Purchasing Power Parity

International price comparisons

based on purchasing power parity

Because exchange rate movements, in general,

tend to be more volatile than changes in national price levels,
the purchasing power parity approach provides the proper basis
for comparing living standards and examining
productivity levels over time

I
Michelle A. Vachris magine you are planning a trip to France and using exchange rates to convert them into a
and and would like to figure out how much cur- single currency would not yield an accurate com-
James Thomas rency you will need during your visit. You parison. Again, the comparison based on ex-
would need to know how much in French francs change rates does not take into account differ-
it would cost for incidentals such as meals, ing prices among the countries.
sightseeing, and souvenirs. What information In each of these scenarios, analysts could con-
would be helpful to you in making your estimate? struct better estimates if they convert the data
You could check the price of, say, a lunch in your into a common currency and value it at the same
hometown and then convert that figure into francs price levels. In September 1998, the Organisa-
using the exchange rate. This type of estimate tion for Economic Co-operation and Develop-
would not be very accurate, however, because it ment (OECD) released price level data and mea-
is likely that a lunch in your hometown costs rela- sures for 1996 as a part of the Eurostat-OECD
tively more or less than a lunch in France. A bet- purchasing power parity ( PPP ) program.
ter estimate would be based on the price of a lunch (Eurostat is the statistical office of the European
in France. Union.) The purpose of this program is to com-
Similarly, if you were opening a subsidiary pare economic data across countries without
company in Japan, how would you determine the using exchange rates. As illustrated in the pre-
salaries for your employees? Again, using the ex- vious scenarios, exchange rates do not neces-
change rate to convert the salary you would pay sarily reflect the relative purchasing powers of
in the United States into yen would not be accu- the different currencies and applying them can
Michelle A. Vachris rate. To adequately compensate employees mov- produce inaccurate comparisons. To accurately
is an associate professor ing overseas, you would need information about compare GDP data across countries, one must ex-
of economics at
Christopher Newport the cost of living in Japan. press the data in a common currency and value
University and Finally, if a government or international orga- it at the same price level. These problems are
James Thomas is a
senior economist nization were comparing national expenditures similar to those encountered in comparisons of
in the Office of Prices across different countries, merely collecting the GDP across time for one country. Of course in
and Living Conditions,
Bureau of Labor Statistics. gross domestic products (GDPs) of the countries the case of comparison across time, the data are

