D1~IF~ SEPTEMBER/OCTOBER 0995
Christopher J. Neely is an economist at the Federal Reserve Bank of St. Louis. Geoffrey F. Wood is a professor at City University Business School in Landon and was a visiting scholar at the Federal Reserve Bank of St. Louis when this article was begun. Kelly M. Morris provided research assistance.
“I argue for the inflation target because I Deflation and fear the consequences of having to aim to deflate the econotny half the time, Real Economic which is what the price level approach requires.” Activity Under Since the end of World War II, year-over- year declines in theprice level have been rare the Gold in the industrialized world; during the period of the gold standard, however, both long Standard downward trends in the price level and much shorter periods of falling price levels were 3 common. Ironically; although Irving Fisher Chrktoplier J. Neely and advocated a price level target precisely to Geoffrey E. Wood avoid the protracted downward (and upward) swings in the price level observed under a In the past few years, several countries gold standard, the experience of this period have announced explicit target ranges for provokes, in part, the criticism of price level I inflation. New Zealand did this in 1990, targeting today Perhaps more important for Canada in 1991, the United Kingdom in 1992, these beliefs about deflation is the deflation- and Sweden and Finland in 1994. Even when ary period (not examined here) from 1929 an inflation target is achieved, the future through 1933, in which the price level fell by price level is not easy to predict because none 20 to 30 percent. Bernanke (1995) argues of these countries has committed itself to persuasively that this price decline, caused reversing the consequences of shocks to the by the U.S. determination to stay on the gold price level. Indeed, in New Zealand there is standard, was a major contributor to the A price level is o weighted overage an explicit commitment root to reverse cer- severity of the Great Depression. This article of prices in o country. Price level tain such shocks. reexamines the facts surrounding temporary targets may be either constant over One alternative to inflation targeting periods of deflation that occurred under the twe (static) or hove a trend. In is price level targeting.’ The adoption of a gold standard from 1870 to 1913. We first this paper, we ose pica level torget~ngto refer to o static price constant price ]evel target would have several describe the behavior of price, money and level target. The sboded insert at advantages over an inflation target. Chief output data, then perform some simple tests pp. 34 ond 35 distinguishes price among these is that consumers and firms to determine whether output growth grew level ond inflation targets. could write simpler contracts and make more slowly during periods of falling prices long-run plans without worrying about and whether knowledge of a falling price 2 Pine Finopciol Times, Jane 24, inflation. Price level targeting also may level would, in fact, have helped predict lower 1994. Note that Fischer refers to static price level target. A price avoid the “time-inconsistency” problem output growth. Although we must be cautious o level target with a positve trend of an inflation targeting regime in that the about drawing conclusions from 100-year—old world on~reqeire the monetary monetary authority would have less incen- data generated under a much different mone- outhority to “disinf late’ half the tive to inflate the economy in a one—time tary regime, another look at this experience twa, thor is, to mr a rote of inflo’ bid to increase output temporarily Under is warranted because several countries have toe below the long’ron trend. a price level target, any “surprise” inflation adopted policies that are likely to he associ- Disinflaton is not the only poten’ must he reversed. ated with temporary periods of deflation. tol deowbock of price level targets. Critics of price level targeting argue The next section briefly reviews why Some oppose them becoose they that making a commitment to reverse sur- deflation may affect real output. A descrip- rnighr lend to groorer short-rev prise increases in the price level is undesir- tion of our data set and an explanation of nolotlity in the inflator rate. able because a fall in the general price level, our statistical tests follow, We then report °Periods in ‘which prices fall one or deflation, can have harmful effects. One the results of our tests, before concluding yeorwrer-yeor basis ore considered such critic, Stanley Fischer, put it this way: with some ideas for future work. pedods of deflaton.
