Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

The effects of food price and income policies on the nutrition of low-income groups: a Philippine case-study

M. A. R. Quisumbing
Assistant Professor, College of Development Economics and Management, University of the Philippines at Los Baņos, Philippines


In recent years food policy analysts have recognized that the normal course of development will not close the nutritional deficits of the lowest income groups in developing countries. While growth- and supply-oriented policies may increase the aggregate supply of nutrients, there is no guarantee that these nutrients will remedy the nutritional deficits of at-risk groups. The policy option of implementing general food price subsidies aimed at protecting consumers from high food prices entails high costs either for fiscal outlays or for producer disincentives. Thus, support has been growing for target-group-oriented policies which have the purpose of increasing the nutrient consumption of at-risk groups. To have a positive impact on the in take of a given nutrient by low-income groups, a commodity must contain the nutrient in adequate amounts and must be targeted specifically to those groups. Moreover, the kinds of commodities in which nutrients are made available, as well as the variations in demand behaviour among consumer groups, must be considered in formulating nutrition policies [31, 39].

The nutritional impact of food policies depends, to a great extent, on how both supply and demand influence household food consumption behaviour and which households- whether poorly nourished or not-are reached by the policies. This study uses Philippine income-stratum-specific demand parameters to estimate the potential impact of food market intervention policies, such as price subsidies, income transfers, and target-group-oriented policies; it also estimates the treasury costs of target-group-oriented policies with the aim of ranking them in terms of cost-effectiveness.

FIG. 1. Distribution of agricultural gross value added in constant 1972 prices by commodity groups in the Philippines, 1955-1980


The importance of the agricultural sector in the Philippine economy and its critical role in the national food supply cannot be overestimated. According to a comprehensive study of the Philippine agricultural sector by C. C. David [11], despite the strong bias toward industrialization in the post-war development strategy' agriculture still dominates the economy, employing about 50 percent of the total labour force and contributing nearly 30 percent of the net domestic product (fig. 1). Between 1955 and

1980, agriculture in the Philippines grew at a rate similar to that of other Association of South-East Asian Nations (ASEAN)countries but about 4 percent higher than that of other middle-income countries. On the other hand, the Philippine manufacturing sector, posting a growth rate of 6 to 7 percent, grew more slowly than that of other ASEAN countries.

The growth rate of the Philippine agricultural sector increased from the mid 1960s and into the 1970s, surpassing the growth rate of manufacturing in the mid 1970s, when a marked slowdown in the latter occurred. The agricultural sector also seems to have performed well relative to the manufacturing sector in the late 1970s, despite the second oil price shock and the world-wide recession, which caused a sharp drop in world prices of the major agricultural export commodities (table 1) [11] .

TABLE 1. Sectoral growth rates of value added in the Philippines. 1955 to 1980 (percentages)

  1956-61 a 1961 -66 1966-71 1971 -76 1976-79 1956-79
Industry 4.7 5.8 5.2 6.9 6.8 5.2
Manufacturing 6.3 5.2 5.6 7.1 5.8 6.0
Services 4.3 4.4 4.6 5.2 4.8 4.8
Agriculture 3.6 4.3 3.5 4.2 4.7 4.0
Forestry 9.8 5.9 40 -46b Lob 2~8b
Fishery 2.9 4.9 7.7 4.6 3.3 4.8
Livestock and poultry -2.6 6.8 1.9 1.7 4.2 2.1
Crops 4.6 3.3 5.0 7.7 6.5 5.3

a. End years are three-year averages centred at the year shown.
b. This low growth rate was due in part to under-reporting of log exports.
Source: References II and 28.

TABLE 2. Average growth rates (AGR) of agricultural production and proportion of food crops to total production. the Philippines, 1971 to 1980

  All crops Food crops  
Year Quantity (1)a AGR (%) Quantity (2)a AGR (%) Proportion: (2) - (1)x100 (%)
1971 15,6121.4   10,773.8    
1972 15,191.1 - 2.75 10,629.4 - 1.34 68.97
1973 15,092.1 - 0.65 9,890.4 - 6.95 69.97
1974 17,546.3 16.26 12,072.6 22.07 65.53
1975 19,807.4 12.87 13,549.0 12.23 68.40
1976 23,539.2 18.84 15,439.9 13.96 65.59
1977 24,506.5 4.11 16,856.9 9.18 68.79
1978 26,095.9 6.49 18,371.0 8.98 70.40
1979 28,609.2 9.63 20,835.8 13.42 72.83
1980 29,566.2 3.35 21,837.0 4.81 73.85
1981 29,507.9 - 0.19 21,748.6 - 0.40 73.70
1982 29,884.0 1.27 22,436.2 3.16 75.08
1983 27,261.4 - 8.76 20,116.9 - 10.34 73.79

a. The quantities are in thousands of metric tonnes.
Source: Reference 28.

