow firms with good investment projects to find the funds they need to invest, leading to an increase in investment spending and therefore to an increase in demand and output.
It is equally clear, however, that such a policy is politically very risky. Restructuring or closing firms and banks leads initially to layoffs, a politically unpopular outcome, especially when unemployment is already high. For this reason, not much was done to solve the banking problem in Japan in the 1990s. Banks continued to lend to had firms and thc proportion of bad loans steadily increased. Since 2002, however, the government has put increasing pressure on banks lo reduce bad loans, and banks, in turn, have put increasing pressure on bad firms to restructure or close. The proportion of bad loans has begun to fall, and good lirms have been increasingly able to finance investment. This is another factor behind the strong increase in investment spending since 2003.
with our discussion in Chapter 9 of whether a central bank can achieve d sinflation at little or no output cost.The answer there was that if the central bank can credibly convince people that inflation will be lower, it may be able to achieve lower inflation at little output cost
A number of other lactors are also helping Japan lo recover. In recent years the Bank of Japan has been intervening in foreign exchange markets, purchasing LIS Treasury bills lo prevent the yen from
strengthening This has kept Japanese exports competitive in world markets. And critically, strong output growth in the rest ol Asia particularly in China, has led to strong export growth in Japan. This is because about 40 per cent ol lapan's exports go to other Asian economies, and about 15 per cent to China. liven if export growth was lo lall, however the regime change in monetary policy, coupled with thc clean-up ol the banking system implies that domestic spending might continue to increase enough lo sustain growth in the luture. This is why most economists were more optimistic in 2007 about future growth in Japan than they had been for a decade. LInlortunately, the country is likely to suffer in the expected 2008-09 global slowdown.
BOX
1.
2.
The lessons learnt from the long slump in lapan may well help policy-makers everywhere to devise more effective policy responses to the 2008 linancial crisis.
Like the Great Depression, the sharp decrease in output growth in Japan in the early 1990s left many firms unable to repay their bank loans.
The situation was made worse by two facts:
In the 1980s, the banks had started losing their best borrowers—the large Japanese firms, which increasingly financed themselves by issuing bonds rather than by borrowing from banks. As a result, banks had made loans to more risky borrowers, some of whom would have a hard time repaying their loans even in the absence of a slump.
Also, for many of the loans, firms had used land as collateral. ('Collateral' stands for any asset that the borrower promises to give to the bank if the loan isn't repaid.) And, together with stock prices, land prices collapsed in the early 1990s, decreasing the value of that collateral.
In the Great Depression, bad bank loans had led to a series of bank failures and bank runs (see the focus box 'Bank runs' in Chapter 4.) Indeed, one of the lessons of the Great Depression was that, to avoid such bank runs, governments should find ways to ensure that depositors had insurance from bank failures or runs. Federal deposit insurance was introduced in 1934 in the United States. And a similar insurance system was later put in place in most countries, including Japan.The issue of government insurance of bank deposits is alive and well as a response to the chaos in financial markets in 2008.Treasuries around the world, including Australia, have rushed into providing guarantees of deposits to try and encourage financial institutions to resume normal lending.
Deposit insurance solves one problem—it eliminates the risk of bank runs. But it creates other problems, which were in evidence in Japan in the 1990s.To understand what these problems are. think of a bank that has the balance sheet described in Table I.
• On the asset side, it has one loan at $ 100,
• On the liability side, it has $50 in deposits.
• The net worth of the bank, the difference between assets and liabilities, is therefore $ 100 - $50 = $50. (Note two differences with the balance sheets we studied in Chapter 4. First, we ignore reserves.They
were important for the arguments developed in Chapter 4; they are not important here. Second, we assumed in Chapter 4 that assets were equal to liabilities, that net worth was zero. Net worth is typically positive, and this plays an important role here.)
Table 1 The bank's balance sheet
Assets Liabilities
Loan: $100 Deposits: $50
Net worth: $50
Now suppose the loan turns bad: the firm to which the loan was made cannot pay any of it back. What should happen?
• The value of the loan is now zero.The bank should write off the loan.The bank still owes $50 in deposits, but cannot pay them back.Thus. deposit insurance should pay $50 to the depositors, and the bank should close.
• But that is unlikely to happen. To keep his job, the manager of the bank may pretend that nothing has happened and the loan is still good. Indeed, the manager may decide to lend further to the firm, so the firm can pay off the old loan, and it looks like business as usual.This is clearly throwing good money after bad, but by doing so the manager buys time and keeps his job, at least for the moment.
• Even the owners of the bank may go along. If the bank closes now. they lose everything. (Net worth is clearly equal to zero.) If there is the slightest chance that the firm will recover and be able to pay the loan back, the bank may end up with positive net worth. (This is known as 'gambling for resurrection .) So, even if the odds are very bad. the bank may continue to lend to the firm.
• Depositors don't care what the bank does.Their deposits are insured, whatever the bank does. Even the regulator, if there is one. may prefer to close his eyes. Acknowledging the existence of bad loans, and the fact that the bank must be closed, may reflect badly on him—again, better to wait.
The result is that banks are likely to renew the bad loans, or even to make new loans to the bad firms, and to do so at the expense of good firms. And so, the more time passes, the worse the problem of bad loans becomes.
This is exactly what happened in Japan in the 1990s. Until 1993, the banks didn't disclose any information about bad loans. Since then, they have reluctantly acknowledged the presence of bad loans on their books. And the total amount of bad loans (self-reported by banks) has steadily increased, from 12 trillion yen in 1993 to 30 trillion yen in 1998 and 44 trillion yen in 2001. But even that amount may have been far below the true number. An estimate by the OECD put the total value of bad loans in 2003 at 237 trillion yen. or close to 45 per cent of GDP.
Progress has since been made. Bank lending increased in 2005 for the first time since 1998.The major Japanese banks improved their balance sheets, and it is hoped that the clean-out that it has already undertaken will mean that it will be less vulnerable to the global financial crisis of 2008 than it might otherwise have been.
SUMMARY
• In general, a recession leads to a decrease in inllation. Given nominal money growth the decrease in inllation leads to an increase in real money growth, decreasing the nominal interest rate and increasing output back to its natural level.
