The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ^ ___________________________ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Business Review Federal Reserve Bank o f Philadelphia March* April 1993 ISSN 0007-7011 Leaning Against the Seasonal Wind: Is There a Case for Seasonal Smoothing of Interest Rates? Satyajit Chatterjee Business Review The BUSINESS REVIEW is published by the Department of Research six times a year. It is edited by Sarah Burke. Artwork is designed and produced by Dianne Hallowell under the direction of Ronald B. Williams. The views expressed here are not necessarily those of this Reserve Bank or of the Federal Reserve System. SUBSCRIPTIONS. Single-copy subscriptions for individuals are available without charge. Insti tutional subscribers may order up to 5 copies. BACK ISSUES. Back issues are available free o f charge, but quantities are limited: educators may order up to 50 copies by submitting requests on institutional letterhead; other orders are limited to 1 copy per request. Microform copies are availablefor purchase from University Microfilms, 300 N. Zeeb Road, Ann Arbor, MI 48106. REPRODUCTION. Perm ission must be obtained to reprint portions o f articles or whole articles. Permission to photocopy is unrestricted. Please send subscription orders, back orders, changes o f address, and requests to reprint to Publications, Federal Reserve Bank o f Philadelphia, Department o f Research and Statistics, Ten Independence Mall, Philadelphia, PA 19106-1574, or telephone (215) 574-6428. Please direct editorial communications to the same address, or telephone (215) 574-3805. MARCH/APRIL 1993 WHAT CAUSES INFLATION? Laurence Ball As suggested in this article, inflation is like the weather: everyone complains about it, but no one does anything about it. And, like the weather, inflation affects almost everybody. Some favorite candidates for the causes of inflation include OPEC, ex cessive government spending, and the climbing costs of medical care. What does cause inflation, and can it be eliminated? In answer to these questions, Larry Ball tells us, "There's good news and bad news." LEANING AGAINST THE SEASONAL WIND: IS THERE A CASE FOR SEASONAL SMOOTHING OF INTEREST RATES? Satyajit Chatterjee Financial and banking panics occurred regularly in the United States in the 19th and early 20th centuries, triggered by a seasonal increase in the demand for money and short-term interest rates. The Federal Reserve System was established, in part, to insulate the financial system from such panics. But the introduction of federal deposit insurance in 1933 greatly lessened the need for a seasonal monetary policy to fight panics. Consequently, the primary effect of such a policy is to smooth the seasonal path of short-term interest rates, which leads Satyajit Chatterjee to pose the question: Is there a case for seasonal smoothing of interest rates? FEDERAL RESERVE BANK OF PHILADELPHIA What Causes Inflation? Laurence Ball* I nflation is universally unpopular; everyone from ordinary consumers to top government officials bemoans the perpetual process of ris ing prices. Frequently, discussions of inflation have an air of resignation. Inflation is like bad weather: we can complain about it, but it seems to be a fact of life. For most people, the causes of inflation are murky. Popular writers lay the blame on a variety of scapegoats: governments ^Laurence Ball is an assistant professor of economics at Princeton University and a visiting scholar in the Research Department of the Philadelphia Fed. that spend too much money, the OPEC cartel, skyrocketing costs of medical care. What causes inflation, and is there any way to eliminate it? Economists have both good news and bad news about inflation. The good news is that we know a lot about its causes and how it could be ended. The bad news— and the reason that inflation has not been ended—is that doing so could be costly. This article describes what economists understand about inflation and what issues remain mysterious. There is a clear consensus about the long-run causes of inflation—the determinants of average infla tion over a decade or more. The short-run 3 BUSINESS REVIEW behavior of inflation— the ups and downs from year to year— is only partly understood. MARCH/APRIL 1993 INFLATION IN THE LONG RUN The year-to-year movements in inflation that make newspaper headlines are small compared with the differences in inflation across different eras or different countries. In the United States, consistently decreases relative to output there is deflation: the price level falls. (The most recent example of deflation in the United States is the early 1930s.) Why does too rapid growth in the money supply cause inflation? To see the answer, consider how the economy responds when the money supply rises. According to mainstream in flation as m easu red by the gro ss-n atio n al- econ om ics, firm s do n o t im m ed iately ad just product deflator averaged 7.4 percent per year from 1970 through 1979, but only 2.4 percent from 1950 through 1959.1 From 1930 through 1939, inflation averaged -1.7 percent per year— the price level was lower at the end of the decade than at the beginning. And these differ ences across periods in the United States, while substantial, are dwarfed by differences across countries. From the 1950s to the mid-1980s, inflation averaged 4.2 percent per year in the United States, only 2.7 percent in Switzerland, but 8.0 percent in Italy, 21.2 percent in Israel, and 54.4 percent in Argentina (see Ball, Mankiw, and Romer, 1988).2 What causes these differ ences in inflation over long periods? The Culprit: Too Rapid Money Growth. While economists disagree about many issues, there is near unanimity about this one: continu ing inflation occurs when the rate of growth of the money supply consistently exceeds the growth rate of output. In the long run, as Milton Friedman puts it, "inflation is always and everywhere a monetary phenomenon." When the money supply grows much more quickly than output of goods and services, inflation is high; when it grows only slightly faster than output, inflation is low; and when it their prices in response to an increase in the money supply. Because prices do not respond immediately, there is an increase in the real money supply—the money supply relative to the price level. The increase in the real supply of money pushes down the price of money—that is, the interest rate. Over time, lower interest rates stimulate borrowing and spending by firms and consumers, and the economy ex pands. The story ends when firms react to the booming economy and their strained capacity by raising prices. Prices rise until they match the increase in the money supply, pushing the real money supply back to its original level and choking off the boom. That is, the long-run effect of a 10 percent increase in the money supply relative to output is a 10 percent in crease in the price level and no change in the ratio of money to prices. It follows that if the money supply increases 10 percent faster than output every year, prices must eventually rise 10 percent per year. The gap between the average rate of money growth and the average growth rate of output determines average in flation.3 aUnless otherwise noted, all inflation figures refer to the percentage change in the GNP deflator. This variable is a broad index of the level of all prices in the economy. The more famous Consumer Price Index covers only prices paid by consumers, not those paid by governments or businesses. 2Citations to all papers mentioned in the text are in cluded in the "References" section at the end of this article. 4 3To be complete, inflation depends on the growth rate of the "velocity" of money— the frequency with which money is turned over—as well as on the gap between the average growth rates of money and output. For the United States, the average growth rate of velocity (for the M2 measure of money) has been zero over the past 40 years. In practice, then, money growth of 2 or 3 percent per year is consistent with stable prices. This rate of money growth matches the natural growth of output and spending. FEDERAL RESERVE BANK OF PHILADELPHIA Laurence Ball What Causes Inflation? In principle, differences in in flation across countries or time periods could be explained by dif ferences in either money growth or output growth, since the gap between the two determines infla tion. In practice, however, the most important factor is money growth, which varies widely, with levels near zero in some countries and over 100 percent per year in others. Variation in output growth is smaller and thus is a secondary factor in explaining differences in the gap between money growth and output growth. As a first approximation, then, differences in inflation across time periods or countries can be explained by dif ferences in money growth. To provide evidence for this FIGURE 1 Money Growth and Inflation (U.S.) (1870 -1980) Inflation Rate (Percent Per Year) 1970s1 11910s 1940s 1980s ■ ■1960s ■ 1900s 1950s ■ 1890s" 1930.sl _ ■1870s ■ 1880s ■ 1920s '4 0 2 4 6 8 10 Growth in Money Supply (Percent Per Year) 12 Recreated from: Milton Friedman and Anna J. Schwartz, Monetary Trends in the United States and the United Kingdom. Chicago: University of Chicago Press, 1982, with permission. p oin t, F ig u re 1 p lots av erage in flation and money growth in the United States for various decades. Figure 2 presents average infla tion and money growth from 1986-89 for a number of coun tries.4 In Figure 1, the decades with the highest inflation, such as the 1910s and the 1970s, are those with the highest money growth. Similarly, Figure 2 shows a close relationship between inflation and money growth across countries. Countries such as Switzerland and France produce low inflation through low m oney grow th; countries such as Turkey and Mexico produce high inflation 4The data for Figures 1 and 2 are taken from Friedman and Schwartz (1982) and Abel and Bernanke (1992), respectively. FIGURE 2 Money Growth and Inflation Across Countries (1986 -1989) Inflation Rate (Percent Per Year) M exico P Poland □ ■J k Zaire □ Turkey M alaw i C ote DTv oire j 1 r £ fi: C had t-C -20 -10 undi .S. E G hana ■ | fiil _ ■ M ada »ascar ® ■N epal “ itzer la id , R w an da, Belgiun Sw Sen egal France tL 0 10 20 30 40 50 60 Growth in Money Supply (Percent Per Year) 70 80 Recreated from: Andrew B. Abel and Ben S. Bernanke, Macroeconomics. Reading, MA: Addison-Wesley, 1992, p. 141, with permission. 