flDERAL RESERVE BANK OF ST. LOUIS
27 lltYIt~ SIPTEMRER/OCTOEEE 1995
3., despite evidence to the contrary, many econ- 4 An evcellent neview of these issoes omists continue to believe that some prices can he found in Mc(ollum 119891, are inflexible downward and that even tem- Chopter 9. Ohoninn and Stockmnn It is now widely accepted that there is porary periods of deflation might reduce 2 (forthcoming) consider rho conso’ no long-term trade-off between inflation and output through this channel. qeences of monetony shocks for the output or employment; the existence of a Bernanke andJames (1991) argue economy when some, hot not nIl, short-run trade-off, on the other hand, is not that deflation might alternatively affect prices one sticky. That paper also general])’ denied. There are several explana- the economy by increasing the real value of sets out severn enplonotions fan price stickiness in odditon to those tions for this trade-off: lags between actual and nominally denominated debt. For example, reviewed in McCollnm. expected inflation (see Flume, 1752; Fisher, a2percent annual deflation would translate 1926; and Friedman, 1968); misperceptions a nominal interest rate of 4 percent into a Borm 119951 finds henef its of about relative and general price shifts (Lucas, real interest rate of 6 percent. Increasing lower inflation in the form of higher 1972); and staggered wage or price setting the real rate of interest might promote debtor long’nnn growth in e cross’coentry (Fischer, 1977; Taylor, 1980).~ None of these insolvency and financial distress. study. Here, vie are concerned with shorr’ron eflects. theories, however, predicts that lowering the The opposition to price level targeting price level is more costly than lowering infla- from those who fear the results of deflation, See Wynne 119951 far a sorvey of tion, Nevertheless, prices have not fallen (by either because of downward price rigidity or price stckiress and Croig (1995) anything more than a trivial amount) in any the consequences of debt-deflation, makes fan evidence on wage rigidity. major economy since 1945. the stud) of the historical association between Advocntes nf this view might point The means by which deflation might output and deflation worthwhile. A review oat that neol woges nose sobston- reduce output, however, are often not explicitly of the eiodence nould he a first step in con- 3 tally dening the severe deflotion of stated. One view is that deflation interferes siderong whether a central bank should now 1 the Greet Depression. with the adjustment of relative prices because adopt a p ice ci el target. Varioes series enisted before pablo nominal wages or some prices do not adjust ceton of thor noleme, hot thoy downward easily If wages and/or prices are hod deficiencies that wene remo’ sticky downwards, a negative demand shock died los well as somo new data will tend to cause persistent unemployment permitted) by Copie nnd Wehbec as prices and wages are slow to fall as required See Copie and Webben fon discvs- to clear markets. With a sufficiently high sion of pnevioes series deficiencies inflation trend, relative prices can adjust to u e tu o set of data Thc fir t con- and how they ere remedied. lhe a negative demand shock without any actual sists of 44 annual oh eri.anons on money~ cneciol point is that these previoos prices having to fall. Because nnarkets work prices, interest rates and output tn the seies contained n sponioes trend. better with a little inflation, according to this United Kingdom fronn 1870 to 1913 The o There has recently been some dis- view output will be less variable over busi- period 1880-1913 is generally considered cossion of the neliobility of that net ness cycle horizons and, perhaps, even the heyday of the cla socal mt rnational pet series—see the intenchenge higher in the long