TRANSITION REPORT 2014 Innovation in Transition
TRANSITION REPORT 2014 Innovation in Transition
Alexander Plekhanov is a Lead Economist at the EBRD Office of the Chief Economist. He has worked on global macroeconomic issues and as a country economist covering a number of countries, including Belarus, Kazakhstan, Mongolia and Russia. Prior to joining the Bank, Alexander worked as an economist at the International Monetary Fund in the Western Hemisphere and Fiscal Affairs Departments.
Innovation is an important driver of improvements in productivity. But what drives innovation itself? This chapter looks at the reasons for the significant variation seen in the rates of innovation of individual countries and sectors, as documented in Chapter 1.
Various factors influence firms’ incentives and ability to innovate, ranging from the prevalence of corruption to the availability of an adequately skilled workforce and access to finance. Some of these factors are internal, reflecting either characteristics of the firm (its size or age, for instance) or decisions made by the firm (such as the decision to compete in international markets or the decision to hire highly skilled personnel). Other factors are external and shape the general business environment in which firms operate (such as customs and trade regulations).
In some cases, the two are closely related: each firm makes personnel decisions that determine its ability to innovate, but these decisions are, in turn, strongly influenced by the prevailing skills mix and the availability of a sufficiently educated workforce in the region where the firm operates. Similarly, Chapter 4 shows that the local banking structure (an element of the external environment) has an impact on firms’ funding structures (an internal aspect), which then affects innovation. Even if firms share the same business environment, they will not necessarily make the same business decisions, and these decisions will influence their innovation activity.
A firm’s willingness and ability to innovate will depend on various characteristics. In particular, young, small firms are often perceived to be the main drivers of innovation. While such firms do make an important contribution to the development of new products, they are not necessarily more innovative than other firms when viewed as a whole.
This is partly because when young, innovative firms are successful, they often grow fast, thereby becoming larger firms. Google and Amazon were once start-ups with just a handful of employees, but they have quickly grown and now employ thousands of people. Innovative start-ups that are not successful, on the other hand, typically run out of funding and exit the market.1 Neither of these types of firm will be categorised as young, small firms in an enterprise survey such as BEEPS V or MENA ES. In addition, not all young, small firms are innovative start-ups. Many will be in conventional service sectors (takeaway restaurants or small convenience stores, for instance).
Young, small firms may tend to innovate less, but start-ups still represent a very important class of innovators. They are the firms that are most likely to come up with innovations that are new to the global market. In some cases, the innovation is the sole reason for the firm’s creation.
In Israel, two-thirds of small firms introduced product innovations that were new to the international market, compared with 48 per cent for larger firms (see Chart 3.3). Moreover, all young firms (defined as companies that were established less than five years ago) introduced at least one new product that was new to the international market, hence the fact that Israel’s start-ups have a reputation as one of the key drivers of economic growth in that country.
Another important characteristic affecting innovation is the type of firm ownership. In general, foreign ownership and the integration of local firms into global supply chains are expected to lead to increased innovation (see Box 3.2). On the other hand, concerns are sometimes raised that multinational companies may conduct all of their R&D activities in their home countries, outsourcing only lower-value-added activities to emerging markets, so foreign takeovers may actually result in reduced spending on R&D.6
In addition to firm-level characteristics such as a firm’s age, size and ownership structure, various decisions made by firms are related to their incentives and ability to innovate. One such decision is whether to compete in international markets.
Firms that export their goods are able to spread the fixed costs of innovation over a larger customer base, so exports can support innovation. By the same token, firms in larger economies with larger domestic markets may find it easier to innovate on account of higher levels of domestic demand for new products.
Another important decision that a firm faces is whether to spend on R&D to support the development of new products. As discussed in Chapter 1, R&D is not a prerequisite for the introduction of new products or processes, as firms may decide to acquire existing knowledge from elsewhere.
At the same time, R&D significantly increases the likelihood of successful innovation. Firms that invest in R&D are an average of 22 percentage points more likely to introduce new products or processes.12 They are also an average of 20 percentage points more likely to introduce marketing or organisational innovations (perhaps because these types of innovation often go hand in hand with technological innovation).
