The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
Federal Reserve Bank of New York Quarterly Review A utum n 1986 Volum e 11 No. 3 1 R e sp o nsive n e ss of In te re st Rate S preads and D e p o sit F low s to C hanges in M arket R ates 11 W age R ig id ity in W est G erm any: A C om parison W ith the U.S. E xp e rie nce 22 The C ycle in P ro p e rty /C a s u a lty Insurance 31 Tax R eform and the M erger and A cq u isitio n M arket: The R epeal of G e n e ra l U tilitie s 36 43 Treasury and Federal R eserve Foreign E xchange O p e ra tio n s A u g u s t-O c to b e r 1986 M a y-Ju ly 1986 The Quarterly Review is p u blished by the Research and Statistics Group o f the Federal Reserve Bank o f New York. Am ong the m em bers o f the staff who contributed to this issue are JOHN WENNINGER (on the responsiveness of interest rate spreads and deposit flows to changes in m arket rates, p a g e 1); LINDA A. BELL (on w age rigid ity in West Germany com pared to the U.S. experience, page 11); R O B E R T ! McGEE (on the cycle in property/casualty insurance, pag e 22); and CHARLES STEINDEL (on the effect o f the repeal of General Utilities on the m erger and acquisition market, page 31). Two quarterly reports on Treasury and Federal Reserve foreign exchange operations for the periods A ugust through O ctober and M ay through July begin on pages 36 and 43. Responsiveness of Interest Rate Spreads and Deposit Flows to Changes in Market Rates Changes in interest rates have long been recognized as an influence on the growth of transactions balances (M1). As market rates rise, depositors have typically reduced their money holdings because the interest income they forgo in holding money balances increases as market rates rise. When the Monetary Control Act of 1980 set a tim etable for a gradual deregulation of interest rates on consumer deposits, it was widely rec ognized that the demand for transactions balances would probably respond differently to changes in market rates than it had in an environment where deposit rates were subject to officially imposed ceilings.1 But it was not certain whether these balances would become more or less sensitive to changes in market rates because it would depend to a much larger degree on the rate-set ting policies of the banks. It now appears, however, that banks have adjusted the rates on deregulated accounts (both on time deposits and transactions accounts) in such a way that the demand for transactions balances has been considerably more interest-sensitive than it was prior to 1980. If these banking practices continue, M1 is likely to speed up or slow down far more than it did in the past in response to decreases or increases in market rates. Deregulation has produced an envi ronment in which changes in market rates have con tinued to affect the attractiveness of holding M1 bal ances relative to m arket instrum ents. In addition, changes in market rates now can affect the attractive ness of holding M1 balances relative to time deposits by causing spreads between the rates paid on time deposits and M1 balances to narrow or widen. ’ For a detailed listing of the steps in the deregulation of consumer deposits, see R. Alton Gilbert, "Requiem for Regulation Q: What It Did and Why It Passed Away,” Review, Federal Reserve Bank of St. Louis, February 1986, p. 31. Since the third quarter of 1984 (when short-term rates peaked) these interest rate spreads have narrowed considerably (see Chart 1). At the same time, M1’s growth rate accelerated from 5.4 percent in 1984 to over 11.5 percent in 1985 and the first half of 1986, and its velocity dropped from a 3 percent increase in 1984 to a negative 5.25 percent over the past year and a half. By comparison, during the 1960s and 1970s M1’s growth averaged about 5 percent and velocity increased about 3 percent per year. Because changes in these rate spreads seem to affect M1 and velocity growth so dramatically, a question arises about how these rate spreads adjust to changes in market rates in a deregulated (flexible-rate) environ ment. In other words, the responsiveness of M1 growth to changes in market interest rates now depends on how rate spreads (between market rates and the rate paid on M1 balances as well as between the rates on time deposits and M1 deposits) adjust to changes in market rates. The large changes in these rate spreads as interest rates fell in 1985 and the first half of 1986 demonstrated that banks do not adjust the rates on various types of deposits in step with market rates, leaving rate spreads (and hence the incentives to shift funds) unaffected as might have been expected in a deregulated structure.2 In a regulated environment, the spreads between the market rate and various consumer transaction and time deposit rates tended to move in step with market rate changes because the rates on bank deposits did not change as long as market rates were above the ceiling 2For more on this aspect of deregulation, see R.G. Davis, "Monetary Targeting in a Zero Balance World," in Interest Rate Deregulation and Monetary Policy, Asilomar Conference, Federal Reserve Bank of San Francisco, November 1982. FRBNY Quarterly Review/Autumn 1986 1 Chart 1 S ele cted In terest Rate Spreads and D eposit Flows As m arket and tim e deposit rates fe ll, rate spreads re la tive to NOW accounts narrowed and M1 grow th accelerated . . . Percent Percent 16 ------------------------ 1 --------------------------1------Growth rate from tw elve m onths e a rlie r 12 - \ 10- Treasury bill rate j / \ \ v ' Six-month * _ small time _ V deposit rate \ \J / \ " \ \ - ............. account rate ■ NOW account rate III ,1 1.J I 1 1 1.1 1 1 1 1 1 11 1 1 1 11 1 1 1984 1985 1 1 1 1 1..J—1—L.. I l l , 1986 . . . re fle c tin g the more rapid grow th in NOW accounts, w h ile sm all tim e deposits sto pp ed grow ing. Percent 30 --------Growth rate from tw elve months earlie r Percent 20- Growth rate from tw elve months e arlie r Small time deposits 10 5l I I I I I I I l I I I 1984 II I I I II I II1 I I I II I I II I I I 1985 1986 -5I l l I I I I I I I I I I I I l l I I I I I 1 l 1984 Sources: Federal Reserve Board, Statistical Release H.6; Federal Reserve Bulletin 2 FRBNY Quarterly Review/Autumn 1986 1985 1986 specified by Regulation Q. Therefore, even though consumers did have incentives to shift funds between market instruments and bank deposits, they had little incentive to shift between time deposits and transactions accounts because these rate spreads tended to remain constant. As long as banks were paying the ceiling rates on deposits, the spreads between the rates on various types of deposits did not change even when market rates increased or decreased. But now that banks can pay the rate the market dic tates on consumer transactions and time deposits, it is important to study the effects that changes in market rates will have not only on (1) the spreads between market rates and the rates paid on bank deposits, but also on (2) the rate spreads between the different types of bank deposits (between time deposits and transac tions deposits, for exam ple). This second point is important, of course, because time deposits are part of the M2 definition of money while transactions balances are in M1. The dramatic slowdown in time deposits in 1985 and the first half of 1986, along with the concur rent acceleration in M1 growth as the spreads between the rates offered on time deposits and transactions deposits narrowed, suggest that substitution between M2 components could cause the demand for M1 to be more interest-sensitive in a deregulated environment (Chart 1).3 Of course, when rate spreads change con sumers can move funds not only between time deposits or market instruments and M1 balances, but also into and out of money market deposit accounts (MMDAs) and money market mutual funds (MMMFs). These components of M2 could be alternatives to holding M1 or time deposit balances as interest rate spreads change. In the first section of this article, the responsiveness of various rate spreads to changes in market rates is reviewed on the basis of some econometric results. By and large, banks, with a lag, have fully adjusted the rate on time deposits to reflect changes in market rates. But they have made only a partial adjustment to their MMDA rates and have been very slow to adjust the rate on deregulated transactions balances (Super NOWs). Hence, even though rates on deposits have been de regulated, consumers still have had an interest rate incentive to reduce their liquidity when market rates rise and increase it when market rates fall.4 In the second section of this paper, we review the problems of estimating the responsiveness of deposit *For more detail, see J. Wenninger and L. Radecki, "The Monetary Aggregates in 1985,” this Review, Winter 1985-86. 4Of course, now that the rates on consumer deposits are competitive with market instruments, the demand for bank deposits might be affected to a larger degree than in the past by changes in inflationary expectations or changes in the expected returns on other investments such as stocks. flows to changes in rate spreads. As a general note of caution, we have had too little time in a deregulated environment to make very precise estimates. But even with more time, the high correlations among the interest spreads that would affect the deposit flows will make estimates difficult. Nevertheless, we can anticipate the direction of response that some of the components of M2 will make to changes in market rates, based on the way interest rate spreads have responded to changes in market rates. Responsiveness of interest rate spreads Ten interest rate spreads are studied in this article: (1) Six-month Treasury bill less six-month time deposit. (2) Six-month Treasury bill less Super NOW. (3) Six-month Treasury bill less MMDA. (4) MMDA less Super NOW. (5) Six-month time deposit less Super NOW. (6) Six-month time deposit less MMDA. (7) Six-month Treasury bill less MMMFs. (8) Six-month time deposit less MMMFs. (9) MMMFs less MMDAs. (10) MMMFs less Super NOWs.5 The table in the box shows the results when the weekly changes in these ten rate spreads are regressed on current and lagged changes in the six-m onth Treasury bill rate. Based on these results, Charts 2 through 6 illustrate the response over time of the rate spreads to changes in market rates. Chart 2 (bottom line) shows the response over time of the spread between the six-month bill rate and the six-month time deposit to changes in the six-month bill rate. In other words, we want to see what happens to the spread between the market rate and the time deposit rate when the market rate changes. The chart shows that initially the spread widens considerably, but after about ten to twelve weeks banks have adjusted the rate on time deposits to reflect completely the change in market rates. At the other extreme, banks are very slow to adjust the Super NOW rate when market rates change.6 *ln theory, the spreads between the rates earned on longer term time deposits and these deposits could be important as well. To keep the number of rate spreads manageable, however, longer term rates on time deposits were not included. For an analysis of longer term deposit rates as well as short-term rates, see Paul O’Brien, "Deregulated Deposit Rate Behavior," Board of Governors of the Federal Reserve System, April 1986, unpublished. •As of March 31, 1986, the distinction between conventional NOW accounts and Super NOWs was no longer meaningful. By that time, the minimum balance requirements for Super NOWs had been eliminated and the interest rate ceilings on savings deposits FRBNY Quarterly Review/Autumn 1986 3 The top line in C h art 2 show s that even a fte r tw elve w eeks the Super NOW rate has changed by only about 25 percent of the change in m arket rates, and therefore changes in m a rket rates have had lo n g -la stin g effe cts on the spread betw een m a rke t rates and the S uper NOW rate. (T his can also be seen from C hart 1.) Between these two extrem es is the responsiveness of the MM DA rate. The ce n te r line in C hart 2 show s that after a tw elve-w eek period, the spread between market and MM DA rates has a d ju ste d about 60 perce n t of the w ay to the ch ange in m arket rates, as com pared with 25 p ercent fo r S uper N O W s and 100 perce n t fo r tim e d e p o s its . T h u s , a lo n g th e liq u id ity s p e c tru m fro m transactions accounts to tim e deposits, there have been in cre a sin g ly fu lle r a d ju s tm e n ts to changes in m arket in te re st rates. A pparently, banks have not m ade rapid ad ju stm e n ts to the Super NOW rate, either due to lack of experience in pricing th ese a cco u n ts, or relu cta n ce to m ake fre qu ent ch a n g e s to the te rm s o ffe re d on tra n sa ctio n s accounts once they set a com bination of fees, minimum balances, and an interest rate on Super NOWs. In other words, banks may have wanted to market Super NOWs not as flexible rate accounts, but as fixed-rate accounts on w hich the te rm s do not change fre q u e n tly but co n sum ers still earn a fa ir rate of return on a verage over the long er run. On the other hand, banks have had considerably more experience with offering flexible rates on time deposits. Indeed, for several years the ceiling rates on six-month tim e d e p o sits w ere linked d ire c tly to changes in the Treasury bill rate. H ence, banks w ere q u icke r to adjust tim e d e p o sit rates to fo llo w m arket rates a fte r the c e il ings rates on tim e deposits were elim inated. In addition, w ith tim e d e p o sits banks are ad ju sting only the rate o ffered on m aturing or new d e p o sits; the rate on the no nm a tu rin g stock re m a in s unchanged. H ence, th e ir c o s t o f fu n d s fro m th is s o u rc e a d ju s ts g ra d u a lly to changes in m arket rates even if they quickly match any change in m a rke t rates w ith a change in tim e dep o sit rates. In co n tra st, w ith S uper NOW s any change in the rate offered by banks affects the entire stock of deposits since S up er N O W s fo r all p ra ctica l purposes do not have a m aturity like tim e deposits do. Therefore, banks may feel th a t the y have b e tte r co n tro l over the cost of fun ds from th is sou rce if th e y prom ote them as fixed- rate accounts or acco u n ts on w hich the te rm s change only infrequently. In any case, it a ppears th a t M1 has retained a sig n ific a n t, if not a la rg e r in te re s t e la sticity o ver the last few years as a re su lt of the w ay banks h ave a d ju s te d th e ra te s on S u p e r N O W s a nd tim e d e p o s its .7 MM DAs, as a c o m b in a tio n sa v in g s /c h e c k in g in s tru m ent, p ro b a b ly in v o lv e a c o m b in a tio n of th e above considerations for banks. MMDAs were introduced as a m eans fo r banks to com pete e ffe c tiv e ly w ith M M M Fs. H ence, th e y w ere view ed from the s ta rt as a fle xib le rate deposit, and banks may have been predisposed to 7This result may not hold in the very long run, of course. If market rates changed and then held steady for a very long period, banks would probably over time adjust the Super NOW rate to reflect this change fully, after allowing for the cost of required reserves. Over shorter periods of time in a less stable interest rate environment, however, it appears that significant changes in rate spreads can occur that strongly affect M1's growth. The overall responsiveness of deposit flows to changes in market rates depends, of course, not only on how rate spreads change but also on how responsive consumers are to a given change in these spreads. In this article, we are focusing primarily on how much rate spreads adjust to changes in market rates. Chart 2 Responses of the Spreads Between the Treasury B ill Rate and the Time Deposit, MMDA, and Super NOW Rates to Changes in the Treasury B ill R ate* Cum ulative percentage points Footnote 6 continued (including conventional NOWs) were no longer effective. For most of the three-year period studied in this article, however, the distinction between Super NOWs and conventional NOWs was important because banks could vary the rate on the former, while there was a ceiling rate on the latter. How banks varied the interest rate on Super NOWs during this period in response to changes in market rates is of interest because it gives some insights into how they are likely to administer all NOW accounts now that they are deregulated. 4 FRBNY Quarterly Review/Autumn 1986 Weeks ^C u m u la tive responses of the spreads between the six-m onth bill rate (TB) and the six-m onth time deposit rate (TD), MMDA rate (MDA), and Super NOW rate (SN) to a change in the bill rate (percentage points). Estim ating the Response o f Rate Spreads To Changes in Market Rates To illu s tra te ho w v a rio u s in te re s t rate s p re a d s have re s p o n d e d to c h a n g e s in m arke t rates, th e ch a n ge in each sp re a d w as re g re sse d on the c u rre n t and lagged ch a n g e s in th e s ix -m o n th T reasury bill rate (see ta b le ). T he sum of th e c o e ffic ie n ts , w h ich rep re se n ts the total resp o n se o v e r th irte e n w eeks, is show n at th e bottom of each c o lu m n . For exa m p le , the to ta l resp o n se of the spread betw een the six-m onth bill rate and the six-m onth tim e de p osit rate to a change in the bill rate is zero. That is, w hen th e b ill rate in cre a se s by a give n am o u n t, so does th e tim e d e p o s it rate, le a vin g the sp re a d a fte r th irte e n w e e ks u n a ffe cte d (colum n 1). In co n tra st, the to ta l re sp o n se o f th e sp re a d betw een the s ix-m o n th bill rate and the S uper NO W rate to a change in the bill rate is 0 .7 7 (c o lu m n 2 ). T h u s , if th e T re a s u ry b ill ra te increases one percentage point, after thirteen weeks the sp re a d b e tw een the T reasury bill rate and the Super NO W rate w ill be a b o ut th re e -q u a rte rs of a p e rce n ta g e point wider than it was before the change in the bill rate. In o th e r w ords, the S u p e r NO W rate o n ly a d ju sts by ab o ut 25 p e rce n t (1-0 .7 7 ) of the ch a n g e in the m arket rate, leaving the spread a b o ut 0 .75 p e rce n ta g e p o in ts wider. The rem a in in g e ig h t co lu m n s in the ta b le show w hat happens to o th e r rate sp re a d s w hen the bill rate ch a n ge s.* "In the table, there are four basic equations (shown in columns 1, 2, 3 and 7) which determine how -the rates on time deposits, Super NOWs, MMDAs and MMMFs adjust to changes in the market rate. The response of the six remaining spreads to changes in market rates can either be estimated, as was done here, or calculated from the results obtained from the four basic equations Response of Various Rate Spreads to Changes in Treasury Bill Rate* (1) Six-Month Bill Less Six-Month Time Dep t t-1 t-2 t-3 t-4 t-5 t-6 t-7 t-8 t-9 t-10 t-1 1 t-12 Total R2 D.W. 0.81 -0 .2 0 -0 .1 3 -0 .0 7 -0 .0 4 -0 .0 5 -0 .0 7 -0 .0 8 -0 .0 7 -0 .0 5 -0 .0 5 0.01 0.01 0.02 0.94 1.48 (41.0) (10.4) (6 3 ) (3.7) (2.1) (2.3) (3.3) (3.8) (3.5) (2.4) (2.4) (0.6) (0.5) (6) Six-Month Time Dep. Less MMDA t t-1 t-2 t-3 t-4 t-5 t-6 t-7 t-8 t-9 t-10 t-11 t-12 Total R2 D.W. 0.15 0.13 0.05 0.03 0.01 - 0 01 0.001 0.01 0.004 0.02 -0 .0 0 8 -0 .0 2 -0.0 1 0.366 0.50 1.32 (7.7) (6.5) (2.6) (1.7) (0 7 ) (0 5 ) (0.8) (0.8) (0.2) (1.2) (0.4) (0.9) (0.6) (2) Six-Month Bill Less Super NOW 0.99 -0.01 -0 .0 3 -0 .0 2 -0.01 -0 .0 3 -0 .0 5 -0 .0 3 -0 .0 3 -0.0 0 2 -0 .0 3 0.02 0.001 0.769 0.96 1.80 (58.0) (0.9) (1-8) (1.1) (0.8) (1.6) (2.8) (1-8) (1.6) (0.1) (1.6) (1.0) (0.06) (7) Six-Month Bill Less MMMFs 1.04 -0 .2 6 -0 .1 3 -0 .1 3 -0 .1 2 -0 .0 8 -0 .0 5 -0 .0 4 -0 .0 7 -0 .0 3 -0 .0 6 -0 .0 2 -0 .0 2 0.03 0.92 2.42 (36.0) (9 1) (4.4) (4 3) (4.1) (2.7) (1.8) (1.3) (2.3) (1 0 ) (20) (0.7) (0.7) (3) Six-Month Bill Less MMDA 0.95 -0 .0 8 -0 .0 7 - 0 04 -0 .0 3 -0 .0 5 -0 .0 5 - 0 06 -0 .0 6 -0 .0 2 -0 .0 5 - 0 006 -0 .0 0 2 0.432 0.94 1.67 (45.8) (3.8) (35) (19) (13) (26) (2.4) (2.9) (3.1) (1.2) (2.6) (0 1) (0.3) (8) Six-Month Time Dep. Less MMMFs 0.23 -0 .0 6 -0.0 0 2 - 0 05 - 0 08 -0 .0 3 0.01 0.04 0.0001 0 02 -0.0 1 -0 .0 3 -0 .0 3 0.008 0.27 2.28 (72) (18) (0.07) (16 ) (2.4) (1.0) (0.4) (1.1) (0.004) (0.6) (0.4) (0.9) (0.9) (4) MMDA Less Super NOW 0.03 0.06 0.04 0.02 0.01 0.03 0.003 0.03 0.04 0.03 0.03 0.02 0.001 0.344 0.43 1.83 (24 ) (5.0) (3.2) (16 ) (09) (2.1) (0.2) (2.3) (2 8 ) (2.0) (2.1) (1.7) (0.1) (9) MMMFs Less MMDA -0 .0 8 0.18 0.05 0.09 0.09 0.02 0.00 -0 .0 2 0.004 0.004 0.004 0.01 0.02 0.452 0.33 2.12 (24) (5 3 ) (15 ) (25) (2.7) (0.7) (00) (0.7) (0.1) (0.1) (0.1) (04 ) (0.5) (5) Six-Month Time Dep. Less Super NOW 0.18 0.19 0.09 0.06 0.03 0.02 0.02 0.04 0.04 0.05 0.02 0.005 -0.0 1 0.735 0.70 1.30 (94 ) (99) (48 ) (2.8) (0.9) (0.9) (0.9) (23) (21) (2.6) (1.0) (0.3) (06 ) (10) MMMFs Less Super NOWs -0 .0 5 0.25 0.10 0.11 0.11 0.05 0.004 0.007 0.04 0.03 0.03 0.04 0.02 0.741 0.49 2.12 (16) (73 ) (2.8) (3.1) (3.1) (1.5) (0.1) (0.2) (1.2) (09 ) (0.9) (1 1 ) (0.5) 1 *T-statistics in parenthesis. Source: Bank Rate Monitor. Estimation period: weekly 10/12/83 to 7/23/86. FRBNY Quarterly Review/Autumn 1986 a d ju s tin g th e ra te on M M D A s w h e n m a rk e t ra te s changed more than the rates on Super NOWs. However, lik e S u p e r N O W s, th e ra te on th e e n tire s to c k of MMDAs changes when banks adjust the rate offered on M M DAs. A gain banks m ight be slow er to adjust the rates on M M DAs than on tim e d e p o sits in an e ffo rt to avoid large flu c tu a tio n s in the costs of fu n d s from this source. On bala nce, it is not su rp risin g th a t the rate on M M DAs has show n a re sp o n se to changes in m arket rates that is between the responses of the tim e deposit and the S up er NOW rates. W hen m arke t rates change, the rate spreads change not on ly betw een m a rke t rates and va rio u s bank lia b il itie s but also betw een the typ e s of bank lia b ilitie s . And ch a nge s in th e se spre a d s m ight induce sh ifts betw een co m p o n e n ts of M2, p erhaps a ffe ctin g the grow th of M1 as a result. C hart 3 (botto m line) illu s tra te s the e ffe ct on the rate spread betw een M M DAs and S uper NOW s w hen the m arket rate change s. Initially, banks are slow to adjust both of these rates to changes in market rates, but after Chart 3 Responses of the Spreads Between the MMDA and Super NOW Rates and Between the Time D eposit and Super NOW Rates to Changes in the Treasury Bill R a te * a tw e lve -w e e k period, the spread changes by about 33 percent of the change in m a rke t rates, c re a tin g an in ce n tive fo r co n su m e rs to s h ift fu n d s betw een M M DAs and S uper NOW s. T hese sh ifts could have som e e ffe ct on th e g ro w th o f M1 b u t w o u ld le a v e M 2 g ro w th unchanged. We w ould exp e ct the e ffe c t on M1 grow th to be sm all because the rate spread does not appear responsive enough to changes in m arket rates to cause large su b s titu tio n s betw een MM DAs and S u p e r NOW s. The top line in C hart 3 show s m uch m ore d ram atic e ffe cts on the rate spread betw een tim e d e p o s its and S uper NOW s w hen the m arket rate changes. As in the previous case, there is little effect in the first week. But a fte r tw e lv e w e e k s , th e ra te s p re a d b e tw e e n tim e deposits and Super NOWs has moved by 75 percent of the am ount th a t the m arket rate changed. T h is re flects the te n d e n cy fo r the tim e d e p o s it rate (w ith a lag of twelve weeks) to follow the m arket rate much more fully than the S uper NOW rate does. As a re su lt, s u b s titu tions between tim e deposits and Super NOW s are likely to have size a b le e ffe cts on M1 grow th w hen m arket Chart 4 Response of the Spread Between the Tim e D eposit Rate and the MMDA Rate to Changes in the Treasury B ill Rate* Cumulative percentage points 1 0-------------------------------------------------------------- Cumulative percentage points 0 .9 ----------------------------------------------------------------------------------------- 0 . 9 ----------------------------------------------------------------------------------------- 0.8 0 . 8 ----------------------------------------------------------------------------------------- 1.0----------------------------------------------------------------------------- 0.7 ----------------------------------------------------------------------------------------0.6 ----------------------------------------------------------------------------------0.5 ----------------------------------------------------------------------------------------- 1 2 3 4 5 6 7 8 Weeks 9 10 11 12 13 Weeks ♦Cumulative responses of the spreads between money market deposit rate (MDA) and the Super NOW rate (SN) and between the time deposit rate (TD) and Super NOW rate to a change in the bill rate (percentage points). 