5 BUSINESS REVIEW through high money growth. Along with the theoretical arguments discussed above, this evidence has convinced economists that trend or average inflation is determined by money growth. Why Is Money Growth Excessive? The question of what causes inflation has, at one level, an easy answer: money growth. This MARCH/APRIL 1993 Budget deficits are not, however, the basic source of inflation in the United States or in most European economies. The U.S. govern ment has, of course, run large deficits over the past decade. But these deficits have been fi nanced primarily by borrowing, not by print ing money. That is, the government covers its deficit mostly by issuing bonds. The Federal answ er, h o w ev er, raises ano th er, d eep er qu es R eserve co n trib u tes to g ov ern m en t rev en u e b y tion: why do policymakers allow the money supply to grow quickly? The Federal Reserve and corresponding monetary authorities in other countries possess effective techniques for controlling the average growth rate of the money supply.5 Policymakers could slow average money growth enough to keep the average inflation rate at zero (although shocks to the economy would cause temporary movements above and below zero). Since both the public and the Federal Reserve dislike inflation, why isn't it eliminated? The answer to this question is different in different types of economies. In some coun tries, the answer is simple: the government prints money at a rapid rate to finance budget deficits. This explains most episodes of very high inflation—the annual inflation of several hundred percent or more that has afflicted South American countries and Israel within the past decade. These countries have had high levels of government spending and have been unable politically to match this spending with tax revenues; thus they have financed their spending by creating new money. Predictably, rapid money creation has produced high infla tion. Inflation has been brought under control only when the underlying budget deficit was reduced. (In Israel, for example, such a stabili zation occurred in 1985.) creating new money, but this "seignorage" is small: less than 1 percent of total revenue. In countries like the United States, policymakers would gladly eliminate inflation through lower money growth if the only cost were a small revenue loss. The deterrent to lowering infla tion must arise from a different source. The reason U.S. policymakers are reluctant to push inflation to zero is that doing so is likely to cause a recession, or at least slower economic growth. This fear is supported by both macroeconomic theory and historical experi ence. Slower money growth reduces inflation in the long run, but there is a lag, as discussed earlier. When money growth falls, firms ini tially continue to raise prices at the rate to which they are accustomed. With money grow ing more slowly than prices, the real money supply falls, causing a recession. Only the experience of the recession causes inflation to fall. This theoretical story fits much of the U.S. experience. One cause of the recession that began in 1990 was, arguably, the Fed's efforts to reduce inflation in the late 1980s. More clearly, disinflation was a major cause of the recession of 1981-82—the worst recession since the 1930s. Paul Volcker, the chairman of the Federal Re serve from 1979 to 1986, moved decisively to eliminate the double-digit inflation of the late 1970s. He succeeded, but at a price: inflation fell from 10.1 percent in 1980 to 4.0 percent in 1983, but unemployment rose from 5.8 percent in 1979 to 9.5 percent in both 1982 and 1983. Research by economic historians has shown that this experience is part of a regular pattern: Specifically, the Fed manipulates the supply of money through "open market operations"—purchases and sales of government bonds. Buying bonds with money adds to the economy's money stock, and selling bonds drains money out of the economy. Digitized for 6 FRASER FEDERAL RESERVE BANK OF PHILADELPHIA What Causes Inflation? Laurence Ball when the Fed slows money growth substan 5.0, 7.6, and 9.6 percent. One source of these tially to reduce inflation, a recession occurs inflation movements is temporary fluctuations almost invariably.6 in the growth of the money supply. In contrast While some policymakers are willing to pay to the long run, however, too rapid money this price to reduce inflation, others are not. growth is not the only, or even the primary, And the Fed's eagerness to fight inflation ap determinant of inflation. Figure 3 plots infla pears to depend on the severity of the inflation tion against money growth for each year during problem. Volcker was sufficiently concerned about double-digit infla tion to implement the monetary FIGURE 3 tightening needed to reduce infla Money Growth and tion. But inflation of around four Inflation During the 1970s (U.S.) percent, the level through much of Inflation Rate (Percent) the 1980s, did not create enough distress to prompt a further tight 10 □ 1974 ening. Thus inflation continued. (For more on this subject, the reader 111979 c >1978 is referred to "W hat Are the Costs n 1975 8 of D is in fla tio n ?" by D ean □ 197 3 a 1977 Croushore, in the May/June 1992 1976 issue of this Business Review.) SHORT-RUN FLUCTUATIONS IN INFLATION Although money growth deter mines average inflation in the long run, the short-run behavior of infla tion is more complicated. Inflation fluctuates around its long-term trend from year to year; for ex ample, annual inflation rates in the second half of the 1980s varied around their average of 3.6 percent, with annual rates from 1985 to 1989 of 3.6, 2.5, 3.3, 4.3, and 4.2 percent. Short-term fluctuations in inflation were larger in the 1970s: the annual rates from 1970 to 1974 were 4.7,5.6, 3 1970 4 6 8 10 12 Growth in Money Supply (Percent) 14 Money Growth and Inflation During the 1980s (U.S.) Inflation Rate (Percent) 12 1980 m IQ 1981 10 19 8 9 b B19!38 19841a 6Romer and Romer (1989) identify six episodes since World War II in which the Fed sharply tightened policy to reduce inflation. In each case, a recession occurred within two or three years. H 1971 H (? ) 1972 a 1987 1982 m 1983 ®1985 ®19J26______ 4 6 8 10 12 Growth in Money Supply (Percent) 14 7 BUSINESS REVIEW the 1970s and 1980s. Clearly, annual inflation can differ considerably from money growth. What causes this short-run divergence? Demand Shocks. One source of short-run changes in inflation is shifts in aggregate demand—in desired spending by government, businesses, and consumers. Suppose that the government spends more to finance a war or businesses become more confident about the future and invest in factories and machines. As the demand for military hardware or for facto ries rises, the economy expands: firms increase production and hire more workers, cutting unemployment. But again, high output and low unemployment eventually spur faster in creases in wages and prices: inflation rises. Similarly, a fall in aggregate demand causes a recession, leading firms to raise prices more slowly. The economy's short-run movements between booms and recessions produce fluc tuations in inflation as well. A good example of inflation arising from a shift in aggregate demand—a shift that was not initiated by monetary policy—is the increase in inflation in the late 1960s. Annual inflation varied from 0.8 percent to 2.3 percent over the period of 1960-64, but rose to 5.3 percent in 1969. The consensus explanation for this expe rience is increased government spending. As the Vietnam War escalated, the Johnson admin istration raised military spending while also continuing the social programs of the "Great Society." As a result, the federal budget deficit grew from $1.4 billion in fiscal year 1965 to $25.2 billion in 1968, and the economy over heated: unemployment fell, but inflation rose. Price Shocks. Until the early 1970s, most economists believed that shifts in aggregate demand were the dominant source of short-run movements in inflation. This view had to be modified, however, after the experience of the 1970s, when price shocks— a.k.a. "supply shocks"—caused large increases in inflation. These shocks were sharp increases in the prices of particular goods, namely food and energy 8 MARCH/APRIL 1993 products, arising ultimately from poor weather and the emergence of the OPEC cartel. These shocks created "stagflation": inflation rose while unemployment rose and real output fell (in contrast to the experience of demand shocks, which push inflation and unemployment in opposite directions). From 1972 to 1974, annual inflation rose from 5.0 percent to 9.6 percent as a result of a rise in food prices and the first OPEC price increase. OPEC II raised