Monthly Labor Review October 1999 3

Purchasing Power Parity

already expressed in one national currency, but the figures must ket of goods and services with the expenditure patterns in the
accommodate changes in the price level for the comparisons country determining the items selected. Likewise, to calculate
to have any meaning. Purchasing power parities are estimates PPPs, one needs to price a representative basket of goods and
derived from the relative price levels in different countries and services across countries. In the interspatial case, however, this
reflect the rate at which currencies can be converted to pur- becomes difficult to implement, as the different countries can
chase equivalent goods and services. Therefore, a PPP is the have very different expenditure patterns. The availability of
rate of currency conversion that equalizes purchasing power common representative products is dependent on the number
of different currencies and so has the dimensions of an ex- of countries, the extent to which their markets and expenditure
change rate as well as a price index. PPPs are preferable to patterns are similar, and the type of specification used to de-
exchange rates for converting national expenditure data into a fine selected products. Even between economies as similar as
common currency because they also adjust for differences in the United States and Canada, there remain important differ-
price levels and reflect only differences in the volume of goods ences in expenditure patterns due to differences in climate,
and services purchased between countries. tastes, packaging, regulations, and the like. Therefore, the ini-
International organizations involved with multilateral com- tial groundwork for calculating PPPs is to determine a list of
parisons of real GDP and its components increasingly base such goods and services and their detailed specifications for pric-
comparisons on PPPs.1 The European Commission also uses ing by each country. The final lists for the 1996 comparison
PPPs to determine funding levels, as well as to adjust staff sala- contained both traded and nontraded goods and services that
ries. Moreover, two recent reports analyzed the strengths and covered around 4,000 items, including about 2,900 consumer
weaknesses of the PPP programs.2 Independently, both came goods and services; 800 pharmaceuticals; 186 capital goods;
to the same conclusion, reaffirming the importance of using 50 motor vehicles; 34 government, education, and health ser-
PPPs for real multilateral comparisons of GDP and related ag- vices; and 20 construction projects.
gregates. The second component needed to calculate purchasing
Historically, the first comprehensive investigation of the power parities is the expenditure patterns for the participating
relative value of money in different countries was published in countries. These figures, expressed in national currencies, are
1940 by Colin Clark. This study compared the purchasing derived from the national accounting data for each country.
power of many currencies across a range of consumption goods. The classification system that provides the framework for the
In the early 1950s, Irving Kravis and Milton Gilbert, at the calculations is the 1968 System of National Accounts and the
Organization of the European Economic Community (prede- 1979 European System of Economic Accounts. The defini-
cessor to the OECD), used national accounts data to compare tion of expenditure starts at the basic heading level, which in-
national incomes of four western European countries and the cludes a group of similar commodities that are representative
United States. The methodology they developed was further of the purchases made in participating countries for which a
refined in benchmark studies for 1970, 1973, and 1975.3 In sample of detailed item specifications can be determined.
the early 1980s, the Eurostat-OECD PPP program was later es- Eurostat had 270 basic headings, while the OECD had 218. The
tablished independently, with its own methodology, timetable, difference between the two headings was attributed to the cat-
and mission to compare national incomes and price levels for egory, household final consumption expenditures; otherwise,
the European Union and OECD member countries. Benchmark the breakdowns were the same.
studies were published under this program for 1980, 1985, Once the detailed specifications of goods and services have
1990, and 1993. been defined within the expenditure categories, each partici-
This article analyzes the 1996 Eurostat-OECD purchasing pating country provides national average prices for as many of
power parities study, which covers 32 countries, including all the items as possible. Possibly, the most important issue in
of the current 29 member countries of the OECD, except one price selection is choosing identical products that are also im-
(Korea), plus four additional countries (Israel, Russian Fed- portant in terms of expenditure patterns. These prices, ex-
eration, Slovak Republic and Slovenia). pressed in national currencies, are for the most part, transac-
tion prices, although in some cases, list prices are used as a
Methodology proxy. Many of the countries perform special pricing surveys
to meet the requirements of the program; however, the United
Purchasing power parities measure relative price level differ- States primarily provides prices calculated from data already
ences for one time period across countries. In that respect, collected by the Consumer Price Index (CPI) and Producer Price
PPPs can be thought of as interspatial price indexes, and the Index (PPI) programs. The Bureau of Labor Statistics does not
methodology and data requirements belong to the methodol- do any special sampling for the OECD item specifications and,
ogy of index number theory. For example, to calculate as a result, the average prices provided are not the same as the
intertemporal price indexes used to measure price change for average price series produced by the CPI. Where possible, BLS
one country over time, one needs to price a representative bas- adjusts the underlying data from the CPI average price series to

