The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
____________ Review____________ Vol. 68, No. 6 June/July 1986 5 Forward Exchange Hates in Efficient Markets: The Effects o f N ew s and Changes in M onetary Policy Regimes 16 H ow Federal Farm Spending Distorts M easures o f Econom ic Activity Federal Reserve Bank of St. Louis Review June/July 1986 In This Issue . . . Recent theoretical explanations and em pirical analyses o f exch an ge rates em p h a size the role o f asset m arkets rather than trade flow s. M a n y argue that forw a rd exch an ge rates and the future spot exch an ge rates th e y m a y p red ict are prim arily d eterm in ed b y interest and inflation rate differentials b etw een co u n tries. In the first article in this Review , “ F orw ard Exchange Rates in Efficient M a rk ets:T h e Effects o f N ew s an d Changes in M o n eta ry P o licy Regim es,” M ack Ott and Paul T.W .M. V eugelers investigate the extent to w h ic h errors in fo rw ard exch an ge rate p red ictio n s o f future spot exch an ge rates have been in flu enced, on the on e hand, b y changes in interest and in flation rates and, on the other, by changes in the p o lic y stance o f the U.S. m o n etary authority. T h e authors fin d that changes in interest differentials explain a p o rtio n o f fo rw a rd rate forecast errors, esp ecia lly d u rin g the p e rio d o f U.S. m on etary aggre gate targeting, O ctober 1979 to S eptem ber 1982, and that changes in the U.S. m on etary p o lic y regim e alter the risk p rem iu m in fo rw a rd exch an ge rates. T h e significant divergen cies betw een the fo rw ard and spot exch an ge rate relations u n d er differen t U.S. m on etary p o lic y regim es suggest that cred ib le goals fo r m on etary p o lic y m a y be as im portan t as the m ech an ical details o f that p o lic y ’s execution. * * * In recen t years, fed eral paym en ts to farm ers fo r both loans and pu rchases o f farm produ cts have set n e w records. In the secon d article in this issue, “ H o w Federal Farm S pen ding Distorts M easures o f E con om ic A ctivity,” John A. T atom explains h o w transactions b y the C o m m o d ity Credit C o rp oration (CCC) are treated in the National In co m e and Produ ct A ccou n ts (NIPA). T a tom sh ow s that the volatile, qu arter-to-qu arter p attern o f CCC paym en ts to farm ers affects m easures o f farm, business, govern m en t and overall e co n o m ic activity. A cco rd in g to th e author, the recen t unusual d evelop m en ts have c o m p li cated the in terpretation o f som e key m easures o f e co n o m ic p erform an ce. He points out that a dju sting fo r these m ovem en ts can alter significan tly con clu sion s about the short-term perfo rm a n ce and eco n o m ic ou tlook fo r fed era l purchases, business in ven to iy investm ent and final sales in the econ om y. T a tom explains that analysts are likely to be m isled about the e c o n o m y ’s shortrun eco n o m ic p erform a n ce unless th ey p ro p e rly adjust the N IP A m easures w h e n large changes in CCC purchases occur. 3 JUNE/JULY 1986 FEDERAL RESERVE BANK OF ST. LOUIS Forward Exchange Rates in Efficient Markets: The Effects of News and Changes in Monetary Policy Regimes Mack Ott and Paul T. W. M. Veugelers s L - J lN C E the late 1970s, theoretical explanations o f exchange rate determ in ation have em p h a sized the asset approach rather than the exp en d itu re a p p ro a ch .1 M ost o f the em pirical research ap p lyin g the asset m odels o f exch an ge rate d eterm in ation also sub sum e the efficient market hypothesis. In this article, w e test three efficien t m arket h ypoth eses bearin g on forw ard exchange rates: First, are fo rw a rd rates u nbi ased forecasts o f future spot exch an ge rates? Second, does “ n e w s ” — in particu lar u nan ticipated changes in n om inal or real interest differentials — explain for- Mack Ott is a senior economist at the Federal Reserve Bank of St. Louis and Paul T. W. M. Veugelers, formerly a professor in the Department of Monetary Economics, Erasmus University, is a private consultant in the Netherlands. This article is the result of research undertaken in 1985 during an exchange o f visits — Mr. Veugelers to this Bank and Mr. Ott to Erasmus University. The authors acknowledge the research assis tance of James C. Poletti and the helpful comments of Clemens Kool. 'One rationale for this shift is the observation that the interest rate parity (IRP) postulate of the asset view has held up substantially better than the purchasing power parity (PPP) postulate of the expenditure view; see Mussa (1979) and Frenkel (1981b). The former refers to the equality of asset yields across currencies, while the latter refers to the equality of purchasing power across curren cies. PPP frequently, and for protracted periods, has been violated by exchange rates; see Frenkel (1981b). Thus, analysts have been faced with either modifying the PPP assumption and diluting its relevance, or accepting the evidence and developing theories to explain it. Indeed, some authors, Bomhoff and Korteweg (1983) and Darby (1981), argue that changing real exchange rates vitiate the relevance of PPP. w a rd rate forecast errors? Th ird, are fo rw a rd rate fo re cast errors affected b y change in the U.S. m on etary p o lic y regim e? Th ese h ypotheses are tested b y exam in ing the forecast errors (the differen ce b etw een the fo rw a rd rate and the su bsequ ently ob served spot rate) fo r the U.S. dolla r fo rw a rd rate against the cu rren cies o f eight in du strialized cou ntries o ver the latest floating-rate era (1973-85). EFFICIENT MARKETS AND FORWARD EXCHANGE RATES T h e fo rw ard exchange rate in an efficien t market reflects all the in form ation possessed by individuals active in that market. Thus, in an o p en market, the fo rw ard rate sh ou ld be an u nbiased p re d ic to r o f the future spot rate.- Hence, a regression o f the observed spot rate at tim e t on the fo rw a rd rate at tim e t — 1 (w h ere exch an ge rates are m easured by natural loga rithm s o f the dolla r prices o f foreign exchange), (11 s, = a + b f,_, + e,, sh ou ld result in an estim ated constant not signifi cantly d ifferen t from zero, an estim ated coefficien t on 2See Dornbusch (1976), Mussa (1979), Frenkel (1981a), Bomhoff and Korteweg (1983) and Edwards (1983b). 5 FEDERAL RESERVE BANK OF ST. LOUIS the fo rw a rd rate; not significantly differen t from 1 .0, and serially u n correlated errors (e,).:l Risk Premium T h e em pirical fin din g o f a significant in tercept has been su fficiently frequent in recent research that it is no lo n g er in terp reted as a departure from market efficiency. T h e question, then, is, w hat d oes the signifi cant in tercep t represent? T h e current v ie w is that the in tercep t represents a return to specu lation .4 For exam ple, if real interest rates on U.S. securities are h igh er than those on fo r eign securities, investors w ill shift th eir p o rtfolios to w a rd the h igh er-yield in g securities d en o m in a ted in U.S. currency; if these investors are risk-averse to un foreseen changes in cu rren cy values, th ey can h edge bv sellin g the h igh er-yield in g U.S. cu rren cy forw ard and bu yin g th eir o w n cu rren cy forw ard. By IRP, the resulting u p w a rd pressu re on the fo rw a rd rate must just offset the h igh er y ie ld o b tain ed o n the U.S. secu ri ties.3 Thus, the fo rw a rd rate in equ ation 1 , in such cases, w o u ld overestim ate the future sp ot rate so that the estim ated in tercep t w o u ld be negative. C on versely, a h igh er rate on non-U.S. securities, by the sam e logic, w o u ld im p ly a positive intercept. 3These propositions about the forward exchange rate have not been supported by recent empirical work. For example, Hansen and Hodrick (1980) find significant evidence of risk premia and explana tory power in lagged errors in both the 1920s and 1970s in one- and three-month forward markets. Baillie, Lippens and McMahon (1983), using a time series model on weekly data reject the hypothe sis that the forward rate is an unbiased predictor of the future spot rate in weekly data. Fama (1984) argues that the risk premium explains much of the error in the forward rate’s forecasts and finds that the risk premium and expected future spot rate are negatively correlated. Jacobs (1982) argued that the forward rate is an imper fect proxy for the expected rate and constructs a time series proxy for the expected rate. Unlike Fama, however, Jacobs found informa tion in the past variables, that is, information not included in the efficiently constructed forward rate at time t - 1 . Jacobs’ emphasis on omitted information is analogous to the decomposition sug gested by Frenkel (1981a) and elaborated in Isard (1983) and Edwards (1983a, 1983b). Edwards (1983b) finds that market effi ciency is not rejected in three out of four currencies in his study once news is included. 