May 30, 2003 JPMorgan Chase Bank, New York Derek Hargreaves (1-212) 834-5536 Derek.K.Hargreaves@jpmorgan.com Carlton M.

Strong (1-212) 834-5612 Carlton.M.Strong@jpmorgan.com

Economic & Policy Research Global Data Watch Page 17

Economic Research note

JPMorgan effective exchange rates: revised and modernized

China weight up from zero to 8.7% in Morgan s s new US REER but it doesn make much difference t Inflation adjustment remains chronic challenge USD down more sharply on JPMorgan new real efs fective index than on Federal Reserve Board measure Effective exchange rate indices (also known also as tradeweighted exchange rates) are widely used by analysts for several purposes: in nominal form (NEERs), as a summary descriptor of a currency fortunes in fx markets and as an s aid in assessing how exchange rate changes bear on the host country inflation outlook; and in real (inflation-adjusted) s form (REERs), to help gauge how exchange rate changes bear on a country international trade competitiveness. s JPMorgan began publishing narrow NEERs for major currencies as long ago as 1971, as the transition from the post-WWII Bretton Woods fixed exchange rate system made clear the need for such navigational aids in the successor floating rate system. The Morgan indices have been updated and revised roughly once a decade since. REERs were developed in 1974, and so too were broad indices encompassing not only major currencies, but also those of many emerging market economies. The latest incarnation, introduced in this note, derives weights from the 2000 global pattern of trade in manufactured goods, recognizes for the first time the dramatically expanded role of China in international trade, and broadens the narrow nominal major currency indices to encompass the currencies of China, the classic four Asian tigers, and Mexico, as well as those of the OECD area. The note concludes with a brief comparison of the Morgan indices with those published, albeit for the US dollar only, by the Federal Reserve Board (FRB), using a substantially similar methodology (and reaching fairly similar results for the US dollar REER). For the US (table, first chart), the indices show the dollar to have undergone a moderate and constructive depreciation of 11-13% to date since its February 2002 peak, which will be a tonic for the US traded goods sector without presenting a significant inflationary threat in an economy operating with substantial slack. For the Euro area (table, second chart), the REER is up some 22% since the trough, a plus for household incomes but a challenge for ECB policy. For China, the REER has depreciated 6% or so over the past two years; that will encourage the country proclivity for s

Exchange rate changes, Feb 27, 2002 - May 28, 2003

% change, + indicates appreciation vs. US dollar or in effective terms US Effective rate % weight in dollar Narrow Broad USD narrow rate nominal real nominal index US -10.9 -13.4 0.0 Canada 16.0 11.9 10.7 19.2 Euro 35.8 22.7 22.0 19.8 UK 15.4 -5.7 -6.1 4.9 Denmark 36.0 8.0 10.5 0.3 Norway 32.8 7.2 5.0 0.2 Sweden 34.7 9.0 5.4 1.1 Switzerland 31.5 5.9 5.7 1.3 Japan 13.4 3.9 -2.9 16.4 China 0.0 -10.9 -6.2 10.0 Hong Kong 0.0 -5.8 -9.2 1.2 Korea 10.2 0.1 4.6 4.8 Singapore 5.8 -4.5 6.4 2.0 Taiwan 1.1 -8.4 -5.9 4.5 Australia 25.7 11.8 16.6 0.7 New Zealand 37.2 20.3 23.6 0.1 Mexico -12.6 -16.6 -7.9 13.4

USD REER - JPMorgan indices vs FRB index

Jan 98=100, broad measure 116 112 108 104 100 96 1998 1999 2000 JPMorgan new JPMorgan old FRB

2001

2002

2003

Selected broad real effective exchange rates - JPMorgan new series

Jan 1998=100 115 110 105 100 95 90 85 1998 China renminbi 1999 2000 2001 Japanese yen 2002 2003 UK pound Euro

trade surplus and compounds the adjustment challenge in coming years.