6 FRBNY Quarterly Review/Autumn 1986 Cumulative responses of the spread between the six-month time deposit rate (TD) and the money market deposit rate (MDA) to a change in the bill rate (percentage points). rates change. As noted earlier, th is w ould in a sense be a new so urce of M1 grow th w hen m arket rates change and could well be contributing to M 1’s increased responsiveness to interest rate changes in recent years. Moreover, changes in m arket rates might also prompt som e sh iftin g of fund s from tim e d e p o sits into more liquid MMDAs. This w ould not affect M1 or M2 but would affect the overall liquidity of the consumer sector. Chart 4 show s th a t the spre ad betw een the tim e d e p o sit rate and the MM DA rate a fte r tw e lve w eeks changes by about 40 percent of the am ount that the m arket rate has c h a n g e d . T h e re fo r e , w h ile c o n s u m e rs m ig h t a ls o respond to low er rates on tim e d e p o sits by in cre a sin g th e ir h o ld ings of M M DAs, th e ir response is not like ly to be very large because the im pact on the rate spread when m arket rates change is quite small (about 50 per c e n t o f th e s iz e o f th e im p a c t on th e ra te s p re a d betw een tim e d e p o sits and S uper NOW s in C hart 3). F in a lly , c h a n g e s in m a rk e t ra te s a ffe c t th e ra te s e arned not only on va rio u s typ e s of bank d e p o sits, but a ls o on a v e ry c lo s e s u b s titu te fo r b a n k d e p o s its , MM M Fs. C hart 5 show s th a t a change in m arket rates does not re su lt in a p e rm anent change in the rate spreads betw een m arket in stru m e n ts and the M M M Fs or betw een tim e d e p o sits and the M M M F s.8 However, more sizeable changes in rate spreads between MMMFs and M M D A s o r S u p e r N O W s h a ve o c c u rre d w h en m a rk e t ra te s c h a n g e (C h a rt 6 ). H e n c e , c h a n g e s in m arket rates could result in som e fu n d s flo w in g into or out of M M M Fs and out of or into NOW acco u n ts or MM DAs. M oreover, since fa irly large sp re a d s betw een the rates on MMMFs and MMDAs have occurred, it does not a p p e a r th a t the rates being offered by M M M Fs are the prim ary fa c to r d e te rm in in g how banks set the rate on MMDAs. 8This result should be expected from the basic way MMMFs operate. That is, as their market instruments mature and are gradually reinvested at the prevailing interest rate, the average rate of return on their overall portfolio gradually moves toward the market rate. Chart 6 Chart 5 Responses of the Spreads Between the MMMFs and MMDA Rates and Between the MMMFs and Super NOW Rates to Changes in the T reasury B ill Rate* Responses of the Spreads Between the Treasury Bill Rate and MMMFs Rate and Between the Time D eposit Rate and MMMFs Rate to Changes in the Treasury Bill R ate* Cum ulative percentage points Cum ulative percentage points 1 . 0 ----------------------------------------------------------------------------------------------- 0.9 \ \ 0 . 9 ----- * _ o .i I___ I___ I___ I___ I___ I___ L 5 6 7 I Weeks 10 11 12 13 W eeks *C u m u la tiv e resp o n se s of the spreads between the six-m onth b ill rate (TB) and money m arket mutual funds ra te (M F ) and betw een the time deposit rate (TD) and MF to a change in the bill rate (pe rce n tag e points) *C um ulative responses of the spreads between the money market mutual funds rate (MF) and the money market deposit rate (MDA) and between MF and the Super NOW rate (SN) to a change in the bill rate (percentage points). FRBNY Quarterly Review/Autumn 1986 7 Responsiveness of deposit flow s In terms of very broad trends, Chart 1 shows how deposit flows have responded to changes in interest rate spreads. However, this section will give some reasons why precise estimates of how strongly deposit flows will respond to changes in these spreads are not possible now. Since MMDAs and Super NOWs were introduced in 1983, we have data for only about three years in which all four flexible-rate instruments were available— too short a period to estimate money demand equations with monthly or quarterly statistics, particularly since the equation for each type of deposit (MMDAs, Super NOWs, MMMFs, and time deposits) would in theory include four interest spreads and seasonal dummies, as well as some other variables as explanatory variables. (Table 1 shows which of the ten rate spreads would appear in each of the demand equations as well as the expected signs on the coefficients.) Even when more statistics become available, serious problems will arise in estimating the responsiveness of deposit flows to changes in the various rate spreads. These rate spreads, since they all respond to changes in market rates in a deregulated environment, tend to be correlated with one another, creating the problem of m u ltic o llin e a rity among the rate spreads used as explanatory variables. Table 2 shows the degree of correlation among the rate spreads that would be used in the equations shown in Table 1. The most striking result from Table 2 is the high degree of correlation among the spreads that would be included in the demand equation for Super NOWs. Since the rates on the other three types of deposits adjust more fully and quickly to changes in market rates than the Super NOW rate, a high degree of correlation exists among the rate spreads that would logically be included in a demand equation for Super NOWs. Indeed, the correlation (multicollinearity) is so high and so extensive that it appears very unlikely that reliable estimates of the responsiveness of Super NOWs to changes in these spreads could be obtained. Table 2 Correlation Between Rate Spreads* Monthly Levels and (Changes) Time Deposits TD-TB TD-MDA TD-MF TD-SN 1.00 (1.00) TD-TB TD-MDA 0.00 (0.02) TD-MF 0.25 (0.07) TD-SN 0.00 (0.00) 1.00 (1.00) 0.02 (0.06) 0.74 (0.85) 1.00 (1.00) 0.06 (0.01) 1.00 (1.00) MDA-TB MDA-TD MDA-MF MDA-SN (1.00) (0.43) (0.16) (0.01) 1.00 (1.00) 0.87 (0.62) 0.45 (0.29) 1.00 (1.00) 0.48 (0.36) 1.00 (1.00) MF-TB MF-TD MF-SN MF-MDA (1.00) (0.39) (0.07) (0.05) 1.00 (1.00) 0.15 (0.11) 0.24 (0.17) 1.00 (1.00) 0.78 (0.88) 1.00 (1.00) SN-TB SN-TD SN-MF SN-MDA (1.00) (0.54) (0.34) (0.20) 1.00 (1.00) 0.98 (0.82) 0.91 (0.68) 1.00 (1.00) 0.90 (0.71) 1.00 (1.00) Money Market Deposit Accounis MDA-TB MDA-TD MDA-MF MDA-SN 1.00 0.53 0.30 0.19 Money Market Mutual Funds Table 1 Rate Spreads fo r Demand Equations* Types of Deposits Rate Spreads SNOWs TD-TB TD-MDA TD-MF TD-SN MDA-TB MDA-MF MDA-SN MF-TB MF-SN SN-TB MMDAs x( -) MMMFs Time Deposits X (-) x(-) x( -) x(-) X( + ) X( + ) X (+ ) X (-) X (+ ) X (+ ) X( + ) 8 = = = = = rate rate rate rate rate on on on on on six-month time deposit six-month Treasury bill MMDA MMMFs Super NOWs FRBNY Q uarterly Review/Autumn 1986 1.00 0.54 0.06 0.10 X( + ) X( + ) X( + ) X( + ) *The x’s indicate which rate spreads should be included in the demand equation for each type of deposit. The + or - signs in parentheses indicate whether a widening in the spread would cause more rapid ( + ) or slower ( - ) growth in a given type of deposit. Where: TD TB MDA MF SN MF-TB MF-TD MF-SN MF-MDA Super NOWs SN-TB SN-TD SN-MF SN-MDA 1.00 0.87 0.80 0.78 ‘ The R2s that result when the interest rate speads that would appear in each of the demand equations are regressed on one another. Since four rate spreads would appear in each demand equation, there are six combinations of possible interest-rate-spread regressions for each type of deposit. The estimation period is from October 1983 to June 1986. TD TB MDA MF SN = = = = = six-month time deposit rate six-month Treasury bill rate MMDA rate MMMF rate Super NOW rate Source: Bank Rate Monitor. The multicollinearity problem is somewhat less severe for the other three categories of deposits, but probably still serious enough to raise questions about whether reliable demand equations could be estimated. In par ticular, Table 2 shows that for the time deposit demand equation there would be strong correlation between the (TD-SN) and the (TD-MDA) spreads. For the MMDA demand equation, there would be a strong correlation between the (MDA-MF) and the (MDA-TD) spreads, and somewhat weaker correlations between the (MDA-TD) and the (MDA-TB) spreads, the (MDA-SN) and the (MDA-TD) spreads, and the (MDA-SN) and the (MDAMF) spreads. And for the MMMFs demand equation, there would be a strong correlation between the (MFMDA) and the (M F-SN) spreads and a somewhat weaker correlation between the (MF-TD) and the (MFTB) spreads.9 This multicollinearity among the interest-rate-spread variables in all the equations is at least in part a by product of a deregulated financial structure. When ceiling rates were fixed in a regulated structure, the spreads between the interest rates on deposits tended not to change when market rates changed. Now all these spreads can change as market rates change, and particularly in the case of Super NOWs, the outcome is an environment where it will be extremely difficult to estimate demand equations using rate spreads. Never theless, general trends (as shown in Chart 1) strongly suggest that these rate spreads are sig n ifica n tly affecting M1. Though we cannot estim ate precisely how much deposit flows will respond to changes in interest rate spreads, we can infer from the responses of interest spreads to changes in market rates the direction that deposit flows are likely to move: (1) Time deposits should grow more rapidly as market rates increase. Since the rate on time deposits adjusts fully to the change in market rates, there should be no net loss of funds into m arket instruments. Likewise, there should be no net inflow or outflow of funds from MMMFs into time •In practice, some of the rate spreads could probably be eliminated in estimating demand equations. At the minimum, the opportunity costs with respect to market instruments and the other components of M2 should be included in each equation. Hence, for the MMDA and Super NOW equations, it probably would not be necessary to include both the spread with time deposits and MMMFs since both of these are components of M2 and fully adjust to changes in market rates with a similar pattern. That is, either rate spread could be used as a general proxy for the spreads that fully adjust to changes in market rates. In the case of Super NOWs, however, serious multicollinearity problems would still remain, whereas for MMDAs the problem would be considerably reduced. In some cases, taking the first differences of the spreads tends to reduce the degree of correlation somewhat, but in other cases it becomes greater. deposits because the rate on MMMFs over a twelve-week period also fully adjusts to changes in market rates. However, time deposits should grow more rapidly as market rates rise because of shifts of funds from MMDAs and Super NOWs into time deposits. The rates earned on MMDAs and Super NOWs do not fully adjust to changes in market rates, causing their spreads with time deposits to change as a result. (2) Super NOWs should grow more slowly as market rates increase. Funds should flow from Super NOWs not only into market instruments but also into time deposits, MMMFs, and MMDAs because the rates on these three other deposits adjust more fully and rapidly to changes in market rates than the Super NOW rate. (3) MMDAs will probably grow more slowly as market rates increase. MMDAs would lose funds to market instruments, time deposits, and MMMFs when market rates rise, but perhaps gain some funds from Super NOWs. (4) MMMFs should grow more rapidly as market rates increase. In the longer run, MMMFs should not lose any funds to market instrum ents or time deposits (the rate on MMMFs fully adjusts to changes in market rates) and should gain some funds from MMDAs and Super NOWs, since the rates on these types of deposits do not fully adjust to changes in market rates. Overall, as market rates increase, time deposits and MMMFs should grow more rapidly and NOW accounts and MMDAs should grow more slowly. Chart 7 shows that these patterns have generally held over the last three years. Time deposits showed their most rapid growth relative to trend at about the time interest rates peaked in 1984 and have slowed since then. In contrast, NOW accounts and MMDAs showed their weakest growth at about the time interest rates peaked and have accelerated as interest rates have fallen. By and large, MMMFs have displayed a pattern sim ilar to tim e deposits, but the chart suggests that the main flows as market rates change are between time deposits and NOW accounts or MMDAs. Conclusions The experience of the last few years offers some gen eral insights into how monetary aggregates are likely to respond to future changes in interest rates (provided that banks continue to behave in the same way) and raises some interesting questions. The demand for M1 has retained a significant, and probably larger, interest rate elasticity even though checking accounts for con sumers have been deregulated. The traditional interest- FRBNY Quarterly Review/Autumn 1986 9 rate channel is still open whereby movem ents in market rates cause changes in the desired level of transactions balance s by a ffe ctin g the spread betw een m arket rates and the rate paid on M1. In a d d ition , the d e re g u late d environm ent has provided a new channel through which changes in m arket rates can narrow or widen the spread betw een the tim e d e p o s it rate, as well as the MMDA and MMMF rates, and the rate on transactions balances. S ince m any of th e se flo w s are w ithin M2, M 2 ’s in te re st Chart 7 Response of D eposit Flows to M ovem ents in In te re s t Rates Percent deviation 0.6 Six-m onth bill rate from mean lagged one quarter . 0 .2 - 0- 0.2 I - 0 .4 0.10 -0.05 0.10 I I I I NOW a cco u n ts from I ---------- i X I I rend . I I U 1 Time d e p o sits from trend 1 0.05 0- -0 .1 0 L 0 .10- I 1 1 1 1 1 . 1...I, 1 1 MMDAs from trend 0.05 - 0 -0.05 - 0 .1 0 L 0 .10- MMMF s from trend 0.05 o-O.O5 IL 1 1983 i j V 1984 m . 1985 : i 1986 1 re s p o n s iv e n e s s has not b een in c re a s e d . In d e e d , it p robably has been co n s id e ra b ly reduced com pared to a re g u la te d e n v iro n m e n t, b e c a u s e to an in c re a s in g degree the rates paid on its co m p o n e n ts respond at least p a rtia lly (and tim e d e p o sits and M M M Fs fu lly) to changes in m arket rates. W hile the exp e rie n ce of the last three years can p ro vide som e in sig h ts, in te re st rates have not m oved over a s u ffic ie n tly large range in both d ire c tio n s fo r th e re to be much co n fid e n ce th a t the p ro ce ss by w hich these rate spreads are affe cte d is w ell unde rsto o d . Indeed, it is likely th a t banks have been learning how to price co nsum er d e p o sits in a d e re g u la te d e n v iro n m e n t over these last few years, and th a t as th e y gain m ore e xp e rience th e y may behave in a d iffe re n t way. In the sam e way, consum ers w ill becom e m ore fa m ilia r o ve r tim e with deregulated deposits and could respond differently in the future. And both banks and consum ers m ight not respond as stro n g ly if rates w ere g ra d u a lly in cre asin g rather than fa llin g by a large am ount as th e y did over the past few years in response to the sharp fa ll in the rate of inflation. In other words, their response might not be sym m e trica l to rising and fa llin g m arket rates, or to g ra d u a l ra th e r th a n la rg e c h a n g e s in m a rk e t ra te s . Moreover, we have no exp e rie n ce w ith how banks and consum ers m ight behave in a situ a tio n w here the yield cu rve fo r m arket in stru m e n ts becam e inve rte d . And even in a stable interest rate environm ent, banks may find it profitable to reprice these various accounts, thus a ffe ctin g M1 as w ell. For exam ple, if banks begin to believe that a large volum e of the funds held in NOW accounts are re la tiv e ly in a ctive sa vin g s ba la nce s that have been sh ifte d into NOW acco u n ts as in te re s t rates fell, they may design c o m b in a tio n s of a cco u n ts w ith tra n s fe r fe a tu re s th a t w ould induce co n su m e rs to hold th e s e in a c tiv e s a v in g s b a la n c e s in n o n tra n s a c tio n s accounts in ord e r to avoid re se rve re q u ire m e n ts. Then M1 could a ppear u n u su a lly w eak re la tive to GNP fo r a period of time, instead of appearing unusually strong as it has in recent years w hen sa vin g s ba la nce s w ere a d d e d to M 1. In d e e d , if b a n k s s h o u ld s tro n g ly e n courage co n su m e rs to keep only fric tio n a l tra n s a c tio n s balances in M1, M 1’s in te re s t e la s tic ity could begin to a ppear very low com pared w ith the e xp e rie n ce of the past few years. W hile we do understand a few features of this new environm ent, it continues to be im portant to monitor changes in the banking system that m ight affect the b e h a vio r of the m onetary agg re g a te s. T here are many reasons to e xp e ct th a t the recent past m ight not be a good guide to the future. John W e n ninger 10 FRBNY Q uarterly Review/Autumn 1986 Wage Rigidity in West Germany A Comparison With the U.S. Experience Even though inflation seems to be well under control in West Germany, many policymakers and economists continue to be pessim istic about the ability of the economy to sustain a substantial domestic demand expansion. To a large extent, such pessimism is based on the view that the German economy is afflicted by severe labor market rigidities which leave virtually no scope for expansionary policies. Presumably the con cern is that any demand expansion, even at today’s record high unemployment levels, would simply rekindle in fla tio n w ith o u t s ig n ific a n t gains in o u tp u t and employment.1 This article provides some fresh evidence on labor market rigidities in West Germany, focusing on one of the most important aspects of these rigidities, namely the behavior of wages. Specifically, using both aggre gate and disaggregate (industry level) data, this article examines the flexibility of wages in West Germany. Although other sources of rigidity may be potentially important, the relatively narrow approach of this article is appropriate, given that the behavior of wages is I wish to thank M. A. Akhtar, A. Steven Englander, Ethan Harris, and Peter Rappoport for helpful discussion, and Elizabeth A. Hall for her excellent research assistance. ’ Several arguments are believed to be relevant, the most common of which rests on the view that wages respond asymmetrically to conditions of excess demand and supply in the labor market. If wage structures are rigid, then wages are unresponsive to unemployment and unlikely to fall in the appropriate market-clearing way. If unemployment is classical (resulting from already too high wage levels) and wages are rigid, then a demand expansion could perversely result in higher wages (and prices) with little or no gain to output and employment. Wage flexibility prevents this scenario from occurring because unemployment exerts continued downward pressure on wages at the same time that the economy is expanding. widely believed to be the driving force for most other labor market rigidities.2 Because it is difficult to gauge precisely what flexibility implies for wage responsiveness, we evaluate wage behavior in West Germany, and wherever possible contrast it with wage performance in the United States. We begin with the assumption, inspired by the literature in this area, that real wages in the United States through the 1970s and 1980s have been flexible and that the pattern of U.S. real wage response has aided output and employment expansion.3 The analysis in this article extends the work of pre vious studies in its consideration of industry wage behavior and in the distinct way it treats pre- and postOPEC aggregate wage behavior. The key finding of our analysis is that wages in West Germany, at both the aggregate and industry level, have been flexible in recent years. As a consequence, the pace of real wage growth in West Germany has moderated, and unit labor costs have grown at* about two-thirds the U.S. rate since 1980. The industry patterns offer new and additional evi dence of wage flexibility. Industry wages were highly responsive to industry-specific performance in West German manufacturing, particularly in the short run. Over the long run the data indicate greater flexibility in 2This article does not evaluate, to any significant extent, more microeconomic aspects of labor market rigidities, such as minimum wage laws, unemployment insurance rules, labor mobility, and the costs of hiring and dismissing workers. *There are many studies which characterize real wages in the U.S. as flexible. The most comprehensive study, and reference to other work in this area, may be found in M. Bruno and J. Sachs, The Economics of Worldwide Stagflation (Cambridge: Harvard University Press, 1985). FRBNY Quarterly Review/Autumn 1986 11 the United States, although to a degree not statistically d istin g u is h a b le from W est G erm any. Taken as a w hole, the d ata o ffe r co n vin cin g e vid e nce th a t in d u stry w ages w ere fle x ib le in W est G erm any. T he next se ctio n o f th is a rtic le co m pares a g g regate wage, productivity, and cost trends in the United States and W est G erm any, and e va lu a te s the la b o r dem and and su p p ly p re ssu re s in flu e n c in g e q u ilib riu m w ages in each country. T he fo llo w in g s e c tio n s e xp lo re a g g regate and in d u s try w a ge fle x ib ility in W est G erm any, draw ing com parisons w ith the U.S. experience. A brief sum m ary of the m ain fin d in g s and th e ir im p lic a tio n s fo r m acroecon om ic p o licy are p re se n te d in the fin a l se ctio n . The aggregate data: labor supply and demand in wage responsiveness W est G erm an u n e m p lo ym e n t, u n like u n e m p lo ym e n t in the U nited S tates, has in cre a se d since 1982 (Table 1, top panel). M ost analysts attribute this divergence to the b e h a vio r o f a g g re g a te w a g e s and co n clu de th a t w ages on average in W est G erm any have been more inflexible d o w nw a rd , p re ve n tin g la b o r m a rke ts from cle a rin g and re su ltin g in re la tiv e ly high u n em ploym ent. S ince we e va lu ate w age re sp o n sive n e ss in both the U nited S ta te s and W est G erm any, it is useful to co m pare la b o r m arket b e h a v io r in the tw o co u n trie s. The firs t q u e stio n is w h e th e r th e re is a n ything in the trend of la b o r s u p p ly or de m a n d gro w th th a t can e xplain d iffe re n c e s in o u tp u t and e m p lo ym e n t grow th in W est G erm any, in d e p e n d e n t of w age fle xib ility. Table 1 Unem ploym ent S ta tistics The Civilian Unemployment Rate (annual averages) 1965-85 United States . . West Germany . 6.2 3.8 1965-73 1974-79 1980-85 4.5 1.1 6.8 4.1 8.1 7.5 1986* 7.0 9.0 Share of Unemployed Prime Age Male Workersf 4ln the United Kingdom and France, for example, female participation rates in 1984 were 59 and 55 percent, respectively. (calculated as a percentage of total unemployed) 1965-85 United States . . West Germany . 23.6 26.6 1965-73 1974-79 1980-85 20.5 22.3 22.3 27.6 28.7 30.7 ‘ Averages include the first ten months of 1986. fM ale workers aged 25-54. ^Includes first ten months of 1986. §1985 figure. 