inflation from 7.1 percent in 1977 to 10.1 percent in 1980. These increases dwarfed the fluctuations in inflation arising from the demand shocks of the previous 20 years. More recently, the spike in oil prices during the gulf crisis raised inflation in the second half of 1990. Why do rises in food and energy prices create inflation? The reader will be forgiven for thinking that the answer is obvious: food and energy are a significant fraction of the economy, and rises in prices are the definition of inflation. Economists, however, believe that the issue is not so simple because of the distinction be tween the overall price level and relative prices. In classical economic theory, the price level is determined by the money supply, as de scribed above. Changes in supply and demand for various products arising from weather con ditions, cartel decisions, and so on affect not the price level but relative prices: OPEC makes oil more expensive relative to other goods. Theo retically, this is accomplished partly by an increase in the absolute price of oil and partly by decreases in all other prices. With these price adjustments, oil can become relatively more expensive while the price level remains un changed at the equilibrium level determined by the money supply. In practice, this is not what happens: OPEC in fact raised the average price level. But it is not obvious why this is so.7 7Writing in 1975, Milton Friedman puts the point this way: "It is essential to distinguish changes in relative prices from changes in absolute prices. The special conditions that FEDERAL RESERVE BANK OF PHILADELPHIA What Causes Inflation? Laurence Ball This issue is the subject of recent research by me and Gregory Mankiw of Harvard Univer sity (Ball and Mankiw, 1992). Our explanation for the inflationary effects of price shocks rests on two ideas. First, there is some inertia in prices. Firms do not instantly adjust prices to every change in circumstances; instead, they adjust only if their desired price change is large enough to justify the costs of adjustment. For example, a mail-order company will print a new catalog to announce a 50 percent sale, but it is not worth the effort to announce a one-cent price change arising from a tiny change in costs; instead, the firm will simply keep its prices fixed. This behavior implies that large shocks have disproportionately large effects on prices: firms adjust to them quickly, while they make smaller adjustments more slowly. The second key idea is that "price shocks" are episodes in which certain relative prices rise or fall by unusually large amounts. In the The large relative shocks to oil-related prices triggered quick upward adjustments. For ex ample, given the large increase in oil prices, gas stations would have suffered huge losses had they not quickly raised prices at the pump. In contrast, while prices of many other goods came under downward pressure, the required price decreases were small and hence occurred more slowly. When consumers spent more money on oil, they had less available for toothbrushes, soft drinks, and all other nonoil goods, creating an incentive for the sellers of these products to reduce prices. But the desired decreases were only a few percentage points because OPEC did not cut heavily into toothbrush or soft drink demand. Thus firms were slow to adjust prices downward. In the short run, oil-related prices rose, and the offsetting decreases did not fully occur. Thus prices rose on average: there was inflation. This th eoretical story exp lain s a large num O P E C ep iso d es, for exam p le, som e relative prices— those for oil-related products—rose 50 ber of the rises and falls in inflation in the percent or more in response to the trebling of oil United States. The oil and food price episodes prices. By definition, other relative prices went in the 1970s are examples. Another example is down to balance these increases: if some prices the large decrease in oil prices in 1985-86. Our are relatively higher, others must be relatively theory predicts that inflation should fall in this lower. It was not the case, however, that episode because the decreases in oil prices equilibrium prices of some nonoil products occur more quickly than the smaller increases needed to fall by more than 50 percent. Instead, in other prices. And, indeed, inflation fell from the relative price decreases were spread over 4.4 percent in 1984 to 2.5 percent in 1986. Our theory also explains episodes before the all nonoil goods: a fraction of relative prices rose a large amount, balanced by smaller rela famous supply shocks of the 1970s. For ex ample, inflation rose above 10 percent in 1951, tive decreases in the majority of prices. Combining this idea with the previous largely due to a demand shock: the Korean one— that large shocks have disproportionate War. Inflation then plummeted to near zero in effects— explains why OPEC was inflationary. 1952, and the cause appears to be a price shock. Specifically, the prices of meat, rubber, veg etable oil, and several other products fell steeply. More generally, my research with Mankiw sug drove up the prices of oil and food required purchasers to gests that a combination of demand and price spend more on them, leaving them less to spend on other shocks explains most of the year-to-year fluc items. Did that not force other prices to go down or to rise tuations in U.S. inflation since 1950. less rapidly than otherwise? Why should the average level Although some relative price increases are of prices be affected significantly by changes in the price of inflationary according to our theory, others are some things relative to others?" 9 BUSINESS REVIEW not. One example is the steady increase in the cost of medical care. These price increases probably have little to do with inflation, despite frequent claims to the contrary in popular dis cussions. A relative price increase affects infla tion only if there is an unusually large shock during a particular year, so that the upward price adjustment occurs more quickly than the offsetting downward adjustments. Medical costs have risen faster than the overall price level for several decades, but the rise has been steady; there are no cases of 50 percent or 100 percent increases within a year, as in the case of oil. This smooth adjustment of relative prices could occur without inflation. If the Federal Reserve pursued noninflationary monetary policy, the average price level would remain steady, with rises in the price of medical care offset by price decreases in other industries. FROM THE SHORT RUN TO THE LONG RUN According to the analysis so far, the average rate of inflation over a long period is deter mined by the amount that average growth of the money supply exceeds average output growth. Inflation fluctuates around its trend from year to year in response to various de mand and price shocks. We have seen that these ideas explain much of the U.S. inflation experience, but they do not capture one aspect: the link between the short run and the long run. Suppose that inflation is proceeding at the level determined by trend money and output growth and that oil prices rise sharply. The theories reviewed so far suggest that this price shock should raise inflation in the short run but that inflation should then return to its long-run trend if trend money growth is unchanged. In fact, shifts in inflation arising from demand or price shocks appear quite persistent. When government spending raised inflation in the late 1960s, and when OPEC raised inflation in the 1970s, there was little sign that inflation would naturally return to its previous level. 10 MARCH/APRIL 1993 Instead, inflation continued until the Federal Reserve became sufficiently concerned to tighten policy, producing a recession. (Such policy tightenings occurred in 1970 in response to the high inflation of the late 1960s and in 1974 and 1978-79 after the OPEC shocks. See Romer and Romer, 1989.) Absent a policy tightening and recession, inflation arising from price or de mand shocks seems to continue indefinitely: short-run shifts in inflation have long-run ef fects on trend inflation. How can this evidence be squared with our earlier theories? Recall the crucial fact that trend inflation is ultimately caused by faster growth in the money supply than in output. Logically, if shocks such as OPEC shift trend inflation, they must induce the Federal Reserve to raise trend money growth (until the point when policymakers decide that inflation is too high and accept the cost of disinflation). Why does a short-run spurt in inflation lead the Fed to raise the average level of money growth? The usual answer to this question focuses on the behavior of inflationary expectations. In past experience, individuals have seen that increases or decreases in inflation usually per sist for a substantial period. Thus, when they see a new rise in inflation (because of an OPEC shock, for example), they expect inflation to stay high. Crucially, this expectation is selffulfilling: the expectation that inflation will stay high causes it to stay high. The reason expecta tions affect actual inflation is that they affect decisions about wage- and price-setting. If everyone expects a 10 percent rate of inflation to continue, workers will demand 10 percent wage increases to keep up. Firms will raise prices 10 percent to match the higher wages they pay and also the 10 percent increases they expect from their competitors. Thus inflation will continue at 10 percent, fulfilling expecta tions. The Federal Reserve is not helpless in the face of this self-fulfilling