A suitably skilled workforce (including strong management skills) is one of the key prerequisites for successful innovation – both innovation at the technological frontier and the adoption of existing technology – as workers are required to develop and learn new production techniques.13
The results in Table 3.1 suggest that while the percentage of employees with a university degree affects the probability of introducing a new product or process and the likelihood of investing in R&D, this impact is fairly small relative to the effect of other firm-level characteristics discussed above. The regression analysis already accounts for the differences between the skill intensities of the various industries, so this finding suggests that differences in human capital across firms within a particular industry do not explain much of the remaining differences in innovation activity.
Firms’ ability to innovate also depends on external factors. As Chapter 2 notes, a poor business environment – widespread corruption, weak rule of law, burdensome red tape, and so on – can substantially increase the cost of introducing new products and make returns to investment in new products and technologies more uncertain. These factors can undermine firms’ incentives and ability to innovate.
The results of BEEPS V and MENA ES confirm this. As part of these surveys, each firm was asked whether various factors, such as access to land or labour regulations, were obstacles to doing business. Firms responded using a scale of 0 to 4, where 0 meant “no obstacle” and 4 signified a “very severe obstacle”.
The previous section shows that innovative firms tend to have a much more negative view of certain aspects of their business environment when compared with non-innovative firms. This raises the question of whether such perceived constraints negatively affect innovation outcomes. Do they inhibit innovation in practice? To answer this question, the impact of various aspects of the business environment is examined in more detail using cross-country regressions.
The business environment is, to a large extent, shaped by a country’s deeper economic institutions, such as the rule of law, control of corruption, the effectiveness of the government and regulatory quality. This can be captured by the average of the relevant Worldwide Governance Indicators, as discussed in Chapter 2. Together with other country-level characteristics, such as income per capita, R&D inputs, financial development and the quality of human capital, the quality of institutions is used in this section to explain the number of patents granted per worker and the innovation intensity of exports in various countries.The results of these cross-country regressions are presented in Table 3.3.
These results indicate that better institutions are associated with increases in patenting and more innovation-intensive exports. The effect of improving institutions is stronger and has greater statistical significance in countries where institutions are relatively weak. This can be seen where the average of the Worldwide Governance Indicators is interacted with (i) a dummy variable that takes the value of one when that average is above the mean for the sample (indicating strong economic institutions); or (ii) a dummy variable that takes the value of one when that average is below the mean for the sample (indicating weak economic institutions; see columns 3 to 8).
An improvement of around half a standard deviation in the quality of economic institutions in a country with below-average economic institutions (say, from the level of Ukraine to that of Albania) is associated with a 60 per cent increase in the innovation intensity of exports. An improvement of this magnitude is also associated with a 40 to 50 per cent increase in patent output. These effects are sizeable, considering that they only capture the direct impact of the quality of institutions, beyond the indirect effect that it may have through a higher level of income and of human capital in the country.
The analysis above shows that innovative firms feel far more constrained by customs and trade regulations than non-innovative firms. At the same time, firms that sell their products in export markets are more likely to innovate. The results of cross-country analysis confirm that both the size of the market (measured by population and GDP per capita) and economic openness (measured by the ratio of exports and imports to GDP) are important for the innovation intensity of exports. An increase in openness to trade totalling 30 percentage points of GDP (say, from the level of Ukraine to that of Latvia) is associated with a 9 to 15 per cent increase in the innovation intensity of exports. At the same time, no strong links are found between patent output and economic openness or the size of the economy.
In addition, there is also a positive (albeit weaker) relationship between the innovation intensity of exports and the financial openness of the economy (as measured by the Chinn-Ito index, where higher values correspond to free cross-border movement of capital and lower values correspond to more restrictive regimes).15 All in all, these results suggest that a country’s ability to commercialise innovations and adopt technologies benefits from openness to trade and a large market.
These results should be viewed as indicating a general correlation between innovation and country-level characteristics, rather than a causal relationship. For instance, the causality may also run from innovation to openness to trade. Indeed, innovation can support exports, as it can help firms to become more productive and improve their competitive positions in international markets, thereby increasing the ratio of exports to GDP. In order to take some account of such reverse causality, similar regressions have been estimated using values for income per capita, openness to trade and dependence on natural resources with a lag of ten years as proxies for their contemporaneous values. The results remain broadly unchanged (see columns 7 and 8).16
Interestingly, an abundance of natural resources – measured by calculating natural resource rents (that is to say, revenues net of extraction costs) as a percentage of GDP – has the opposite effect to economic openness. Reliance on commodities does not appear to have an impact on the patent output of an economy, but the exports of countries that are dependent on natural resources tend to be significantly less innovation-intensive than those of other countries (see Table 3.3).