12 FRBNY Quarterly Review/Autumn 1986 On the supply side, labor market demographic trends do not explain the pattern of West German unemployment. W hile an in flu x of fem ale, p a rt-tim e , you th , or fo re ig n workers could conceivably lead to greater unemployment for any given level of aggregate demand, the evidence sug gests that these changes have not been the leading cause of unemployment in West Germany. C onsider for exam ple women in the labor force. W hile fe m a le labor fo rce p a rtic ip a tio n grew ra p id ly o ve r the 1970s in the U nited S tates (in cre a sin g from 49 pe rcent in 1970 to 65 p e rce n t in 1984), in W est G e rm a n y over the same period it increased just one percentage point. With only 49 percent of women in the labor force, West G erm any has one of the lo w e st fe m a le p a rtic ip a tio n rates in the m ajor E uropean O E C D c o u n trie s .4 In the sam e w ay th a t the e n try of w om en m ay a ffe ct the shape or position of the labor supply curve, changes in the mix of p a rt-tim e and fo re ig n w o rke rs m ay a lte r a g g regate supply. T h is su g g e sts th a t the e m p lo ym ent data should be adjusted for these workers, to see if the e m p lo y m e n t re co rd o f e ith e r c o u n try is q u a lita tiv e ly altered. W hile this adjustm ent results in a stronger trend decline in West German em ploym ent, it has only modest e ffe cts on the pattern of e m p lo ym e n t grow th in the U nited S tates (C h a rt 1). As a tool e ith e r fo r sm o o th ing em p lo ym e n t or fo r m in im izin g e m p lo ye r co sts, W est G erm an firm s have e m ployed c o n s id e ra b ly m ore parttime workers than U.S. firm s.5 Excluding teenagers and older workers from the em ploym ent analysis, so that we consider the unem ploym ent patterns of prime age male w o rke rs only, leads to the sam e c o n c lu s io n — the core of the W est G erm an u n em ployed are p e rm a n en t labor force m em bers (Table 1, bottom panel). The broad d e m o g rap h ic d ata do not su g g e st m ajor d iffe re n c e s in a g g re g a te la b o r s u p p ly b e h a v io r and th e re fo re p ro b a b ly do n o t e x p la in re la tiv e ly h ig h e r un e m ploym ent in W est G erm any than in the U nited S tates. However, o th e r su p p ly -re la te d fa c to rs are re le vant. One o b vio us source of d iffe re n c e could be the u n e m p lo y m e n t in s u ra n c e sy s te m . In W est G erm any, incom e re p la ce m e n t ra tio s from u n e m p lo ym e n t in s u r ance are, on average, about two tim es greater, and the period of e n title m e n t is about three tim e s lo n g e r6 than 1986 30.1* 32.2§ 5Part-time workers typically are paid less than full-time workers and have fewer fringe benefit provisions. In addition, certain payroll tax exemptions are associated with part-time workers. 6The replacement ratio in West Germany for a single worker with average earnings is approximately 65 percent of previous earnings, and benefits last at this rate for three years By contrast, the same worker in the United States receives an average first-year replacement ratio equal to 35 percent of his base earnings, and benefits are exhausted, on average in the United States, after 52 weeks. in the U nited S tates. As a conse q u e n ce , costs to the u n em ployed w o rke r are lo w e r in W est G erm any than in the U nited S tates, and the in centive to rem ain unem p lo y e d is th e re fo re m u ch g re a te r. W h ile th e W est G e rm an u n e m p lo y m e n t sy s te m m ig h t a c c o u n t fo r a h ig her o ve ra ll level of u n e m p lo ym e n t at any p oint in tim e, it can not exp la in the trend through tim e. W est G erm an b e n e fits have tra d itio n a lly been g enerous, and the cu rre n t law has been in effect, ro ughly w ith o u t c h a n g e , s in ce 1969. T h e re fo re , un le ss a s ig n ific a n t change has o ccu rre d in the a ttitu d e of W est G erm an workers tow ard work, higher unem ploym ent must reflect e ith e r lo w e r e x p e c te d b e n e fits of s e e k in g w o rk or greater inability of the unem ployed to find suitable jobs. S ta tistics on u n e m p lo ym e n t du ra tio n su p p o rt the view th a t there is ch ro n ic excess supply in W est G erm an labor m arkets. W hile only 8.5 perce n t of W est G erm an workers were unemployed for one year or more in 1973, by 1985, 31 percent w ere u nem ployed fo r lo n g er than one year. By co n tra st, only a bout 15 p e rce n t of unem ployed workers in the United States were idle for longer than six m onths in 1985. F inally the re la tio n sh ip betw een u n e m p lo ym e n t rates and jo b v a c a n c ie s in W e st G e rm a n y (th e s o -c a lle d Beveridge Curve, Chart 2), suggests that the historically high recent rates of unem ploym ent in West Germ any are not su p p ly-sid e induced. S h ifts of this cu rv e are a s so c ia te d w ith s tru c tu ra l c h a n g e s and s tru c tu ra l u n e m ploym ent and are taken to re fle ct a m ism atch of jobs and w o rke r skills; m ovem ents along the cu rve re flect Chart 1 E m ploym ent G rowth M illions of persons M illions of persons 2 8 .0 ------------------------West G erm any 1 1 0 ------------------------- United States 27.5 27.0 55L..I 1965 I I 1.1. 1...1...1..i .. I 67 69 71 73 1 1 1 1 I I 75 77 79 1....1 „J 81 83 85 1965 67 69 71 73 75 77 79 81 83 85 * Data rep re se n t total employment (as m easured by the C urrent Population S urvey) minus p a rt-tim e employment. + Data rep re se n t total employment minus employm ent of foreign and sh o rt-te rm workers. Sources: U.S. Departm ent of Labor, Current Population Survey; Deutsche Bundesbank, S tatistical Supplem ent to the Monthly Report, and staff estim ates. FRBNY Quarterly Review/Autumn 1986 13 de m a n d -in d u ce d chan g e s. W hile the cu rv e a ppears to have sh ifte d in the U nited S tates, it is m ore sta b le in W est G erm any. However, the la b o r m arket situ a tio n in W e st G e rm a n y, as re fle c te d by th e p o s itio n on th e B eve ridge C u rve in the m ost recent years, has su b s ta n tia lly w o rsen ed. In 1962, fo r exam ple, there w ere m ore than tw o va ca n t jo b s fo r e ve ry u nem ployed W est G erm an w orker, but by 1985, fo r every tw o va ca n t jobs th ere w ere roug hly 50 un e m p lo ye d w o rk e rs .7 Because supply-side developm ents do not explain why the e m p lo ym e n t situ a tio n is re la tiv e ly w orse in W est G erm any than in the U nited S tates, we next c o n sid e r the degree to which wages, productivity, and costs may have adversely influenced West German labor demand. Since the mid-1960s real wages in West Germany have grown at nearly four tim es the U.S. rate. More moderate 7Vacancy data are unreliable and need to be interpreted with care. In the U.S., no vacancy series exists as a time series; we have used the Medoff technique of adjusting the help-wanted advertising data as a proxy for vacancy rates. In West Germany, vacancies are registered through the German central agency. It is likely that more vacant jobs will go unreported when unemployment is high, since available openings are filled immediately and with ease by employers. For a discussion of the stability of the U.S. Beveridge Curve, see J. Medoff, “ U.S. Labor Markets: Imbalance, Wage Growth, and Productivity in the 1970s,” Brookings Papers on Economic Activity, Vol. 1 (1973), pp. 87-128. nominal w age settlem ents in West Germ any since 1980 have caused the pace of change in real wages to be more equal in the two countries (Table 2), although the latest available data indicate some change in early 1986. Despite the recent slowdown of real wage growth in West German m anufacturing, rapid acceleration in the mid1960s and early 1970s led to real w ages that are now, on average, tw ice their 1965 level. By contrast, U.S. m anufacturing w orkers now earn w ages only about 20 percent higher than th e ir 1965 level.0 W hile the a cce le ra tio n of w ages th ro u g h the 1970s suggests real w age levels th a t are both re la tiv e ly and a b so lute ly high in W est G erm any, fo cu sin g e x c lu s ive ly on the b e h a vio r of the w age se rie s m asks m ore im p o r tant labor market trends. The growth of unit labor costs represents the excess of wage over productivity growth and p ro b a b ly is a b e tte r m easure of the p re ssu re s on 8Another potential source of difference between real wage patterns in the two countries results from the importance of minimum wages in influencing the pattern of real wage movement. In West Germany, the union-legislated minimum wage sets an effective floor on real wages at a relatively high level which is binding on the employer. In the United States, by contrast, the legislated minimum wage has been allowed to erode considerably in real terms and nearly all full time workers in manufacturing currently receive wages well in excess of this level. As a consequence, the U.S. minimum wage is, in practice, not binding on the employer. Chart 2 B everidge Curve Vacancy ra te * Vacancy rate 3 .0 ---------------- 2.0 W est Germ any United States 2.5 74 79 0 8 '------------------------ 1-------------------------1 ------------------------ 1 ------------------------- 1 ----------------------- 1 ------------------------ 1 -------------------------1 ' 3 4 5 6 7 8 Unemployment rate 9 10 J _____ I_____I_____I---3 4 5 6 7 Unemployment rate *T h e vacancy rate was constructed by taking the average of the monthly help-wanted index fig u re s for each year (1967 = 100) and dividing by the number of em ployees on non-agricultural payrolls. Sources: The Conference Board; U.S. Department of Labor, Current Population Survey and Establishm ent Survey; Deutsche Bundesbank, S ta tistica l Supplement to the Monthly Report, Series 4. 14 FRBNY Quarterly Review/Autumn 1986 prices stem m ing from labor m arket conditions. From an e m p lo y e r’s p e rsp e ctive , fa s te r real w age gro w th does not m atte r ne a rly as m uch if it is o ffse t by la b o r p ro d u c tiv ity advan ce. T his a p p e a rs to be the lo n g er run trend supported by the growth pattern of unit labor costs in W est G erm a ny (Table 3). W h ile w ages a c ce le ra te d ra p id ly in W est G e rm a n y in the late 1960s and e arly 1970s, and un it lab o r co sts grew so m e w h a t fa s te r than in the U nited S ta tes, by the m id-1970s m ore m odest w age ga ins in W est G erm a n y had reversed the e a rlie r tren d. As a result, the c u m u la tiv e grow th of un it labor costs since 1965 has been s lo w e r in W est G erm any than in the U nited S tates. F inally the e vid e nce on w age grow th in the n on m a n u fa ctu rin g s e c to r is also in c o n s is te n t w ith w ages being the key determ inant of em ploym ent trends in West Germany. A lthough nonm anufacturing wage growth was som ew hat fa s te r in W est G erm any than in the U nited States until the late 1970s, it has been more modest in W est G erm an y th an in the U nited S tates since 1980 (Table 4). M oreover, the pace of n o n m a n u fa ctu rin g re l ative to m a n u fa ctu rin g w age grow th has been c o n s is t e n tly slo w e r in W est G e rm a n y (Table 4). D espite these trends, nonm anufacturing em ploym ent in West Germany has grow n at a fa r w e a k e r pace than in the U nited S tates. In sum , the da ta do not reveal ra d ic a lly d iffe re n t changes in labor m arket conditions in the United States and W est G erm an y since the 1970s. We next exam ine w h e th e r d iffe re n ce s in w age fle x ib ility m ay lie behind W e s t G e r m a n y ’s p o o re r o u tp u t p e rfo rm a n c e a nd e m ploym en t gro w th record. The fle x ib ility of aggregate wages In th is se ctio n , we apply som e standard m easures of w age fle x ib ility to the U.S. and W est G erm an d a ta to gauge w age re sp o n sive n e ss in each country. D espite the c o n tin u in g d ebate a bout the re la tio n s h ip betw een w a g e fle x ib ility and e m p lo y m e n t p e rfo rm a n c e , m o st a n a lysts assum e th a t a m ore fle x ib le w age system w ill allow faster and more com plete adjustm ent to econom ic shocks and will therefore perm it faster econom ic growth and g re a te r le ve ls of em ploym ent. We c o n sid e r firs t the sim p le s t and m ost s tra ig h t forw ard m easure of w age fle xib ility. In a fle x ib le real w age system , w ages a d ju st fre e ly to sh ifts in labor supply or dem and, w ith the re su lt th a t m a rke ts e q u ili brate quickly. T his im plies th a t d u ring p eriods of la bo r m arket flux wages should be more variable in a flexible than in an inflexible system. Given this description, one test of wage flexibility is to calculate the variance of real w age grow th around tre n d . A fle x ib le system should produce high va ria tio n generally, w ith in cre a sin g v a ri ation during p e riods of u n sta b le ag g re g a te dem and or su p p ly behavior. Two m e a s u re s o f re a l w a g e s a re u s e d : (a ) th e re a l c o n s u m p tio n w a g e , m e a s u re d as th e ra tio of nominal wages to the personal consumption deflator and (b) the real product wage, measured as the ratio of nominal w ages to the p ro d u ce r price index. These m easures show that the variability of real consum ption and product w age changes in m a n u fa ctu rin g has been g e n e ra lly greater in West Germ any than in the United States; the cle a re st d iffe re n ce betw een the tw o co u n trie s re su lts from the relative long-run stability of real wage changes Table 2 Table 3 Wages in M anufacturing Unit Labor Costs in M anufacturing Nominal Wage Growth in Manufacturing Growth in Unit Labor Costs (average annual rates of change in average hourly earnings) (average annual rates of change in unit labor costs in manufacturing) 1965-85 United States . . West Germany . 6.5 8.5 1965-73 1974-79 1980-85 5.5 10.3 8.7 8.9 5.9 5.4 1986* 2.1 4.5 1965-85 United States . West Germany . 4.7 4.2 1965-73 1974-79 1980-85 3.5 5.3 8.2 4.2 3.2 2.4 Real Wage Growth in Manufacturing! Cumulative Unit Labor Cost Growth (average annual rates of change) (1965 = 100) 1965-85 United States . . West Germany . 0.7 4.3 1965-73 1974-79 1980-85 1.2 6.3 0.6 4.1 0.1 1.6 1986* 0.2 5.3 *For the U.S. the figure represents 1986-111/1985-III; for West Germany the figure represents 1986-11/1985-11. fR eal wage is calculated by deflating the average hourly earnings index in manufacturing by the implicit PCE deflator. United States . West Germany 1986 0.1 3.7 1972 1975 1979 1986 129.5 136.0 164.7 175.1 210.0 189.3 261.0+ 224.4* ‘ For the U.S. the figure represents 1986-111/1985-III; for West Germany the figure represents 1986-11/1985-11. fThree-quarter average. i:Two-quarter average. FRBNY Quarterly Review/Autumn 1986 15 Table 4 Wage G rowth in N on-M anufacturing Industries Wage Growth Relative Wage Levels* 1971-73 1974-79 1980-85 1985 1971-73 1974-79 1980-85 1985 United Statest Wholesale t r a d e ............. Retail t r a d e ...................... B a n k in g ............................. Insurance ......................... ............ ............. ............ ............ 5.8 5.3 4.8 5.0 7.8 6.6 7.2 6.5 6.5 4.9 8.1 6.6 4.6 1.5 5.7 3.6 0.9 1.4 1.2 0.8 0.9 1.5 1.2 0.9 1.0 1.7 1.2 0.9 1.0 1.8 1.2 0.9 West Germany^ Wholesale t r a d e ............ Retail t r a d e ...................... B a n k in g ............................ Insurance ......................... ............. ............. ............. ............. 10.8 10.5 10.4 10.6 7.4 7.2 7.2 8.4 4.2 3.5 4.7 5.1 3.2 2.7 4.5 5.2 1.3 1.4 1.1 1.2 1.3 1.4 1.2 1.1 1.3 1.5 1.2 1.1 1.3 1.5 1.2 1.1 •Relative wages calculated by dividing the wage level in manufacturing by the wage level in each non-manufacturing industry. fWage and salary workers. ^Data represent a weighted average of male and female earnings. in the U nited S tates (Table 5). T his is tru e generally, both b efore and a fte r the O P E C oil shocks. W hile the v a ria b ility o f w age changes in the United States and W est G erm any is relatively unaffected by the choice of deflator, the c o n su m p tio n and product w age m e a s u re s o ffe r in d e p e n d e n t in fo rm a tio n a b o u t th e in s titu tio n a l b e h a vio r of w ages. Product w ages more a c c u ra te ly m easure e m p lo ye r co sts; high v a ria b ility in this series may indicate greater flexibility on the part of e m p lo y e rs in s e ttin g w a g e s . T h e re a l w a g e s e rie s d e fla te s n om inal w age grow th by the personal co n sum ption deflator in each country and therefore reflects the p u rch a sin g po w e r gains of nom inal w age s e ttle m ents. If, as m any econ o m ic m odels suggest, w orkers d e s ire a c o n s ta n t s tre a m o f re a l e a rn in g s o v e r th e cou rse o f th e ir w orking liv e s ,9 then va riatio n in this se rie s m ay be a signal of w eakness on the part of w o rkers or unions in se cu rin g real wage gains. In any case, the substantial degree of variability in the pattern of W est G erm an w age grow th is co n siste n t w ith there being som e fle x ib ility in w age setting. V ariation in real w age grow th does not n e cessarily im p ly th a t w a g e s w e re fle x ib le in any e c o n o m ic a lly m e a n in g fu l w ay. E v id e n c e o f w a g e re s p o n s iv e n e s s requires a system atic link between movements in wages and key econom ic variables. Real wage variability alone does not e xp la in the source of w age m ovem ents and th e re fo re ca n n ot pro vid e evid e nce of any such link. im p lic it contract models of the labor market are based on this assumption. While these models have been criticized on many grounds, including their failure to make an adequate distinction between real and nominal wages, most subsequent work has assumed that it is constancy in real earnings that workers seek in their bargaining demands. 16 FRBNY Quarterly Review/Autumn 1986 However, our e ffo rts at eva lu ating the source of w age m o v e m e n ts re v e a l a u n ifo rm in c re a s e in re a l w a g e fle x ib ility in W est G erm an m a n u fa c tu rin g in re c e n t ye ars. To d em onstrate th is fle xib ility, we estim ate equ a tio n s lin k in g g ro w th ra te s o f a v e ra g e h o u rly e a rn in g s in m anufacturing to inflation and unem ploym ent rates. Real wage fle x ib ility is m easured by com paring the re sp o n siveness of w ages to expected price in fla tio n and the unem ploym ent rate; in a fle x ib le real w age stru ctu re , nominal wages react weakly and with a lag to expected p ric e m o v e m e n ts b u t s tro n g ly to m o v e m e n ts in th e u n e m p lo y m e n t ra te . T h is m e a s u re c o m b in e s th e standard view tha t real w age fle x ib ility re su lts from inertia in the response of nom inal w ages to p rice s and the view th a t w ages should be responsive to excess d e m a n d o r s u p p ly in th e la b o r m a rk e t. In a d d itio n , combining the two flexibility criteria in a single equation y ie ld s a m e a s u re o f th e d e g re e o f a c c o m m o d a tio n necessary to keep the in fla tio n rate c o n s ta n t.10 According to this flexibility measure, real wages have been increasingly responsive in W est Germ any in recent years (Box 1). The main reason is the lack of any strong response of w age grow th to p rices in W est G erm any over the recent period, m ost like ly reflectin g the d e te rioration of real w age grow th. In general, the larg e st difference between the pattern of wage response in the 10The standard framework for evaluating real wage responsiveness, based on the concept of nominal wage inertia, is discussed in J. Sachs, “ Wages, Profits, and Macroeconomic Adjustment. A Comparative Study," Brookings Papers on Economic Activity, Vol. 2 (1979), pp. 269-319. An alternative approach, which stresses the role of unemployment, is discussed in D. Grubb, R. Jackman, and R. Layard, "Wage Rigidity and Unemployment in the OECD Countries,” European Economic Review, Vol. 2 (1983), pp. 11-40. two countries is in the reaction of nominal wages to expected price movements, not in the overall respon siveness of real wage changes to the unemployment rate. In fact, the response of wages to unemployment in the two countries is similar. Additional tests of aggregate wage flexibility, based on commonly used variants of the general specification reported here, broadly confirmed these results.11 Our analysis shows that wages in West Germany, while more rigid than in the United States in the early 1970s, were quite flexible by the late 1970s. Previous studies may have failed to isolate this tendency because they did not distinguish the pattern of wages in West Germany in the most recent years.12 The fle x ib ility o f in d u s try wages There is no stra igh tfo rw ard relationship between aggregate and industry wage flexibility. Aggregate wage flexibility does not necessarily imply that industry wages are free to vary; aggregate wages may be flexible at the same time that institutional restraints prevent industry wages from moving to equilibrate labor markets. Box 1: F le xib ility o f Aggregate Wages Data from 1966-1 to 1985-IV were used to analyze wage flexibility in the United States and West Germany. The full period data were analyzed over subperiods chosen to capture the recent changes in nominal wage patterns. Alternative specifications were estimated over each subperiod; the results presented here were chosen for general fit. The equations are specified in four-quarter growth rates of both the dependent and independent variables. The dependent variable in all equations is the change in the natural log (In) of average hourly earnings in manufacturing. The price expectations variable in all equations is estimated as a fitted lag on past price changes (see A.S. Englander and C. Los, Federal Reserve Bank of New York Research Paper, No. 8305, August 1983), and the unemployment rate is the rate for the economy as a whole. All equations have been cor rected for fourth-order serial correlation. Standard errors appear in parentheses below the estimated coefficients. Table A "Includ in g both a productivity growth variable and a dummy variable to serve as a proxy for shifts in the structural Phillips Curve relation did not significantly affect the coefficient estimates on inflation and unemployment in either country. While the productivity term did figure significantly in the West German equations and did raise the explanatory power of the equation, it did not affect the size or significance of either the price expectations or the unemployment variable. 12More recent work has concluded that greater wage flexibility characterizes the West German economy today. See F. Klau and A. Mittelstadt, “ Labour Market Flexibility,” OECD/ESD Working Papers, No. 24 (July 1985). Table 5 Variation* in M anufacturing Wage Growth Real Real Product Wagesf Consumption Wagesf United West West United States Germany Germany States 1965-85 . 2.8 3.5 1.2 3.4 1965-78 1979-85 . . . . 3.2 2.1 2.8 2.4 1.1 0.8 3.3 1.7 1965-73 1974-85 . . 2.6 3.0 2.6 2.6 1.3 1.0 3.5 2.5 Mean wage growth (1965-85) 0.9 4.6 0.7 4.3 'Calculated as the standard deviation in the arithmetic annual growth rate of average hourly earnings in manufacture. -(-Deflated by producer price index. ^Deflated by the personal consumption deflator in manufacture. Dependent Variable: Change in AHE (In AHEt- In AHEt^) United States West Germany (2) (3) (4) (5) (6) (1) 1966-85 1974-85 1979-85 1966-85 1974-85 1979-85 .813 (.158) .949 (.232) .954 (.031) 1.18 (.335) .761 (.193) .392 (.215) InU, . . . . -.0 3 4 (.011) -.0 5 0 (.016) -.0 5 2 (.005) -.0 3 4 (.005) -.0 4 6 (.006) -.0 4 7 (.005) Flexibility coefficient 23.91 18.98 18.35 34.71 16.54 8.34 .387 .001 1.97 .980 .0002 2.23 .351 .017 1.73 .700 .005 1.89 .760 .003 2.01 •p,e R2 . . . . SSE DW . . . . .255 .003 1.95 The wage flexibility coefficient listed in Table A above is calculated by taking the ratio of the long-run elasticity of wages with respect to past price inflation and the elasticity of wages with respect to the unemployment rate. This statistic may be interpreted for any given price change as the change in the In of unemployment nec essary to keep the nominal wage constant (i.e., to ensure a fall in the real wage). It is sim ilar to the measure adopted in D. Coe, “ Nominal Wages, the NAIRU and Wage Flexibility,” OECD Economic Studies, No. 5 (Autumn 1986). FRBNY Quarterly Review/Autumn 1986 17 The eco nom ic sho cks of the 1970s affe cte d sp e cific in d u strie s d iffe re n tly. The ones th a t relied to a large e xtent on oil as an inpu t to p ro d u ctio n w ere m ade p a r tic u la rly vu ln e ra b le . In a w ell fu n c tio n in g c o m p e titive econom y, the re spo nse of in d u s try w ages to sh o rt-te rm d is tu rb a n c e s sh o u ld re fle c t s p e c ific p e rfo rm a n c e — in industries particularly affected by the OPEC oil shocks, w ages should have fa lle n . There are good reasons to suppose that the industrial w age stru c tu re of W est G e rm a n y m ight be rigid. W hile collective bargaining is highly decentralized in the United S ta te s (w ith th o u s a n d s of in d iv id u a l e s ta b lis h m e n ts setting wages), it is highly centralized in West Germany (w ith n e a rly all ba rg a in in g ta kin g place at the in d u stry and reg io n a l le ve ls). W h ile unions are a m in o rity p re s ence in the U.S. w o rkp la c e (w ith less than 25 perce n t of U.S. w o rke rs covered by a union co n tra ct), th e y are a p ow erful m a jo rity pre se n ce in W est G erm any (w here m ore than 90 pe rce n t of w o rke rs are em ployed in s e c tors covered by collective bargaining agreem ents).13 The co m b in a tio n of th e se tw o fa cts im plies, all else the sam e, th a t the stru c tu re of w ages am ong in d u strie s is m ore like ly to be rigid in W est G erm any than in the U nited S tates. To te st fo r w age rig id ity am ong W est G erm an in d u s tries, we have assem bled m anufacturing data at the twod ig it level fo r the U nited S tates and W est G erm any on w ages, p rices, productivity, and e m p lo y m e n t.14 These d a ta allow fo r a new a n a lysis of w age fle x ib ility w ithin the econom y that yields independent and more detailed in fo rm a tio n abou t the b e h a vio r of labor m arkets than can be le arned from the a g g re g a te data. To the exte n t that industry wages are flexible, structural labor market problem s are more likely to be short-lived, since wages help to allocate labor appropriately among industries in the long run. Ind ustrial w age fle x ib ility is defined as the re sp o n s iv e n e s s o f in d u s try w a g e s to in d u s try -s p e c ific p e r form ance. A rigid in d u s tria l w age stru ctu re has fixed 13For the statistics on union coverage and membership, see the chapter on West German collective bargaining in OECD, Collective Bargaining and Government Policies in Ten OECD Countries (Paris, 1979), In West Germany there is a statistically large difference between the number of workers who are union members (which for years has been slightly greater than one-third of all workers) and those who are covered by union contracts. By contrast, the difference between union coverage and union membership is modest in the United States. 