inflationary spiral. The spiral can continue only as long as it is FEDERAL RESERVE BANK OF PHILADELPHIA What Causes Inflation? "accommodated" by the Fed—as long as the Fed raises money growth as much as inflation has risen. However, a price shock such as that caused by OPEC is not only inflationary for the U.S., it also is contractionary. Because the higher price of imported oil leaves Americans with less of their incomes to spend on domestic goods and services, it causes output and em ployment to fall, at least temporarily. The Federal Reserve could bring inflation back down by slowing money growth. The result will be to reduce output further, causing a recession that eventually forces inflation down. Over sub stantial periods, however, such as the 1970s, the Fed has been unwilling to impose this cost on the economy. Thus, once a shock such as OPEC raises inflation, it can stay high for a long period before a Paul Volcker takes charge and disinflates. The price shock creates a vicious circle in which persistence in inflation creates the expectation of persistence, which in turn creates persistence. While this story is widely accepted, it is not airtight. At an empirical level, it appears true that changes in inflation are expected to persist. Surveys of the expectations of forecasters and of ordinary citizens show that a rise in current inflation leads to higher forecasts of future inflation. At a deeper level, however, it is not clear why expectations behave that way. Since the expectation of persistence is self-fulfilling, it proves itself correct. But there are other expectations that would also be self-fulfilling. Suppose that a price shock raised inflation in one year, but everyone expected that inflation would return to its original level in the next year. With the expectation of moderate infla tion, workers would moderate their wage de mands, and firms would moderate their price increases. Thus the expectation of low inflation would also prove itself correct. Since expecta tions of either persistent or nonpersistent infla tion are self-fulfilling, it appears that either expectation would be rational. The U.S. economy has settled into a situation in which Laurence Ball people expect inflation to persist, perhaps only because it has in the past. CONCLUSIONS The behavior of inflation is one of the betterunderstood areas of macroeconomics. There is a wide consensus about the long-run determi nants of inflation and, arguably, a consensus about much of its short-run behavior. The average inflation rate over long periods is de termined by the extent to which the average rate of money growth (which, in the United States, is chosen by the Federal Reserve) ex ceeds the average growth rate of real output. Short-run inflation fluctuates around its longrun average because of demand shocks, such as large increases in government spending, and supply shocks, such as sharp rises in the prices of food and energy. Some countries have persistently high infla tion because they continuously create new money to finance large, ongoing budget defi cits. Such countries are unable to reduce money growth enough to halt inflation because their governments have been unable to eliminate budget deficits and because they do not have effective alternatives for financing those defi cits. In the United States, however, the govern ment budget deficit is financed almost entirely with Treasury debt, not money creation. The United States had low average inflation in the 1980s because money growth, on average, only slightly exceeded output growth. Finally, the distinction between short-run and long-run determinants of inflation is blurred by the fact that short-run changes often influ ence the long-run trend. When a demand or price shock raises short-run inflation in the United States, expectations of future inflation rise. Historically, the Fed often accommodated these expectations by allowing money growth to rise, so expectations were fulfilled. Not allowing money growth to rise would have slowed output growth and perhaps caused a recession. 11 BUSINESS REVIEW These conclusions— a summary of the think ing of mainstream economists—partly fit ideas that are popular among journalists and the public and partly contradict such ideas. It is common, for example, to blame inflation on excessive deficit spending by the government. This view is on target for the case of Argentina, but not for the United States. Little of the U.S. d eficit is financed by p rin tin g m oney . Thus it was possible for U.S. inflation to fall between the 1970s and the 1980s even though the U.S. budget deficit rose substantially. On the other hand, the view that government spending fuels U.S. inflation has a grain of truth. There are periods, notably the Vietnam era, when too much spending overheats the economy, pro MARCH/APRIL 1993 ducing inflation that persists as long as mon etary policy is accommodative. Perhaps the most common scapegoats for inflation are the particular prices that the public observes to rise most rapidly. In some eras, these are oil or food prices; a current favorite is medical care. When journalists and citizens blame individual prices for inflation, they con fuse average and relative prices. Particular prices could rise just as much in relative terms even if the overall price level were constant. Again, however, there is a grain of truth in conventional thinking. Particularly sharp in creases in prices, such as OPEC shocks, are inflationary. Abel, Andrew, and Ben Bernanke. Macroeconomics. Addison-Wesley, 1992. Ball, Laurence, and N. Gregory Mankiw. "Relative-Price Changes as Aggregate Supply Shocks," mimeo, Princeton University (April, 1992). Ball, Laurence, N. Gregory Mankiw, and David Romer. "The New Keynesian Economics and the Output-Inflation Trade-off," Brookings Papers on Economic Activity (1988:1). Croushore, Dean. "W hat Are the Costs of Disinflation?" this Business Review (May/June, 1992). Friedman, Milton. "Perspectives on Inflation," Newsweek (June 24,1975). Friedman, Milton, and Anna J. Schwartz. Monetary Trends in the United States and the United Kingdom. University of Chicago Press, 1982. Romer, Christina, and David Romer. "Does Monetary Policy Matter? A New Test in the Spirit of Friedman and Schwartz," NBER Macroeconomics Annual, 1989. 12 FEDERAL RESERVE BANK OF PHILADELPHIA Leaning Against the Seasonal Wind: Is There a Case for Seasonal Smoothing of Interest Rates? s ince it began in 1914, the Federal Reserve System has followed a policy of allowing the supply of money to vary over the seasons. At present, the Fed allows the money supply to grow faster than average in the third and fourth quarters of each year to meet the seasonally high demand for money during summer and the holiday shopping season and forces it to grow slower than average in the first and sec ond quarters. In other words, the Fed injects * Satyajit Chatterjee is a senior economist in the Research Department of the Philadelphia Fed. Satyajit Chatterjee * additional money into the economy during the last two quarters of a year, then withdraws this addition during the first half of the following year. This seasonal pattern in the growth rate of money supply and the Fed's role in generating it is evident in Figure 1. The two lines show the seasonal deviations in the quarterly growth rate of M l and the monetary base (the sum of bank reserves and currency in circulation) from their average quarterly growth rates in the post-WWII period. The Federal Reserve, through its open-market operations, increases the growth rate of the monetary base in the 13 BUSINESS REVIEW MARCH/APRIL 1993 seasonal monetary policy. In his well-known lecture A Program for M onetary S tability , M ilton Seasonal Pattern in Quarterly Growth Friedman saw no "objection to Rates of Monetary Base seasonal fluctuations in short and Money Stock term interest rates" and recom 1948 I - 1980 IV mended that the Fed desist from following such a policy. More recently, Gregory Mankiw and Jeffrey Miron, in an article titled "Should the Fed Smooth Interest Rates? The Case o f Seasonal Mon etary Policy," have also raised doubts about the wisdom of such a policy. Indeed, why should the Fed accom m od ate seaso n al changes in money demand and stabilize short-term interest rates? After all, the increase in rental 1 2 3 4 rates for vacation properties on Quarters the New Jersey shore in August is third and fourth quarters as money demand a natural outcome of market forces and does rises, then reverses this increase in the follow not call for a program of rent stabilization by ing two quarters when money demand shrinks. the government. By the same token, why Correspondingly, the seasonal growth rate in shouldn't the Fed tolerate an increase in the M l is above average in the third and fourth rental price of money (interest rates) caused by quarters when the quantity of money demanded natural forces in the third and fourth quarters rises quickly and falls below average in the first of each year? In this article I examine this question by two quarters when quantity of money de looking first at the historical reason underlying manded shrinks. Because it accommodates seasonal variation the Fed's seasonal monetary policy and deter in the demand for money, the Fed's seasonal mining whether the historical rationale is still monetary policy has the effect of reducing valid. In light of the major institutional changes seasonal variation in short-term interest rates. Indeed, by some measures, there does not ap pear to be any evidence of seasonal movements 1In a recent article, Robert Barsky and Jeffrey Miron in short-term interest rates in the post-WWII report the absence of seasonal movements in the threeperiod.1If the Federal Reserve were to stop this month T-bill rate over the period 1948:2-1985:4. However, * seasonal variation in the growth rate of the there have been periods when short-term interest rates have monetary base, short-term interest rates would shown some seasonal fluctuations. Stanley Diller, in an rise in the third and fourth quarters in response article written in 1971, used measures of seasonality differ to the higher demand for money during these ent from the ones employed by Barsky and Miron and documented that T-bill rates showed some seasonality in times and fall in the first two quarters in re the 1950s, but this seasonal pattern all but disappeared in sponse to the lower demand for money. the 1960s. Citations may be found in the "References" Economists have questioned the need for a section at the end of this article. 