I. Individual Crop Response to Price

The purpose of this section is to provide information on the key parameters estimated from empirical supply equations. Many problems are encountered in econometrically assessing the responsiveness of crop output to prices and other variables. First, the data are of poor quality or simply are not available. Second, a positive correlation between the producer price and marketed output indicates very little about the causal relationships underlying the choices of economic agents between leisure and production, food crops and cash crops, and wage work and work on owned farms. Third, empirical results depend on the choice of output and price variables. For example, either harvested tonnage or planted acreage could be used for the output variable. To the extent that acreage is used instead of the quantity produced, there will be a tendency to underestimate the supply elasticity, as it takes no account of variations in the supply of inputs other than land per unit of harvested area. With regard to price variables, one problem facing the researcher is how to calculate the real producer price, and which input and alternative crop prices to include in the estimating equation.

In addition to relative prices, another important variable that affects a farmer’s cropping decision is his perception of risk. In particular, risks related to weather patterns and dependency on markets affect agricultural output. The risk of drought may lead a farmer to plant drought-resistant crops that could be marketed easily. The risk of being dependent on unreliable markets for supplies of inputs and food needs and uncertain receipts owing to price fluctuations may discourage farmers from growing non-traditional or cash crops. Other nonprice constraints include poor transportation networks, inadequate research and extension facilities, unavailability of credit, shortages of fertilizer and other inputs, and the lack of consumer goods on which the farmer can spend his income. Measuring the impact of these factors is usually quite difficult.

Before deciding on the level of production of his crop, the farmer has to choose between consumption and production bundles. The three consumption activities that he can choose from are goods that are produced and consumed within the family, goods that are purchased with cash, and leisure. In his production activities, he has four alternatives: (i) producing food crops and other goods for own consumption; (ii) producing cash crops; (iii) offering his labor for a wage (or participating in other cash-earning activities); and (iv) taking more leisure. Changes in the prices of any of these consumption or production activities will affect the production of food crops or export crops. The way in which crop production is affected by a change in one of these prices depends upon the relative weights of the substitution and income effects in the consumption process, and upon whether or not subsistence goods are inferior in consumption.¹

It follows that the effect of producer prices on the production of cash crops cannot be considered in isolation from other variables; government policies also alter these other variables and thereby change economic incentives. For example, government policy that changes money wages in the industrial sector (or money rates of return from other cash-earning activities) will produce a change in the production of cash crops. Government policies that overprice consumer goods sold for cash also have a strong effect on the production of cash crops. If prices of consumer goods are exorbitantly high in relation to cash received for crop production, producers may have little incentive to respond to changes in producer prices for cash crops.

On a crop-by-crop basis, a great deal of empirical evidence has been accumulated. Positive supply responses have been observed for a wide variety of smallholder cash crops in sub-Saharan Africa, such as cocoa, coffee, cotton, groundnuts, palm kernels, palm oil, rubber, sisal, and tobacco. The survey presented here is on a crop-by-crop, country-by-country basis.

Most of the estimates of price responsiveness are obtained from functional forms based on the model presented here, which includes both acreage functions and yield functions.²

acreage equations

yield equations

where

A_t = actual acreage under cultivation at time t

$A_t^D$ = acreage desired to be under cultivation at time t

P_t = actual real producer price at time t (nominal producer price deflated by consumer price index)

$P_t^e$ = expected real producer price at time t

$Y_t^{*}$ = potential yield at time t

Y_t = actual yield at time t

Z_t = other exogenous factors, including prices of substitutes and complementary crops affecting supply at time t

δ_i = potential yield per acre in year t of crop planted in year i

β, γ = expectations and adjustment coefficients

k= age at which plant (tree) begins to bear

This basic model has been employed in a variety of ways in empirical work. Its modification depends upon the crop under investigation, alternative crops that compete for land and labor, the inclusion of other factors of particular importance for the country or crop being investigated, and the time horizon that must be considered for each crop.

Conceptually, the short run can be defined as the period in which productive capacity cannot be changed but output can be varied by using the fixed capacity with different amounts of variable factors. In the context of particular crops, however, this definition becomes blurred. For annual crops, defined as those whose sowing-to-sowing cycle is one year or less (e.g., cotton), output in the short run can be altered by varying either the acreage under cultivation or the intensity of farming, or both. In this case, $A_t^D$ = A_t, $P_t^e$ = P_t and $Y_t^{*}$ = δA_t (equation (4))—that is, only acreage under cultivation in the current year is important. For perennial crops, however, the stock of trees corresponds to a fixed factor, so that in the short run output can be changed only by varying the intensity of farming, and in the long run capacity can be changed by varying the acreage under planting. For an annual crop, there is likely to be a high correlation between the elasticity of acreage and the elasticity of output with respect to price. For a perennial crop, on the other hand, there is likely to be little relationship between the elasticity of acreage and the elasticity of output, because it may take many years for a tree to mature and to reach peak output.

