Abstract
Purpose
The purpose of this study is to examine the relative importance of climate normals (average long-term temperature and precipitation) in explaining net farm revenue per hectare (NRh) for supplementary irrigated and rainfed cocoa farms in Nigeria.
Design/methodology/approach
NRh was estimated for 280 cocoa farmers sampled across seven Nigerian states. It was regressed on climate, household socio-economic characteristics and other control variables by using a Ricardian analytical framework. Marginal calculations were used to isolate the effects of climate change (CC) on cocoa farm revenues under supplementary irrigated and rainfed conditions. Future impacts of CC were simulated using Six CORDEX regional climate model (RCM) ensemble between 2036-2065 and 2071-2100.
Findings
Results indicate high sensitivity of NRh to Nigerian climate normals depending on whether farms use supplementary irrigation. Average annual temperature increases and precipitation decreases are associated with NRh losses for rainfed farms and gains for supplementary irrigated cocoa farms. Projections of future CC impacts suggest a wide range of NRh outcomes on supplementary irrigated and rainfed farm revenues, demonstrating the importance of irrigation as an effective adaptation strategy in Nigeria.
Originality/value
This paper uses novel data sets for simulating future CC impacts on land values in Nigeria. CORDEX data constitute the most comprehensive RCMs projections available for Africa.
Keywords
Citation
Fonta, W.M., Kedir, A.M., Bossa, A.Y., Greenough, K.M., Sylla, B.M. and Ayuk, E.T. (2018), "A Ricardian valuation of the impact of climate change on Nigerian cocoa production: Insight for adaptation policy", International Journal of Climate Change Strategies and Management, Vol. 10 No. 5, pp. 689-710. https://doi.org/10.1108/IJCCSM-05-2016-0074
Publisher
:Emerald Publishing Limited
Copyright © 2018, William M. Fonta, Abbi M. Kedir, Aymar Y. Bossa, Karen M. Greenough, Bamba M. Sylla and Elias T. Ayuk.
License
Published by Emerald Publishing Limited. This article is published under the Creative Commons Attribution (CC BY 4.0) licence. Anyone may reproduce, distribute, translate and create derivative works of this article (for both commercial and non-commercial purposes), subject to full attribution to the original publication and authors. The full terms of this licence may be seen at http://creativecommons.org/licences/by/4.0/legalcode
1. Introduction
Agriculture is a very climate-sensitive sector, and climate-smart agriculture is the way forward to increase agricultural productivity in a sustainable manner. According to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, climate change (CC) will amplify existing stresses on agricultural systems, particularly those in Africa for several reasons (IPCC, 2014). First, about 80 per cent of African agriculture is still mainly rainfed and therefore highly vulnerable to changes in climatic conditions such as droughts, higher temperatures and reduced precipitation levels (World Bank, 2008; Hassan, 2010). Second, African agriculture is mostly extensive and practiced on relatively fragile environments and poor quality soils, with little use of modern inputs and farming methods to cope with CC impacts (Mendelsohn and Dinar, 1999; Mano and Nhemachena, 2006). Third, the presence of multiple stresses such as endemic poverty, weak institutions, inadequate health services, limited access to capital and markets, poor infrastructure and technology and conflicts over natural resources reduce farm households’ adaptive capacity to manage the numerous vagaries of CC (IPCC, 2007; Odingo, 2008; Hassan, 2010). Finally, most African governments devote meager financial resources to their agricultural sectors, thereby reducing investment in scientific research needed to better understand and respond to CC impacts (Hassan, 2010). This investment neglect is persistent across many African countries despite the well-recognized poverty-reducing impact of agriculture (Christiaensen et al., 2006).
There is growing concern that CC will be more severe in the west African region because of the problems of huge spatio-temporal rainfall variability, population pressure and limited adaptation and mitigation capacities (Berg et al., 2009; Ibrahim et al., 2014; IPCC, 2014). The impact is expected to be even more severe for a country like Nigeria, where rainfed agriculture is the main source of livelihood for about 50-60 per cent of the population and accounts for over 25.5 per cent of the nation’s gross domestic product in real terms (National Bureau of Statistics, 2017). Most of the agricultural value added in Nigeria is mainly from the plantation tree crops sub-sector (i.e. cocoa, rubber, coffee and palm produce) which is highly climate-sensitive. For instance, between 2006 and 2012, a total of about 1.2 million tons of output from plantation tree crops were exported from Nigeria, representing over 40 per cent of the total export value derived from agricultural exports alone (Central Bank of Nigeria, 2014). Cocoa’s contribution to the nation’s economic development is not in doubt. No other agricultural export commodity has earned more foreign exchange than cocoa. The sub-sector offers substantial employment, both directly and indirectly, and supplies significant volumes of raw materials to cocoa-producing states, as well as revenue to their respective provincial governments. Because of its importance, the Federal Government of Nigeria, concerned with diversifying the nation’s export base, has identified cocoa as the most important export tree crop.
However, there are growing concerns that CC may be impacting negatively on cocoa production in Nigeria. The first real evidence of CC impacts on cocoa production was during the 1972/73 drought, when production declined from about 216,000 metric tons to less than 150,000 metric tons (Central Bank of Nigeria, 2014). Thereafter, the production trend has consistently declined despite the agricultural conditionality of the structural adjustment program that was introduced immediately after the major drought occurrence. Some of the key questions that still remain unanswered and of great concern to policy makers are follows:
What proportion of change in cocoa production is due to the impacts of CC?
What are the economic implications of CC on cocoa production in Nigeria?
How do cocoa farmers in Nigeria adapt to changing climatic conditions?
As climate variables worsen, what is the likely future for this important plantation tree crop in the country?
The academic literature on CC impact on Nigerian cocoa agriculture in general is still very scanty, despite the importance of cocoa as a major export earner. In fact, a review of the extant literature suggests that there are few peer-reviewed documented studies for Nigeria. These include the works of Lawal and Emaku (2007), Omolaja et al. (2009), Ajewole and Iyanda (2010), Ajayi et al. (2010), Fonta et al. (2011), Nwachukwu et al. (2012) and Oyekale (2017). Lawal and Emaku (2007) find that while temperature is positively correlated with cocoa yield, rainfall and relative humidity are negatively correlated with cocoa yield. Omolaja et al. (2009) find that increasing precipitation and favorable temperature promote the flowering intensity of cocoa in Nigeria. For Ajewole and Iyanda (2010), cocoa yield is less sensitive to decreasing rainfall but more sensitive to increasing temperatures. In line with these findings, Ajayi et al. (2010) find cocoa yield also less sensitive to annual rainfall levels, compared to temperature increases. Contrastingly, Nwachukwu et al. (2012) finds cocoa productivity to be more rainfall-sensitive. Similar findings were obtained by Fonta et al. (2011) and Oyekale (2017). In both papers, net cocoa revenue per household farm hectare and yields was more sensitive to increasing precipitations.
Given the dearth of empirical literature on the impacts of CC on Nigerian agriculture, this study seeks to contribute to the academic literature by investigating the economic impacts of CC on Nigerian cocoa agriculture. It intends to do so by using a climate-land value analytical framework (i.e. a theoretical Ricardian cross-section model). The framework is applied to seven different Nigerian cocoa-producing states to:
assess the potential economic impacts of CC on Nigeria cocoa agriculture under supplementary irrigated and rainfed farming conditions;
evaluate the importance of irrigation (supplementary) as an alternative pathway to CC mitigation in Nigeria; and
examine the likely future impacts of CC on cocoa farming in Nigeria.
The rest of the paper is structured as follows. In Section 2, the authors present the Ricardian model (RM) developed for the analysis; Section 3 describes the data and methodology, followed by a description of empirical results in projections with climate scenarios in Section 4. Section 5 concludes the paper with potential policy implications of the findings and suggests policy response options or recommendations to help mitigate the effects of CC on Nigerian tree crop agriculture.