4Fama (1984) and Hodrick and Srivastava (1985). Hodrick and Hansen (1983) find that significant premia are both common and time varying. Frenkel (1981a) finds that news explains some of the risk premium while Edwards (1983b) finds that the combination of news and a system estimation technique eliminates the significant intercept. investors are concerned about after-tax real rates of return; through out this article we ignore the possibility that long-run real interest differentials may persist due to different tax rates on interest and investment income. Since our tests are on the effects of unantici pated changes in interest differentials, this possibility does not affect our results. JUNE/JULY 1986 News o f Interest Rate Changes Frenkel (1981a) argues that ch anges in exp ecta tion s b etw een the tim e that the fo rw a rd rate p red ictio n is m ade and the spot rate is o b s e iv e d explain the for w a rd errors. T h ese changes in expectation s, w h ic h he calls n ew s, are based o n in form ation revealed after the fo rw a rd contracts are m ade but before the spot rates are realized . Thus, u n an ticipated ch anges in interest rate differentials b etw een tim e t — 1 and t, — on e exam ple o f n ew s — explain part o f the residu al b e tw een the fo rw a rd rate forecast fj_, an d the re a lized spot rate s,. In co rp o ra tin g this m o d ifica tio n in to eq u a tion 1 yie ld s (21 s, = a + b f, _, + c|(i — i*l, - K ,_ ,li - i*),l + ej, w h e re i is an interest rate o f the sam e term as the fo rw a rd rate w ith asterisks in d ica tin g non-U.S. vari ables (interest rates are not in logs). O n ce again, riskneutrality a n d efficien t m arkets w o u ld im p ly an in sig nificant in tercep t and a slop e coefficien t o f unity; the sign o f the co efficien t on the n ew s variable, how ever, w o u ld d e p e n d u p o n w h e th e r the rise in the interest differential w e re du e to a relative rise in U.S. inflation — in w h ic h case it w o u ld be positive — o r a relative rise in U.S. real interest rates — in w h ic h case it w o u ld be n egative.'1 Fren kel’s p rox y fo r the e x p ected interest rate d iffer ential w as obtain ed from a regression o f the interest differential on its o w n lagged values a n d the lagged fo rw a rd exch an ge rate. Estim ating this m o d e l o ver 1973-79 fo r the p o u n d sterling, deu tsch em ark and franc, he fo u n d the in tercep t to be insignificant and the coefficien t o f the la gged fo rw a rd rate not signifi can tly differen t from one; these findings are consistent w ith the efficient m arket hypothesis. M oreover, the coefficien ts on the n ew s variable — the u nan ticipated interest rate ch ange — w e re positive, w h ic h he in ter p reted as prim arily reflectin g the relatively h igh and rising U.S. ex p ected inflation rate d u rin g this period . THE ROLE OF NEWS IN THE FORWARD EXCHANGE MARKET An im portan t insight o f the asset-m arket approach to exchange rate determ in ation is the em phasis on expectations. A sset prices are m u ch m o re d ep en d en t 6An increase in the expected inflation rate differential implies that, in the future, the dollar price of foreign currency will rise faster, and fewer dollars will be demanded because of their higher holding cost; hence, s, would rise. An increase in the U.S. real interest rate relative to foreign rates would increase the value of the dollar; hence, s, would fall. FEDERAL RESERVE BANK OF ST. LOUIS than current go o d s prices on the an ticip ated cou rse o f future events. Consequently, the role o f n ew s is m ost aptly captu red in the change o f expectations, not the error b etw een the ex p ected and realized y ie ld d iffer entials. By an application o f IRP and the efficien t fo rw ard market h ypothesis for foreign exchange, w e can obtain an alternative form o f the n ew s equ ation 2 estim ated by Frenkel. T h e alternative m o d el takes the form (see sh aded insert on the next p a g e ): (3) s, - f,_, = a + p A fp , + a>,. This m o d el has the advantage o f u sing a marketim p lied interest differential as w e ll as d irectly em b o d yin g the change in expectation s rather than the em p irica lly derived, exp ecta tion erro r p ro x y u sed b y Frenkel. The Distinction Between Real and Nominal News Frenkel claim ed that the positive coefficien t on the interest rate n ew s he fou nd du rin g 1973-79 reflected the relatively high and rising U.S. inflation rate du rin g this p eriod . Since the U.S. inflation rate has fallen both absolutely and relative to o th er nations in the years since 1979, the estim ated coefficien t on the ex p ected n om inal interest differential sh ou ld be unstable o ver the full p erio d 1973-85. O ne w a y to deal w ith this problem is to break the p erio d into sm aller units, each o f w h ic h have u niform relative U.S. inflation rates. W e, instead, separate the real and inflation com p o n en ts o f the n om inal n ew s variable. Th at is, w e w ill v ie w the change in the nom inal interest differential as the sum o f a change in the ex p ected real y ie ld differential and the change in the ex p ected inflation differential. Th ese co m p on en ts o f the n ew s sh ou ld have differen t effects on the fo rw ard rate errors. A rise in the real y ie ld on investm ents in on e countiy relative to those elsew h ere, in the absence o f ca p i tal restrictions, w ill cause an im m ediate appreciation in its exch an ge rate and result in a negative error in equation 3. Such appreciation s are transitory because capital in flow s w ill bring d o w n the in itially h igh er yields, w h ile the con com ita n t ou tflow s raise the yield s elsew here, until equ ality o f yie ld s is re sto red .7 C on se q u e n tly the very rise in the relative y ie ld that causes a 7See Dornbusch (1976), Isard (1983), and Edwards (1983a). None theless, the existence of risk premia implies that interest differences have persisted for some time in open capital markets; see Fama (1984). Hodrick and Hansen (1983) find these risk premia to be nonconstant and that their time variation is not summarized by nominal interest rate movements. JUNE/JULY 1986 cu rren cy to a ppreciate also creates the an ticip ation o f its subsequent d ep recia tion as y ie ld d ifferen ces go to zero. In contrast, an in crease in the ex p ected inflation differential prim arily alters the rate o f d ep recia tio n o f the exchange rate b y ch anging its PPP level; a rise in the inflation differential causes the exch an ge rate to rise faster o ver tim e bv the am ount o f the inflation increase. T h e d ep recia tion o f the spot rate also w ill reflect the p erceived in crease in the h o ld in g costs o f the cou ntry's cu rren cy w h ich redu ces the quantity d em anded. Thus, express the n om in al n ew s as the sum o f its real and inflation com pon ents, (4 1 A lp , = A ir, - r,*l + A l 77 — i t *), w h e re r, = ex p ected real interest rate, and it, = ex p ected in flation rate. Then, substitute the right-hand-side o f equ ation 4 into equ ation 3, to obtain (51 s, — l',_, = a + (3, All', — r*l + [3, A l77, — 77*1 + f,,. In equ ation 5, a is n on -zero in the p resen ce o f a risk prem ium , (3, is negative (since an u n an ticipated rela tive rise in U.S. real rates lo w ers s„ im p lyin g s, — f,_, < 01, [3, is positive but sm aller than (3, (since a rise in the relative U.S. inflation rate w ill cause a ch ange in the rate o f dep recia tio n o f the dollar, and, throu gh d e creased dem an ds fo r transaction balances, som e d e clin e in its level), and e, is a serially u n correla ted disturbance term . Another Kind o f News: Changes in Monetary Policy Regimes T h e estim ated param eters o f an e co n o m ic relation reflect the p erceived p o licy stance o f the govern m en t and m on etary authorities. Thus, as Lucas (19741 ar gued, changes in policy, e ith er b roa d goals such as the d esired inflation rate o r n a rro w er ones such as the m eth o d in w h ic h the p o licy is im p lem en ted , m a y alter the p u b lic’s respon se to prices and o th er in fo rm a tio n .'1 8We abstract from changes in the long-run real exchange rate in this analysis. That is, different rates of capital or human capital invest ment will cause different rates of productivity growth, or resource price changes that can alter the real exchange rate; see Darby (1980), Bomhoff and Korteweg (1983). Also, a reduction in the security of property rights can make investment in one currency less attractive than investments in other currencies, depreciating the currency and raising its real yields; see Dooley and Isard (1980). An apt application of the Dooley-lsard hypothesis may be the change in the French government in 1981, which was followed by significant nationalizations — especially in the banking sector. In our analysis, the only structural change considered is the U.S. monetary policy regime. 