14Data for West Germany were kindly provided by the West German Bundesbank for 29 manufacturing industries over the period 197082. The U.S. data, except where otherwise noted, cover 22 manufacturing industries at the two-digit level and are from the National Income and Products Accounts Series We focus on industry wage patterns over the period 1970-82 in this section because more recent data for West German industries are not available. 18 FRBNY Quarterly Review/Autumn 1986 re la tive w ages, so th a t e xistin g w age d iffe re n c e s are p re se rve d across in d u strie s th rough tim e. In the m ost rigid s tru ctu re , w a g e s am ong in d u strie s w ould respond e q u a lly to e co n o m y-w id e p ro d u c tiv ity shocks but w ould show little or no re sp o n se to in d u s try -sp e c ific p ro d u c tivity movements. As a consequence, wages, on average, would grow equally among industries through time. One obvious in d ic a to r of a c o u n try ’s in d u stria l w age fle x ib ility is th e d e g re e o f d is p e rs io n , a s ta tis tic a l m easure of the in e q u a lity of w ages am ong in d u stries, adjusted fo r the m ean w age level. One w ay of d e riving th is m easure is to ca lcu la te the sta n d a rd d e v ia tio n of the natural log of wages among industries in each year; a more flexible system should produce greater variation generally, with a trend of rising dispersion during periods of econom ic flux. Higher levels of wage dispersion imply g re a te r in d u stry w age fle x ib ility because w ages re fle ct the specific circum stances of each industry. By contrast, modest and constant levels of dispersion signal equality in wage response characteristic of an inflexible system. In dustrial w age d isp e rsio n has risen in the U nited S ta te s a nd W e s t G e rm a n y s in c e 1 9 7 0 ( C h a rt 3 ). While the rise is far more pronounced in both level and trend in the United States, the rise in West Germany stands out, p a rtic u la rly in lig h t of its highly ce n tra liz e d system of co lle c tiv e b a rg a in in g . The tre n d in the W est G erm an Chart 3 Wage D ispersion in M a n u fa ctu rin g Standard deviation in log wage 0.28--------------------------------------------------------------- ................................................................I 1970 72 74 76 78 I I 80 I 82 * Data for West Germany available in even-num bered years only. Sources: U.S. Departm ent of Commerce, National Income and Product Accounts; Deutsche Bundesbank. series is also unusual in comparison with other major European countries, where industry wage dispersion has either remained roughly stable or declined somewhat.15 While rising wage dispersion is consistent with wages being flexible, it does not indicate that wages are flex ible in responding to economic events. Rising dispersion can occur due either to a mixing of industries within the wage ranking, or to an increase in the differential between wages paid to workers in high and low wage sectors. It may reflect institutional changes in union concentration or bargaining power within industries, or it may be economically motivated by structural changes in industry-specific productivity performance. Therefore, we measure the responsiveness of industry wages to evaluate the importance of economic factors in deter mining industry wage patterns. Industry wage movements are im portant in labor markets because they send signals to workers about where and how to supply their labor. If workers have complete information about wages in other industries and can move freely among industries to higher paying jobs, then wages will be responsive to specific produc tivity developments in the short term but will respond only to economy-wide productivity shifts over longer periods. This pattern results because over the long run the mobility of workers should be sufficient to equalize inter-industry wage differences. With this basic model of labor market behavior as a guide, we measure statistically the responsiveness of industry wages to specific performance using data on wages, productivity, and output prices at roughly the two-digit level for both U.S. and West German manu facturing industries. Equations linking annual changes in industrial wages to annual changes in industrial pro ductivity and output prices show that wages in West Germany were very responsive to short-run shifts in industrial performance (Box 2). While the magnitude of this effect may appear larger in West Germany than in the United States (based on the productivity estimates), these differences are not statistically significant. In any case, the U.S. and West German regressions are not strictly comparable—the industry samples differ, and the periodicity of the data is not the same.16 For each 1sSee L. Bell and R. Freeman, "Does a Flexible Industry Wage Structure Increase Employment?: The U.S. Experience," National Bureau of Economic Research Working Paper, No. 1604 (April 1984). The analysis in this section for the U.S. replicates the Bell-Freeman tests and methodology. ’ •Because West German industry wage, productivity, and price data were available for even-numbered years only, we report two-year changes in wages and sectoral performance in Box 2. While statistical tests may be unreliable due to lack of comparability in data and specification across countries, standard tests reject the notion that wage response was different across countries, largely due to the relative imprecision of the West German estimates. country, analysis reveals that short-run movements in industry wages were highly responsive to industry per formance, suggesting that industry labor markets func tioned efficiently in both West Germany and the United States in the short term. If workers are not free to move across industries, then industry wages should be correlated with specific pro ductivity and price movements over the long- as well as the short-run. For example, in a labor market with many barriers to switching jobs and obtaining training, labor mobility will be constrained. Thus in institutional settings where union rules govern the workplace, industry wage patterns may reflect both short- and long-term industry productivity trends. A second test of industrial wage flexibilty confirms the view that industrial performance influences wages over longer periods of time (Box 2). Over the period 197082, we found that long-run industry wage movements were related to long-run industry productivity movements in both the United States and West Germany. To the extent that industrial productivity movements reflect inter-industry changes in the skill mix, sex mix, or occupational structure, the link between productivity and wages can be understood as a competitive labor market revaluation of the rewards to work. If instead productivity movements reflect long-run labor demand shifts or movements along the demand schedule, then the explanation must be either that worker mobility is imperfect across sectors or that wage patterns reflect at least some noncompetitive factors. If worker mobility is limited across industries, then any shift in the industry labor demand curve must result in wage movement in the same direction—thus wage and employment growth should be positively associated among industries. In the United States, this does not seem to be the case, suggesting that the pattern of industry wage response is unlikely to be driven by longrun constraints to worker mobility. In West Germany, by contrast, industries with above average wage growth over the period 1970-82 exhibited above average employment growth. Although suggestive at best, these results imply that workers may, in fact, be inhibited from moving freely across industries in West Germany.17 While both disaggregate tests reveal that industry wages in West Germany were flexible over the 1970s 17To evaluate the relationship between wage and employment response, we correlated long-run changes among the 22 U.S. manufacturing industries over the period 1970-82, and performed the same analysis among the 29 West German industries. The correlation statistics in these tests were equal to - .0 9 for the United States and .24 for West Germany. More complete analysis of the implications of these tests for labor market performance can be found in L. Bell, "Essays in Labor Market Efficiency and Comparative Macroeconomic Performance," Ph D dissertation, Harvard University (June 1986). FRBNY Quarterly Review/Autumn 1986 19 Box 2: F le x ib ility o f In du stry Wages Industry-specific data for manufacturing industries from 1970-82 were used to analyze industrial wage flexibility in the United States and West Germany. The data were estimated linking both short-run changes in wages by industries (a panel study of 29 industries for West Ger many and 22 U.S. industries) and long-run 13-year changes in wages among industries (a cross-sectional study of wage behavior). Annual data were used for the analysis in all cases. The dependent variable is the change in the natural log of the wage in the manufac turing industry, and the independent variables include industrial productivity and output price changes, which serve as proxies for industrial performance. The gen erally lower explanatory power of each test is standard to cross-sectional wage regression. Standard errors appear in parentheses below the estimated coefficients. 1970-82 to long-run changes in specific performance, measured in the same way as above. This test shows that value productivity movements by industry influenced industry wages in both countries. Decomposing these effects, we find that while industrial output prices influ enced wages in the United States (column 2), they were insignificant in determining wage patterns in West Ger many (column 4). In both countries, industry wage movements were positively associated with industry productivity movements. Table C: Dependent Variable: Change in wage by industry (In wM982-ln wM#70) United States (1) Short-run flexibility Standard competitive theory requires that industry wages be responsive to short-run movements in industrial per formance, as a means of allocating labor efficiently across sectors. To test for short-run industry wage flex ibility, we link annual changes in industry wages to annual changes in industry value productivity (two-year changes for West Germany due to data limitations), which we decompose into industry productivity per worker and output prices. Our results (Table B) indicate a statistically significant degree of wage responsiveness in both West Germany and the United States. Table B: Dependent Variable: Short-run change in wage by industry (Inwu-lnw,^) United States A In (VA/L)tt (1) 101 (.017) A In (Q/L)it . . . . West Germany (3) .238 (.031) .118 252 .077 .021 .235 252 .077 .021 (4) .334 (.044) .184 (.035) .039 (.019) .191 (.021) A In P „ ................ R2 ...................... N ......................... . . Mean (A In w) S.D. (A In w) . . (2) .262 174 .155 .049 .299 174 .155 .049 Long-run flexibility Although the standard competitive model assumes that industrial wages will be linked only to aggregate per formance in the long term, industrial performance may in fact influence industrial wages over longer periods of time. To test the extent to which industrial performance matters for industry wages over the long run, we link long-run changes in industry wages over the period FRBNY Quarterly Review/Autumn 1986 A In (V A /I)i......................384 (.113) A In ( Q /L ) ,................ A In P ) ...................... R2 .................................. 378 N ............................... 21 Mean (A In wage) . .077 S.D. (A In wage) . . .009 (2) West Germany (3) .143 (.078) .206 (.081) .083 (.080) .338 (.142) .411 (.125) .388 21 .077 .009 (4) .110 29 .078 .006 .230 29 .078 .006 Pooled analysis The equations listed below test for the statistical equiv alence of the industry wage-productivity link in the United States and West Germany by pooling data from the two countries. The equations are estimated using Weighted Least Squares (with the assigned weights equal to the standard errors from each of the individual regressions), and the relevant explanatory statistics have been recal culated to conform with the scaling procedure. We cannot reject, based on the pooled regression results below (data are from equations 2 and 4, Table C above), the hypothesis that the relationship between industry wages and industry-specific performance is statistically the same in the United States and West Germany over this period. The implication of this test is that industry wages were equally flexible in the two countries. Pooled U.S.-W.G: (R2 = .234) (1a) A In (Wj) = .277 A in (Q t/I_i) + .209 A In (Ps) (.075) (.072) Individual U.S.-W.G.: (R2 = .322) (1b) A In (w() = .206 A In {QJU)wg + .338 A In (Qi/Lt)us (.090) (.125) + .083 A In (P ^g + .411 A In (P ^ , (.089) (.110) The F-statistic for this test is 2.85. and early 1980s, they do not explore the relative flex ibility of the U.S. and West German industrial wage structures. To make this comparison, it is necessary to pool the data for West Gemany and the United States and estimate a single equation. In fact, the link between industry wage and value productivity movements was statistically indistinguishable in the two countries (Box 2). Even though industry output price movements had a stronger impact on wages in the United States, the combined impact of the industry performance vari ables in influencing wages was the same in the two countries.18 Therefore, the response of industry wages to industry performance was, on average, just as strong in West Germany as in the United States. In sum, the indus trial wage structures in both countries were flexible.19 1*This test is not perfect. Missing variables may be of greater significance in determining wage behavior in one country than in another and therefore may bias the coefficient estimates and the statistical tests. For example, changes in the inter-industry mix of skill, sex, or age that are correlated with both industry wages and industry productivity will bias the point estimates on the industry performance variables. If the omission of these controls is more important in one country, then the pooled results will be biased as well. ’T his section explores inter-industry wage patterns, but does not evaluate /nfra-industry wage flexibility. There is reason to believe that the pattern of wages within industries among establishments is far more rigid in West Germany than in the United States. While firms in the United States vary their wages according to size, with large firms paying 25 to 30% more than small ones, firms in West Germany are generally forced to pay the union scale wage, and there is no sizeable difference between what small and large firms pay their workers. For a comprehensive discussion of firm size effect, see OECD Economic Outlook (September 1985). Conclusion The key finding of our analysis is that wages, at both aggregate and industry levels, have been flexible in West Germany, at least since the late 1970s. There appear to be strong similarities in both real and nominal aggregate wage flexibility in the United States and West Germany. While real wages responded equally to unemployment rates in both countries, nominal wages responded to prices differently. Although price inflation was an important influence on nominal wage growth in West Germany through the late 1970s, it has been rel atively unimportant recently. At the industry level, wages were flexible in the United States and West Germany as well. In the two countries, over both the short- and long-run, industry wage movements reflected changing industrial performance and showed some variation. Despite major differences in industrial structure and collective bargaining institu tions in the United States and West Germany, the degree of wage responsiveness at the industry level was similar. Since wages in the United States and West Germany have behaved similarly in recent years, wage rigidity seems unlikely to be the dominant cause of persistently high West German unemployment. Even more impor tantly, with reduced wage rigidity since the late 1970s the West German economy may be able to sustain a faster demand expansion over the next year or two without risking a resurgence of inflation. Linda A. Bell FRBNY Quarterly Review/Autumn 1986 21 The Cycle in Property/Casualty Insurance Property/casualty insurance companies hold about five percent of all financial assets in the United States. Currently they are recovering from one of their worst cyclical downturns in the post-World War II period. The industry is divided about evenly between personal and commercial lines of insurance. It is very competitive with fairly easy entry and exit and there are now more than 3000 companies operating in the United States. The vast majority of property/casualty coverage is written by a few hundred of these companies. But no single com pany supplied as much as five percent of the $85 billion of coverage written in the first half of 1986. Deviations from competition tend to be the result of regulation, particularly rate regulation, which is extensive in per sonal and workers’ compensation lines. Commercial lines other than workers’ compensation, and especially commercial reinsurance, are the focus of the current problems in the industry. Reform efforts are introducing rate regulation into these traditionally less regulated lines as well. This article focuses on the underlying reasons for the p rofitability cycle in the property/casualty industry. Changes in interest rates are the primary force behind the recurrent swings in the industry’s profitability. After describing the link between interest rate fluctuations and the insurance cycle, we look more closely at the most recent cycle. Its relative severity was primarily the result of the industry’s response to the unprecedented swings The author would like to thank Paul Bennett for many helpful discussions while developing this article. 22 FRBNY Quarterly Review/Autumn 1986 in interest rates over the past ten years. Consequently, the return of interest rates and inflation to more normal historical levels should eventually ease the "crisis” in the industry. Interest rates and the insurance cycle The cyclical behavior of the property/casualty insurance industry results from the extreme interest-sensitivity of the competitive price for insurance. The key to this sensitivity is the basic nature of the insurance product. Companies receive money (premiums) in exchange for promises to pay future claims. As interest rates rise, companies can lower premiums to meet the same future claims because the interest accumulated with premiums will be greater. As a starting point for analyzing the insurance cycle, it is helpful to think of the insurance market as char acterized by a fairly stable demand curve and a supply curve that shifts with interest rates. As rates rise, the supply curve shifts to the right, companies are willing to offer more insurance at the same price, and prices fall until enough new demand is induced and/or sup pliers withdraw to clear the market. This fundamental economic relationship between policy pricing and interest rates implies that insurance companies will raise prices when interest rates fall, and lower them when interest rates rise. The magnitude of these price changes will vary with the magnitude of interest rate changes. It is not a coincidence that the intense price competition of the late 1970s and early 1980s came at the same time as the unprecedented increase in interest rates. Likewise, the enormous pre mium rate increases of recent years have coincided with the large declines in interest rates.1 The magnitude of these price changes also depends on how far competitive pressures push these firms beyond prudent underwriting practices when interest rates rise. Because the industry is quite competitive with easy entry and exit, it tends to overshoot the price level dictated by changes in interest rates.2 Inflation also has an impact on the relationship between the competitive price of insurance and interest rates. If costs of settling claims are expected to rise through time, a higher premium or investment return will be necessary to cover future costs. To the extent that rising interest rates reflect anticipated inflation, they should not affect insurance premiums. The insurance company must therefore incorporate expectations of future inflation, or more specifically future claims costs, into its pricing policy. Uncertainty about the inflation outlook can amplify the cycle in premium pricing by widening the range of inflation expectations. Firms with lower than average expectations about future inflation will price policies more cheaply than those that expect higher rates of inflation. The lower price will draw an increasing market share to companies that anticipate low inflation, unless other firms match their prices. In either event, prices will tend toward the level dictated by a lower than average inflation outlook.3 If the average level of inflation expectations is more near the mark, prices will end up too low and the extent of the ultimate industry shakeout will vary directly with the gap between actual inflation and the lower range of inflation expectations. A corollary to the basic inverse relationship between interest rates and competitive premium pricing is greater volatility of premiums in longer tailed lines of insurance 'O ther factors besides interest rates affect insurance prices. The trends toward wider liability and higher settlements are obvious factors. Insurance prices declined despite these trends, when interest rates were at the high levels of the late 1970s and early 1980s. With interest rates lower now, the trend toward higher claims costs exacerbates the rise in competitive insurance prices. Policy measures to contain increases in the scope and size of insurance settlements could conceivably act as a partial offset to the interest rate pressure for higher premiums. Unfortunately, price data is not widely available. Constructing price data is difficult because there is no standardized unit of insurance. For example, deductibles can be increased and coverage limits lowered in lieu of raising the premium. 2See Paul L. Joskow, "Cartels, Competition and Regulation in the Property-Liability Insurance Industry,” The Bell Journal of Economics, Vol. 4, No. 2 (Autumn 1973), for a discussion of competition in the property/casualty industry. 