• One reason why this adjustment hack to the natural level of output may fail is that the decrease in inflation may lead lo an increase in the real interest rate. If expected inflation decreases more than the nominal interest rate, the real interest rate will increase. As spending depends on the real interest rate, lhe increase in the real interest rate will lead to a further decrease in output.
• Monetary policy can be used to decrease lhe nominal interest rate further, which helps to increase output. Monetary policy, however, cannot decrease the nominal interest rale below zero. When this happens, the economy is said to be in a liquidity trap.
• The combination of the liquidity trap and deflation can transform a recession into a slump or a depression. It the nominal interest rate is zero, and the economy is experiencing a deflation the real interest rate is positive and may be too high to lead lo an increase in spending and output. Output may continue to decline, leading to higher deflation a higher real interest rate, and so on.
On the Great Depression in the United States
• The unemployment rate increased Irom 3.2 percent in 1429 to 24.9 percent in 1933.
• Thc initial cause ol this increase in unemployment was a large adverse shilt in demand, brought about by the stock-market crash of 1929 and the resulting increase in uncertainty about thc luture.
• The result of high unemployment was a large deflation from 1929 to 1933.
• The favourable effect of the decrease in thc price level on the real money stock was ollset, however, by a roughly equal decrease in nominal money. This decrease in nominal money was due to bar,к failures and a decrease in the money multiplier. Thc main ellect ol dellation was a large increase In the real interest rate, leading to a further decrease in demand and output.
• Recovery started in 1933. Average growth was high, 7.7 per ccnt per year from 1933 to 1941. Unemployment decreased, but it was still equal to 9.9 per ccnt in 194 I. In contrast to thc predictions ol the Phillips curve, deflation turned to inflation from 1934 on, despite a very high unemployment rate.
• Many questions remain about the recovery. What is dear is lhat high nominal money growth, leading to high real money growth was an important lactor in the recovery.
On the Japanese slump
• After a long period of very high growth, lapan has had very low growth since 1992. This long period of low growth is callcd the Japanese slump.
• Thc slump was triggered by thc lall ol lapancse stock prices at the end of the 1980s, which led to a sharp decrease in investment spending, and in turn to a decrease in output.
• Monetary policy was used to try to increase output in thc 1990s. But Japan got itself into a liquidity trap, with the nominal interest rale very close to zero. Because Japan was experiencing deflation, the real interest rate was positive.
• Fiscal policy was also used to try to increase output in the 1990s. But after more than a decade of deficits government debt has increased to 100 percent of C.DP, and the lapanese government was reluctant to increase its debt further.
• Great Depression, 535
• New Deal. 540
• Turning dellation into inflation would decrease the real interest rate and help thc lapancse economy to recover. From 2003 until mid-2008 there was reasonable output growth. The main lactors behind this are a change in the monetary policy regime, strong export growth and an improvement in the banking system.
KEYTERMS
• depression, 527
• slump, 527
• liquidity trap, 532
QUESTIONS AND PROBLEMS
Quick check
I. Using the information in this chapter, label each of the following statements 'true', 'false' or
'uncertain'. Explain briefly.
a. The stock market crashed in 1929 because financial investors realised that the Great Depression was coming.
b. Ccntral banks could have done more to prevent, or at least to limit, thc scope of thc Great Depression.
c. The decrease in the unemployment rate explains why deflation ended during the Great Depression.
d. The Japanese slump ol the 1990s and 2000s was triggered by the sharp lall ol Japanese slock priccs at the end of the 1980s.
c. The Japanese central hank can help the Japanese economy to recover by keeping inflation very low.
2. Active monetary policy
a. Consider an economy with output below the natural level and a nominal interest rate equal to zero. Illustrate your answer in an IS-LM diagram.
b. Again suppose that output is below the natural level This time, however, assume that the central bank does not change monetary policy. Under normal circumstances, how does the economy return to its natural level of output? Illustrate your answer in an IS-LM diagram.
c. Considering your answer to part (b), if the ccntral bank docs nothing, what is likely to happen to expected inflation? How does this change in expected inflation allect the IS-LM diagram? Does output move closer to the natural level?
d. Consider the following policy advice: 'Because the economy always returns to the natural level ot output on its own, the central bank docs not need to concern itscll with recessions.' Do your answers to parts i a > to (c) support this advice?
3. Monetary versus fiscal policy
a. Consider an economy with output below the natural level and a nominal interest rate equal to zero. Illustrate this cconomy in an IS-LM diagram.
b. Under normal circumstances, how docs the cconomy return to thc natural level ol output? Docs this adjustment mechanism work when thc nominal interest rate equals zero?
c. Suppose that the central bank wants to use monetary policy to return the cconomy to its natural level of output? Can it do so when the nominal interest rate is equal to zero? What happens if the central bank tries to use expansionary monetary policy? Illustrate your answer in an IS-LM diagram.
d. In principle, can fiscal policy he used to restore the economy to its natural level of output when thc nominal interest rate equals zero? II so, explain how the appropriate policy affects output. II not. explain why not.
e. Consider the following policy advice: 'Sincc the central bank can act to keep the economy at the natural level ot output, the government should never use fiscal policy to stimulate the economy. Do your answers to this question support this advice?
Dig deeper
4. LoH^-ffrm unemployment ami the natural rate
Suppose that the markup is 5 per cent, so that the price-setting equation is given by
l_
P ~ I + 0.5 Suppose that the wage-setting equation is given by
VV
— = I - (MS + 0,5UL)
where us is the ratio of the number of short-term unemployed to the labour force, ami uL is the ratio of the number of long-term unemployed to the labour force. Suppose further that the proportion of unemployed who are long-term unemployed is equal to [3, so U| = (5u and u$ = (1 - (3)u where u is the overall unemployment rate and 0 < p < 1 ■
a. According to thc wage-setting equation, which type of unemployment has a greater impact on wages—long-term or short-term? Docs this make sense to you? Explain.
b. Derive thc natural rate of unemployment. (Hint: Substitute iij (in and iu I - /3)н in the wage-setting equation. The natural rate will depend on fi.)
c. Calculate the natural rate il j8 = 0Д- 0.4,- 0.8. Explain.