This is, of course, partially a reflection of the fact that commodity sectors inevitably account for a larger share of such countries’ exports. However, this negative relationship may also arise because the economy’s dependence on natural resources reduces the average firm’s economic incentives to innovate, as a large percentage of the value added in the economy is derived from activities that are less reliant on continuous innovation.
For instance, while constant innovation and the adoption of cutting-edge technologies is a prerequisite for maintaining a competitive position in the automotive sector, a firm’s competitive edge in terms of natural resource exports is dependent primarily on natural resource endowments.17 At the same time, the availability of natural resource rents may enable governments (as well as universities and firms) to finance research, which offsets any negative impact that natural resources may have on patent output, but does not necessarily strengthen incentives to commercialise innovations.
The third aspect of the business environment that constrains innovative firms particularly strongly is the availability of the right skills. In country-level regressions (such as those reported in Table 3.3) measures of human capital – including the percentage of the population that has completed secondary or tertiary education, the average number of years of schooling and the average number of years of tertiary education – are not consistently found to be significant determinants of innovation. However, a higher average number of years of university education is generally associated with a higher patent output. This weaker correlation may be due to the fact that enrolment ratio-type measures predominantly capture the quantity – rather than the quality – of education.18
A more nuanced measure of the quality of education and basic skills is available for a sample of 65 OECD and non-OECD economies, based on the Programme for International Student Assessment (PISA) conducted by the OECD. PISA is a standardised international assessment of 15-year-old students’ abilities in the areas of reading, mathematics and science. It has been conducted every three years since 2000, with a sample of schools chosen at random in each country. Higher average scores across all students in all three subjects generally correspond to a higher quality of education in a given country.
For the sub-sample of countries participating in PISA, the average scores achieved by these 15-year-old students are positively and significantly correlated with innovation, in terms of both patent output and the innovation intensity of exports (see Chart 3.11). This relationship is particularly strong for patent output (with the correlation coefficient standing at around two-thirds), highlighting the role that the quality of education plays in facilitating innovation at the technological frontier.
The effect that R&D has on innovation outcomes, which was examined earlier at the level of individual firms, can also be observed in cross-country data (see Table 3.3). Furthermore, the results of cross-country analysis reveal that the distribution of R&D spending across firms, academic institutions and government also plays an important role. Both business R&D spending and government R&D spending are associated with increases in patent output, with the impact of an additional US$ 1 of R&D spending estimated to be higher for government R&D than for business R&D. However, only business R&D appears to have a positive impact on the innovation intensity of exports. This could be because of the poor links between science and industry in transition countries (see Box 5.3).
This discussion of the links between innovation and R&D in the various sectors also highlights the complexity of the innovation process, which requires a variety of general and specialist inputs. For this reason, countries that are at a more advanced stage in their development (measured, for instance, by GDP per capita at purchasing power parity) may be better placed to innovate. The cross-country results presented in Table 3.3 confirm that rich countries do tend to patent more.
However, there does not appear to be any correlation between income per capita and the innovation intensity of output. This may be due to the fact that firms in less developed countries have become increasingly successful at adopting existing technology over the last few decades.
Overall, the various factors discussed above explain between 60 and 90 per cent of variation in innovation outcomes across countries. The analysis also suggests that, given their income per capita, economic openness, human capital, economic institutions, R&D spending and other characteristics, transition economies innovate at around or slightly above the level that would be expected of them, in terms of both patent output and the innovation intensity of their exports.19
Successful innovation relies on a supportive business environment. A poor business environment can substantially increase the cost of developing new products and make returns to innovation much more uncertain, undermining firms’ incentives to innovate. In some cases it may prompt start-ups and other innovative firms to move their activities elsewhere, resulting in an “innovation drain”.
Strikingly, firms that have recently introduced a new product tend to regard all aspects of the business environment as a greater constraint on their operations and growth than firms that do not innovate. These differences between the views of innovative and non-innovative firms are particularly large when it comes to corruption, the skills of the workforce and customs and trade regulations.