Trade flows shift hugely, the indices much less

Experience shows that even quite large shifts in weights have a relatively modest impact on calculated REERs. With

May 30, 2003 JPMorgan Chase Bank, New York Derek Hargreaves (1-212) 834-5536 Derek.K.Hargreaves@jpmorgan.com Carlton M. Strong (1-212) 834-5612 Carlton.M.Strong@jpmorgan.com

Economic & Policy Research JPMorgan effective exchange rates: revised and modernized Page 18

JPMorgan new narrow nominal indices for the US dollar and other major currencies s
The dollar index from JPMorgan that is reported daily in The Wall Street Journal and elsewhere is a narrow version of the dollar broad nominal effective exs change rate limited to major currencies only. Hitherto, the set of major currencies has been fixed as those identified in 1971 when Morgan launched the indices, namely those viewed at the time as spanning the industrial world (the United States, Canada, Western Europe, Japan, Australia, and New Zealand). This set also largely coincided with the Federal Reserve swap network with other major s central banks, and indeed, the FRB s major currency index for the dollar, as revised in 1998, covers essentially that same currency set. The world, of course, has changed enormously over the past three decades. Many, though not all, currency markets have been more or less liberalized, and many more currencies than before are actively traded outside their country of issue. More importantly, China and other countries of Emerging Asia have grown hugely in relative economic and financial size. The new JPMorgan narrow NEERs broaden the previous calculations to include the currencies of China, the four Asian tigers (Hong Kong, Korea, Singapore, and Taiwan), and Mexico. Nominal indices, however, are soon distorted if they embrace currencies of countries prone to rapid inflation, and for that reason, the new Morgan NEERs do not include the currencies of, for example, Argentina and Brazil. (Partner-currency weights in the narrow NEERs are proportional to those in the broad indices for included currencies, zero otherwise, and normalized to sum to unity.) Whether through market forces or policy design, the US dollar exchange rates of the currencies newly added to the Morgan NEERs tend to be relatively stable (in fact, CNY and HKD are unchanged). That could change in the future. For now, however, their inclusion causes the new Morgan US dollar NEER to be a lot less volatile than the old Morgan index or the FRB current major currency index. s

USD NEER - narrow measures

Jan 1998=100 115 110 105 100 95 90 1998 1999 2000 2001 2002 2003 JPMorgan new FRB "major currencies" JPMorgan old

* Partner currency weights are listed in the rightmost column of the table on the first page of this note.

one exception, this is true too of the present updating of the JPMorgan index weights from the 1990 manufactures trade pattern to the 2000 pattern. The exception is the new indicesinclusion of China. China role on the world trade s stage has exploded over the past decade or so. Thus, whereas the old 1990-based Morgan US dollar indices, had they covered China, would have assigned the renminbi a fairly negligible weight of under 3%, the new Morgan US dollar indices give it a weight of 10% in the narrow NEER (table, previous page, and box above) and 8.7% in the broad REER (table, final page of this note). The renminbi has been virtually pegged to the US dollar since 1994, so its inclusion considerably reduces the volatility of the new Morgan US dollar NEER and the latter deprecias tion since February 2002. The counterparts of the large new weight for China s renminbi in the Morgan US dollar REER (and in the REERs for most other currencies) are reduced weights for the Canadian dollar and, especially, for the euro and the yen and a reduced weight for the US dollar in the REERs of other currencies. Interestingly, the yen weight has risen in the REERs for the euro and other West European currencies, yet has fallen in those for most other currencies. In 1990, West European economies were a lot less open to imports from Japan than they have since become. Note too that, compared with 1990, Latin currency REERs now give still modest, but increased weight to other currencies within the region (reflecting some increase in intraregional integration) with the exception of the Mexican peso REER, for which the NAFTA factor has dominated.

USD REER - nominal exchange rate contribution

Jan 98=100, broad measure 116 112 108 104 100 96 1998 1999 2000 2001 2002 2003 FRB JPMorgan new

USD REER - relative inflation contribution

Jan 98=100, broad measure 102 100 98 96 94 [JPMorgan new (consumer prices)] 1998 1999 2000 JPMorgan new (core prices) 2001 2002 2003 FRB (consumer prices)

Inflation adjustment: a challenge without end

Inflation differentials and volatility have retreated markedly in much of the world in recent years, but not everywhere. For that reason, inflation adjustment of effective exchange rate indices remains essential to permit meaningful assessment of international trade competitiveness. Achieving comparable inflation adjustment internationally is notori-