5. Long-term unemployment ami inflation Recall equation t8.b) in Chapter 8:
7Г,- 7T,_| = i PL + z) - «и,
a. Interpret the equation. Why does higher unemployment lead to lower inllation given past inflation? Draw the change in inflation against the unemployment rate.
/1s in problem 4, write thc overall unemployment rate, u, as u - us + U|_. with us lhe short-term unemployment rate IIhe ratio of lhe short-term unemployed to the labour forcel and U| the long- tern: unemployment rale (the ratio of the long-term unemployed to the labour force). Using this definition, substitute for u( in equation 18.6).
b. Now assume that the long-term unemployed have no effect on wage bargaining. Show how the equation above should be modified to fit this assumption.
c. Assume further that i/f = fiu and its = (I - j8ti( as in problem 4(b). Substitute the appropriate- equality into your revised equation in part
d. Suppose that the proportion ol long-term unemployed in unemployment increases that is, fi increases). Show what happens to the line you drew in part d).
e. Look at the line you drew in part (e). Suppose that policy-makers want to reduce inflation (so that the desired change in inflation is negative). Il the proportion ol long-term unemployed increases, how will this allcct the unemployment rate required to achieve a given reduction in inflation? Does the cost ot disinflation increase or decrease?
f. Now suppose that policy-makers lear disinflation and are hoping that thc inflation rate will increase. Does an increase in thc proportion of long-term unemployed tend to imply a larger or a smaller increase in inflation? How might your analysis be relevant to our understanding ol the recovery from the Great Depression?
Explore further
6. Japan's future
As discussed in Chapter 15, the yield curve will slope up (down) when financial market participants expect short-term interest rates to increase Idecrease) in the future. Co to the website of The Economist magazine lwww.economist.com). Find the 'Markets &' Data' section. In the section 'Economic and Financial Indicators', you should be able to obtain three-month, two-year and ten- year interest rates for Japan.
a. What is the three-month interest rate in Japan? Is it slill near zero? What is the two-year interest rate in Japan? Does the ditlerence between the two-year and three-monih interest rates suggest thai financial market participants expect Japan to escape the liquidity trap in thc near future?
b. Now compare thc ten-year interest rate in Japan with thc three-month interest rate. Does the dillerence between these two interest rates suggest that investors expect that Japan will escape the liquidity trap within a decade?
c. Now look at the most recent data on inflation and growth lor Japan (probably under the title Output Prices, and lobs' on The Economist website) and the forecasts for these variables lor the near future Do these data suggest that Japan will escape the liquidity trap soon?
d. Now look at data for economies other than Japan. Consider advanced and emerging economies. Is inflation low or negative ior forecast to be low or negative) in economics other than Japan? II so, look at the growth rates and short-term interest rates for these economies. Are any ol these economies close to a liquidity trap?
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 issues in this chapter.
FURTHER READINGS
CHAPTER ф
High Inflation
I
n 1913 the value of all currency circulating in Germany was six billion marks.Ten years later, in October 1923. six billion marks was barely enough to buy a one-kilogram loaf of rye bread in Berlin. A month later, the price had increased to 428 billion marks.
The German hyperinflation of the early 1920s is probably the most famous hyperinflation. (Hyperinflation simply means very high inflation.) But it isn't the only one.Table 24.1 summarises the seven major hyperinflations that followed the world wars. They share several features.They were all short (lasting a year or so) but intense, with inflation running at 50 per cent per month or more. In all. the increase in the price level was staggering. As you can see, the largest price increase was actually not reached during the German hyperinflation but in Hungary after World War II. What cost one Hungarian pengo in August 1945 cost 3,800 trillions of trillions of pengos less than a year later.
Table 24.1 Seven hyperinflations of the 1920s and 1940s
Country Beginning End P,IPo Average monthly inflation (%) Average monthly money growth (%)
Austria Oct. 1921 Aug. 1922 70 47 31
Germany Aug. 1922 Nov. 1923 1.0 310'° 322 314
Greece Nov. 1943 Nov. 1944 4.7 3I06 365 220
Hungary 1 Mar. 1923 Feb. 1924 44 46 33
Hungary 2 Aug. 1945 Jul. 1946 3.8 3I027 19,800 12,200
Poland Jan. 1923 Jan. 1924 699 82 72
Russia Dec. 1921 Jan. 1924 1.2 310s 57 49
PT/P(> Price level in the lost month of hyperinflation divided by the price level in the first month.
SOURCE: Philip Cagan. 'The moneta-y dynamics of hyperinflation', in Milton Friedman (od.). Studies in the Quanlty Theory of Money (Chicago: University of Chicago Press. 1956).ТаЫе I. © University of Chicago Press (1958).
Such rates of inflation hadn't been seen before. Today inflation is low in nearly all countries. The only exception is Zimbabwe. The rate of inflation has been growing exponentially for the last few years. In mid-2008 the Reserve Bank of Zimbabwe reported an inflation rate of about 2.2 million per cent per year. Here is a note printed sometime in early 2008. Note that it has an expiry date of 30 June 2008. It's unlikely that it would have reached that date with any value other than the paper it was printed on.
In a desperate but hopeless move in August 2008, the government changed the currency denomination by simply removing ten zeroes from all monetary values. But by October 2008 local newspaper reports indicated that the inflation rate had reached 219.8 million per cent per year. This is about 600.000 per cent per day. Even your computer will struggle to compute how much more something will cost after twenty-four hours have passed. Zimbabwe's currency is out of control.
Many Latin America countries experienced high inflation in the last quarter of the twentieth century.
From January 1984 to September 1985. Bolivian inflation averaged 40 per cent per month, implying a roughly thousand-fold increase in the price level over twenty-one months. (With an inflation rate of 40 per cent per month, the price level at the end of twenty-one months is (1 + 0.4)2' = I,I7I times the price level at the beginning.) Table 24.2 gives average monthly inflation rates for four other Latin American countries since 1976. All four had at least five years with average monthly inflation running above 20 per cent a month. Both Argentina and Brazil had monthly inflation rates in excess of 10 per cent a month for more than a decade. All four countries have now returned to low inflation. Inflation has even turned to deflation in the case of Argentina.