14 FIGURE 1 FEDERAL RESERVE BANK OF PHILADELPHIA Leaning Against the Seasonal Wind that have occurred in the banking industry since 1933,1 argue that the historical reason for the Fed's seasonal monetary policy is now much less relevant. On the other hand, im provements in economists' understanding of the different ways in which monetary policy could affect the functioning of the economy suggest benefits and costs of a seasonal mon etary policy that were not apparent in 1914. In the rest of the article, I discuss the nature of these costs and benefits. THE HISTORICAL RATIONALE Throughout the latter part of the 19th cen tury and the early years of the 20th, the U.S. financial system was plagued by recurrent cri ses. Edwin Kemmerer, a Cornell University scholar who testified before the National Mon etary Commission in 1910, listed no less than six major crises and 15 minor crises in financial markets between the years 1890 and 1910. These fin an cial p an ics were a co m bin atio n of bank failures, bank runs, and stock-market crashes. Kemmerer, as well as other contemporary schol ars, believed that most of these crises had a seasonal connection.2 The United States, at that time a still heavily agricultural nation, experi enced large increases in the demand for cur rency and short-term loans during early spring and autumn when farmers were planting and harvesting. The increased demand for currency drained cash from country banks precisely at a time when their farming customers clamored for loans. As a result, the country banks would call in their reserves with the city banks and thereby transmit the seasonal pressure on bank reserves to the city banks as well. To try to accom m odate having fewer reserves and greater loan demand, many banks tried to make do with reserve-deposit ratios that were pre- 2In addition to Kemmerer's testimony, see, for example, the testimony of O.M. W. Sprague and the book by Laurence Laughlin. Satyajit Chatterjee cariously low and left them vulnerable to unex pected cash withdrawals. Bankers and deposi tors were quite aware that during these times the banking system's ability to absorb unex pected adverse shocks was low. Thus, an unexpected loan default or an unexpectedly heavy withdrawal that caused a city or a coun try bank to fail would generate panic with drawals from other banks as well. Even if the withdrawal or default did not lead to a bank failure, the episode made banks nervous enough to call in more of their loans, many of which were stock-market call loans, which, in turn, led to sharp drops in stock prices.3 The seasonal element in these financial pan ics is evident in the historical record. Of the 21 financial panics documented by Kemmerer, seven occurred in September and October and another seven between March and May. Thus, fall and spring accounted for all but a third of the total number of panics between 1890 and 1910. While these panics differed in severity, some were quite serious. For instance, the panics of May 1893 and October 1907 resulted in the suspension of convertibility of deposits into currency. In general, these disturbances were considered disruptive enough to warrant seri ous attention and led to the creation of the National Monetary Commission to investigate the source of the problem facing the U.S. bank ing industry.4 The deliberations of the commis- 3See Jeffrey Miron's 1986 article for a description of the connection between seasonality and financial panics. 4Unfortunately, there is no quantitative estimate of the disruption caused by these financial panics. Jacob Hol lander, a professor of political economy at Johns Hopkins University who also testified before the National Monetary Com m ission, noted the im portance of bank loans collateralized by stock certificates in the financing of busi ness activity in the U.S. This suggests that U.S. businesses probably faced considerable difficulty in carrying out their normal operations during times when panic conditions made such collateralized loans unattractive. 15 BUSINESS REVIEW MARCH/APRIL 1993 sion, published in 1910, identified the seasonal reserves, respectively.5 In Figure 2, the vertical pressure on bank reserves as one of the princi axis measures the average difference in call pal contributory factors in these panics. Three money rates across adjacent months in the preyears later, the Federal Reserve Act of 1913 Fed and post-Fed era. For example, in the preestablished the Federal Reserve System and Fed era, call money rates were, on average, 1.02 charged it with the task of eliminating the percentage points per annum higher in Septem seasonal pressure on bank reserves by allowing ber than in August and 1.15 percentage points banks to borrow additional reserves and cur per annum higher in December than in Novem rency (“to furnish.... an elastic currency") dur ber. In contrast, in the post-Fed era, call money ing times of increased seasonal demand for rates were only 0.13 percentage points per annum higher in September than in August and currency. Thus, a seasonal monetary policy came to be only 0.089 percentage points per annum higher one of the key goals of the Federal Reserve in December than in November. More gener System. The policy was remarkably successful ally, Figure 2 clearly shows that the call money in that in the 15 years following November rate was considerably more seasonal in the pre1914, there were no financial crises in the U.S. Fed era than in the post-Fed era. Figure 3 shows the other side of the same Aside from eliminating panics triggered by seasonal shortages of liquidity, the Fed's sea sonal monetary policy also had another impor tant effect. Because of the seasonal pressures 5The information on which these plots are based was on bank reserves, the period before the found obtained from Truman Clark's 1986 article, Table 2 (p. 82) ing of the Fed was characterized by prominent and Table 4 (p. 84). seasonal fluctuations in short-term interest rates. As bank reserves FIGURE 2 tightened in the fall and spring and the commercial banks called Seasonal Pattern in Call Money Rates in their loans, short-term interest Before and After Founding rates rose. Then, as the seasonal of Federal Reserve System pressure on reserves ebbed, short term interest rates declined in the winter and summer. After the Fed went into operation in 1914 and eliminated the seasonal pres sure on bank reserves, it also elimi nated the seasonal fluctuation in short-term interest rates. Thus, the seasonal smoothing of short term interest rates that continues to characterize Federal Reserve policy to this day originated in the battle against financial panics. Figures 2 and 3 display the preand post-Fed seasonal patterns in the call money rate (an overnight interest rate) and commercial bank Digitized 16 FRASER for FEDERAL RESERVE BANK OF PHILADELPHIA Leaning Against the Seasonal Wind Satyajit Chatterjee cial tranquility ended rudely with the stock market crash in October 1929; the terrible years of the Great Depression followed. A seasonal monetary policy notwithstanding, five major banking crises occurred between the years 1929 and 1933.6 The experience of the Great Depression con vinced American business and legislative com munities that monetary policy alone was inad equate to insulate the economy from financial and economic disasters. In a far-reaching insti tutional change, the Banking Act of 1933 intro duced federal insurance of bank deposits, which made bank deposits completely safe for the majority of depositors. While deposit insur ance does not cover all commercial bank depos its, the FDIC has acted in the past to protect all deposits, even the so-called uninsured ones. Typically, in the event of a bank failure FDIC policy is to merge the failed institution with an ongoing one. This way, the liabilities of the failed bank become the liabilities of the ongoing institution, and uninsured deposi tors emerge unscathed as well. FIGURE 3 However, what often goes unnoticed is that the existence of Seasonal Pattern in Bank Reserves deposit insurance greatly reduces Before and After Founding the need for a seasonal monetary of Federal Reserve System policy to fight banking panics. Seasonal pressures on bank re serves and short-term interest rates may cause some unlucky or ineptly managed banks to fail, but because of explicit and de facto deposit insurance such fail ures are unlikely to lead to bank runs or financial panics. Since coin. The vertical axis measures the average difference in bank reserves (in millions of dol lars) across adjacent months. In the pre-Fed era, the bank reserves declined by about $10.68 million in September, reflecting the withdrawal of currency for farm expenditures. A decline of similar magnitude is also evident in the month of February. In contrast, bank reserves rose $22.79 million in September in the post-Fed era, fueled by Federal Reserve purchases of Trea sury securities from banks. This increase in reserves allowed banks to meet the currency drain and, at the same time, expand the volume of their agricultural loans. In general, the Fed eral Reserve's seasonal monetary policy made bank reserves much more responsive to the pace of commercial activity and thereby elimi nated the pronounced seasonal pattern in short term interest rates. Seasonal pressures on currency and credit demand, alas, are not the only reason for finan cial disruptions. The 15-year stretch of finan 6Milton Friedman and Anna Schwartz provide in-depth discussions of the ori gins and dynamics of these panics in their book A Monetary History o f the United States 1867-1960. Miron's article discusses the seasonal factors that may have contrib uted to these banking crises. 