Maize Yield Variability

A typical time-series of FAO crop yields consists of random variations about a slowly varying trend. The trend can be fit by a smooth curve, which is close to linear in many cases. The residual yield variation (difference between actual yield and trend yield) for top maize producering countries between 1961 and 2007 is shown below. In this case the trend fit was a locally weighted regression (lowess parameter = 0.7).


Computed values of  maize yield volatility for the top six producer countries are shown in the table below.

Maize Yield Volatilities

US 8.9%
China 6.4%
Brazil 8.0%
Mexico 7.4%
India 8.7%
Argentina 11.9%
Global 5%

Yield volatility for the world’s top producer (the US) is surprisingly large at 9%. It is also apparent from the plot that there is little correlation in yield variations between countries, at least for the top six producers. This explains why the global yield volatility (5%) is substantially lower than individual country volatilities. Just as in finance, diversification lowers risk.


What drives these relatively large annual fluctuations in maize yield? One of the most important factors influencing crop yields is climatic variability. Changes in precipitation, water vapour, temperature, cloud cover, frost, hail, snow-cover etc may affect harvest, depending on region and crop. However, management practices, such as fertiliser use, may also vary from season to season. Sometimes this is a large effect. For instance, crop yields in many Eastern block countries collapsed after break-up of the Soviet Union in 1991. The graph below illustrates this in the case of Belarus maize. Yields collapsed in 1993 and 1994, but have recovered strongly since. For the interval 1961-1991 the plot shows aggregate USSR data, because FAO does not have data for individual FSU states prior to 1992.



Zimbabwean maize yields (below) began to fall in the early 1980’s. While the fall is less dramatic than the case of Belarus, yields have not recovered due to local political factors. In contrast to neighbouring Zimbabwe,  South Africa shows a steady improvement typical of the Green Revolution.


Despite the different trends in Zimbabwe and South Africa, annual variations in yield are correlated. The reason is that both countries are strongly influenced by droughts associated with the El Nino Southern Oscillation (Equatorial Pacific sea surface temperature anomaly). This was pointed out in 1994 in an influential article article by Cane et al . The yield variations about trend line reflect climatic variability in Southern African countries.

Conclusions? It appears reasonable to decompose crop yield time-series into a smooth trend (reflecting the Green Revolution) and fluctuations about the trend. Annual variations are driven primarily by climate (e.g. Southern Africa maize) and sometimes by abrupt changes management practices (e.g. Belarus maize).

This entry was posted in Agriculture. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *