Data‎ > ‎

Ghana Cocoa Farmers Survey

The Ghana Cocoa Farmers Survey is a long
-running panel of cocoa farmers in three regions of Ghana:  Ashanti, Brong Ahafo, and Western.  This panel dataset represents a collaboration between the Centre for the Study of African Economies, Oxford, and the Ghana Cocoa Board.  Data have been collected every other year since 2002.  The data are available in Stata format for public use, and the CSAE is very happy for these to be used.  I only ask that you contact me to let me know if you are planning to make use of these data. 

A useful starting point for accessing the data is this guide, which was written based on the 2002 - 2006 data, but which remains broadly accurate for the remaining rounds of the questionnaire.  Key to putting together the data is to understand the basic file structure.  In most years, the data were collected and compiled as 6 separate files.  The central file, which is typically named with the prefix gcfs in each survey round, contains one observation per farmer; farmers are uniquely identified by the variable s0far The other components of the survey potentially include many observations per farmer.  These survey instruments include:  current and former household members; sales of cocoa and other crops; plots owned and cultivated; sales to Licensed Buying Companies; and agricultural shocks. 

In addition to the raw data, we have put together a few, illustrative files to demonstrate how the data may be linked across modules and across rounds.  The Stata do-file "sample descriptives.do" puts together simple descriptive statistics for the 2002 - 2006 survey rounds, resulting in the demo.dta data file. 

Finally, this is an extensive dataset, and, inevitably, individual users will encounter areas that have not been extensively cleaned, where these have not been used in analysis to date.  While the data are provided on an 'as is' basis, I am happy to provide guidance in order to ensure that consistent approaches to cleaning the data are undertaken across rounds.  I would also be grateful if users share any routines developed to further clean unused sections of these data; where appropriate, these will be incorporated in the publicly available data.

GCFS 2002

GCFS 2004

GCFS 2006

GCFS 2008

GCFS 2010