Short-run price elasticities of acreage and yields can be derived from the separate equations specified earlier, namely, from equation (3)—after substituting equations (1) and (2) in equation (3)—and from equation (5)—after substituting equation (4) in equation (5). To obtain the long-run elasticities, the model must be solved simultaneously. In some empirical work, the long-run elasticities have been obtained by using the acreage and yield equations simultaneously. In others, the acreage and yield equations are used to obtain the coefficients separately, and these coefficients are merely added.

The summary of the own-price elasticities of supply for individual crops for sub-Saharan Africa is presented in Table 2 for cocoa, coffee, cotton, groundnuts, palm kernels, palm oil, rubber, sisal, and tobacco. The focus is on the elasticity of response of individual crops to changes in relative producer prices. Where equations were estimated in level form, the resulting coefficients have been converted to elasticities for purposes of comparison. Information on individual supply responses does not necessarily provide information on the elasticity of aggregate agricultural production. Nevertheless, to the extent that many of the studies presented here do give information on the cross-price substitution effects, own-price responses provide some information on how total agricultural output might respond to an overall increase in agricultural producer prices. The supply elasticities are discussed later on a crop-by-crop basis.

Table 2.

Ten Sub-Saharan African Countries: Cash Crop Supply Elasticities

(By crop and region)

¹ Asterisk indicates that the estimate is significantly different from zero at the 5 percent level of significance.

² In these equations, acreage rather than quantity produced was used as the dependent variable.

Table 2.

Ten Sub-Saharan African Countries: Cash Crop Supply Elasticities

(By crop and region)

Crop and Region		Period	Author	(Year)	Short-Run Elasticity (First year)1	Long-Run Elasticity1
cocoa
	Ghana	1947–64	Behrman	(1968)	—	0.71*
	Ghana (old areas)	1949–62	Bateman	(1965)	0.39	0.77
	Ghana (medium areas)	1949–62	Bateman	(1965)	0.42–0.51	1.28
	Ghana (new areas)	1949–62	Bateman	(1965)	0.61–0.87	1.06
	Nigeria	1947–64	Behrman	(1968)	—	0.71*
	Nigeria	1947–64	Behrman	(1968)	—	0.45*
	Ivory Coast	1947–64	Behrman	(1968)	—	0.80*
	Cameroon	1947–64	Behrman	(1968)	0.68*	1.81*
coffee
	Kenya	1946–64	Maitha	(1970)	0.64*	1.33*
	Kenya (estates)	1946–64	Maitha	(1970)	0.66*	1.38*
	Kenya (smallholders)	1946–64	Maitha	(1970)	0.64*	1.48*
	Kenya	1946–64	Ford	(1971)	—	1.07*
	Kenya (estates)	1946–64	Ford	(1971)	—	1.18*
	Kenya (smallholders)	1946–64	Ford	(1971)	—	1.55*
	Africa	1947–73	de Vries	(1975)	0.12*	0.44*
cotton
	Nigeria2	1948–67	Oni	(1969 b)	0.23–0.38	0.28
	Nigeria	1950–64	Diejomaoh	(1973)	0.67*	0.67*
	Sudan	1951–65	Medani	(1970)	0.39*	0.50*
	Uganda	1922–38	Frederick	(1969)	0.25*	0.25*
	Uganda Buganda	1922–38	Frederick	(1969)	0.67*–0.73*	0.67*–0.73**
	Uganda Buganda2	1945–66	Alibaruho	(1974)	0.50	0.63
	Uganda (Eastern region)2	1945–66	Alibaruho	(1974)	0.23	0.44
	Uganda (Western region)2	1945–66	Alibaruho	(1974)	0.26	0.62
	Uganda (Northern region)2	1945–66	Alibaruho	(1974)	0.02	0.07
groundnuts
	Nigeria	1948–67	Olayide	(1972)	0.24–0.79	0.24–0.79
palm kernels
	Nigeria	1949–64	Oni	(1969 a)	0.22–0.28	0.22–0.28
	Nigeria (Eastern)	1949–64	Oni	(1969 a)	0.28–0.39	0.28–0.39
	Nigeria	1950–64	Diejomaoh	(1973)	0.25*	0.25*
palm oil
	Nigeria	1950–64	Diejomaoh	(1973)	0.81*	0.81*
	Nigeria	1949–63	Helleiner	(1966)	0.41*	0.41*
	Nigeria	1949–64	Oni	(1969 a)	0.29–0.35	0.29–0.35
	Nigeria (Eastern)	1949–64	Oni	(1969 a)	0.41*–0.70*	0.41*–0.70*
rubber
	Liberia	1950–72	Ghoshal	(1974)	0.14	0.22
	Nigeria	1952–72	Olayemi & Olayide	(1975)	0.04	1.75*
sisal
	Tanzania	1945–67	Gwyer	(1971)	0.06*	0.48*–0.76*
tobacco
	Malawi	1926–60	Dean	(1966)	0.48*	0.48*
	Nigeria2	1945–64	Adesimi	(1970)	0.60*	0.82*

¹ Asterisk indicates that the estimate is significantly different from zero at the 5 percent level of significance.

² In these equations, acreage rather than quantity produced was used as the dependent variable.

View Table

Table 2.