run. gold standard. We end ahe sample before hetweenlrensley 11986, 19891 Critics of the theory of downward price the beginning of World \\ar I in 1914. The and Feinsteir (19891, and the dis’ rigidity point out that many wages and prices source for the monetars sero s o Capie and cnssion in Crafts, Leyboenne end Wehher (1985). ihe anterest atc is a short- Mills (19891—bet there seems to do, in fact, decrease, and that the extent to be general agreement that whotev’ which prices are sticky depends on whether term one from the last quarter of each year. er its deficiencies, it is the best people expect inflation. An atmosphere of The output sene is Feunstemns (1972) com- ovoiloble. overall price stability will make people more promise estnnn~teof GDP and, therefore, We dropped Jopon from the sample willing to accept reductions in their wages ho tcnplacnt prncc deflator Os used as rhe proce or prices. scrod All data are annual to conform to hecoose it did not hone a metelic There is mixed evidence from rnicroeco- standard dying the 19th centur~nnd thc n des nty of u ing annual GDP data. leceese its nonionol honking ond nonnic data on the idea that prices are sticky; Tb econd data set consost of 44 annual finenciol system man lost forming certainly, some prices change tnore frequently output and onflatcon ohs rs utnon 1870 to (see 8ockos and Kehne, 1992). than others. There is, however, little evidence 1013) Ironu ninc of 10 on lu trnaln ccl cc untries 3 Uniqnely, lopan’s gnnwth onder of asynnmetry in price stickiness. Blinder onnpalcd fo comparoson of nnt rnational lolling pices (5.4 percent) mos (1991) presents the results of a survey in which hu one cicles hi Backu md 1< hcc (1992), sobstontinily higher non ts grreti firnus report asyonrnetric price rigidity He frocn hicln onor complete d scrnptnon of ondnr rising prices 11.5 pencenol. finds greater upward rigidity Nevertheless, the dat as -vuanlablc
EEDEEAL BIEERVI BANK OF ST. LOUIS
28 II FYI F~ SEPTEMSER/UCTOBEE1 995
Time Series of the Levels of the United Kingdom Data
Output Implicit Price Deflator 110 too 90 80 70 60 50 40
Interest Rates M3 6.4 1200- -: - 4/’ ~<~~ ,, --‘ 4 2’ <3. 2 5.6 1100 - 4- <3< ~‘344 ,,< 2 3 <3 4- ‘43 <3 ~ -\ 4.8 1000 - <1’3’< \ :< 2’ .3 /‘ ‘2 ‘<44. 2 4.0 900 ‘4’ ‘2 - - <<<3)’ 44 34- , 3 800 - 3.2 --
2 ,‘< - 2.4 700 - 2-3 \/
1.6 600- ‘1 , ‘ <‘ 4 4 0.8 500-t A-. i i,;i..i r~.,. . 1870 74 78 82 86 90 94 98 02 06 101913 1870 74 78 82 86 90 94 98 02 06 101913
TL y . )enes 4 42 States and France returning to the gold stan- Figures 1 and 2 display the time series dard, raising the demand for and price of gold. of the log levels and log differences of the four The nominal interest rate seems to display United Kingdom series from 1870 to 1913. typical cyclical fluctuations around a The shading in the figures represents periods stationary mean. in which the price level fell (not periods of recession). The monetary series, M3, and the output series generally grew over time, The 3’ —‘ ~< price deflator series does not display the consistent rise typical of modem price indices; Figure 3 shows the higher average rates rather, periods of risingand declining prices of inflation, displaying a scatterplot of the mean seem to he nearly equally common, The long rates of output growth vs. mean inflation rates downward trend in the price level until 1896, for each ofnine countries from the Backus and followed by an upward swing through the end Kehoe data set for each of the two subperiods of the sample in 1913, was caused hy fluctu- (1870-96 and 1897-1913). The figure shows ations in the world supply of and demand for that average inflation rates were uniformly gold. For example, the downward drift in lower in the first period (1870-96) than they prices until 1896 was partly due to the United were in the second period (1897-1913).
FEOIRAL RESEEVE BANK OF St LOUIS
29 A[YI1~ SEPTEMSER/OCTOEER 1995
~
Time Series of the Differences of the United Kingdom Data
Output Growth Inflation 0.075
0.050
0.025
0.000
—0.025
-0.050
Consistent with the idea that deflation reduces growth over short periods. To see this, we output growth, the mean levels of output sort the data on output growth by the rise or growth also appear to he lower during the first fall of prices. For the purpose of categoriza- period. Curiously across countries there seems tion, we define a deflationary period as any to be a negative relationship between output year in which prices fell; we make no dis- and price changes in the first period and a pos- tinction between the episodes on the basis itive relationship in the second. of length, severity or cause. For the United Kingdom data, five of nine deflationary episodes lasted more than one year, and Output Growth and Deflation three lasted more than two years. Over Shart Horizons Table 1 (page 32) provides some Examining inflation and output growth summary’ statistics for data from the nine over the two long subperiods is a convenient countries used by Backus and Kehoe for the way to examine the relationship between period 1870-1913, The first two columns average inflation and average output growth provide the unconditional means of output over longer periods. It does not, however, growth and inflation, The third column get directly at the question of whether price shows the percentage of the time that prices declines were associatetl with lower output were rising during the sample period. Mean
FEDERAL RESERVE BANK OF ST. LOUIS
30 DFYIFW SEPTEMBEB/OCTOBERI 995
price declines were of comparable magnitude to mean price rises, and periods of mild price rises were only slightly more common Mean Output Growth In the First Period (1870-96) and the Second than periods of declining prices; the data Period (1897-1913) show that prices rose about 46-67 percent Output growth of the time during the sample. 6.5 - Figure 4 is analogous to Figure 3in Canada 2 + that it depicts mean output growth for the 5,5 - nine countries from the Backus and Kehoe 4.5 - sample, conditioned on whether prices rose Australia 2 4-1152 +1151 + or fell, Again, the means of output growth 3,5 - Denmark 2+ during periods of rising prices appear gener- Denmark I + Germany !~. Germany 2 4-Sweden 2 + + .g’Norway2 2.5 - Australia 1 +Sweden I ally higher than the means during periods Canada r’ Italy 2 UK I 4-Norway I of falling prices. This positive relationship 1.5- ~UK2 between price changes and output growth + Naly 1 is again consistent with the idea that defla- 0.5 — — —2.0 -1.5 -1.0 -0.5 0 0.5 1.0 1.5 2.0 2.5 tionary periods were associated with Influfion relatively hard times. Nate: Sample It 1870-96, sample 2 = 1897-1913.
<04447244ni3 344337720 2r0’4/4c4’334i272n<5343r72< <<‘<0< <.4’4<4’4’<’<<~4-. 4323 <<“<‘8<33<017<737< <3433<’330 33<02033<3073 3300.0 nz.i FEDERAL RESERVE RANK OF St LOUIS 31 I11~I[~ SEPTEMRER/OCTORRR 1995 International Output Growth and Inflation Statistics Under Rising and Falling Prices Proportion of Years Prices Unconditional are Rising Statistics (percent) Rising Prices Falling Prices Mean Mean Mean Mean Output Mean Output Mean Output Inflation Growth Inflation Grawth Inflation Growth Australia 0.23 310 51.16 3.3/ 3.60 —3.06 25/ Canada 059 3.90 61.44 2.37 3.98 351 3./5 Denmark 01-I 3.20 48.84 1.67 3.5! —1.86 2.91 Germany 0.61 266 58.11 2.69 279 —2.29 2.48 Italy 0/4 115 58.14 360 2.58 3.24 -013 Norway 011 2.1/ 5814 3.35 2.68 —2.96 145 Sweden 066 2/3 60.47 3.07 34/ —3.07 1.60 United Kingdom 0.08 1.88 51.16 183 2.16 1.16 158 United States 0.26 4.03 4651 280 4.03 —2.93 4.03 Tests of the Equality of Mean Output Growth Under Inflation VSR Deflation Test of Equality of Mean Output Test of Equality of Mean Output Giowtk Between the Subperiods Growth Conditioned on 1870-96 and 1897-1913 Inflation or Deflation Test Statistir p-value Test Statistic p-value Australia I 44 0.23 1.06 0.30 (anada 5.45 000 0.05 082 Denmark 0.32 0.5/ 036 055 Germany 000 1.00 0.10 0.76 lnth 3.19 0.07 7.36 001 Nonvai 0.82 03/ 1.52 022 Sweden 066 0.42 350 006 UK 0.09 0.11 033 0.56 US. 0.16 069 000 1.00 Aggreqate 1101 028 1532 0.08 than 0.1 or 0.05 are usually interpreted as the overall mean output growth for all nine meaning that we can reject the idea that the countries for the second period is the same means are the same. A lower