7 FEDERAL RESERVE BANK OF ST. LOUIS JUNE/JULY 1986 Forward Exchange Rate Errors, Efficient Markets and the News: the Role of the Forward Premium In its strong form , the efficien t market h ypothesis im p lies that the in tercep t in equ ation 1 w ill b e z ero and the coefficien t o f the lagged forw a rd rate w ill be unity. Consequently, the e rro r term, e„ is sim p ly the error o f the fo rw ard rate’s forecast o f the spot rate, (2.1) s, — f,_, = e,. Frenkel's insight co n cern in g the role o f n ew s is to argue that this erro r is du e to in form ation revealed after t — 1 (but before t) w h ic h alters expectation s and, hence, s,: ... current exchange rates already reflect current expectations about the future, while changes in the current exchange rate reflect primarily, changes in these expectations which, by definition, arise from new information.1 Frenkel’s specification, equ ation 2, em p loys the d if feren ce b etw een the re a liz e d interest differential and the exp ected differential; h ow ever, his argu m ent im p lies that the n ew s variable sh ou ld be the ch ange in the ex p ected differential b etw een t — 1 and t. That is, 'Frenkel (1981b), pp. 700-701, emphasis added. Frenkel notes (see footnote 31, p. 701) that Gustav Cassel, “the most recog nized proponent of the purchasing power parity doctrine” also recognized this forward-looking aspect: The international valuation of the currency will, then gener ally show a tendency to anticipate events, so to speak, and become more an expression of the internal value that the currency is expected to possess in a few months, or per haps in a year’s time (Cassel 1930, pp. 149-50). (2.2) e, = <t>(E, (i - i*)M+1 - E,_, (i - i*), IRP im p lies that the an n u alized o n e-m on th fo rw ard prem ium , (2.3) fp, = 12(f, - S,), is equal to the interest differen tial ex p ected to p re vail du rin g t through t + 1 , (2.4) fp, = E, (i - i*),.1+„ w h ere the term to m aturity o f the interest rates is equal to the h old in g p eriod in fp. If this equ ality d id not hold, riskless opportu n ities fo r profitable arbi trage w o u ld exist.2 Thus, substituting the relevant fo rw ard prem ia from equ ation 2.4 fo r the exp ected interest differentials in equ ation 2.2 and then sub stituting this expression fo r the erro r term in eq u a tion 2 .1 , w e obtain (2.5) s, — f,_, = 4> (Afp,), w h ich can be w ritten in an estim able form as (3) s, — f,_, = a + |3 Afp, + co,. 2This is known as the covered arbitrage condition. For example, if the fp, < (ius - iJK)„ an investor could sell pounds and buy dollars at time t, use the proceeds to buy a U.S. security; by buying forward pounds at t, the investor removes any exchange rate risk and obtains a higher yield than he would have in U.K. securities. Since this yield differential is riskless, arbitrage should drive it to zero and, in the process, ensure the equality shown in equation 2.4. For a fuller discussion and many instructive examples, see Wood and Wood (1985), pp. 378ff. Therefore, regression estim ates o f equ ations 2, 3 o r 5 such a m o n e ta iy p o licy regim e, the Fisher h ypothesis m ay be sensitive to changes in p o lic y goals and re holds, so that real interest rates are sim ply the d iffer gim es. en ce b etw een n om inal interest rates and an ticipated In particular, the h ypotheses fo r real and inflation n ew s su m m arized above are d ep en d en t on the m o n e tary p o lic y regim e. F or exam ple, w h e n the m on etary authority targets m o n eta iy grow th , interest rates w ill be d eterm in ed b y the private and public d em a n d fo r loanable funds; u n foreseen changes in that d em an d w ill cause changes in interest rates. Interest rates also w ill reflect private expectation s about inflation. In inflation; consequently, equ ation 4 holds, w h ile equ a tion 5 fo llo w s as an im p lica tion o f equ ations 3 and 4.'J In contrast, co n s id er a m o n e ta iy p o licy regim e o f “However, a critical caveat in evaluating equation 5 (or 5', see below) is Fama’s assertion that, when complete PPP does not hold, uncer tainty and differential tastes combine “to strip the Fisher equation of its meaning” (1984, p. 323). FEDERAL RESERVE BANK OF ST. LOUIS targeting interest rates."’ U n d er such a p o lic y stance, m ovem en ts in interest rates are, to som e extent, p o licy determ in ed in the short run since changes in the n om inal interest rate in d u ce offsettin g changes in the m o n ey su pply throu gh a p o licy-rea ction feedback. Consequently, changes in interest rates u n d er a regim e o f targetin g interest rates co n vey differen t in form ation than d o interest rate changes u n d er a re gim e o f targeting m on etary aggregates. A real interest differential u n d er interest-rate targetin g cannot be clo sed b y capital flow s alon e if the m o n e ta iy authority ch ooses to m aintain a particu lar n om inal target rate w h ic h m aintains the differential. O ver tim e, an interest rate target b e lo w the market rate w ill in crease the inflation differential. T h e adjustm ent process then dep en d s totally u p o n the relative inflation rates to restore PPP. And, again, the risk p rem iu m em b o d ie d in the in tercep t shou ld be sm aller du rin g an interest-rate regim e d u e to the red u ced short-run, interest-rate uncertainty. This p o licy regim e h ypothesis can be tested by an Ftest on the restriction im p licit in both equ ation 3 and 5 that the coefficien ts — a, (3, {$,, (3, — are stable o ver changes in m on etary p o licy regim es. T h e restriction is tested b y a d d in g in tercep t and slope d u m m y variables to get equations 3' and 5', com p u tin g the F-statistic on the change in the residuals b etw een the estim ates o f the restricted and u nrestricted equations: (3'I s, - f, , = a„ + a„,D + P„Afp, + p,„ DAlp, + co,' (5') s, = f,_, = a„ + a,,,]) + p, A(r, - if) + P,,DA(r, - r*) + p . A ( tt, — tt*) + P ,,D A ( tt, — it *) + e ', ( 1 if October 197!) =£ t « September 1982 where D = < 0 otherwise. Summary o f Testable Implications T h e im plication s o f the analysis in equ ations 3' and 5' are w o rth su m m arizing before rep ortin g the estim a tion results. First, n ew s about the real interest differ 10Only two U.S. monetary policy regimes are distinguished in this study— the October 1979-September 1982 period and the remain ing period before and after. Implicitly, this assumes that both the pre-October-1979 and the post-September-1982 periods are based on interest-rate targeting procedures; support for this characteriza tion of these two periods is offered in Gilbert (1985), Kaufman (1982) and Rasche (1985). The foreign monetary policy stance might also be argued to be relevant; while this is a possibility for a refinement on the estimates reported in this study, there do not appear to have been substantial changes during the period 1974-83 in six of the eight countries. The policy procedures of six of the eight non-U.S. countries (excluding Italy and Netherlands) are reviewed in Johnson (1983). JUNE/JULY 1986 ential causes negative forecast errors, s, — f,_„ w h ile changes in the inflation differential cause positive forecast errors. I f there are p eriod s d o m in a ted by relative volatility in inflation and o th er p eriod s d o m i nated b y real y ie ld volatility, then equ ation 3, w h ich restricts the coefficien ts to equality, sh ou ld be re jected b y an F-test in com p a rison w ith equ ation 5 w h ich d oes not restrict these coefficien ts to equality. Second, the th e o iy u n d erlyin g equ ation 5 im plies that n ew s about the e x p ected inflation differential w ill cause forecast errors, s, — f,_„ w h o se m agn itu de d e pen ds on the sensitivity o f m o n ey d em a n d to changes in the inflation rate. T h e co efficien t sh ou ld have the sam e sign as the ch ange in the inflation differential. Given the shortness o f the observation p e rio d — one m on th — the regression co efficien t fi, in equ ation 5 sh ou ld be positive but m a y n ot b e significant. Third, since the interest rates (hence, fo rw a rd p re mia) are assum ed to be d eterm in ed w ith o u t a m o n e tary p o licy reaction fu n ction in the analysis re p re sented in equ ation 5, m o n eta iy p o lic y based on interest-rate targets affects these h ypotheses. If the m o n e ta iy p o licy regim e affects the market valuations, i.e., spot and fo iw a rd exch an ge rates, h en ce forw ardrate forecast errors, then the restrictions in equ ation 5 w h ic h are rem o ved in equ ation 5' w ill be rejected by an F-test on the im p roved fit o f equ ation 5' relative to equ ation 5. Fourth, since it is w e ll k n ow n that the variances o f U.S. interest rates, both n om inal and real, have been h igh er du rin g m on etary target regim es than alterna tive regim es, there is a greater lik elih o od o f m isforecasting interest rates u n d er a m o n e ta iy target re gim e." T h e risk prem iu m m easured b y the intercept, w h ich prim arily is d eterm in ed b y this risk, sh ou ld be negative, larger and m o re significant du rin g p eriod s o f m o n eta iy targeting than du rin g period s o f interestrate targeting. This h ypothesis can be tested b y the significance o f the in te rcep t’s d u m m y variable in equations 3' or 5'. Finally, u n d er the efficien t market h ypoth esis e m b o d ied in equations 3, 5, 3' and 5', the erro r term s should be serially u ncorrelated. C orrelation in the disturbance term im p lies in co m p lete use o f past in for m ation and failure to exhaust p rofit opportu nities. Alternatively, if markets are efficient, serially co rre lated residuals im p ly a m isspecification o f the estim at in g equation. "S ee Roley (1983) and Rasche (1985). 