Ten Sub-Saharan African Countries: Cash Crop Supply Elasticities

(By crop and region)

Crop and Region		Period	Author	(Year)	Short-Run Elasticity (First year)1	Long-Run Elasticity1
cocoa
	Ghana	1947–64	Behrman	(1968)	—	0.71*
	Ghana (old areas)	1949–62	Bateman	(1965)	0.39	0.77
	Ghana (medium areas)	1949–62	Bateman	(1965)	0.42–0.51	1.28
	Ghana (new areas)	1949–62	Bateman	(1965)	0.61–0.87	1.06
	Nigeria	1947–64	Behrman	(1968)	—	0.71*
	Nigeria	1947–64	Behrman	(1968)	—	0.45*
	Ivory Coast	1947–64	Behrman	(1968)	—	0.80*
	Cameroon	1947–64	Behrman	(1968)	0.68*	1.81*
coffee
	Kenya	1946–64	Maitha	(1970)	0.64*	1.33*
	Kenya (estates)	1946–64	Maitha	(1970)	0.66*	1.38*
	Kenya (smallholders)	1946–64	Maitha	(1970)	0.64*	1.48*
	Kenya	1946–64	Ford	(1971)	—	1.07*
	Kenya (estates)	1946–64	Ford	(1971)	—	1.18*
	Kenya (smallholders)	1946–64	Ford	(1971)	—	1.55*
	Africa	1947–73	de Vries	(1975)	0.12*	0.44*
cotton
	Nigeria2	1948–67	Oni	(1969 b)	0.23–0.38	0.28
	Nigeria	1950–64	Diejomaoh	(1973)	0.67*	0.67*
	Sudan	1951–65	Medani	(1970)	0.39*	0.50*
	Uganda	1922–38	Frederick	(1969)	0.25*	0.25*
	Uganda Buganda	1922–38	Frederick	(1969)	0.67*–0.73*	0.67*–0.73**
	Uganda Buganda2	1945–66	Alibaruho	(1974)	0.50	0.63
	Uganda (Eastern region)2	1945–66	Alibaruho	(1974)	0.23	0.44
	Uganda (Western region)2	1945–66	Alibaruho	(1974)	0.26	0.62
	Uganda (Northern region)2	1945–66	Alibaruho	(1974)	0.02	0.07
groundnuts
	Nigeria	1948–67	Olayide	(1972)	0.24–0.79	0.24–0.79
palm kernels
	Nigeria	1949–64	Oni	(1969 a)	0.22–0.28	0.22–0.28
	Nigeria (Eastern)	1949–64	Oni	(1969 a)	0.28–0.39	0.28–0.39
	Nigeria	1950–64	Diejomaoh	(1973)	0.25*	0.25*
palm oil
	Nigeria	1950–64	Diejomaoh	(1973)	0.81*	0.81*
	Nigeria	1949–63	Helleiner	(1966)	0.41*	0.41*
	Nigeria	1949–64	Oni	(1969 a)	0.29–0.35	0.29–0.35
	Nigeria (Eastern)	1949–64	Oni	(1969 a)	0.41*–0.70*	0.41*–0.70*
rubber
	Liberia	1950–72	Ghoshal	(1974)	0.14	0.22
	Nigeria	1952–72	Olayemi & Olayide	(1975)	0.04	1.75*
sisal
	Tanzania	1945–67	Gwyer	(1971)	0.06*	0.48*–0.76*
tobacco
	Malawi	1926–60	Dean	(1966)	0.48*	0.48*
	Nigeria2	1945–64	Adesimi	(1970)	0.60*	0.82*

¹ Asterisk indicates that the estimate is significantly different from zero at the 5 percent level of significance.

² In these equations, acreage rather than quantity produced was used as the dependent variable.

A great variety of studies have been conducted on the response to price incentives of West African cocoa growers. Once cocoa reaches maturity, it does not compete with food crops (maize and cassava) for labor inputs, because these food crops can be planted and harvested throughout the year in alternate rows with cocoa plants. Before planting, however, cocoa competes with other perennial crops, such as coffee. The estimates of the short-run (first year) supply elasticities vary greatly. In the only study in which the estimate of the short-run elasticity is statistically significant, the magnitude of the estimated elasticity is about 0.7. With respect to the long-run supply response, practically all studies conclude that, allowing for the time for a tree to reach maturity, there is a large output response to real producer prices. The average of the estimates of the long-run elasticity is about 0.9; the long-run output responses for the Ivory Coast, Cameroon, and Ghana are larger than those for Nigeria.

In both Bateman’s (1965) and Behrman’s (1968) supply equations, the quantity produced is used as the dependent variable because of the lack of available data on acreage. Equations similar to equations (1) and (3) are solved for A_t (acreage),³ and then substituted in the quantity-produced equation similar to equations (4) and (5). The long-run elasticity is the response after newly planted trees have come into full bearing and all other adjustments have occurred.

The estimates of the cross-price elasticities between cocoa and coffee (at the planting stage) are large, negative, and significantly different from zero at the 5 percent significance level in both studies; the size of the parameter estimates is almost twice as large as the estimates of own-price effects. Bateman concludes that cocoa production responds to both cocoa and coffee prices, and to the length of time that a region has been planted to cocoa. The Bateman and Behrman studies, however, both omit some of the important variables that farmers may consider when making their production choices, such as the money wage outside the agricultural sector and the availability of cash goods.