17 BUSINESS REVIEW combating financial panics was the main rea son for a seasonal monetary policy in the first place, has this policy outlived its usefulness? SEASONAL MOVEMENTS IN INTEREST RATES ARE COSTLY Even though seasonal changes in interest rates probably would not cause financial panics today, a seasonally varying short-term interest rate results in a loss of economic efficiency. This loss of efficiency, while not as dramatic and severe as that imposed by a banking panic, could nevertheless provide a rationale for con tinuing a seasonal smoothing of short-term interest rates. To understand how this effi ciency loss occurs, we need to understand how changes in short-term interest rates affect indi viduals' and corporations' demand for money. Consider the case of Sadie Wherebucks, who must decide how much money to hold, on average, in her wallet or checking account and how much to put in a time deposit or a short term security such as T-bills. When short-term interest rates are low, the convenience pro vided by holding money is more important to Sadie than the small amount of income that she gives up by holding money instead of interestbearing assets, so Sadie will hold more money. When short-term interest rates are high, Sadie will reduce the amount of money she holds so that she can hold greater time deposits or invest in T-bills. If Sadie holds more financial wealth in a savings account or in the form of T-bills, she will have to use her bank or broker more often to convert her assets into money to meet her daily expenses. This will impose additional costs on Sadie either because she has to make frequent trips to her bank or because she has to pay her broker's commission fees more often. Therefore, one effect of an increase in short term interest rates will be to increase Sadie's transaction costs as she attempts to make do with smaller average holdings of money. What is true of Sadie as an individual is even 18 MARCH/APRIL 1993 more true of corporations. Because firms deal with large flows of funds, higher short-term interest rates present them with even greater inducement to tighten up on their cash manage ment. They spend considerably more time and resources on making sure that they reduce their holdings of currency or checking account bal ances. At the other end, because of the increased flow of customers, banks would probably be forced to incur additional expenses. For in stance, a bank might have to hire an extra teller or put up an extra ATM. Similarly, as individu als and corporations use the services of their brokers more frequently, brokerage firms would have to spend more resources to deal in a timely fashion with the additional business. This means that if the Fed were to stop accommodating the seasonal variation in money demand and thereby let short-term interest rates rise during Christmas and summer and decline other times of the year, it would in crease the level of transaction costs during Christmas and summer and lower it at other times of the year. However, the net effect of this move would be to increase the level of transaction costs over the course of a whole year. The reason for this is intuitive and quite simple. By letting short term interest rates rise at a time when the economy is in greater need of money, the Fed would force individuals and corporations to conserve on money holdings at a time when the cost of doing so is high. In contrast, by letting interest rates fall when the demand for money is low, the Fed would encourage individuals and corporations to relax their conservation efforts at a time when conserving money bal ances is relatively less costly. In other words, the Fed would be withdrawing money from circulation when the economy has more need for it and would be putting it back in circulation when it has less need for it. Clearly, such a policy would impose an additional cost on the economy. FEDERAL RESERVE BANK OF PHILADELPHIA Leaning Against the Seasonal Wind How big might this cost be? How much more time, effort, and resources would be used to conserve on money holdings if the Federal Reserve did not accommodate the seasonal increase in demand for money? The answer depends on how much short-term interest rates would rise during the period of seasonally high money demand: if interest rates need to rise a lot to induce people to hold interest-bearing assets such as bonds instead of money, it indi cates that the value of resources used up in reducing money balances (the cost of trips to the bank, brokers' fees) is large. Empirical studies typically find that people adjust their money holding very little in response to changes in short-term interest rates.7 This result sug gests that the gains from following a seasonal monetary policy (or the cost of following a nonseasonal policy) may be worth worrying about. But that is not the whole story. A nonsea sonal monetary policy would cause not just seasonal changes in short-term interest rates, but also seasonal adjustments in the price level. Those price adjustments would reduce the size of seasonal changes in interest rates and also reduce the extra transaction costs generated by following a nonseasonal monetary policy. For this reason, and because there is skepticism about the reliability of estimates of how sensi tive the demand for money is to changes in interest rates, it is difficult to draw firm conclu sions about how big the costs imposed on the economy by a nonseasonal monetary policy would be. In any event, regardless of the size of the benefit from a seasonal monetary policy, we now have an answer to the question we posed in the introduction: what is the difference be tween seasonal variability in the rental price of 7For a review of the empirical literature on the sensitiv ity of money demand to interest rates and other variables, see Judd and Scadding's 1982 article. Satyajit Chatterjee shore property and seasonal variability in the rental price of money? In the former case, the seasonal rise in rents reflects a real scarcity of rental space during times of high demand, and the increase in rents is an efficient way of allocating the limited amount of available space to families that value it most. In contrast, the scarcity of money is artificial in that the Federal Reserve can change the quantity of money available at very little cost. Therefore, since families and corporations gain more from a lower rental price of money during Christmas and summer than they lose from a higher rental price of money at other times of the year, it makes sense for the Fed to smooth the rental price of money over the seasons. ARE SEASONAL MOVEMENTS IN INTEREST RATES COSTLY IN OTHER WAYS? Macroeconomists agree that a nonseasonal monetary policy will increase the overall level of transaction costs, but they do not agree on whether there are other costs of following a nonseasonal policy. To see where these disagreements come from, let's take a closer look at the statement that a nonseasonal monetary policy would raise short term interest rates during Christmas and sum mer and lower it at other times of the year. So far we have talked about interest rates without being specific about what type of inter est rates we mean. In reality, there are two distinct types of interest rates, and it is impor tant that we keep them separate. The type that people are most familiar with is the money, or nominal, interest rate reported in the financial columns of newspapers. For instance, if the interest rate on a one-year Treasury bill is listed as 3.4 percent, then each $1 invested in a T-bill today will fetch $1,034 in a year. The nominal interest rate does not adjust for change in the purchasing power of the dollar; that is, it does not take into account that the purchasing power of a dollar available a year from now may be 19 BUSINESS REVIEW less than that of a dollar given up today because the general level of prices in the economy may be higher a year from now. In contrast, the real interest rate does take changes in the purchas ing power of the dollar into account. The real interest rate is calculated by subtracting the inflation rate expected over the maturity pe riod of the asset from its nominal interest rate. For instance, if the annual inflation rate is ex pected to be 3.0 percent, the real interest rate on the one-year T-bill is only 0.4 percent. This distinction between nominal and real interest rates raises two questions about our previous discussion. First, when we asserted that people's demand for money depends on short-term interest rates, which interest-rate concept did we mean? Second, would a