9 FEDERAL RESERVE BANK OF ST. LOUIS JUNE/JULY 1986 EMPIRICAL TESTS Tests o f Forward Market Efficiency T h e m od els sp ecified in equations 1, 3, 5, 3' and 5' w ere estim ated using m o n th ly data from O cto b er 1973 Table 1 reports the results o f estim atin g equ ation 1 du rin g the fu ll sam ple p eriod , O cto b er 1973 throu gh June 1985. For six o f the eight cu rren cies con sidered, market efficien cy is not rejected; fo r Japan a n d S w itz erland, how ever, the m arket efficien cy h yp oth esis is rejected at the 5 p ercen t level. F o r all eight, th e Durbin-W atson statistic in dicates that h ypoth esis o f seri ally u n correla ted disturbances is n ot rejected . Thus, ex cep t fo r Japan and Switzerland, the results in table 1 in dicate that the n ew s m o d e l sp ecified in eq u a tion 3 is an a ppropriate em p irica l m odel. throu gh June 1985, using the U.S. d olla r spot and onem on th -forw ard p rices o f the cu rren cies o f Canada, France, Germ any, Italy, Japan, the Netherlands, S w itz erland and the U nited K in gdom . T h e tests are n ested in that equ ation 3 is obtain ed from equ ation 1 by im p osition o f the efficien t m arket h ypothesis. Equa tion 1 also contains both the restriction to suppress the real interest rate vs. inflation rate d e c o m p o s itio n and the restriction to suppress the effects o f ch anging m on etary p o lic y regim es on the regression c o e f ficien ts’ values. W e first test the sim ple efficien t market h ypothesis bv estim atin g equ ation 1. Next, w e esti m ate the sim ple n ew s m o d el w ith the ch ange in the n om inal fo rw a rd prem iu m , equ ation 3. Th is m o d el contains both the n om in al n ew s and the p o lic y regim e restrictions above. W e can then test these restrictions b y estim atin g 5', w h ic h is u n restricted and c o m p a rin g it through F-tests w ith equ ations 5 and 3'. F-tests on equ ation 5' vs. equ ation 5 and 5' vs. 3' determ ine, respectively, w h e th e r the p o licy regim e o r n om inal fo rw a rd prem iu m restrictions can be rejected. Data T h e spot and 30-day fo rw a rd exch an ge rates u sed in the estim ates are N e w York o p en in g market (10 a.m. m idpoin ts) fo r the last business day o f the m on th as co m p ile d b y the Bank o f Am erica. T h e change in the real interest d ifferential w as o btain ed from the change in the fo rw a rd p rem iu m : First, the fo rw a rd prem iu m w as con verted to an a n n u alized rate; the ch ange in this ann u alized fo rw ard prem iu m is the n ew s — that is, the change in the ex p ected n om in al interest differ ential. Second, an ex p ected an n u alized inflation rate fo r the one-m on th h orizo n w as co m p u ted fo r each cou n try from its m o n th ly CPI series.12 T h e change in the differential, U.S. m inus foreign inflation, is the change in the inflation differential u sed in estim ating equations 5 and 5'. T h e change in the real interest differential is then the ch ange in the ann u alized, n o m F o r Japan an d Sw itzerland, eq u a tion 1 w as reesti m ated b y su bperiods before, du rin g and sin ce the U.S. m on etary aggregate target regim e o f O cto b er 1979 throu gh S eptem b er 1982. For each country, the h y pothesis o f serially u n correla ted residuals w as not rejected in any su bperiod. F o r each o f the subperiods, the efficien t m arket h yp oth eses bearin g on th e c o e f ficients fo r S w itzerlan d w ere not rejected . F or Japan, the earlier tw o su b p eriod estim ates d o n ot reject m ar ket efficiency, but the recen t su b p erio d rejects m arket efficien cy both in term s o f a significant in tercep t and the deviation from u nity o f the la gged forw ard rate co effic ien t.13 Consequently, fo r n eith er S w itzerlan d n o r Japan is the estim ation o f equ ation 3 ju stified sin ce equ ation 3 is d erived from equ ation 1 assum ing a u nit co efficien t on f,_,. Yet, equ ation 3' o r equ ation 5' m ay be ju stified fo r Sw itzerlan d since the d u m m y variables can a c cou nt fo r the nonstable coefficient. F or Japan, the failure o f the efficien t m arket h ypoth esis in the last su bperiod is not offset by any o f o u r variables, and it is consistent w ith this failure that Japan rejects each o f the specification s equ ation s 3', 5 and 5' as re p orted in tables 2 and 3. Tests o f News Model with U.S. Monetary Regimes Not Distinguished Table 2 reports the results o f estim atin g equ ation 3, inal, one-m on th -forw ard p rem iu m m inu s the change the n ew s m o d el w ith the ch ange in the n om inal fo r w a rd prem ium , o ver the full period , O cto b er 1973- in the ex p ected inflation differential. June 1985. In sharp contrast to the results in table 1, w h ic h su pport this specification, the estim ates u n i form ly reject this m o d el: n o co efficien t is significant at l2Clemens Kool of Erasmus University computed this series using a multi-state Kalman filter. A simple Kalman filter is a forecasting method based on assumptions about the forecasted variable’s relation to current and lagged data on itself and or other series. A multi-state Kalman filter allows this relation to vary according to a feedback or adaptive error loop; the multi-state modifier refers to the alternative sets of assumed weights. A concise description and illustrative example are contained in the statistical appendix to Bomhoff and Korteweg (1983). Digitized for 10FRASER 13The October 1982-June 1985 estimates for Japan are very curious. The estimated intercept is huge in comparison with the earlierperiod Japanese estimates, the Swiss estimates or any of the estimates in table 1: a = - 1.192 (s.e. = 0.548), (i = 0.783 (s.e. = 0.100). FEDERAL RESERVE BANK OF ST. LOUIS JUNE/JULY 1986 Table 1 Tests of Forward Exchange Market Efficiency for U.S. Dollar, October 1973-June 1985 (U.S. Monetary Regimes Not Distinguished) Summary Statistics Coefficients1 Currency Intercept Canada -0 .0 0 2 (0.002) -0 .0 0 2 (0.018) -0 .0 2 0 (0.016) 0.012 (0.052) -0 .2 9 8 (0.112)* -0 .0 1 3 (0.017) -0 .0 3 4 (0.013)* 0.001 (0.009) France Germany Italy Japan Netherlands Switzerland United Kingdom Test f.-. R2 DW 0.998 (0.012) 1.001 (0.010) 0.981 (0.018) 1.002 (0.007) 0.946 (0.020)* 0.991 (0.018) 0.961 (0.016)* 0.994 (0.014) 0.981 2.16 0.730 0.985 2.07 0.573 0.954 2.03 1.800 0.992 1.87 0.500 0.940 1.80 3.734* 0.957 2.01 1.486 0.962 1.92 3.537* 0.974 1.82 0.483 F2 'Standard errors of estimated coefficients appear in parentheses; asterisks indicate rejection at 5 percent level of individual efficient market hypotheses — intercept is zero, slope coefficient = 1.0. 2F-test of joint efficient market hypothesis that intercept is zero and slope coefficient is unity; asterisk indicates rejection at 5 percent level. any reasonable co n fid en ce level and the adju sted R- is negative fo r six o f the eight cu rren cies tested. C onsis tent w ith the efficien t market h ypothesis, h ow ever, the h ypothesis o f serially u n correlated disturbances is not rejected. N onetheless, the results requ ire an investiga tion o f alternative explanations fo r this m o d e l’s u ni form failure. Decomposition o f Nominal Forward Premium Also rep orted in table 2 is the F-statistic fo r testing w h eth er d e c o m p o s in g the change in the n om inal for w a rd prem iu m in to innovations in its ex p ected real Tests o f News Model with U.S. Monetary Regimes Distinguished As d iscussed above, the U.S. m on etary p o licy regim e can be ex p ected to affect the relationship b etw een the d olla r’s exchange rates and U.S.