4 Monthly Labor Review October 1999

match the OECD questionnaire specifications. If insufficient Table 1. Purchasing power parities, exchange rates,
average price data exist to estimate OECD specifications, then and price level indexes, 1996
(National currency per U.S. dollar)
list prices are used in lieu of, or to supplement average price
data when necessary. In particular, all of the OECD transporta- Purchasing Price
Country power Exchange level
tion survey, most of the clothing survey, and some of the con- parity rate indexes
sumer durables and furniture surveys contain list prices for the
United States. In addition, by using regression methods and United States ......................... 1 1 91
rent data collected for the CPI, the United States is able to esti- OECD 1....................................... 1.1 1 100
mate prices for the rent specifications priced by the other par- Australia ................................ 1.3 1.278 92
Canada .................................. 1.19 1.363 79
ticipating countries. Ideally, the price surveys are completed Czech Republic ..................... 11.7 27.15 39
on a flow basis over 3 years, and prices are moved forward to Hungary ................................ 72.6 152.6 43
Iceland ................................... 76.8 66.71 104
the benchmark year with intertemporal price indexes provided Japan .................................... 166 108.8 138
by each country. Mexico ................................... 3.79 7.601 45
New Zealand ......................... 1.48 1.455 92
From the matrices of prices submitted by each country, Norway .................................. 9.11 6.459 128
unweighted bilateral price ratios (parities) are calculated. The Poland ................................... 1.36 2.661 46
Switzerland ............................ 2.05 1.236 151
resultant bilateral comparisons illustrate the quantities of Turkey .................................... 39.3 81.4 44
“country A” at the prices of “country B” and the quantities of
European Union 2.................... .922 .788 106
“country B” at the prices of “country A.” Obviously, all item
specifications cannot be priced by all countries due to differ- Austria ................................... 13.6 10.59 116
Belgium ................................. 36.8 30.97 108
ences in what is available on the market, so missing parities Denmark ............................... 8.33 5.799 130
are imputed from the existing data. These unweighted pari- Finland .................................. 5.89 4.593 116
France ................................... 6.57 5.117 116
ties are then aggregated using the expenditure data provided Germany ............................... 2.03 1.505 122
by the countries. Finally, the purchasing power parities, thus Greece .................................. 214 240.8 81
Ireland .................................. . .673 .625 98
calculated, are used to convert the national expenditures into Italy ....................................... 1583 1544 93
a common currency and price level, as well as to construct Luxembourg .......................... 39.7 30.97 116
Netherlands ........................... 2.04 1.686 110
price level indexes for analytical purposes. A country’s price Portugal ................................. 122 154.3 72
level index is used to interpret whether average prices are Spain ..................................... 124 126.7 89
Sweden ................................. 9.68 6.71 131
high or low, relative to a base country or a country grouping. United Kingdom ..................... .641 .6441 91
The price level index is derived by simply taking the PPP-to-
Israel ..................................... 3.25 3.192 92
exchange-rate ratio. Russian, Federation .............. 2191 5124 39
Slovak Republic ..................... 12.2 30.65 36
Slovenia ................................ 96 135.4 64
Presentation and analysis of results
1
The OECD presently includes 29 countries: the European Union (15),
The 1996 purchasing power parities, exchange rates, and price Australia, Canada, Czech Republic, Hungary, Iceland, Japan, Korea, Mexico,
New Zealand, Norway, Poland, Switzerland, Turkey, and the United States.
level indexes for all of the participating countries are pre- Korea is not included in this analysis.
2
sented in table 1, with the U.S. dollar as the reference cur- Currently, there are 15 countries in the European Union: Austria,
Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Lux-
rency. Although the U.S. dollar is the reference currency, the embourg, Netherlands, Portugal, Spain, Sweden, and the United Kingdom.
associated PPPs refer to the price structure of the group as a SOURCE: OECD.
whole, not just the U.S. domestic price structure. With the
OECD equal to 100, the most expensive country was Switzer-
land, at 151, and the least expensive was the Slovak Repub- in a lower income country than in the United States.
lic, at 36. As shown in chart 1, prices in the United States One explanation for low-income countries having prices
were slightly below those of the OECD as a whole in 1996, lower than high-income countries is that prices are cheaper
with an index value of 91. If a country’s price level index is for nontraded items, such as housing and services. This idea
less than 91, then purchases made with a dollar converted at is based on the theory that the differences in prices broadly
market exchange rates in that country are less expensive, com- reflect differences in productivity; therefore, high-income
pared with making the same purchases in the United States. countries have more productive labor forces than low-income
Likewise, a value greater than 91 means U.S. prices are rela- countries. The differences in productivity are particularly sig-
tively cheaper. In general, the currencies of the lower income nificant for traded goods, and are reflected in higher wages
countries—including the former communist countries as well and prices. As a country develops, theoretically, its produc-
as Turkey, Mexico, Portugal, and Spain—were undervalued, tivity in the traded goods sector increases faster than that for
relative to the U.S. dollar. In other words, a U.S. dollar con- nontraded goods; nonetheless, firms making nontraded goods
verted at market exchange rates has greater purchasing power must also pay higher wage rates. The higher productivity for

Monthly Labor Review October 1999 5

Purchasing Power Parity

Chart 1. Price level indexes by country, 1996

Index (OECD = 100)
0 20 40 60 80 100 120 140 160

United States

1
OECD
Australia
Canada
Czech Republic
Hungary
Iceland
Japan
Mexico
New Zealand
Norway
Poland
Switzerland
Turkey

2
European Union
Austria
Belgium
Denmark
Finland
France
Germany
Greece
Ireland
Italy
Luxembourg
Netherlands
Portugal
Spain
Sweden
United Kingdom
Israel
Russia
Russian Federation
Slovak Republic
Slovenia

0 20 40 60 80 100 120 140 160

Index (OECD = 100)

11 The Organisation for Economic Co-operation and Development includes 28 member countries:
The OECD presently includes 29 countries: the European Union (15), Australia, Canada, Czech Republic, Hungary, Iceland,
Japan, Korea, Mexico, New Zealand, Norway, Poland, Switzerland, Turkey, and the United States. Korea is not included in
2this
The European Union includes 15 countries:
analysis.
2
Currently, there are 15 countries in the European Union: Austria, Belgium, Denmark, Finland, France, Germany, Greece,
Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, Sweden, and the United Kingdom.