Supply models for coffee have been estimated by de Vries (1975), Maitha (1970), and Ford (1971) for both Kenya and all of Africa. These studies show that even after allowing for the substitution (cross-price) effects of changes in the relative price between coffee and cotton (which is regarded as the major alternative crop for coffee), coffee production responds strongly to a rise in its producer price. For the period 1943–73, estimates of the short-run elasticity of production average about 0.5, while the long-run elasticities average about 1.1. The studies on Kenya by Maitha and Ford contain equations for both acreage and yield per acre. The short-run supply elasticity is obtained from the equation for the yield per acre, since in the short run coffee acreage is fixed. The long-run supply elasticities are derived by summing the long-run price elasticities for acreage and for yield per acre. The long-run elasticities estimated for industry and estate coffee by Maitha are considerably larger than those estimated by Ford. One important reason for this result is that Ford estimated the coefficients from the acreage and yield equations separately and then simultaneously determined the long-run elasticities. This simultaneous determination was not attempted by Maitha. Both these studies demonstrate that in the long run most of the supply responsiveness for coffee comes from the change in yield per acre rather than from the change in acreage, demonstrating that even if the producer price has little effect on the acreage under coffee production, it has a strong effect on yield.

Studies of cotton production and acreage in Nigeria, Sudan, and Uganda show significant short-run and long-run supply responses, as indicated in Table 2. Cotton (an annual crop) competes with other crops, in particular coffee and groundnuts, so that, in the short run, yield as well as acreage can vary. Frederick (1969), Medani (1970), and Diejomaoh (1973) estimated output equations for cotton and found that after taking account of the cross-price substitution effects with coffee and groundnuts, there is a marked own-price effect. The average of the estimates of the short-run own-price yield elasticity is about 0.5. As might be expected with an annual crop, the long-run elasticity is not much larger than the short-run elasticity.⁴ The average of the estimates of the long-run price elasticity is also about 0.5. These results are generally supported by the estimates of acreage response obtained by Oni (1969 b) and Alibaruho (1974). However, Alibaruho finds the own-price elasticity to be much larger in the long-run than in the short run. The evidence from the cross-price elasticities also suggests that an increase in the producer price of cotton relative to that of coffee would induce a rise in cotton output that would be significantly larger than the fall in coffee output.

Sisal is a perennial crop that has been grown in Tanzania since the beginning of the century. Gwyer (1971) finds that sisal planters in Tanzania are influenced by prices in the year of planting and in the two previous years. Gwyer estimates separate equations for acreage and yield. The short-run supply elasticity, derived from the yield (output per hectare of mature acreage) equation, is small (0.06) but significant. The long-run elasticity, the sum of the acreage and yield elasticities, is large (0.62).⁵ Gwyer also finds that in the short run producers react to any current price stimuli by varying the frequency of the harvesting, rather than by heavy harvesting and overcutting—a reaction that would lead to a decline in future yields. Since there is no shortage of land suitable for sisal cultivation, most of the long-run response comes from producers who are changing their planting decisions.

Most studies of the supply responses of palm products (oil and kernels) show significantly positive own-price effects. Although palm oil and palm kernels are tree crops, the trees tend to grow in the wild, and producers’ responses are essentially related to harvesting decisions rather than to planting decisions. Consequently, in econometric work, yields of palm products can be modeled as for an annual crop; for this reason, the long-run elasticity does not differ from the short-run yield elasticity. Most of the studies for palm products have been carried out on data between the late 1940s and the mid-1960s for Nigeria. The average of the elasticity estimates is about 0.2 for palm kernels and 0.5 for palm oil. Olayide (1972) has noted that Nigerian palm producers respond very strongly to world prices, and has suggested that Nigeria’s export marketing policies be revised so that producer prices closely reflect world prices.

Rubber is a perennial crop, which begins to produce only after five years. Most of the statistical evidence from Africa—mainly from Liberia and Nigeria—suggests that the short-run response of rubber production to price is inelastic. Estimates of short-run yield elasticities for rubber range from 0.04 to 0.14. Olayemi and Olayide (1975) use a polynomial lag model to show that while output response to a price change may be inelastic for the first few years, the price response rises to a peak at about the fifth year when new plantings come into production. In subsequent years, the price effect declines. A total price elasticity of 1.75 was calculated over a period of eight years, indicating that, when all production adjustments are made, the long-run price response is large. The effects of the price change are twofold: the price increase induces the farmer not only to undertake new plantings but also to take more intensive care of mature trees. At first, only the effect of more intensive care shows up, but after five to six years new plantings come into production and add their own effect to output.

Studies on tobacco, an annual crop, find a marked response to the real producer price. A study by Adesimi (1970) focuses on the acreage response in Nigeria’s western provinces, where tobacco competes for land and labor inputs with yams, cassava, and maize. The study shows that the tobacco acreage is positively related to the price of tobacco and negatively related to the price of yams, and that if the prices of both products increase, tobacco acreage will still increase. A study by Dean (1966) on tobacco production in Malawi focuses directly on the output equation and obtains a somewhat smaller supply response.

