86114 MARCH 2014 • Number 137 Access to Finance, Product Innovation, and Middle-Income Growth Traps Pierre-Richard Agénor, Otaviano Canuto, and Michael Jelenic After experiencing an initial period of rapid growth, many developing countries have fallen into the middle-income “trap”—stuck between low-wage, low-technology markets and high-income, innovation-based developed economies. As previous literature has demonstrated (Agénor and Canuto 2012), public policies aimed at improving access to advanced information and telecommunications (ITC) infrastructure, protecting intellectual property rights, and reforming labor markets to reduce rigidities can help developing countries avoid such low-growth equilibria. As a complement to these policies, which create an enabling environment for learning and innovation, this note draws on more recent work (Agé- nor and Canuto 2014) that emphasizes the role of access to finance in supporting the innovative activities that in turn can help countries climb the ladder to high-income status. In particular, this note argues that inadequate access to finance has an adverse effect on innovation, directly, through the financing of fewer research and development (R&D) projects, and also indirectly, as fewer individuals may choose to invest in the skills necessary to work in R&D fields. These dual effects highlight the need for public policies aimed at alleviating credit market imperfections to promote the production of ideas and increase the incentives for workers to invest in higher skills. An empirical comparison of countries in East Asia that were able to escape the middle-income trap with less successful counterparts in Latin America provides a poignant example of how access to finance influences innovation outputs and long-term economic growth. Frictions on Innovation offer limited collateral. Notably, expenditures on salaries and wages for scientists and researchers, which often represent a The impacts of financial constraints on innovation have been large fraction of innovation-related expenditures, cannot be the subject of much debate in recent years. The conventional properly collateralized by financial intermediaries. For the view is that firms engaging in innovative activities, particu- borrower, this creates a moral hazard because there are dimin- larly R&D, may suffer from a variety of frictions that may ished consequences for nonperformance. For the lender, there limit their access to debt finance from the private sector. Two is limited recourse to recoup the initial investment if the R&D key drivers of external financing constraints on R&D activi- efforts do not produce a profitable output. ties are limited collateral, because assets held by firms engag- Second, information asymmetries may further limit the ing in innovation are mainly intangible, and asymmetric lev- possibilities of external financing for R&D activities. To pro- els of information between the firm engaging in R&D tect their proprietary information, firms may be unable or activities and potential external lenders. unwilling to offer fully transparent signals about the effective- First, by their very nature, R&D activities are compara- ness of their intended innovation programs to potential lend- tively difficult to finance because these types of investments ers. Indeed, divulging R&D activities with enough detail to 1 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK    www.worldbank.org/economicpremise assuage the concerns of potential lenders increases the risk are particularly evident in studies that used direct indica- that ideas developed through R&D activities may simply be tors based on firms’ own assessments. appropriated and replicated by competitors. An Overlapping Generations Approach: Given this lack of information, as well as the aforemen- Links between Finance, Innovation, and tioned issues relating to lack of collateral and the moral haz- Growth ard it creates, it is perhaps not surprising that firms with capi- tal to invest may discriminate against R&D activities. When To better understand the relationship among access, or lack firms do decide to invest in such activities, a high degree of thereof, to finance, product innovation, and labor supply in information asymmetry may induce lenders to demand a a growing economy, Agénor and Canuto (2014) developed higher rate of return than in the case of investments in physi- an overlapping generations (OLG) model of horizontal in- cal assets, since potential lenders would need to invest greater novation in the tradition of Romer (1990), with an endog- time and resources to monitor investment performance of enous distribution of skills, credit market imperfections, R&D activities. Thus, although