6 Monthly Labor Review October 1999

Table 2. Gross domestic product per capita 1996 vided to describe why exchange rates and PPPs diverge. One
important difference between the data published by the
Nominal Nominal Real Real
Country
GDPper capita ranking GDP per capita ranking Eurostat-OECD program and theoretical purchasing power
parities is that the equilibrium exchange rate PPPs only refer
Luxembourg .................. 40,821 2 35,113 1
United States ................. 27,831 9 30,694 2 to domestically produced tradable goods and services that are
Norway .......................... 35,968 4 28,113 3 valued at export prices. In contrast, the PPPs presented in this
Switzerland .................... 41,424 1 27,504 4
Japan ............................ 36,509 3 26,447 5 article cover the entire range of final goods and services in-
Iceland ........................... 27,000 10 25,882 6 cluded in GDP estimates; many of which are not tradable and
Denmark ....................... 33,229 5 25,521 7
Canada .......................... 19,767 17 25,078 8 some of which are imported. Furthermore, the prices of traded
Belgium ......................... 26,403 11 24,490 9 goods could diverge for a number of reasons such as price
Austria ........................... 28,383 8 24,406 10
discrimination; incomplete exchange rate pass through5; cur-
Germany ....................... 28,738 7 23,525 11 rency fluctuations related to financial assets; or barriers to
Australia ........................ 21,611 15 23,440 12 trade, such as quotas, taxes, tariffs, or transportation costs.
Netherlands ................... 25,507 13 23,199 13
France ........................... 26,189 12 22,488 14 The primary purpose of the Eurostat-OECD estimated PPPs,
Italy ............................... 20,764 16 22,332 15 however, is not to predict future exchange rate movements,
Sweden ......................... 28,755 6 21,987 16
United Kingdom ............. 19,611 19 21,537 17 but rather to convert national expenditure data into a common
Finland .......................... 24,424 14 21,019 18 currency. Table 2 presents GDP per capita for all countries,
Ireland ........................... 19,632 18 20,130 19
New Zealand ................. 17,850 20 19,379 20 expressed in both nominal and real terms. The nominal values
Israel ............................. 17,679 21 19,131 21 are converted into U.S. dollars using exchange rates, while the
real figures are converted using PPPs. Just as real-inflation-
Spain ............................. 14,892 22 16,822 22
Portugal ......................... 10,960 24 15,236 23 adjusted GDP per capita is recognized as the appropriate mea-
Slovenia ........................ 9,471 25 14,728 24 sure of living standards within a country, adjusted price dif-
Greece .......................... 11,745 23 14,581 25
Czech Republic ............. 5,473 26 14,017 26 ferences across countries are just as important because such
Hungary ........................ 4,431 27 10,281 27 measures are used to adjust for price differences over time.
Slovak Republic ............. 3,533 29 9,792 28
Mexico ........................... 3,580 28 7,920 29 The differences between the nominal and real figures illus-
Poland ........................... 3,530 30 7,593 30 trate the importance of correcting for differences in price lev-
Russian Federation ....... 2,906 31 7,497 31
Turkey ............................ 2,894 32 6,617 32 els when making international comparisons. The results show
that even within a group of similar economies, there are sig-
European Union 1............ 22,986 … 21,660 …
OECD 2............................... 21,903 … 21,903 … nificant differences between nominal and real per capita GDP.
For the United States, the conversion using PPPs results in a
1
Currently, there are 15 countries in the European Union: Austria, GDP per capita estimate that ranks second among the 32 par-
Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Lux- ticipating countries, as opposed to ninth, using the nominal
embourg, Netherlands, Portugal, Spain, Sweden, and the United Kingdom.
2
The OECD presently includes 29 countries: the European Union (15), data. The top 10 countries in real and nominal terms are illus-
Australia, Canada, Czech Republic, Hungary, Iceland, Japan, Korea, Mexico, trated in the box. These rankings of the countries in terms of
New Zealand, Norway, Poland, Switzerland, Turkey, and the United States.
Korea is not included in this analysis. GDP per capita can be greatly influenced by the use of the PPP
SOURCE: OECD.