Another study of African tobacco production by Dean (1965) for Malawi bears indirectly on the issues considered here. Dean’s main interest was to determine whether the tobacco grower’s labor supply curve (and hence that for tobacco itself) is backward bending. The rise in tobacco production that is associated with a rise in the tobacco price is interpreted by Dean as indicating that increased amounts of labor are allocated to tobacco production when there is an increase in the return to labor in this activity, and that this result is consistent with the hypothesis that the supply curve of labor for tobacco production is upward sloping. The negative coefficient on the wage variable indicates that a rise in wages obtainable abroad decreases the amount of labor allocated to tobacco and constitutes evidence against the presence of a backward-sloping supply curve.

The principal findings from the studies from sub-Saharan African countries can be summarized as follows. First, with few exceptions, the own-price elasticities are positive and significant from an economic and statistical standpoint even in the short run. The size of the short-run elasticities are not large, however. Second, in general, the long-run own-price elasticities tend to be larger than those for the short run and of fairly sizable magnitudes. Since many of the equations also contain cross-price effects, these own-price responses provide some support for the view that overall agricultural output does increase with an overall increase in agricultural producer prices.⁶ Third, the findings tend to suggest that spare capacity suitable for crop cultivation often exists and that increasing producer prices can lead to increased productivity per acre. Finally, the findings also contradict the view that peasant farmers have only income goals and that they therefore react perversely to changes in producer prices.

II. Overall Output Responses

Section I dealt with the findings on the response of individual crops to price changes, but a more important question is that of how aggregate agricultural production responds to price changes. Aggregate agricultural production can change in response to (i) a change in the relative prices of export versus food crops and (ii) a change in the aggregate agricultural price index. Changing the price of export crops relative to food crops will have important implications for the balance of payments, as well as for overall agricultural production. For most African countries, food as well as export crops are traded goods, so that a change in the terms of trade in favor of export crops could either improve or deteriorate the current account, depending on the size of the import and export sectors and the output response to price changes for each crop. How aggregate production responds to changes in the aggregate agricultural price index depends on a number of factors, such as the movement of labor and materials from the urban to the rural sector, the question of the trade-off between leisure and production, the amount of unused resources in the economy, and how much new investment is undertaken in agriculture. In many of the countries covered by this paper, there will be considerable unused resources or inefficient use of resources, and consequently movement of resources from the urban to the rural sector need not take place in order for aggregate agricultural production to rise.

Reliable and conclusive evidence of aggregate supply responses in Africa is difficult to find. Very few tests have been made of how aggregate supply—production or marketed surplus of either individual farmers or the agricultural sector as a whole—responds to the general agricultural price level or to the urban/rural terms of trade. It would seem reasonable to believe that the elasticity of aggregate production is lower than that of individual crops; the latter involves the redeployment of resources within the agricultural sector, whereas the former includes transferring resources to the agricultural sector from other sectors, or utilizing unemployed or underemployed resources existing within the sector. Although it is widely accepted that leisure has gradually taken a smaller proportion of the time of the agricultural sector as the cash economy has made inroads into African rural society (Helleiner (1975)), this assumption has not been verified empirically.

One method sometimes used to assess overall supply responses is to study the impact of devaluation on a country’s overall exports, since in most of these countries exports come predominantly from the agricultural sector. Export changes following devaluation may be quite large because exports may be rechanneled from illegal selling activities,⁷ or from existing large stocks, into official channels. However, there are serious problems associated with equating the response of exports to a devaluation with the response of overall agricultural supply to aggregate producer prices. For the following reasons, studies on the response of exports to a devaluation are not considered in this paper.

1. Increases in production of agricultural exports may be offset by a reduction in the production of food crops, so that this does not necessarily imply an overall increase in output following devaluation.

2. Devaluation may not necessarily have an effect on agricultural producer prices if the authorities regulate prices and output changes following devaluation.

3. The price of primary products varies considerably from year to year and cannot be held constant while the effect of a devaluation is assessed.

4. The medium-term and longer-run effects of a devaluation depend largely on the supporting policies taken by the authorities.

A recent study on sub-Saharan African countries by Acharya (1981) considers both export and food crop production together. Acharya concludes that failure to maintain incentives in these countries’ agricultural sectors was the major reason for their poor performance in agricultural production: Africa’s per capita food production declined by 7 percent between 1967 and 1978 in contrast to increases of 7 to 8 percent in Asia and Latin America. Nevertheless, Acharya also shows that there were sharp differences in outcome between market-oriented economies (namely, Ivory Coast, Malawi, and, to a lesser extent, Kenya), which sustained the most favorable environments for private peasant agriculture, and the etatist economies (Ghana, Sudan, and Tanzania), which introduced a high level of state ownership and tried to force the pace of industrialization by skewing the structure of incentives against agriculture. The evidence in Table 3 shows that Ivory Coast, Kenya, and Malawi achieved massive increases in export volume despite the fall in their terms of trade, and were able to more than quadruple the purchasing power of their exports over the period 1954–56 to 1975–77. In contrast, the purchasing power of exports remained nearly constant for Ghana, while Sudan and Tanzania recorded modest gains. The export volume of Ivory Coast, Kenya, and Malawi increased sevenfold, sixfold, and fourfold, respectively, between 1954–56 and 1975–77, whereas the increase for Ghana and Tanzania was 75 percent and for Sudan was 70 percent over the same period. Thus, Acharya concludes that the market-oriented countries clearly generated increases in volume that outweighed the adverse price trends, whereas the etatist countries did not and therefore had to rely more and more on foreign aid.

