192

Manfred J. M. Neumann

Manfred J. M. Neumann is the director of the Institute tor Inter- national Economics and a professor ot economics at the University of Bonn,

I Commentary

MICHAEL BORDC) PROVIDES US with a WHICH REGIME PERFORMED comprehensive, scholarly study of the history of BEST? the three main international monetary regimes: the gold standard, the dollar standard, and the It is natural to evaluate the welfare implica- floating exchange rate. lie focuses on two tions of monetary regimes by asking what important questions. First, which regime provided different regimes achieve with respect to the the best performance with regard to the levels level and stability of inflation and real growth. of inflation and real growth? Second, what Any monetary regime can be described as a makes an international monetary regime viable? mechanism or device that delivers an average rate of monetary expansion and a variance of Because I am not a historian, I will limit my money growth. With respect to economic per- comments to two areas, I will first discuss the formance, the essential difference is whether a comparative evidence on the performance of particular m-egime pr’ovides governments with the three monetary regimes and use Bordo’s more or less freedom to manipulate the aver- statistics to infer a little more information on age rate of and the variance of monetat-y expan- the role of demand shocks under the different sion. It follows that regime differences should regimes. Thereafter I will concentrate on the be reflected in inflation levels and variances of important issue of determining a monetary inflation and per capita growth. system’s credibility. I find Bordo’s thoughtful Table I draws from Bot-do’s tables I and 4. 1 discussion of the issue useful. I should add, consider the Group of Seven countries as a however, that sometnnes he takes the literatui-e whole, the United States, Germany and Fi-ance too seriously—especially the affirmative litera- and concentrate on the three major periods: the ture on the European Monetary System (EMS). pre-World War I gold standard, the Bret ton \‘evertheless, Bordo forces us to consider which Woods system of the 1950s and 1960s, and the monetary system or standard, if any, can solve floating exchange rate in place since the the credibility problem in terms of firml inid~197Os.’Note that, in contrast to Bordo, 1 do anchoring market expectations about its not separate out the favorable performance of viability. the Bretton Woods convertible subperiod

‘The Group of Seven countries are Canada, France, Germany, Italy, Japan, the United Kingdom and the United States.

FEDERAL RESERVE BANK OF ST, LOUIS 193

/ fl__

~ / ~ _/ -I ~r _, /_ ~-~-c — ~:- / ~-~c ~~ ~ ~_-t~ ,___ ~ ~ ~ - w~ ~ v~~Asc/ / ~~ ~ A ~rir4~

~ ~‘ / ~ ~ ~ / c’~ ~ ‘A ~~\\ N~ ~tr~: —~r~-Jra -~ t~- -- - * ~ - —— -::~~---~~-‘ \t~ ~ - ~~~~/~\_ ~~ ~ rj~t;/~ ~ ~‘t~1~r~tç~ ~t~c~:4;~ ~ / ~ ~ :~~

~ N ~ ~- V~ ~ S~A*~~

~ ~ t~a~:c~~~~ p ~ ~s~/ ~ ~‘ \ ~___-c_ , c~~ ~ ~ -,~-- ~ r’-~-~-°~~ 4 ~ ~ ‘~ -7 / /~ -~ ~-\ 4 ~- ~~ ~ \ ~,, ~ / \\~‘~~ ~ ~-/ \-t- a>/ ~ //1-~ ~ ~ — ~ \ ~ ~:~-s-~> - ~ac~~_s~~s~:/~ / /~t~sJ~-~ \ \ t- 4 - ~~~~~ ; -~--~ $-/\ :~-

~tt-’ ~ c~r~ \ç/~ ~

~ - ~s~iie~ ~: : ~ c:~ ~ - ~ ~H: ~1 t~- c~ ~\/Y\~ a-~ ~ \~C ~- ~~ ~ ‘7 i-.~ 7 7 ~//p tc ,~/ ~ 7~-/’~~’ ~ — -,,~ / ~, f 7 7 - / -, - ‘7 4 ~7’ 7/ :4-- - — - ‘7 -‘7”-? -- -

