Does fertiliser use improve household welfare? Evidence from Ghana’s cocoa industry

ABSTRACT

This article analyses welfare impacts of fertiliser adoption using data collected from 838 cocoa farm households in the four key cocoa-producing regions in Ghana. Using propensity score matching (PSM), the study indicates that application of fertiliser leads to significant gains in farm yields, farm income, consumption expenditure, consumption expenditure per capita, and value of productive farm assets. The article concludes that cocoa-specific programmes such as Cocoa High Technology (Cocoa Hi-tech), initiated to intensify the application of improved farm technologies such as fertiliser, should be strengthened through effective and efficient management systems.

KEYWORDS:

• Environment (built and natural) – Agriculture
• Food security
• Labour and livelihoods – Economics
• Sub-Saharan Africa

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Gideon Danso-Abbeam is a PhD student in the School of Agricultural, Earth and Environmental Sciences, University of KwaZulu-Natal, Pietermaritzburg, South Africa.

Lloyd J.S. Baiyegunhi is an Associate Professor in Agricultural Economics in the School of Agricultural, Earth and Environmental Sciences, University of KwaZulu-Natal, Pietermaritzburg, South Africa.

ORCID

Gideon Danso-Abbeam http://orcid.org/0000-0002-7971-4676

Lloyd J. S. Baiyegunhi http://orcid.org/0000-0001-7717-3610

Notes

1. CRIG is a subsidiary of Ghana Cocoa Board (COCOBOD), a state institution monopsony that manages Ghana’s cocoa industry.

2. Matured cocoa trees are trees between 6 and 30 years old.

3. The study uses “fertiliser adoption” and “fertiliser use or application” interchangeably.

4. The study defined fertiliser users as farmers who applied CRIG-recommended fertiliser, whether granular or foliar. It is important to note that fertiliser users in the data sample applied only CRIG-recommended subsidised fertiliser, while non-users did not apply any fertiliser.

5. The cocoa-producing regions referred to here are the administrative regions of Ghana and not the regional classification by COCOBOD.

6. The proportionate sampling was based on the intensity of production levels. Western region contributed about 53.4% while Ashanti, Brong-Ahafo, and Eastern regions contributed about 17.2%, 9.7%, and 9.6%, respectively of cocoa production in the 2015/16 cocoa season. The selection of districts and communities were also based on production levels and were chosen with the help of the regional and districts offices of COCOBOD.

7. Welfare indicators were measured as follows: (1) farm income is gross revenue from the sale of the cocoa beans; (2) annual consumption expenditure is the household expenditure for the preceding year covering 12 months; (3) per capita consumption expenditure is consumption adjusted for adult equivalents. The study used the “square root scale” used by the Organization of Economic Co-operation and Development (OECD) to compute consumption expenditure per capita; (4) productive farm asset is the aggregated value of farm assets such as motorised and hand-spraying machines, tricycles, motorbikes, and so on.

8. Western region was the reference category due to its dominance in cocoa production in Ghana and being the largest sample size in the dataset.

9. From the survey and interactions with farmers, the study found that one contributing factor of fertiliser adoption is farmers’ perception of the soil fertility status of their cocoa farms.

10. Note that all the four PSM techniques (nearest-neighbour, kernel-based, radius, and stratification) provide similar results for the common support condition and the balancing tests on covariates. The study, therefore, presents results from the nearest-neighbour matching.

11. Total bias reduction is calculated as $\mathrm{BR}=100\times \left(1-{\displaystyle \frac{B_{\mathrm{after}}}{B_{\mathrm{before}}}}\right)$where $B_{\mathrm{before}}$ and $B_{\mathrm{after}}$denote bias before and after matching, respectively (Caliendo and Kopeinig 2008).