Table 3.

Six Sub-Saharan African Countries: Export Volume, Terms of Trade, and Purchasing Power, 1954–56, 1964–66, and 1975–77

(1970 = 100)

Source: United Nations Conference on Trade and Development, Handbook of International Trade and Development Statistics, Sales No. E/F.79.II.D.2 (New York, 1979).

Table 3.

Six Sub-Saharan African Countries: Export Volume, Terms of Trade, and Purchasing Power, 1954–56, 1964–66, and 1975–77

(1970 = 100)

		1954–56	1964–66	1975–77
Ivory Coast
	Export volume	22	83	152
	Terms of trade	139	82	102
	Purchasing power	31	68	155
Kenya
	Export volume	17	81	103
	Terms of trade	154	103	106
	Purchasing power	26	83	112
Malawi
	Export volume	29	77	118
	Terms of trade	111	97	113
	Purchasing power	32	74	133
Ghana
	Export volume	52	96	92
	Terms of trade	148	66	92
	Purchasing power	77	64	84
Sudan
	Export volume	42	80	76
	Terms of trade	119	90	115
	Purchasing power	50	71	86
Tanzania
	Export volume	39	95	69
	Terms of trade	123	100	113
	Purchasing power	47	95	78

Source: United Nations Conference on Trade and Development, Handbook of International Trade and Development Statistics, Sales No. E/F.79.II.D.2 (New York, 1979).

View Table

Table 3.

Six Sub-Saharan African Countries: Export Volume, Terms of Trade, and Purchasing Power, 1954–56, 1964–66, and 1975–77

(1970 = 100)

		1954–56	1964–66	1975–77
Ivory Coast
	Export volume	22	83	152
	Terms of trade	139	82	102
	Purchasing power	31	68	155
Kenya
	Export volume	17	81	103
	Terms of trade	154	103	106
	Purchasing power	26	83	112
Malawi
	Export volume	29	77	118
	Terms of trade	111	97	113
	Purchasing power	32	74	133
Ghana
	Export volume	52	96	92
	Terms of trade	148	66	92
	Purchasing power	77	64	84
Sudan
	Export volume	42	80	76
	Terms of trade	119	90	115
	Purchasing power	50	71	86
Tanzania
	Export volume	39	95	69
	Terms of trade	123	100	113
	Purchasing power	47	95	78

Source: United Nations Conference on Trade and Development, Handbook of International Trade and Development Statistics, Sales No. E/F.79.II.D.2 (New York, 1979).

To provide stronger evidence of the response of aggregate output to changes in aggregate real producer prices in sub-Saharan African countries, a simple empirical equation has been estimated for nine predominantly agricultural African countries—Ghana, Ivory Coast, Kenya, Liberia, Madagascar, Senegal, Tanzania, Uganda, and Upper Volta. Any empirical testing of the responsiveness of aggregate output to relative price changes involves numerous difficulties. First, the aggregate function is obtained by adding across the supply functions for individual crops. Individual crops respond quite differently to price changes, as can be seen from Table 2, and the estimator from grouped data is always less efficient than the estimator from ungrouped data. Adding across supply functions also has important implications for the lag structure. Second, increasing the real producer price index will lead to a reduction in the proportion of agricultural output growth in the subsistence sector—an effect that is not captured in this study. Third, for some countries, the time series are short, and for most countries the quality of the data is poor.

The equation derived here is based on the methodology of Nerlove (1958). The actual changes in per capita total agricultural output (Q_t) from the previously existing level (Q_t) is only some fraction of the change required to achieve the equilibrium level (Q_t) Assuming that the proportion achieved is β, the equation can be written

Equilibrium output (Q_t) depends on the aggregate real producer price (PR_t), measured as the average producer prices of the major agricultural products deflated by the consumer price index, a time trend (t) to take into account the effect of long-run structural changes on equilibrium output, and a dummy variable (Z_t) to capture the influences of unusual weather patterns:

By substituting equation (7) in equation (6), equation (8) is obtained:

Large values of β imply relatively fast adjustment so that the lower the estimate of b₂, the faster equilibrium output adjusts.

In constructing equation (8), it is necessary to note data limitations. Data on per capita agricultural output obtained from the Food and Agriculture Organization (FAO) are based on a crude methodology and are hence subject to large errors in measurement. On the other hand, good data for producer prices are available for only some of the countries and some of the crops.