Top 10 countries in terms of gross domestic

traded goods in high-income countries, therefore, results in product per capita, 1996
higher wages and prices in both the traded and nontraded goods
sectors, relative to wages and prices in low-income countries. Real GDP Nominal GDP
It is important to note that the PPPs published by the Eurostat-
OECD program are not intended to be used as proxies for equi- Luxembourg Switzerland
librium exchange rates. They are calculated to facilitate inter- United States Luxembourg
national comparisons of prices and volumes for GDP and its Norway Japan
components. Purchasing power parities as equilibrium ex- Switzerland Norway
change rates have long existed in international trade theory.4 Japan Denmark
PPP theory predicts that in equilibrium, a dollar should buy Iceland Sweden
the same basket of goods and services in all countries, as free Denmark Germany
trade causes prices in the various countries to converge. In Canada Austria
the long run, then, the exchange rate between two countries Belgium United States
should reflect the ratio of the price levels. The relationship Austria Iceland
between exchange rates and PPP has been studied numerous SOURCE: OECD.
times over the years and a variety of reasons have been pro-

Monthly Labor Review October 1999 7

Purchasing Power Parity

Chart 2. Gross domestic product of Japan, Germany, and Canada as percentages of U.S. GDP, 1970 97
Percent Japan Percent
80 80

70 Exchange rate 70
Purchasing power parity
60 60

50 50

40 40

30 30

20 20

10 10

0 0
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996

Percent Germany Percent

80 80

70 Exchange rate 70

Purchasing power parity

60 60

50 50

40 40

30 30

20 20

10 10

0 0
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996

Percent Canada Percent

80 80
Exchange rate
70 70
Purchasing power parity
60 60

50 50

40 40

30 30

20 20

10 10

0 0
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996

8 Monthly Labor Review October 1999

Table 3. Comparison of real expenditure patterns as a narrows when purchasing power parities are used. As discussed
percent of gross domestic product, selected earlier, the exchange rates of less developed countries are typi-
countries, 1996
cally undervalued. Therefore, in poorer countries, labor-in-
tensive services such as haircuts are cheaper than such ser-
Country Private Government Gross fixed
consumption consumption capital formation vices in rich countries, and using exchange rates to convert
local currency GDPs into dollars understates the value of their
United States ........... 70.1 12.4 17.5 output relative to rich countries.
OECD 1
....................... 64.2 15.2 20.6
The relationship between the nominal and real figures does
Australia .................. 61.5 18.0 20.4 not remain stable over time. Exchange rate movements are
Canada .................... 60.4 18.9 20.7
Czech Republic ....... 44.2 34.6 21.2
much more volatile than PPPs, which tend to move gradually
Hungary .................. 49.4 37.1 13.5 with relatively small fluctuations. In particular, the estimated
Iceland ..................... 55.1 26.4 18.5
Japan ...................... 56.3 10.9 32.8
nominal figures in 1996 are quite different from previous-year
Mexico ..................... 65.0 21.0 14.0 estimates. An advantage of level comparisons based on PPPs is
New Zealand ........... 62.8 16.7 20.5
that actual output of goods and services in different countries
Norway .................... 50.7 22.6 26.7 does not change every time exchange rates between countries
Poland ..................... 54.2 31.8 14.1
Switzerland .............. 64.2 11.2 24.6 change as comparisons based on market exchange rates would
Turkey ...................... 61.1 20.5 18.4 imply. Comparative price level indexes, which incorporate
European Union2 ..... 63.0 17.2 19.8 market exchange rates in their calculation, subsequently are
Austria ..................... 55.0 19.7 25.3
Belgium ................... 65.3 15.0 19.7
also more volatile.
Denmark ................. 54.4 27.3 18.3 When exchange rate movements are more rapid and of
Finland .................... 54.3 24.9 20.8
France ..................... 60.4 19.3 20.4
greater magnitude than changes in relative prices, using them
Germany ................. 67.7 10.7 21.6 as conversion factors to estimate real expenditures becomes
Greece .................... 64.9 17.0 18.0
Ireland ..................... 60.2 18.1 21.7
meaningless. This is clearly demonstrated in chart 2 (top panel),
Italy ......................... 64.5 17.0 18.6 which shows the GDP of Japan expressed as a percentage of
Luxembourg ............ 65.4 10.6 24.0 the United States GDP for selected years. The comparisons are
Netherlands ............. 63.7 15.0 21.3 based on the conversion of GDP at current exchange rates and
Portugal ................... 56.1 24.8 19.1
Spain ....................... 62.9 17.0 20.1 current PPPs, with the extent of divergence between the two
Sweden ................... 54.6 28.1 17.3 bases over the 1970–97 period. In this example, if the 1995
United Kingdom ....... 61.5 21.8 16.7
ratios based on exchange rates are compared with those for
Israel ....................... 48.4 26.0 25.6 1985, Japan’s GDP relative to the United States, more than
Russian Federation . 48.9 36.6 14.4
Slovak Republic ....... 44.1 33.6 22.3 doubled, increasing 120 percent. Using the PPP results, how-
Slovenia .................. 53.0 27.5 19.4 ever, yields a much more modest increase of 12 percent over
1
The OECD presently includes 29 countries: the European Union (15), the 10-year period. Moreover, if market exchange rates are
Australia, Canada, Czech Republic, Hungary, Iceland, Japan, Korea, Mexico, used as the conversion factor, then the implication is that the
New Zealand, Norway, Poland, Switzerland, Turkey, and the United States.
Korea is not included in this analysis. relative standard of living of Japan fell 27 percent over the
2
Currently, there are 15 countries in the European Union: Austria, Bel- following 2-year period from 1995 to 1997. The PPP results
gium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxem-
bourg, Netherlands, Portugal, Spain, Sweden, and the United Kingdom. show a dip of only 2.7 percent over that same period.
SOURCE: OECD. Similarly, the movement of the Deutsche Mark versus the
U.S. dollar has resulted in large fluctuations in the nominal
data, as opposed to exchange rates. As can be seen from table GDP of Germany expressed as a percentage of U.S. GDP. (See
2, the PPP-based measures of real GDP per capita for Japan chart 2, center panel.) Between 1970 and 1978, the GDP of
and the West European countries are not as high, relative to Germany relative to the United States increased 57 percent,
the United States, as the nominal market exchange rate-based based on market exchange rates, whereas the same compari-
measures would suggest. son based on PPPs, resulted in a modest 4-percent decline in-
For the poorer countries, the conversions using PPPs gener- stead. Moreover, paralleling the depreciation of the mark,
ally do not yield huge differences in the rankings; however, nominal GDP as a percentage of the U.S. GDP fell 22 percent
the differences in the GDP per capita levels for these countries between 1995 and 1997, compared with a 4.5-percent real
can be quite large. In some cases, for example, Mexico, the decline. On the other hand, a comparison between Canada
real GDP per capita is more than double that of the nominal and the United States reveals exchange rate movements that
value. Although comparing the less developed countries with are far less pronounced, resulting in a more stable relation-
the most developed is analogous to comparing real output be- ship between nominal and real GDP movements. (See chart 2,
tween distant time periods, generally, it can be said that the bottom panel.) Because exchange rate movements, in gen-
gap between high-income countries and low-income countries eral, tend to be more volatile than changes in national price