What causes hyperinflations? You saw in Chapter 9 that inflation ultimately results from or is associated with nominal money growth.The relation between nominal money growth and inflation is confirmed by the last two columns of Table 24.1. Note how. in each country, high inflation was associated with correspondingly high nominal money growth.Why was nominal money growth so high? The answer turns out to be common to all hyperinflations: nominal money growth is high because the budget deficit is high.The budget deficit is high because the economy is affected by major shocks that make it difficult or impossible for the government to finance its expenditures in any way other than money creation.
In this chapter we look at this answer in more detail, relying on examples from various hyperinflations.
• Section 24.1 looks at the relation between the budget deficit and money creation.
• Section 24.2 looks at the relation between inflation and real money balances.
• Section 24.3 puts the two together, and shows how a large budget deficit can lead to high and increasing inflation.
• Section 24.4 looks at how hyperinflations end.
10000000 RESERVE BANK OF ZIMBABWE Q00 000
Pi) ibe Ьсдга on iln-inH
TEN MILLION \
DOLLARS |
on or before 30th June 2008
Cot tm ban Uackuf
Щ /
Issue itttr. 1st Itauiy 200*
BEARER CHEQUE
Of . • • k AJ584 91 22
10000000
©
• Section 24.5 draws conclusions from our two chapters on pathologies—depressions and slumps in Chapter 23, and high inflation in this chapter.
Table 24.2 High Inflation in Latin America, I976-2000
Average monthly inflation rate (%)
1976-1980 1981-1985 1986-1990 1991-1995 1996-2000
Argentina 9.3 12.7 20.0 2.3 0.0
Brazil 3.4 7.9 20.7 19.0 0.6
Nicaragua* 1.4 3.6 35.6 8.5 0.8
Peru 3.4 6.0 23.7 4.8 0.8
* Nicaragua: last column is for 1996-99. SOURCE: IMF. International Financial Statistics, various issues.
24.1 BUDGET DEFICITS AND MONEY CREATION
A government can finance its budget delicit in one ol two ways:
• It can borrow, the way yon would. We borrow by taking a loan. Governments borrow by issuing bonds.
• It can do something that you cannot do. It can, in effect, finance the deficit by creating money. We say in effect because, as you will remember Irom Chapter 4, governments don't create money,- the central bank creates money, lint with the ccntral banks cooperation, the government can. in ellect, linance itscll by money creation. Il can issue bonds and ask the central bank to buy them. Thc central bank then pays the government with money it creates, and the government uses that money to finance its deficit. This process is callcd debt monetisation.
Most of the lime, and in most countries, dclicils arc financed primarily through borrowing rather
than through money crcaiion. But, at ihe siari of hyperinflations, two changes usually lake place:
• There is a budget crisis. I he source is typically a major social or economic upheaval. It may be a civil war or a revolution that destroys the state's ability to collect taxes. This was ihe case, for example, in Nicaragua in thc 1980s.
It may come, as in the case ol thc hyperinflations alter World War I and World War II. Irom the aftermath of a war that leaves ihe government with both smaller tax revenues and thc large expenditures needed lor reconstruction. This is what happened in Germany in 1922 and 1923. Burdened with payments called war reparations' that it had to make to the Allied forces. Germany had a budget deficit equal lo more than two-thirds of its expenditures.
It mav come Irom a large adverse economic shock—lor example, a large decline in the price of a raw material that is the country's major export and its mam source of revenues. As you will see in ihe focus box entitled The Bolivian hyperinflation of thc 1980s this is what happened in Bolivia in thc 1980s. Thc decline in the price of tin Bolivia's principal export was one of the main causes ol Bolivian hyperinflation.
It may come from a bad policy decision. This is ihe case for Zimbabwe, where thc decision to redistribute land away from white farmers in 2000 led to a catastrophic decline in agricultural output, and, in turn, a large fall in GDP and a large increase in the budget deficit.
• The government becomes increasingly nimble to borrow from the public or from abroad to finance its deficit. I he reason for this is the size of thc deficit itscll. Worried that the government may not be able lo repay the debt in the future, potential lenders start asking ihe government lor higher and higher interest rates. Sometimes, loreign lenders decide lo slop lending to ihe government altogether. As a result, thc government turns increasingly lo the other source ol finance- money creation. Eventually, mosi ol the deficit is financed by money creation.
How large is thc rale ol nominal money growth needed to finance a given amount ol revenues?
• Let M be the nominal money stock, measured, say, ai the end ol cach month. In thc case ol hyperinflation, things change so quickly thai it is useful to look ai whai happens from month to month, rather than trom quarter to quarter or Irom year to year.) Let AM be ihe change in the nominal money stock from the end of lasi month lo thc end ol this month—nominal money creation during thc month.
We are taking a shortcut here.What should be on the right-hand side of the equation is the change in the monetary base—the money created by the central bank—not the change in the money stock (which includes both currency and current account deposits).We ignore the distinction here;
it doesn't play an important role in the argument that follows. ►
• Thc revenue, in real terms that is in terms of goods), thai the government generates by creating an amount of money equal to AM is therefore equal to A.M/P—nominal money creation during the month, divided by the price level. This real revenue Irom money creation is called seignorage. Thc word is revealing. Thc right to issue money was a precious source ol revenue for the seigneurs' of the past. They could buy the goods they wanted by issuing their own money and using il to pay for the goods.
(24.
Wc can summarise what we have just learned by writing
AM
Seignorage =
Seignorage is equal lo money creation, divided by the price level. To see what rate of nominal money growth is required to generate a given amount ot seignorage, note that we can rewrite ЛМ/Р as
A,VI A,VI M_
p = ~лГ p
In words: We can think of seignorage, AM I', as the product of the rale of nominal money growth
AM/M, times real money balances, M/P. The larger the real money balances held in the economy, the Real money balances' is
larger the amount ol seignorage corresponding to a given rate ol nominal money growth Replacing this ji-st another name for
expression in equation (24.1) gives the real топеУ stock-
AM M
Seignorage = —— — (24.2'
M P
a flow. 1 here is real income.