/7- / ‘7 \- ‘7 /-//~/,/ ~ ‘7 ‘7’- “7?“ 4, -~ -~ -

- ‘7 - ‘-~ ,,7”’7 -‘ --i’~~~ ‘7~ ~r —~ ~ ~ >-~- ‘ ~-~“ ‘‘-‘77/- (1 959-197(J) in terms of inflation and output. • Third, the Bretton Woods regime exhibited The subperiod looked good on the surface; how- the highest variability of inflation, whereas ever, it was in fact the period when the break- output variability was closer to its level down of Bretton Woods was programmed. More under the float than under the gold generally speaking, for- ant’ i-egime we nught standard. find ey post a good looking suhperiod.2 The first observation on average inflation per- As Bordo and others have pointed out, the formance is well known and understood. It is data permit the following observations: widely accepted that the classical gold standard • First, average inflation was negligible under prevented the manipulation of monetary expan- the gold standard and highest under the sion by enforcing a direct link between the base floating exchange rate. money stock, the national stock of gold and the balance of payments. Though devaluation was • Second, the variabihtv of inflation, as possible by raising the gold parity in national well as that of real growth, was higher currency, it was rare. Thus the gold standard dcliv- under the gold standard than under the wed the lowest average rate of inflation, given that floating exchange i-ate. the available gold stock did not grow much.

2As Anna Schwartz pointed out in the discussion, an evalu- when the EMS came close to collapse, would be seriously ation of the EMS that bypasses the most recent period, misleading.

MARCH/APRIL 1993 194

member country of the European snake or EMS. At the other extreme, fiat money cum floating does not put any external constraint on domes- Checking the empirical validity of these con- tic money production. ‘I’hus governments are jectures requires estimating the variance of ft-ce to use money production to collect inflation nominal demand shocks. Bordo’s study pi-ovides tax and to dampen the business cycle. The addi- us with some valuable information in this nonal advantage to governments of the floating respect. Following Blanchard and Quah (19891 exchange rate is that the regime spares them and liavoumi and Eichengr-een (I 99Z) in particu- the political cost of negotiating devaluation. In lar, he has estimated for each country andl each sum, the floating exchange rate is the monetary monetary regime a bivariate vector autoregres- regime most conducive to inflationary policies. sion (VAR) for the rates of change of the price Finally, the I3retton Woods system was in level and output. The lower panel of table I between the gold standat-d and the floating provides the variances of the estimated aggregate exchange rate in that it started as a gold supply and demand shocks? Under the straight- exchange standard but was permitted to forward assumption that the distribution of real degenerate into a pure fiat money standard (the demand shocks was the same over the different dollat- standardl dut-ing the early 1960s when monetary regimes, differences in demand shock the United States gold stock fell short of the variability can he attributed to the operational value of outstanding dollar liabilities. differences of the regimes. More interesting than the average inflation The empirical findings are mixed. The data performance is the observed difference in the reject our I’irst conjecture. For the Group of volatility of inflation and output growth among Seven countries demand shock variability was regimes. But to what extent can this volatility he highest under the gold standard and lowest attributed to the operation of the different under the potentially permissive floating exchange monetary systems? rate regime. The most puzzling aspect is the high demand variability during the gold stand- EXPLORING THE ROLE OF ard period because not only was monetary pol- icy discretion constrained by the rules of the DEMAND SHOCKS regime, but also fiscal discretion was negligible, Apart from determining the level of inflation, at least by today’s standards. monetary regimes differ with respect to nomi- Our second conjecture, in contrast, is con- nal demand shock variability. I propose the firmed. Demand shock vai-iability was lowest in following conjectures. the United States during the Bretton Woods First, nominal demand shock variability is period and in Germany during the float. More- highest under the floating exchange rate and over, it can be shown for Germany using an lowest under the gold standard. This reflects F-test that the demand shock variance of the the diffeiences in the limits to monetary discre- float differed significantly from its value under tion. Because the degree of monetary discretion Bretton Woods at the 1 percent level of sig- is close to zero under the gold standard and nificance. In the United States the level of unlimited under the floating exchange rate, we significance was 10 percent. should observe that the variance of inflation Though we have not seen any test statistics of was caused predominantly by nominal demand Bordo’s VAR estimates, let us assume that the shocks under the floating exchange rate but by estimates are clean. On this assumption we may supply shocks under the gold standard. use them to investigate the contribution of the aggregate demand shocks to the variability of Second, in a fixed-exchange rate system the inflation and output growth under the different system leader sets the floor for nominal demand monetary regimes. To do so requires a model of shock variability. Consequently, for the Bretton aggregate supply and demand to determine the Woods period we should observe that nominal unknown price elasticities of aggregate demand demand shock variability was lowest in the and supply. United States. Similarly, during the floating rate period nominal demand shock variability should Table 2 provides the bare bones of such a have been lower in Germany than in any other model. The model is written in logs and has a 3 See Bordo’s table 4.