-foreign interest differ entials. Thus, the statistical results rep orted in table 2 m ay be invalid because th ey d o not distinguish changes in the U.S. m on etary p o licy stance. T o test fo r such p o licy regim e effects, equations 3' and 5', w ere estim ated to isolate the p erio d o f U.S. m on etary aggr e gate targeting, from O cto b er 1979 to S eptem b er 1982, w ith slope and in tercept dum m ies. and inflation co m p on en ts is statistically w arranted. Table 3 reports estim ates o f equ ation 5' and the F- T h e F-statistic is obtain ed from the d ifferen ce in the statistics to test the effect o f m on etary regim e changes explanatory p o w e r o f equ ation 5 w ith respect to equ a tion 3; the critical value fo r rejectin g the restriction in equation 3 (that fi,, (3, in equ ation 5 are equal) is 3.92. and the equ ality restriction im p licit in equ ation 3' and rem oved in equ ation 5'. T h e estim ates presen t a sub stantia] contrast to those in table 2. Canada and Italy O nly the N eth erlan ds result rejects the restriction. reject the nom inal forw a rd prem iu m restriction (last 11 FEDERAL RESERVE BANK OF ST. LOUIS JUNE/JULY 1986 Table 2 Tests of News Model Using Change in Nominal Annualized Forward Premium on U.S. Dollar, October 1973-June 1985 (U.S. Monetary Regimes Not Distinguished) Coefficients' Summary Statistics Test Currency Intercept Afp R2 DW F Canada -0.00 1 (0.001) -0 .0 0 3 (0.003) -0 .0 0 4 (0.003) -0 .0 0 2 (0.002) -0 .0 0 2 (0.003) -0 .0 0 5 (0.003) -0 .0 0 4 (0.003) -0 .0 0 2 (0.003) -0 .0 6 9 (0.102) -0 .0 7 6 (0.066) -0 .1 4 6 (0.188) -0 .0 1 6 (0.027) -0 .0 1 7 (0.035) 0.031 (0.031) -0.011 (0.179) 0.006 (0.123) -0 .0 0 4 2.16 0.455 2.059 0.002 2.04 1.347 2.937 -0 .0 0 3 2.06 0.600 1.898 -0 .0 0 5 1.86 0.332 0.053 -0 .0 0 6 1.80 0.236 1.254 0.000 2.03 1.004 4.164* -0 .0 0 7 1.92 0.004 0.326 -0 .0 0 7 1.83 0.002 1.888 France Germany Italy Japan Netherlands Switzerland United Kingdom F2 'Standard errors of estimated coefficients appear in parentheses. ?F-statistic for testing the equality restriction on the coefficients of the change in the real and the inflation differentials (components of the change in the nominal forward premium); asterisk indicates rejection at 5 percent level. colum n, F-test) but, in contrast to table 2, the N eth er lands does not w h e n the U.S. m on etary regim e shift is accou n ted for. C on sid erin g the a ppropriate sp ecifica tion, equations 3' o r 5', six o f the eight equ ations are significant in term s o f their overall fit (F-statistics) at the 5 percen t level, France is significant at the 6 per cent level, and seven o f eight cou ntries reject the restriction o f stable coefficients across m on etary re gim e changes at the 10 p ercen t level o r better. O nly Japan fails the F-test fo r the significance o f the m odel. U nited Kingdom , this entails a sw itch from a positive and significant co efficien t du rin g the U.S. n on -m on etary targeting regim e. Thus, fo r each o f the seven cu rren cies fo r w h ic h the market efficien cy criteria are met, the U.S. m on etary p o licy regim e has a significant effect on the errors in the fo rw a rd rate forecasts. M o re specifically, tw o g e n eralizations can be advan ced based on the results in table 3. First, the greater interest rate volatility du rin g U.S. m on etary aggregate targetin g sh ow s up in a sig In term s o f the in dividu al coefficients, six o f the eight countries evid en ce a significant negative risk nificant risk p rem iu m ten d in g to strength en the d olla r against six o f the eight currencies. Second, given the p rem iu m (10 p ercen t o r better) du rin g the U.S. m o n e failure to reject the n om in al fo rw a rd p rem iu m restric tary aggregate targetin g period, w h ile the in tercep t is tion o f equ ation 3', the negative significan ce o f the slope d u m m y im p lies that the interest differential u n iform ly non sign ifican t du rin g the o th er U.S. m o n e tary p o licy regim e, O cto b er 1973—S eptem b er 1979 and O ctober 1982-June 1985. T h e im pact o f the different regim es is also notable in the slop e in teraction dum m y. Th e coefficien t on the change in the real fo rw a rd p rem iu m is negative and significant fo r Can ada, Germany, the Netherlands, S w itzerlan d and the U nited K in gdom . F or Germany, S w itzerlan d an d the Digitized for12 FRASER n ew s w as prim arily in terp reted as an increase in the inflation differential du rin g U.S. n on -m on etary aggre gate targeting p eriod s and as an in crease in real in ter est differentials d u rin g U.S. m on etary aggregate target ing. In o th er w ords, the d olla r a p p recia ted a lon g w ith unan ticipated increases in the fo rw ard prem iu m d u r in g O ctober 1979 to S eptem b er 1982, but d ep recia ted JUNE/JULY 1986 FEDERAL RESERVE BANK OF ST. LOUIS Table 3 Tests of News Model Using Unrestricted Specification, October 1973-June 1985 (U.S. Monetary Regimes Distinguished)______________ Coefficients' Canada -0 .0 0 2 (0.001) 0.001 (0.003) -0.001 (0.003) 0.000 (0.003) 0.001 (0.003) -0.001 (0.003) 0.001 (0.004) -0.001 (0.003) France Germany Italy Japan Netherlands Switzerland United Kingdom Dl2 A (r-r*) 0.000 0.299 (0.003) (0.188) -0 .0 1 5 -0 .0 2 3 (0.105) (0.006)* -0 .0 1 4 0.540 (0.299)* (0.006)* 0.012 -0.011 (0.006) + (0.033) -0 .0 1 2 0.029 (0.006) + (0.044) 0.047 -0 .0 1 3 (0.006)* (0.029) -0 .0 1 8 0.433 (0.007)* (0.218)* 0.382 -0 .0 0 6 (0.149)* (0.006) Dr2 -0 .4 3 6 (0.223) + -0 .0 7 7 (0.135) -1 .1 3 7 (0.382)* -0 .0 6 7 (0.056) -0 .2 5 5 (0.200) -0 .8 1 6 (0.243)* -1 .2 1 7 (0.356)* -0 .9 5 0 (0.238)* * t= I Intercept < Currency Summary Statistics -0 .3 4 3 (0.209) + -0 .3 6 8 (0.193) + -0.31 1 (0.395) 0.139 (0.100) 0.163 (0.124) -0 .2 8 0 (0.161) + 0.506 (0.263) + 0.319 (0.180) + Tests Dtt2 R2 DW F F3 F4 - 0.776 (0.253)* 0.211 (0.282) - 0.980 (0.511) + -0 .5 6 8 (0.181)* -0 .4 3 4 (0.292) -0 .5 3 6 (0.343) -1.19 1 (0.437)* -1 .0 8 7 (0.296)* 0.056 2.19 2.728* 3.658* 0.045 2.12 2.297* 2.352 + 2.133 0.081 2.10 3.445** 4.837** 0.605 0.059 1.92 2.731* 4.415** 3.990* 0.016 1.89 1.443 1.899 1.093 0.107 2.06 4.324** 5.313** 2.098 0.088 2.03 3.688** 6.025** 0.162 0.097 1.79 3.979** 5.933** 1.052 4.027* 'Standard errors of estimated coefficients appear in parentheses; asterisk indicates significance at 5 percent level and plus sign indicates significance at 10 percent level. 2DI, Dr and Dir equal 1.0 during period of U.S. monetary-target policy regime, October 1979-September 1982 and zero otherwise. 3F-statistic for testing restriction that coefficients are stable across different monetary regimes; double asterisk indicates rejection at 1 percent level, asterisk indicates rejection at 5 percent level, and plus indicates rejection at 10 percent level. 4F-statistic for testing the equality restriction on the coefficients of the change in the real and the inflation differentials (components of the change in the nominal forward premium); asterisk indicates rejection at 5 percent level, plus indicates rejection at 10 percent level. w ith such n ew s du rin g the rest o f the floatin g rate p eriod . This is consistent w ith F ren kel’s (1981a) results fo r 1973-79. Finally, tin: D u rbin-W atson statistics in table 3 d o not in d ica te serial correlation in the re sid uals, consistent w ith the m ain tain ed h ypoth esis o f market efficiency. T h ere rem ain tw o p u zzlin g results: (1) T h e esti m ated coefficien ts o f the change in the inflation d iffer ential du rin g the m on etary regim e are gen era lly n ega tive, refuting the h ypoth esis em b o d ie d in equ ation 5; this negative coefficien t is significant at the 10 percen t level or better in five countries. (2) M oreover, the d e co m p o sitio n o f the n om inal interest differential is sig nificant o n ly fo r Canada and Italy. Th is irrelevan ce o f the distin ction b etw een real and n om inal interest differentials m ay sim p ly be a con firm ation o f Fam a’s (1984) assertion that, w ith risk aversion o r w ith o u t PPP, the Fisher equ ation d oes not h old (see fo o tn o te 9). In deed, fo r six o f the eight currencies, the F-test does not reject the im p licit restriction o f equ ality o f changes in the n om inal interest differential s tw o com p o n en ts d isp layed in table 3. The Implications o f Monetary Regimes: A Closer Look T h e negative coefficien t on the inflation differential du rin g the 1979-82 m o n e ta iv regim e is b oth pervasive and p u zzlin g. T w o possible explanations are w orth considerin g. First, the o n e-m on th h o rizo n o f the esti mated, an ticip ated CPI inflation rates u sed in estim at ing equ ation 5' m ay be too short, o r the estim ated ex p ected inflation series sim p ly m ay be bad proxies. Second, the market 11133' have d eterm in ed that the U.S. m o n eta iy au th ority and the adm inistration w ere co m m itted to lo w e rin g the U.S. inflation rate. C on se 13 FEDERAL RESERVE BANK OF ST. LOUIS quently, a short-term increase in the U.S. ex p ected inflation rate w o u ld lead market participants to expect a tigh ten in g o f m o n e ta iy g r o w th .14 I f so, a short-term increase in U.S. inflation w ou ld lead to increases in the U.S. real interest rate as the market an ticip ated the m onetary au th ority’s reaction. This explanation, co n sistent w ith research by C orn ell (1982), has not been tested here, but it is consistent w ith the d e c o m p o s i tion o f changes in the n om in al interest differential generally not in creasin g the ex p la n a to iy p o w e r o f the equation for six o f the eight cu rren cies.1'' CONCLUSION W e have tested the efficien cy o f fo rw a rd exchange markets fo r the dolla r against eight m a jo r cu rren cies during the floating period. T h e regression estim ates clearly dem on strate that failing to accou nt fo r changes in the p o licy procedu res o f the U.S. m o n e ta iy au thor ity entails m isspecification. M o n e ta iy regim e changes alter the risk prem ia that market participants require on fo iw a rd contracts and affect the d irection o f errors im p lied b y n om inal and real news, that is, u n foreseen events occu rrin g b etw een the tim e o f contract and its maturity. T h e im p lication s o f the standard m o d el o f exchange rate beh avior w'ere substantiated fo r n om i nal n ew s u n d er a m on etary target regim e, but its im p lication for inflation differentials w as refuted. W h ile a closer m o d