From a geographical perspective, they tend to be larger in Central Asia, the EEC region and Russia. In the CEB region, by contrast, these differences are less pronounced, suggesting that the overall environment there may be more supportive of innovation.
Firm-level and cross-country analysis has identified a number of factors that play an important role in shaping firms’ incentives and ability to innovate, as well as innovation outcomes at country level. In the case of the latter, the factors that determine a country’s patent output are not necessarily the same as those that determine the innovation intensity of a country’s exports. For example, countries that are rich in natural resources tend to have less innovation-intensive exports, despite patenting levels that are comparable to those of other countries.
Overall, the analysis in this chapter suggests that efforts to further improve the innovation potential of firms and economies in the transition region should primarily target reductions in corruption, greater openness to international trade and cross-border investment (including effective customs and trade regulations) and improvements in the skills of the workforce. Other factors, such as improved access to finance and the upgrading of ICT infrastructure, also play an important role.
This analysis also reveals the relative scarcity of innovative start-ups in the transition region. While larger firms that have been around for a longer period of time tend to innovate more – particularly in high-tech manufacturing sectors, where innovation is more dependent on R&D – smaller and younger firms are often the ones developing products that are new to the global market.
In Israel, young, small firms are more likely to introduce world-class innovations than larger, established firms, but in the transition region this is not the case. On the contrary, innovations introduced by young, small firms in the EBRD region are less likely to target the global technological frontier than those of larger firms.
The analysis in this chapter supports the view that R&D activities increase the likelihood of successful innovation, but are by no means a prerequisite for innovation. The impact that R&D activities have on the likelihood of a new product being introduced is particularly large in high-tech manufacturing sectors. Meanwhile, R&D in low-tech sectors can help to optimise production processes. Lastly, while both business R&D and government R&D increase a country’s patent output, only business R&D has a significant positive impact on the innovation intensity of a country’s exports.
The transition region’s most successful innovative entrepreneurs and small firms often move to London, Berlin, Silicon Valley, Boston, New York and other innovation hubs at the earliest available opportunity in order to take advantage of the resources available there. The investors, mentors, advisers and clients located in these places help them to develop products faster and more efficiently (thanks to the benefits of agglomeration and clustering), while at the same time increasing the value of their businesses.20 The legacy of socialism means that entrepreneurship does not have a long tradition in the transition region, so marketing and business development still lag behind advanced economies.
Since a country’s development prospects are partly dependent on its capacity for innovation – which, in turn, depends on human capital – such “innovation drain” may be damaging. Indeed, research suggests that the emigration of highly skilled individuals weakens local knowledge networks.21
Over the past two decades, the increased prominence of global value chains (GVCs) has transformed the world economy. The declining cost of communication and international shipping has caused production processes to be broken down into ever smaller parts and spread across vast geographical areas. As a result, international commerce is now dominated by trade in intermediate – rather than final – goods and services. This box looks at how GVCs stimulate innovation among manufacturing firms in the transition region.26
There are several reasons why participation in GVCs can help firms in emerging economies to learn and innovate. First, being part of a GVC means that a firm has to satisfy the chain’s requirements in terms of the quality of products and the efficiency of processes.27 To do so, managers may need to adapt their production methods or acquire technology via licensing arrangements. Second, serving foreign clients may require improved logistical solutions or delivery methods, as delivery at the appropriate time is essential for a smooth supply chain. Third, importing intermediate goods can itself be a channel for the diffusion of technology where firms import state-of-the-art technology that has not previously been available in the domestic market. Importing new technologies can also enhance the technical skills of the workforce if this necessitates further training. These increases in human capital may, in turn, enable companies to introduce innovative products of their own.
Does stronger competition in product markets boost or hamper technological advances? The relationship between competition and innovation is complex, as multiple countervailing forces are at work.
On the one hand, concentrated markets with less competition may be more conducive to innovation. Large firms with substantial market power may be more willing to carry out innovation-oriented R&D activities because the scarcity of competitors will allow them to reap higher rents from newly introduced products if those innovations turn out to be successful. Market power may also help firms to finance R&D activities using retained earnings.