Monthly Labor Review October 1999 9

Purchasing Power Parity

Table 4. Comparative price levels for final expenditure on gross domestic product at international prices
(OECD = 100)

Expenditure item Australia Austria Belgium Canada Denmark Finland France Germany Iceland Ireland

Private final consumption expenditure .............. 94 118 111 78 136 127 122 121 117 101
Food, beverages and tobacco ....................... 93 111 109 83 148 133 114 109 143 114
Clothing and footwear ................................... 104 125 142 100 134 149 145 134 145 112
Gross rent, fuel and power ............................ 103 105 113 84 118 112 129 141 79 78
Household equipment and operation ............ 103 112 108 74 117 109 116 116 113 102
Medical and health care ................................ 78 116 93 63 128 117 93 111 91 81
Transport and communication ....................... 90 139 118 89 158 139 126 124 116 124
Education, recreation and culture ................. 94 127 119 78 135 134 126 112 138 94
Miscellaneous goods and services ............... 97 130 110 65 167 143 137 122 143 119

Net purchases abroad .................................. 100 100 100 100 100 100 100 100 100 100
Government final consumption expenditure ..... 88 125 111 90 128 111 121 145 83 90
Gross fixed capital formation ............................ 93 109 101 69 124 98 103 119 99 92
Construction ................................................. 81 114 107 61 131 82 100 126 101 84
Machinery and equipment ............................ 110 103 95 82 119 121 106 111 97 106
Increase in stocks ............................................. 117 133 128 93 160 155 135 133 149 127
Balance of exports and imports ........................ 100 100 100 100 100 100 100 100 100 100
Gross domestic product ................................... 92 116 108 79 130 116 116 122 104 98

Euro-
Expenditure item Luxem- Nether- New Switzer- United United
Italy Japan Norway Sweden pean
bourg lands Zealand land Kindom States
Union1