elin g o f the p o lic y p roced u re m ay explain this rejection, it rem ains a p rom in en t p u zzle in this study. N onetheless, on e in terpretation o f these results is that market participants regarded the U.S. m o n e t a r y p o lic y r e g im e o f 197 9-82 as ant i inflationaiy. I f this is correct, it fo llow s that credible goals o f m on etary p o lic y m a y be as significant fo r market participants as the m echan ical details o f that p o lic y ’s execution. JUNE/JULY 1986 REFERENCES Baillie, Richard T., Robert E. Lippens and Patrick C. McMahon. “Testing Rational Expectations and Efficiency in the Foreign Ex change Market,” Econometrica (May 1983), p. 553-63. Bomhoff, Eduard, and Pieter Korteweg. “ Exchange Rate Variability and Monetary Policy Under Rational Expectations,” Journal of Monetary Economics, 1983. Cassel, Gustav. Money and Foreign Exchange After 1914 (New York: McMillan, 1922). Cornell, Bradford. “ Money Supply Announcements, Interest Rates and Foreign Exchange,” Journal of International Money and Fi nance (August 1982), pp. 201-08. Darby, Michael R. “ Does Purchasing Power Parity W ork?” Pro ceedings of Fifth West Coast Academic/Federal Reserve Eco nomic Research Conference, University of California-Los An geles, National Bureau of Economic Research (December 1981). Dooley, Michael P., and Peter Isard. "Capital Controls, Political Risk and Deviations from Interest Rate Parity,” Journal of Political Economy (April 1980), pp. 370-84. Dornbusch, Rudiger. "Expectations and Exchange Rate Dy namics,” Journal of Political Economy (December 1976), pp. 1161— 76. Edwards, Sebastian. “ Comment on Isard,” Jacob A. Frenkel, in Exchange Rates in International Macroeconomics (University of Chicago Press, 1983). ________ _ “ Floating Exchange Rates, Expectations and New In formation,” Journal of Monetary Economics (May 1983b), pp. 3 2 1 36. Fama, Eugene F. “ Forward and Spot Exchange Rates,” Journal of Monetary Economics (November 1984), pp. 320-38. Frenkel, Jacob A. “ Flexible Exchange Rates, Prices and The Role of “ News": Lessons From the 1970s,” Journal of Political Economy (August 1981a), pp. 665-705. ________ _ “ The Collapse of PPP in the 1970s,” European Eco nomic Review (1981b), pp. 145-65. Gilbert, Alton. “ Operating Procedures for Conducting M onetary Pol icy,” this Review (February 1985), pp. 13-21. Hansen, Lars Peter and Robert J. Hodrick. “ Forward Exchange Rates as Optimal Predictors of Future Spot Rates: An Econometric Analysis," Journal of Political Economy (October 1980), pp. 8 2 953. ________ _ “ Risk Averse Speculation in the Forward Foreign Ex change Market: An Econometric Analysis of Linear Models” in Frenkel, Ed., Exchange Rates in International Macroeconomics (University of Chicago Press, 1983). 14The U.S. CPI inflation rate was 13.3 percent in 1979,12.4 percent in 1980, 8.9 percent in 1981 and 3.9 percent in 1982. There is also some support for this view in the impact of lagged reserve account ing during the monetary targeting period. As Kaufman (1982) notes, this results in more volatility of both money and interest rates since a decision to maintain a target growth path when the money supply has exceeded the path requires a subsequent reduction of reserve growth. Since banks already will have increased their required reserves, real rates will vary with the money supply errors and, perhaps, short-run inflation expectations. 15Cornell (1982) finds that unexpected monetary supply increases are correlated with an appreciation in the dollar, not the depreciation that an anticipated simple link with increased inflation would imply. Cornell suggests that the explanation is an anticipated policy reac tion, a tightening of the money supply growth rate. Digitized for14 FRASER Hodrick, Robert J., and Sanjay Srivastava. “ The Covariation of Risk Premiums and Expected Future Spot Exchange Rates,” Journal of International Money and Finance (March 1986 Supplement), pp. 5 - 21. Isard, Peter. “ An Accounting Framework and Some Issues for Modelling How Exchange Rates Respond to News,” in Jacob A. Frenkel, Exchange Rates in International Macroeconomics (Univer sity of Chicago Press, 1983). Jacobs, Rodney L. “The Effect of Errors in Variables on Tests for a Risk Premium in Forward Exchange Rates,” Journal of Finance (June 1982), pp. 667-77. Johnson, Karen. “ Foreign Experience with Targets for Monetary Growth,” Federal Reserve Bulletin (October 1983), pp. 745-54. Kaufman, George. “The Fed’s Post-October 1979 Technical Oper ating Procedures Under Lagged Reserve Requirements: Reduced FEDERAL RESERVE BANK OF ST. LOUIS Ability to Control Money,” The Financial Review (November 1982), pp. 279-94. Lucas, Robert E., Jr. Eds. Karl Brunner and Alan Meltzer, 1974, “ Econometric Policy Evaluation: A Critique,” The Phillips Curve and Labor Markets, Carnegie-Rochester Conference on Public Policy, Vol. 1, pp. 19-46. Mussa, Michael L. “ Empirical Regularities in the Behavior of Ex change Rates and Theories of the Foreign Exchange Market,” Vol. II Carnegie-Rochester Conference Series on Public Policy, 1979. JUNE/JULY 1986 Rasche, Robert H. “ Interest Rate Volatility and Alternative Mone tary Control Procedures,” Economic Review, Federal Reserve Bank of San Francisco (Summer 1985), pp. 46-63. Roley, Vance. “ The Response of Short-Term Interest Rates to Weekly Money Announcements,” Journal of Money, Credit and Banking (August 1983), pp. 344-54. Wood, John, and Norma L. Wood. Financial Markets (Harcourt Brace Jovanovich, Inc., New York) 1985. 15 FEDERAL RESERVE BANK OF ST. LOUIS JUNE/JULY 1986 How Federal Farm Spending Distorts Measures of Economic Activity John A. Tatom D URING the 1980s, federal purchases o f farm produ cts b y the C o m m od ity Credit C orp oration (CCC) have exhibited relatively large qu arterly sw ings that have significantly affected h o w w e in terpret eco n om ic d evelop m en ts.1 A lthou gh these purchases increase the g o vern m en t’s in ven to iy o f farm products, they are treated as final sales to the governm en t, instead o f in ven to iy transactions, in the National In co m e and Product A ccou nts (NIPA). As a result, a CCC purchase increases federal purchases and final sales in the eco n o m y and redu ces m easured investm ent in farm inventory. Similar private sector transactions, w h ich redistribute farm produ cts from one o w n e r to an other, result in offsettin g changes in farm and busi ness in ven toiy; these transactions affect n eith er bu si ness in ven tory investm ent n or final sales. This article explains the im pact o f CCC] purchases and exam ines the d istortions that th ey can p rod u ce in quarter-to-quarter m ovem en ts o f som e im portant N IPA measures. It show s that a dju sting fo r the effect o f CCC purchases can alter conclu sion s about the sh ort term perform a n ce and ou tlook fo r federal purchases, the farm sector and aggregate p rod u ctio n and em ploym ent. T h e largest sw ings in CCC purchases on record w ere record ed at the en d o f 1985 and early this year; hence, these recent sw ings have h ad the greatest im pact on m easures o f in ven to iy investm ent, federal purchases and overall final sales. A m o re useful p er spective on N IPA m easures can be o btain ed by adjust in g these m easures du rin g quarters w h e n large changes in CCC purchases occur. John A. Tatom is an assistant vice president at the Federal Reserve Bank of St. Louis. Michael L. Durbin provided research assistance. ’The significance of such swings, especially as a major source of changes in federal purchases, was first noted by the Bureau of Economic Analysis (1982). 16 CCC PURCHASES, SALES AND INVENTORY CHANGES T h e C o m m od ity Credit C orporation, establish ed in 1933 as part o f the D epartm ent o f Agriculture, carries out the federal go vern m en t’s p rice su pport p rogra m s.2 T h ese program s in clu d e both “ n on recou rse loan s” and direct purchases o f farm produ cts. T h e fo rm er are called n on recou rse loans becau se the fa rm er is free to deliver the p le d g e d crop, w h ic h serves as collateral, in o rd er to settle the loan :' T h e p rice o f the c o m m o d ity at w h ic h the loan is advanced is called the loan rate; it establishes a m in im u m p rice fo r the com m od ity. W h en the govern m en t m akes such a loan, the transac tion is treated in the NIPA as a pu rchase o f farm products. As a result, these loans in crease federal purchases and redu ce farm in v e n to iy h oldin gs. Re paym en t en tries.4 o f the loan reverses these a cco u n tin g Direct purchases o f farm produ cts are treated in the 2More extensive discussion of the CCC can be found in the Council of Economic Advisers (1986), Herman (1978), Bureau of Economic Analysis (1982) and Wakefield (1986). The former also details other features of U.S. agricultural policy. Nonrecourse loans to farmers are based on the government-set loan rate for each farm product and the amount of the current or past product pledged against the loan as collateral. If the producerborrower cannot sell his product for more than the loan rate plus the accumulated storage costs and interest on the