information asymmetries and the involvement of financial intermediaries. In the basic matter for external financing of all types of investments, they model, financial intermediaries provide uncollateralized may be particularly significant in limiting financing of innova- loans to researchers in the design sector to finance wages. tion due to the complexity and specificity of the innovation These intermediaries also monitor borrowers at a cost that process. depends solely on the amount lent and the level of effort Taken together, limited collateral value, information fric- required from researchers. The model also assumes that tions, and increased monitoring costs may help explain why there is a one-to-one relationship between the level of effort some firms do not rely on external debt finance and instead and the probability of success of a research project. In the fund most of their R&D investments with their own equity second stage, the model accounts for the possibility that the resources. However, funding through equity is costly—espe- intensity of monitoring costs may fall, at least initially, with cially for firms whose values are mostly determined by their the number of successful projects. This reduction in costs growth opportunities and hence are severely exposed to asym- may occur because the expertise acquired in evaluating metric information frictions—or simply not available, as is of- profitable (and eventually successful) projects tends to re- ten the case for younger and smaller firms. Moreover, while duce, to a certain extent, the cost of screening and monitor- internal financing might be an option for large firms that have ing future borrowers. the resources to draw on to support their R&D activities, it is Crucially, this analysis highlights the fact that high inter- particularly difficult for smaller firms to compete, as large mediation costs may adversely affect innovation activity and sunk costs for innovative activities may act as a barrier to en- the long-run rate of economic growth. In addition, if the cost try in certain markets (Baumol 2002). Accordingly, if financ- of borrowing is high, wages in the design sector will be rela- ing constraints are binding for a sufficient number of innova- tively low, which may result in fewer individuals choosing to tive firms, and innovation stagnates, growth may be adversely invest in their skills and engage in design activities. Thus, lack affected. of access to finance not only exerts a direct, adverse effect on The theoretical relationship between access to finance innovation activity (by constraining the number of research and innovation is further supported by a large body of em- projects that are implemented), but also an indirect adverse pirical research that suggests financial constraints hamper effect as well. This demonstrates that the degree to which innovation activities for developed and developing coun- firms innovate and the distribution of skills in a population tries alike (Savignac [2008]; Ang [2010, 2011]; Ayyagari, (which conditions the ability to engage in innovation activi- Demirgüç-Kunt, and Maksimovic [2011]; Ilyina and Sa- ties) are jointly determined—as such, lack of skills and poor maniego [2011]; Brown, Martinsson, and Petersen [2012]; access to finance are linked. Efthyvoulou and Vahter [2012]; Gorodnichenko and Another key implication of the analysis is that if the in- Schnitzer [2013]; Hottenrott and Peters [2012]; Maskus, tensity of monitoring costs depends on the number of re- Neumann, and Seidel [2012]; Popov and Roosenboom search projects that have received financing (as a result of an [2012]; Silva and Carreira [2012]; and Hsu, Tian, and Xu information externality, for instance), then multiple equilib- [2013]). These recent studies, which address a range of ria may emerge. One such equilibrium is characterized as a econometric problems that plagued the earlier literature middle-income trap—a phenomenon that has been well docu- (inadequate controls, endogeneity bias, direction of causal- mented in the recent literature on developing countries, ity, sampling issues, and others), largely support the hy- where (barring a few exceptions) the cost of financial interme- pothesis that financial barriers and lack of finance produce diation remains high, the quality of the labor force is weak, significant negative effects that impact firms’ abilities—es- and R&D activity is limited. This potential outcome is espe- pecially smaller firms’—to innovate. These negative effects cially relevant because the focus of this formal analysis is on 2 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK    www.worldbank.org/economicpremise middle-income countries, where labor market effects of