The title of the lecture is ”Trade, Markets and Economic Growth”. As a way of introduction, let me expand a little on the title.
Economic growth is of course an important subject. In my mind, it is the most important subject of economics as an academic discipline. To understand the causes of economic growth must be of the greatest importance, since it can help to design institutions and policies so that the resources available to us are used efficiently, both in each period in time and across time periods. This is not to say that economic growth should be as high as possible. Growth is not without cost; it requires giving up consumption today for consumption in the future.
Trade is commonly regarded as an engine of economic growth. I will attempt to explain in what ways trade is thought to promote growth in economic theory, and account for some of the empirical research that attempts to establish the extent to which trade promotes growth. I hope to make clear, that although we can list a number of ways in which trade is an engine of growth in theory, to find empirical support requires careful thinking and sophisticated methods, and the results are not always what we expect.
To put the concept of markets in the title is not as self-evident. I want to make the point that trade is an engine of economic growth, but only if markets work well. Markets cannot exist in a vacuum. They are always embedded in an institutional environment, which includes laws, regulations, policies, beliefs, organizations and other institutions that affect the behavior of economic agents. My point is that both trade and a well designed institutional infrastructure are necessary to achieve the desired rate of growth with the minimum of resources.
Some simple arithmetic
Relatively small differences in growth rates add up to large cumulative differences in levels of GDP. To design good institutions and policies is therefore of great importance for future generations.
Consider the following simple arithmetic. An economy whose GDP per capita grows at 1 per cent per year will double its GDP per capita in about 70 years, or in about three generations.
Assume instead that GDP per capita grows at 3 per cent per year, which happens to be slightly below the per capita growth rate of Sri Lanka during the three decades from 1970 to 2000. A 3 per cent growth rate doubles real income per capita in about 23 years, or in about one generation. Note that the example is in terms of GDP per capita. Sri Lanka’s growth rate has been about 5 per cent per year during the same period, with the difference made up of population growth.
Three per cent growth in GDP per capita and 5 per cent growth in GDP are historically very high rates. Looking at the experience of today’s most developed economies, we find that per capita growth rates during the period 1870, the beginning of industrialization in a majority of the countries, up to 1973, the year of the first oil crisis, generally was below 2 per cent per year. Only Japan, Sweden and Finland had per capita growth rates above 2 per cent.
But more recently, several economies have experienced much higher growth rates both in GDP and GDP per capita. GDP growth in Singapore, Taiwan, South Korea and Hong Kong – to take the most celebrated cases – has been in the range of 7-10 per cent for decades. With a growth rate of 7 per cent per year, GDP is doubled in about 10 years, and raised four times in 20 and eight times 30. With a 10 per cent growth rate, GDP is doubled in 7 years and raised four times in just 14 years.
The perhaps most striking example of transformation is that of Singapore. Its GDP per capita grew at more than 7 per cent in the three decades from 1965 to 1995. In that short time period, Singapore was transformed from being relatively poor, with a per capita income of about 2 300 US dollars to a per capita income of about 23 000 dollars. In other words, Singapore became a developed economy with a level of income similar to that of countries in Western Europe in a little more than one generation.
The same rapid transformation is now taking place in several other countries in Asia, with China and India as the most published examples.
All of the countries mentioned have been very open to international trade. That is frequently cited as an important, if not the most important factor, for their economic success.
Standard growth theory and growth accounting
The standard textbook model of economic growth posits that the level of GDP is determined by the amount of resources used for production and by the efficiency with which they are used. The term for efficiency of resource use is total factor productivity.
Since we ultimately are interested in per capita income and the consumption possibilities that individuals have, we divide GDP and the stocks of productive resources by the population and consider per capita GDP and per capita GDP growth instead of aggregate GDP and aggregate growth. Countries differ in GDP per capita because they have accumulated different amounts of factors of production per capita and because they have different total factor productivity. Factors of production include man-made resources in the form of physical capital – machines, infrastructure and so on – and human capital – accumulated through education and learning-by-doing – and natural resources.