loan, the farmer forfeits the pledged crop and the loan obligation is discharged. The farm products that are covered by the loan program include wheat, corn, barley, oats, rice, cotton, honey, peanuts, sorghum, soybeans, rye, tobacco and sugar. “Even when the farmer pays off the loan, he reaps a benefit in the form of a short-term credit subsidy, since the interest rate on such loans is less than market rates. The CCC also supports prices of farm products by directly purchasing certain products at official support prices when such prices exceed market levels. Chief among these are such dairy products as cheese, butter and dry milk. FEDERAL RESERVE BANK OF ST. LOUIS JUNE/JULY 1986 Chart 1 CCC Purchases 1973 74 75 76 77 78 79 exact same w a y in the NIPA. Thus, c o m m o d ity loans and direct co m m o d ity purchases hv the federal g o v ernm ent result in offsettin g changes in federal pu r chases o f g o o d s and seivices and business (farm) in v e n to r y in v e s tm e n t. G N P is u n a ffe c te d b v the transactions because they result in no change in prod u ction .3 Chart 1 show s both n om inal and real (19X2 prices) CCC in ven to iy purchases from 1973 to the secon d qu arter o f 1986. A lthou gh the n om inal purchases a p 80 81 82 83 84 85 1986 the quarter-to-quarter sw ings are som etim es quite large in com parison to G NP m ovem ents. For exam ple, in the fourth qu arter o f 1985, such purchases rose $20.8 billion, o r 36.5 p ercen t o f the total increase in GNP du rin g the sam e quarter. It is also evid en t from the chart that m ovem en ts in CCC purchases have b ecom e substantially larger in the 1980s, w ith the biggest sw ings occu rrin g at the en d o f 1985 and in early 1986. In part, these in creased fluctuations reflect the g ro w in g role o f federal farm program s. pea r sm all relative to current G N P o f o ver $4 trillion, 5The independence of GNP from CCC purchases is based on two assumptions: (1) that the coverage, timing and seasonal adjustment of changes in farm inventory and CCC purchases are consistent and (2) that farmers, in general, cannot or do not respond to CCC purchases within the quarter by altering production. The former point has been made by the Bureau of Economic Analysis (1982). These second-order considerations are ignored below in order to focus solely on the measurement principles involved. CCC AND FEDERAL PURCHASES OF GOODS AND SERVICES Q uarterly m ovem en ts in CCC purchases have had a sizable im pact on the pattern o f g ro w th o f federal purchases du rin g som e quarters in the 1980s. Chart 2 show s the gro w th rates o f real federal purchases and adju sted real federal purchases (w h ich exclu d e CCC 17 FEDERAL RESERVE BANK OF ST. LOUIS JUNE/JULY 1986 Chart 2 The G ro w th Rate of Real Federal Purchases w ith an d w ithout CCC Purchases Compounded annual rate Compounded annual rate 50 i II n ii 40 IIii iii ii iii ii ii ii ii 1 i A i 1 ii1 1 A1 1 i i fj] /i A\ \i f ! i1 \ i1f » I1i \iJ 19 f 30 i/ 20 Real fe d e ra l lu rc h a s 10 r\ 1 t j l A 1' 1 1 ii A [ \ A f 9U V A d ju s t ed rea A n Ii l\1vA\V** *1 lh \▼V N I V\l1y, » Ii fmJl I I V* \l/ -10 /V AA\ / \ * \ \ \\ federa / iI iii i i 5S i\• A ''' i»\\ r\ » \i l V U / ' '' #/iI V 1 1 40 30 lI iI I II IIII 11 II fMl1 J 1 if 111 1 11 11 111 1 | _ \ i \II 1 Vi J purch ases 20 I l1 Ijm -20 10 -10 ■20 1 -30 1973 74 75 76 77 78 79 80 81 82 83 84 85 1986 -30 purchases) since 1973." In the 1980s, the differen ce in the grow th rates often lu;s been qu ite large and m ore variable. Since 11),SO, the fed eral govern m en t gen erally has been accu m u lating in ven tory o f farm products, but in 1983 and early 1984, the Paym en t-In-K ind IPI K) program led to large sales fo r fou r quarters.7 T h ese sw ings in CCC purchases had a m a jor im pact on the grow th rate o f federal purchases, gen era lly d ep ressin g it in 1983 and early 1984 and su bsequ ently raising it. fed eral purchases d u rin g recession p eriods, w h ile d e pressing the grow th o f federal purchases d u rin g the initial stages o f expansions. T h is effect has resu lted in the appearan ce o f a negative relationship b etw een GNP and federal purchases, a relationship that disap pears w h e n federal purchases are adju sted fo r CCC purchases. For exam ple, from 1/1980 to 11/1986, the correlation b etw een the grow th rate o f real federal T h ese sw ings m ake it difficult fo r analysts to interpret chases and o f real GNP is negative ( — 0.15); w h e n real CCC purchases are o m itted from govern m en t p u r trends in federal spending. A n o th e r coin ciden tal effect o f CCC purchases in recent years has been to raise the grow th rate o f purchases o f go o d s and services in clu d in g CCC pu r chases, h ow ever, the correlation is positive (0.04). W h ile n eith er correlation is statistically significant, distortions caused by volatile CCC purchases can bias statistical tests o f fiscal p o licy's general effectiveness. 6Since nominal and real CCC inventory changes are not substantially different over the period since 1973, attention throughout this article is focused on real measures. Movements in the nominal counter parts of real measures provide no additional insight and so are ignored here. 7A description and analysis of the PIK program that was in effect in 1983 and early 1984 can be found in Belongia (1983) and Rosine (1984). Digitized for18 FRASER CCC PURCHASES AND CHANGES IN FARM INVENTORY Federal purchases o f farm produ cts are offset in the JUNE/JULY 1986 FEDERAL RESERVE BANK OF ST. LOUIS Table 1 The Change in Farm Inventory and CCC Purchases (billions of dollars, 1982 prices) Change in farm inventory 1/1980 II III IV $ - 0 .3 5.5 - 0 .2 - 2 .0 -7 .0 -1 0 .5 3.8 1/1981 II III IV 1.6 - 0 .8 5.5 9.1 1/1982 II III IV Annual mea standard deviation Annual mean/ standard deviation Change in farm inventory and CCC $ - 4 .7 6.09 $ - 5 .3 - 1 .5 -1 0 .7 1.8 $ - 3 .9 5.37 4.6 11.2 5.0 -1 .3 4.9 5.11 6.2 10.4 10.5 7.8 8.7 2.09 10.8 0.7 7.9 17.2 -4 .1 4.0 3.2 -8 .9 - 1 .5 6.16 6.7 4.7 11.1 8.3 7.7 2.71 1/1983 II III IV 3.8 -0 .1 -3 .1 -1 7 .2 -9 .1 -6 .9 -1 5 .7 6.5 - 6 .3 9.32 - 5 .3 -7 .0 -1 8 .8 -1 0 .7 -1 0 .5 6.01 1/1984 II III IV -1 5 .9 3.1 3.4 0.8 16.4 1.8 1.3 0.0 4.9 7.72 0.5 4.9 4.7 0.8 2.7 2.4 1/1985 II III IV 3.2 2.0 11.5 32.3 6.4 7.8 -0 .7 -2 1 .3 - 2 .0 13.43 9.6 9.8 10.8 11.0 10.3 0.7 1/1986 II 6.4 4.5 2.9 4.1 I Ul o CCC purchases — 9.3 8.6 — GNP accounts by redu ction s in farm inventory.8 Thus, CCC purchases can distort the short-run in terpreta tion o f changes in farm and business in ven toiy. W hen the CCC purchases (sells) farm goods, farm and busi ness in ven to iy investm ent falls (rises), giving the a p pearance o f an in ven to iy change. O f course, such an Table 1 sh ow s quarterly real CCC purchases and changes in both real farm in ven to iy and real farm inventory plus real CCC purchases since 1979." Th e m ean and standard deviation o f each series also are sh ow n for each year. T h e pattern o f changes in the overall m easure o f farm in ven to iy is m u ch sm ooth er appearan ce is deceptive; in fact, in ven to iy holdin gs w h e n CCC purchases are in clu d ed than w h e n they are not. This is esp ecia lly true w h en relatively large have sim ply m oved from private to federal govern m en t ow nership, or vice versa. changes in CCC purchases o ccu r. At these times, farm in ven tory investm ent sw ings w id e ly in the o p p o site direction, such as in IV/1982, IV/1983, 1/1984 and the “An inverse relationship between business inventory investment and government purchases of goods has been noted by Weidenbaum (1959) and (1961). His analysis emphasizes the time pattern of production and delivery and the NIPA accounting of such programs. The implied lack of a contemporaneous relationship of GNP and such spending was first pointed out in these articles. 9For the period shown in table 1, the correlation between changes in CCC purchases (1982 prices) and changes in farm inventory invest ment is - 0.56, which is statistically significant at the 1 percent level. 