3This assumes that there is adequate capacity among firms expecting low inflation to absorb more market share. Firms with a low-inflation outlook and capacity to write more business will be among the most aggressive price cutters. (e.g., general liability) versus shorter tailed lines (e.g., auto liability). Tail length refers to the amount of time between the premium payment and the expected claims payout. Other things equal, the longer the time between the premium payment and the expected claims payout, the bigger the effect of interest rate changes on the competitive price of insurance. This corollary helps explain why certain insurance lines are more cyclical than others. It also provides a theoretical basis for the greater cyclicality of reinsurance compared with primary insurance. Reinsurers typically have a longer tail length or emergence pattern in their claims payments than primary insurers.4 Combined ratios and the interest rate-insurance cycle Property/casualty companies’ profitability is divided into two broad categories— underwriting profits and invest ment income. Rising interest rates increase the invest ment income from each premium dollar. As discussed earlier, this higher investment income allows firms to charge a lower premium for the same level of coverage. Premium cutting due to rising interest rates erodes underwriting profits. Underwriting profitability is judged by a measure called the combined ratio. This measure adds together the ratio of losses incurred over premiums earned and the ratio of commissions and other expenses incurred over premiums written and multiplies the result by 100.5 It shows the cash outflow from underwriting operations relative to the cash inflow. When the combined ratio is greater than 100, it means underwriting expenses exceed revenues. Unless investment income makes up the difference, the firm will lose money. A practice called "cashflow underwriting" relies on investment income to meet part of underwriting expenses and causes the combined ratio to exceed 100. Tradi tionally, this practice has been regarded as unsound. Investment income, in this view, is considered a buffer against unexpected underwriting losses, not a source of cashflow for anticipated claims costs. Property/casualty company aversion to cashflow underwriting was seriously ♦Reinsurance is insurance for insurers. It allows them to cede parts of the risk they assume to other insurers. Emergence patterns show the time path of the cumulative claims associated with policies written at a particular time. For example, if 10 percent of the claims ultimately made on a set of policies are paid out each year over a ten-year period the emergence pattern would show 10 percent after one year, 20 percent after two years, and so on, reaching 100 percent at the end of the tenth year. ‘ Premiums written include earned premiums and an unearned premium reserve. The earned portion of premiums written is the property of the insurance company and is based on the expired portion of the policy period. For example, an annual premium of $400 paid in advance would initially be allocated to the unearned premium reserve. After six months, $200 or half the payment would remain in the unearned premium reserve. FRBNY Quarterly Review/Autumn 1986 23 underm ined by the unusually high level of interest rates in the late 1970s and ea rly 1980s. The com b in ed ratio is ve ry cyc lic a l and m oves (w ith v a ry in g la g s) a c c o rd in g to in te re s t rate m o ve m e n ts. C yclica l peaks in the co m b in e d ratio are u su a lly a sso ciated w ith in te re st rate peaks. As rates de clin e , the co m bined ra tio tend s to d e clin e (C hart 1).6 B e ca u se p re m iu m in co m e is w e ig h te d to w a rd the present com pared with investm ent income, the trade-off o f c u rr e n t (p re m iu m ) in c o m e fo r im p ro v e d fu tu re (in ve stm e n t) incom e raises the com bined ratio, as c u r re n t e x p e n s e s ris e re la tiv e to c u rre n t u n d e rw ritin g in c o m e . F u rth e rm o re , b e c a u s e g a in s in in v e s tm e n t incom e o ccu r w ith a lag, exp e n se s also rise re la tive to to ta l revenue . T hus, d e c lin in g u n d e rw ritin g incom e is a sso cia te d w ith d e clin in g to ta l incom e (C hart 2). E ventually, how ever, as in ve stm e n t incom e increases to ta l incom e should im p ro ve . But if un d e rw ritin g s ta n d a rd s d e te r io r a te in re s p o n s e to c o m p e titiv e p ric e 6The cycle beginning in 1976 is somewhat different because the combined ratio continued to go up even as interest rates began their descent in 1982. This may reflect the large increases in expensive, long-tailed liability settlements. These underwriting losses pressure, the im provem ent in investm ent income will be in s u ffic ie n t to s e rvice the increased cla im s a sso cia te d w ith taking on g re a te r u n d e rw ritin g risks. If the price cutting is excessive, total income will deteriorate despite the rise in in ve stm e n t incom e. O nly w hen losses fo rce more prudent u n d e rw ritin g and an in d u stry sh a ke out occurs does the cycle reverse and incom e im prove. T he s ig n ific a n c e o f in te re s t ra te m o v e m e n ts fo r underwriting perform ance has increased dram atically in the past 25 years. U ntil the p ro tra cte d rise in in te re st rates th a t began in the late 1960s, in te re s t ra te s had fluctuated around a sufficiently low level that investm ent in c o m e re m a in e d a m uch le s s im p o rta n t s o u rc e of cashflow than it becam e in the 1970s. As the level of in te re st rates rose th rough each su cce ssive b u sin ess cycle, the im p o rta n ce of in te re st incom e fo r c a sh flo w increased threefold. In 1967, prem ium income w as over eighteen tim e s in te re st incom e. By 1985, th is ratio had dropped to less than seven. Footnote 6 continued emerge later on average, and therefore the combined ratio continues to deteriorate for a longer time. Special factors associated with the most recent cycle are discussed in more detail below. Chart 2 Chart 1 P roperty/C asua lty Incom e as a Percent of P o lic y h o ld e rs ’ S u rp lu s* Three Month Treasury B ill Rates and C om bined Ratios fo r P roperty/C asualty Insurers Combined ratio Percent Percent 3 0 ------- 1 2 5 -----------------------------------------------------------------------------------------------------16 4 0 1— I— I---- 1— 1— I---- 1— I— I---- 1---- 1— I— I— I___ I___1___I___I__ I___ 1967 1935 40 Source: 24 45 50 55 60 65 70 75 B e s t’s Aggregates and Averages. FRBNY Q uarterly Review/Autumn 1986 80 85 69 71 73 75 77 79 81 83 ♦P olicyholders’ surplus is the net worth of an insurer as reported in its annual statement. Source: Best's Aggregates and Averages. 85 If the disinflation-induced decline in interest rates over the past fo u r yea rs m arks a return to m ore norm al h is torical levels, prem ium s will rise simply because interest incom e w ill co ve r a m uch s m a lle r part of the o verall costs of insu ra n ce . If the c u rre n t lo w -in fla tio n scenario persists, the co m bined ra tio fo r p ro p e rty /c a s u a lty co m p anies co uld return to the lo w e r average levels that p revaile d p rio r to the 1970s. A lso, if the sharp in te re st rate flu c tu a tio n s of the late 1970s and ea rly 1980s are replaced by the m ild e r flu c tu a tio n s of e a rlie r years, the c y c lic a lity of the in d u stry co u ld dim inish. M aturity stru ctu re of claim s and the interest rate-insurance cycle The e ffe cts of in te re s t rate flu c tu a tio n s on prem ium s should be g re a te r in lines of in surance w ith longer in te rv a ls b e tw e e n th e re c e ip t o f p re m iu m s and th e paym ent of claim s. Consequently, the com bined ratio for long-duration lines of insurance should move more than the ratio fo r s h o rt-d u ra tio n lines over the in te re st rate cycle, and the m ix of in su ra n ce by lines w ill a ffe ct the tim ing and v o la tility of the p ro p e rty /c a s u a lty cycle. T he d u ra tio n of the cla im s payout is illu s tra te d by em ergence patterns (C hart 3). For example, autom obile lia b ility in su ra n ce claim s are g e n e ra lly s e ttle d sooner after the insured event than general liability or w orkers’ com pensation claim s, w hich m ight not even be reported until years after the prem ium s are paid (e.g., asbestosis claim s). W ithin th re e ye a rs of occu rre n ce, a bout 75 percen t of a u to m o b ile lia b ility claim s have been paid, while only about 25 to 40 percent of general liability and w o rk e rs ’ co m p e n sa tio n cla im s have been paid. Even a fter nine years, on ly a bout half of w o rk e rs ’ co m p e n sation cla im s have been paid out. Over the past 10 years, the com bined ratio for general lia b ility insu ra n ce has risen m ore than fo r a u tom obile lia b ility insu ra n ce w hen in te re s t rates rose, and has fallen more when rates fell. Since general liability is the lo nger taile d line, this is co n s is te n t w ith the notion that lines w ith a slo w e r em e rg e nce pattern w ill be more in te re st-se n sitiv e . O th e r fa c to rs can c o m p lic a te th is p rin c ip le . For exam ple, w o rk e rs ’ co m p e n sa tio n lines are lo n g -ta ile d , b u t th e ir c o m b in e d r a tio d o e s n o t b e h a v e as th e in creased in te re s t-s e n s itiv ity p rin cip le w ould suggest. A m o n g th e in c e n tiv e s th a t w o rk e rs ’ c o m p e n s a tio n insurers offer to prom ote safety is the return of premium d o llars to e m plo ye rs w ith a fa vo ra b le loss record. As a result, w o rk e rs ’ c o m p e n sa tio n in su re rs pay a large part of all the dividends property/casualty insurers return to p o licyh o ld e rs each year.7 S ince the com bined 7For more on special factors affecting various insurance lines, see 1985-86 Property Casualty Fact Book, Insurance Information Institute. ratio is based on premium income before distribution of dividends, underwriting perform ance in this line is often less fa vo ra b le than the co m bined ratio w ould indicate. F u rth e rm o re , re g u la tio n is m uch m o re s trin g e n t in w o rk e rs ’ co m p e n sa tio n than in g e neral lia b ility in s u r ance, perhaps re stra in in g co m p e titive e x c e s s e s .8 In this case, interest rate effects are outweighed by other factors. In past cycles the differences in perform ance between co m m e rcia l and p e rsonal lines, prim a ry in su re rs and re in su re rs, and lo n g -ta ile d and sh o rt-ta ile d lines w ithin 8There are extensive laws at the state level providing performance standards for workers' compensation insurance. In recent years these laws have changed to meet standards recommended in 1972 by the National Commission on State Workmen's Compensation Laws. Chart 3 Emergence Patterns by Line of Business Percentage of ultimate loss payments ✓ wy -----------------------------------------------------------0I______I______ I______I______ i______i______I______ I______ I 1 2 3 4 5 6 7 8 9 Report pe rio d (Y ears) Explanatory notes: 1 Emergence is defined as the sum of two items: (1) claim s outstanding at the end of each report period and (2) the cumulative payments that have been made up to the end of that report period. 2 The ratio of emergence to ultimate loss payments is expressed in percentage form and is plotted for each report period shown. 3 The usual range of points is 0 percent at inception to 100 percent (ultim ate), a point which is reached after many years or report periods. Source: 1980 Reinsurance Association of Am erica Loss Development Study. FRBNY Quarterly Review/Autumn 1986 25 these categories were much less pronounced. But sharp d iffe re n ce s arose w ith the u n u su a lly high and v o la tile level of in te re st rates d u ring the past 10 years. For exam ple, th e p e rfo rm a n ce of co m m e rcia l lines d e te ri orate d m a rke d ly in rela tio n to personal lines o ve r the past se ve ra l years. C o m m e rcia l insurance w as very a g g re s s iv e ly p riced in the last cycle and has e x p e ri enced som e of the big g e st rate in cre a se s in the past two years. Both the longer tailed nature of the business and rising costs of insu ra n ce se ttle m e n ts have c o n trib uted to th is vo la tility. Likew ise, re in s u re rs ’ p e rfo rm a n ce d e te rio ra te d much m ore than the ove ra ll p e rfo rm a n ce of p rim a ry insurers in the last cycle. An im p o rta n t d iffe re n ce betw een re in surers and p rim a ry in su re rs is the large am ount of “ e x c e s s -o f-lo s s ” co ve ra g e th e y retain. T hat is, re in surers are m ore expose d to cla im s th a t exceed large d e d u c tib le s o r s o m e lim it th a t a n o th e r in s u re r is re sp o n sib le for. Thus re in su ra n ce is lo n g er ta ile d since the e x c e s s -o f-lo s s com p o n e n t of losses is g e n e ra lly slo w e r to deve lop . As a re su lt, prem ium s in th e se lines are m ore in te re s t-s e n s itiv e . Beyond the cyclica l issue th e re is also an im p o rta n t longer term issue. There has been a secular lengthening in loss em ergence across lines. For example, the claims payout on m any re in su ra n c e lines has slow ed s u b s ta n tia lly in recent years. C la im s are com ing la te r and in Table 1 Stages of the P/C Insurance Business Cycle* Insurers' View Stage Number Consumers* View Upturn. Rising reve nues, lower combined ratios, lower average risk. 1 Crisis. Scarcity, rapid price increases, un availability of some lines. Peak. Best underwriting results, highest overall profit. 2 Consolidation. Fixing of new price plateau, highest ratio of price to actual cost of providing protection. Decline. Influx of new capital lured by high profits, price cutting, lower earnings. 3 Upturn. Easing of prices, greater avail ability, more willingness to tailor products to consumer demands. Crisis. Massive under writing losses, ruinous price competition, major risk of insolvencies. 4 Peak. Rampant price cutting, ample avail ability, full buyer’s market. *From Insuring Our Future, April 7, 1986, Report of the Governor's Advisory Commission on Liability Insurance, New York State. 26 FRBNY Quarterly Review/Autumn 1986 b ig g er am ounts than actu a ria l c a lc u la tio n s based on h isto rica l exp e rie n ce w ould in d ica te. As a re su lt, loss re s e rv e s have g e n e ra lly been in a d e q u a te in re ce n t years. W he th e r th is is due to the ch a n g e s in legal and social a ttitu d e s to w a rd in su ra n ce or to o th e r fa c to rs, it implies that the industry is more sensitive to interest rate flu c tu a tio n s than in the past. To reduce th is lo n g -ta ile d exposure, in su re rs have begun to w rite m ore co ve rage on a c la im s-m a d e b a sis.9 C la im s-m a d e p o lic ie s m ake the in d u stry less se n s itiv e to in te re s t rate flu c tu a tio n s. Stages of the cycle Table 1 show s how the in te re st rate s e n s itiv ity of p re mium pricing tra n s la te s into the sta g e s of the p ro p e rty/ ca su a lty insurance cycle. The upturn (stage 1), w here the in d u stry is now, is the re co ve ry stage w hen prices m ove back into line w ith costs and a v a ila b ility of c o v erage is a problem, as bigger risks are dropped. For the consum er, this is the problem phase of the cycle. For the insurer, it is the im p ro ve m e n t phase. It g e n e ra lly co in cid e s w ith fa llin g in te re s t rates. At the peak (stage 2), p ro fita b ility fo r the in su re r is highest, se ttin g the stage fo r the d e clin e (stage 3) as new capital comes in and price com petition reverses the p ro fit cycle. H igher in te re s t rates, should th e y em erge, w ould provide fu rth e r im petus fo r price cu ttin g at this stage. In the crisis stage, p rice c u ttin g gets out of c o n trol and com panies begin to fa il (stage 4). The m ost recent cycle began stage 1 from a trough in 1975, rose to a peak around 1978, and began a d e clin e th a t c o n tin u e d into 1984. D uring the d e c lin in g phase, prices in som e co m m e rcia l lines w ent dow n 50 pe rce n t or m ore. P ersonal lines w ere not as se rio u s ly a ffe cte d . Som e firm s attempted to raise rates in the declining phase but lost m arket share as a result. By 1984, price in cre a ses and th e re c o v e ry p h a s e had b e g u n . T h e b ig p ric e increases have been in the lines where com petition was m ost e xce ssive in the c ris is stage. The rate of return on p ro p e rty /c a s u a lty c o m p a n ie s ’ ca p ita l tra cks this cycle q u ite closely. The peak in p ro f ita b ility around 1978 a ttra cte d m any new firm s into the industry, se ttin g off the price w ars (C hart 4). What made the recent cycle different? The cycle that began in 1976 was longer than usual and more p ronounced. A tra d itio n a l rule of th u m b fo r the pro p e rty /c a s u a lty cycle is three ye a rs up and three dow n. The m ost recent cycle w as th re e ye a rs up and 9A claims-made policy covers only claims initiated during the policy period. Traditionally, coverage has been on an occurrence basis, so that an insurer covering the policy period 1987 would still be liable for claims filed in 1995 based on damages arising out of incidents occurring in 1987. Under a claims-made policy, any claims would have to be initiated by the end of 1987. six dow n, b otto m ing in 1984. W hat m ade the d ow nturn so long and sharp? Five fa cto rs seem to a cco u n t fo r the s e v e rity of the re cen t cycle and su g g e st th a t the recovery phase w ill also ta ke longer than usual: • the unusually large sw ings in interest rates over the last 10 years; • the re la tiv e ly w orse c y c lic a l p e rfo rm a n ce in lo n g tailed co m m e rcia l lines; • the u n e xp e cte d ly rapid grow th in cla im s costs; • the 1979 e n try of c a p tiv e insurers into th ird party business fo r tax re a s o n s ;10 and • the large inflow of fo re ig n re in su ra n ce capital. The m agn itu de of in te re s t rate changes alone w ould have guaranteed a longer and sharper cycle since 1975. B ecause the peak phase of the in surance cycle g e n e ra lly does not o ccu r until in te re s t rates have begun their cyclical upswing, the recovery phase of the current property/casualty cycle will probably be extended by the unusual behavior of interest rates in the present general business expansion. N orm ally interest rates would have begun to rise this fa r into an e xpansion. If rates are in a secular return to a more norm al pattern, the insurance reco very phase m ay be e xte n d e d as firm s c o n tin u e to raise prem ium s to o ffse t fa llin g in ve stm e n t incom e. 10Captive insurers are set up by firms to provide themselves with inhouse insurance services. Chart 4 Average Annual Rates of Return on Net W orth A fte r Taxes Percent of net w orth 2 0 ---------------------------- The relatively greater severity of the last cycle in long ta ile d co m m e rcia l and re in su ra n ce lines also w orks to prolong the re co ve ry phase of this cycle. C laim s a sso ciated with policies underw ritten in the aggressive pricecu ttin g phase of the last cycle (1 9 7 8 -83 ) w ill c o n tinue to haunt in su re rs w ell into the fu tu re . To som e extent c om panies have prepared fo r th is by h olding large loss reserves. But the g e neral co n se n su s seem s to be that th e in d u s try has in s u ffic ie n t re s e rv e s fo r the fu tu re claims arising from coverages written in the last cycle.11 B esides the usual cyclica l fa llo u t from poor u n d e r w ritin g , the u n e xp e cte d ly rapid grow th in cla im s costs in the long-tailed com m ercial and reinsurance lines has e xa ce rb a te d the situ a tio n . A re ce n t Rand C o rp o ra tio n study blam es the rising co sts of the p e rsonal injury system ra th e r than the volum e of cases fo r the e x p lo sio n in in s u ra n c e lo s s e s . A c c o rd in g to th is study, dam age aw ards and in su ra n ce se ttle m e n ts in personal injury cases have increased on average twice as rapidly as in fla tio n during the past five y e a rs .12 The vo lu m e of la w su its file d in creased an average of only 3.9 p e rcen t a ye a r during th is period, a ccording to the study. In essence, the high nominal interest rates of recent years were insufficient to protect insurers against rising claim s costs. The u n u su a lly high level of in fla tio n in the late 1970s and e arly 1980s m eant th a t the real return on insurers’ investm ents was much lower than the nominal return. The even greater rate of increase in claim s costs m ade the problem th a t m uch m ore severe. M any a n a lysts also a ttrib u te the s e v e rity of the last cycle to the special role of ca p tive in su re rs. In 1979, ca p tive s w ere forced to seek third party b u sin ess to maintain their special tax status as insurers. Some claim these relatively inexperienced insurers pushed prices too low by a g g re s s iv e ly bidding fo r o u tsid e business. Finally, the U.S. has tra d itio n a lly been a net im p o rter of re in su ra n ce from W estern E urope, w ith about onefourth of reinsurance coverage supplied from abroad. In the late 1970s high re tu rn s to ca p ita l in the p ro p e rty/ c a s u a lty in d u s try c o m p a re d w ith o th e r in d u s trie s attracted an in flo w of E uropean capital th a t put a d d i tional com petitive pressure on prem iums and contributed to the seve rity of the cycle. Capital adequacy and failure The am ount of p o lic y h o ld e rs ’ surplus (ca p ita l or net w orth) in the p ro p e rty /c a s u a lty in d u stry m ore than QI-------1------ 1-------1-------1------ 1-------1------ 1------ 1-------1------- 1----1975 76 77 78 79 80 81 82 83 Source: 1985-86 P roperty/C asualty Fact Book, Insurance Information Institute. 84 85 11See Insuring Our Future, April 7, 1986, Report of the Governor's Advisory Commission on Liability Insurance. New York State. See also, "Second Thoughts About Loss Reserves,” Institutional Investor (May 1986). 12See James S. Kakalik and Nicholas M. Pace, “ Costs and Compensation Paid in Tort Litigation: Testimony Before the Joint Economic Committee of the U.S. Congress,” Rand Corporation, Institute for Civil Justice (July 1986). FRBNY Quarterly Review/Autumn 1986 27 dou bled as it rose from the cyc lic a l low point of 1974. The sta ndard re g u la to ry m easure of ca p ita l ad e q u a cy is the prem ium -to-surplus ratio.13 A com pany with a low ra tio is in a p o s itio n to w rite a d d itio n a l b u s in e s s . A lth ough it has increased s o m e w h a t as m ore le verage has been accepte d ove r the ye a rs, the rule of thum b is a ratio of three- or four-to-one. Regulators may consider a com pany w ith a h igh er ratio to be o v e re xte n d e d and not in a positio n to w rite new b usiness. O f course, the s ig n ifica n ce of any p a rtic u la r p re m iu m -to -s u rp lu s ratio depe nds on o th e r fa c to rs as w e ll. For exam ple, a co m pany w ith am ple loss re s e rv e s is in a b e tte r p o sition to expand its b u sin ess p ru d e n tly than a com pany w ith in a d e q u a te re s e rv e s . B a s ic a lly , th e u n d e rre s e rv e d co m pa ny has o v e rsta te d its true capital. P ro b le m s w ith the p re m iu m -to -s u rp lu s ra tio as an indicator of capital adequacy arise because the volume of prem ium s is an im p e rfe c t m easure of p o te n tia l loss e xposure. The sam e a m o u n t of prem ium s could re fle ct e ith e r a la rge a m ount of co ve ra g e at a low price or a small am ount of coverage at a high price. Obviously, the la tte r s itu a tio n is less risky than the form er. In the c u r rent re co ve ry phase of the cycle, w ith prices high and coverage hard to get, p re m iu m s have risen re la tive to the am ount of coverage. A high prem ium -to-surplus ratio is less w o rriso m e un de r th e se co n d ition s. W hile there is w ide variation in the prem ium -to-surplus ratios of in d ivid u a l co m p a n ie s, the a g g regate ratio fo r the in d u stry d e clin e d th ro u g h the last cycle until 1983. T he re co ve ry phase of the in su ra n ce cycle is m arked by a c a p a c ity sh o rta g e w h ile th is ratio in cre a se s and co m p a n ie s are c o n stra in e d from w ritin g new business. T his helps the re co ve ry of prices. The prem ium -to-surplus ratio masks another important d im ension of risk. Two co m p a n ie s may have the sam e p re m iu m -to -su rp lu s ratio , yet one may have a much shorter average tail length on its policies than the other. C om pan ies w ith lo n g e r ta ile d business w ill g e n e ra lly carry a larger proportion of loss reserves to assets than co m pan ies w ith s h o rte r ta ile d b usiness. T h e ir ca p ita l is more leverage d as a re su lt. To m easure th is asp e ct of capital adequacy, analysts use the ratio of loss reserves 13The idea behind the premium-to-surplus ratio as an indicator of capital adequacy is straightforward. The presumption is that the amount of risk that may be safely assumed by an insurance company should be related in some way to its net assets. Policyholders' surplus is the capital cushion firms have to pay policyholders' claims if premiums prove insufficient to cover future claims costs. A few words about property/casualty accounting conventions may be useful at this point. A major liability on the books of property/ casualty companies is loss reserves. Property/casualty loss reserves are fundamentally different from loss reserves at life insurance companies or other financial institutions like banks. Property/casualty loss reserves are set up after events causing losses have occurred. Life insurance and bank loss reserves are set up in anticipation of events causing losses. 