May 30, 2003 JPMorgan Chase Bank, New York Derek Hargreaves (1-212) 834-5536 Derek.K.Hargreaves@jpmorgan.com Carlton M. Strong (1-212) 834-5612 Carlton.M.Strong@jpmorgan.com

Economic & Policy Research Global Data Watch Page 19

ously problematic. The Morgan REERs ideally use core prices for finished manufactured goods, excluding food and energy. The rationale for the food exclusion is that most countries are so protective of domestic agriculture that international prices have little bearing on domestic pricing. The rationale for excluding energy is that many countries engage in administered pricing of energy products, and international energy prices can be highly volatile. In practice, however, available national price series permit only an approximation to the core price ideal. From time to time, the approximation can be improved, as new price data become available or an earlier choice proves unreliable; for Canada, for example, the new Morgan REER uses consumer prices for durables and semidurables, in preference to a PPI-based core measure that had turned out to be erratic. Even with best efforts, though, compromises must be made that in some cases are unsatisfactory, especially for smaller countries. Often, for example, general PPI or WPI measures have to be used, without exclusion of food and energy; or even general CPI measures, which additionally contain a high proportion of services that, directly, are little affected by exchange rate movements.

Algebra of the indices

A currency real effective exchange rate is defined as a geometric weighted s average of the ratios of its own domestic price level to its trading partnersdomestic price levels, where all the price levels are restated in a common currency via the relevant nominal exchange rates. For convenience, but without analytical significance, the resulting geometric average is indexed to a base period (2000 average = 100 in the case of JPMorgan indices). s The present weighting methodology of the Morgan indices was adopted in 1983 and conceives the overall weight of a partner-country currency as a weighted sum of a bilateral import component and a modifiedbilateral export component. The modification aims to take account of third country competition in which, say, US firms exporting to Brazil compete in the Brazilian market not only against Brazilian firms but also rivals from Japan, Germany, and elsewhere. Formally, the weights are derived as follows:

X Xx + Mx M ij where i includes j but not x (2) W x = xi . xj d i X x M i + M i M jx M xj ()Wxj= m Mxj 3 M (3) W xj = M x Mx where W xj Trade weight for the importance of country j to country x.
W xjx Modified export component of final trade weight.
m Wxj Import component of final trade weight. X x Total exports of country x.

(1) W xj = aW

x xj

+ (1 a)W

M xj

, where a =

Xx

JPMorgan USD REER down more than FRB index s

Given the ambiguities and pitfalls in inflation adjustment, it is reassuring that they may not matter hugely in the final calculated REERs. The first chart in this note compares the US dollar REERs as calculated by Morgan and the Federal Reserve Board, while the charts on the facing page decompose each REER measure into a nominal exchange rate contribution (in fact, the broad NEER) and a relative inflation contribution. The broad NEERs turn out to be quite close to identical. But the FRB US dollar REER shows a distinctly smaller depreciation since the peak set back in February 2002 than does the Morgan REER 5.2%% vs. 9.9% as of April 2003 primarily because the FRB inflas tion adjustment relies on overall CPIs, in contrast to Morgan core prices. (Compare also, in the relative ins flation contribution chart, the Morgan core price measure with a similarly weighted CPI measure.) Finally, observe that the Morgan core price-based US dollar REER has been falling on the whole relative to the FRB CPI-based measure since late 1998 and more particularly since late 2001. Over the last five years, the US has consistently grown faster than its trading partners and has experienced slightly higher consumer price inflation but mainly in housing, medical care, and other nontraded sectors, whereas the US traded goods sector has seen pricing power collapse in the face of weak demand from trading partners and, at least until recently, a strong currency.

M x Total imports of country x. X xi Exports of country x to country i. M ij Imports of country i from country j.

M id Manufacturing GDP of country i.

Weights in the new Morgan indices are based on UN data for trade in manufactures (SITC 5-9) in 2000 and do not change from year to year (the old Morgan indices, still in use for the pre-1994 period, were based on the 1990 pattern). In 1998, the Federal Reserve Board revamped its indices for the US dollar, adopting essentially the same methodology as JPMorgan. The Board currently bases its weights on 2001 data for total exports and non-oil imports, with the weights rolled forward as new full-year data become available.

