Growth, distance to frontier and composition of human capital
GAICIL Insights
More schooling, slower growth?1 That puzzling result for rich countries turns out to reflect a measuring mistake about what kind of schooling counts. Growth, distance to frontier and composition of human capital, by Jérôme Vandenbussche, Philippe Aghion and Costas Meghir, examines human capital's contribution to economy-wide technological improvement through two channels, innovation and imitation. Far from the frontier, imitation is the main engine of total factor productivity growth. Nearer the frontier, a country relies increasingly on innovation, requiring labor to be reallocated between the two activities.
The model rests on a reasonable assumption: innovation is more skill-intensive than imitation. Then skilled labor has a higher growth-enhancing effect closer to the technological frontier. Reallocation creates a complementarity between skilled labor and frontier proximity. The closer to the edge, the more the top of the skills distribution matters.
The puzzle dissolves. Previous research found the relationship between schooling level and growth in rich countries slightly negative. Revisiting the question with a panel of 19 OECD countries between 1960 and 2000, the authors find two things. First, the growth-enhancing margin in OECD countries is skilled human capital, with total human capital mattering less. Second, skilled human capital has a stronger growth-enhancing effect in economies closer to the technological frontier. Averages hid the action at the top.
Composition is the crucial distinction. Average years of schooling may miss what matters for growth in advanced economies. What counts is the stock of highly skilled workers capable of innovation. Countries near the frontier need advanced education, in addition to broad basic schooling, to sustain growth. A frontier economy measured by its median diploma is measured wrongly.
The authors acknowledge the model is stylized. Labor-force skill composition is taken as exogenous. Next steps would model schooling decisions explicitly and allow cross-country migration, capturing how individuals respond to the returns to skill and how talent moves across borders.
Where does this leave AI? AI is the archetypal skill-intensive frontier technology. Frontier countries depend on skilled researchers and engineers to innovate in AI. Investment in advanced education and research training yields high growth returns for them. For a frontier nation, the AI talent pipeline is a growth policy.
Countries further from the frontier face a different path. Imitating and adopting existing AI technologies can drive productivity growth and is less skill-intensive. Early on, broad education enabling workers to use and adapt AI tools may be the priority. As these countries near the frontier, emphasis should shift toward higher education. The right AI education policy depends on where a country stands.
Migration is the live extension. Skilled AI talent is highly mobile, and countries compete for it. Cross-country migration of skilled workers affects the stock of skilled human capital available for innovation. Immigration policies attracting skilled researchers strengthen frontier innovation capacity. A visa office can be an AI lab's best recruiter.
Education policy should weigh composition as well as quantity. Advanced-economy policymakers expanding university and graduate education in frontier-relevant fields, including AI, can strengthen the innovation channel.
The empirical approach adds credibility. Distinguishing types of human capital and letting effects vary with distance to the frontier uncovers relationships that aggregate measures had missed. Education's effects on growth depend on context.
Read as a whole, human capital composition matters for growth, and its importance depends on distance to the frontier. Skilled human capital drives innovation and has stronger effects closer to the frontier. As AI defines the new technological frontier, countries seeking to lead will need to invest in the advanced skills innovation requires, while those further behind can gain from adopting AI and building toward innovation over time.
For frontier-economy education ministries, the sting is this: a strategy built only on raising average attainment may be aimed at the wrong margin. In the AI era, the growth return may sit in the graduate seminar and the immigration file.
Introduction
We examine the contribution of human capital to economy-wide technological improvements through the two channels of innovation and imitation. We develop a theoretical model showing that skilled labor has a higher growth-enhancing effect closer to the technological frontier under the reasonable assumption that innovation is a relatively more skill-intensive activity than imitation. Also, we provide evidence in favor of this prediction using a panel dataset covering 19 OECD countries between 1960 and 2000 and explain why previous empirical research had found no positive relationship between initial schooling level and subsequent growth in rich countries.
Takeaway
Technological progress is a dual phenomenon which makes a different use of labor inputs at different levels of development. Far from the technological frontier, imitation of technologies is the main engine of total factor productivity growth. As a country gets closer to the frontier, it relies more and more on innovation, which implies reallocating labor from one activity to the other. Using an endogenous growth model, we demonstrate that this reallocation process can create a complementarity between skilled labor and proximity to the frontier. We use this theoretical insight to revisit the empirical relationship between schooling level and growth in rich countries, which previous research had found to be slightly negative. Using a panel of 19 OECD countries between 1960 and 2000, we obtain two main results. First, the growth-enhancing margin in OECD countries is that of skilled human capital rather than that of total human capital. Second, skilled human capital has a stronger growth-enhancing effect in economies which are closer to the technological frontier. Our model is stylized and could be enriched in several dimensions. First, we have taken the skill composition of the labor force to be exogenous. A further step would be to endogeneize it by modeling explicitly schooling decisions and by allowing for cross-country migrations. The current research by Aghion-BoustanHoxby-Vandenbussche (2004) on Education, Migration, and Growth across US states, is a first attempt in that direction .1
Our empirical estimates in this paper can be used to derive policy prescriptions on the optimal composition of education spending in developed countries , however endogenizing educational achievement would make this objective more rigorously reachable. Second, for simplicity we abstracted from international trade considerations. Developing a dynamic Ricardian model around the core idea presented in this paper would make it possible to study cross-sectoral allocation of skilled and unskilled labor in a context of international specialization. Such analysis would certainly yield further insights on the relationship between distance to frontier, composition of human capital and economic growth.
Notes
- Jérôme Vandenbussche, Philippe Aghion & Costas Meghir, Growth, distance to frontier and composition of human capital, 11 J. Econ. Growth 97 (2006), https://doi.org/10.1007/s10887-006-9002-y. Takeaway source: Conclusion, reproduced from a free author or repository copy. ↩
Cite as: Jérôme Vandenbussche, Philippe Aghion & Costas Meghir, Growth, distance to frontier and composition of human capital, 11 J. Econ. Growth 97 (2006), https://doi.org/10.1007/s10887-006-9002-y
Header image: Chemistry lab (6146864098).jpg by Barry Pousman from New York, NY, CC BY 2.0, via Wikimedia Commons.
.jpg){kind=link}