GDP per capita in the standard model will eventually find an equilibrium, constant level, given the level of total factor productivity and the share of income allocated to investment. The reason is that capital is assumed to have diminishing productivity. In equilibrium, the return to investment in new capital is just sufficient to replace capital that becomes obsolete. Beyond this point, it does not pay to invest.
At the equilibrium level of the per capita capital stock, GDP per capita is constant. Further growth requires that total factor productivity increases.
Of course, if an economy is below the equilibrium level of the capital stock, it will grow both because of factor accumulation – the return to investment is sufficient to cover replacement of obsolete capital and additions to the capital stock – and because of increases in factor productivity.
We observe that economies with the highest levels of GDP per capita, such as Canada, the United States, Western Europe, Japan and Australia, grow at lower rates than countries with markedly lower levels of GDP per capita. It is common to assume, based on historical experience, that the “natural rate” of per capita growth in the most developed economies is about 2 per cent. This is the assumed long-run rate of productivity increase. We also observe that many less developed countries have much higher rates of growth of GDP per capita. Some, such as China, approach 10 per cent, which is extremely high in historical perspective. The observed difference in growth rates between the most developed and a number of developing countries can be explained by much higher rates of capital accumulation in the developing countries, which presumably are far below the equilibrium level of their capital stocks, and also have much higher rates of productivity increase.
How important is capital accumulation relative to productivity increases in explaining the actual growth record? This question is usually answered by so-called growth accounting. Growth accounting uses data on stocks of physical and human capital collected for each country. The stock of a particular kind of capital – including human capital – is calculated by adding up investments over a long period of time and allowing for depreciation. A weighted average growth rate of capital is then calculated using the income share of GDP of each kind of capital as weights. The weighted average of capital accumulation is always considerably lower than the recorded growth of GDP. On average, it can usually account for no more than about 60 per cent of GDP growth. The difference between recorded GDP growth and factor growth is attributed residually to growth in total factor productivity.
Within the framework of growth accounting, the level of a country’s GDP depends on its stock of physical and human capital and on its level of total factor productivity, and its growth rate of GDP depends on the weighted growth of capital stocks and factor productivity growth.
We can observe large measured differences between countries both in terms of the level of total factor productivity and in terms of its the rate of increase. The level is of course much lower in relatively poor countries almost by definition. What is interesting – and intriguing – is that it also differs very substantially between countries on about the same level of development. For example, the level of total factor productivity of Sweden was only about two thirds of that in the United States or Canada and only half of Hong Kong’s during the period 1960-1985.
We find even larger measured differences between countries on about the same level of development when it comes to rates of growth in total factor productivity. In fact, differences are so large that one must wonder whether they can be accurate. For example, for the period 1971 to 1995 the growth rate for the United States, Canada and Switzerland was about 0.5 per cent, while the United Kingdom’s was 1.3 per cent and Finland’s and Norway’s was higher than 2 per cent. Even after taking account of the fact that Finland and Norway were poorer than the United States in 1970, these are surprisingly large differences.
It is also interesting to note the surprising differences in the contribution of total factor productivity to growth in Singapore, South Korea, Taiwan and Hong Kong. Total factor productivity is calculated to have contributed almost nothing to Singapore’s growth at one extreme and to about one third of Hong Kong’s growth at the other extreme.
The growth accounting research leads us to conclude that, 1) growth in total factor productivity seems to explain at least 60 per cent of economic growth on average, and 2) it is important to understand what determines the growth of total factor productivity.
Learning and innovation
One determinant of total factor productivity growth is learning by doing. Production engineers employ so-called learning curves, which show the rate at which labor hours or cost fall with each additional unit produced. An example is the assembly of aircraft, where it is expected that a doubling of output results in a 15 per cent reduction in the number of working hours per unit, as workers learn how to use tools, managers to coordinate operations, and so on. Another often cited example in the literature is an iron-smelting plant in Sweden where productivity increased by 2 per cent per year over a 15-year period despite the fact that no new investments were made.