Currency weights in USD broad effective exchange rate calculations

percent JPMorgan New (2000) Old (1990) 100.0 100.0 16.7 18.3 17.2 22.3 4.3 5.5 2.5 3.8 1.7 0.4 14.2 20.1 8.7 0.0 18.2 18.5 11.6 6.3 2.9 3.5 2.1 1.3 Federal Reserve Board Current 1990 100.0 100.0 17.0 17.1 17.4 18.5 4.4 5.1 2.5 3.7 0.9 0.0 12.1 19.7 9.0 3.1 20.4 21.7 10.6 5.3 3.9 4.0 1.9 1.9

Total Canada Euro area UK Other W. Europe EM Europe Japan China Other EM Asia/ANZ Mexico Other Latin America Other n.i.e.

May 30, 2003 JPMorgan Chase Bank, New York Derek Hargreaves (1-212) 834-5536 Derek.K.Hargreaves@jpmorgan.com Carlton M. Strong (1-212) 834-5612 Carlton.M.Strong@jpmorgan.com

Economic & Policy Research JPMorgan effective exchange rates: revised and modernized Page 20

Currency weights in JPMorgan's broad effective exchange rate indices: new vs. old - selected currencies
weights in percent based on 2000 manufactures trade pattern; changes in percent points, 2000 pattern less 1990 pattern [ ( ) signifies negative]; for partner currencies, read downwrd; indices are calculated for 48 currencies (of countries listed in the core price tabulation below) US Canada Euro UK Switzerland Poland Russia Japan New wt. Ch. New wt. Ch. New wt. Ch. New wt. Ch. New wt. Ch. New wt. Ch. New wt. Ch. New wt. US 0.0 0.0 64.8 (3.8) 20.0 (4.2) 14.8 (0.4) 11.6 1.7 6.2 11.3 26.4 Canada 16.7 (1.6) 0.0 0.0 1.7 (0.2) 1.8 0.3 0.9 0.3 0.5 1.1 2.4 Euro area 17.2 (5.1) 7.3 (2.1) 0.0 0.0 48.1 (7.6) 55.0 (8.6) 59.5 39.6 15.6 UK 4.3 (1.2) 2.5 0.1 16.1 (3.5) 0.0 0.0 6.4 (0.1) 5.7 6.3 3.3 Other W. Europe 2.5 (1.3) 1.3 (0.5) 12.4 (7.1) 6.6 (2.3) 2.5 (0.9) 6.8 5.0 2.5 Emerging Europe 1.7 1.3 0.7 0.6 13.9 12.3 3.9 3.4 4.9 4.4 8.6 8.7 1.5 Japan China Other Asia/Pac Mexico Other Latin America Other n.i.e. 14.2 8.7 18.4 11.6 2.9 1.9 (5.9) 8.7 (0.3) 5.3 (0.6) 0.8 6.1 4.2 7.7 4.1 0.8 0.6 (2.2) 4.2 0.6 2.6 (0.1) 0.4 Ch. (6.4) (0.5) (4.9) (0.5) (1.2) 1.4 (7.9) 11.8 6.2 0.5 0.6 1.0 9.3 6.3 13.4 1.1 2.3 3.3 (5.2) 6.3 0.8 0.3 (0.5) 0.9 6.0 (1.2) 3.3 3.3 11.8 3.4 0.6 0.4 0.8 (0.1) 2.4 0.9 5.7 3.2 6.6 0.6 1.2 1.5 (1.5) 3.2 1.0 0.3 0.2 0.1 2.9 2.9 5.2 0.3 0.7 0.7 6.7 8.9 9.2 0.5 1.0 1.6

Ch. (9.0) (0.4) (8.4) (1.3) (1.7) 1.1

0.0 0.0 14.0 14.0 29.6 4.7 1.6 0.7 1.6 (0.3) 1.6 0.6 Ch. (1.6) 0.2 (7.0) (1.4) (1.0) 1.7 (4.1) 3.2 1.5 0.8 7.3 0.5

US Canada Euro area UK Other W. Europe Emerging Europe Japan China Other Asia/Pac Mexico Other Latin America Other n.i.e.