nonseasonal monetary policy lead to seasonally varying short-term real interest rates or season ally varying short-term nominal interest rates or both? The first question is easy to answer. People's demand for money depends on nominal inter est rates. Consider, again, the case of Sadie Wherebucks, who must decide how much money to hold in her wallet or checking account and how much to invest in T-bills. As an investor, Sadie is concerned with the real inter est rate she expects to receive on her T-bill investments. By holding money instead of Tbills, she forgoes this real interest rate, and, in addition, her money loses value over time be cause of inflation. Consequently, the total cost to her of holding a dollar is the real interest rate she could have received on the T-bill plus the inflation rate she expects. But this sum of the real interest rate and expected inflation rate is simply the nominal interest rate. Therefore, in deciding how much money to hold it is the nominal interest rate that counts. Unfortunately, answering the second ques tion is not as easy and opinions differ. The classical view is that a change in monetary policy affects only price levels and inflation rates. Real variables, such as real interest rates, 20 MARCH/APRIL 1993 real output, and real investment, are unaf fected by such changes. Therefore, a classical economist would argue that the increase in the short-term nominal interest rates at Christmas that would accompany a nonseasonal mon etary policy would result from lower prices during the Christmas season but higher prices in winter—after Christmas—and in spring. In his view, it is the faster rate of price increase expected between Christmas and spring that leads to the rise in the short-term nominal interest rate during the Christmas season. Simi larly, a classical economist also would expect a nonseasonal monetary policy to cause prices to drop in summer, then rise in autumn, resulting in an increase in the short-term nominal inter est rate in the summer. He would argue, however, that real variables such as output and employment would be unaffected by these sea sonal price changes (see Seasonal Monetary Policy: The Classical View). Since real variables are not affected, no additional costs or benefits result from pursuing nonseasonal monetary policy. Hence, from the classical perspective, seasonal smoothing of interest rates is desirable because it saves on transaction costs without disrupting real economic activity. This conclusion is not shared by monetar ists. Since monetarists adhere closely to classi cal views in regard to their perception of how money supply changes affect the economy, the rejection of seasonal monetary policy by econo mists such as Milton Friedman and Robert Lucas, Jr., is at first surprising.8 However, their reasons for jettisoning a seasonal monetary policy has to do with their views on how the Federal Reserve should conduct its businesscycle policy. Monetarists believe that a sound monetary policy involves implementing steady growth in the supply of money, with a view to 8For a more recent and more emphatic denial by Friedman of the usefulness of seasonal monetary policy, see his 1982 article. For Robert Lucas's view, see his 1980 article. FEDERAL RESERVE BANK OF PHILADELPHIA Leaning Against the Seasonal Wind Satyajit Chatterjee Seasonal Monetary Policy: The Classical View Does the Fed's choice of seasonal monetary policy affect the real interest rate? Classical economists, who see the real interest rate as being determined primarily by real factors, such as the population growth rate, the rate of technical progress, and people's propensity to save, argue that changes in the money supply, after agents have adjusted to it, do not have any effect on the real interest rate. In other words, they argue that while an unexpected increase in the money supply can reduce the real interest rate for a considerable length of time, the rate returns to its original level as the extra money diffuses through the economy. Once the economy adjusts to the new level of money supply, the only effect of a higher money stock is a higher price level. Applied to the choice of seasonal monetary policy, this argument suggests that in the immediate aftermath of a shift to a nonseasonal policy, there will be a period when the real interest rate will be affected. However, as the economy gets used to the new policy, the real interest rate will return to its original level, and the only change will be in the seasonal path of prices. To see how this works, consider the following numerical example. For simplicity, imagine that there are only two seasons: Christmas and spring. Suppose that when the Fed follows a seasonal monetary policy, the consumer price index is 100 and the real interest rate is 3 percent in both seasons. Since there is no change in the price level from one season to the next, the expected rate of price increase is zero for both seasons. Therefore, the real interest rate is 3 percent in both seasons as well. Now suppose that the Fed switches to a nonseasonal monetary policy and refrains from increasing the money supply during the Christmas season. Since the money supply during the Christmas season is now lower than before, the level of Christmas prices will be lower. Suppose that Christmas prices fall by 2 percent, to a level of 98. Because this nonseasonal policy lowers the average stock of money over the year, it will exert downward pressure on prices in the spring as well and those prices will fall, although not by as much as the Christmas price level.3 Suppose then that the spring price level falls by one-half percent, to a level of 99.5. With these new price levels, the expected rate of price increase from Christmas to spring will be 100 x (99.5 - 98)/98 = 1.53 percent, and the expected rate of price increase going from spring into Christmas will be 100 x (98-99.5)/99.5 = -1.50 percent. Since real interest rates do not change, the nominal interest rate will rise to 4.53 percent during the Christmas season and fall to 1.50 percent in spring. aSuppose that the seasonal monetary policy involved a money supply of 100 in spring and 105 during the Christmas season. With the move to a nonseasonal policy, the money stock will be 100 in all seasons, which would make the average money stock over a year 100 as opposed to 102.5 with the seasonal monetary policy. If the monetary authorities moved to a nonseasonal policy but raised the constant stock of money to 102.5, then relative to the prices that prevailed in the presence of seasonal monetary policy, prices during the Christmas season would fall and those in spring would rise. keeping the inflation rate steady and predict able. They view seasonal adjustments to the growth rate of money supply as a nuisance that distracts attention from the more important task of keeping the money supply growing smoothly over time. Thus, monetarists argue that the benefits of a seasonal monetary policy are small compared with the costs of poten tially erratic movements in the money supply occasioned by attempts to "fine-tune" the growth of money stock to match the seasonal movements in money demand. Keynesian economists also differ with the classical view of a seasonal monetary policy, but for entirely different reasons. A Keynesian economist would disagree with both classical economists and monetarists concerning the likely consequences of a move to a nonseasonal 21 BUSINESS REVIEW monetary policy. In the Keynesian view, changes in interest rates that result from an imbalance between the demand for and supply of money show up in both nominal and real interest rates because prices are not perfectly flexible in the short run. The resulting changes in real interest rates affect the aggregate output of the economy by changing aggregate de mand. Therefore, by following a nonseasonal monetary policy, the Fed would drive up real interest rates and thus reduce the real output of the economy to below current levels during the Christmas season and in summer. Therefore, in the Keynesian view, a move to a nonseasonal policy would result in greater seasonal variability in short-term real interest rates and a lesser seasonal variability of output and employment. Would these changes im pose ad d ition al costs on the econom y? Keynesian economists would argue that a non seasonal policy almost certainly imposes costs on the economy that go beyond the transaction costs discussed earlier, but whether it imposes more costs than existing seasonal monetary policy is more difficult to know. To appreciate the Keynesian point of view, it is important to recognize that Keynesian econo mists regard the classical view on the function ing of a market economy as the ideal toward which actual market economies tend, but which they seldom attain. Because of various frictions in the operation of markets, Keynesian econo mists believe that the outcome of an unregu lated market economy is typically quite differ ent from the outcome depicted by classical economists. Consequently, Keynesian econo mists perceive a need for government policies designed to steer market outcomes toward the classical ideal. In the present context, as already noted ear lier, Keynesian economists would challenge the classical assumption that prices are fully flexible over the seasons. They would argue that if the Fed were to stop accommodating the seasonal demand for money, the price level 22 MARCH/APRIL 1993 would tend to fall during the Christmas season and tend to rise in the spring, but not by as much as in the classical argument. Therefore, short-term real interest rates will