Table 4 contains the regression results of equation (8) for the nine African countries mentioned earlier.⁸ The procedure used for estimation is that of ordinary least-squares multiple regression and the number in parentheses is the t-statistic for each of the respective coefficients. For each country, the time period for which the investigation was carried out is given along with the results.

Table 4.

Nine Sub-Saharan African Countries: Elasticity Coefficients for Aggregate Agricultural Supply Responses, 1963–81

¹ Corrected for first-order autocorrelation ρ = 0.36 (1.66).

² Corrected for first-order autocorrelation ρ = −0.72 (2.59).

Table 4.

Nine Sub-Saharan African Countries: Elasticity Coefficients for Aggregate Agricultural Supply Responses, 1963–81

Country	b0	b1	b2	b3	b4	D-W	R2	SEE	Years
Ghana	1.74 (2.27)	0.20 (3.16)	0.42 (2.55)	− 0.01 (1.70)		2.11	0.824	0.06	1963–81
Kenya	2.52 (2.93)	0.10 (3.06)	0.36 (1.91)	−0.003 (1.86)	0.04 (2.61)	2.24	0.676	0.02	1966–80
Ivory Coast	4.06 (5.42)	0.13 (0.77)		−0.01 (1.43)	0.04 (2.56)	2.56	0.636	0.03	1969–78
Liberia	3.97 (2.28)	0.10 (0.88)	0.08 (0.25)	−0.01 (1.48)		2.07	0.182	0.04	1966–80
Madagascar1	2.70 (2.91)	0.10 (0.63)	0.31 (1.48)	0.0001 (0.03)	0.07 (4.27)	1.95	0.774	0.02	1968–81
Senegal2	3.68 (1.45)	0.17 (0.36)		0.0002 (0.009)	0.16 (3.83)	2.30	0.781	0.12	1970–79
Tanzania	4.66 (21.27)	0.03 (0.69)		−0.02 (8.98)		2.28	0.961	0.01	1972–81
Uganda	3.06 (2.12)	0.05 (0.54)	0.30 (0.88)	−0.02 (1.97)		2.13	0.864	0.04	1968–78
Upper Volta	3.25 (1.89)	0.22 (1.16)	0.08 (0.32)	−0.01 (1.92)		1.83	0.470	0.047	1964–80

¹ Corrected for first-order autocorrelation ρ = 0.36 (1.66).

² Corrected for first-order autocorrelation ρ = −0.72 (2.59).

View Table

Table 4.

Nine Sub-Saharan African Countries: Elasticity Coefficients for Aggregate Agricultural Supply Responses, 1963–81

Country	b0	b1	b2	b3	b4	D-W	R2	SEE	Years
Ghana	1.74 (2.27)	0.20 (3.16)	0.42 (2.55)	− 0.01 (1.70)		2.11	0.824	0.06	1963–81
Kenya	2.52 (2.93)	0.10 (3.06)	0.36 (1.91)	−0.003 (1.86)	0.04 (2.61)	2.24	0.676	0.02	1966–80
Ivory Coast	4.06 (5.42)	0.13 (0.77)		−0.01 (1.43)	0.04 (2.56)	2.56	0.636	0.03	1969–78
Liberia	3.97 (2.28)	0.10 (0.88)	0.08 (0.25)	−0.01 (1.48)		2.07	0.182	0.04	1966–80
Madagascar1	2.70 (2.91)	0.10 (0.63)	0.31 (1.48)	0.0001 (0.03)	0.07 (4.27)	1.95	0.774	0.02	1968–81
Senegal2	3.68 (1.45)	0.17 (0.36)		0.0002 (0.009)	0.16 (3.83)	2.30	0.781	0.12	1970–79
Tanzania	4.66 (21.27)	0.03 (0.69)		−0.02 (8.98)		2.28	0.961	0.01	1972–81
Uganda	3.06 (2.12)	0.05 (0.54)	0.30 (0.88)	−0.02 (1.97)		2.13	0.864	0.04	1968–78
Upper Volta	3.25 (1.89)	0.22 (1.16)	0.08 (0.32)	−0.01 (1.92)		1.83	0.470	0.047	1964–80

¹ Corrected for first-order autocorrelation ρ = 0.36 (1.66).

² Corrected for first-order autocorrelation ρ = −0.72 (2.59).

The overall fit of the equation is reasonably good except for Liberia. Omitted variables and errors in data measurement are likely to be the important causes of the poor fit. With the exception of Senegal and Upper Volta, the estimate of the constant term is significant at the 95 percent level for all countries. The constant could be interpreted as subsistence output, or the level of output that would be produced regardless of the producer price. This coefficient is higher for Tanzania than for any of the other countries.

For all countries, the estimated relative price coefficient is positive. The estimates show that the relative price is an important determinant of overall agricultural output, particularly for Ghana and Kenya. The size of the coefficient for Ghana shows that a rise of 10 percent in the real overall aggregate producer price will result in an increase of 2 percent in overall agricultural output after one year. The average of the estimated price elasticities for the nine countries is 0.12, which is lower than the estimates obtained for individual crops (Table 2).

Agricultural output exhibits a negative time trend for most of the countries. The estimated coefficients for the time trend are negative for Ghana, Kenya, Ivory Coast, Liberia, Tanzania, Uganda, and Upper Volta; it is particularly strong in the equation for Tanzania, reflecting the unfavorable climate for agricultural investment that has been created, the deterioration of the infrastructure and support services available to the rural farmer, and the growing lack of availability of cash goods to the rural sector.

The dummy variable for weather conditions is the most significant factor in the equations for Ivory Coast, Madagascar, and Senegal. It would appear that for these countries the weather was a more important determinant of total agricultural output than any other variable.

The partial adjustment coefficient (b₂), and the short-run (b₁) and long-run (a₁) price elasticities are given in Table 5. Ghana, Kenya, Liberia, Madagascar, Uganda, and Upper Volta all have long-run price elasticities that are greater than the short-run price elasticities, but for the majority of these countries the confidence interval is well below the 95 percent significance level. Despite the considerable differences for individual countries, the average long-run price elasticity for nine countries that were examined is only slightly larger than their average short-run price elasticities.

Table 5.

Nine Sub-Saharan African Countries: Short-Run and Long-Run Price Elasticities and Partial Adjustment Coefficients for Aggregate Agricultural Supply

Table 5.

Nine Sub-Saharan African Countries: Short-Run and Long-Run Price Elasticities and Partial Adjustment Coefficients for Aggregate Agricultural Supply

Country	Short-Run Price Elasticity (b1)	Partial Adjustment Coefficient (b2)	Long-Run Price Elasticity (a1)
Ghana	0.20	0.42	0.34
Kenya	0.10	0.36	0.16
Ivory Coast	0.13	—	0.13
Liberia	0.10	0.08	0.11
Madagascar	0.10	0.31	0.14
Senegal	0.54	—	0.54
Tanzania	0.15	—	0.15
Uganda	0.05	0.30	0.07
Upper Volta	0.22	0.08	0.24
Average	0.18	0.17	0.21

View Table

Table 5.

Nine Sub-Saharan African Countries: Short-Run and Long-Run Price Elasticities and Partial Adjustment Coefficients for Aggregate Agricultural Supply

Country	Short-Run Price Elasticity (b1)	Partial Adjustment Coefficient (b2)	Long-Run Price Elasticity (a1)
Ghana	0.20	0.42	0.34
Kenya	0.10	0.36	0.16
Ivory Coast	0.13	—	0.13
Liberia	0.10	0.08	0.11
Madagascar	0.10	0.31	0.14
Senegal	0.54	—	0.54
Tanzania	0.15	—	0.15
Uganda	0.05	0.30	0.07
Upper Volta	0.22	0.08	0.24
Average	0.18	0.17	0.21

The findings that come from these empirical results and from the Acharya (1981) study do provide considerable evidence for positive aggregate supply responses to changes in aggregate real producer prices. On the basis of intercountry comparisons, Acharya demonstrates that the structure of incentives influences agricultural performance. The estimates reported here support this view. However, structural changes have, for nearly all the countries, hindered agricultural production, and for a few countries, weather patterns have dominated all other variables. Long-run price elasticities have proved to be difficult to measure, but for some countries there is a clear indication that these differ in size from short-run price elasticities.

III. Conclusion

In many sub-Saharan African economies, governments need considerable information in order to review their agricultural pricing and support policies. This paper presents evidence from studies of both individual crops and overall crop production for a number of African countries. It is found that for both individual crops and aggregate production, supply responses are positive. In general, for individual crops, the longer-run price elasticities tend to be larger than those for the short run, and are fairly sizable. For aggregate crop production, there appears to be some evidence that farmers do respond to aggregate real producer prices. Turning the terms of trade in favor of the agricultural sector does lead to increased agricultural production. Where there are unused resources in the economy, increasing agricultural production does not necessarily imply a movement of resources from other sectors. However, in economies where resource use is close to its potential, there may be a movement of resources from other sectors into the agricultural sector. As far as substitution effects are concerned, the evidence from individual crop studies suggests that for some crops an increase in production can take place without an equivalent fall in production of substitute crops.

The target-income hypothesis, which implies a backward-sloping supply curve for labor, is not supported by the evidence presented in this paper. The studies also show that neither labor nor land presents major constraints on agricultural expansion. Individual crop studies tend to suggest that spare capacity suitable for crop cultivation often does exist, and that when producer prices increase, productivity per acre is also likely to rise.

A realistic pricing policy is only one element in a comprehensive package of government action that would be required to ensure a sound agricultural base. Government action would also need to include a policy decision to switch the fiscal burden from producer pricing to other sources, or to depreciate the exchange rate to provide the means to raise producer prices. Prices of consumer goods and inputs to the agricultural sector as well as improvements in extension services, infrastructure, input and availability of consumer goods, and credit services also need to become a part of government action. A package of changes may elicit a quite different response from farmers than a price change alone. For example, an increase in the producer price together with availability of consumer goods and an improvement in extension services will offer more of an output incentive to the farmer than a mere increase in the producer price. Therefore, governments need to be informed of the likely effects of changes in these other variables, along with those of changes in pricing policies, in order to assess more accurately the possible outcome of a variety of policy packages.