Another determinant of total factor productivity increase – probably more important than learning – is innovation. Most innovations are a result of investment in research and development, R&D. The amount of resources allocated to research and development varies greatly across countries. It is positively correlated with GDP per capita, but even among countries at about the same level of income, the variation is surprisingly large. If we subtract weapons and other defense related research and development, we find that among the G-7 countries, Japan tops the list. Its R&D spending relative to GDP is about 3 per cent. Japan is followed by France, Germany and the United States at about 2 per cent. Italy only spends 1 per cent on non-defense R&D. About 95 per cent of all research and development is carried out in a handful of developed countries.
The large differences in R&D spending relative to GDP between developed and developing countries and between developed countries are however not reflected in similar differences in productivity growth and in economic growth itself. For example, Italy’s growth rate does not lag behind Germany’s during the 1980’s and 1990’s, when Italy’s spending on R&D was less than half of Germany’s.
It seems clear that although R&D is the most important factor behind innovation, and innovation in turn is the most important factor behind long-term economic growth, the relation between an individual country’s spending on R&D and its growth is not particularly strong. This is where trade comes into the picture.
Trade and growth in theory
The standard textbook model of economic growth is one of a closed economy, having no trade. However, more recent models of economic growth – within the class of so-called endogenous growth models – often assign a crucial role to trade.
Before I outline the various ways in which trade affects economic growth in such models, let me just enumerate the positive effects on the level of GDP and welfare that trade is thought to have.
First, trade allows a more efficient use of resources through specialization according to comparative advantage. Such specialization raises the level of GDP. It should be noted that geographically small countries can be expected to gain more than geographically large countries, since their resource endowments should deviate more from the world average composition of resources.
Second, trade means access to a greatly extended market and thereby makes possible to take greater advantage of increasing returns to scale in production.
Third, trade exposes producers to international competition, forcing them to reduce slack and lower prices.
Fourth, trade gives consumers and producers access to a wider variety of goods. Greater choice is a clear gain for consumers and can increase productivity for producers by providing them with more specialized and more appropriate inputs.
All of these effects are static; they will raise GDP and consumer welfare once and for all. They do not increase the rate of growth of GDP. However, a country that reduces trade barriers can expect to reap the gains over a relatively long period of time. Essentially static gains from trade will then be registered as an increase in the rate of growth during the transition.
So what are the ways in which trade is thought to increase the rate of growth? More specifically, in what ways can trade raise productivity growth?
One clue is provided by the substantial differences in R&D spending across countries that I gave examples of before. I made a point of the fact that although Germany had spent more than double as much resources on R&D relative to GDP than Italy during two decades, their growth rates had been similar. This can be explained by free information flows and free trade; these have allowed Italy to take advantage of learning and innovation that has taken place in Germany and elsewhere.
In a world economy where learning and innovation are freely available to all in all countries, it would not matter where learning and innovation takes place. Every country would benefit to the same degree, and per capita growth rates would tend to converge to the same rate. However, learning and innovation are not costless even if they are freely available. Some learning is hard to transfer from brain to brain, and many innovations are protected by patents. Moreover, even if in the cases where innovations are freely available, to find, transfer and implement relevant knowledge is costly.
Research and development are likely to give rise to significant externalities or spillovers; it is impossible to entirely keep new knowledge secret or to protect its use by others by legal means. Such spillovers are not contained within national borders. The spillovers from R&D can be exploited directly by competitors. They also add to an ever-expanding and freely available pool of knowledge, and makes new research and development and new innovation less costly. This virtuous loop of knowledge spillovers from research and development onto future research and development is a crucial element in most endogenous growth models.