This gives us the relation we wanted between seignorage the rate ol nominal money growth and real money balances. To think about relevant magnitudes it is convenient to divide both sides of equation (24.2) by real income, У measured at a monthly rate): < Remember income is
Seignorage AM M
- "Г ' M ~PY (24 3)
Suppose that the government is running a budget delicit equal to 10 per ccnt of real income and decides to finance it through seignorage so deficit У = seignorage /У =01. Suppose that people hold real balances equal to two months of income, so МЛРУ 2. This implies that nominal money growth must satisfy
AM AM
= ~м x 2 ^Г " ()'05
To tinancc a delicit ol 10 per ccnt ol real income through seignorage, the monthly growth rale ot nominal money must be equal to 5 per cent.
Does this imply that the government can tinancc a deficit equal to 20 per cent of real income through a rate ol nominal money growth of 10 per cent, a delicit of 4(1 per cent ot real income through a rate of nominal money growth of 20 per cent and so on? No. As nominal money growth increases so does inflation. And. as inllation increases, the opportunity cost ot holding money increases, leading people to reduce their real money balances. In terms of equation 24.2 an increase in nominal money growth AM/M. leads lo a decrease in real money balances, M/P. so that an increase in nominal money growth will generate a less proportional increase in seignorage. What is crucial here is how much people adjust their real money balances in response to inflation, and ii is thc issue lo which we lurn next
24.2 INFLATION AND REAL MONEY BALANCES
What determines thc amount of real money balances that people arc willing to hold: And how does ihis amount depend on nominal money growth?
Let's go back to the LM relation derived in Chapter 5:
M
у = wo (-)
Higher real income leads people m hold larger real money balances. A higher nominal interest rate increases ihe opportunity cost of holding money rather than bonds and leads people to reduce their real money balances.
This characterisation holds in hoth stable economic times and times of hyperinflation. Hut in times of hyperinflation wc can simplify it further. I lere's how:
Recall, from Chapter 14. ► that r = i- it'. Equivalently. / = r + тг'.
In describing the Austrian hyperinflation of the 1920s. Keynes noted:'In Vienna, during the period of collapse, mushroom exchange banks sprang up at every street corner, where you could change your krone into Zurich francs within a few minutes of receiving them, and so avoid the risk of loss during the time it would take you to reach your usual bank.' ►
One of the hopes of the European Union is that the euro will replace the US dollar as the foreign currency of choice. (Why would the European Union want this to happen!) If it happens, we may have to speak of Euroisation' rather than 'dollarisation'.
This decrease in real money balances explains why. in Table 24.1, average inflation is higher than average nominal money growth in each of the seven postwar hyperinflations.
That real money balances. MIP. decrease during a hyperinflation implies that prices. P. > must increase more than M—average inflation must be higher than average nominal money growth.
• hirst, rewrite the LM relation using the relation between the nominal interest rate and the real interest rate, / ^ r + it''-.
M
т -YUr
Real money balances depend on real income. V, on thc real interest rate, r, and on expected inflation, 7r'".
Second note that while all three variables (V, r and тг') are likely to vary over time during a hyperinflation, expected inflation is likely to move much more than the other two variables. During a typical hyperinflation, actual inflation—and, presumably, expected inflation may move from 0 per cent to 50 per cent a month or more.
(24.4)
if)
So, it isn't a bad approximation to assume that both income and the real interest rate are constant, and focus just on thc movements in expected inllation. St), we write
M
-J -
(
)
where the bars over У and r mean that we now take both income and the real interest rate as constant. In times of hyperinflation equation (24.41 tells us we can think ol real money balances as depending primarily on expected inflation. As expected inflation increases and it becomes more and more costly to hold money, people will reduce their real money balances.
During a hyperinflation, people indeed lind many ways of reducing their real money balances. When the monthly rate of inllation is 100 per cent, lor example, keeping currency lor a month implies losing halt ot its real value because things cost twice as much a month later . Barter the exchange ol goods lor other goods rather than for money, increases. Payments for wages become much more frequent—often twice weekly. Once people arc paid, they rush to stores to buy goods. While the government often makes it illegal to use other currencies than the one it is printing, people shilt to foreign currencies as stores of value. And even if it is illegal an increasing proportion of transactions take place in loreign currency. During the Latin American hyperinflations ot the l9X0s. people shifted to US dollars. The shift to LIS dollars has become so widespread in the world that it has a name: dollarisation (the use ol LIS dollars in another country's domestic transactions -.
By how much do real money balances actually decrease as inflation increases? Figure 24.1 examines the evidence from the Hungarian hyperinflation of the early 1920s and provides some insights.
• Panel (a) plots real money balances and the monthly inflation rate Irom November 1922 to February 1924. Note how movements in inflation are reflected in opposite movements in real money balances. The short-lived decline in Hungarian inflation from luly to October 1923 is reflected in an equally short-lived increase in real money balances. At the end ol the hyperinflation in February 1924, real money balances are roughly hall what they were at thc beginning.
• Panel (b) presents the same information as panel 'a), but in the form of a scatter diagram. It plots monthly real money balances on the horizontal axis against inflation on the vertical axis. (We don't observe expected inflation, which is the variable we would like to plot, so we use actual inflation instead.) Note how the points nicely describe a downward-sloping demand for money. As actual inflation—and, presumably, expected inflation as well—increases, the demand lor money strongly decreases.
To summarise, increases in inflation lead people to decrease their use ol money, and lead to a decrease in real money balances.
Figure 24.1 Inflation and real money balances in Hungary. November 1922 to February 1924
0.300 -,
0.275 -
250-
225'
Real money balances (M/P) (scale at left)
—I 1 1 1 1 1 1 1 1 1 1 1 1 1 1
Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb
1922 1923 1924
-10
-2
(a)
io.
о a:
0.200 - 0.175-
70- 60-
Г 50-
>
! 40- I 30- Г 20- 10' 0'
(Ь)
о 2.
-10- 0.175
-r
1
0.300
-T-
0.200
0.225 0.250
Real money balances
0.275
(24.5)
AM M
At the end of the Hungarian hyperinflation, real money balances stood at roughly half their pre-hyperinflation level.