19 FEDERAL RESERVE BANK OF ST. LOUIS en d o f 1985. T h e standard d eviation fo r farm in ven tory investm ent each y e a r is sh arply h igh er than that fo r the total farm p rod u ct in ven tory change. Th is occurs because the m ovem en ts o f CCC purchases are offset b y o p p o site m ovem en ts in farm in ven tory purchases. O f course, this sm ooth in g effect also occu rs fo r the overall change in in ven tory — the sum o f business (non-farm and farm ) in ven tory change an d CCC purchases. CCC PURCHASES AND FINAL SALES W h ile fed eral purchases o f farm p rod u cts d o not affect GNP — the value o f final go o d s and services p ro d u ce d in the e co n o m y — th ey d o affect the m ea surem ent o f final sales, w h ic h equals G NP less the change in business in ven tory.10Analysts often focu s on final sales in o rd er to assess the strength and ou tlook fo r in com e, outpu t and em p loym en t. Assessm ents o f final sales are im portan t both because in ven tory and prod u ctio n decision s are based on expectation s o f such sales and because u n exp ected changes in sales are absorbed b y in ven tory fluctuations. Thus, m o ve m ents in final sales relative to p rod u ctio n provid e in form ation on future p rod u ctio n changes and can give rise to an in ven tory cy cle." W h en sales are less than produ ction, for exam ple, the u n sold produ cts increase inventory. If the rise in in ven tory is u n d esired and u nplanned, it w ill be elim in ated b y redu cin g p rod u ctio n grow th tem p orarily relative to that o f ex pected sales. M oreover, if m ovem en ts in G NP reflect tem p orary changes in p rod u ctio n to adjust inventory, (inal sales can be a m ore useful gauge o f the ou tlook than current p rod u ctio n o r GNP. CCC purchases have substantial quarter-to-quarter effects on the m easurem ent o f final sales. This occurs JUNE/JULY 1986 because such purchases affect the ch ange in business in ven tory but leave G NP u naffected. W h en CCC p u r chases increase, fo r exam ple, m easu red final sales ten d to rise because business (farm ) in ven to ry d e clines. Yet such purchases sim p ly represen t an oth er w a y o f h o ld in g farm inventory, n ot a significant in crease in overall sp en d in g on g o o d s and services that w ill likely lea d to in creased p rod u ctio n . Thus, if the change in business in ven tory is adju sted to in clu d e CCC purchases, the adju sted final sales m easure o b tain ed can m o re clo sely gau ge the actual final p u r chases o f go o d s and services b y consum ers, business, govern m en t and foreign purchasers. Chart 3 show s real final sales g ro w th both w ith o u t an adjustm ent and w ith CCC purchases subtracted from final sales. T h e largest differen ces in the grow th o f final sales, adju sted fo r CCC purchases, o ccu r after 1981. In the secon d h a lf o f 1982, relatively large CCC purchases contribu ted to final sales grow th . From the secon d to the fou rth qu arter o f 1982, real final sales ex p a n d ed at a 2.1 p ercen t rate, h igh er than the 1.1 p ercen t rate fo r adju sted real final sales. Subsequent redu ction s in the go vern m en t’s h o ld in g o f farm p rod u ct in ven to ry throu gh the PIK p rogram led to an u nd erstatem en t o f final sales grow th . From the fou rth qu arter o f 1982 to the fou rth qu arter o f 1983, real final sales ex p a n d ed at a 3.7 p ercen t rate, but this w as b e lo w the 4.8 percen t rate o f adju sted real final sales grow th . In effect, the transfer o f farm p rod u ct in ven tory from the go vern m ent to the private secto r a p p ea red o n ly as a net business in ven tory change, w h ic h u nd erstated the grow th o f final sales. O f c o u r se, these p eriod s m atch the en d o f the 1981—82 recession and early part o f the current expansion. Thus, the cyclical sw in g in m ea sured final sales gro w th understates the actual a ccel eration in adju sted final sales that took place. T h e m ost recen t CCC purchases, esp ecia lly in the fou rth qu arter o f 1985, are the largest o n record. In the 10While the assumed independence of CCC purchases and farm output within the quarter seems satisfactory, it might be argued that such purchases contribute to higher farm output than would other wise occur. To test these views, “ Granger-causality" tests were conducted on the quarterly change in farm sector output and the change in CCC purchases, both in 1982 prices, for the period 11/1973 to 11/1986. Optimal lags on the lagged dependent variable were chosen via sequential F-tests. The results indicate “ bidirec tional causality” : past CCC purchases negatively and significantly affect farm output; past changes in farm output positively and signifi cantly raise CCC purchases. When the contemporaneous value of the change in CCC purchases is included in the farm output equa tion, there is no significant past CCC effect and the contemporane ous CCC term is not significant for lags on the change in CCC purchases up to 10 quarters earlier. 11 The inventory cycle and its significance in U.S. business cycles from 1948 to 1976 is discussed in Tatom (1977). Digitized for20 FRASER secon d qu arter o f 1985 a nd the s econ d quarter- o f 1986, real CCC purchases w e re $2 b illio n and $4.5 billion, resp ective ^ . Thus, in each quarter, the final sales m easure w as little affected b y CCC purchases; o ver the w h o le year, real final sales and real final sales adju sted fo r CCC purchases rose 2.7 and 2.6 percent, re sp ec tively. M oreover, the pace o f overall in ven tory in vest m ent w as about the sam e in each quarter, so that real G NP g r e w at about the sam e rate o ver the year. But the patterns o f real GNP, real final sales and adju sted real final sales w e re qu ite differen t d u rin g the year. Table 2 sh ow s these gro w th rates. Both final sales series sh ow that p rod u ctio n g re w faster than sales in FEDERAL RESERVE BANK OF ST. LOUIS JUNE/JULY 1986 Chart 3 CCC Purchases and Real Final Sales G ro w th Compounded annual rate Compounded annual rate 15 the last qu arter o f 1985 and first qu arter o f 1986. So, not surprisingly, p rod u ctio n g ro w th slo w ed tem p orarily in the secon d qu arter o f 1986 to elim inate excess in ven toiy. Both final sales series also sh ow that sales grow th accelera ted in the secon d qu arter o f 1986. T h e princip al differen ces in table 2 are that sales grow th in 1986 w as stron ger a cco rd in g to the a dju sted series and that it a ccelerated fo r tw o quarters rather than one. T h e stro n ger sales g ro w th on an adju sted basis suggests stron ger grow th in aggregate d em a n d and m ore in cen tive fo r firm s to in crease produ ction and em p loy m en t than the u nadju sted data indicate. Also, the secon d qu arter acceleration in final sales appears less likely to be a fluke using the adju sted series. T h e acceleration sim p ly reinforces the pattern set in the previous quarter, instead o f appearin g to be the first sign o f positive sales grow th since the en d o f 1985, as in dicated in the u nadju sted data. 15 SUMMARY W h ile m ovem en ts in CCC purchases can be rela tively large, they have had no m ajor effects on final sales and o th er N IPA m easures until the past fe w years. D uring recent years, the pattern o f CCC p u r chases has had relatively large effects on m easured in ven to iy change, federal purchases and ex p e n d i tures, and final sales. In 1982 and 1983, the effect w as to raise the grow th o f both fed eral sp en d in g and final sales du rin g the last tw o quarters o f the recession and to lo w e r their grow th in the fii'st five quarters o f the subsequent expansion. M o re recently, record net pu r chases by the CCC in the last h a lf o f 1985 have given rise to a distorted pattern o f sales grow th , suggesting gen erally w eaker sales than the adju sted data in d i cate. Analysts w h o focus on u nadju sted data, a cco rd ingly, w o u ld u nderstate the recen t strength o f aggre gate d em an d and the short-run prosp ects fo r grow th . 21 JUNE/JULY 1986 FEDERAL RESERVE BANK OF ST. LOUIS Table 2 Growth Rates of GNP and Final Sales over the Previous Year Quarter ending Real GNP Real final sales Final sales less CCC purchases 111/1985 4.1% 6.1% 5.0% IV/1985 2.1 2.7 0.4 1/1986 3.8 -1 .3 1.6 11/1986 0.6 3.4 3.6 11/1985-11/1986 2.6 2.7 2.6 For p o licy purposes, fluctuations in CCC purchases can distort quarter-to-qu aiter m ovem en ts in im p o r Council of Economic Advisers. Economic Report of the President (U.S. Government Printing Office, February 1986), pp. 129-58. tant N IPA measures, p rovid in g a m islea d in g in d ica tion o f the strength o r w eakness o f fed eral spending, farm in ven tory investm ent and final sales. Fa ced w ith such distortions, analysts w ill fin d it useful to take m ore care in accou n tin g fo r these qu arterly m o ve m ents in CCC purchases and th eir effects o n key Herman, Shelby W. “The Farm Sector,” Bureau of Economic Analy sis, U.S. Department of Commerce, Survey of Current Business (November 1978), pp. 18-26. m easures o f eco n o m ic perform an ce. Tatom, John A. “ Inventory Investment in the Recent Recession and Recovery,” this Review (April 1977), pp. 2-9. REFERENCES Wakefield, Joseph C. “ Federal Farm Programs in 1986-90,” Bu reau of Economic Analysis, U.S. Department of Commerce, Sur vey of Current Business (April 1986), pp. 31-35. Belongia, Michael T. “ Outlook for Agriculture in 1983,” this Review (February 1983), pp. 14-24. Bureau of Economic Analysis, U.S. Department of Commerce. “ Special Note — The Commodity Credit Corporation in the Na tional Income and Product Accounts,” Survey of Current Business (January 1982), pp. 6-7. 22 Rosine, John. “The Farm Sector and GNP,” paper presented to the Federal Reserve Committee on Agriculture and Rural Develop ment, Board of Governors of the Federal Resen/e System, June 1, 1984, processed. Weidenbaum, Murray L. “The Timing of The Economic Impact of Government Spending,” National Tax Journal (March 1959), pp. 79-85. ________ _ “The Government Spending Process and Economic Activity,” The American Journal of Economics and Sociology (Janu ary 1961), pp. 169-79.