access sectoral composition of R&D funding, the business sector to finance (or lack thereof) are particularly important to un- in East Asia plays a disproportionally large role, accounting derstanding the interactions among financial intermediation, for a little more than 54 percent, with public sector and innovation, and growth. In these countries, the supply of high- universities accounting for just over a third of R&D expen- ly qualified labor remains relatively limited, creating another ditures (table 2). constraint on innovation activity. By contrast, in developed Looking at the same period in Latin America, it is obvi- countries, the issue may be less about the quantity of highly ous that its R&D spending and intensity pales in comparison skilled workers and more about the allocation of talent and with the selected countries in East Asia. Over the same peri- provision of adequate incentives to motivate workers to engage od, R&D spending in Latin America totaled US$21.7 bil- in risky entrepreneurial activities. Nevertheless, the analysis lion—only 2.6 percent of world R&D expenditures—with a also has some relevance for slow-growing industrial countries. 2002 R&D intensity only half of that of East Asia, at just 0.6 Indeed, there is sufficient and compelling empirical evidence, percent. In addition to these divergent rates of spending and including several of the contributions mentioned earlier, to intensity, the sectoral distribution of Latin American R&D suggest that, in many of these countries, access to finance re- funding is highly skewed toward the government and univer- mains an equally important constraint on the innovation ac- sity sectors, which, at the turn of the 21st century, collectively tivity of small and medium-sized firms. Inadequate access to supplied almost 65 percent of R&D funding, with the busi- finance may therefore be the source of a slow-growth equilib- ness sector funding just one-third of total R&D activities. rium, just as with the middle-income trap. R&D Outputs: Patents, Researchers, and Access to Finance and Innovation: Publications Examples from East Asia and While expenditures and intensity provide a partial view of Latin America the resources an economy allocates to innovation activities, Given the implications of the model noted above, East Table 1. R&D Expenditures Asia and Latin America provide a good illustration of the R&D spending, 2002 R&D as % of GDPa relationship between finance for innovative activities, particularly R&D, and the innovative potential of an US$ % of Region billions world 1992 2002 economy. Several fast-growing economies in East Asia— Hong Kong SAR (China), Japan, the Republic of Korea, East Asia 111.7 13.5 0.7 1.2 Singapore, and Taiwan (China)—demonstrate a strong NIEs 36.4 4.4 1.6 2.2 correlation between R&D inputs and innovative outputs, Hong Kong SAR (China) 1.1 0.1 0.3 b 0.6 which have underpinned a sustained growth trajectory Korea, Rep. of 20.8 2.5 1.9 2.5 from middle- to high-income status. A poignant counter- Singapore 2.2 0.3 1.2 2.2 example comes from Latin America, where credit market Taiwan (China) 12.2 1.5 1.8 2.3 frictions have limited access to finance for R&D activities, Southeast Asia 3.3 0.4 0.1 0.2 which has hampered innovation capacities and limited Indonesia 0.3 0.0 0.1 c 0.1d growth in the later part of the 20th century. Malaysia 1.5 0.2 0.4 0.7 R&D Inputs: Expenditures, Intensity, and Sectoral Philippines 0.4 0.0 0.2 0.1 Composition of Funding Thailand 1.1 0.1 0.2 0.2 East Asia has successfully managed to allocate resources China 72.0 8.7 0.8 1.2 toward R&D, accounting for almost three quarters of World 829.9 100.0 1.7 1.7 developing countries’ increase in R&D over the last de- cade of the 20th century (Gill and Kharas 2007). Table Developed countries 645.8 77.8 2.3 2.3 1 shows that R&D expenditures in East Asia reached Japan 106.4 12.8 2.9 3.1 more than US$111 billion in 2002, or 13.5 percent of United States 275.1 33.1 2.6 2.6 the world total. Over the same period, R&D intensity— Developing countries 184.1 22.1 0.6 0.9 that is, the ratio of R&D spending to GDP—nearly dou- Latin America 21.7 2.6 0.5 0.6 bled, from 0.7 percent in 1992 to 1.2 percent in 2002. Emerging Europe 30.3 3.7 1.0 1.2 Leading the pack are Korea, Singapore, and Taiwan Source: UNESCO (2004, 2006); reprinted from Gill and Kharas (2007). (China), which by 2002 devoted between 2.2 and 2.5 a. Regional data are the sum of R&D divided by the sum of GDP. b. 1995. percent of GDP to R&D spending, rivaling the R&D in- c. 1994. tensity of many developed countries. In terms of the d. 2001. 