China Korea New wt. Ch. New wt. 17.0 22.9 1.9 1.9 15.2 13.3 2.8 2.8 2.5 1.9 2.4 1.8 22.9 0.0 31.5 1.0 1.3 1.3 21.5 11.8 17.4 1.3 1.9 1.6

Turkey New wt. Ch. 9.6 (1.4) 0.7 (0.3) 51.4 (6.4) 7.1 0.0 5.9 (0.2) 6.8 6.8 4.1 3.1 7.5 0.3 0.7 2.6 (3.8) 3.1 1.5 0.2 (0.5) 1.0

Singapore Taiwan Australia Mexico Brazil New wt. Ch. New wt. Ch. New wt. Ch. New wt. Ch. New wt. 18.7 (6.6) 21.0 (12.4) 21.3 (5.1) 68.4 3.9 28.9 1.1 (0.3) 1.9 (0.8) 1.6 (0.4) 4.2 1.7 2.5 12.1 (4.0) 12.7 (4.4) 15.1 (2.4) 7.6 (7.4) 23.6 2.7 (1.5) 2.6 (0.5) 6.0 (0.6) 1.1 (0.9) 2.8 2.2 (0.7) 1.9 (0.9) 3.1 (1.3) 1.3 (1.6) 3.5 0.9 0.7 1.2 0.9 0.5 0.4 0.4 0.4 1.9 18.4 (2.6) 7.2 7.2 34.5 7.0 0.8 0.5 0.5 (0.3) 1.0 0.7 22.6 13.2 19.1 1.2 1.0 1.6 (2.3) 13.2 5.4 0.8 (0.2) 1.2 15.9 8.9 24.7 0.6 0.6 1.7 (6.7) 8.9 6.1 0.3 (0.3) 1.1 5.1 2.9 6.5 0.0 1.9 0.5 (0.8) 2.9 2.5 0.0 (1.1) 0.4 7.0 3.2 9.0 2.6 13.7 1.4

Core prices used in JPMorgan REERs

US Canada Euro area UK Denmark Norway Sweden Switzerland Russia Turkey Japan China Hong Kong India Indonesia Korea Malaysia Philippines PPI - finished goods ex food and energy, sa CPI - durable and semidurable goods, nsa PPI manufactured goods ex food, beverages, tobacco, coke, refined petroleum products, and nuclear fuel, nsa PPI - output - manufacturing ex food, beverages, tobacco and petroleum products CPI less food, beverages, energy, nsa PPI - manufacturing less food, beverages, tobacco, and refined petroleum products, nsa PPI - manufacturing less food, beverages, tobacco, coke, refined petroleum products, and nuclear fuel, nsa CPI less food, beverages, energy & fuel, rent, nsa PPI - manufacturing, nsa WPI - manufacturing less manufactured food, beverage, tobacco, coke, and refined petroleum products, nsa CGPI - manufacturing less petroleum & coal products and processed foodstuffs, nsa Industrial goods prices, nsa Retail price index - derived from retail sales survey, nsa WPI - manufactured products ex food products, nsa WPI - manufacturing, nsa PPI - all commodities less food and energy, nsa PPI - manufactured goods, nsa PPI - manufactures ex food and petroleum products, nsa

Singapore Taiwan Thailand

PPI - manufactured goods, nsa WPI - manufacturing less processed foods and petroleum and coal products, nsa

PPI - manufactured goods less food, beverages, tobacco and petroleum products, nsa Manufacturing output prices - less components for food, beverages, tobacco and petroleum & coal Australia New Zealand CPI - all items excluding food, nsa PPI - general excluding petroleum, nsa Mexico WPI - domestic manufactured goods, nsa Argentina Wholesale price index - all items, nsa Brazil WPI - domestic industrial goods, nsa Chile WPI - domestic items, nsa Venezuela Manufacturing PPI less component for food, beverages and tobacco, sa Israel South Africa Manufacturing PPI less component for food, sa
Czech Republic, Croatia, Hungary, Poland, Romania, Slovakia, Slovenia, and Colombia: PPI - all items, nsa Kuwait, Pakistan, Peru, Egypt: WPI - all items, nsa Bulgaria, Ecuador, Panama, Morocco, Nigeria, Saudi Arabia: CPI - all items, nsa

Historical data and other information on JPMorgan effective exchange s rate indices are available through the Internet: URL: http://www2.jpmorgan.com/MarketDataInd/Forex/currIndex.html.