On the other hand, a lack of competition, while enabling firms to enjoy higher rents from new products, may also lead to complacency. In other words, firms may have more incentives to innovate in a competitive environment, in order to get ahead of their rivals and increase their market share.30
Consultancy firms can play a vital role in facilitating innovation by acting as conduits for external know-how and providing information about customers’ preferences.37 They can help a firm adapt its organisational structure and management practices to changing industry needs, help it refine its design and packaging in order to appeal more effectively to its target groups, or provide market research underpinning the development of new products that better satisfy customers’ needs. For instance, consultants have helped a Swedish bank to introduce internet banking.38 Consultants can also help firms’ managers to analyse the pros and cons of developing new products and processes.39
While the percentage of firms using consultants varies greatly across the countries of the transition region – ranging from just 4 per cent in Azerbaijan to 54 per cent in Ukraine – consultants are more likely to be used by innovative firms in almost all countries (see Chart 3.4.1).
D. Acemoğlu, U. Akcigit and M.A. Celik (2014)
“Young, restless and creative: Openness to disruption and creative innovations”, NBER Working Paper No. 19894.
P. Aghion, N. Bloom, R. Blundell, R. Griffith and P. Howitt (2005)
“Competition and innovation: an inverted U relationship”, Quarterly Journal of Economics, Vol. 120, No. 2, pp. 701-728.
A. Agrawal, D. Kapur, J. McHale and A. Oettl (2011)
“Brain drain or brain bank? The impact of skilled emigration on poor-country innovation”, Journal of Urban Economics, Vol. 69, No. 1, pp. 43-55.
K. Arrow (1962)
“Economic Welfare and the Allocation of Resources to Invention”, in R.R. Nelson (ed.), The Rate and Direction of Economic Activity, Princeton University Press.
Y. Back, K. Praveen Parboteeah and D.I. Nam (2014)
“Innovation in emerging markets: The role of management consulting firms”, Journal of International Management, forthcoming.
J.R. Baldwin and W. Gu (2004)
“Trade liberalization: Export-market participation, productivity growth and innovation”, Oxford Review of Economic Policy, Vol. 20, No. 3, pp. 372-392.
R. Barro and J.W. Lee (2013)
“A new data set of educational attainment in the world, 1950-2010”, Journal of Development Economics, Vol. 104, pp. 184-198.
N. Bloom, M. Draca and J. Van Reenen (2011)
“Trade induced technical change? The impact of Chinese imports on innovation, IT and productivity”, NBER Working Paper No. 16717.
M. Bruhn, D. Karlan and A. Schoar (2012)
“The Impact of Consulting Services on Small and Medium Enterprises: Evidence from a Randomized Trial in Mexico”, Economic Growth Center Working Paper No. 1010, Yale University.
P. Buccirossi, L. Ciari, T. Duso, G. Spagnolo and C. Vitale (2013)
“Competition Policy and Productivity Growth: An Empirical Assessment”, The Review of Economics and Statistics, Vol. 95, No. 4, pp. 1324-1336.
W. Carlin, M. Schaffer and P. Seabright (2004)
“A minimum of rivalry: Evidence from transition economies on the importance of competition for innovation and growth”, B.E. Journal of Economic Analysis & Policy, Vol. 3, No. 1, pp. 1-43.
M.D. Chinn and H. Ito (2006)
“What matters for financial development? Capital controls, institutions, and interactions”, Journal of Development Economics, Vol. 81, No. 1, pp. 163-192.
D. Coe, E. Helpman and A. Hoffmaister (2009)
“International R&D spillovers and institutions”, European Economic Review, Vol. 53, No. 7, pp. 723-741.
W.M. Cohen and D. Levinthal (1989)
“Innovation and learning: The two faces of R&D”, Economic Journal, Vol. 99, No. 397, pp. 569-596.
P. Conway, D. De Rosa, G. Nicoletti and F. Steiner (2006)
“Regulation, Competition and Productivity Convergence”, OECD Economics Department Working Paper No. 509.
Cornell University, INSEAD and WIPO (2014)
The Global Innovation Index 2014: The Human Factor in Innovation, Fontainebleau, Ithaca and Geneva.
G. Crespi and P. Zuñiga (2012)
“Innovation and productivity: Evidence from six Latin American countries”, World Development, Vol. 40, No. 2, pp. 273-290.
J.P. Damijan, Č. Kostevc and S. Polanec (2010)
“From innovation to exporting or vice versa?”, World Economy, Vol. 33, No. 3, pp. 374-398.
EBRD (2010)
Transition Report 2010: Recovery and Reform, London.