3 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK    www.worldbank.org/economicpremise Table 2. R&D Expenditures (by sector) Sector of performance Sector of funding Higher Higher Region Business Government education Business Government education East Asia 62.2 21.7 14.4 54.3 35.2 2.3 NIEs 63.0 11.7 18.8 58.7 35.9 1.7 Hong Kong SAR (China) 33.2 3.1 63.6 35.3 62.8 0.2 Korea, Rep. of 76.1 12.6 10.1 74.0 23.9 1.7 Singapore 63.8 10.9 25.4 54.3 36.6 2.3 Taiwan (China) 62.2 24.8 12.3 63.1 35.2 0.0 Southeast Asia 51.3 22.1 15.7 46.6 35.4 6.2 Indonesia 14.3 81.1 4.6 14.7 84.5 0.2 Malaysia 65.3 20.3 14.4 51.5 32.1 4.9 Philippines 58.6 21.7 17.0 59.7 24.6 7.5 Thailand 43.9 22.5 31.0 41.8 38.6 15.1 China 62.4 27.1 10.5 60.1 29.9 .. Developed countries (21) 62.9 13.3 27.0 49.2 33.6 2.1 Japan 75.0 9.3 13.7 74.5 17.7 6.3 United States 70.1 12.2 13.6 63.7 31.0 .. Latin America (11) 29.0 27.2 32.7 32.9 37.3 27.4 Emerging Europe (9) 42.7 29.8 20.1 38.3 54.2 0.5 Source: UNESCO 2006; reprinted from Gill and Kharas (2007). Note: Table covers 2002–5 or latest available year and shows medians for regions and subregions. ". ." = negligible. The number of countries involved is shown in parentheses. patents are an indicator of an economy’s innovation outputs. 368 over the same period, with only 0.08 patents per A large body of literature corroborates the “significant rela- 100,000 of population. tionship between innovation inputs, such as R&D expendi- In addition to the number of patents produced in real tures, and innovation outputs, such as patent counts” (Gill and per capita terms, the OLG model suggests that access to and Kharas 2007, 158). In addition to patent counts, given finance for R&D has a secondary effect on labor compensa- the extensions of the OLG model used above, R&D invest- tion (or the return to education) in the design sector, and ments may have a second generation effect on the composi- hence the number of individuals pursuing a research or tech- tion of the labor force, particularly on the number of individ- nical career. Accordingly, the number of R&D researchers and uals choosing to pursue careers as researchers or technical technicians—and the journal articles they produce—varies experts in the design sector. For inter-regional comparison, it widely across countries in East Asia and Latin America. Ac- is instructive to examine this labor dimension of R&D invest- cording to the 2008 World Development Indicators, the suc- ment as well as the nonpatent outputs these individuals gen- cessful countries in East Asia have a relatively high ratio of erate through scientific publications. individuals working as R&D researchers and technicians. Sin- Considering the relationship between R&D inputs gapore, Japan, and Korea are leading the pack, with an average and outputs, it is perhaps not surprising that countries in of more than 5,000 researchers per million people. Converse- East Asia have far outpaced countries in Latin America in ly, Latin America’s largest economies—Argentina, Brazil, terms of patents generated. As confirmed by table 3, the Chile, and Mexico—average far less than 1,000 researchers per average annual number of patents granted in East Asian million people. The disparities between these two regions is economies was at 12,108 per year during 2000–2004. As further highlighted by the number of scientific and technical a result of their prolific patent growth, Taiwan (China) and journal articles published, with Korea producing more in Korea had by 2004 become the fourth and fifth biggest re- 2008 than Argentina, Brazil, Chile, and Mexico combined. cipients of U.S. Patent and Trademark Office patents in the Reasons for Divergence world after the United States, Japan, and Germany (Gill Both theoretical and empirical evidence suggest that finan- and Kharas 2007, 154). In stark comparison, the average cial market imperfections, such as information asymmetries number of annual patents in Latin America averaged only and transaction costs, are likely to be especially binding on 4 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK    www.worldbank.org/economicpremise Table 3. Number of Patents Granted, Selected Regions and Countries (annual averages) Number of patents Patents per 1,000 of population Region 1990–94 2000–2004 1990–94 2000–2004 % change East Asia (9) 2,239 12,108 0.14 0.72 17.6 NIEs 2,159 11,601 2.93 14.74 17.5 Hong Kong SAR (China) 184 616 3.15 9.32 11.4 Korea, Rep. of 633 4,009 1.44 8.67 19.7 Singapore 36 382 1.09 9.87 24.6 Taiwan (China) 1,307 6,593 6.30 30.17 17.0 Southeast Asia 31 140 0.01 0.04 15.3 Indonesia 6 15 0.00 0.01 8.8 Malaysia 13 64 0.07 0.28 15.3 Philippines 6 18 0.01 0.02 10.4 Thailand 6 43 0.01 0.07 20.9 China 48 368 0.00 0.03 22.9 World 107,361 182,523 1.98 2.95 4.1 Developed countries (21) 104,170 168,017 12.88 19.58 4.3 Japan 22,647 35,687 18.23 28.54 4.6 United