Appropriate Growth Policy: A Unifying Framework
GAICIL Insights
There is no such thing as a growth policy, only growth policies that fit.1 Appropriate Growth Policy: A Unifying Framework, by Philippe Aghion and Peter Howitt, uses growth theory in which growth comes from quality-improving innovations to elaborate a theory of growth policy and explain the growth gap between Europe and the United States. The framework systematizes the case-by-case approach to growth policy design.
Something for international organizations issuing AI policy checklists to sit with: one checklist for every country may be the clearest sign of a policy that has not been thought through. Ask first how far a country is from the frontier and how deep its finance runs, then prescribe.
Three policy areas get attention: competition and entry, education, and macroeconomic policy. Higher entry and exit, meaning higher firm turnover, and greater emphasis on higher education are more growth-enhancing in countries closer to the technological frontier. Countercyclical budgetary policies are more growth-enhancing in countries with lower financial development. Every lever has a context where it works best.
Interactions are the key point. Growth-maximizing policies should vary with a country's or sector's distance to the technological frontier and with its financial development. Empirical support implies policy does matter for growth, even if no linear relationship between policy and growth appears once institutions are controlled for. Appropriate growth policies vary with institutions, and such effects show up only when nonlinear interactions are accounted for. Regressions that look for one universal policy effect will report nothing, and wrongly conclude that policy is powerless.
A common macroeconomic separation is challenged. The distinction between short-run macroeconomic policy and long-run growth may mislead. Countercyclical macro policy promotes growth, mainly in countries with relatively low financial development. Stabilization and growth are linked. The business cycle and the growth trend are one subject.
The framework also adjudicates between models. AK and product variety models fail to account for evidence on the relationships among competition, education, volatility and growth, and so cannot deliver relevant prescriptions in these areas. The quality-improving innovation framework generates precise testable predictions. A theory earns policy relevance by being falsifiable where it matters.
The parallel with today's technology is close. Artificial intelligence policy should be equally context-dependent. Frontier countries may benefit most from policies promoting entry, competition and advanced education, which support AI innovation. Countries further from the frontier may benefit from different emphases, such as facilitating adoption. Copying the leader's AI strategy may be the wrong strategy for a follower.
Firm turnover is the frontier lever. Entry and exit let new AI firms challenge incumbents and let resources flow to more productive uses. For frontier economies, reducing entry barriers and letting failing firms exit can strengthen AI innovation. An economy that protects its incumbents protects its past.
Higher education matters more near the frontier. AI research and development require advanced skills, and frontier economies investing in universities and graduate training may reap larger growth benefits.
Macro policy protects AI investment. Innovation investment, including in AI, can be disrupted by downturns, especially where financial markets are less developed. Countercyclical policies sustaining investment in recessions protect long-term growth.
Europe remains the case study. Policies suited to catching up, which served Europe well after the war, may be less suited to an economy near the frontier. As AI becomes central to growth, Europe's policy mix in competition, education and macroeconomic management will influence its ability to innovate. A continent can outgrow the policies that made it rich.
The message is to tailor. A single prescription for all countries is unlikely to work. Knowing a country's distance to the frontier and financial development helps identify which policies will promote growth.
Read as a whole, appropriate growth policy depends on context, and a unifying framework can guide its design. Competition, entry, education and macroeconomic policy interact with a country's position and institutions. As AI shapes the frontier, context-dependent policy will help each economy capture the technology's growth potential.
Introduction
In this lecture, we use Schumpeterian growth theory, where growth comes from qualityimproving innovations, to elaborate a theory of growth policy and to explain the growth gap between Europe and the US. Our theoretical apparatus systematizes the case-by-case approach to growth policy design. The emphasis is on three policy areas that are potentially relevant for growth in Europe, namely: competition and entry, education, and macropolicy. We argue that higher entry and exit (higher firm turnover) and increased emphasis on higher education are more growth-enhancing in countries that are closer to the technological frontier. We also argue that countercyclical budgetary policies are more growth-enhancing in countries with lower financial development. The analysis thus points to important interaction effects between policies and state variables, such as distance to frontier or financial development, in growth regressions. Finally, we argue that the other endogenous growth models, namely the AK and product variety models, fail to account for the evidence on the relationship between competition, education, volatility, and growth, and consequently cannot deliver relevant policy prescriptions in the three areas we consider.
Takeaway
“In this lecture, we have argued that the Schumpeterian paradigm provides a unifying framework for thinking about and designing appropriate (context-dependent) growth policy. In particular, the paradigm produces precise testable predictions as to how growth-maximizing policies (e.g., competition and entry policies, the allocation of education funding, or the design of macroeconomic policies) should vary with a country’s or sector’s distance to the technological frontier, and/or with the country’s level of financial development. The empirical support we have presented for these predictions implies that policy does indeed matter for a country’s growth performance even if, as Easterly (forthcomming) has argued, no linear relationship between policy and growth can be found in the data once institutions are controlled for. This is because appropriate growth policies vary with a country’s institutions. For example, the results reported in Section 5.3 imply that countercyclical macro policy promotes growth, but mainly in countries with relatively low financial development. Such effects can be detected only by taking into account the nonlinear interactions predicted by Schumpeterian theory. We have also emphasized that the separation between (short-run) macroeconomic policy and long-run growth, often assumed in macro textbooks or in policy circles, is theoretically questionable and at odds with cross-country panel regressions; not only does long-run growth respond to fiscal or monetary policy over the cycle, but in addition there is complementarity between the growth effects of a more countercyclical budgetary policy and those of structural reforms such as product market or labor market liberalization.1
When comparing the Schumpeterian paradigm to the AK or product diversity models of endogenous growth, we have concluded that the former does a better job at delivering systematic and yet context-dependent policy prescriptions.”
Notes
- Philippe Aghion & Peter Howitt, Appropriate Growth Policy: A Unifying Framework, 4 J. Eur. Econ. Ass'n 269 (2006), https://doi.org/10.1162/jeea.2006.4.2-3.269. Takeaway source: Philippe Aghion, Peter Howitt, Appropriate Growth Policy: A Unifying Framework, Conclusion. ↩
Cite as: Philippe Aghion & Peter Howitt, Appropriate Growth Policy: A Unifying Framework, 4 J. Eur. Econ. Ass'n 269 (2006), https://doi.org/10.1162/jeea.2006.4.2-3.269
Header image: Cloud cover, MN State Capitol-01.jpg by Myotus, CC BY 4.0, via Wikimedia Commons.