24.3 DEFICITS, SEIGNORAGE AND INFLATION
The lirsi line repeats equation 12-1.2'. Seignorage equals the rale ot nominal money growth times real money balances. And the second line replaces real money balances by their expression in terms of expected inllation, Irom equation 1 24.4
Using ihe relation between seignorage, the rate ol nominal money growth and the expected rale ol inflation implied by the second line ol (24.5 we can now show how the need lo finance a large budget delicit through seignorage can lead not only lo high inflation but also, as is ihe case during hyper¬inflations, to high ami increasing inflation.
The case of constant nominal money growth
Suppose that the government chooses a constant rate ol nominal money growth and maintains lhal rate forever. (Clearly this isn't what happens during hyperinflations, where the rate ol nominal money growth typically increases over the course of the hyperinflation.- we will get more realistic laier. • I low much seignorage will ihis constant rate ol nominal money growth generate?
Recall that in the medium run (equation [9.8]):
g, = 0 => - = g„.
lives
II nominal money growth is constant torcver, then inllation and expected inllation must eventually be constant as well. Assume that output growth equals zero. Then actual inflation and cxpectcd inflation must both equal nominal money growth:
AM M
Replacing — by AM/M in equation '24.5' g
AM M
AM
YL f
'24.6i
M
Seignorage -
Note thai nominal money growth AM/M, enters thc equation in two places and has two opposite effects on seignorage:
• C.iven real money balances, nominal money growth increases seignorage. This ellect is captured by
the lirsi term in AM/M in equation (24.6>.
• An increase in nominal money growth increases inflation and thus decreases real money balances.
This ctfcct is captured by AM AI in the second term on the right of equation (24.6).
So the net elfect ol nominal money growth on seignorage is ambiguous. 1 he empirical evidence is lhat the relation between seignorage and nominal money growth looks as shown in Figure 24.2. The relation is hump-shaped.
At low rates ol nominal money growth, such as we observe in Australia, Europe or the United Slates today, an increase in nominal money growth leads to a small reduction in real money balances. Thus, higher money growth leads lo an increase in seignorage.
When nominal money growth and therefore inflation) is very high, however, the reduction in real money balances induced by higher nominal money growth becomes larger ami larger. Eventually, there is a rate ot nominal money growth—point A in Figure 24.2—beyond which further increases in nominal money growth decrease seignorage.
The shape ol thc relation in Figure 24.2 may look lamiliar to those of you who have studied the economics of taxation. Income tax revenues equal the lax rate on income times income—the lax base. At low lax rates the tax rate has little influence on how much people work, and lax revenues increase with the tax rate. But as lax rates increase further, some people start working less—or stop declaring part ol their income—and the tax base decreases. As the income tax reaches very high levels, increases in the tax rate lead to a decline in tax revenues. Obviously, lax rates ol 100 per cent lead to no tax revenue at all. Why work il the government takes all your income?
AM/M i -> Seignorage T ►
mm т => тг т => л* t *
M/P -I Seignorage 1
This relation between tax revenues and the lax rate is often callcd the Laffer curve after the economist Arthur l.afler, who argued in the early 1980s that a cut in US tax rates would lead to more tax revenues. He was clearly wrong about where thc United Slates was on the curvc. I he effect of
the decrease 11 tax rates was to lower tax revenues not increase them. Hut the general point slill stands: when tax rates are high enough, a further increase in the tax rate can lead to a decrease in tax revenues.
There is more than a simple analogy here. Inllation can he thought ol as a tax on money balances. The lax rate is the rate ol inllation тг, which reduces thc real value of money holdings. The tax base is real money balances M/P. The product ol these two variables тг'М'Р is called the inflation tax. There is a subtle difference Irom other forms ol taxation. What the government receives Irom money creation at any point in time isn't the inllation tax but rather seignorage AM VI Л1/Р1. However, the two arc closely related. When nominal money growth is constant, inllation must eventually be equal to nominal money growth, so lhat
M
Inllation lax - тг~
AM M M P
= Seignorage
What rate of nominal money growth leads to the most seignorage, and how much seignorage does it generate? I hese are lhe questions that Philip ( ngan asked in a classic paper on hyperinflations written in 1956. In one ol the earliest uses ol econometrics C'.agan estimated the relation between the demand lor money and expected inllation (equation [24.1 during each ol the hyperinflations in Table 24.1 Then, using equation (24.6 he calculated the rate ol nominal money growth that maximised seignorage, and the associated amount ol seignorage. The answers he obtained are given in the lirst two columns ol Table 24.3. The third column repeats the actual nominal money growth numbers from Table 24.1.
See the focus box 'Monetary contraction and fiscal expansion: The United States in the early 1980s'. in Chapter 20.
If the inflation rate is S per cent, you lose 5 per cent of the value of your real money balances. It is as if you were paying a tax of 5 per cent on these balances.
Seignorage Is first an increasing function, then a decreasing function, of nominal money growth.
HIGH INFLATION chapl 24
Figure 24.2
Seignorage and nominal money growth
This table shows something very interesting. In all seven hyperinflations actual average nominal money growth (column 3 lar exceeded the rale ol nominal money growth that would have maximised seignorage 'column I Compare the actual rate ol nominal money growth in I lungary alter World War If 12,200 per cent to lhe rate ol nominal money growth lhat would have maximised seignorage ?2 per cent. This would seem to be a serious problem lor the story we have developed so lar. II the reason for money creation was to finance the budget deficit why was the actual rate ol nominal money
Table 24.3 Nominal money growth and seignorage
Rate of money growth maximising seignorage (% per month) Implied seignorage (% of output) Actual rate of money growth (% per month)
Austria 12 13 31
Germany 20 14 314
Greece 28 1 1 220
Hungary 1 12 19 33
Hungary 2 32 6 12,200
Poland 54 S 72
Russia 39 1 49
I 1
Monthly rate of nominal money growth, in percentage.
SOURCE: Philip Cagan.'The monetary dynamics of hyperinflation', in Milton Friedman (cd.). Studies In the Quantity Theory of Money (Chicago: University of Chicago Press. 19561.© University of Chicago Press (1958).
growth so much higher than the number thai maximised seignorage? The answer lies in the dynamics of the economy's adjustment to high nominal money growth. Wc now turn to lhat.
Dynamics and increasing inflation
Return to the argument just developed. I f maintained forever, a higher rate ol nominal money growth will eventually lead to a proportional increase in both actual inflation and expected inflation, and so lead to a decrease in real money balances. II nominal money growth is higher than the amount that maximises seignorage, the increase in nominal money growth will lead to a decrease in seignorage.
The crucial words in the argument are il maintained forever' and eventually'. Consider a govern¬ment that needs to finance a suddenly much larger delicit, and decides to do so by creating money. As thc rate ol money growth increases, it may take a while for inflation and expected inllation to adjust. Even as expected inflation increases it will take a while longer for people to fully adjust their real money balances. Creating barter arrangements lakes time, the use of foreign currencies in transactions develops slowly, and so on.
Lets state this conclusion more formally. Recall our equation for seignorage:
AiM M
Seignorage = — —
4 In the short run: gm T =» / -l and
«Д-уТ.
So. for both reasons.
«— 77 = r -
The following joke was told in Israel during the high inflation of the 1980s. Why is it cheaper
* to take a taxi rather than a bus? Because in the bus you have to pay the fere at the beginning of the ride. In the taxi, you pay only at the end.'
4 We have discussed here the costs of very high inflation. The discussion today in OECD countries is about the costs of. say. S per cent inflation versus 0 per cent.The issues are quite different in that case, and we return to them in Chapter 25.
There is also another effect at work, which we have ignored until now. We have taken the delicit as given. But as inflation becomes very high thc budget delicit typically becomes larger. Part of the reason has lo do with lags in lax collection. This ellect is known as the Tanzi-Olivera effect, for Vito Tanzi and Julio Olivera. two economists who have emphasised its importance. As taxes are collected on past nominal incomc their real value goes down with inflation. For example, if income taxes are paid this year on income received lasi year, and il the price level this year is icn times higher than last year's price level, the actual lax rate is only one-tenth of the official tax rate. I hus, high inflation typically decreases real government revenues, making ihe delicit problem worse. The problem is often compounded by other effects on the expenditure side. Governments oilen try to slow inllation by prohibiting firms under state control from increasing their prices although their costs arc increasing with inllation. The direct effect on inflation is small at best, but the lirms then run a dclicit that must in turn be financed by the government, further increasing the budget deficit. As the budget deficit increases, so docs thc need lor more seignorage. and so does the need lor even higher nominal money growth.
Hyperinflations and economic activity
We have focused so far on movements in nominal money growth and inllation—which clearly dominate thc economic scene during a hyperinflation. But hyperinflations alfeel the economy in many other
ways.
Initially, higher nominal money growth leads to an increase in output. It takes some lime lor increases in nominal money growth to be reflected in inflation, and during that time the effects ol higher nominal money growth are expansionary. As you saw in Chapter 14, the initial ellect of a monetary policy expansion is actually to decrease the nominal interest rate and real interest rates and increase the money supply leading to an increase in demand and an increase in output. Bui as inflation becomes very high, the adverse ellects ol hyperinflation dominate:
• Thc transaction system works less and less well. One famous example of inellicicnt exchange is the story ol people using wheelbarrows to earn,' all the currency needed lor transactions al the end of the German hyperinflation.
• Price signals become less and less useful. Because prices change so often, il is dillicull lor consumers and producers to assess the relative prices of goods and to make informed decisions. Thc evidence shows lhat ihe higher the rate of inflation, the higher the variation in the relative prices of different goods. Thus, thc pricc system, which is crucial to the functioning of a market cconomy, also becomes less and less efficient.
• Swings in ihe inflation rale become larger. It becomes harder to predict what inflation will be in thc near future—whether il will be, say, 500 or 1.000 per cent over the next year. Borrowing ai a given nominal interest rate becomes more and more a gamble. II you borrow at, say, 1,000 per cent lor a year, you may end up paying a real interest rate ol 500 per cent or 0 per ccnt. A large difference! The result is thai borrowing and lending typically come to a near stop in the last months of hyperinflation. leading to a large decline in investment.
So, as inflation increases and its costs become larger, there is typically an increasing consensus that il should be slopped. This lakes us to the next section, how hyperinflations actually end
24.4 HOW DO HYPERINFLATIONS END?
()n the expenditure side reform typically implies reducing the government subsidies that have often mushroomed during the hyperinflation. Obtaining a temporary suspension ol interest payments on loreign debt also helps to decrease expenditures. An important component ol stabilisation in Germany in 1923 was thc reduction in reparation payments'—precisely those payments that had triggered the hyperinflation in the lirst place.
()n the revenue side, what is required isn't so much an increase in overall taxation as a change in the composition ol taxation. This is important. As you saw, during a hyperinflation, people are in effect paying a tax—namely, the inflation tax. Stabilisation implies replacing the inflation tax with other taxes. The challenge is to put in place and collect these other taxes. This cannot be done overnight hut it is essential that people become convinced that it will be done and lhat the budget deficit will be reduced.
• Ihe central bank must make a credible commitment that it will no longer automatically monetise the government debt This credibility can he achieved in several ways. The central bank can be prohibited, by decree. Irom buying any government debt, so that no moneiisaiion ol the debt is possible. ()r the central bank can peg the exchange rate to thc currency ol a country with low inflation. An even more drastic step is to dollarise it) make a foreign currency such as the US dollar the country's official currency. This step is drastic, because it implies giving up seignorage altogether, and is olten perceived as a decrease in the country's independence.
• Are other measures needed as well? Some economists argue that incomes policies—that is wage and price guidelines or controls—should be used in addition to tiscal and monetary measures, to help the economy reach a new lower rate ol inflation. Incomes policies, they argue, help to coordinate expectations around a new lower rate ol inllation. II firms know that wages won't increase, they won . increase prices. II workers know that prices won't increase, they won't ask lor wage increases, and inflation is eliminated more easily.
Others argue that credible deficit reduction and central bank independence are all that is required. They argue lhat the appropriate policy changes, it credible, can lead to drastic changes in expectations and therefore lead to the elimination ol expected and actual inflation nearly overnight. They point lo the potential dangers ol wage and price controls. Governments may end up relying on the controls, and may not take the painlul but needed liscal and policy measures, leading ultimately to failure. Also il the structure ol relative prices is distorted to start with, price controls run the risk of maintaining these distortions.