FEDERAL RESERVE BANK OF ST. LOUIS 195

- - ?~~1-‘7~’7o47//<~~~~C <~~J1~,’ ‘7-?-. ‘-4~’71-~ ~ ‘7s~’74 ~ -~ -~‘--41

, 47/-/’~— ‘7/ ft ~ ~ ~:

//

“7.‘7”t ‘7’7”-’7t’71~‘~ ‘~ —‘7 —--- —

: ,- -. ,, ‘1, ~Gs4’S*M4$P*— titslSo4i/ ‘7

‘ 4 ‘7 ‘72-2— _ a4,~14~ <‘g’ ~a rn.~~4’

:‘7 “-‘7 t~~’I A4ii~Is$~sS 4fl ‘7’ ~‘~“ ~ ~— ,~r ~“7‘ 4 ~ ‘---,~/-C-’7’ 4/___ ~ - --‘7/--?-, — / / ~7/~~-~ ~/4/~/_ ~ / -/ “7 a’,-, / 4’ 4 — ‘7 / ‘7<’ ‘7< —a - <-‘7--’7 <‘7 ?~ - ,~. c--4c --- -/*‘7/ “7< --/ t----< 2 t’7_f~ ~‘2- ~ ~<-2--,’-12-,2--:,’~t/-l ‘7Y-~-—4--;?

51 a’ ‘7 4/,,,-, /7 4 ‘7?

4 ~ “-a:~’~s’7’ ‘7 ‘‘7/ -/ ‘7 ‘ ‘-“7- ~ ‘fl’4. ‘7t8~,,<4\ ‘7 - _~&_z±___ / ~ ~ ‘7~<

MARCH/APRIL 1993 196 cases—notably the Bretton Woods convertible the volatility of aggregate supply shocks. subperiod.~Given the estinates of the slope This is a little surprising. Are we prepared coefficients, we can compute the contribution of to accept that systematic differences in the the variance of aggregate demand shocks to the level of demand shock variability are not observed vaiiances of inflation and real growth a characteristic feature of international in table 3. monetary systems?