rise above the classical ideal during the Christmas season and will fall below it in spring; correspondingly, real output and employment will be below the classical ideal during the Christmas season and above it in spring. Consequently, Keynesian economists would feel the need for a monetary policy that works to reduce seasonal fluctua tions in short-term real interest rates. In other words, they would perceive the need for a seasonal monetary policy.9 That having been said, it does not follow that Keynesian economists would necessarily en dorse the Fed's existing seasonal policy. The Keynesian objective is to get to the classical ideal, but the Fed's current policy may result in too much seasonal variability in output and employment and too little seasonal variability in short-term real interest rates relative to it. In the absence of quantitative information on what the ideal seasonal pattern of short-term real interest rates and output really is, it is not possible for a Keynesian to know whether the Fed's existing seasonal monetary policy is the best one. CONCLUSION The Federal Reserve's policy of accommo dating seasonal movements in money demand originated in an attempt to eliminate recurrent financial panics. At the time the Federal Re serve System was established, it was widely felt that the seasonal outflow of bank reserves that occurred in the fall and spring jeopardized the liquidity of the banking system and raised fears on the part of depositors that banks would be unable to honor their deposit liabilities. A key 9See Gregory Mankiw and Jeffrey Miron's 1991 article for a detailed discussion of the Keynesian view on the usefulness of seasonal monetary policy. FEDERAL RESERVE BANK OF PHILADELPHIA Leaning Against the Seasonal Wind objective of the Federal Reserve System was to allow banks to borrow additional reserves dur ing these months of heavy currency demand so that the natural pace of commercial activity would cease to be a threat to the banking system. This practice of increasing bank reserves and the supply of currency during a time of season ally high demand continues to the present. However, given the institutional changes that have occurred in the banking environment since 1914, most notably the introduction of federal deposit insurance, it is doubtful whether a seasonal monetary policy is needed to protect the banking system from panics. Therefore, a different justification of seasonal monetary policy is needed. This article has suggested that a justification for seasonal monetary policy may be found in Satyajit Chatterjee the argument that such a policy, by smoothing the path of short-term nominal interest rates, serves to reduce transaction costs. The article also pointed out that a classical economist would view the reduction in trans action costs as the only significant impact of a seasonal monetary policy and would therefore argue in favor of such a policy. In contrast, monetarists would argue in favor of eliminat ing seasonal monetary policy on the grounds that it interferes with what they see as the more important task of keeping the money stock growing smoothly and predictably over time. Keynesian economists would concede that ex isting policy may be too seasonal but would argue that some degree of seasonality in money supply is desirable; therefore, they would cau tion against abandoning such a policy. 23 REFERENCES BUSINESS REVIEW MARCH/APRIL 1993 Barsky, Robert B., and J.A. Miron. "The Seasonal Cycle and the Business Cycle," Journal of Political Economy 97 (1989), pp. 503-34. Clark, Truman A. "Interest Rate Seasonals and the Federal Reserve," Journal of Political Economy, 94 (1986), pp. 76-125. Diller, Stanley. "The Seasonal Variation in Interest Rates," in Essays on Interest Rates, Jack M. Guttentag, ed., National Bureau of Economic Research, New York, 1971, pp. 35-133. Friedman, Milton, and A. J. Schwartz. A Monetary History of the United States 1867-1960. Princeton: Princeton University Press, 1963. Friedman, Milton. A Program for Monetary Stability. New York: Fordham University Press, 1959. Friedman, Milton. "Monetary Policy—Theory and Practice," Journal of Money, Credit and Banking, 14 (1982), pp. 98-118. Hollander, Jacob H. Bank Loans and Stock Exchange Speculation. National Monetary Com mission, S. Doc. 589, 61st Congress, 2nd Session, 1910. Judd, J.P., and J.L. Scadding. "The Search for a Stable Money Demand Function," Journal o f Economic Literature, 20 (1982), pp. 993-1023. Kemmerer, Edwin W. Seasonal Variation in the Relative Demand for Money and Capital in the United States. National Monetary Commission, S. Doc. 58 8 ,61st Congress, 2nd Session, 1910. Laughlin, Laurence J. Banking Reform. Chicago: National Citizens League for the Promotion of Sound Banking, 1912. Lucas, Robert E. Jr. "Rules, Discretion and the Role of the Economic Advisor," in Rational Expectations and Economic Policy, Stanley Fischer, ed., Chicago: The University of Chicago Press, 1980, pp. 199-210. Mankiw Gregory N., and J. Miron. "Should the Fed Smooth Interest Rates? The Case of Seasonal Monetary Policy," Camegie-Rochester Conference Series on Public Policy, 34 (1991), pp. 41-70. Miron, Jeffrey A. "Financial Panics, the Seasonality of the Nominal Interest Rate, and the Founding of the Fed," American Economic Review, 76 (1986), pp. 125-40. Sprague, O.M.W. History of Crises Under the National Banking Act. National Monetary Commission, S. Doc. 538, 61st Congress, 2nd Session, 1910. 24 FEDERAL RESERVE BANK OF PHILADELPHIA Philadelphia/RESEARCH Working Papers The Philadelphia Fed's Research Department occasionally publishes working papers based on the current research of staff economists. These papers, dealing with virtually all areas within economics and finance, are intended for the professional researcher. The papers added to the Working Papers series in 1992 and thus far this year are listed below. To order copies, please send the number of the item desired, along with your address, to WORKING PAPERS, Department of Research, Federal Reserve Bank of Philadelphia, 10 Independence Mall, Philadelphia, PA 19106. For overseas airmail requests only, a $2.00 per copy prepayment is required; please make checks or money orders payable (in U.S. funds) to the Federal Reserve Bank of Philadelphia. A list of all available papers may be ordered from the same address. 1992 No. 92-1 Leonard I. Nakamura, "Commercial Bank Information: Implications for the Structure of Banking." No. 92-2 Shaghil Ahmed and Dean Croushore, "The Marginal Cost of Funds With Nonseparable Public Spending." No. 92-3 Paul S. Calem, "The Delaware Valley Mortgage Plan: An Analysis Using HMD A Data." No. 92-4 Loretta J. Mester, "Further Evidence Concerning Expense Preference and the Fed." (Super sedes No. 91-6) No. 92-5 Paul S. Calem, "The Location and Quality Effects of Mergers." No. 92-6 Dean Croushore, "Ricardian Equivalence Under Income Uncertainty." (Supersedes No. 90-8) No. 92-7 James J. Me Andrews, "Results of a Survey of ATM Network Pricing." No. 92-8 Loretta J. Mester, "Perpetual Signaling With Imperfectly Correlated Costs." No. 92-9 Mitchell Berlin and Loretta J. Mester, "Debt Covenants and Renegotiation." No. 92-10 Joseph Gyourko and Richard P. Voith, "Leasing as a Lottery: Implications for Rational Building Surges and Increasing Vacancies." No. 92-11 Sherrill Shaffer, "A Revenue-Restricted Cost Study of 100 Large Banks." (Supersedes FRBNY Research Paper No. 8806) No. 92-12 Paul S. Calem, "Reputation Acquisition and Persistence of Moral Hazard in Credit Markets." (Supersedes No. 91-5) No. 92-13 Sherrill Shaffer, "Structure, Conduct, Performance, and W elfare." (Supersedes No. 90-27) 25 Philadelphia/RESEARCH Working Papers No. 92-14 Loretta J. Mester, "Efficiency in the Savings and Loan Industry." No. 92-15 Dean Croushore and Shaghil Ahmed, "The Importance of the Tax System in Determining the Marginal Cost of Funds." No. 92-16 Keith Sill, "An Empirical Investigation of Money Demand in the Cash-in-Advance Model Framework." No. 92-17 Sherrill Shaffer, "Optimal Linear Taxation of Polluting Firms." No. 92-18 Leonard I. Nakamura and Bruno M. Parigi, "Bank Branching." No. 92-19 Theodore M. Crone, Sherry Delaney, and Leonard O. Mills, "Vector-Autoregression Forecast Models for the Third District States." No. 92-20 William W. Lang and Leonard I. Nakamura, "'Flight to Quality' in Bank Lending and Economic Activity." No. 92-21 Theodore M. Crone and Richard P. Voith, "Estimating House Price Appreciation: A Compari son of Methods." No. 92-22 Shaghil Ahmed and Jae Ha Park, "Sources of Macroeconomic Fluctuations in Small Open Economies." No. 92-23 Sherrill Shaffer, "A Note on Antitrust in a Stochastic Market." No. 92-24 Paul S. Calem and Loretta J. Mester, "Search, Switching Costs, and the Stickiness of Credit Card Interest Rates." No. 92-25 Gregory P. Hopper, "Can a Time-Varying Risk Premium Explain the Failure of Uncovered Interest Parity in the Market for Foreign Exchange?" No. 92-26 Herb Taylor, "PSTAR+: A Small Macro Model for Policymakers." 1993 No. 93-1 Gerald Carlino and Robert DeFina, "Regional Income Dynamics." No. 93-2 Francis X. Diebold, Javier Gardeazabal, and Kamil Yilmaz, "On Cointegration and Exchange Rate Dynamics." No. 93-3 Mitchell Berlin and Loretta J. Mester, "Financial Intermediation as Vertical Integration." No. 93-4 Francis X. Diebold and Til Schuermann, "Exact Maximum Likelihood Estimation of ARCH Models." Digitized for 26 FRASER FEDERAL RESERVE BANK OF PHILADELPHIA FEDERAL RESERVE B A N K O F P H ILA D E LP H IA BUSINESS REVIEW Ten Independence Mall, Philadelphia, PA 19106-1574 Address Correction Requested