28 FRBNY Q uarterly Review/Autumn 1986 to surplus. T his m easure show s the size of exp e cted losses in relation to capital or surplus. It has more than doubled over the past 20 years. The in cre a se in the ratio of loss re se rve s to su rp lu s is the prim a ry fa cto r behind the declining capital-to-asset ratio in the industry. The re s e rve -to -su rp lu s m easure of ca p ita l a d equacy is also im perfect. If a firm d e lib e ra te ly u n d e rre s e rve s, it w ill a ppear to be in b e tte r shape than it a c tu a lly is. But as a rough in d ica to r th is ratio is useful. F irm s g e n e ra lly re se rve w ith in a s u ffic ie n tly clo se m argin of th e ir actual needs to m ake large d iffe re n c e s betw een firm s ’ r e s e r v e - t o - s u r p lu s r a t io s u s e f u l f o r c o m p a r i son purposes. There is wide dispersion in the ratios of loss reserves to surplus am ong p ro p e rty /c a s u a lty co m p a n ie s. A c o m pany w ith a high ratio has less m argin fo r e rro r in its loss re se rve co m p u ta tio n . For exam ple, a firm w ith a tw o -to -o n e ratio, u n d e rre se rve d by 10 percent, w ould have 20 percent less capital than reported, while a firm w ith a fiv e -to -o n e ratio, u n d e rre se rve d by 10 percent, w ould have 50 pe rce n t less ca p ita l than its acco un ts would show. Because there is a general consensus that firm s are u n d e rre se rve d fo r the cla im s like ly to re sult Chart 5 Involuntary R etirem ent of P ro p e rty/C a su a lty Insurers and C om bined Ratios Combined ratio Number of retirem ents 1 2 0 ------------------------------------------------------------------------------------- 30 ■ Retirement I S c a le ------ ► Combined ratios 1965 67 69 71 73 75 77 79 81 Source: B est’s Aggregates and A verages; Best's Management Reports. 83 85 from cove ra g e w ritte n in re ce n t years, firm s w ith high ra tio s o f lo s s re s e rv e s to s u rp lu s d e s e rv e s p e c ia l atten tion . Failures or involuntary retirem ents in the property/cas ualty industry generally move with the com bined ratio. This has been especially true in the high interest-rate environm ent of the past 12 years. The two most recent cyclical low points (1975 and 1984) coincide with peaks in the num ber of failures in the industry (Chart 5). Chart 6 D is trib u tio n of Best’s P roperty/C asua lty Ratings ♦ -O th e r* '•*—Fair F airly good ♦ -G ood O utlook The in d u stry is c u rre n tly e n jo yin g a s e lle rs ’ m arket. S u rp lu s in c re a s e d s u b s ta n tia lly o v e r th e p a s t year, m ainly as a re su lt of e q u ity issuance and ca p ita l gains from the stock and bond m a rke t rallies. E arnings also c o n trib u te d to su rp lu s as firm s c o n tin u e d to increase prem ium s and stopped w riting coverages in areas where legal u n c e rta in tie s p re clu d e sound a ctu a ria l e va lu atio n of risks. A v a ila b ility of co ve ra g e p ro b le m s are c o n fin e d p ri marily to product liability, d ire cto rs’ and officers’ liability, p ro fe ssio n a l liability, and e n viro n m e n ta l dam age c o v erage. In th e se lines co ve ra g e above c e rta in am ounts is now often w ritte n on a cla im s-m a d e ra th e r than an o c c u rre n c e b a s is . T h is e lim in a te s th e lo n g e r ta ile d exp osu re by c o n fin in g in su ra n ce com pany losses to c la im s m ade during the p o licy period. A look at the distribution of property/casualty com pany ratings also su g g e sts th a t the w o rst of the in d u s try ’s problem s m ay be ove r (C h a rt 6). The s te a d y d e te rio ration from 1981 to 1985 s ta b ilize d in 1986, and the strong e a rn in g s re porte d th is ye a r su g g e st the 1987 d istrib u tio n of ra tin g s w ill show som e im provem ent. The m ost v u ln e ra b le area of the in d u stry is re in s u r ance, w here high cla im s aw ards have hit hardest. T his less re gulated area has also been the fo cu s of fraud in the industry which, as in other financial industries, is an im porta nt cause of in so lv e n c ie s . U nfortunately, these problem s have cre a te d u n c e rta in ty a bout the q u a lity of reinsurance on the books of many prim ary insurers. The a d ve rse c o n s e q u e n c e s of u n c o lle c ta b le re in s u ra n c e w hich erode s su rp lu s and lim its ca p a city to w rite new b usiness should re in fo rc e the effe cts of lo w e r in te re st rates in p ro lo n g in g the re co ve ry phase of this cycle. P a rtly o ffs e ttin g th is , h o w e v e r, are th e e x c e p tio n a l o p p o rtu n itie s fo r new e n tra n ts and th o s e firm s th a t escaped the w o rst co n se q u e n ce s of the last dow nturn. The com bination of unusually high and volatile interest rates w ith the o th e r fa c to rs cited e a rlie r— the re la tive ly poor p e rfo rm a n ce of lo n g -ta ile d lines, the grow th in cla im s costs, the role of ca p tive s, and the inflow of fo reig n re in su ra n ce c a p ita l— seem s to acco u n t fo r the se ve rity of the re ce nt re ce ssio n in the industry. C o n sequently, the next d ow n tu rn should be less severe if -•—Very good E xcellent -Superior 1982 1983 1984 1985 1986 *T his category includes but is not limited to com panies w ith large reinsurance exposures. Source: A.M. Best Company. Chart 7 Loss and Loss A d ju stm e n t Expense Reserves fo r all U.S. P roperty/C asua lty Insurers and Reinsurers Percent of nominal GNP 4 .0 -------------------------------- Source: B est’s Aggregates and Averages. FRBNY Quarterly Review/Autumn 1986 29 three current trends continue. First, the return of inflation to low and less volatile levels should clarify the outlook for future claims costs. High and volatile inflation rates mask the real cost of future claims, making it more likely that some firms will price insurance inadequately to meet future obligations. Competition pressures other firms to make the same mistake. By reducing uncer tainty about future costs, price stability eliminates one important source of volatility in the industry. Second, and critically related to the inflation outlook, stability in interest rates around lower, more normal historical levels will reduce the pressure for excessive cashflow underwriting. The relationship between pre miums and claims is less variable when rates are stable. Market determined prices are more likely to match the costs of providing coverage when the cloud of interestrate uncertainty is lifted. Finally, the legal uncertainty surrounding future claims costs is also a barrier to efficient pricing. The broad ening of the legal concepts of liability and damages over 30 FRBNY Quarterly Review/Autumn 1986 the past 25 years is associated with an ever-growing share of national output devoted to insurance losses (Chart 7). More than 30 states have adopted elements of tort reform to stem this long-run increase in the real burden insurance costs place on the United States economy. These reforms incorporate recommendations from consumer groups and the insurance industry. Similar efforts are under way at the federal level. The unexpected claims costs associated with the broadening scope and size of insurance settlements contributed to the severity of the most recent down cycle. Successful reform efforts should mitigate the next down cycle. Taken together, these trends, along with the shift of extraordinary risks to claims-made policies, should aid the industry as it continues to improve its financial condition. Robert T. McGee Tax Reform and the Merger and Acquisition Market: The Repeal of General Utilities The 1986 Tax Reform Act repeals the so-called General Utilities doctrine— the principle that corporations liqui dating their businesses are not subject to capital gains tax on the appreciation in the value of their assets.1 This change, along with the new corporate tax rate structure, reduces the benefits and raises the costs of many mergers and acquisitions (M&A’s), especially those involving firms with undervalued assets. The repeal of General Utilities takes effect after the end of 1986 (except for generous transition rules), and along with other tax changes, may help to explain the surge in M&A activity in the second half of 1986 (Chart 1).2 A liquidating corporation, using General Utilities, escapes the tax liability that comes with appreciated assets. This can be an important element of a liqui dation, since the purchaser of the firm’s assets will wish to acquire them with an increased (stepped-up) tax value (basis) in order to claim larger depreciation and ’ The General Utilities doctrine derives its name from a 1935 Supreme Court case, General Utilities and Operating Co. v. Helvering. The Court’s decision in the case was ultimately incorporated into the Internal Revenue Code. For an overall look at the pre-reform tax implications of mergers, see Joint Committee on Taxation, Federal Income Tax Aspects of Mergers and Acquisitions (JCS-6-85), March 29, 1985. “Other elements of tax reform have affected M&A activity. The January 1, 1987 increase in the personal long-term capital gains tax rate— raising the maximum from 20% to 28%—created an important incentive to accelerate the completion of sales from 1987 to 1986. New rules on the transfer of net operating loss carryforwards and changes in the corporate minimum tax will make complex changes to the tax implications of many proposed mergers, favoring some and impeding others. depletion allowances. Ordinarily, a step-up implies that a corporation will incur a capital gains tax liability (the corporate capital gains tax rate, currently 28%, will rise to 34% in 1987). General Utilities is relevant to the M&A market because, under Section 338 of the Internal Revenue Code, the purchaser of at least 80% of the stock of a corporation may treat the transaction, for tax purposes, as liquidation of the corporation and purchase of its assets.3 By using General Utilities and Section 338, a corporation can obtain the advantages of a basis stepup without paying capital gains tax and without truly liquidating assets—a firm can stay in the same business with the same capital stock, managers, and workers. The tax saving arises solely from the change in own ership of a firm ’s stock. The prospect of such tax sav ings has been an important spur to the M&A market. A step-up in the basis of an acquired firm’s assets may not always be in a purchaser’s interest, however. Even though an acquired firm escapes capital gains tax on the appreciation of its assets, it still has to pay tax on that part of the basis step-up that represents the “ recapture” of past depreciation allowances. That is, because depreciation allowances are intended to cap ture the decline in an asset’s value, sale of an asset for an amount greater than the depreciated book value implies that allowances taken in the past overstated the 3The transaction must subject the selling shareholders to capital gains tax on the appreciated value of their stock. In general, a takeover involving the exchange of securities for stock is not taxable while a cash purchase is. FRBNY Quarterly Review/Autumn 1986 31 Chart 1 Chart 3 L a rg e S to c k T ra n s a c tio n s N um ber* Billions of dollars A c tu a l D e p r e c ia tio n C o m p a re d to D e p re c ia tio n o n a S te p p e d -u p B a s is Billions of dollars 5 1____ I____ I____ I____ I____ I____ I------- 1-------1-------1-------1-------1------ lo J F M A M J J A S O N D 1986 Includes p u b lica lly-d isclose d tra n sa ctio n s completed or cu rre n tly pending for sto ck or assets of U.S. nonfinancial corporations. Data through Novem ber 19, 1986. * D o lla r amount greater than $250 million. Source: Securities Data Company. Chart 2 F ix e d R e p r o d u c ib le T a n g ib le A s s e ts C urrent c o st less h isto ric cost B illions of dollars 100 0 -------------------------------------------------------------------------U.S. non finan cia l co rp o ra tio n s Source: Federal Reserve Board, Balance Sheets for the U.S. Economy 1946-1985, October 1986. 32 FRBNY Quarterly Review/Autumn 1986 gQ |-------1------- 1------- 1------- 1------- 1-------1------- 1------- 1------- 1------- 1 1976 77 78 79 80 81 82 83 84 85 Source: Federal Reserve Board, F low -of-Funds: Federal Reserve Bank of New York staff estim ates. true decline in value. This recapture tax may offset the advantages of the basis step-up.4 Because detailed study of a corporation’s assets is often necessary to calculate the costs and benefits of a basis step-up, purchasers are allowed one year to decide on carrying out a Section 338 liquidation. In the case of those buyouts involving a firm’s management, the acquirers are likely to know the costs and benefits of a Section 338 liquidation well in advance of sale. Although the Treasury Department has no data on the overall use of Section 338, the device seems to be used often in the aftermath of leveraged buyouts. But in general, Section 338 and General Utilities is more advantageous the greater the proportion of the acquired firm ’s overall purchase price that can be assigned to its depreciable and depletable assets, and the larger the basis step-up relative to the original cost of the assets. Many firms in the manufacturing and natural resource sectors fit this description. The inflation of the 1970s greatly increased the difference between the market value of tangible corporate assets and their tax basis, and in conjunction with the acceleration of depreciation schedules in 1981, allowed for dramatically increased depreciation allowances on existing assets. ♦Investment tax credits and certain other deductions, along with depreciation, are also recaptured. The w ave of large-scale mergers and leveraged buyouts in the last fe w years, e s p e c ia lly in m a n u fa ctu rin g and m ining, is p a rtly due to th e a ttra ctio n of the ta x -fre e basis ste p -u p un der S e ctio n 338 and G e n e ra l U tilitie s. D uring the h ig h -in fla tio n era o f the late 1970s and early 1980s, the potential size of basis step-ups surged. C h art 2 p lots the d iffe re n c e betw een the va lu e of nonfinancial corporate plant and equipm ent on a current (or re p ro d uctio n ) co st basis and on a d e p re cia te d h isto ric cost basis. C u rre n t co st can be co n sid ere d an a p p ro x im atio n of m a rke t va lu e and h is to ric co st an a p p ro x i m ation of the b a s is .5 W h ile th e se a p p ro xim a tio n s are ro u g h , th e d iffe re n c e b e tw e e n th e tw o p ro v id e s an in d ica tio n o f th e p o te n tia l a m o u n t of ste p -u p a va ila b le on p la n t and e q u ip m e n t. T h e c h a rt s h o w s th a t th e potential step-up did increase substantially in the 1970s, peaking at n e a rly $1 trillio n in 1981. The d iscre p a n cy has been red uced so m e w h a t in re ce n t ye a rs as old, u n d e rva lu e d c a p ita l has been re tired from se rvice . sThe historic cost data used in Chart 2 are derived from expenditures on new capital and straight-line depreciation schedules. The tax basis of capital would be calculated from expenditures on new and used capital and actual depreciation schedules, which can be Table 1 Effect of Merger (Before Tax Reform) Pre-merger First year Cashflow from operations . . . . Depreciation . . . Interest ................ Pre-tax profits . . . Taxes ...................... After-tax cashflow^ Post-merger* $2,000,000 $ 500,000f $ 0 $1,500,000 $ 690,000 + $2,000,000 $3,850,000 $2,200,000 ($4,050,000) ($1,863,000) (on profits) $1,380,000 (on recaptured depreciation) = ($ 483,000) $1,130,000 $ 283,000 Following years Cashflow from operations . . . . Depreciation . . . Interest ................ Pre-tax profits . . . Taxes ...................... After-tax cashflow^ $2,000,000 $ 500,000 $ 0 $1,500,000 $ 690,000 $1,310,000 $2,000,000 $3,850,000 $2,200,000 ($4,050,000) ($1,863,000) $1,663,000 Present value of after-tax cashflow s§. . . . $5,518,197 $5,703,274 ’ Purchase price $22 million, financed at 10%. fO riginal purchase price $10,000,000; current basis $7,000,000. ^Cashflow from operations less interest and taxes. §Years 1 to 5, evaluated at a 6% rate. A firm w ill w ish to step up the basis of asse ts to obtain higher depreciation allowances. Chart 3 plots the a ctual d e p re cia tio n a llo w a n ce s ta ke n by n o n fin a n cia l c o rp o ra tio n s in th e last d e cade and co m p a re s them to an estim a te of the firs t y e a r’s a llo w a n ce th a t could be o b ta in e d by ste p p in g up the basis of p la n t and e q u ip m ent to c u rre n t cost, and d e p re c ia tin g u nder p re va iling ru le s . T h e p o te n tia l b e n e fit o f a s te p -u p g ra d u a lly increased d u ring the late 1970s as in fla tio n heated up. More importantly, the sharp reduction in taxable service lives in tro d u ce d in 1981 d ra m a tic a lly in cre a se d the tax a d va n ta g e s of a ste p p e d -u p b a s is .6 In c o n ju n c tio n w ith the p o st-1 9 8 2 d e clin e in in te re s t rates, w hich reduced the cost of raising the funds used to finance takeovers, th is o p p o rtu n ity to a cce le ra te d e p re c ia tio n and avoid capital gains tax has probably facilitated many mergers. An e xam ple can illu s tra te in m ore d e ta il how the sp e c ific c h a ra c te ris tic s of the pre -re fo rm tax law c o n trib u te d to the fe a s ib ility of c e rta in d eals. The ta rg e t c o rp o ra tio n d e scrib e d in Table 1 pu rch a se d its assets fo r $1 0 ,0 0 0 ,0 0 0 , and its cu rre n t basis in th e se assets is $7,0 0 0 ,0 0 0 . The c o m p a n y ’s pre -ta x p ro fit (ca sh flow fro m o p e ra tio n s le s s in te re s t and d e p re c ia tio n ) is $ 1,5 0 0 ,0 0 0 , and it pays ta xe s at the p re -re fo rm 46% ra te .7 The sto ck o f the ta rg e t is p urchased by a n o th e r c o r po ra tio n fo r $ 2 2 ,0 0 0 ,0 0 0 . The a c q u ire r b o rro w s this m oney at 10%. The ta x basis of the a cq u ire d firm ’s assets is stepped up from $ 7 ,0 0 0 ,0 0 0 to $22,0 0 0 ,0 00, and using post-1981 rules, the d e p re c ia tio n rate on th e se assets is in creased from 5% to 17.5% (fo r s im plicity, d e p re cia tio n is assum ed to be on a s tra ig h t-lin e s c h e d u le ).8 T h u s, th e a n n u a l d e p re c ia tio n d e d u c tio n rises from $500 ,0 0 0 to $ 3 ,8 5 0 ,0 0 0 . The in cre a se in Footnote 5 continued considerably different from hypothetical straight-line schedules. Thus, the historic cost data is likely a very rough approximation of the true basis. •The hypothetical depreciation line in Chart 2 is based on the first year’s depreciation from accelerated schedules. The depreciation deductions in subsequent years tend to decline. In present value terms, the depreciation benefits enacted in 1981 were less substantial than the surge in the hypothetical line may suggest. 7For simplicity, the slight progressivity in the corporate tax schedule is ignored. •A reasonable estimate, one used in the Federal Reserve Board's macroeconomic model, is that the useful life of equipment for tax purposes was reduced from an average of 10.5 years prior to 1981 to 4.6 years today, and structures from 40 years to 19 years. Thus, the example's assumption of a 12.5 percentage-point increase in the tax depreciation rate is a bit high, but not unrealistic. On a straightline basis, the first-year depreciation rate on a capital stock equally divided between equipment and structures is now 14% as compared to 6% prior to 1981. Tax reform will increase the useful life of most structures to 30 years and slightly increase the lives of some categories of equipment. These changes will further reduce the attractiveness of stepping up the basis of assets following a merger. FRBNY Q uarterly Review/Autumn 1986 33 d e p re cia tio n , co upled w ith the in te re st expense of the borrow ed money, re su lts in a d e duction a g a inst the acquirer’s earnings of more than $4 million and a credit of nearly $2 m illion ag a in st its tax liability. In the year follow ing the merger, the target firm (which is technically selling its assets to the acquirer) will have to pay tax, at the ordinary corporate rate of 46%, on $3 Table 2 Effect of New Tax Rates (General Utilities D octrine Intact) Pre-merger F irst year Pre-tax profits . . . $1,500,000 Taxes ...................... $ 510,000 Post-merger After-tax cashflow* ($4,050,000) ($1,377,000) (on profits) + $1,020,000 (on recaptured depreciation) = ($ 357,000) $1,490,000 $ 157.000 Follow ing years Taxes ...................... After-tax cashflow* $ 510,000 $1,490,000 Present value of after-tax c a s h flo w s !. . . $6,276,422 ($1,377,000) $1,177,000 $3,995,688 'Cashflow from operations less interest and taxes. fYears 1 to 5, evaluated at a 6% rate. m illio n w o rth of re c a p tu re d d e p re c ia tio n .9 T h is ta x reduces the com bined firm ’s tax cre d it in the firs t year after the m erger to ju s t under $ 5 0 0 ,0 0 0 ; the a fte r-ta x c a s h flo w a c c ru in g to th e c o m b in e d firm fro m th e fin a n c in g o f th e ta k e o v e r and th e o p e ra tio n of th e acquired firm w ill be $283,0 0 0 . In the fo llo w in g years the after-tax cashflow will am ount to $1.7 m illion— larger th a n th e flo w to th e ta rg e t firm b e fo re th e m e rg er, d espite the rise in in te re st e xpense a sso cia te d w ith fin a n cing the deal. In present value term s, over a fiv e -y e a r h orizon the a fte r-ta x ca sh flo w th e se assets yield to the com bined firm is greater than that to the target before the merger. (The 6% d isco u nt rate used is arbitrary, but is ro ughly the a fte r-ta x return earned by a h ig h -in co m e in d ivid u a l who can invest 10% pre-tax and pays federal, state, and lo c a l in c o m e ta x ) . T h u s , if e q u ity m a r k e ts p ric e according to fiv e -y e a r e x p e cta tio n s, the e q u ity va lu e of the com bined firm w ill be g re a te r than the sum of the e q u ity v a lu e s o f th e tw o firm s b e fo re m e rg e r. T h is com es about even though no in crease in the ca sh flow from o p e ra tio n s of e ith e r firm has been assum ed, and even though in te re st on the debt raised to fin a n ce the purchase exceeds the cashflow from the acquired firm ’s o p e ra tio n s . A fte r six y e a rs , w h e n th e d e p re c ia tio n allowances are assumed to expire, the merged firm will need to a ugm ent its cashflow , sell a ssets, or re fin ance 9The actual rules on depreciation recapture are more complex than those in the example. Furthermore, the example ignores the recapture of any investment tax credit taken on the purchase of these assets. Table 3 Effect of General Utilities Repeal No change in tax rates Post-merger Pre-merger $1,500,000 $ 510,000 After-tax c a s h flo w * ............................... ($4,050,000) ($1,863,000) (on profits) + $1,380,000 (on recaptured depreciation) + $3,360,000 (on $12 million capital gain) = $2,877,000 $1,310,000 ($3,007,000) F ollow ing years Taxes ......................................................... After-tax cashflow* $ 690,000 $1,310,000 F irst year Pre-tax p ro fits ......................................... Taxes ......................................................... Present value of after-tax cashflows! $1,500,000 $ 690,000 . $5,518,187 ‘ Cashflow from operations less interest and taxes. tYears 1 to 5, evaluated at a 6% rate. 34 New tax rates Pre-merger FRBNY Quarterly Review/Autumn 1986 Post-merger ($4,050,000) ($1,377,000) (on profits) + $1,020,000 (on recaptured depreciation) + $4,080,000 (on $12 million capital gain) = $3,273,000 $1,490,000 ($3,923,000) ($1,863,000) $1,663,000 $ 510,000 $1,490,000 ($1,377,000) $1,177,000 $2,533,463 $6,276,422 $ 141,914 to cover the interest expense. In