p-value means as the overall mean output growth for the that it is less likely that the means are the first period. The p-value from such a test same. These tests of equality of means reject is 0.28 (see the third column, last row of the idea that the conditional means are equal Table 2). which strongly suggests that it is for Canada and Italy hut not for the other very possible that the data were generated countries if our criterion for rejection is a by processes with equal means. That is, for p-value less than 0.1. only two countries could we conclude that If we pooi the observations from all aggregate mean output growth in the second the countries, we can test the hypothesis that period was statistically significantly higher FEDERAL RESERVE BANK OF ST. LOUIS 32 DF~IFW SEPTEMBER/OCTOBERI 995 ~8~ ~ 4 Fit of Asymmetric Vs. Symmetric Tests of Linear Forecasting Models of Prices and Output Ability of Price Changes and Growth Preferred Model Under the Output Akaike Criterion Schworz Criterion Granger Causality Statistics (p-value) Australia symmetric symmetric lest that Price Changes Test That Output Growth Canada symmetric symmetric Do Not Help Forecast Does Nat Help Forecast Denmark asymmetric asymmetric Output Growth Price Changes Germany asymmetric symmetric Italy symmetric symmetric Australia 20.53/ / 36/ Norway symmetric symmetric (0.000) (0.010) Sweden symmetric symmetric Canada 3.8/5 8.848 UK asymmetric symmetric (0018) (0.000) US symmetric symmetric Denmark 1.848 5.168 (0 1/2) 10.00/) Germany 5493 0.343 than the mean of output growth in the (0.024) (0562) first period. Italy 2.262 0.06! Columns four and five of Table 2 present 10119) 10941) results of similar tests for equality of means Norway 134/ / 245 for the data in Figure 4. For Italy and Sweden, 10.253) (0.010) we reject the idea that the mean of output Sweden 0210 5011 under inflation was the same as that during (0.649) (0030) deflation. For this test, however, aggregating US 2.020 1443 the observations across countries leads to the (0.130) (0248) conclusion that output growth was signifi- UK 1.346 2.089 cantly lower in a statistical sense during peri- (0253) (0.156) ods of deflation. The p-value for the test of that hypothesis is 0.08 (see the fifth column, last row of Table 3). own lagged values. VARs are a commonly used, general method of modehng the dynamic Have an relationship between macroeconomic variables. ‘C” In the first system of equations, we treat positive and negative price changes as two The previous analysis described the different variables and allow them to influence period-by-period relationship between aver- output growth (and each other) differently” age output growth and average price changes In the second system, we treat price changes conditioned on the sign of the price changes. as one variable, forcing positive and negative Macroeconomic variables, however, influ- changes to have mirror-image effects on out- ence each other not just contemporaneously put growth. Then we examine which model hut also over time. The symmetry of the fits the data better. 0< JIve three veiohles in the system dynamic relationship between output growth We judge the fit of the systems accord- ore output growth, positive price and price changes is important, because an ing to two commonly used criteria: the changes (ItIFtOP) and regotrve essential implication of the idea that deflation Akaike information and the Schwarz infor- pricechanges IDEFOPi, where is harmful to output is that output reacts mation criteria. These measures of the fit INFLOP FEDERAL RESERVE BANK OF St LOUIS 33 ll1tdI[~ SEP?EMEEE/OCTOBER 1995 PRICE LEVEL VS. INFLATION TARGETING Price stability has attracted a lot of attention lately Unfortunately the important choice between inflation and price level targeting has been neglected. Either would lead to a lower and more stable inflation rate than we have observed over the past 25 years, hut there is a fundamental distinction between the two. Price level targeting “corrects” past errors in monetary policy while inflation targeting ignores them. To make this distinction more concrete, consider the problem of a monetary authority with an inflation target of zero to 2 percent in which the 1995 inflation rate is 3 percent, 1 percentage point above the target range. In choosing monetary policy for 1996, the authority will aim, as usual, for an inflation rate of zero to 2 percent. It will