Private final consumption

expenditure ......................................... 94 149 109 112 94 136 138 151 95 87 108
Food, beverages and tobacco ........... 106 187 105 101 102 166 139 142 105 80 107
Clothing and footwear ....................... 113 150 161 118 128 132 137 136 97 73 123
Gross rent, fuel and power ................ 70 174 107 124 96 106 135 176 80 90 104
Household equipment and
operation ........................................ 105 184 117 106 110 120 128 130 89 82 108
Medical and health care .................... 79 87 103 91 80 121 136 149 73 115 97
Transport and communication ........... 104 117 107 128 103 158 143 146 113 86 119
Education, recreation and culture ..... 109 135 109 104 95 142 142 141 92 88 109
Miscellaneous goods and services ... 101 168 112 124 76 158 148 155 108 83 113
Net purchases abroad ...................... 100 100 100 100 100 100 100 100 100 100 100

Government final consumption

expenditure .......................................... 95 124 166 111 82 131 131 201 88 114 107
Gross fixed capital formation ................ 91 127 112 110 97 111 120 129 85 88 102
Construction ..................................... 86 136 121 121 96 113 128 135 74 87 101
Machinery and equipment ................ 99 115 103 99 99 110 114 123 99 88 104
Increase in stocks ................................. 121 173 129 125 125 161 154 154 117 97 123
Balance of exports and imports ............ 100 100 100 100 100 100 100 100 100 100 100
Gross domestic product ....................... 93 138 116 110 92 128 131 151 91 91 106

1
Currently, there are 15 countries in the European Union: den, and the United Kingdom.
Austria, Belgium, Denmark, Finland, France, Germany, Greece,
Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, Swe- N OTE : Data are from OECD.

levels, the PPP approach obviously provides the proper basis government expenditures and 17.5 percent of its GDP on gross
for comparing living standards and examining productivity fixed capital formation; these figures are relatively low, com-
levels internationally over time. pared with the OECD as a whole. In terms of government spend-
PPPs also can be used to examine differences in the struc- ing per capita, Luxembourg, Germany, and Japan had the low-
ture of real expenditures across countries. Expenditure pat- est figures reported in the OECD (10.6 percent for Luxem-
terns for selected components of GDP are presented in table 3. bourg, 10.7 percent for Germany, and 10.9 percent for Ja-
To compute the structure of real expenditures, the data are pan). Predictably, government spending was higher among
converted to common currency using PPP, then the percent- the transition economies of Poland, Slovakia, the Czech Re-
ages are computed. More than two-thirds (70.1 percent) of public, Russia, and Hungary than that for the other participat-
the U.S. GDP is attributed to private consumption expendi- ing countries. Japan reported the highest level of private in-
tures, the second highest percentage of all of the participating vestment spending, by far, at 32.8 percent of GDP.
countries. The United States spent 12.4 percent of its GDP on In addition, the components of national expenditure have