Another way in which trade promotes growth is by providing a larger market and thereby greater incentives for spending resources on innovation. Holding everything else equal, a larger market potentially means greater sales and therefore greater profits. This is of special significance for small countries, which otherwise would be at a serious disadvantage relative to large countries with their large domestic markets.
The effect of a larger market is not necessarily positive for innovation. It also brings more intense competition, which tends to depress profits and thereby the diminish the willingness invest in R&D. The effect of the degree of competition on innovation is a long-standing issue in the economics of innovation, both theoretically and empirically. At the same time as increased competition can reduce the incentives to innovate, it can force producers to innovate in order to stay in the marketplace and to increase profits by achieving temporary monopoly positions. Empirical research does not give a clear-cut answer to the question of whether increased competition causes more or less innovation.
Finally, the creation of a world market through trade tends to eliminate redundancy in research and development. Worldwide competition forces producers to specialize and to concentrate their efforts in areas and on products where they think that they can obtain a monopoly position. Without trade and international competition we would probably see much more duplication of research and development.
Empirical evidence on the effect of trade volumes on growth
So much for theory. Can we find any empirical evidence for positive effects of trade on either the level or the rate of growth of GDP?
It may seem that this is a question with an obvious answer. Looking around, we can observe many countries, both in the past and in the present, which have had very high growth rates at the same time as they have had open and export-oriented trade regimes. Is it not self-evident that, for example, Hong Kong or Singapore have achieved their economic success due to trade?
One needs to be a bit careful when assuming, first, that there must be a clear and positive relation between the degree of openness, measured as exports and imports relative to GDP, and the rate of growth, and, second, that trade causes growth.
First, the degree of openness of an economy depends on its resource endowments, technologies, preferences and market structure. A geographically large country tends to have resource endowments that are closer to the world average than a small country, and should therefore be less open. If consumers in some country have preferences that are very biased towards domestic goods and services, the country will be less open than a country where consumers are biased towards foreign goods and services.
Second, trade is the difference between domestic supply and demand. Domestic supply and demand are affected by a host of features of the economy, including economic policy, as well as other, more exogenous features, such as history, geography and resource endowments. In economics terminology, trade is very much an endogenous variable. The direction of causality between trade and growth is therefore not certain. Growth may cause trade, or trade may cause growth, or both may be caused by some other, exogenous factor.
When economists first tried to find empirical evidence for the effect of trade on growth, they took cross-section data for large samples of countries over some time period, and explained growth by openness to trade and a great number of other factors using statistical methods. The estimated effect for trade was usually statistically significant and positive, but only when investment in physical capital was left out as an explanatory variable. When investment was included, there was no significant effect of trade. This led some to conclude that investment was the driving factor; if investment was made to increase, it would bring with it an increase in trade and also in growth. Others concluded that trade only had an indirect effect on growth; it served to increase investment and through its effect on investment it increased growth.
More sophisticated methods are required to overcome the endogenous nature of trade and to sort out causality. One such method was employed by two American researchers, Jeff Frankel and Paul Romer. In a first step, they estimated the amount of trade that each country in a large sample of countries should have on the basis of geographical factors, including distance to other countries. Geographical factors are basically not affected by the economy. The predicted trade flows were then used to estimate the effect of trade as a share of GDP on GDP per capita. In statistics terminology, the predicted trade flows were instrumental variables for trade.
Frankel and Romer used two samples, one with 98 countries and one with 150 countries. They found a strong effect of the trade share – imports plus exports relative to GDP – on GDP per capita; an increase of 1 per cent in the trade share led to a 2 per cent increase in GDP per capita. They also found that most of the effect of trade on per capita income was due to increased total factor productivity. Only a smaller part was due to investment in physical and human capital.