not try to make up for past errors. In contrast, if the same monetary authority has targeted a static price level (zero percent inflation on average) and observes 1 percent inflation, it will have to try to reduce the price level by I percent in the years ahead. This difference makes price level targeting a long-run commitment in ways in which inflation targeting is not. There are three major consequences of this divergence between the two, First, the average rate of inflation over a long horizon can be predicted very well under a price level targeting regime; it is less certain under an inflation targeting regime.r Advocates of price level targeting often point to the greater certainty of the price level (average inflation rate) in the long run as an advantage. As the accompanying chart shows, uncertainty about the future price level associated with an inflation targeting range of zero to 2 percent increases as the time horizon grows. In contrast, the level of uncertainty asso- ciated with aprice level target is constant (and small), even over long time horizons, For example, an investor evaluating the real return on, or the present value of, a project can do so much more easily because the price level can be predicted over long periods. Second, an important theoretical advantage of the long-run nature of price level targeting is that by being a multi-period commitment, it does not suffer from the time-inconsistency problem described by Barro and Gordon (1983). In their model, a monetary authority has an incentive to produce a one-time monetary stimulus that results in a burst of output Tire expected predictor error for future overoge iofloton would go to zero under a price level targeting regime as the time horizon increases, while it would remain constant under an of lotion targeting regime. 2 is favored.’ These tests provide mixed evi- test directly whether the deflation itself was dence on the hypothesis that price changes the cause of lower growth, we can test whether have an asymmetric effect on output for the the falling price level helped to forecast it. countries considered here. Such a test of hnear forecasting ability is called a test of Granger causality If price changes improve the forecasts of output growth, they 4/Ut> Uu0<22Ui2U2U ‘Ut,> are said to “Granger-cause” output growth. The idea is that if a falling price level causes 0< leroose the Akoihe and Schworz Previously we showed that, under the lower growth, then it should precede output criterio ore rov’resred model selec. gold standard, output growth tended to be growth and he useful in forecasting it. Note, tior cniterir, thep ore not formal lower than average during periods of deflation. however, that if a third factor is causing both stotstcol tests and do not hove ‘sigrificonce levels.’ Instead, frep Then we showed at least some evidence in deflation and lower growth, this statistical irformolly test for stotstcel sigrifi favor of the hypothesis of an asynnmetric procedure can find that deflation helps fore- once by perolizirg more complex dynamic relationship between price changes cast lower growth, even when it is not the models. and output growth. Although we cannot cause of lower growth. FEDERAL RESERVE RANK OF ST. Louis 34 SEPTEMBER/OCCOSERI 995 growth and inflation. Because the puhlnc under- stands this incentive, it Price Level reacts in such a40 way that the authority 35 inflates each period hut 30 fails to increase output. 25 A price level target solves the time-inconsis- 15 tency problem by requiring 10 the monetary authority’ to Area of inFlation targeting certainty correct past errors. The authonty has no xncentrve ‘~0 have to reduce the price - level hack to the target 1 level, l’herefore. a price level target should he onore credible than an inflation target. A third major difference motivates the subject of this article. A static price level target requires the monetary authority to reduce the price level in response to surprise increases. While an inflation rate target may produce occasional reductions in the price level accidentally, they will he rare if the average inflation rate is high relative to the volatility in inflation. In contrast, under a static price level target. price changes will he negative roughly half the tione. A hybrid of targeting inflation and targeting a static price level is targeting a small upward trend in the price level. Such a system has