effect, the increase in depreciation expense gives the combined firm a long grace period to achieve operating economies. Tax reform greatly reduces the incentives for this transaction to take place, both by changing the cor porate tax rate structure and by repealing General Utilities. Table 2 shows the effect of reducing the cor porate tax rate from 46% to 34%. The after-tax cashflow of the target firm rises (because its tax bill falls), while that of the merged firm falls (the after-tax value of the interest and depreciation deductions declines as the tax rate falls). The present value of the cashflow of the combined firm falls below the sum of the cashflows of the two firms separately. Repeal of General Utilities sharply reduces the value of the combined firm (Table 3). The end of General Utilities means that capital gains tax is levied on the $12 m illion of the basis step-up that is not subject to recapture tax. The first year after-tax cashflow of the combined firm falls substantially, given the pre-reform corporate capital gains tax rate of 28%. Moreover, combining the new tax schedule (which, as mentioned above, includes a 34% rate on corporate capital gains) with the repeal of General Utilities produces a dis counted cashflow for the combined firm only slightly larger than for the acquirer alone. This example overemphasizes the impact of the General Utilities doctrine and its repeal— M&A activity is also motivated by non-tax factors and tax consider ations other than depreciation. Nonetheless, elimination of General Utilities may harm investors who have taken positions based on the assumption that a corporation will be liquidated, since the end of General Utilities will raise the costs of liquidation. It is not clear how great the impact will be, and whether any investors will experience outright losses. On the other hand, tax reform could cause buyers to be more interested in the underlying earnings potential of merger candidates than in their tax attributes.10 10Repeal of General Utilities, along with the increase in the personal capital gains tax rate, may also mean that cash deals—which are usually necessary to use Section 338 but subject selling shareholders to capital gains tax—will become less common. Charles Steindel FRBNY Quarterly Review/Autumn 1986 35 (This report was released to the Congress and to the press on December 8, 1986) Treasury and Federal Reserve Foreign Exchange Operations August-October 1986 After declining without interruption for nearly a year and a half, the dollar steadied during the period under review. Although the dollar continued to ease against most of the industrialized countries’ currencies through August, it moved up subsequently to close the threemonth period mixed on balance (Chart 1). From August to October, it appreciated against some currencies— 61/4 percent against the Japanese yen, 53U percent against sterling, and 21A> percent against the Swiss franc. It declined, however, by about 1 percent against the German mark and other currencies of the European Monetary System (EMS). There were dollar purchases by foreign central banks, but no intervention by the U.S. authorities during the period. As the period opened early in August, the dollar was declining. Market participants had come increasingly to question whether the major industrialized countries would be able to work together to redress their large external imbalances. The huge trade deficit of the United States and the enormous trade surpluses of Japan and Germany had shown little adjustment, not withstanding the considerable movements in exchange rates between the dollar and both the Japanese yen and German mark. Moreover, there was growing disap pointment that the sharp, $20-per-barrel drop in oil prices that occurred between November 1985 and July 1986 was failing to provide much of a boost to business activity in the oil importing industrialized countries. A report by Sam V. Cross, Executive Vice President in charge of the Foreign Group at the Federal Reserve Bank of New York and Manager of Foreign Operations for the System Open Market Account. 36 FRBNY Quarterly Review/Autumn 1986 Doubts developed that our major trading partners were likely to expand domestic demand vigorously enough to provide a global environment within which the United States could markedly improve its balance of payments position. Market participants considered seriously the possibility that the U.S. authorities might welcome a continued decline in the dollar on the grounds that central banks abroad might thert cut interest rates in their countries more quickly. Under these circum stances, market participants expected the trend toward lower interest rates to con tinue, with the United States setting the pace and other industrial countries perhaps following later on. Although there were already a few signs that the U.S. economy was regaining some momentum from the slow first half of the year, market participants still were struck by the areas of weakness in U.S. economic performance. Output and investment remained sluggish, manufac turing employment continued to decline, and retail sales were generally stagnant. At the same time, prospects for price and wage stability appeared to be good for the short term, despite some concern about the longer term inflationary implications of recent rapid monetary growth. In this environment, expectations resurfaced from time to time throughout the first few weeks of August that the Federal Reserve might again cut its discount rate, per haps operating unilaterally as it had done in July (Chart 2). As a result, in August interest rates on deposits denominated in U.S. dollars fell, and their decline was sharper than the decline in interest rates in other currencies (Chart 3). The Federal Reserve did cut its discount rate by one-half of one percentage point, Chart 1 During the period, the d o lla r paused in its long-term decline a ga inst other m ajor currencies. Percent * 5 -1 5 -2 0 -2 5 -3 0 N D 1985 J F M A M J 1986 J A S O * Percentage change of weekly average rates for dollars from the average rate for the week ending November 1, 1985. Figures calculated from New York noon quotations. to 5 1/2 p e rc e n t, e ffe c tiv e A u g u s t 21. T he e x c h a n g e market reaction was muted, partly because many m arket p a rticip a n ts expected the a u th o ritie s in G erm any and Japan to provide som e fu rth e r s tim u lu s to th e ir e c o n omies— either with m onetary or other measures— before the annual m eetings of the International Monetary Fund (IM F) and W orld Bank at the end of S eptem ber. E conom ic s ta tis tic s released in m id -A u g u st began to paint a c o n tra s tin g p ictu re betw een the G erm an and Japanese economies. The German economy, which had co n tra cte d sh a rp ly e a rly in the year, seem ed to be staging a robust re co ve ry; and o fficia l G erm an p ro je c tio n s of an a cce le ra tio n in grow th began to be given widespread credence in the financial markets. Japan, on the o th e r hand, appeared to be having much m ore d if fic u lty a d ju s tin g to th e a p p re c ia tio n of its cu rre n cy. A lthough both the mark and the yen had risen by about the sam e am o u n t a g a in st the d o lla r since early 1985, on a tra d e -w e ig h te d basis the y e n ’s a p p re cia tio n had been much g re a te r than the a p p re cia tio n of the m ark (C hart 4). W hereas G erm an m a n u fa ctu re rs lost little com petitiveness in their m arkets in other EMS countries, J a panese e xp o rt in d u strie s w ere hit hard. T h e y lost c o m p e titive n e ss not only in the U nited S ta te s but also in im portant East Asian markets. With business statistics released in August showing continued stagnation in the Chart 2 Chart 3 The Federal Reserve reduced its d isco u n t rate in August. At the end of O ctober, the Bank of Japan announced a reduction in its lending rate. S hort-term in te re s t rates closely m irrored exp e cta tio n s of in te re st rate cuts. Percent 8.5 Percent 8 ---------------------------------------------------------------------------- 7.0 5 - — i ---------m, I Japan I----- - I Germany 4.0 O N D 1985 J F M A M J J 1986 A S O _ 1 ____ I____ I____ I____ I____ I____ I____ I____ I____ I____ I J F M A M J J 1986 A S O N Weekly average of interest rates on three-month E urocurrency deposits. FRBNY Quarterly Review/Autumn 1986 37 Ja pane se m a n u fa ctu rin g s e c to r (C hart 5), m arket p a r tic ip a n ts began to q u e stio n w h e th e r the yen should a p p re cia te m uch m ore. In th e s e c irc u m s ta n c e s , tra d e rs b e g a n to s e n s e around m id -A u g u st th a t th e d o lla r had m ore room to declin e a g a in st th e G erm an m ark and the o th e r c u r rencie s o f co n tin e n ta l E urope than a g a in st the yen. W hen a large U.S. trade deficit for July was announced at the end o f A ugust, tra d e rs sold d o lla rs a g g re ssive ly a g a inst both m arks and S w iss fra n cs. T he d o lla r co n tin u e d to d e c lin e a g a in s t th e E u ro p e a n c u rre n c ie s thro u g h the end of A u g u st, even though it sta b ilize d a g a inst the yen. By m id -S e p te m b e r, th e re w a s fu rth e r e v id e n c e of im p ro ve m e n t in the U.S. e co n o m ic ou tlo o k. G ains in e m p lo y m e n t d u rin g A u g u s t w e re m o re b a la n c e d , industrial activity w as a little firmer, and retail sales were m o re b u o y a n t. T h e s e d e v e lo p m e n ts , to g e th e r w ith c o n firm a tio n of stro n g g ro w th fo r the G erm an econom y in th e s e c o n d q u a rte r, s e e m e d to s u g g e s t th a t an a tm o sp h e re s u p p o rtive of renew ed co o p e ra tio n w ould surround th e m e e tin g s o f the G roup of Five (G -5) and Group of Seven (G-7) industrial countries in Washington at the end o f the m onth. W ith Ja panese p ro d u ctio n for e xp ort d e clin in g , G erm an d o m e stic dem and replacing exports as the m ajor source of growth, and U.S. output app e a rin g to grow at a m ore s a tis fa c to ry pace, the process of a d ju stm e n t app e a re d to be underw ay at long last. In response to these developm ents, foreign exchange dealers concluded that the need for the U.S. authorities to seek further exchange rate adjustm ent had lessened, and the im m e d ia te p re ssu re on d o lla r exch a n g e rates sub sided . A t th e sam e tim e , in the w ake of repeated c o m m e n ts by G e rm a n o ffic ia ls , m a rk e t p a rtic ip a n ts becam e reconciled to the view that the Bundesbank was u n lik e ly to ease m o n e ta ry p o licy soon. As a result, e xp e cta tio n s of a fu rth e r reduction of in te re st rates fa d e d — not only in G erm any, but also in the U nited S tates and o th e r co u n trie s . U.S. in te re st rates a ctu a lly backed up so m ew h at. As d o lla r e xchange rates and interest rates both started to move up, foreign exchange p ro fe s s io n a ls b e g a n to c o v e r s iz e a b le s h o rt d o lla r po sitions. B idding fo r d o lla rs becam e intense, at tim es exagg e ra te d by rum ors th a t u n re a lis tic a lly good U.S. e c o n o m ic s ta tis tic s w e re a b o u t to be re le a s e d . By S e p te m b e r 12, the d o lla r w as sw ept up to DM 2.1030 to m atch its high e a rly in the th re e -m o n th period. A fte r m id-S e ptem ber, th e d o lla r show ed little trend. M arket p a rtic ip a n ts re m a in e d ske p tica l that, o ve r the longer term, the dollar had declined sufficiently to correct the U.S. balance of payments deficit. But over the shorter term, market participants perceived the dollar to be con solidating its position around mid-September rate levels. 38 FRBNY Quarterly Review/Autumn 1986 Chart 4 In tra d e -w e ig h te d term s, the a p p reciatio n of the Japanese yen since S eptem ber 1985 has been far greater than the a p p re cia tio n of the German mark . . . Index February 1985 = 100 150 T rade-w eighted e x ch a n g e rat ss 140 Japanese f yen r j / 130 / / 120 / / ----- J / 110 mark i i 100 i i .L. I I 1985 I I J 1 ..1 1986 1 T rade-w eighted value of the Japanese yen and the German mark v is-a -vis 17 other industrial co u n trie s derived from the International M onetary Fund’s M ultilateral Exchange Rate Model (MERM); indices resca le d to February 1985=100, the month of the lowest value during the rece n t period. Source: International Financial Statistics. Chart 5 . . . and the im pact on the Japanese m anufacturing in d u stry was also greater. Percent Index 1980=100 104 C ap acity utilization in manuf acturing S Germany m V S c a le ------► .. V v 100 86 N -83 s \ \ \ Japan \ ■*------Scale 96 1— ........1 ......... 1 1985 1 , 1986 Sources: Japan, Ministry of Industry and Trade; Germany, Ifo Econom ic Trends Survey. J80 They were sensitive to any evidence that U.S. and other m o neta ry a u th o ritie s w ould be w illin g to su p p o rt such a sta b iliz a tio n of exch an g e rates. In th is e n viro n m e n t, th e y to o k n o te o f s ta te m e n ts s u c h as th e o n e by C h a irm a n V o lc k e r on S e p te m b e r 24 th a t c u rr e n t excha nge rate re la tio n s h ip s place our in d u stry in a fa r b e tte r c o m p e titiv e p o s itio n th a n fo r s o m e y e a rs . Accordingly, the dollar fluctuated w ithout clear direction. But it w as so m e tim e s su b je ct to a b ru p t m ovem ents, especially against the mark between a range of DM 2.00 and DM 2.08. T hese abru p t s h ifts cam e in response to sta te m e n ts, a ctio n s, or rum ors of a ctio n s th o u g h t to re fle ct o ffic ia l a ttitu d e s to w a rd e xchange rates. The vie w th a t the d o lla r w as e ntering a period of g re a te r s ta b ility w as ca lle d in to q u e stio n se ve ra l tim es between m id-S eptem ber and mid-October. The first such o ccasio n cam e in respon se to s ta te m e n ts th a t b rought o ffic ia l a ttitu d e s abo ut e xch a n g e rates into que stio n. Bundesbank President Poehl was reported to have said that the Bundesbank w ould not cut its interest rates but th a t G e rm a n y w o u ld a c c e p t a s tro n g e r m a rk as its c o n trib u tio n to in te rn a tio n a l e c o n o m ic a d ju s tm e n t. S u b s e q u e n tly , T re a s u ry S e c re ta ry B a k e r s a id th a t, althoug h it w as p re fe ra b le not to rely on exch a n g e rate a d ju stm e n ts alone to reduce trade im b a la n ce s, th e re w ould need to be fu rth e r e xchange rate ch a n g e s in the a b s e n c e o f a d d itio n a l m e a s u re s to p ro m o te h ig h e r grow th abroad. In response, the d o lla r m oved down Table 1 Federal Reserve R eciprocal Currency Arrangem ents In millions of dollars Institution Austrian National Bank National Bank of Belgium Bank of Canada National Bank of Denmark Bank of England Bank of France German Federal Bank Bank of Italy Bank of Japan Bank of Mexico Netherlands Bank Bank of Norway Bank of Sweden Swiss National Bank Bank for International Settlements: Dollars against Swiss francs Dollars against other authorized European currencies Total Amount of Facility October 31, 1986 250 1,000 2,000 250 3,000 2,000 6,000 3,000 5,000 700 500 250 300 4,000 600 1,250 30,100 decisively, declining on S eptem ber 19 to DM 1.9845 and ¥ 151.77, its low fo r the p e riod a g a inst the yen. But it soon recovered m ost of th is d e clin e a fte r a European C om m unity (EC) m eeting of fin a n ce m in iste rs and c e n tral bank g o ve rn o rs at G le n e a g le s, S cotland, the fo l low ing day. M arket p a rtic ip a n ts in te rp re te d sta te m e nts about th a t m eeting as in d ica tin g th a t the EC co u n trie s had agreed to use exch a n g e m arket in te rv e n tio n , if necessary, to p ro te ct the EMS from s tra in s w hich they fe lt w ere a sso cia te d w ith th e d e clin e in the dollar. The next p oint of u n c e rta in ty o ccu rre d at the end of Septem ber. The w eekend G-5 and G -7 m eetings in W ashington ended w ith o u t a s p e cific agreem ent, w hich some observers had been looking for, that G erm any and Japan w ould cut in te re s t rates in return fo r a U.S. c o m m itm e n t to s ta b iliz e the dollar. M arket p a rticip a n ts, se nsing th a t no a rra n g em e n t w as in place to pre ve n t a resum ption of the dollar’s decline, moved to reestablish sh o rt d o lla r p o sitio ns. As a result, the d o lla r d e clined s h a rp ly a g a in s t th e c o n tin e n ta l E u ro p e a n c u rre n c ie s throughout the first half of October, hitting its low against the G erm an m ark of DM 1.9690 on O cto b e r 17. M eanw hile, the d o lla r had co n tin u e d to trade in a re la tive ly narrow range a g a inst the Ja p a n e se yen. In early September, news of a meeting between Secretary of the Treasury B aker and J a p a n ’s Finance M in ister M iyazaw a g e n erated som e a n tic ip a tio n th a t an a g re e m ent on e xchange rates m ight be fo rth co m in g . Later in S e p te m b e r, fo r e ig n in v e s to rs , d is c o u ra g e d by th e w o rse n in g b u sin ess clim a te in Japan, began to sell h oldings of sh a re s on the Tokyo sto ck m arket. This outflow , com bined w ith a gro w in g p essim ism about the likelihood of a reduction in the Bank of Japan’s discount rate, co n trib u te d to a sharp drop in the Tokyo stock m arket in the m iddle of O ctober. Ja p a n e se in stitu tio n a l investors, a tte m p tin g to o ffs e t the re su ltin g losses on th e ir yen e q u ity p o rtfo lio s p rio r to th e e n d -O c to b e r re p orting date, re a lize d p ro fits on th e ir d o lla r-d e n o m i nated assets by u n w inding hedges th a t had been put in place w hen the d o lla r w as m uch higher. These v a r ious factors generated a demand for dollars throughout most of October and reinforced sentim ent that the dollar had reached a ne a r-te rm bottom a g a in st the Japanese currency. Late in O cto b e r evid e nce w as a ccu m u la tin g th a t the U.S. economy had strengthened significantly during the third q u a rte r and th a t the U.S. tra d e p o sition had at least begun to stabilize. A prelim inary estim ate showing that real GNP in cre a se d by 2.4 pe rce n t in the third q u a rte r w a s fo llo w e d by a re p o rt th a t U .S . d u ra b le goods orders had in cre a se d 4.9 perce n t in S eptem ber. M oreover, p re lim in a ry tra d e s ta tis tic s fo r S e p te m b e r indicated a second m onth of d e clin e in the U.S. trade de ficit. FRBNY Quarterly Review/Autumn 1986 39 A t th e s a m e tim e , m a rk e t p a rtic ip a n ts b e c a m e increasingly impressed with European officials’ apparent intention to buy dollars to resist depreciation of the U.S. cu rre n cy and a sso cia te d s tra in s on the EMS. There were several reports of B undesbank and other European central bank intervention to buy dollars during October. In a d d ition , repo rted s ta te m e n ts from G erm an o fficia ls th a t any fu rth e r declin e of the d o lla r th re a te n e d e c o nom ic grow th in E urope c o n trib u te d to the perception th a t there m ight also be a lim it to the d o lla r’s d e p re c ia tio n a g a in s t the c o n tin e n ta l c u rre n c ie s . A c c o rd in g ly , w hen the dem and fo r d o lla rs a g a inst the yen s tre n g th e n ed la te in O c to b e r, a nd th e d o lla r b e g a n to firm ag a inst th a t currency, it also firm e d som ew hat against the E uropean cu rre n cie s. As th e pe riod drew to a close, the d o lla r received a final boost of support from the announcem ent of a onehalf p e rce n ta g e p oint cut in the Bank of J a p a n ’s d is count rate and an econom ic policy accord between U.S. Treasury S e cre ta ry B aker and Ja panese Finance M in iste r M iyazaw a. The accord ou tlin ed fisca l policy in iti atives, in clu d in g tax reform plans in Japan, and u n d e r scored the U.S. co m m itm e n t to reducing the budget d e ficit. The tw o c o u n trie s ju d g e d the e xchange rate re a lig n m e n t ach ie ved b e tw een th e ir c u rre n cie s since S e p te m b e r 1985 to be b ro a d ly c o n siste n t w ith present underlying econom ic fundam entals, and they reaffirm ed a w illin g n e ss to co o p e ra te on exch a n g e m a rke t issues. N o tw ith sta n d in g s ta te m e n ts by Treasury o ffic ia ls tha t U.S. intervention policy had not changed, some market p a rticip a n ts in te rp re te d the accord to be a pact for concerted in te rv e n tio n to su p p o rt the dollar. Thus the dollar continued to rise through end-October. T h is ris e in d o lla r e x c h a n g e ra te s w a s le d by an in c re a s e a g a in s t th e yen b u t w a s a c c o m p a n ie d by increases ag a inst o th e r m ajor c u rre n cie s. The increase in the d o lla r at the end of the period left it h ig h er on balance against some currencies and limited its decline ag a inst the G erm an m ark. On the tra d e -w e ig h te d basis of the Federal R eserve Board d o lla r exch a n g e rate index, the d o lla r closed the period 13/s p e rce n t higher than at the end of July. The pound ste rlin g w as the only cu rre n cy a g ainst w hich the d o lla r rose c o n s is te n tly d u ring the period under review. Som e of s te rlin g ’s d e clin e w as seen in fo re ig n exchange m arkets as re fle ctin g the im p a ct of w eak oil prices on B ritish e xp o rt re ve n u e s and g o v e rn m ent incom e. But m arket p a rtic ip a n ts w ere also co n cerned about the d ire ctio n of the g o v e rn m e n t’s o ve rall m onetary and fisca l p o licie s, as w ell as a bout p re e le c tio n p o litic a l u n c e rta in tie s . W ith th e a u th o ritie s d eciding fo rm a lly to abandon m o n e ta ry ta rg e ts as a Table 2 Drawings and Repayments by Foreign Central Banks under Regular Reciprocal Currency Arrangem ents In millions of dollars; drawings ( + ) or repayments ( - ) Central Bank Drawing on the Federal Reserve System Outstanding as of August 1, 1986 August September October Outstanding as of October 31, 1986 0 + 210.2 -6 6 .8 0 143.4 Bank of Mexico Data are on a value-date basis. Table 3 Drawings and Repayments by Foreign Central Banks under Special Swap Arrangem ents w ith the U.S. Treasury In millions of dollars; drawings ( + ) or repayments ( - ) Central Bank Drawing on the U.S. Treasury Amount of Facility Outstanding as of August 1, 1986 August September October Outstanding as of October 31, 1986 100.0 75.0 273.0 37.0 * 75.0 * * -7 5 .0 + 211.0 0 * -6 7 .0 * * 0 * 0 + 22.2 0 * 144.0 22.2 Central Bank of Bolivia Central Bank of Ecuador Bank of Mexico Central Bank of Nigeria Data are on a value-date basis. 'N o facility 40 FRBNY Quarterly Review/Autumn 1986 policy tool, expectations strengthened that the govern ment might adopt an exchange-rate guide for policy instead. As a result, discussion of sterling’s joining the intervention arrangements of the EMS became even more widespread than before, both in the press and in financial markets. But no new policy initiatives along these lines emerged during the period under review. By the end of October, sterling had depreciated by almost 6 percent against the dollar and by even more against the continental European currencies. During the period, the exchange rate mechanism of the EMS was at times subject to strain. The Irish pound was caught between the decline of sterling on the one hand and the rise of continental currencies on the other. With Irish exporters experiencing a loss of competi tiveness in the United Kingdom, Ireland’s primary export market, the Irish authorities devalued the Irish pound on August 2 by 8 percent against the bilateral central rates of the other EMS currencies. Later on, as the German mark appreciated against the dollar, it also moved up against other currencies. By late August the mark reached the top of the narrow band, a position it held throughout the remainder of the period. At times during September and to a lesser extent during October, the narrow band was fully stretched to the 21/4 percent intervention limit as the mark benefitted more than the others from the d o lla r’s decline. In response to these pressures, EC finance minister and central bank governors, at their Gleneagles meeting, agreed to try to stem the rise of the member currencies against the dollar, largely in an effort to preserve sta bility within the EMS. By late October, tensions within the EMS joint float had subsided substantially. * Chart 6 In August, the German mark m oved to the top of the narrow EMS band. P e rce n t' 3.0 2.5 2.0 15 1.0 0.5 * * * At the beginning of the three-month period, the only drawing outstanding on the credit arrangements of the U.S. Monetary Authorities was $75.0 million drawn on May 16, 1986 by the Central Bank of Ecuador against a $150.0 million U.S. Treasury Exchange Stabilization Fund (ESF) short-term swap facility. On August 14, the swap arrangement was terminated pursuant to the agreement. In the period from July through October, the U.S. Monetary Authorities provided short-term bridging facilities to Bolivia, Nigeria, and Mexico: Bolivia. The U.S. Treasury through the ESF on Sep tember 17 extended a $100.0 million financing facility to the Central Bank of Bolivia. There were no drawings made against this facility during the period under review. 