Other researchers subsequently used the same method employed by Frankel and Romer to estimate the effect of openness – the trade share – on the rate of growth of GDP per capita. They found that openness and also population size as a measure of country size have a positive effect, and that openness and size interact with one another. The effect of trade on per capita income growth is particularly strong for small countries and ceases in large countries the size of France. These results confirm the theoretical predictions that 1) market size is an important factor for innovation and 2) that small countries in particular gain from spillovers of learning and innovation through trade.
A note of caution is in place. The variation in openness and in rates of growth is great across countries. The results cannot be applied to a particular country to say that a given increase in trade would result in some per cent increase in the level and growth of GDP per capita; they are averages for a large set of countries. Also, we need more studies, using other methods and other data sets, before we can say that we have robust empirical support for positive effects of trade on growth.
The effects of trade policy on growth
Attempts to establish the influence of trade policy on per capita growth involve even greater difficulties than attempts to establish the effects of trade itself.
First, trade policy is usually part of a more general policy stance; countries with restrictive trade policies tend to have restrictive policies in other areas as well, and also a relatively extensive government involvement in the economy. If we try to explain the effect of trade policy on growth using cross-country data, we have to control for other policies, as well as a host of other factors as well. This is hard, if not impossible, to do.
Second, trade policy is usually quite complex; it can include a great number of measures to restrict and regulate trade. It is difficult to capture the complexity in one single measure.
Third, it is doubtful whether we can interpret the statistical results of a cross-country study. Assume hypothetically that trade policy has been put in place for one of two reasons. In the first case, the government was concerned about the bad state of the economy and thought that trade policy could improve the situation. In the second case, the government wanted to use trade policy to distribute quasi-rents and favors to importers and import-competing industries. In either case, a cross-country study would show a negative effect on trade policy on growth, but we would not know the underlying mechanism. In the first case, we cannot necessarily conclude that trade policy is bad for growth, while in the second case that is more likely.
Despite the great difficulties in trying to establish the effects of trade policy, one well-known empirical study should be mentioned, by the American economists Jeffrey Sachs and Andrew Warner. The study used a binary measure of trade policy restrictiveness, based on a number of indicators, which of course constitutes a very crude representation of trade policy. Sachs and Warner found that economies that were open – the trade policy variable had a value of 1– on average had a growth a rate that was 2.5 percentage points higher than closed economies – for which the variable had a value of 0.
The Sachs-Warner study has been criticized by many for some of the reasons that I have mentioned, but other, more sophisticated studies are affected by the same basic problems. All of them show that trade policy has negative effects on growth. Most economists would agree with the findings, including myself, but so far our conviction lacks robust empirical support.
International spillovers of learning and innovation
I stressed earlier that more than half of the international variation in the level of GDP per capita is due to variation in total factor productivity and that considerably more than half of the variation in the rate of growth of GDP per capita is due to variation in total factor productivity growth. Total factor productivity is determined by learning and by innovation, and innovation is in turn basically determined by spending on research and development.
The empirical research trying to establish a causal link between trade and growth is based on the assumption that trade plays an important role for the transfer of knowledge across national borders. The reasoning is that the goods themselves and the information flows surrounding goods transactions are channels through which knowledge is transferred.
Some research attempts to find a more direct link between R&D spending abroad and the domestic level of total factor productivity. A notable study was done by David Coe and Elhanan Helpman. They constructed data on the stocks of research and development in 22 developed countries by adding yearly spending over time and allowing for depreciation. The variation in R&D stocks was found to be very great; the average ratio between two countries was about 1 to 10. For every country a stock of foreign R&D stock was calculated by constructing a weighted average of the stocks of its trade partners, using the shares of total trade with each partner as weights. This means, for example, that a country that had a large share of its trade with the United States would tend to have a large foreign R&D stock.
Total factor productivity of the 22 countries was then regressed on domestic and foreign R&D stocks, and also on openness and on a combination of openness and the foreign R&D stock. Coe and Helpman could explain about 60 per cent of the variation in the level of total factor productivity in this way. They found that more open economies and economies with larger foreign R&D stocks had higher levels of total factor productivity.
