AK Endogenous Growth Model

Evidence and data

Use saving, investment, human capital, research intensity, infrastructure quality, depreciation, and productivity to judge broad-capital accumulation.

Calibration or measurement

A is the return to broad capital. If institutions or allocation lower A, higher saving does not guarantee higher growth.

Boundaries

• No diminishing returns is the key assumption.
• Capital is broadly defined and hard to measure.
• The model can overstate policy effects when investment quality is poor.

Use guidance

When sufficient

Endogenous-growth intuition when the key claim is that capital accumulation, broadly defined to include human capital and knowledge spillovers, does not hit diminishing returns at the aggregate level. The model supports the argument that policy can permanently raise the growth rate, not just the level, because a higher saving rate lifts the constant return A rather than moving along a diminishing-returns curve (Romer 1986 JPE; Rebelo 1991 JPE).

When sketch only

Do not use for detailed innovation policy, sector-level productivity analysis, or any question where identifying the specific mechanism of growth is the task. The model collapses all growth-sustaining forces into a single constant A, which hides the distinction between physical capital, human capital, and knowledge that policy needs to see to act on.

When to switch

Switch to Romer (1990 JPE) when the question is about the incentives for intentional innovation and the returns to research investment. Switch to Aghion-Howitt (1992 Econometrica) when creative destruction and Schumpeterian replacement of incumbents is the relevant channel. Switch to augmented Solow (Mankiw-Romer-Weil 1992) when cross-country level comparisons are the goal rather than permanent growth-rate differences.

Falsification signal

Cross-country convergence among economies with similar institutions and open capital accounts: if AK held literally, richer countries with higher K would have the same marginal product of capital as poor ones and capital would not flow toward developing economies. Sustained conditional convergence at the rates observed in the postwar data implies some diminishing returns to capital, contradicting the constant-marginal-product assumption.