0 - 0 .5 Table 4 - 1.0 - 1 . 5 I______I______ I______ I_____ I______ I______I______ I______ I______ I______I______ I______l---------- 1----------1 1 8 15 22 29 Aug 5 12 19 Sept 1986 26 3 10 17 24 31 Oct Weekly averages of 9 a.m. rates in New York fo r the weeks ending on dates shown. ♦Percentage deviation of each cu rre n cy from its ECU central rate. Dotted lines correspond to the System's 214 percent lim it on movement from bilateral central exchange rates fo r all particip a tin g currencies except the Italian lira. The lira may fluctuate 6 percent from its central rates against other EMS currencies. "I"On August 2, the Irish pound was devalued by 8 percent against the b ila te ra l central rates of the other EMS cu rre n cie s. Net P rofits ( + ) or Losses ( - ) on United States Treasury and Federal Reserve Current Foreign Exchange Operations In millions of dollars United States Treasury Exchange Federal Stabilization Reserve Fund Period August 1, 1986— October 31, 1986 Valuation profits and losses on outstanding assets and liabilities as of October 31, 1986 0 0 + 1,341.3 + 1,290.1 Data are on a value-date basis. FRBNY Quarterly Review/Autumn 1986 41 Nigeria. The U.S. Treasury, through the ESF, agreed on October 24 to provide a $37.0 million short-term facility to the Central Bank of Nigeria as part of a $250.0 million multilateral facility organized under the leadership of the Bank of England. On October 31, a drawing of $22.2 million was made on the U.S. portion. Mexico. On August 27, the U.S. Monetary Authorities agreed jo in tly to a m ultilateral arrangement in the amount of $1.1 billion with the Bank for international Settlements (acting for certain central banks) and the central banks of Argentina, Brazil, Colombia, and Uru guay to provide a near-term contingency support facility for Mexico’s international reserves. Drawings on the fa cility were made available in light of agreement between Mexico and the IMF concerning a proposed stand-by arrangement, the expected receipt by Mexico of disbursements under loans from the International Bank for Reconstruction and Development (IBRD), and the agreement by Mexico to apply drawings from the IMF and disbursements from the IBRD to the balances on outstanding drawings on the facility. On August 29, $850.0 million was made available to Mexico. On this date Mexico drew $211.0 million from the Treasury through the ESF and $210.2 million from the Federal Reserve through its regular swap facility with the Bank 42 FRBNY Quarterly Review/Autumn 1986 of Mexico. On September 30, Mexico repaid $67.0 mil lion to the ESF and $66.8 m illion to the Federal Reserve. During this period the Federal Reserve and the ESF realized no profits or losses from exchange transactions. As of October 31, cumulative bookkeeping or valuation gains on outstanding foreign currency balances were $1,341.3 million for the Federal Reserve and $1,290.1 for the Treasury’s ESF. These valuation gains represent the increase in the dollar value of outstanding currency assets valued at end-of-period exchange rates, com pared with the rates prevailing at the time the foreign currencies were acquired. The Federal Reserve and the ESF invest foreign currency balances acquired in the market as a result of their foreign operations in a variety of instruments that yield market-related rates of return and that have a high degree of quality and liquidity. Under the authority pro vided by the Monetary Control Act of 1980, the Federal Reserve invested $2,868.0 million equivalent of its for eign currency holdings in securities issued by foreign governments as of October 31. In addition, the Treasury held the equivalent of $3,980.1 million in such securities as of the end of October. (This report was released to the Congress and to the press on September 4, 1986) Treasury and Federal Reserve Foreign Exchange Operations May-July 1986 Report The dollar declined against most major currencies during the three months ended July. The dollar’s downward movement proceeded against the background of slug gish U.S. economic growth, expectations of continued monetary easing in the United States, and doubts that large payments im balances among the developed countries were being reduced. There was no interven tion by the U.S. authorities during the period but there were sizable dollar purchases by some other central banks. The dollar’s depreciation was temporarily inter rupted in May only to resume in June and July. By the end of July, the dollar was at its low point of the period, having declined approximately 9 percent against the Japanese yen and Swiss franc, and nearly 5 percent against the German mark and other Continental Euro pean currencies. Com ing into the period, the d o lla r had already declined substantially from its highs of February 1985. Market participants had noted that officials in several foreign industrial countries were expressing concern over the adjustments that their own industries were beginning to experience. In the face of the appreciation of their currencies, foreign exporters increasingly com plained of a squeeze on profits as they sought to maintain market shares. Indeed, a number of commen ta tors questioned w hether increases in dom estic demand in Germany and Japan would be sufficient to offset the decline in export orders and sustain prospects for economic growth in these two countries. A report by Sam Y. Cross, Executive Vice President in charge of the Foreign Group at the Federal Reserve Bank of New York and Manager of Foreign Operations for the System Open Market Account. Many in the exchange markets anticipated that the governments of the seven major industrial countries m ight use the occasion of the Economic Summit meeting in Tokyo during early May to outline measures to stabilize dollar exchange rates. The Tokyo Economic Declaration noted that there had been a significant shift in the pattern of exchange rates which better reflected fundamental economic conditions. It stated that the Group of Seven (G-7) countries had agreed to develop a process to review trends for a number of economic variables, including exchange rates, in order to achieve more effective policy coordination. But the declaration did not call for specific measures or concerted actions to prevent the dollar from declining further. Instead, there were reported remarks by some G-7 officials which seemed to imply that there was still room for further appreciation of nondollar currencies, especially the Japanese yen. In reaction to the absence of an announcement of specific measures, the dollar resumed its decline after the Tokyo Summit. It depreciated most against the Japanese yen, trading as low as ¥159.99 on May 12, some 381/z percent below its peak of about a year before. Contributing to this decline in the dollar was the narrowing of favorable long-term interest differentials. In addition, the dollar was undermined by the persistent current account imbalances manifested by a large U.S. deficit and Japanese surplus. Market participants per ceived that the U.S. Administration hoped that a high level of economic activity and rising imports abroad would set the stage for a sizable narrowing of the U.S. trade deficit, given that the dollar had already declined substantially during the past year. But the most recent data were seen by the market as showing little progress in redressing the trade imbalance. Strong protectionist FRBNY Quarterly Review/Autumn 1986 43 sentiments persisted in the U.S. manufacturing industry, even as the U.S. authorities sought to reduce restrictive trading practices abroad and resist pressures for pro te ctio n is t m easures at home. Market participants believed that so long as the imbalances were not diminishing, market pressures in favor of the yen would remain strong and that the authorities, at least in the United States, would accept further declines in dollar exchange rates. In early May, the dollar’s decline against the German mark was more muted than its decline against the yen. Political and econom ic uncertainties following the Chernobyl nuclear accident of late April weighed against the mark for a time. There were also heavy reflows of funds into the French franc and Italian lira following an April realignment of the European Monetary System (EMS) and, in the case of the franc, in response to the exchange market’s favorable reaction to initial plans for privatization of French public-sector firms. Thus, the mark traded at the bottom of the EMS. Before long, however, many in the market came to interpret official views as indicating that a period of consolidation was appropriate. Dealers anticipated that Chart 2 Interest d iffe re n tia ls favorable to the dollar in itia lly widened but then resumed th e ir d e clining trend. Percent Ten-year U.S. Treasury bonds / V • German public-sector hnnrtc; Japanese Government b o n d s \ '" \ s' EIII I I II I I I 11 I I I I 1I ! I I 11I I 111l 1l 111l l l i 1 l 111 l 1 11 l I I l 1 J A S O N D J F M A M J J 1985 1986 C hart 1 The d o lla r continue d to d e cline against most m ajor fo re ig n currencies. Percent * Pound sterling Swiss franc French franc ------ A ------20 German —mark Japanese yen - 3 5 .......... ............ ....... A S O N 1985 M A 1986 M J *Percentage change of weekly average rates for dollars from the average rate for the week ending August 2, 1985. Figures calculated from New York noon quotations. 44 FRBNY Quarterly Review/Autumn 1986 J many of the governm ents abroad, facing local or national elections, would welcome a period of exchange market tranquility. Also, time was needed to evaluate the effects on economic activity and trade flows of the changes in exchange rates and declines in interest rates that had occurred during the preceding year. After mid-May, perceptions about the relative strength of the U.S. economy temporarily brightened, expecta tions of further drops in U.S. interest rates faded, and the dollar appreciated more or less steadily for the rest of the month. Faster-than-expected growth in U.S. monetary aggregates appeared to lessen the scope for a near-term easing of U.S. monetary policy. Repeated denials of any need to ease monetary policy by officials of the Bundesbank and the Bank of Japan led dealers to believe that there was little chance of a coordinated cut in interest rates. For the first time in several months, dollar interest rates increased, with the rate on threemonth Eurodollar deposits exceeding 7 percent. A strong upward revision in first quarter real GNP and other statistics on U.S. economic activity were inter preted favorably by the exchanges. By June 2, the dollar reached ¥177.05 and DM2.3445, levels which were the highs for the dollar during the period under review. But the dollar began to edge down again in early June as new evidence suggested that the anticipated boost to U.S. exports and growth was not being sustained and e xp e cta tio n s of a n o th e r d ow nw ard a d ju stm e n t in U.S. in te re st rates w ere revived. A fte r the s ta tis tic s of late May, an in cre a se in U.S. une m p lo ym e n t cam e as a d isa p p o in tm e n t and w as the s ta rt of a se rie s of fig u re s p o in tin g to o n ly la c k lu s te r U .S . e c o n o m ic a c tiv ity . S ta tem e nts by C ha irm an V olcker w ere in te rp re te d as running co u n te r to the idea th a t the Federal R eserve needed to w ait to cut its discount rate again until central banks in other countries eased monetary policy. Market p a rticip a n ts sta rte d to c o n s id e r the p o s s ib ility th a t the U.S. authorities m ight w elcom e a renewed decline in the d o lla r on the g rou nds th a t ce n tra l banks abroad m ight cut th e ir in te re st rates m ore q u ickly in such an e n vi ronm ent. In the m eantim e, th e re w ere co n ce rn s that som e of the h e a vily inde bte d Latin A m erican co u n trie s w ere co n sid e rin g im po sin g a d ebt se rvice m oratorium o r lim itin g d e b t p a y m e n t to a p e rc e n ta g e of e x p o rt earnings. Thus, for dom estic and international reasons, m arket participants thought that a further easing of U.S. m o n e ta ry p o lic y m ig h t be im m in e n t. W ith the p o s s i b ility th a t such a U.S. m ove m ig h t not be m a tch e d Chart 3 Attention in the fo re ig n exchange m arket continued to focus on the persistent large trade im balances of the United States, Germany, and Japan. Billions of dollars Monthly trade b a la n c e s * elsew here, the d o lla r cam e under dow nw ard pressure. For several weeks in June, pressures to sell the dollar w ere well con ta in e d . D ealers p erceived th a t a u th o ritie s abroad w ere prepared to in te rve n e to p re ve n t a fu rth e r d ecline in d o lla r rates for a w hile. In particular, there were numerous reports of dollar purchases by the Bank of Japan, and m arket p a rtic ip a n ts seem ed to believe th a t th e J a p a n e s e c e n tra l b a n k w o u ld s tre n u o u s ly atte m p t to lim it the y e n ’s rise before Ja p a n e se p a rlia m entary e le ctio ns on July 6. D ealers also th o u g h t that the Bundesbank might intervene if the mark threatened to rise too strongly. In July, th e d o lla r b e g a n to m o ve d o w n q u ic k ly , e s p e cia lly a g a inst the Ja p a n e se yen and the Swiss fra n c . M a rk e t p a r tic ip a n ts d o u b te d th e J a p a n e s e a u th o ritie s w ould be able to co n ta in fo r long the y e n ’s rise in the face of m ounting trade su rp lu se s. (B ecause of the su b sta n tia l d e p re cia tio n of the d o lla r since Feb ruary 1985 and the d e clin e in w orld oil prices, J a p a n ’s trade surplus co n tin u e d to grow in d o lla r term s, even though Japanese e xports in 1986 w ere a ctu a lly low er in volume term s than in the previous year.) As a result, tra d e rs started to e sta b lish large long p o sitio ns in yen and com m ercial leads and lags swung in favor of Japan. The Swiss fra n c also began to be view ed as a p a rtic ularly a ttra ctive a lte rn a tiv e to the dollar. It w as not as Chart 4 Oil prices again moved low er. Dollars per barrel United States [ ^ \ \ | Japan | | Germany 1 0 -------------------------------------------------------- v\ \ \v X\ v West Texas interm ediate crude >\ / \ ^ ' \ 'A J K ' V Brent crude Y \ __ * Balance of payments basis for the United States. Census data for Japan and Germany. + U.S. data for June on a balance of payments basis were unavailable at the time of publication. Li i i i I i i i Nov Dec 1985 1 1 1m 1 1 1 1 1 m 1 1 1 1 1 1 1 1 1 1 1 1 1 1 11 Jan Feb Mar Apr May Jun Jul Aug 1986 j j FRBNY Quarterly Review/Autumn 1986 45 a ffected as the G erm an m ark by p o litica l u n ce rta in tie s, and by June had d evelop e d an in te re st rate advantage over the mark. Moreover, m arket participants felt that the S w iss N a tio n a l B ank w o u ld m a in ta in re la tiv e ly tig h t m o n e ta ry co n d itio n s w h a te v e r the in te rn a tio n a l e n v i r o n m e n t a n d w a s n o t lik e ly to in te r v e n e in th e exch a n g e s to lim it the a p p re c ia tio n of its currency. The G erm an m ark, too, began to gain m ore strength as the d o lla r d e clin e d d u ring July. A fter the Federal R eserve cut its d is c o u n t rate a half of one percentage point, e ffe ctive Ju ly 11, a num ber of G erm an o ffic ia ls co m m ented th a t a fu rth e r d e clin e in G erm an in te re st rates would be inappropriate inasm uch as their dom estic econom y had picked up in the second q u a rte r and the grow th of ce n tra l bank m oney rem ained above target. In a d d ition , the G erm an g o v e rn m e n t indicated it w ould not d e p a rt from its e a rlie r fis c a l ta rg e ts. The m ark also strengthened against other European currencies around this time. Flows into France that had occurred after the A pril EMS re a lig n m e n t and had w eighed on the m ark began to subside as F rench re sid e n ts re p o rte d ly too k advantage of an easing of exchange controls. The mark also b enefited from s h ifts in in v e s to r p re fe re n ce aw ay from sterling-denom inated assets, previously viewed as a principal alternative to dollar investments. As B ritain’s e conom ic o u tlo o k dim m ed w ith oil p rice s reaching new lows and the g o ve rn m en t of P rim e M in iste r T h a tch e r fa c in g c o n s id e ra b le p o litic a l c ritic is m , in v e s to rs and tra d e rs both sh ifte d fu n d s in c re a s in g ly out of ste rlin g and into m arks. D uring July, the G erm an m ark m oved from near the bottom to near the top of the EMS to em erge as the th ird s tro n g e st c u rre n c y in th a t a rra n g e ment; it also gained 7 1A> p e rce n t a g a in st ste rlin g . In late July, the d o lla r’s d e clin e a c c e le ra te d . T here w as p re s s c o m m e n ta ry to th e e ffe c t th a t, fo r o th e r in d u s tria liz e d c o u n trie s , th e b o o s t to re a l in c o m e resulting from the oil price decline was not yet showing thro u g h ; these co u n trie s w ere going to have to expand more q u ickly and im p o rt m ore vig o ro u s ly fo r the U nited States to achieve a su b sta n tia l balance of paym ents Chart 5 Chart 6 In July, the Federal Reserve low ered its disco un t rate but was not jo ine d by some other c e n tra l banks. The German mark strengthen ed w ith in the EMS jo in t flo a t. Percent * Percent * 9.0 1 1 8.5 1 1 8.0 Bank of France 7.5 7.0 Federal Reserve 6.5 6.0 5.5 5.0 1 Bank of Japan 1---------------I 1 tm—mmmm Bundesbank 4.5 4.0 1 1 1986 3.5 3.0 1 1 Jan | 1 Feb Mar . Apr 1986 1 i May 1 Jun 1 Jul ♦Percentage change in the discount rates of the central banks of the United States, Germany, and Japan and the money market intervention rate of the Bank of France. 46 FRBNY Q uarterly Review/Autumn 1986 W eekly averages of 9 a.m. rates in New York for the weeks ending on dates shown. * Percentage deviation of each currency from its ECU central rate. Dotted lines correspond to the System's 2% percent limit on movement from bilateral central exchange rates for all participating currencies except the Italian lira. The lira may fluctuate 6 percent from its central rates against other EMS currencies. Table 1 Federal Reserve Reciprocal Currency Arrangements In millions of dollars Amount of facility Amount of facility July 31, 1985 July 31, 1986 Institution Austrian National Bank National Bank of Belgium Bank of Canada National Bank of Denmark Bank of England Bank of France German Federal Bank Bank of Italy Bank of Japan Bank of Mexico Netherlands Bank Bank of Norway Bank of Sweden Swiss National Bank Bank for International Settlements: Swiss francs-dollars Other authorized European currencies-dollars 250 1,000 2,000 250 3,000 2,000 6,000 3,000 5,000 700 500 250 300 4,000 250 1,000 2,000 250 3,000 2,000 6,000 3,000 5,000 700 500 250 300 4,000 600 600 1,250 1,250 Total 30,100 30,100 as c a lcu la ted by the Federal R eserve B oard, the d o lla r closed the period 3 percent below its level of end-April. On May 14, the U.S. Treasury, thro u g h the E xchange S ta b iliza tio n Fund (E S F), agreed to p rovide sh o rt-te rm fin a n cing to the C entral Bank of E cuador to ta lin g $150 m illion until E cuador could fin a liz e n e g o tia tio n s fo r a new fin a n cin g fa c ility from co m m e rcia l banks and a d d i tio n a l loans from in te rn a tio n a l fin a n cia l in s titu tio n s. On May 16, the C entral Bank of E cuador m ade a draw ing of $75 m illion. T he F e d e ra l R e s e rv e and th e E SF in v e s t fo re ig n currency balances acquired in the market as a result of their foreign exchange market operations in a variety of instrum ents that yield market-related rates of return and th a t have a high d e g re e of q u a lity and liquidity. U nder the a u th o rity provided by the M onetary C ontrol A ct of 1980, as of July 31 the Federal R eserve had invested $ 2 ,9 4 1 .2 m illio n e q u iv a le n t o f its fo re ig n c u rre n c y holdings in securities issued by foreign governm ents. In a d dition, the Treasury held the e q u iva le n t of $ 4 ,0 8 3 .6 m illion in such se c u ritie s as of the end of July. Table 2 a d justm en t. Yet a U.S. o ffic ia l’s call for s tro n g e r grow th abroad had elicited replies from German and Japanese officials indicating that stim ulative policies would not be fo rth co m in g in the near term . As fo r the U nited S tates, rapid grow th in the U.S. m o n e ta ry a g g re g a te s and a su sta in e d d e clin e in U.S. in te re s t rates in d ica ted that m onetary p o licy w as not a co n s tra in t on U.S. grow th. But long-term U.S. interest rates had actually firm ed as short-term rates eased during the last half of July. Under th e s e c irc u m s ta n c e s , m a rk e t o b s e rv e rs w o n d e re d w h e th e r fo reign dem and fo r U.S. se c u ritie s w as being su stained s u ffic ie n tly to fin a n ce the U.S. d e fic its and th e re b y avoid a n o th e r sh a rp d e clin e in d o lla r rates or a fu rth e r rise in in te re s t rates. S im ultaneously, release o f U.S. trad e s ta tis tic s su g g e stin g the d e fic it had w id e n e d in J u n e re in fo rc e d th e v ie w th a t th e d e s ire d a d ju stm e n ts w ere slow in m a te ria lizin g . As m arket p a r tic ip a n ts in c re a s in g ly q u e s tio n e d w h e th e r th e m a jo r in d u stria liz e d c o u n trie s w ould be able to w ork to g e th e r to redress th e ir large eco n o m ic im balances, the d o lla r declined to close the period at DM2.0890 and ¥153.65. A t the end of July, the d o lla r had d e clin e d 9 percent a g a inst the Jap a n e se yen and S w iss franc, as well as a lm o st 5 percent a g a inst the G erm an m ark and other E M S c u rre n c ie s . It h ad re m a in e d s ta b le , h o w e ver, a g a inst the C ana dian d o lla r and had risen a g a inst the pound ste rlin g . T h e re fo re , on a tra d e -w e ig h te d basis against the currencies of the m ajor industrial countries, Net Profits ( + ) or Losses ( - ) on U.S. Treasury and Federal Reserve Current Foreign Exchange Operations In millions of dollars Period May 1, 1986July 31, 1986 Valuation profits and losses on outstanding assets and liabilities as of July 31, 1986* Federal Reserve U.S. Treasury Exchange Stabilization Fund 0 0 + 1,398.6 + 1,470.4 Data are on a value-date basis. 'Valuation gains represent the increase in the dollar value of outstanding currency assets valued at end-of-period exchange rates, compared with the rates prevailing at the time the foreign currencies were acquired. Table 3 Drawings under Special Swap Arrangements with the U.S. Treasury In millions of dollars; drawings ( + ) or repayments ( - ) Drawings on U.S. Treasury facilities for Central Bank of Ecuador Total facility 150 May 16, Outstanding 1986 July 31, 1986 + 75 + 75 Data are on value-date basis. FRBNY Quarterly Review/Autumn 1986 47 Single-copy subscriptions to the Quarterly Review (ISSN 0147-6580) are free. Multiple copies are available for an annual cost of $12 for each additional copy. Checks should be made payable in U.S. dollars to the Federal Reserve Bank of New York and sent to the Public Information Department, 33 Liberty Street, New York, N.Y. 10045 (212720-6134). Single and multiple copies for United States and for other Western Hem isphere subscribers are sent via third- and fourth-class mail, respectively. All copies for Eastern Hemisphere subscribers are airlifted to Amsterdam and then forwarded via surface mail. Multiple-copy subscriptions are packaged in envelopes containing no more than ten copies each. Quarterly Review subscribers also receive the Bank’s Annual Report. Quarterly Review articles may be reproduced for educational or training purposes only, providing they are reprinted in full, distributed at no profit, and include credit to the author, the publication, and the Bank. Library of Congress Catalog Card Number: 77-646559 FRBNY Quarterly Review/Autumn 1986