Penn tables

The World Economy, 1950–2001

HS–7: The World Economy, 1950–2001

Tables 7a–7c show annual estimates of economic activity in 7 regions and the world for the year 1900, and annually 1950–2001. They aggregate the detailed estimates by country in HS–1 to HS–6 and there are analytical tables showing percentage year–to–year movement in real terms. Three basic ingredients are necessary for theseestimates. These are time series on population which we have for 221 countries, time series showing the volume movement in GDP in constant national prices for 179 countries, and purchasing power converters for 99.3 per cent of world GDP in our benchmark year 1990. With these converters we can transform the GDP volume measures into comparable estimates of GDP level across countries for every yearbetween 1950 and 2001. For countries where all three types of measure are available, the estimates of per capita GDP level are derivative. However, to arrive at a comprehensive world total, we need proxy measures of GDP movement for 42 countries, and proxy per capita GDP levels for 48 countries for the year 1990. These proxies collectively represent less than 1 per cent of world output.

a) WorldPopulation Movement 1950–2003
There are two comprehensive and detailed estimates of world population which are regularly updated and revised. They both provide annual estimates back to 1950 and projections 50 years into the future. No other source provides such comprehensive detail, length of perspective, or causal analysis of birth and death rates, fertility and migration. Here I have used thelatest (October 2002) estimates of the US Bureau of the Census (USBC) for all countries except China, India and Indonesia. In Maddison (2001), I used the USBC 1999 version for 178 countries, OECD sources for 20 countries and Soviet sources for 15 countries. USBC estimates are available at http://www.census.gov/ipc. The United Nations Population Division (UNPD) is the alternative. Its latestestimates, World Population Prospects: The 2000 Revision, were prepared in February 2001; the previous version was issued in 1998. The UN shows estimates for quinquennial intervals, but annual country detail is available for purchase on a CD ROM. Table 7a shows my estimates, based mainly on USBC. Table 7* shows UNPD figures with the same regional breakdown. The easiest way to compare the two sources isthe ratio of the two alternatives shown in Table 7**. On the world level the differences are minimal, and the regional differences are not very large after 1973. On the country level there are larger differences between the two sources. These are biggest for small countries and the UNPD omits Taiwan. Both sources provide a similar long–term view, showing the fastest demographic momentum in Africaand a general reduction in the pace of growth in the 1990s. Differences are mainly due to use of different sources or conjectures for cases where evidence is poor. It is clear from inspection of the country detail that the USBC takes better account of short–term interruptions due to war, flight of population or natural disasters. Their impact is smoothed by UNPD interpolation between censusintervals. One example is the genocide and exodus from Rwanda: USBC shows a 25 per cent fall of population in 1993–5, UNPD 9 per cent. USBC shows a 55 per cent fall in Kuwait in 1991 during the Gulf war, UNPD 2 per cent. USBC shows a 70 per cent fall in Montserrat in 1998 due to volcanic activity, UNPD tapers this decline over several years. In fact, a major objective of the UN is to provide alternativeprojections of population trends which are of fundamental importance in assessing prospects for its development programmes. USBC is probably more interested in monitoring past and present performance.
225

The World Economy: Historical Statistics

b) Movement in Volume of GDP 1950–2001
Table 7–1 shows the coverage of our GDP estimates for five benchmark years since 1820. For 2001 there...