Abdel Senhadji

In the vast empirical growth literature, research on the sources of economic growth, based on the growth accounting framework, has received particular attention since the widely publicized papers of Alwyn Young and Paul Krugman in the mid-1990s.¹ These authors generated a heated debate on the sources of real GDP growth in East Asian countries, by asserting that the “Asian Miracle” was a myth because the engine that drove the spectacular growth in the region came essentially from capital accumulation and not growth in total factor productivity (TFP). Why does the source of growth matter? The answer hinges on the important assumption of diminishing returns in physical capital in the neoclassical growth mode, which implies that capital accumulation cannot sustain long-term growth while TFP can. Thus, the source of growth is crucial to the long-term outlook of a country. This article provides a brief overview of recent IMF research on sources of growth using the growth accounting framework.

Since the mid-1990s, there has been an abundant literature extending Young’s growth accounting exercise to more countries and/or refining his methodology. The growth accounting framework was first presented in Solow (1957).² It is based on an aggregate production function and decomposes real GDP growth in terms of contributions from factor inputs and total factor productivity (TFP). Therefore, the results of the growth accounting exercise depend on the specification of the production function.

Several IMF studies have contributed to the debate on sources of growth in East Asia. Using internationally comparable data and factor shares, based on the industrial structure, and the level of development of economies, Sarel (1997) finds, contrary to Young, that the TFP growth has been impressive in Singapore, Thailand, and Malaysia.³ These results are corroborated by Crafts (1999), using different data sets, and Iwata, Khan, and Murao (2002), using nonparametric methods for estimating a very general production function.⁴

Numerous other studies have focused on other regions. Hu and Khan (1997) find that the contribution of TFP in China’s growth has been increasing, especially during the reform period.⁵ El-Erian and Helbling (1997) argue that growth in the Arab countries region was overly reliant on volatile external sources of funding and TFP growth was too low. The importance of market-friendly institutions are emphasized, in Dhonte, Bhattacharya, Yousef (2000), in absorbing the demographic explosion in the Middle East and sustaining adequate growth in GDP per capita.⁶ Ghura (1997) finds empirical evidence of increasing returns to scale for the aggregate production function for Cameroon stemming from positive externalities in physical and human capital.⁷ Hugo (1999) shows that despite a large investment-GDP ratio in Honduras, growth has been hampered by low TFP due to deficient levels of human capital and inadequate composition of investment.⁸ According to De Broek and Koen (2000), the severe contractions in Russia, the Baltics and the other countries of the former Soviet Union reflected not only collapsing investment and shrinking employment but also sharp declines in productivity because of the transition to a market economy.⁹ Even after correcting for factor utilization during the transition process, Dolinskaya (2001) finds a similar result for Russia.¹⁰ Cerisola and Chan-Lau (2000) find that investment-specific technical change is the major underlying cause of the pickup in productivity in Canada and narrowing of the productivity gap with the United States.¹¹ Calderon (2001) finds that the differential in productivity growth across OECD countries could be accounted for, to a large extent, by measurement errors.¹² Cardarelli (2002) shows that most of the divergence in income per capita between Australia and New Zealand is the result of lower capital accumulation, and to a lesser extent, lower TFP growth in New Zealand.¹³

While the basic growth accounting exercise which decomposes real GDP growth into contributions from factor inputs and TFP is a useful initial exercise, it does not, however, explain why certain countries enjoy faster TFP growth than others. Some of the empirical literature on sources of growth has explored the determinants of TFP. Bayoumi, Coe, and Helpman (1996) show that a country can raise its TFP not only by investing in research and development (R&D) but also by trading with countries with high levels of R&D.¹⁴ Similarly, Coe, Helpman, and Hoffmaister (1997) find empirical evidence to support that developing countries could boost their TFP growth through R&D spillovers by increasing trade with industrial countries and, in particular, by importing intermediate products and capital equipment embodying foreign knowledge.¹⁵ Hakura and Jaumotte (1999) find that this spillover is stronger through intra-rather than inter-industry trade because countries are likely to absorb foreign technologies more easily if they produce products similar to the imported goods.¹⁶ Senhadji (2000), using econometric estimates of factor shares for a large sample of countries, finds that initial conditions, institutional factors, and macroeconomic policy variables explain most of the cross-country differences in TFP levels.¹⁷ Salgado (2002) finds that the effects of structural reforms on TFP in the OECD countries are weakly positive, or even negative, in the short run but significantly positive in the long run.¹⁸ Jonsson and Subramanian (2001) find econometric evidence for a positive relationship between trade and TFP growth both over time and across sectors in South Africa.¹⁹

The growth accounting framework has also been used to compute potential output. Jonsson (2001) stresses the importance of TFP for sustaining growth in Thailand and recommends measures for stimulating TFP growth.²⁰ Ma (2001) argues that higher TFP growth from structural reforms is also essential in Korea in order to sustain growth close to historical levels.²¹ Senhadji (2002) develops a framework based on growth accounting for analyzing alternative scenarios of poverty reduction in Bangladesh.

More recently, the growth accounting methodology has been used to analyze the TFP acceleration in some countries, particularly in the United States, which enjoyed strong growth with low inflation and robust labor productivity growth during the second half of the 1990s. The acceleration in labor productivity in the United States—which coincided with increasing production and use of information and communication technologies (ICT)—has spurred a growing empirical literature aimed at identifying the sources of this acceleration, its longevity, and the potential for diffusion across sectors and across countries.

While the large number of empirical studies for the United States suggests that there is a positive link between investment in ICT and TFP growth, there is, however, no clear evidence that the acceleration is permanent. De Masi (2000) provides a survey of this literature.²² Acceleration of TFP and labor productivity growth in the second half of the 1990s has been more pronounced in the United States than in Europe, partly because the ICT sector share in GDP is generally lower in European countries. De Masi, Estevâo, and Kodres (2001) provide an excellent summary of the evidence for Canada, France, Germany, the United Kingdom and the United States.²³ In Finland, where the ICT sector plays a dominant role, Wagner (2001) documents a significant acceleration of TFP and a substantial contribution of the ICT sector to GDP growth.²⁴ Finally, Cardarelli (2001) shows that the rapid accumulation of ICT capital over the last two decades has played a significant role in explaining the impressive acceleration in productivity in Australia since 1995.