Tahle 3 presents for each monetary regime We cannot, however, rule out that these findings the measured variances of inflation and real are statistical artifacts enforced by an inability gi-owth, as well as adjusted variances, which to separate demand from supply innovations exclude the contribution to volatility of the accurately in the VAR estimation. Bordo himself aggregate supply shocks, The numbers printed has noted that in some cases the overidentifying in parentheses indicate the percentage share in resti-iction (according to which positive supply the measured variance of the contribution of shocks should permanently raise output and the demand shock variance. Note that data from driye down the price level) is not satisfied. only four of the Group of Seven countries are included; data from Italy, Japan and the United Another indication of a possibly insufficient Kingdom had to he removed because it was identification is the estimated change in the impossible to compute the slope coefficients for slopes of aggregate supply and demand between these countries tin at least one subperiod). regimes. Figure 1 presents the average slopes of aggregate supply and demand for the four Table 3 can be summat-ized as follows: Group of Seven countries. What effect do we • First, for the United States, the leader of expect monetary regimes to have on these the Bretton Woods regime, we find that slopes? the variances of inflation and real growth were dominated by the volatility of demand Consider the model printed in table 4 which provides more structure than the model in table 2. shocks during this period. About 62 per- cent of the variance of inflation and 77 Because the model is linear in logs, the size of the alpha and beta coefficients depends on the pe’’cem~tof the variance of output growth can he attributed to the variance of demand agents’ price responsiveness, as well as on the shocks. Similarly, we find that for Germany shat-e in output of the respective input in the under the floating regime the inflation production function or of the respective and the output variance were dominated expenditure. by demand shocks, which accounted for Comparing the regimes of the gold standard 86 percent of the inflation variance and and Bretton Woods periods, we find that both 72 percent of the output variance. aggregate supply and demand schedules were • Second, for the four Group of Seven steeper under Bretton Woods. I would have countries we find that demand shocks expected the opposite on the argument that the produced a higher inflation variance economies were generally more open to interna- under Bretton Woods (5.2) than undet tional trade under Bretton Woods than before. the gold standard (4.8) or the floating Comparing the Bretton Woods regime with exchange rate (2.5). The result probably reflects the differential performance of the float, we observe that the aggregate supply schedule became steeper under the floating the two leading countries, exchange rate but the aggregate demand sched- • Third, for the four Group of Seven countries ule became more fiat. The first observation is in as a whole, the variance of demand shocks line with the model in table 4 because the posi- did not dominate the inflation variance tive dependence of the nominal exchange rate undet any of the three monetary regimes. (its log is denoted by e) on the domestic pt-ice Its contribution never exceeded 45 per- level implies a steeper aggregate supply schedule. cent. Thus we find over all regimes that For the same reason, the demand schedule should the inflation variance was dominated by also be steeper under floating. l-Iowever, the data

~Ofthe up to four real solutions for each case, Ichose the one which combines a negative slope of aggregate demand with a positive slope of aggregate supply.