the long-tenn predictability of a static price level target hnmt does not require the monetary authority to correct past upward deviations in the price level with deflation. This orgrwerre ossemes that even onticipeted debtors will be os cosfy Os the benefit geined from the initial irforion, To test whether price changes innprove this period. We should emphasize that rejec- the forecasts of output growth, we first fore- tions of Granger causality tests are a neces- cast output grosvth using only its own lags. sary’ hut not sufficient condition to determine Then we add lagged price changcs as another that output growth is not “caused” by price explanatory variable to see if their inclttsion changes. Once again, the data provide us with rinproves the forecasts. The second column mixed results on the idea that price changes of Table 4 displays the test statistic and p-value have an asymmetric effect on output. (significance level) froon the tests that price \Ve can also investigate svhether changes do not Granger-cause (help forecast) output growth helps forecast price changes output growth. For Australia, Canada and in this system. Fconomnic commentators Germany we reject the null hypothesis than commonly suggest that price pressures lagged values of price changes do not improve (or the lack thereof) are due to the level of the forecasts of output growth. ln other words, output growth, employment, capacity utiliza- the data suggest that price changes do help tion or some other real variable. The test sta- forecast output growth for three countries in tistics and p-values from the tests that output FEDERAL RESERVE BANK OF ST. LOUIS 35 V [\‘I [~ SEPTEMRER/OCTOBRR 1995 growth does not help forecast future price From 1931 to the trough of the Depression, changes are in the third column of Table 4. the price level fell by 20 percent to 30 per- These statistics indicate that output growth cent in countries that stayed on the gold does help forecast price changes for Australia, standard, while falhng less than 2 percent Canada, Denmark, Norway and Sweden. in Sweden (from 100 in September 1931, Although these results do not shed light when the Rikshank started targeting the directly on a possible asymmetric response price level, to 98.4 in October 1.933). of output to price changes, they are consis- Unlike a gold standard, price level targeting tent with traditional Phillip~curve explana- permits control of the price level through tions of inflation. the money supply e.tt,lnmx*nar4/= ‘O Because we have only a small sample, A number of countries, including the predictive power of one variable on another New Zealand, Canada and the United must be very strong for tests for Granger Kingdom, have recently announced explicit causality to find a relation. Weaker but target ranges for inflation, Such a policy important relations may not be found at all. has also been suggested for the United Statisticians would say that tests of Granger States, Others have suggested that we target causality may have “low power.” Another the price level instead of the rate of inflation. complication is that both price and output One potential reason to oppose this sugges- changes may result from some third factor, tion is that such a policy would necessitate which has been left out of the analysis. that the monetary authority reduce the level No matter how confident we are that we of prices, that is, deflate the economy to off- understand how these economies functioned set any transient, positive shocks to the price TOO years ago, we must be cautious about level. The historical association between using historical data to answer pohcy questions deflation and bad economic performance today For example, economic structures such has led some economists to reject price as the wage-setting mechanism, the degree of level targeting as had policy flexibility of the labor market and credit allo- We find that lower output growth cation mechanisms—~-allof which may influ- was associated with periods of deflation ence how changes in the money supply in nearly all the countries examined. For translate to changes in the price level—have a majority of the countries, the dynamic changed a great deal in the last century relationship between price changes and Even methods of data collection are much output growth appeared to be synnmetric, different now. and price changes did not help forecast Finally we remind the reader that output growth. There is more evidence, the economists who observed this episode however, that output growth forecasts first-hand believed that deflation was a price changes. disruptive factor causing lower output Ultimately a final conclusion about growth. Many recommended a price level the desirability of a price level target requires target as a remedy for that prohlem.’