The same method was later applied to a set of 77 developing countries. These countries had very little R&D spending themselves – as I mentioned earlier, 95 per cent of all R&D spending is done in a handful of the most developed countries. It was found that these developing countries benefited substantially from R&D carried out in the developed countries; foreign R&D stocks could explain about 20 per cent of the variation in total factor productivity of the developing countries.
Knowledge is transmitted internationally through other channels than trade, notably through foreign direct investment. A huge body of empirical research has established that foreign direct investment has significant positive spillovers in host countries. Wolfgang Keller, a German economist, has tried to decompose the international transmission of research and development into three parts: trade, foreign direct investment and language skills. Surveys of why firms choose one country to invest in over others have shown that language is an important factor. Keller found that about 70 per cent was due to trade, 15 per cent to foreign direct investment and 15 per cent to language skills.
It seems clear that countries with little spending on research and development benefit substantially from R&D in a small number of developed countries. The benefit is not confined to total factor productivity and GDP. R&D in developed countries also benefits consumers in other countries directly, by providing them with better and cheaper products.
However, it also seems clear that the greatest benefit accrues to those countries where R&D is done. In other words, the very uneven distribution of R&D tends to widen the income gap between developed and developing countries.
Markets and institutions
Even after we account for differences in accumulated physical and human capital and other resources, and for differences in total factor productivity, substantial differences across countries in the level of GDP per capita and in its rate of growth still exist. How can these remaining differences be explained?
Markets do not operate in a vacuum. They are imbedded in an institutional environment of laws and a judiciary, regulations and regulatory bodies, policies, government and private organizations, beliefs and culture. Economic historians and economists who have studied economic development and growth over longer periods of time are in widespread agreement that if markets are to function well, they have to be supported by a good institutional environment. In particular, they stress the importance of how institutions affect incentives that determine the willingness to invest in physical and human capital and in knowledge-creation. Institutions can encourage or discourage such investment, and they can give stronger or weaker incentives for rent-seeking and outright corruption.
Perhaps the most influential proponent of this view has been economic historian Douglas North, Nobel Prize Winner in 1993. He has consistently emphasized the importance of institutions that define, protect and enforce property rights as a condition for well functioning markets and the willingness to undertake investments in physical and human capital and in making innovations. He argues that it was the formation of such institutions that was the decisive factor behind the Industrial Revolution.
This is not to say that there is a unique, optimal set of institutions that all countries should adopt and that is appropriate for all times. What works well in a particular country and time probably depends to a large degree on that country’s history, culture and existing stage of development. Any program of institutional reform has to take account of the particular circumstances.
Moreover, institutions are not solely a matter of form. It is not sufficient to adopt a particular set of laws and regulations, or a particular form of organization, for the right outcomes to be assured. The recent experience of Russia’s reforms is a case in point. Property rights are protected by law, but expropriation in various form nevertheless takes place, and discourages investment. On the other hand, until recently, private ownership of firms did not exist in law in China, but the amount of investment has nevertheless been enormous. Investors in China seem to feel quite secure, probably because local public and political organizations have a stake in the return to those investments. What matters in the end, is the security that the individual or firm feel in undertaking investments. Sometimes, only a small change is required, such as the attitude of the government, or the change of a particular policy or regulation.
The empirical research trying to find regularities by statistical methods in the relation between institutions and economic growth is fairly recent. One problem has been to isolate the institutional framework from other factors, such as geography and climate. Another has been to try to determine what set of institutions are better at achieving a desired set of outcomes.
The issue of whether climate or geography is an important determinant of economic development is long-standing. Recent research, using large samples of developing countries, seem to show that climate or geography has had an influence on institutions and through them on present GDP levels, but not a direct and independent effect.