FEDERAL RESERVE BANK OF ST. LOUIS 197

Figure 1 G-4

yd: Bretton Woods 14— yd: Floating yd~Gold Standard 12-

10-

8-

6- ys: Gold Standard 4-. ys: Bretton Woods ys: Floating 2 Output

Figure 2 United States 16—

14—

yd: Bretton Woods 12— yd: Floating

10—

8-

6- ys: Floating ys: Bretton Woods

4-

Output

MARCH/APRIL 1993 SiflOl IS dO NNVS 3A~3S381V83Q3d

cp~o~UJP uiojsAs uqi JO S~lJ3UflO3Jaqulaul uqi fl -PuLfad 30(1300 UPLUJOf) soaupaJ siqj SJP!JOIP[U 1 8 o~q~paJa~q IBM 8010.1 0141 aJq!poJJ UJP SOJUJ M MUJ fX)IJ0dU1! fO S3!Jd Lu3I?IOJ 011 ~Iui~hinqa Ji 9~qPTsaq IT!’~~uIa3SAS AJIT3000W ~eUoflPUJaflfl fljaOt3S cjddns use~CuPulJa’)U! do noqs piM 4 t UP U?143 ~P’~P~’°~ 03 Ulifi spuuj 8dU1~ UJ~cJnouow 014] JOAk)MOIf ‘0W!3 OUTPS 0143 fl! JO AJO1SI 014T Jo UO!TPUTUIPXO It1J011~) SOpJOq -IOUJOCI soaijd uuuJJaf) f)tIl? JcI!JUJOCIUTaT 30(13fl() 14 UPIUJO’) SOSIPJ luqi )130145 pUPWOf) P SU 401110 ~wajs csaqj JO A1!I!qep~ 1 -Jot) U~dn 410145 ~ pa~jaqi Aq f)aaUpOJd )330L38 ~n ~ suoqp3cIodxo flqJIiw ~kI!JOq3Ut? AIWJ!J ~UP00)p {PIIJW0U P ‘ajclwexa J0J AIIPU0!IPU 0 08008 aqj i-~wa qoad AiqiqIpoJo ~qi 0A~OSUPO 1 J03U! p~3I14LISUPJI OJP 5130138 aU1!~QJAJPI0110I\J ‘ALIP 41 ~majszs ia~~@uouiwuonPaiu~ qJ!qA~ 800000 3UapUadapU! sP sa!J11I003 3P9J3 3OUUPJ JO U013800b [P3110WPpUflJ aip do a~uiMOO I OM iPq] SaTfdLU! 14311441 ‘sOW!~aJ fPUO!JPUJQ1U1 WWIHOUd AIflIEHU3UD ~1FLL ~Ui~cpnis OJP a~ 30141 azis~qdwa jsnw aw!~aJ ATJPI000W aqi Xq pasooa loU ~ ,aqi so!ul000cja ~~JOM uo asodwi iCam AiqrqmJ~’~aq3 08003014 IUPAOJOJJ! 030 qaUJM spoqs f)UPIUOp ~ 00010 03 300(ISOJ If1LM JQJJIp AifeJ1]PwU3SATS sawd~aj,Cjwi ~U1? ~~pnis 03 148141 aM 143114M ‘s)fflOqS P l~ -auolu fPuoquuJaIlUI leqi OJOIJOIUOO 0141 ~3Ui]13WT] JPU!WOU 000Mlaq a1euUaJaJJ!P 1001103 UM 1014130 5!0 ~e iduiaiw pat~aqspjdosaiow iwijiou~ )faeqMPJp SHoIJUS P ‘sauiØ~aa AJtn000W JO 1 J4 0 0)~?W03 Oi%PL4 aM 0”a~~0q~ ‘UO!5H10U00 UI 1P!30 JOJJII) aqT JOIJU UJO aM asnoaaq ‘JUAOOJO]iV 0014000 TUOJW 1411M p041014 aq p~noqsso~ew AUPwJO9 U! )fDOlJS A d 1 -1180 14\~ASflj T~141OpJOq 1313M OOJ$P I ‘0105 Uf -dos rsp sajt~gpaiiu~0141 UJOJJ J30148 poe~uap I~~~1PU0U0141 AJLSSI14D IBM SO3PLUIJSO UVA sOpJofl Z UJfl~hJ009 TPO3J 0143 JapiJo aqi ‘zCttuanbasooj iCfdul! JO 5U0130!J3SOJ ~u!1cJ!3Uap~ IOu UJOW OJP safnpaqas qloq TPIJT Plop ~cfluauewJadseaud opwJarj sosiuj pu~ cpuau 50p119 pallufi aqi TWill OJOU 081%! A dwoa lOU Of) 1 1

