Agricultural Productivity and Economic Growth

Abstract

In most poor countries, large majorities of the population live in rural areas and earn their livelihoods primarily from agriculture. Many rural people in the developing world are poor, and conversely, most of the world's poor people inhabit rural areas. Agriculture also accounts for a significant fraction of the economic activity in the developing world, with some 25% of value added in poor countries coming from this sector. The sheer size of the agricultural sector implies that changes affecting agriculture have large aggregate effects. Thus, it seems reasonable that agricultural productivity growth should have significant effects on macro variables, including economic growth. But these effects can be complicated. The large size of the agricultural sector does not necessarily imply that it must be a leading sector for economic growth. In fact, agriculture in most developing countries has very low productivity relative to the rest of the economy. Expanding a low-productivity sector might not be unambiguously good for growth. Moreover, there are issues of reverse causation. Economies that experience growth in aggregate output could be the beneficiaries of good institutions or good fortune that also helps the agricultural sector. Thus, even after 50 years of research on agricultural development, there is abundant evidence for correlations between agricultural productivity increases and economic growth but little definitive evidence for a causal connection. This chapter reviews theoretical arguments and empirical evidence for the hypothesis that agricultural productivity improvements lead to economic growth in developing countries. For countries with large interior populations and limited access to international markets, agricultural development is essential for economic growth. For other countries, the importance of agriculture-led growth will depend on the relative feasibility and cost of importing food.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 97 citations exist upstream.