The proposed mechanism is the following. Some regions had a favorable climate and geography for European settlement. They were not unusually dangerous to the health of settlers, and they were mostly sparsely populated and had land that was suitable for cultivating wheat and other grains. In regions where large numbers of Europeans settled, they established institutions modeled on those in the countries from where they came and which were conducive to economic development. Other regions were more adverse for the health of Europeans and were more suitable for growing valuable cash crops, such as coffee and sugar, or for the extraction of minerals. Relatively few Europeans settled in these regions, and the institutions were designed to protect the ruling elite and create privileges. Such institutions provided weak protection of property rights and they were not as conducive to industrialization. In other words, climate and geography had an influence, but only through institutions, not independently.
I should add that many countries were not colonized by Europeans, and that the variation in present per capita income is almost as great among these as among the countries that are former colonies. Colonization produced institutions that were better or worse for economic development, but so did conditions in the countries that were not colonized.
So much for the origins of institutions. An example of the research trying to find differences in the ability of different institutions to promote growth and development, is the research on whether the English common-law or the French civil-law system is better in supporting investors and markets. The common-law system uses broad principles that have been established over time through court decisions and oral argument, while the latter uses legal codes and written records. Empirical research on large samples of countries indicates that the common-law system is better at protecting and enforcing investor rights, at resolving disputes and regulating the start-up of new firms.
Let me summarize. I have tried to give a brief account of what economic theory and empirical research can say about the determinants of economic growth and the large differences that exist across countries in levels and rate of change of GDP per capita. It is worth to emphasize that if we can learn enough to raise the rate of growth even by a fraction of one per cent, the cumulative effects over time would be very large.
The standard, textbook model of economic growth states that growth in per capita output is caused by accumulation of physical and human capital and by increasing their productivity. The productivity of capital is assumed to be falling as more capital is accumulated. An equilibrium will be reached, where the return to investments is only sufficient to replace capital that becomes obsolete and where the capital stock ceases to grow. Further growth in per capita GDP can only result from a continuous increase in the productivity of physical and human capital, what is more generally called technological progress. Countries that have not yet reached the equilibrium level of capital stocks are accumulating capital and can grow at higher rates.
The great variation in GDP levels that are observed across countries, can be explained by differences in the stocks of physical and human capital and in total factor productivity. Total factor productivity is the most important cause of variation; it explains more than half of the variation in levels of GDP per capita. The great variation in GDP per capita growth that can be oberved is explained by differences in the rate of capital accumulation and differences in rates of total factor productivity growth.
Total factor productivity growth is driven by innovations and innovations are made through research and development. Newer theory make innovations endogenous. One version of the theory posits that research and development generates knowledge that adds to the common, freely available knowledge base, which lowers the cost of subsequent research and development and therefore starts additional investment in R&D, and so on, in a virtuous loop. The results of one firm’s R&D benefits both domestic and foreign firms. One channel through which new knowledge is transmitted is trade. Also, R&D gives us better and cheaper products and benefits consumers and producers in all countries.
Empirical research confirms that trade has a positive effect both on the level and growth rate of GDP. Research also shows that countries benefit directly from research and development that their trading partners undertake, and that the benefits increase with more openness. This is particularly true for smaller countries.
Differences in physical and human capital per capita and in total factor productivity and international transmission of knowledge can still not account for all of the variation in levels and rates of growth of GDP per capita. Much variation remains.
There is considerable agreement among researchers that free trade by itself is not sufficient to generate high rates of growth. Well functioning markets are also required. For markets to work well, they have to be embedded in a favorable institutional environment – a set of laws, rules, organizations and policies – that define, protect and enforce property rights so that individuals and firms are encouraged to undertake investment in physical and human capital and in research and development, and are discouraged to engage in rent-seeking and other unproductive activities.
1 Public lecture at the Centre for Banking Studies of the Central Bank of Sri Lanka, July 27, 2006. The lecture draws on Elhanan Helpman, “The Mystery of Economic Growth,” MIT Press, 2004, which includes other topics as well.