Stabilisation programs that don't include incomes policies are called orthodox,- those thai do are called heterodox because they rely on both monetary liscal changes and incomes policies . The hyperinflations of Table 24.1 were all ended through orthodox programs. Many ot the l.atin American stabilisations ol the 1980s and 1990s have relied on heterodox programs.
Can stabilisation programs fail?
Stabilisation programs can lail, and they ollen do. Argentina went through live stabilisation plans Irom l )84 to 1989 before succeeding in stabilising inflation in the early 1990s. Brazil succeeded in 1995, but it was on iis sixth attempt in twelve years.
Sometimes lailure comes Irom a hall-hearted etfori ai stabilisation. A government puts wage controls in place but doesn i take the measures needed to reduce the deficit and nominal money growth. Wage controls cannot work if nominal money growth continues, and the stabilisation program eventually fails.
This is whit Argentina ► did in 1991. adopting a currency board and fixing the exchange race at one US dollar for one peso. See die discussion of currency boards and the evolution of the Argentine economy since 1991 in Chapter 21.
This argument was ► particularly relevant in the stabilisations in Eastern Europe in the early 1990s where, because of central planning, the initial structure of relative prices was very different from the structure of relative prices in a market economy. Imposing wage or price controls would have prevented relative prices from adjusting to their appropriate market value.
See. for example, the ► failed stabilisation attempt in April 1984 in Bolivia described in the focus box in diis chapter.
pi 24
Sometimes failure comes from political opposition. II social conflict was one ot the causes ol the initial budget delicil and thus was at the root ol the hyperinflation ii may still he present and iust as hard lo resolve at the time ot stabilisation. Those who lose Irom the liscal reform required to decrease the deficit will oppose the stabilisation program and may force the government to retreat. Often, workers who perceive an increase in the price ol public services or an increase in taxation, but don't fully perceive the decrease in the inflation tax, go on strike or even riot, leading to failure ol thc stabilisation plan.
4 This is a variation on the theme of self- fulfilling exchange rate crises developed in Chapter 21.
Remember that the rate of real money growth equals the rate of nominal money growth minus the rate of inflation. If inflation decreases by less than nominal money growth, this implies negative real money growth—a 4 decrease in the real money stock.This decrease in the real money stock then leads to high interest rates, which can trigger a recession.
All rich and most middle-income countries in the world have low inflation at this point A few, such as Japan, have deflation. The two middle-income countries with the highest inflation rate as of mid-2008 are Venezuela, with an inflation rate of 25 per cent, and Turkey, with an inflation rate of 10 per 4 cent.
After ten years of low inflation under a currency board 4 arrangement. Argentina is again at risk.The currency board has collapsed, the peso has depreciated and the budget deficit is getting larger. By 2003, inflation had reached 35 per cent after a mere - I per cent in 2001. In 2008, inflation was down to 8.5 per cent.
Failure can also come from the anticipation of failure. Suppose that the exchange rate is fixed to the dollar as part ot the stabilisation program. Also suppose that participants in linancial markets anticipate that the government will soon be forced to devalue. To compensate lor the risk of devaluation, they require very high interest rates to hold domestic bonds rather than US bonds. These very high interest rates cause a large recession. The recession forces the government to devalue, validating thc markets initial fears. If, instead, markets had believed lhat the government would maintain the exchange rate, the risk of devaluation would have been lower, interest rates would have been lower, and the government would have been able to proceed with stabilisation. To many economists, thc succcsses and lailurcs of stabilisation plans appear to have an element of self-fulfilling prophecy. F.ven well-conceived plans work only if they arc expected to work. In other words, luck and good public relations both play a role.
The costs of stabilisation
You saw in Chaptcr 9 how thc Australian disinflation of the early 1990s was associated with a recession and a large increase in unemployment. Similarly, disinflation in Europe in thc 1980s was also associated with a large increase in unemployment. We might therefore expect the much larger disinflations associated with thc end ol a hyperinflation to be associated with very large recessions or even with depressions. This is typically not the case.
To understand why. recall our discussion ot disinflation in Section 9.3. We argued that there were three reasons why inflation might not decrease as fast as nominal money growth, leading to a recession:
• Wages arc typically set in nominal terms lor some period of time (up to three years in the United States, though usually only one year in Australia) and, as a result, many ol (hem are already determined when the decision to disinllate is made.
• Wage contracts are typically staggered, making it difficult to implement a slowdown in all wages at thc same time.
• Thc change in monetary policy may not be fully and instantaneously credible. Flypcrinllation eliminates the first two problems. During hyperinflation, wages and prices arc
adjusted so often that both nominal rigidities and the staggering of wage decisions become nearly irrelevant.
But the issue of credibility remains. Thc lact that even coherent programs may not succeed implies that no program is fully credible from the start. If, lor example, thc government dccides to lix thc exchange rate, a high interest rate may be needed initially to maintain the parity. Programs that turn out to be successful arc those where increased credibility leads to lower interest rates over time. But, even when credibility eventually arrives, the initial high interest rate often leads to a recession. Overall, the evidence is that most, but not all, hyperinflations involve some cost in output.
How should a stabilisation package be designed so as to reduce this output cost? Should the stabilisation program be orthodox or heterodox? Should there be restrictions on nominal money growth, or should thc exchange rate be fixed? At this point, few countries are experiencing high inflation, so the questions aren t at thc top of policy-makers' agendas. But, if history is any guide, some countries will again lose control of their budget, finance thc budget deficit through money creation and experience high inllation, if not hyperinflation. Such questions are then sure to come back.
24.5 CONCLUSIONS
Monetary and fiscal policy may prove unable to help. In a slump, monetary policy may be con¬strained by the liquidity trap. Nominal interest rates cannot be negative. The government cannot run budget deficits to sustain higher demand and higher output forever,- if it tries, the increase in government debt eventually becomes a problem in itself.
And governments may lose control of both fiscal policy and monetary policy. Faced with major adverse shocks—war, civil war, a collapse of their exports, a social explosion—they may lose control of their budget, am a larger and larger budget deficit, and have no other choice than to finance the deficit through money creation. The result of this loss of control may be high inflation or even hyperinflation.
Комментариев нет:
Отправить комментарий