/ 27 2 N ~ 2 NN ~ NN/ N N N 2NN N~ ~7N~t 7~*N — ~N N NNN N NIN

- 44 N N_N N ,~ N~

44 N >/N~ IN N~N ~~*~P~øfl*bIt /

N ~ ~ N /NN 7 ~/ N 2~ NN/ N N ~ N N 4÷ N? N

N N NN NN NN NNN/ N/ / I ~N N~ N 7 NN~~*øS~* NN

N NNN N N N/N

N NN N I/N N N N N NNN7 N N 4 N N NN NN NN N /NNNNNN/N N N NNNNN 2 N~ N /NNNNN/ NN N NN dNtWN

N N N NN NNN N/ N N N tN N N

NN N \N N44_/fl~ ++~~ N NN N N N N N N N N NNNN NN N4~ I

N N N N

N N N N4 N N N lmwlfleu s~s~~~i~ii put s~dotsN NN

NN N N ~ N/N / NN N /

I NNNN~N 2I~~ ~ N~IN~Y1N~ ~NIN N2~ NNNNNN / 7 NN/N NNNN/N2NN1NNN~ 4NN/ N NN/ N NN/N\NNN N/7NNNN//1NNN NN N

861- 199 to honor the rules. And member countries will vis-~-visthe center currency by writing the honor the rules if there is a center country that fixed exchange i-ate into the country’s constitu- enforces the rules. Accordingly, the classical tion as Sweden did during the gold standat-d. gold standard did not break down because the The alternative international regime is created United Kingdom, its center country, was com- by an agreement that all governments precom- mitted to convertibility. In contrast, the United mit to price stability at home by providing their- States, as the center country of Bretton Woods, central banks with constitutional independence. was not committed to convertibility and main- tenance of price stability. Bretton Woods Which of the two regimes is preferred? The consequently broke down. Finally, Germany’s first regime provides price stahility for all coun- commitment to price stability made the EMS a tries in the medium to long run. The precom- successful and viable system. Unfortunately, the tnitment to fixed exchange rates by ti-I members latter prediction held only until last September. implies that idiosyncratic shocks will be dis- tributed over member countries at full force, as Though Bordo’s reasoning makes a lot of was the case under the classical gold standaid. sense, it fails to address two essential questions. Because fiscal policy is an important source of First, by which means or under what conditions idiosyncratic shocks, the regime will hardly be will the center country be able to enforce the attractive without a (enforceable) rule prohibit- j-ules? Second, and more fundamentally, what ing public deficits. conditions are required to make the center country keep its commitment? The altet-natiye regime of uniform precommit- ment to price stability at home might be rejected In my view, any international monetary sys- by some as a nonsystem. But semantics apart, the tem that is based on commitment to rules will setup is not to be equated with unconstrained be fragile. Commitment should be replaced by floating. The regime will provide nominal ex- precommitment. The game theory reformulation change rate stability though not fixity. Depending of the pathbreaking analysis by Kydland and on the judgment of central bankers, the regime Prescott proves that governments cannot com- 5 might eyolve into an adjustable peg system where mit to price stability. In contrast to commit- up to n-I central banks unilaterally peg their ment, pi-ecommitment does not depend on a currencies to the cutrency of a center country government’s good will or interest. Instead it is in a flexible manner-. ‘I’his means that in the ad- created by setting up an external mechanism vent of a sizable country-specific shock at home or that effectively ties the hands of current and in the center country. they will per-nut exchange future governments. rate adjusting. In contrast to the non-precommitted An international monetary regime will he sta- governments in Europe, the independent central ble and therefore durable if it provides the hankers will have no interest in defending mis- institutional constraints fot- a subgame-perfect aligned exchange rate parities. supergame. The fundamental constraint is effec- In conclusion, the Bretton Woods system and tive precommitment by all membet- govern- the EMS broke down because both systems were ments. There are two types of precomnnitment: built on unenforceable commitment instead of precommitment to pt-ice stability at home and precommitmnent. precommitment to a fixed exchange rate vis-~- vis another currency. Consequently, we can REFERENCES design two alternative regimes. Bayoumi, Tamin, and Barry Eichengreen. ‘Shocking Aspects A first negime resembles the EMS but commit- of European Monetary Unification,” NBER Working Paper ment is replaced by precommitment. The center No. 3949 (January 1992). country precommits on price stability at home Blanchard, Olivier, and Danny Ouah. ‘The Dynamic Effects of by providing its central bank with the constitu- Aggregate Demand and Aggregate Supply Disturbances,” American Economic Review (September 1989), pp. 655—73. tional status of independence from government. Elsewher-e I have laid out a sufficient set of Kydland, FE., and E.C. Prescott- “Rules Ratherthan Discretion: The Inconsistency of Optimal Plans,” Journal of Political institutional elements that provides an incentive- Economy, (June 1977), pp. 473—91. compatible status of independence.G The other Neumann, M.J.M. “Precommitment by Central Bank Inde- countries precommnit on a fixed exchange rate pendence,” Open Economies Review (1991), pp. 95—112.

5 See Kydland and Prescott (1977). 6 See Neumann (1991).

MARCH/APRtL 1993