The Value of Accurate Crop Production Forecasts

October 2, 2010 - Author: T.S. Jayne and Shahidur Rashid

IDWP 108. T.S. Jayne and Shahidur Rashid. 2010. The Value of Accurate Crop Production Forecasts

EXECUTIVE SUMMARY:
Crop production forecasts are widely recognized as an important input into food balance
sheets and for anticipating production shortfalls. However, the role of accurate crop
production forecasting systems in mitigating food price instability and transitory food
insecurity is often under-appreciated. This paper explains how crop production forecasting
systems affect price instability and risks, and how they can be improved to stabilize the food
system.

There are two basic kinds of crop production forecasting systems in Africa. The most
common and longstanding approach relies on the large administrative network of Ministry of
Agriculture extension workers to make area and yield estimates in their local areas, which are
then aggregated up to district, provincial and national level production estimates. The second
approach, which is increasingly utilized in the region, is the use of nationally representative
annual crop forecast surveys. These surveys are generally implemented by the national
statistical office with survey design support from the Ministry of Agriculture. The strengths
and weaknesses of these two approaches are discussed.

Some countries, for example Ethiopia, Mozambique, and Malawi, produce estimates from the
Ministry of Agriculture each year and periodically augment these with nationally
representative survey-based estimates. The method based on Ministry of Agriculture
extension workers generally produces staple food production estimates that are substantially
higher than the statistically-based survey method. The main reasons for the generally higher
crop production estimates produced by Ministry of Agriculture extension worker approach
are identified. Regardless of which set of estimates are more accurate, discrepancies in
national food production estimates of 35% or more can cause great uncertainty in actual
production, create mistrust and second-guessing of import and export requirements resulting
from food balance sheet exercises, and lead to overshooting or undershooting of actual
import, export, and food aid decisions. All of these problems exacerbate food price risks and
market unpredictability. Two examples of such problems are discussed based on recent
experiences in Ethiopia and Malawi.

Our assessment is that statistically-based survey methods derived from recent agricultural
censes have the greatest potential to provide unbiased and reasonably precise cereal
production forecasts. However, poor implementation at certain stages of the process can lead
to major forecasting errors, and the potential for improved forecasting in most cases remains
great. The number of enumerators required to carry out a census is so large that it is very
difficult to ensure that all of them are adequately trained and supervised. Data entry errors, if
unchecked, can cause problems. The most important source of bias in survey-based
production forecasts is generally in the computation of weighting factors to extrapolate from
the surveyed sample to the farm population. External consistency checks can and should be
used to crosscheck production estimates with other economic data. For example, forecasts
indicating a major production expansion would seem to be inconsistent with rising prices,
assuming relatively stable demand. However, there are often alternative explanations for
apparently contradictory results.

Procedures for improving the accuracy of national survey-based crop forecasting systems
would include the following:

i. Invest in long-term capacity building of the national statistical organization to design and
carry out agricultural censes and surveys. Despite the costs involved, the foregoing
sections have provided examples of how potentially great the costs can be – in terms of
unanticipated food price shocks, hunger and food insecurity – due to inaccurate crop
production forecasts.

ii. Ensure major attention to achieving full listings of farm households in each
administrative unit, which is essential to generate correct weighting factors to enable
accurate extrapolation of results from the sampled farm households to the population.
iii. Recognize that demand fluctuates annually as well as production. In drought years, for
example, an increasing proportion of rural farm households become buyers of grain,
raising the demand for grain. Changes in annual demand, due to either weather or regional
or world market events should be more explicitly incorporated into the methodology of
food balance sheets.

iv. Ultimately, food balance sheets can become more accurate by moving to a more
sophisticated approach that recognizes how requirements are not a fixed number; they
respond to changes in prices. Nor is supply fixed either. Even in the short-run, supplies of
staple foods can increase or decrease markedly in response to price movements and
expectations about future price movements, which are in turn influenced by expectations
about future government actions. For example, in the event of drought, farmers in dual
staples zones may dig up supplies of cassava for home consumption, enabling them to sell
more maize than they would otherwise to take advantage of high prices that usually
accompany a drought year. These cross-commodity dynamics may produce major errors
in estimates of marketed supplies unless properly understood and reflected in food
balance sheets.

v. Production forecasting and food balance sheets can benefit from accounting for changes
in regional and global market conditions, changes in trade and marketing policies, as well
as changes in consumption and production behaviors. The key adjustment parameters will
include revised estimates of marketable surplus, changes in consumption behaviors,
requirements for national food assistance programs, and the analyses of the domestic and
international price transmissions.

Tags: cross-country, idwp


Related Topic Areas

Cross-country


Authors

Thomas Jayne

Thomas Jayne
517-432-9802
jayne@msu.edu


For more information visit:

Food Security Group

Michigan State University Michigan State University Close Menu button Menu and Search button Open Close