States 59,024 97,104 23.00 33.56 3.9 Developing countries Latin America (11) 173 368 0.04 0.08 6.3 Emerging Europe (9) 205 348 0.07 0.12 5.6 Source: Data of the USPTO; reprinted from Gill and Kharas (2007). Note: The number of countries involved is shown in parentheses. talented individuals and small enterprises that lack collater- (Pagés 2010). Moreover, “lack of financing points directly at al, credit histories, and connections. Indeed, it may be more Latin America’s deficit in private financial intermediaries, difficult for small firms to secure outside finance (due to such as venture capital or angel investors, as well as public fi- more severe problems of information asymmetries) or use nancing directly aimed at encouraging private sector innova- internal funds to finance an R&D project. Moreover, credit tion, especially by small and medium businesses” (Pagés availability and financial sector depth also have important 2010, 14). influences on innovation, because they help meet the vari- Although public sector and university-based R&D ex- ous financing requirements of the R&D projects undertaken penditures constitute the majority of outlays for R&D in the by firms. Latin America region, evidence suggests that R&D at the firm When comparing the countries in East Asia—those able level is largely self-financed. For example, the IDB estimates to escape the middle-income trap—and the experiences of that “internal sources constitute the main source of innova- Latin America’s largest economies, there is indeed a large di- tion financing, representing more than 70 percent of total fi- vergence between their R&D inputs and R&D outputs. This nancing (reinvestment represents 74 percent of total financ- divergence begs the question of why there is such a large dis- ing in Argentina and 76.5 percent in Uruguay), followed by connect between the two regions in R&D spending and inten- commercial bank financing” (Pagés 2010, 10). sity in the first place, and why is the sectoral allocation of R&D Conversely, firms in East Asia benefit from highly inte- funding so skewed toward the business sector in East Asia and grated financial markets, where capital is more readily avail- the government/university sectors in Latin America? able to allocate to such R&D activities. However, this has not A key reason is that access to finance, particularly for always been the case. In earlier stages of development, the R&D activities, has been especially binding on innovation in- governments of many countries in East Asia provided essen- vestment in Latin America. According to the Inter-American tial public resources to target the innovative capacities of Development Bank (IDB), this might partly reflect problems strategic sectors of the economy. The rise of the developmen- in the functioning of the region’s financial markets in general, tal state in East Asia almost certainly played a crucial role in as Latin America has the highest cost of capital in the world mobilizing resources for innovation and managing rents so 5 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK    www.worldbank.org/economicpremise that private and semiprivate entities could undertake inno- If new entrants and startup firms are the most affected vative activities. by the lack of finance, it is important for governments to pro- While access to finance can have a great effect on R&D vide some form of assistance to these firms. However, this intensity and outputs, other key factors can influence techno- analysis does not explicitly provide support for the provision logical innovation and learning. As mentioned in an earlier of direct subsidies to R&D firms, especially in the context of policy note (Agénor, Canuto, and Jelenic 2012), developing developing countries where targeting may be particularly dif- countries can rely on a triad of policy measures to create posi- ficult due to weak institutional environments. Measures tive spillovers, including developing advanced infrastructure aimed at reducing the cost of collecting, processing, and dis- in the form of high-speed communications networks to share seminating information about potential borrowers may prove disembodied knowledge, improving the enforcement of prop- more effective in some developing countries. Ensuring that erty rights through patent protections, and reforming labor firms know where to find available funding sources may lead markets to ensure that rigidities do not prevent the efficient to more intensive innovation, which in turn may result in the allocation of labor resources. Fundamentally, these policies more rapid development of new goods and technologies and attract more high-ability workers into the design sector, im- convergence to the technological frontier. prove productivity and wages in that sector, and increase a In addition to these public sector initiatives, there is also country’s capacity for innovation. a greater role for private financial intermediation to play. Al- In addition to access to finance and a network of enabling though this analysis does not explicitly introduce capital mar- policies to facilitate technological spillovers, there are a num- kets other than credit into the model, it does suggest that ber of underlying systemic conditions that can boost a coun- there may be a greater role for private finance in developing try’s ability to innovate. Among other factors, literature high- countries. Credit bureaus, which are designed to share infor- lights the important roles played by macroeconomic stability, mation about potential borrowers, can and have been created competition, openness, and an educated workforce (Gill and by private agents in the past, although free-rider problems Kharas 2007). While these four elements are not the focus of and weak institutions may make government intervention this note, they are worth noting as additional underlying driv- necessary during an initial stage. Many developed countries ers of an innovative economy. have a private sector venture capital industry that provides fi- Policy Implications nancing for new and young innovative firms. However, this is not the case in most developing countries. In these countries, The broad policy implications of the foregoing analysis are regulatory reform of financial markets should focus on the fi- in line with a number of recent academic and policy con- nancing of innovative startups through venture capital mar- tributions for both high- and middle-income countries, kets, the securitization of innovation-related assets (namely, which have argued that lack of access to external finance intellectual property rights), and the provision of appropriate may hamper the development of innovative firms. Because incentives for risk-taking. larger firms tend to be less constrained in their operation and growth because of their ability to obtain external fi- Conclusion nance—a possible reflection of their ability to pledge col- From both a theoretical and empirical perspective, access to lateral—promoting innovation requires a particular focus finance greatly impacts the resources devoted to R&D. Be- on improving access to finance for small and medium firms (Beck, Demirgüç-Kunt, and Maksimovic 2005; World cause lack of collateral, information asymmetries, and high Bank 2008, 2012; Dinh and Clarke 2012), in which a dis- monitoring costs make external investments in R&D less ap- proportionate share of innovative research continues to be pealing than other potential investment choices, public po- conducted. lices can be employed to direct more public and private re- Based on the findings of the OLG analysis, an aggressive sources to support R&D. Where private capital is not available, policy aimed at alleviating credit market imperfections and such as in some developing countries, the public sector can increasing innovators’ access to finance may allow a country play a crucial role in supporting innovation and R&D initia- to avoid the middle-income trap and lead it to an equilibrium tives. The East Asian experience and that of selected econo- characterized by a high share of skilled workers, high growth, mies in Latin America empirically illustrate this point, con- and high productivity in design sectors. In such conditions, firming that directing resources to innovation is an important promoting access to finance is also essential to escape from a element in escaping the middle-income trap. low-growth equilibrium and put the economy on a path that About the Authors would allow it to converge to a high-growth, high-innovation equilibrium. To reach this state, there are key roles that both Pierre-Richard Agénor is Hallsworth Professor of International the public and private sectors can play. Macroeconomics and Development Economics, School of Social 6 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK    www.worldbank.org/economicpremise Sciences, University of Manchester, and Co-Director, Centre for Efthyvoulou, Georgios, and Priit Vahter. 2012. “Financial Con- Growth and Business Cycle Research. Otaviano Canuto is Se- straints and Innovation Performance: Are All Firms Similar?” Unpublished, University of Birmingham. nior Advisor on BRICS Economies in the Development Eco- Galor, Oded, and Omer Moav. 1998. “Ability Biased Technologi- nomics Department at the World Bank. He previously served cal Transition, Wage Inequality, and Economic Growth.” as the World Bank’s Vice President and Head of the Poverty Brown University, Department of Economics, Working Paper Reduction and Economic Management (PREM) Network. Mi- 98-14. chael Jelenic is an Operations Analyst in the PREM Network of Gill, Indermit Singh, and Homi Kharas, with Deepak Bhattasali et the World Bank’s Africa Region, and a visiting fellow at The al. 2007. An East Asian Renaissance: Ideas for Economic Growth. Washington, DC: World Bank. Graduate Institute of International and Development Studies Gorodnichenko, Yuriy, and Monika Schnitzer. 