EBRD (2013)
Transition Report 2013: Stuck in Transition?, London.
The Economist (2014)
“The British diaspora: And don’t come back”, The Economist, 9 August 2014. Available at: www.economist.com/news/britain/21611102-some-5m-britons-live-abroad-country-could-do-far-more-exploit-its-high-flying-expats-and (last accessed on 5 September 2014).
R. Feenstra, R. Lipsey, H. Deng, A. Ma and H. Mo (2005)
“World trade flows: 1962-2000”, NBER Working Paper No. 11040.
J. Fortwengel (2011)
“Upgrading through integration? The case of the Central Eastern European automotive industry”, Transcience – A Journal of Global Studies, Vol. 2, No. 1, pp. 1-25.
L. Franssen (2014)
“The effect of global value chains on the relative skill-employment of firms in emerging economies”, EBRD working paper, forthcoming.
R. Griffith, E. Huergo, J. Mairesse and B. Peters (2006)
“Innovation and productivity across four European countries”, Oxford Review of Economic Policy, Vol. 22, No. 4, pp. 483-498.
A. Hoecht and P. Trott (2006)
“Innovation risks of strategic outsourcing”, Technovation, Vol. 26, No. 5-6, pp. 672-681.
D. Hummels, J. Ishii and K.-M. Yi (2001)
“The nature and growth of vertical specialization in world trade”, Journal of International Economics, Vol. 54, No. 1, pp. 75-96.
I. Khrennikov (2013)
“Ex-Soviet programmers take on India in $48 billion market”, Bloomberg, 17 September 2013. Available at: www.bloomberg.com/news/2013-09-16/ex-soviet-programmers-take-on-india-in-48-billion-market.html (last accessed on 5 September 2014).
A. Lileeva and D. Trefler (2010)
“Improved access to foreign markets raises plant-level productivity… for some plants”, Quarterly Journal of Economics, Vol. 125, No. 3, pp. 1051-1099.
R.R. Nelson and E. Phelps (1966)
“Investment in humans, technological diffusion and economic growth”, American Economic Review, Vol. 56, No. 1-2, pp. 69-75.
G. Nicoletti and S. Scarpetta (2005)
“Regulation and Economic Performance: Product Market Reforms and Productivity in the OECD”, OECD Economics Department Working Paper No. 460.
P. Nightingale and A. Coad (2013)
“Muppets and gazelles: political and methodological biases in entrepreneurship research”, Industrial and Corporate Change, Vol. 23, No. 1, pp. 113-143.
OECD (2009)
Innovation in Firms: A Microeconomic Perspective, Paris.
OECD (2010)
Investment Reform Index 2010: Monitoring policies and institutions for direct investment in South East Europe, Paris.
H. Pavlínek, B. Domański and R. Guzik (2009)
“Industrial upgrading through foreign direct investment in Central European automotive manufacturing”, European Urban and Regional Studies, Vol. 16, No. 1, pp. 43-63.
C. Pietrobelli and R. Raballotti (2011)
“Global value chains meet innovation systems: Are there learning opportunities for developing countries?”, World Development, Vol. 39, No. 7, pp. 1261-1269.
I. Sample (2014)
“If Pfizer’s AstraZeneca takeover succeeds, bad news for UK research”, The Guardian, 28 April 2014. Available at: www.theguardian.com/business/2014/apr/28/pfizer-astrazeneca-takeover-bad-news-uk-research (last accessed on 25 August 2014).
M. Stankovic, B. Angelova, V. Janeska and B. Stankovic (2013)
“Science and innovation policy in Southeast Europe: Brain drain as brain gain”, International Journal of Technological Learning, Innovation and Development, Vol. 6, No. 3, pp. 262-282.
B. Szabo (2013)
“How Central Eastern Europe is transforming from outsourcing to a real tech hub”, Forbes, 10 February 2013. Available at: www.forbes.com/sites/ciocentral/2013/10/02/how-central-eastern-europe-is-transforming-from-outsourcing-to-a-real-tech-hub/ (last accessed on 5 September 2014).
N.J. Thrift (2005)
Knowing Capitalism, SAGE Publications, Thousand Oaks, California.
H. Welsch (2008)
“Resource dependence, knowledge creation, and growth: Revisiting the natural resource curse”, Journal of Economic Development, Vol. 33, No. 1, pp. 45-70.