2. Analytical framework
In his pioneering work on the theory of “economic rents” in 1817, David Ricardo (1817) observed that land values reflect land productivity at a site under a perfectly free competition market structure without monopoly. This invariably implies that any factor of production (e.g. capital, labor, technical progress, conducive weather or their combination) that influences land productivity will be reflected in sale value of land or land rent. As climate affects land productivity through cultivation, it is intuitive to assume that the value of land (rents) contains information about the value of climate as one attribute of land productivity. This is the theoretical basis of the Ricardian approach pioneered by Mendelsohn et al. (1994), which is adopted to predict the potential economic damage to US agriculture from CC. Specifically, Mendelsohn et al. (1994) showed that by regressing land value (rents) on climate, household covariates and other environmental control variables, it is possible to measure the marginal contribution of each variable to land rents as capitalized in land value. However, in specific cases like in most developing countries where land values (rents) are difficult to compute because of the absence of well-functioning land markets, Dinar et al. (1998) and Mendelsohn et al. (2001, 2009) suggest the use of net revenue per farm hectare (NRh). Hence, this principle is usually captured through a net farm revenue equation (Mendelsohn et al., 2009) of the form:
Because values far above or below preferred climatic conditions reduce crop productivity, the relationship between NRh and the climatic variables (C) is non-linear and postulated to be inverted U-shaped (Reinsborough, 2003; Seo et al., 2005; Kurukulasuriya and Mendelsohn, 2008; Deressa and Hassan, 2009; Mendelsohn et al., 2009). To control for this potential non-linear relationship in the empirical specification of the NRh function, the authors adopt a quadratic functional form on the CC indicator variable, C:
One major advantage of using NRh is that it accounts for the direct impacts of climate on yields of different crops, as well as the indirect substitution of different inputs, the introduction of different activities and other possible adaptations by farmers to different climatic shocks (Mendelsohn and Dinar, 1999). However, the RM has been extensively criticized on several grounds, first because crops evaluated under the method are not subject to controlled experiments across farms as with the production function model, the agro-economic model or the agro-ecological zone model (Hassan, 2010). To control for this concern, the authors included other important variables such as soil quality and market access in the model, as suggested by Mendelsohn and Dinar (1999). Second, it fails to account for future change in technology, policies and institutions (Mendelsohn et al., 2009). Third, it assumes that input and output prices remain constant, introducing a potential bias in the analysis (Mendelsohn et al., 2009), and fourth, it does not account for the effect of factors consistent across space, e.g. carbon dioxide concentrations being beneficial to crops (Seo et al., 2005; Mendelsohn et al., 2009; Hassan, 2010). Finally, the model may overestimate potential damage caused by CC to farmland values when irrigation is not factored into account (Reinsborough, 2003; Seo et al., 2005).
This study recognizes the various problems highlighted above and attempts to address them to generate reliable results. However, one also should not expect drastic price changes due to CC. Equally, to make the analysis less prone to potential empirical concerns, one can include carbon dioxide exogenously in the analysis, new technology and irrigation as important control variables.
3. Data and econometric procedure
3.1 Data description
Temperature and precipitation data are key to climate-land value analysis. Because CC involves long-term trends, the use of monthly climate normals is usually preferable (Reinsborough, 2003). To determine seasonal temperatures and precipitations, this study used January-December monthly means from 1981 to 2010 from 32 weather stations around the country from Nigeria’s Meteorological Agency. Data on soil types for the seven cocoa-producing states in Nigeria were obtained from the Food and Agricultural Organization (FAO). The FAO soil statistics include information about the major and minor soils in each location, as well as the slope and texture. In all, there exists four types of soil in the respective states, and all of them were used in the analysis. Farm-level data on NRh and its determinants were collected from 280 cocoa farmers spread over the various agro-ecological zones of the seven selected states – Cross River, Abia, Edo, Ondo, Ekiti, Oyo and Ogun (Figure 1). These states represent the major cocoa-producing regions in the country with significant variations in temperatures and precipitations driven primarily by elevation. Four enumeration areas (EAs) were used in each state, for a total of 28 EAs. From each EA, ten farmers were randomly selected from farmers with a cocoa production record of more than 20 years (i.e. long enough to have experienced the CC effects). To collect data, enumerators with extensive fieldwork experience were recruited through Nigeria’s National Bureau of Statistics.
The questionnaire contained seven sections that gathered information on household demographic characteristics, the employment status of the household head, land tenure and household labor composition for farming activities, including labor costs. In the fourth section, responses from more in-depth questions concerning farming activities were collected, such as primary crops grown, harvested and sold in the past twelve months; average yields obtained in a normal year; water source for farming, including the specific irrigation system used for farming; animal ownership in terms of the total number of livestock, poultry or other farm animals kept, sold, lost and consumed during the last growing season; and various costs associated with seeds, fertilizer and pesticides purchased, as well as those related to the use of farm machinery (light, heavy and animal power) and the cost of farm buildings. The last sections of the questionnaire helped to collect data on access to sources of information on farming activities and costs involved in obtaining information; estimates of the household’s total income (for both farming and non-farming activities), taxes paid and subsidies received and farmers’ perceptions of short- and long-term CC effects and their adaptation strategies in response to these changes.
Based on the answers to these questions, it was possible to calculate, for each of the 280 sampled households, the net farm revenue per cocoa hectare (NRh), defined as the difference between gross revenue per farm hectare and the per hectare costs of production, which includes seeds, fertilizer, pesticide, insecticide, herbicide, farm labor, depreciation on machineries and other farming costs.
3.2 Econometric procedure
In the empirical estimation procedure, NRh was regressed on climate, household socioeconomic characteristics and other environmental and control variables. As none of the explanatory variables (e.g. climate variables, household socio-economic condition variables and soil type variables) are endogenous, a log-linear ordinary least squares technique was used to generate the coefficient estimates. Therefore, the data at hand and the nature of the variables considered do not necessitate the use of instrumental variables (IVs) or other econometric models that are designed to account for potential endogeneity for identification of estimated coefficients. Following Reinsborough (2003) and Seo et al. (2005), after several trials and iterations of different definitions of seasonal climate, three seasons were used to define our climate variables that correspond to the three predominant seasons experienced in Nigeria, that is, the dry season (average precipitation and temperature data for October-March), the rainy season (average precipitation and temperature data for April-September) and the Harmattan period[1] (average precipitation and temperature data for December-February).
As indicated earlier, both linear and squared terms of the temperature and precipitation variables were included in the empirical specification of equation (2) to capture the optimum temperature and precipitation levels of cocoa and account for a potential non-linear relationship between CC and net farm revenue. When the coefficient of the quadratic term is positive with a negative linear term, the climate response function is U-shaped, and when the linear term is positive and the quadratic term is negative, the function is hill-shaped or bell-shaped (Reinsborough, 2003). However, note that because seasonal climate variables are often used, the process is somehow complex and likely to result in a mixture of positive and negative quadratic coefficients across seasons (Mendelsohn et al., 2009).
In addition, the climate variables were specified in their interactive forms to check whether climate effects on NRh are dependent on the season under consideration. Hence, the analysis includes interacting dry season temperature with precipitation, rainy season precipitation with temperature and Harmattan temperature with precipitation. With regards to other control variables, urban/rural characteristics, farmland altitude and irrigation, among others, have been shown to significantly affect land value or net farm revenue other than CC alone (Reinsborough, 2003; Seo et al., 2005; Ouedraogo et al., 2006; Kurukulasuriya and Mendelsohn, 2006; Mendelsohn et al., 2009). For instance, to account for demographic and other important factors in the empirical specification of the climate response function [equation (2)], population density was incorporated to capture urban/rural characteristics, main source of water (irrigation or not) and farm altitude.
Furthermore, two sets of climatic scenarios were used to examine the likely future impacts of CC on Nigerian cocoa agriculture. The first set, comprising two representative concentration pathways, i.e. RCP4.5 and RCP8.5 (Moss et al., 2010), and two future time slices (2036-2065 and 2071-2100), was generated from a multimodel ensemble of regional climate models (RCMs) run under the recent coordinated regional climate downscaling experiment (CORDEX) initiative (Giorgi et al., 2009). The CORDEX RCMs downscale many global climate models (GCMs) from the Coupled Model Intercomparison Project, Phase 5 (Taylor et al., 2012; Sylla et al., 2016). The RCMs are run over the whole of Africa at 50 km resolution. To date, the CORDEX data constitute the most comprehensive RCMs projections available for the continent. The authors are not aware of any African application that has utilized CORDEX projections.
Figure 2 shows absolute precipitation and temperature changes (Future minus Historical; in mm d-1 and Kelvin, respectively) for each of the CORDEX RCMs averaged over the whole of Nigeria and for both RCP4.5 and RCP8.5 during the two future time slices (2036-2065 and 2071-2100). In general, an increase in precipitation is projected in all cases, except for the Climate Limited-area Modelling Community (CCLM), which produces a substantial decrease for the RCP8.5 during 2071-2100. The highest precipitation increase of more than 1 mm d-1 is simulated by RegCM4 of the International Centre for Theoretical Physics, Italy. Although the multimodel ensemble projects more rainfall, it is clear that the various CORDEX RCMs produce different magnitude and sign of the precipitation changes over Nigeria. However, the temperature change (i.e. Figure 2) shows a consistent warming across all RCMs with the highest change (more than 4°C) projected during 2071-2100 for the RCP8.5.
Because the CORDEX ensemble data (Table I), which show increases in all variables (rainfall and temperature) regardless of the scenarios (RCP4.5 and RCP8.5), balance out the decreases in precipitation predicted by the regional climate model CCLM, the authors not only use CORDEX but also base the second set of scenarios on hypotheses of 2.5°C increase in temperature and 5 per cent decrease in precipitation from the historical observations (1981-2010). Similar hypotheses (i.e. +2.5 to 5°C and 7 to 14per cent decrease in precipitation) were tested earlier for African croplands by Kurukulasuriya and Mendelsohn (2006) and Mano and Nhemachena (2006). The projected changes from the different climate scenarios generated are summarized in Table I.
4. Results and discussion
4.1 Sample statistics
The summary statistics of the key variables used in the analysis are reported in the Appendix. On average, the NRh was Nigerian Naira 42,558.9 (hereafter NGN) for all farms, NGN45,339.12 for supplementary irrigated farms and NGN31,386.7 for rainfed farms. Mean temperatures and precipitations corresponding to the three seasons were 32.4°C and 47.7 mm for the hot dry season, 25.7°C and 169 mm for the heavy rainy season and 19.6°C and 12.7 mm for the Harmattan season. The soil type on which the farmers operate is a function of geographical location. The soil types are:
La: These are ferralitic soils of the coastal plain sand and escarpment; dominant color is yellowish brown (not differentiated).
Jc: These are ferruginous tropical soils on crystalline acid rocks.
Li: These are ferrallite soils, and the dominant color is red on loose sandy sediments.
LVf: These are ferric luvisol soils, which are seen as reddish sandy clay loam.
More than 42.9 per cent of the sampled farmers cultivated on La soil type, about 28.7 per cent of them farmed on the Jc soil type and about 28.4 per cent of the farmers were split between the Li and LVf soil types.
The average area devoted to cocoa cultivation was about 2.4 hectares from a total average household farm size of 6.5 hectares, suggesting that a non-negligible proportion (about 31 per cent) of the total available farming land is dedicated to cocoa. Men-headed households dominated the sample (95 per cent), with an average age for the entire sample of about 55 years, with 22 years of farming experience and about 9 years of schooling (i.e. junior secondary school). On average, nearly three household members worked in cocoa farming as their primary occupation from a total household size of about seven members. Only about 20 per cent of sampled farmers reported using irrigated farming (i.e. supplementary irrigation); the rest relied only on rain-fed agriculture. Less than 33 per cent of the farmers acknowledged having access to any form of credit facilities; on the other hand, about 14 per cent reported receiving farm subsidies. Yearly average use of fertilizer came to about 776 kg for the whole farm, and more than 93 per cent of farmers reported using pesticides. Average farm visit time from an agricultural extension worker was estimated to be less than 2 h a week; however, more than 66 per cent of sampled farmers reported having received advice from an agricultural extension worker in the last 12 months preceding the survey. More than 72 per cent of farmers reported using multiple farmlands for cocoa farming, and about 74 per cent reported practicing mixed farming. Finally, more than 52 per cent sold their cocoa produce in urban areas, with a mean market distance of about 90 km.
4.2 Ricardian estimates
Table II presents three log linear regressions or models on determinants of NRh for Nigerian cocoa farms. In Column two (Model 1), NRh is regressed on climate variables alone. Though this model fits well, as many of the climatic variables including their squared terms are statistically significant, it does not account for the effects of household socio-economic characteristics on NRh, an important category of independent variables reflecting farmlands’ potential for alternative uses (Reinsborough 2003). Notwithstanding, Model 1 highlights the direction and magnitude of climate impacts on NRh. The implications of these findings are taken up later.
In the second regression or Model 2 (i.e. Column 3 of Table II), net farm revenue is regressed on climate and climate interaction effect variables conditional on seasons (dry, rainy and Harmattan). Note that many of the climatic variables which were statistically significant in the first model are dropped because of multicollinearity. Equally, it was also observed that many of the squared terms for temperature and precipitation remained statistically significant as in the first model and unchanged in sign and magnitude. However, the squared term of rainy season precipitation is no longer significant, though negative, as in Model 1. This implies that there is an optimal level of precipitation during the rainy season that, above and beyond, NRh decreases. Of the temperature–precipitation interaction effect variables, rainy season temperature–precipitation interaction has a very large significant negative effect on net farm revenue, whereas Harmattan temperature–precipitation interaction has a significant positive effect on net farm revenue. The former suggests that during the rainy season, net farm revenue decreases for hotter and wetter cocoa farms; while during Harmattan, net farm revenue increases for hotter and wetter cocoa farms.
Column 4 or Model 3 corresponds to the climate response function [equation (2)], where NRh is regressed on climate variables (C), their squared terms (C2), soil types and three important control variables (S), including household socio-economic characteristics (Z). This model, which comprehensively includes as many determining variables as possible, provides the best fit, with results one would expect, and is therefore used for all subsequent climate calculations. The adjusted R2 for the model has a higher goodness of fit value of 0.46. The effects of some of the significant variables from the climate “only” Model 1 are unchanged in sign and quite similar in magnitude. Dry season temperatures and precipitations are significant, with large negative effects on NRh as in Model 1. The squared terms for dry and rainy season remained significant and unchanged in sign as in Model 1, indicating that above optimal levels of temperature–precipitation NRh decreases. The four soil types all have a large significant negative effect on net revenue per cocoa farm hectare as expected and consistent with most previous Ricardian findings in Africa. This is perhaps so because African farmers often make extensive use of chemical fertilizers and pesticides, including agricultural practices that enhances soil degradation and worsens soil infertility. This partly explains why net farm revenue is often negatively correlated with poor quality soils in Africa (Kurukulasuriya and Mendelsohn, 2006; Mano and Nhemachena, 2006; Mendelsohn et al., 2009).
Of control and socio-economic variables, irrigation, population density, access to credit and distance to urban markets are significant as expected. Farmers with greater access to credit facilities have higher NRh, a predictable empirical finding considering the capital-intensive nature of cocoa farming. This has an important policy implication with regard to the financial resource allocation by the banking sector for agriculture. Similarly, farmers closest to urban markets have higher NRh, probably because of market proximity and lower transportation cost. Hence, distance to urban markets has a statistically significant negative coefficient. Irrigation has a beneficial effect on NRh, suggesting its potential use as an adaptive intervention to mitigate the impact of CC. On the other hand, population density was found to be detrimental to NRh, as its significant negative coefficient shows.
Column 2 of Table III reports the marginal impacts of climate on NRh of the full model (i.e. Model 3) reported in Table II (i.e. Column 4). The marginal analysis simply shows the effect of an infinitesimal change in temperature and precipitation on cocoa farming in Nigeria. As observed (i.e. Row 6 of Column 2 in Table III), the annual temperature impact is close to NGN−5,698.4 or about US$38.0 per degree celsius evaluated at the mean of the sample. However, the most harmful temperature effects are associated with the dry and rainy season. Conversely, for the precipitation impacts, the marginal calculations also revealed that annual precipitation impact is NGN40 or about US$0.30 increase in net revenue per mm/month, and the most harmful precipitation effect is largely due to dry season precipitation. The marginal impact analysis equally reveals that while temperature increase is strictly detrimental to cocoa farming in Nigeria, infinitesimal precipitation increase is strictly beneficial to cocoa farms in general. This may be linked to the use of supplementary irrigation by cocoa farmers. The implication of this can be profound for Nigeria and other African and developing countries, and we discuss it in the subsequent paragraphs.
Table IV presents a comparison of cocoa farms with supplementary irrigation versus rainfed farms. In the irrigated farm model, the inclusion of the quadratic terms complicated the estimation as many climatic variables were dropped because of collinearity. Similarly, inclusion of the temperature–precipitation interaction effect and control variables produced unrealistic confidence intervals for the estimates. By omitting the C2 and S variables, more sensible and precise estimates for irrigation and rainfed farms were obtained; these are presented in Column 2 and Column 3, respectively. Many significant differences are obtained between the estimates of the two models. The irrigated model, for example, shows that high temperatures in the rainy and Harmattan seasons are best for NRh, whereas dry season precipitations are best for rainfed farms. Irrigated NRh is significantly affected by Li soil types and rainfed NRh is significantly affected by La, whereas both farm types are affected by Jc soil types. Of the socio-economic variables, number of years of farming cocoa is more influential for irrigated farm revenues than rainfed farms, while distance to urban markets significantly influences rainfed NRh more than irrigated NRh, as does farm altitude.
The marginal calculations of Columns 3 and 4 of Table III show the marginal impacts of climate on the irrigated and rainfed regression models based on the results of Table IV. The annual temperature marginals for rainfed farms are much larger than those for supplementary irrigated farms, with an impact of NGN5,747 per degree celsius, in constrast to NGN10,393.3 per degree celsius for rainfed farms. Harmful temperature effects are associated with rainfed farms in all the three seasons, while in the irrigated model, they occur largely in the dry season, with beneficiary effects in the rainy and Harmattan seasons. Similarly, annual changes associated with increased precipitation impacts show net gains of NGN191.6 per mm per month for irrigated farms and NGN128.9 per mm per month for rainfed farms. In general, irrigated farms are more sensitive to increased precipitation during the rainy season, while rainfed farms are most vulnerable to decreased precipitation during both dry and rainy seasons.
4.3 Projections with climate scenarios
There are many contentious issues surrounding the use of estimated Ricardian results to predict how temperature and precipitation will affect future net farmland revenues (Reinsborough, 2003, p. 31-32). First, estimating beyond the range of observed data may be problematic, especially when the estimated relationships are not exactly as expected. For instance, one expects negative square terms of the climatic variables, assuming cocoa has optimum temperature and precipitation levels. However, while the squared terms of dry and rainy season temperature and precipitation variables exhibit a hill-shaped relationship, Harmattan temperature and precipitation squared exhibit a U-shaped relationship, possibly making an accurate forecasting of future CC effects on the cocoa farms’ NRh problematic. Second, variables such as soil types and farmers’ socio-economic characteristics, which are assumed to be constant with temperature and precipitation changes, will certainly be affected in reality. For instance, increased rainfall and sunshine affect soil moisture content and hence plant growth and productivity (Deressa and Hassan, 2009). However, as the changes to be estimated are quite moderate, using the results of the full model (i.e. Table II, Column 3) and those of Table IV to predict future climate impacts on NRh for all farms, irrigated and rainfed, should be not be too problematic. The aim of the projection is not to examine how NRh actually changes, but it is simply to isolate the statistically significant effect of CC on NRh, assuming all other conditions (e.g. price changes investment, population and the use of technology) are held constant (Mendelsohn et al., 2009).
The first set of CC scenarios (i.e. Column 1 of Table I), generated from the CORDEX Ensemble of six RCMs, were fitted into the results of Table II’s full model (Column 4) and those of Table IV (Columns 2 and 3) to examine how future CCs would affect NRh under both irrigated and rainfed conditions. The simulated results are reported in Tables V and VI. Annual and seasonal impacts vary a great deal between the different farms. In the RCP4.5 scenario, the calculations reveal marked variations in NRh across the different farms. Both scenarios predict drastic dry season declines in NRh between 2036-2065 and 2071-2100 for rainfed farms. Seasonal losses range from NGN−58.0 to −83 for all farms and from NGN−110.9 to 148.8 for rainfed farms. A striking prediction from the CORDEX Ensemble suggests that irrigated farm revenues will increase across all seasons, with the most beneficial effects expected during the Harmattan season. However, in terms of annual revenue losses, only rainfed farms are expected to record future NRh decline, with expected losses of NGN−25.2 and NGN−29.1 for 2036-2065 and 2071-2100, respectively. The projections suggest that the increase in benefits to farmers can be gained incrementally via adaptive interventions such as irrigation.
In the second scenario (Table VI), first, a 2.5°C increase in temperature only is associated with seasonal net farm revenues per hectare loss of NGN−126.2 ($−0.84) for all farms and NGN−153.9 ($−1.03) for rainfed farms during the period (2050-2100). These losses are expected also during the dry season as initially predicted by the six CORDEX RCMs ensemble. Second, reducing rainfall by 5 per cent is equally associated with net farm revenues loss of NGN−108.5 ($−0.72) for all farms and NGN−151.2 ($−1.0) for rainfed farms during the dry season per hectare of farmland. Similar seasonal losses are associated when a simultaneous 2.5°C increase in temperature and a 5 per cent reduction in rainfall are considered. The combined effect is a reduction in NRh of about NGN−125.8 ($−0.84) for all farms and NGN−173.9 ($−1.2) for rainfed farms during this same period. Irrigated farms are, in general, less sensitive to any of these changes, as was predicted by the six CORDEX RCMs ensemble. However, note that while the six CORDEX RCMs ensemble predicts that future net revenues decline only for rainfed farms, the second-case climate scenarios considered in the paper predict that annual net revenues decline for all farms and rainfed farms.
5. Conclusions
Using 20 years of data, this paper assesses the economic implications of CC (i.e. average long-term temperature and precipitation changes) on NRh in Nigerian cocoa farms under rainfed and irrigated (supplementary) conditions. The results indicate a high sensitivity of NRh to climate normals, dependent on whether cocoa farms are irrigated. In general, rainfed farms’ revenues were more sensitive to marginal changes in both temperature and precipitation than irrigated farms. For instance, the annual temperature marginals for irrigated cocoa farms is about NGN5,747 ($38.3) per degree celsius compared to NGN−10,393.3 ($−69.3) per degree celsius for rainfed farms. Similarly, the annual changes associated with precipitation change are a net gain of NGN191.6 ($1.3) per mm per month for irrigated farms compared to NGN128.9 ($0.86) per mm per month for rainfed farms. Furthermore, the marginal analysis reveals that irrigated farms are sensitive to decreased precipitation only during the rainy season, whereas rainfed farms are most vulnerable in both dry and rainy seasons.
Two sets of future climate impact projections are included in the analysis. The first set of scenarios, based on the CORDEX ensemble, suggests a wide range of outcomes on NRh for all farms between 2036-2065 and 2071-2100. Dry season losses range from NGN−110.9 ($−0.74) to NGN−148.8 ($−1.0) for rainfed farms and from NGN−58.0 ($−0.39) to NGN−83 ($−0.55) for all farms. Irrigated farms are generally less sensitive to the different CORDEX scenarios. The results suggest the need to disaggregate farms by type for a reliable CC impact analysis so that policy can be directed where impacts are felt strongly.
The second hypothesized scenario set also predicts drastic declines in NRh per hectare for rainfed farms and all farms between 2050 and 2100. The combined effect of a simultaneous 2.5°C temperature increase and a 5 per cent rainfall reduction are associated with reduction in NRh of NGN−173.9 ($−1.2) for rainfed farms and about NGN−125.8 ($−0.84) for all farms, and again, irrigated farms are generally less sensitive to any changes. It is not surprising to find similar impacts for rainfed and all farms as there are more rainfed farms (224) in the overall sample farms compared to irrigated farms (56). Hence, the statistical results of all farms are mainly driven by the results that pertain to rainfed farms. A study on larger number of sampled irrigated farms will be beneficial in future.
The results clearly demonstrate the importance of supplementary irrigation as an adaptation strategy to reduce harmful CC effects on Nigerian cocoa agriculture. Serious neglect by the national and sub-national Nigerian governments of irrigation farming systems may be partly responsible for the country’s declining trend in cocoa productivity. As a concrete future policy direction, this demands for more investment (not less investment) in irrigation to support local farmers’ management of CC uncertainty. The regression results suggest that access to credit/working capital and proximity to markets by improving the road infrastructure that minimize the distance to final destination of agricultural output create the right developmental incentives for cocoa farmers to generate better farm revenues and improve their welfare. In irrigated farms, the experience of farmers was found to be beneficial. In addition to policy measures that increase investment in irrigation, extension services can be used as a vehicle to share experiences among farmers on best practices of cocoa production and environmental management. This may include pilot initiatives such as weather index-based crop insurance that can support farmers in mitigating adverse and heterogeneous effects of CC and natural catastrophes encountered during farming activities (Barnett, 2014; Barnett and Mahul, 2007; Kunreuther, 1996; FAO, 2011; IFAD, 2011; World Bank, 2010). The results clearly demonstrated the detrimental impact of distance to markets. Therefore, the obvious policy implication is the urgent need to provide conducive market infrastructure by improving access to market nodes such as urban areas. This enhances farmers’ productivity, the welfare of their respective households and the dynamism of the cocoa sub-sector in Nigeria, with potential implications for farmers in other African and other developing countries across the globe.
In conclusion, the authors want to highlight the fact that they are not arguing that irrigation is the only adaptation strategy among other alternative adaptation strategies (e.g. developing drought resistant crops) that can be deployed to cushion the impact of CC. However, the authors believe that irrigation in the Nigerian context has advantages, is manageable in terms of technological capability of the farmers and is relatively cost-effective route to adapt to CCs compared to, for instance, breeding new crops that can withstand climate stresses such as water shortage and drought. Generating and using new breed of crops is based on long-term research and development investments. However, high yielding varieties are not readily available for farmers who cannot wait for many years or decades of field experiments which are required to validate the new crops. In this context, irrigation provides a practically feasible alternative to the harsh CC that Nigerian farmers find themselves in.
Acknowledgments
The fieldwork for the study was generously funded by the African Economic Research Consortium (AERC) through Research Grant No.: RT10509. The insightful comments from three anonymous reviewers and the deputy editor of the journal are gratefully acknowledged. The authors specially thank the German Federal Ministry of Education and Research (BMBF) for the financial support through the West African Science Service Center on CC and Adapted Land Use (WASCAL), where the work was carried out.
Figures
Monthly seasonal temperature and precipitation levels
| Scenario 1 (1) | Scenario 2 (2) | |||||||
|---|---|---|---|---|---|---|---|---|
| (a) CORDEX (temperature projections) | 2.5% precipitation decrease from observations (see Table II) | |||||||
| Historical | RCP4.5 | RCP4.5 | RCP8.5 | RCP8.5 | Tb + 2.5°C | Tb + 2.5°C | Tb + 2.5°C | |
| (all farms) | (rainfed farms) | (irrigated farms) | ||||||
| Period | (1981-2010) | (2036-2065) | (2071-2100) | (2036-2065) | (2071-2100) | (2050-2100) | (2050-2100) | (2050-2100) |
| Dec-Feb | 22.58 | 24.19 | 24.98 | 24.93 | 26.88 | 22.1 | 27.5 | 28.4 |
| Apr-Sep | 25.2 | 26.8 | 27.59 | 27.53 | 29.45 | 28.2 | 27.6 | 27 |
| Oct-Mar | 24.46 | 26.12 | 26.89 | 26.84 | 28.85 | 34.9 | 31.1 | 30.5 |
| (b) CORDEX (precipitation projections) | 5% precipitation decrease from observations (see Table II) | |||||||
| Historical | RCP4.5 | RCP4.5 | RCP8.5 | RCP8.5 | Pb: 5% | Pb: 5% | Pb: 5% | |
| (all farms) | (rainfed farms) | (irrigated farms) | ||||||
| Period | (1981-2010) | (2036-2065) | (2071-2100) | (2036-2065) | (2071-2100) | (2050-2100) | (2050-2100) | (2050-2100) |
| Dec-Feb | 2.79 | 10.5 | 10.53 | 3.24 | 3.42 | 12.065 | 20.14 | 17.1 |
| Apr-Sep | 174.12 | 183.78 | 183.09 | 177.66 | 178.92 | 160.55 | 132.24 | 143.165 |
| Oct-Mar | 30.51 | 38.19 | 38.22 | 30.78 | 32.19 | 44.365 | 35.435 | 32.015 |
Notes: N.B., where Tb: baseline temperature: Pb: baseline precipitation; (a) monthly seasonal temperature (in °C) as generated from CORDEX and hypotheses of +2.5°C from observations (baseline period of 1981-2010); (b) monthly seasonal precipitation (in mm/month) as generated from CORDEX and hypotheses of −5% precipitation decrease from observations (baseline period of 1981-2010)
Loglinear specification results of Ricardian model for all farms
| Climate only (2) | Climate with CC interactiona (3) | Full model without interactionb (4) | ||||
|---|---|---|---|---|---|---|
| Variables | Coef. | t-stat. | Coef. | t-stat. | Coef. | t-stat. |
| Dry season temp. | −0.917 | −0.27 | −1.193 | −2.78 | ||
| Dry season temp.2 | −0.023 | −9.10 | −0.021 | −8.27 | −0.058 | −5.64 |
| Rainy season temp. | −6.568 | −3.52 | −0.079 | −0.71 | ||
| Rainy season temp.2 | −0.0045 | −1.27 | −0.0045 | −1.27 | −0.0032 | −0.96 |
| Harmattan temp. | −8.805 | −7.72 | −8.805 | −7.72 | 0.534 | 1.68 |
| Harmattan temp.2 | 1.877 | 7.79 | 1.877 | 7.79 | 0.075 | 10.47 |
| Dry season precip. | −0.086 | −3.25 | −0.113 | −3.46 | ||
| Dry season precip.2 | 0.00075 | 2.41 | −0.00023 | −5.76 | −0.0010 | −4.49 |
| Rainy season precip. | −0.0123 | −0.40 | 0.088 | 0.22 | ||
| Rainy season precip.2 | −0.00023 | −5.40 | −0.00005 | −0.58 | −0.0009 | −4.80 |
| Harmattan precip. | −0.0035 | −0.12 | −0.210 | −0.56 | ||
| Harmattan precip.2 | 0.00019 | 5.65 | 0.0017 | 6.32 | 0.0061 | 4.82 |
| Dry season temp. × precip. | ||||||
| Rainy season precip. × temp. | −0.0027 | −3.19 | ||||
| Harmattan temp. × precip. | 0.0015 | 2.23 | ||||
| La_soil_type | −3.005 | −7.35 | ||||
| Jc_soil_type | −2.432 | −6.93 | ||||
| Li_soil_type | −2.576 | −7.10 | ||||
| LVf | −2.416 | −6.74 | ||||
| Farm altitude | 0.014 | 1.82 | ||||
| Irrigation | 0.191 | 3.24 | ||||
| Population_density | −0.030 | −3.24 | ||||
| Access_credit | 0.199 | 4.05 | ||||
| Education_head | −0.034 | −1.64 | ||||
| Farm_experience | −0.0037 | −0.60 | ||||
| Crop_area | 0.0045 | 0.43 | ||||
| Household_ size | −0.018 | −1.02 | ||||
| Market_distance (km) | −0.446 | −2.85 | ||||
| No of visit by ext. worker | 0.335 | 1.24 | ||||
| Total farm area (hectare) | −0.067 | −1.53 | ||||
| Constant | 12.9 | 7.64 | 13.1 | 7.72 | 26.4 | 10.4 |
| F-statistics | 12.1 | 12.1 | 13.6 | |||
| Adjusted R2 | 0.29 | 0.27 | 0.46 | |||
| Observations | 280 | 280 | 280 | |||
aShowing estimated results for model with only climatic variables used as the independent variables.
Results for climate response function specified in equation (2); all figures in italic signifies statistical significance of variables at 1% and 5% levels of confidence
Marginal impacts of climate on cocoa NRh (Naira)
| Full model | Irrigated | Rainfed | |
|---|---|---|---|
| Variables | Marginal (2) | Marginal (3) | Marginal (4) |
| Temperature | |||
| Dry season | −9,302.7 | −349.9 | −13,528.7 |
| Rainy season | −7,165.2 | 12,403.8 | −16,118.7 |
| Harmattan season | −627.2 | 5,187.2 | −1,547.6 |
| Annual | −5,698.4 | 5,747.0 | −10,398.3 |
| Precipitation | |||
| Dry season | −218.8 | 192.96 | −368.92 |
| Rainy season | 88.8 | −742.51 | −11.16 |
| Harmattan season | 250.1 | 1,124.34 | 766.94 |
| Annual | 40.0 | 191.60 | 128.95 |
Note: N.B. Marginals of each climate variable calculated at the mean of the sample
Loglinear specification of rainfed and supplementary irrigated farm samples
| Irrigated model (2) | Rainfed model (3) | |||
|---|---|---|---|---|
| Variables | Coef. | t-stat. | Coef. | t-stat. |
| Dry season temp. | −0.0035 | −0.18 | −0.902 | −0.62 |
| Dry season temp.2 | −0.135 | −3.06 | ||
| Rainy season temp. | 0.124 | 3.53 | −0.601 | −1.40 |
| Rainy season temp.2 | −0.012 | −3.67 | ||
| Harmattan temp. | 0.519 | 2.30 | −0.315 | −0.50 |
| Harmattan temp.2 | 0.034 | 8.51 | ||
| Dry season precip. | 0.0019 | 0.32 | 0.104 | 3.23 |
| Dry season precip.2 | −0.00005 | −1.17 | ||
| Rainy season precip. | −0.0074 | −1.75 | 0.191 | 0.02 |
| Rainy season precip.2 | −0.00003 | −2.06 | ||
| Harmattan precip. | 0.0112 | 1.82 | −0.808 | −0.65 |
| Harmattan precip.2 | 0.0290 | 0.02 | ||
| Rainy season precip. × temp. | ||||
| Harmattan temp. × precip. | 0.00060 | 2.00 | ||
| La | 2.927 | 9.62 | ||
| Jc | −0.106 | −2.47 | −0.239 | −5.75 |
| Li | 2.927 | 9.62 | ||
| LVf | ||||
| Farm altitude | −0.025 | −9.23 | ||
| Access_credit | 0.083 | 3.82 | 0.180 | 3.10 |
| Farm_experience | −0.011 | −2.63 | −0.032 | −1.91 |
| Cropland | −0.0020 | −0.69 | −0.213 | −3.12 |
| Household_size | 0.0036 | 0.75 | −0.0150 | −1.08 |
| Market_distance (Km) | −0.051 | −1.89 | −0.045 | −2.85 |
| Constant | 16.4 | 2.20 | 21.4 | 8.56 |
| F-statistics | 37.2 | 17.1 | ||
| Adjusted R2 | 0.87 | 0.56 | ||
| Observations | 56 | 224 | ||
Note: All figures in italic signifies statistical significance of variables at 1% and 5% levels of confidence
Impacts of CORDEX ensemble (six RCMs) scenarios on cocoa NRh
| CORDEX scenarios | |||||
|---|---|---|---|---|---|
| Historical | RCP4.5 | RCP4.5 | RCP8.5 | RCP8.5 | |
| Seasonal variables | (1981-2010) | (2036-2065) | (2071-2100) | (2036-2065) | (2071-2100) |
| All farms | |||||
| Dec-Feb (Harmattan) | 76.16 | 81.67 | 85.00 | 85.71 | 94.30 |
| Apr-Sep (rainy season) | 10.41 | 7.76 | 7.73 | 9.03 | 8.23 |
| Oct-Mar (dry season) | −58.01 | −68.53 | −72.92 | −71.28 | −83.44 |
| Annual | 9.52 | 6.97 | 6.60 | 7.82 | 6.36 |
| Rainfed farms | |||||
| Dec-Feb (Harmattan) | 29.59 | 28.39 | 29.45 | 32.36 | 35.07 |
| Apr-Sep (rainy season) | 30.98 | 30.76 | 29.65 | 28.75 | 26.51 |
| Oct-Mar (dry season) | −110.94 | −134.72 | −140.97 | −129.93 | −148.77 |
| Annual | −16.79 | −25.19 | −27.29 | −22.94 | −29.06 |
| Supplementary irrigated farms | |||||
| Dec-Feb (Harmattan) | 28.15 | 29.07 | 29.48 | 29.37 | 30.39 |
| Apr-Sep (rainy season) | 18.24 | 18.36 | 18.47 | 18.50 | 18.73 |
| Oct-Mar (dry season) | 16.37 | 16.38 | 16.38 | 16.36 | 16.36 |
| Annual | 20.92 | 21.27 | 21.44 | 21.41 | 21.83 |
Impacts of +2.5°C and −5% changes in temperature and precipitation on cocoa NRh
| IPCC Scenarios based | ||||
|---|---|---|---|---|
| Baseline Tb and Pb |
Tb + 2.5°C and Pb |
Tb and Pb −5% Pb |
Tb + 2.5°C and Pb − 5% Pb | |
| Seasonal variables | (1981-2010) | (2050-2100) | (2050-2100) | (2050-2100) |
| All farms | ||||
| Dec-Feb (Harmattan) | 64.00 | 73.15 | 64.03 | 73.19 |
| Apr-Sep (rainy season) | 11.42 | 10.79 | 13.19 | 12.56 |
| Oct-Mar (dry season) | −108.94 | −126.22 | −108.46 | −125.75 |
| Annual | −11.17 | −14.09 | −10.42 | −13.33 |
| Rainfed farms | ||||
| Dec-Feb (Harmattan) | 30.68 | 34.35 | 30.26 | 33.94 |
| Apr-Sep (rainy season) | 24.76 | 21.68 | 23.49 | 20.40 |
| Oct-Mar (dry season) | −153.93 | −176.34 | −151.19 | −173.59 |
| Annual | −32.83 | −40.10 | −32.48 | −39.75 |
| Supplementary irrigated farms | ||||
| Dec-Feb (Harmattan) | 30.04 | 31.34 | 30.03 | 31.33 |
| Apr-Sep (rainy season) | 18.32 | 18.63 | 18.38 | 18.69 |
| Oct-Mar (dry season) | 16.37 | 16.36 | 16.36 | 16.35 |
| Annual | 21.58 | 22.11 | 21.59 | 22.12 |
Notes: N.B: Tb: baseline temperature; and Pb: baseline precipitation
Summary statistics of the sampled cocoa farms
| All farms | Rainfed | Irrigated | ||||
|---|---|---|---|---|---|---|
| Variable definition and measurement | Mean | SD | Mean | SD | Mean | SD |
| Socioeconomic variables | ||||||
| NRh (in Naira) | 42,558.9 ($283.7) | 46,774.8 ($311.8) | 31,386.7 ($209.2) | 17387.2 ($115.9) | 45,339.1 ($302.3) | 51,202.7($341.4) |
| Age of household head (years) | 55.3 | 12.72 | 56.7 | 13.3 | 52.4 | 11.7 |
| Education (years) | 9.1 | 4.1 | 6.5 | 5.01 | 12.5 | 10.6 |
| Farming experience (cocoa) | 22.8 | 10.8 | 23 | 3.95 | 21.9 | 5.0 |
| Household size (number of persons) | 7.5 | 3.8 | 5.3 | 2.19 | 6.2 | 2.7 |
| Agricultural variables | ||||||
| Total farm area (hectares) | 6.5 | 4.8 | 5.3 | 2.2 | 7.9 | 5.3 |
| Visit from extension worker (number) | 2.5 | 2.6 | 2.6 | 2.9 | 2.2 | 2.4 |
| Farm labor (cocoa farming) | 2.9 | 2.1 | 3.6 | 2.8 | 2.3 | 2.2 |
| Crop area (cocoa hectares) | 2.4 | 1.0 | 1.4 | 1.4 | 1.6 | 1.0 |
| Distance to urban market (km) | 90.5 | 142.7 | 86.4 | 78.5 | 82.9 | 62.6 |
| Aggregate measures (proportions) | ||||||
| % of farms headed by male | 95.0 | 90.0 | 97.0 | |||
| % of farmers with access to credit | 33.6 | 39.0 | 37.0 | |||
| % of farms that received subsidy | 14.0 | 5.5 | 45.5 | |||
| % of farms using suppl. irrigation | 19.9 | 0 | 100 | |||
| % of farmers keeping livestock | 46.7 | 40.0 | 37.0 | |||
| % of farms using pesticides | 93.0 | 95.0 | 97.0 | |||
| % of farms over 5 hectares | 11.5 | 7.4 | 22.6 | |||
| % of farms with extension contacts | 66.0 | 56.9 | 55.6 | |||
| % of farms on La soils | 42.9 | 52.2 | 19.5 | |||
| % of farms on Jc soils | 28.7 | 21.3 | 46.2 | |||
| % of farms on Li soils | 14.2 | 14.2 | 25.6 | |||
| % of farms on LVf soils | 14.2 | 12.3 | 8.8 | |||
| Climate variables | ||||||
| Monthly dry season temp. (°C) | 32.4 | 4.82 | 28.6 | 1.12 | 28.9 | 1.05 |
| Monthly rainy season temp. (°C) | 25.7 | 1.13 | 25.1 | 0.91 | 24.5 | 0.82 |
| Monthly Harmattan temp. (°C) | 19.6 | 6.5 | 25.0 | 3.0 | 25.9 | 3.41 |
| Monthly dry season precip. (mm) | 46.7 | 33.9 | 37.3 | 23.8 | 33.7 | 21.15 |
| Monthly rainy season precip. (mm) | 169.0 | 15.9 | 139.2 | 28.4 | 150.7 | 22.9 |
| Monthly Harmattan precip. (mm) | 12.7 | 21.4 | 21.2 | 13.1 | 18 | 14.5 |
| Control variables | ||||||
| Population density | 251.9 | 19.93 | 251.7 | 19.9 | 252.3 | 20.4 |
| Irrigation | 19.9 | 0.40 | ||||
| Altitude (m) | 1,238.8 | 139.3 | 1,217.3 | 146.3 | 1,332.8 | 110.2 |
| Number of observations | 280 | 224 | 56 | |||
Note
The Harmattan season is peculiar to the West African subcontinent. It usually occurs between the month of December and early March and signifies the beginning of the dry season. It is often characterized by harsh weather conditions such as cold, dry and dust-laden wind, very low humidity and relatively lower temperatures.
Appendix
References
Ajayi, I.R., Afolabi, M.O., Ogunbodede, E.F. and Sunday, A.G. (2010), “Modelling rainfall as a constraining factor for cocoa yield in Ondo State”, American Journal of Scientific and Industrial Research, Vol. 1 No. 2, pp. 127-134.
Ajewole, D.O. and Iyanda, S. (2010), “Effect of climate change on cocoa yield: a case of cocoa research institute (CRIN) farm, Oluyole local government Ibadan, Oyo state, Nigeria”, Journal of Sustainable Development in Africa, Vol. 12 No. 1, pp. 350-358.
Barnett, B. (2014), “Multi-peril crop insurance: successes and challenges”, Agricultural Finance Review, Vol. 74 No. 2, pp. 200-216.
Barnett, B. and Mahul, O. (2007), “Weather index insurance for agriculture and rural areas in lower income countries”, American Journal of Agricultural Economics, Vol. 89 No. 5, pp. 1241-1247.
Berg, A., Quirion, P. and Sultan, B. (2009), “Weather-index drought insurance in Burkina-Faso: assessment of its potential interest to farmers”, Weather, Climate, and Society, Vol. 1 No. 1, pp. 71-84.
Central Bank of Nigeria (2014), Annual Statistical Report, Abuja.
Christiaensen, L., Lionel, D. and Jesper, K. (2006), “The role of agriculture in poverty reduction – an empirical perspective”, World Bank Policy Research Working Paper No. 4013, World Bank, Washington, DC.
Deressa, T. and Hassan, R. (2009), “Economic impact of climate change on crop production in Ethiopia: evidence from cross-section”, Journal of African Economies, Vol. 18 No. 4, pp. 529-554.
Dinar, A., Mendelsohn, R., Evenson, R., Parikh, J., Sanghi, A., Kumar, K., McKinsey, J. and Lonergan, S. (1998), “Measuring the impact of climate change on Indian agriculture”, World Bank Technical Paper No. 402, Washington, DC.
FAO (2011), Agricultural Insurance in Asia and the Pacific Region, FAO (Food and Agricultural Organization), RAP Publication 2011/12.
Fonta, W.M., Ichoku, E.H. and Urama, E.N. (2011), “Climate change and plantation agriculture: a Ricardian analysis of farmlands in Nigeria”, Journal of Economics and Sustainable Development, Vol. 2 No. 4, pp. 63-75.
Giorgi, F., Jones, C. and Asrar, G. (2009), “Addressing climate information needs at the regional level: the CORDEX framework”, WMO Bulletin.
Hassan, R. (2010), “Implications of climate change for agricultural sector performance in Africa: policy challenges and research agenda”, Journal of African Economies, Vol. 19 No. 2, pp. ii77-ii105.
Ibrahim, B., Karambiri, H., Polcher, J., Yacouba, H. and Ribstein, P. (2014), “Changes in rainfall regime over Burkina Faso simulated by 5 regional climate models”, Climate Dynamics, Vol. 42 Nos 5/6, pp. 1363-1381.
IFAD (2011), “Weather index-based insurance in agricultural development: a technical guide”, available at: www.ifad.org/documents/10180/2a2cf0b9-3ff9-4875-90ab-3f37c2218a90 (accessed 10 January 2016).
IPCC (2007), Impacts, Adaptations, and Vulnerability, Fourth Assessment Report, IPCC (Intergovernmental Panel on Climate Change), Cambridge University Press, Cambridge, MA.
IPCC (2014), “Climate change 2014: synthesis report and summary for policymakers”, available at: www.ipcc.ch/report/ar5/syr/ (accessed 20 February 2016).
Kunreuther, H. (1996), “Mitigating disaster losses through insurance”, Journal of Risk and Uncertainty, Vol. 12 Nos 2/3, pp. 171-187.
Kurukulasuriya, P. and Mendelsohn, R. (2006), “A Ricardian analysis of the impact of climate change on African cropland”, CEEPA Discussion Paper No. 8, CEEPA (Center for Environmental Economics and Policy in Africa), University of Pretoria.