2013. “Financial (IHEID), Geneva. Constraints and Innovation: Why Poor Countries Don’t Catch Up.” Journal of European Economic Association 11 (5): Note 1115–52. 1. The baseline year for comparison was 2008 because it was Hottenrott, Hanna, and Bettina Peters. 2012. “Innovative Capabil- ity and Financing Constraints for Innovation: More Money, the most recent year for which information was available for More Innovation?” Review of Economics and Statistics 94 (4): all countries considered in this analysis. 1126–42. Hsu, Po-Hsuan, Xuan Tian, and Yan Xu. Forthcoming. “Financial References Development and Innovation: Cross-Country Evidence.” Jour- Agénor, Pierre-Richard, and Otaviano Canuto. 2012. “Middle-In- nal of Financial Economics. come Growth Traps.” Policy Research Working Paper No. 6210, Ilyina, Anna, and Roberto Samaniego. 2011. “Technology and World Bank, Washington, DC. Financial Development.” Journal of Money, Credit and Banking ———. 2014. “Access to Finance, Product Innovation, and Middle-In- 43 (August): 899–921. come Traps.” Policy Research Working Paper No. 6767, World Im, Fernando G., and David Rosenblatt. 2013. “Middle-Income Bank, Washington, DC. Traps: A Conceptual and Empirical Survey.” Policy Research Agénor, Pierre-Richard, Otaviano Canuto, and Michael Jelenic. Working Paper No. 6594, World Bank, Washington, DC. 2012. “Avoiding Middle-Income Growth Traps.” World Bank Maskus, Keith E., Rebecca Neumann, and Tobias Seidel. 2012. Economic Premise 98, Washington, DC. “How National and International Financial Development Af- Ang, James B. 2010. “Research, Technological Change and Financial fect Industrial R&D.” European Economic Review 56 (January): Liberalization in South Korea.” Journal of Macroeconomics 32 72–83. (March): 457–68. Pagés, Carmen, ed. 2010. “The Importance of Ideas: Innovation ———. 2011. “Financial Development, Liberalization and Tech- and Productivity in Latin America.” In The Age of Productivity, nological Deepening.” European Economic Review 55 (June): chapter 10. Washington, DC: IDB. 688–701. Popov, Alexander, and Peter Roosenboom. 2012.“Venture Capital Ayyagari, Meghana, Asli Demirgüç-Kunt, and Vojislav Maksimovic. and Industrial Innovation: Evidence from Europe.” Economic 2011. “Firm Innovation in Emerging Markets: The Roles of Policy 27 (July): 447–82. Governance and Finance.” Journal of Financial and Quantitative Romer, Paul M. 1990. “Endogenous Technological Change.” Journal Analysis 46 (September): 1545–80. of Political Economy 98 (October): s71–s102. Baumol, William J. 2002. The Free Market Innovation Machine: Savignac, Frédérique. 2008. “Impact of Financial Constraints on Analyzing the Growth Miracle of Capitalism. Princeton, NJ: Innovation: What Can Be Learned from a Direct Measure?” Princeton University Press. Economics of Innovation and New Technology 17 (August): Beck, Thorsten, Asli Demirgüç-Kunt, and Vojislav Maksimovic. 553–69. 2005. “Financial and Legal Constraints to Firm Growth: Does Silva, Filipe, and Carlos Carreira. 2012. “Do Financial Constraints Firm Size Matter?” Journal of Finance 60 (March): 137–77. Threaten the Innovation Process? Evidence from Portuguese Brown, James R., Gustav Martinsson, and Bruce C. Petersen. Firms.” Economics of Innovation and New Technology 21 (Novem- 2012. “Do Financing Constraints Matter for R&D?” European ber): 701–36. Economic Review 56 (November): 1512–29. World Bank. 2008. Finance for All: Policies and Pitfalls in Expanding Dinh, Hinh T., and George R. G. Clarke, eds. 2012. Performance of Access. Policy Research Report, Washington, DC: World Bank. Manufacturing Firms in Africa: An Empirical Analysis. Washing- ———. 2012. World Development Report 2013: Jobs. World Bank ton, DC: World Bank. Washington, DC. The Economic Premise note series is intended to summarize good practices and key policy findings on topics related to economic policy. They are produced by the Poverty Reduction and Economic Management (PREM) Network Vice-Presidency of the World Bank. The views expressed here are those of the authors and do not necessarily reflect those of the World Bank. The notes are available at: www.worldbank.org/economicpremise. 7 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK    www.worldbank.org/economicpremise