10 Monthly Labor Review October 1999

different underlying price levels and can be disaggregated into Table 5. Country rank and index of benchmark
various personal consumption categories such as clothing, purchasing power parities for 1993 and 1996
education, and transportation, as well as government expen-
ditures, construction costs, and capital equipment purchases. Country Rank, Volume index, Rank, Volume index,
1993 1993 1996 1996
Table 4 presents comparative price levels for final expendi-
ture categories of GDP for the top 20 countries, based on GDP Luxembourg ............. 1 144 1 160
per capita ranking. Among the expenditure breakdowns, the United States ............ 2 128 2 140
Switzerland ............... 3 121 4 126
index for private final consumption of goods and services is Japan ....................... 4 107 5 121
the most interesting for bargain-hunting tourists. The index, Canada ..................... 5 102 8 114
Belgium .................... 5 102 9 112
however, covers a number of items, such as rent and medical Denmark .................. 7 101 7 117
care that reflect domestic demand, rather than the expendi- Austria ...................... 7 101 10 111
Norway ..................... 9 100 3 128
ture pattern of the average tourist. It would be necessary to France ...................... 10 99 14 103
adjust the weights to reflect the spending patterns of a casual
visitor to use the index for international bargain shopping. Iceland ...................... 10 99 6 118
Germany .................. 12 97 11 107
Netherlands .............. 13 93 13 106
Italy .......................... 14 92 15 102
Comparison of 1996 results Australia ................... 15 91 11 107
United Kingdom ........ 16 89 17 98
Every 3 years, the OECD-Eurostat PPP program estimates new Sweden .................... 17 88 16 100
Finland ..................... 18 82 18 96
benchmark PPPs. Although the 1996 results are generally con- New Zealand ............ 19 81 20 88
sistent with findings from previous PPP studies, the data from Ireland ...................... 20 73 19 92
the 1996 survey are not directly comparable with past sur- Spain ........................ 21 70 22 77
veys. The purpose of the PPP program is to compare the same Portugal .................... 22 62 23 70
Greece ..................... 23 57 24 67
market basket for several countries at the same time, hence Turkey ....................... 24 28 32 30
(1) (1)
some differences in results across surveys are to be expected. Israel ........................
(1) (1)
21 87
Slovak ...................... 24 67
Table 5 shows the ranking based on volume and the associ- Czech Republic ........ (1) (1)
26 64
(1) (1)
ated volume indexes for 1993 and 1996 for all countries par- Slovak Republic ........
(1) (1)
28 45
Hungary ................... 27 47
ticipating in the survey. The volume indexes are calculated Poland ...................... (1) (1)
30 35
(1) (1)
using PPPs as the conversion factor. The general reasons for Russian Federation ..
(1) (1)
31 34
Mexico ...................... 29 36
the differences in results between the two benchmarks are
1
changes in expenditure patterns, relative price structure, prod- Did not participate in the 1993 survey.
S OURCE: Key Indicators of the Labour Market, International Labor Orga-
uct lists, and methodology. Norway and Iceland had the larg- nization.
est upward movement in ranking from 1993 to 1996. Norway’s
gain was largely attributable to an increase in expenditures, troublesome comparisons from an international standpoint,
whereas the rise for Iceland was primarily due to a drop in in part, because of the vast differences in government subsi-
relative prices. It should be noted that domestic changes in dization of these services. Pricing high-technology goods, such
volumes, prices, and expenditures, as well as prices of other as computers, is complicated by the fast pace of technologi-
countries will affect the ranking. cal change which makes it difficult to identify and price simi-
lar products in countries that may be at differing levels of
development. Participants at the May 1998 meetings agreed
Future plans
to undertake a closer examination of these and other areas in
Even before the 1996 data were published, the participating an attempt to improve the comparisons.
countries began pricing for the 1999 benchmark study. The While representatives of the European countries partici-
pricing process includes updating the lists of product specifi- pate in regular meetings under the auspices of Eurostat, the
cations to ensure that representative items from all participat- May 1998 meetings provided a unique opportunity for repre-
ing countries are included in the exercise. In May 1998, the sentatives from the non-European countries of the study to
International Price Program of the Bureau of Labor Statistics exchange ideas and experiences. Future meetings of this work-
hosted a meeting of representatives from the U.S., Canada, ing group are planned to build on this progress with the next
Mexico, Australia, and the OECD to develop and discuss pos- scheduled meeting in November 1999.
sible updates to the product specifications. Many of the items
discussed will be included in the 1999 pricing round. Notes
Also discussed at these meetings were possible areas for 1
For example, PPPs are used in numerous publications, such as the World
methodological research that could be used in future bench- Economic Outlook, International Monetary Fund; Human Development Re-
mark studies. For example, education and health care are port, United Nations; World Development Indicators, World Bank; National

Monthly Labor Review October 1999 11

Purchasing Power Parity

3
Accounts and Main Economic Indicators, the OECD; and National Ac- These studies are published under the auspices of the United Na-
counts and Comparisons in Real Terms of the Aggregates of the Euro- tions Statistical Agencies, the University of Pennsylvania, Eurostat,
pean System of Economic Accounts, Eurostat. and the OECD . See Irving Kravis and Milton Gilbert An International
2
See Ian Castles, “Review of the OECD-Eurostat PPP Program” (Consult- comparison of National Products and the Purchasing Power Curren-
ant to the OECD, June 1997); and Jacob Ryten, “The evaluation of the cies (Organization for European Economic Cooperation, Paris, 1954).
International Comparison Project (ICP)” (Consultant to the International 4
Monetary Fund, Statistic Division of the Department of Economic and Gustav Cassell, “Abnormal Deviations in International Exchanges,”
Social Affairs of the United Nations Secretariat, and the World Bank, Economic Journal, December 1918, 413–15.
September 1998). 5
Catherine Mann, “Price, Profits, Profit Margains and Exchange rates,
“Federal Reserve Bulletin, 1986, pp. 366–79.

Where are you publishing your research?

The Monthly Labor Review will consider for publication studies of the labor force,
labor-management relations, business conditions, industry productivity, compensa-
tion, occupational safety and health, demographic trends, and other economic devel-
opments. Papers should be factual and analytical, not polemical in tone. Potential
articles should be mailed to: Editor-in-Chief, Monthly Labor Review, Bureau of La-
bor Statistics, Washington, DC 20212, or by e-mail to: mlr@bls.gov

12 Monthly Labor Review October 1999