Get access



  • Note: We are grateful to Paul Rhode and John Devereux for pointing us to useful data sources. Our special thanks go to Angus Maddison for his careful reading and discussion of our various drafts. Our paper benefited from comments by Stephen Broadberry, Leandro Prados de la Escosura, Riitta Hjerppe, Yumiko Moriyama, Hak Kil Pyo, Jean Pascal Bassino, Prema-Chandra Athukorala, Konosuke Odaka and the two referees and the editor of this Journal; and the participants of the UC Davis workshop, June 1, 2005; the Vice-Presidential Session of the International Economic History Congress at Helsinki, August, 2006; and seminars at the London School of Economics and Oxford University. This paper is partly funded by the NSF Global Prices and Income 1350-1950 project headed by Peter Lindert at UC Davis and the Ministry of Education 21st Century COE Program, Research Unit for Statistical Analysis in Social Sciences, headed by Osamu Saito at Hitotsubashi University. All errors are the responsibility of the authors.

*Debin Ma, Economic History Department, London School of Economics, Houghton Street, London WC2A 2AE, U.K. ( or


This article provides estimates of purchasing power parity (PPP) converters for expenditure side GDP of Japan/China, Japan/U.S. and China/U.S. in 1934–36 through a detailed matching of prices for more than 50 types of goods and services in private consumption and about 20 items or sectors for investment and government expenditure. Linking with the earlier studies on the price levels of Taiwan and Korea relative to Japan, we derive the mid-1930s benchmark PPP adjusted per capita income of Japan, China, Taiwan and Korea at 32, 11, 23, and 12 percent of the U.S. level respectively. These estimates correct the consistent downward bias in East Asian income levels based on market exchange rate conversions. Compared with Angus Maddison's estimates based on the 1990 benchmark back-projection, our current-price based result are 18 and 44 percent lower for Japan and Korea, and 4 and 10 percent higher for Taiwan and China respectively in the mid-1930s. We develop a preliminary theoretical and empirical framework to examine the possible source of the biases in the back-projection method. The article ends with a discussion on historical implications of our findings on the initial conditions and long-term growth dynamics in East Asia.