Between 1990 and 1998, the American economy grew at an average annual rate of over 3%, increasing to over 4% since 1998, and peaking with a record yearly growth rate of 5.2% in the first semester of 2000. The recent downturns in the US, notably the collapse of the Nasdaq index and the closures of a substantial number of 'dotcoms', will certainly result in substantially slower growth. Yet the expectations are for a 'soft landing', with predicted growth rates remaining above those in Europe over the next decade. Even if the US has experienced a speculative bubble in high-tech stocks, a comparison with the East Asian crisis of 1997 shows important differences: there has been little speculation in the housing market; while private sector firms have a debt level which is much lower and, crucially, is denominated in the domestic currency. This means that the American economy is in a better position to overcome the current turbulence in the stock market and avoid a major recession.

The French economy has followed a growth pattern similar to that in the US. GDP growth has increased from an annual average of 1.4% in the period 1990-98, to 2.9% in 1999 and 2.5% in the first semester of 2000. What is surprising in these figures is not so much the similarity in trends but rather the large difference in levels: over the last decade the US will have grown twice as fast as France. Why did this happen?

Part of the answer can be found in the different evolution of employment in the two countries. The rate of output growth is, by definition, the sum of labour productivity and the rate of growth of employment, and French employment levels have been dwindling for the past decade. Over the period 1990-95, private-sector employment grew at an annual rate of 1.3% in the US, while it fell by 0.7% in France. Although the negative trend reversed in the next five years, an annual growth rate of 0.8% seems poor when confronted with the 2.4% experienced in the US. Yet the second half of the 1990s has also witnessed a mediocre performance of French productivity growth. Although output per worker grew faster in France than in the US over the period 1990-95 (1.6% against 1.2% per annum), the next three years saw an acceleration to 2.1% in the US and no change in France.

This acceleration of US productivity growth has coincided with the start of the Information Technology (IT) boom. France lags well behind the US in this respect, as indicated by the low levels of R&D expenditure (particularly in the private sector) and of IT investment. Private-sector R&D expenditures represent only 1.37% of GDP in France, compared with 2.08% in the US. The differences are even more marked when looking at IT investment as a share of gross fixed capital formation – OECD research suggests figures of 6% and 13.4% in France and the US, respectively. Moreover, in the period 1990-96, expenditures in high-technology equipment grew twice as fast in the US as in France. All this seems to indicate that both the creation and the diffusion of new technologies is taking place at a much slower rate in France.

Modern growth theory argues that the main engine for growth in mature, industrial countries is not the accumulation of physical capital but innovation, particularly intentional innovation carried out by firms. This has shifted attention away from households’ saving decisions and into the institutional environment that can provide (or not) the incentives to innovate. The institutional framework becomes particularly important after the arrival of a new 'general purpose technology' (GPT). A GPT is a technological invention (or breakthrough) that affects the entire economic system, but which needs to be adapted to particular industries through further research and development. The new information technologies are such a general purpose technology. Two sets of institutions seem to play a major role in allowing countries to fully exploit IT: those that affect the incentives to innovate, and those that create the possibility of innovating and implementing new technologies.

A key aspect affecting firms' incentives to innovate is the extent to which intellectual property rights are protected. Agents will only be willing to invest in innovation if they know that once they have created a new product or process they can reap the rewards. For this to occur, patent legislation has to provide sufficient protection to innovators. At the beginning of the 1980s, the US approved a number of laws designed to reinforce intellectual property rights both domestically and abroad (most notably the Bayh-Dole Act) and created a federal Appeal Court dealing with patent-related law suits. As a result, the number of resident patent applications by US firms grew by 91% between 1981 and 1997 compared with only 22% in France over the same period. Furthermore, the rate of growth in the number of American patents purchased by French firms lags behind the average for the EU (9.6% compared with 10.2 over 1990-96).

A second factor affecting the incentives to innovate is the degree of product market competition. Contrary to the Schumpeterian argument that monopoly is necessary in order to allow a firm to recoup the fixed costs incurred when investing in R&D, recent empirical and theoretical work emphasizes that competition is more likely to foster the creation and implementation of new technologies. The basic argument is that under strong competitive pressure, firms are forced to adopt the latest technologies in order to survive, and to create ever newer technologies if they want to stay ahead of their competitors. France also lags behind the US in this respect. The public sector is three times larger, and, more importantly, the barriers to trade and foreign investment are much higher. A recent study by the OECD ranks countries according to the degree of barriers to entry. The UK exhibits the lowest barriers, with an indicator of 0.43, while the US has an index of 0.87, and France of 1.03. In other words, the threat of entry by possible foreign competitors is substantially weaker in France than in the US, thus reducing the incentives of French firms to adopt new technologies.

France seems to have another major handicap due to the difficulty in creating new enterprises: a recent NBER working paper suggests that the average number of working days needed to complete the administrative procedures in order to establish a new firm is 66 in France, and only seven in the US. Entry by competitors is consequently slower, and hence poses less of a threat to established firms. Administrative barriers to entry are also problematic because new firms are likely to adopt the latest technologies (as there is no opportunity cost of adoption) and are therefore a source of technological change. For example, research shows that in the US more than 50% of trials on new pharmaceutical drugs are done by newly created biotech firms, while in Europe 90% of the new trials are done by established firms. This shows the importance of entry in the recent technological revolution in the US, and the high cost that such barriers may have in terms of forgone technological possibilities.

Together with the administrative barriers to the entry of new firms, access to capital markets is a crucial factor affecting the incentives and the possibilities to innovate. R&D activities and the implementation of new technologies are, by nature, risky. Innovating firms often find it difficult to obtain loans and must therefore raise capital in the stock market. The US lies well ahead of France in this respect, despite the substantial improvements that the latter has experienced in the last three years. In 1999, market capitalization (as a share of GDP) was over 55% larger in the US than in France. Meanwhile, venture capital represented only 0.065% of GDP in France, compared with 0.16% in the US.

Finally, the creation and the implementation of new technologies requires a sufficiently qualified labour force. The fraction of the labour force that holds university degrees is 20.6% and 34.9% in France and the US, respectively. Moreover, the number of scientists and engineers per thousand employees is one-and-a-half times larger in the US. Such differences can be partly explained by the education system, in particular the greater emphasis placed by American universities on entrepreneurship and business-oriented education, but it seems that migration has also been an important factor. Between 1992 and 1997 the number of qualified workers entering the country grew by 45% in the US, and by only 16% in France. This is partly because of the lack of de facto mobility of workers within the Schengen area as well as the constraints imposed by French legislation on non-EU immigrants. For example, the average duration of working visas extended to skilled workers is only nine months in France compared with three years in the US.

In order to benefit fully from the IT revolution and experience an acceleration in growth rates, France will require substantial institutional changes. This does not mean that US institutions should be reproduced. A crucial issue is whether such changes would necessarily be accompanied by an increase in income inequality. The challenge for policy-makers is to develop innovation, immigration, and growth-oriented policies that can be compatible with an improvement (or at least no deterioration) in the distribution of income. Recent research has shown that a substantial fraction of the increase in wage inequality induced by the technological revolution is within educational groups, and that this fraction affects the temporary component of income inequality. In contrast, inequality between educational groups affects the permanent component of income inequality. Hence, a policy that would increase the innovation rate while reducing skill differentials (or educational differences) among individuals would meet the challenge of increasing productivity growth without increasing the permanent component of income inequality.

The above article was written by Philippe Aghion (Harvard University, University College London and CEPR), Eve Caroli (Laboratoire d'Economie Appliquée (INRA) and CEPREMAP) and Cecilia García-Peñalosa (GREQAM and CNRS).