Kurukulasuriya, P. and Mendelsohn, R. (2008), “A Ricardian analysis of the impact of climate change on African crop land”, African Journal of Agriculture and Resource Economics, Vol. 2 No. 1, pp. 1-23.
Lawal, J.O. and Emaku, L.A. (2007), “Evaluation of the effect of climatic changes on cocoa production in Nigeria: cocoa research institute of Nigeria (CRIN) as a case study”, African Crop Science Conference Proceedings, Vol. 8, pp. 423-426.
Mano, R. and Nhemachena, C. (2006), “Assessment of the economic impacts of climate change on agriculture in Zimbabwe: a Ricardian approach”, CEEPA Discussion Paper No. 11, CEEPA (Center for Environmental Economics and Policy in Africa), University of Pretoria.
Mendelsohn, R. and Dinar, A. (1999), “Climate change, agriculture, and developing countries: does adaptation matter?”, World Bank Research Observer, Vol. 14 No. 2, pp. 277-293.
Mendelsohn, R., Arellano-Gonzalez, J. and Christensen, P. (2009), “A Ricardian analysis of mexican farms”, Environment and Development Economics, Vol. 15 No. 2009, pp. 153-172.
Mendelsohn, R., Dinar, A. and Sanghi, A. (2001), “The effect of development on the climate sensitivity of agriculture”, Environment and Development Economics, Vol. 6 No. 1, pp. 85-101.
Mendelsohn, R., Nordhaus, W. and Shaw, D. (1994), “The impact of global warming on agriculture: a Ricardian analysis”, American Economic Review, Vol. 84, pp. 753-771.
Moss, R.H., Edmonds, J.A., Hibbard, K.A., Manning, M.R., Rose, S.K., van Vuuren, D.P., Carter, T.R., Emori, S., Kainuma, M., Kram, T., Meehl, G.A., Mitchell, J.F., Nakicenovic, N., Riahi, K., Smith, S.J., Stouffer, R.J., Thomson, A.M., Weyant, J.P. and Wilbanks, T.J. (2010), “The next generation of scenarios for climate change research and assessment”, Nature, Vol. 463 No. 7282, pp. 747-756.
National Bureau of Statistics (2017), Nigerian Gross Domestic Product Report: Q42016, National Bureau of Statistics, Abuja.
Nwachukwu, I.N., Ezeh, C.I. and Emerole, C.O. (2012), “Effect of climate change on cocoa productivity in Nigeria”, African Crop Science Journal, Vol. 20 No. s2, pp. 487-491.
Odingo, R.S. (2008), “Climate change and economic development: issues, challenges and opportunity for Africa in the decades ahead”, AERC Senior Policy Seminar Paper No. 10, pp. 1-22.
Omolaja, S.S., Aikpokpodion, P., Adedeji, S. and Vwioko, D.E. (2009), “Rainfall and temperature effects on flowering and pollen productions in cocoa”, African Crop Science Journal, Vol. 17 No. 1, pp. 41-48.
Ouedraogo, M., Some, L. and Dembele, Y. (2006), “Economic impact assessment of climate change on agriculture in Burkina Faso: a Ricardian approach”, CEEPA Discussion Paper No 24, CEEPA (Center for Environmental Economics and Policy in Africa), University of Pretoria.
Oyekale, S.A. (2017), “Impact of climate change on cocoa agriculture and technical efficiency of cocoa farmers in South-West Nigeria”, Journal of Human Ecology, Vol. 40 No. 2, pp. 143-148.
Reinsborough, M.J. (2003), “A Ricardian model of climate change in Canada”, Canadian Journal of Economics/Revue Canadienne d`Economique, Vol. 36 No. 1, pp. 21-40.
Ricardo, D. (1817), The Principles of Political Economy and Taxation, John Murray, London.
Seo, N., Mendelsohn, R. and Munasinghe, M. (2005), “Climate change and agriculture in Sri Lanka: a Ricardian valuation”, Environment and Development Economics, Vol. 10 No. 5, pp. 581-596.
Sylla, M.B., Elguindi, N., Giorgi, F. and Wisser, D. (2016), “Projected robust shift of climate zones over west africa in response to anthropogenic climate change for the late 21st century”, Climatic Change, Vol. 134 Nos 1/2, pp. 241-253.
Taylor, K.E., Stouffer, R.J. and Meehl, G.A. (2012), “An overview of CMIP5 and the experiment design”, Bulletin of the American Meteorological Society, Vol. 93 No. 4, pp. 485-498.
World Bank (2008), World Development Report 2008, Washington, DC.
World Bank (2010), “Agricultural insurance”, Primer Series on Insurance, No. 12, Washington DC.
Further reading
Ajetomobi, O.J., Abiodun, A. and Hassan, R. (2011), “Impacts of climate change on rice agriculture in Nigeria”, Tropical and Subtropical Agroecosystems, Vol. 14 No. 2011, pp. 613 -622.
Deressa, T.T. (2006), “Measuring the economic impact of climate change on Ethiopian agriculture: Ricardian approach”, CEEPA Discussion Paper No. 25, CEEPA (Center for Environmental Economics and Policy in Africa), University of Pretoria, SA.
Druyan, L.M. (2011), “Studies of 21st-century precipitation trends over west africa”, International Journal of Climatology, Vol. 31 No. 10, pp. 1415-1424.
Eid, M.H., El-Marsafawy, S.M. and Ouda, S.A. (2006), “Assessing the economic impacts of climate change on agriculture in Egypt: a Ricardian approach”, CEEPA Discussion Paper No. 16, CEEPA (Center for Environmental Economics and Policy in Africa), University of Pretoria, SA.
Gbetibouo, G. and Hassan, R. (2005), “Measuring the economic impact of climate change on major South African field crops: a Ricardian approach”, Global and Planetary Change, Vol. 47 Nos 2/4, pp. 143-152.
Hassan, R. and Nhemachena, C. (2008), “Determinants of climate adaptation strategies of African farmers: multinomial choice analysis”, African Journal of Agricultural and Resource Economics, Vol. 2 No. 1, pp. 83-104.
Jain, S. (2006), “An empirical economic assessment of impacts of climate change on agriculture in Zambia”, CEEPA Discussion Paper No. 27, CEEPA (Center for Environmental Economics and Policy in Africa), University of Pretoria.
Jones, C., Giorgi, F. and Asrar, G. (2011), “The coordinated regional downscaling experiment (CORDEX): an international downscaling link to CMIP5”, CLIVAR Exchanges, Vol. 56 No. 16, pp. 34-40.
Mariara, K.J. and Karanja, F.K. (2006), “The economic impact of climate change on Kenyan crop agriculture: a Ricardian approach”, CEEPA Discussion Paper No. 12, CEEPA (Center for Environmental Economics and Policy in Africa), University of Pretoria.
Molua, E.L. and Lambi, C.M. (2006), “The economic impact of climate change on agriculture in Cameroon”, CEEPA Discussion Paper No. 33, CEEPA (Center for Environmental Economics and Policy in Africa), University of Pretoria.
Onyekuru, A.N. and Marchant, R. (2016), “Assessing the economic impact of climate change on forest resource use in Nigeria: a Ricardian approach”, Agricultural and Forest Meteorology, Vol. 220, pp. 10-20.
Sene, M.I., Diop, M. and Dieng, A. (2006), “Impacts of climate change on the revenues and adaptation of farmers in Senegal”, CEEPA Discussion Paper No. 20, (Center for Environmental Economics and Policy in Africa), University of Pretoria.
Seo, N. and Mendelsohn, R. (2006), “The impact of climate change on livestock management in Africa: a structural Ricardian analysis”, CEEPA Discussion Paper No. 23, CEEPA (Center for Environmental Economics and Policy in Africa), University of Pretoria.
World Bank (2014), World Development Indicator, World Bank, Washington, DC.