~ more complete economic modeling than Presumably the finite sample variance of we have attempted. What we have presented the price level would be much different are some simple facts about deflation and under a price level targeting regime than it output that are touted as reasons to reject a was under the gold standard. Some evidence particular type of price stability Economists in favor of this view can be found hy com- who support price level targeting must paring Sweden~experience with prices ntake the case that the teanporary periods of during the Great Depression with that of deflation necessary to maintain long-term countries that stayed on the gold standard. price stability would be fundamentally nx See StoNe Money: A History of the Sweden left the gold standard in 1931 and different than those observed under the Moveenneertby Inning Fisher, 1914. began to target the consumer price index. gold standard. FSDSEAI. RESERVE SANK OF ST. LOUIS 36 VF~I[~ SEPTEMSES/OCTOSER1 9,5 “=“u”me 0O”, ii.tt.mrsItmm~iyi.CES Fischer, Stanley. ‘Lang-lirm Contracts, Rational Expectations, and the Optional Money Supply Rule,” Journol of Political Economy Ruckus, David K., tad Patrick]. Kehoe, “Internalionrnl Etidence on the (February 1977), pp. 191-205, Historical Properties of Business Cycles,” The American Economic Review (September 1992), ~ 864-88. FisheçAdelphiIrving.Company,Stoble 1934.Money: AHfstory of the Movement Borro, Roberti. “Inflation and Economic Growth,” Bank of England Quarterly Review (May 1995), pp. 166-76. . The Purchasing Power of Money. Matmillon, 1911. ______and David B. Gordon. “Rules, Discretion tad Reputation Friedman, Milton. “The Role of Monetary Policy,” The American in a Model of Monetary Policy,” Journalof Monetary Economics Economic Review (March 1968), pp. P17. (July 1983), pp. 10P2l. Greasley, David. “British Wages and Income, 1856-1913: A Revision,” Bernonke, Ben S. “The Macroeconomics of the Great Depression: Explorou~six Economic Htitory (Apri 1989), pp. 248-59. A Camporotite Approach,” Journal of Money Credit and Banking 8 (February 1995), pp. 1-28. ______, ‘British Economic Growth: The Paradox of the l BOs and the Timing of the Climacteric,” Explorations in Economic History ______and Harold James. “The Gold Standord, Deflation, and (October 1986) pp. 41 6-44. Financial Crisis in the &eat Depression: Alt International Comparison,” in R. Glenn Hubbard, ed., Financial Markets and Financial Crises. Hume, David. OfMoney in Political Discourses. 1752. University of Chicago Press, 1991, pp. 33-68. Lucas, Robert E., Jr. “Expectations and the Neutrality of Money,” Blinder, Alan S. “Why ore Prices Sticky? Preliminary Results from Journal of Economic Theory (April1972), pp. 1 O3~24. on Interview Study,” The American Economic Review (May 1991). McCallum, Bennett. T. Monetary Economics: Theory and Policy ta. 89-96. Macmillan, 1989. Copie, F. H., and Alan Webber. A Monetary History ofthe UK: l81O~ Ohanion, lee F., Alan C. Stockman and Lutz Kilion. “The Effects of 7982, VoL / Dater: Surveys and Methods. Allen gad Unwin, 1985. Monetary and Real Shocks in a Business Cycle Model with Some Crafts, N. E. R., S. J. leybourne ond Terence C. Mills. ‘The Climacteric Sticky Prices,” Joumol ofMoney Credit, and Banking (forthcoming). in late Victorian Britain and France: A Reappraisal of the Evidence,’ Tayloi John B. ‘Aggregate Dynamics and Staggered Contmcts,” Journal ofApplied Econometrics (April/June 1989), w~103-17. Journal of PoliticalEconomy (February 1980), pp. 1-23. Craig, Ben. ‘Are Wages Inflexible?,” Federal Reserve Bank of Cleveland Wynne, Mark A. ‘Sticky Prices: What is the Evidence?” Federal Reserve Economic Commentary (April 1, 1995). Bank of Dallas Economic Review (first quarter 1995), pp. 1-12. Feinstein, Chafes H. “Wages and the Pomdon of the 1 880s: Comment,” Explorations in Economic History (April 1989), pp. 23T42, ______National Income, Expenditure and Output over the United Kingdom: 1855~l965, Cambridge University Press, 1972. FEDERAL RESERVE BANK OF ST. LOUIS 37