TABLE 3. Annual growth rates of major agricultural crops in the Philippines, 1955 to 1980

  1956-61a 1961-66 1966-71 1971-76 1976-79 1956-79
Rice 2.9 1.4 5.2 3.1 5.7 3.5
Corn 7.2 2.8 2.7 2.6 5.1 5.8
Sugar 8.7 1.7 6.7 5.1 - 2.5 4.4
Coconuts -0.2 5.3 1.9 13.9 8.9 5.6

a. End years are three-year averages centred at the year shown.

An economic crisis occurred in 1983, and restrictions on imports of the intermediate goods needed for manufacturing caused the industrial sector to stagnate. The government recently expressed a renewed, if belated, interest in the agricultural sector as the basis for balanced agro-industrial development.

Although the proportion of food crop production to total crop production has increased from about 69 percent in 1972 to 74 percent in 1983 (table 2), the food sector's performance in the 1980s has lagged behind that of the 1970s, posting negative growth rates from 1980 to 1981 and 1982 to 1983. Total crop production has also fallen by about 9 percent from 1982 to 1983. While growth in rice production has accelerated since 1966 with the introduction of the modern seed-fertilizer technology and the expansion of irrigated land area (table 3), and the Philippines was able to generate a small production surplus and minimal rice exports in 1978,2 the country has suffered setbacks in rice production in recent years. In 1984, because of the prolonged effects of the 1983 drought and the insufficient supply of farm inputs caused by import restrictions, the rice production growth rate in the first quarter slowed down, bringing the rice inventory of the National Food Authority and other commercial sources to around 327,500 tonnes. To increase the country's buffer stock, the government imported 150,000 tonnes of rice in June. The Philippines, therefore, has reverted to its position as an importer after being a marginal exporter of rice in the late 1970s.

With respect to export crops, coconut output expanded rapidly in the 1970s. There was also a shift in land use from grains to exportable crops after the 1962 devaluation [48], which explains in part the remarkable increase in other food crops from the mid 1960s on-mainly bananas and pineapples, and, in recent years, coffee and mangoes for export [11] .

FIG. 2. Available supply of calories and proteins, after NEDA, Philippine Food Balance Sheets 1281

TABLE 4. Comparison of supply and consumption indicators of relative self-sufficiency with respect to calories and protein, the Philippines, 1970 to 1981

  Available calories per capita per day (kcal)a Food energy consumption per capita per day (kcal)b Per cent available calories of AERc Per cent caloric intake of AERc Available protein per capita per day(g)a Protein consumption per capita per day(g)b Per cent available protein of RDAd Per cent protein intake of RDAd
Year (1) (2) (3) (4) (5) (6) (7) (8)
1970 - 1,929.1 - 94.47 - 46.25 - 90.68
1971 2,123 1,954.0 104.27 95.97 54.9 47.29 107.65 92.73
1972 2,047 1,899.3 100.54 93.29 52.8 46.03 103.53 90.25
1973 2,108 1,806.2 103.54 88.71 54.4 43.21 106.67 84.73
1974 2,259 1,859.0 110.95 91.31 56.6 44.52 110.98 87.29
1975 2,290 1,986.4 112.48 97.56 66.8 47.64 130.98 93.41
1976 2,328 2,053.9 114.34 100.88 66.8 47.97 130.98 94.06
1977 2,418 2,028.7 118.76 99.64 62.8 46.74 123.14 91.65
1978 2,520 2,020.2 123.77 99.22 64.6 46.75 126.67 91.67
1979 n.a. 2,054.4 - 100.90 n.a. 47.48 - 93.10
1980 2,692 2,072.9 132.22 101.81 68.6 47.98 134.51 94.08
1981 2,637 2,066.2 129.52 101.48 67.7 48.76 132.75 95.61

a. National Economic and Development Authority, Philippine Food Balance Sheets [281.
b. Policy Analysis Staff 1321. The crop year data were converted into calendar year estimates.
c. AER = 2,036 kcal/day.
d. RDA = 51 g/day.

The shift from food grain to export production has its nutritional implications, for even if the aggregate supply of nutrients has increased, these will not necessarily be consumed domestically. We note only in passing that whether this development will improve nutritional status depends on whether export crop production will generate higher incomes for deficit groups and whether income increases will be spent on food.

Food Balance Sheet (FBS) data from the National Economic and Development Authority indicate the available supply of calories and protein per person per day for the whole economy from 1953 to 1984 (fig. 2) [28, 29]. The FBS data include domestic production of food, imports minus exports, inventories at the beginning and at the end of each year, food products used for seed, animal feed, industrial and other non-human consumption purposes, and an allowance for waste. Table 4 seems to indicate that the country has achieved relative self-sufficiency: for the 1971 to 1978 period and for the years 1980 and 1981 the average per capita supply of calories and protein is greater than the recommended daily allowance IRDA) for protein and the estimated average energy requirement (AER).

However, FBS data and trends in aggregate nutrient availability should not be interpreted to signify food self-sufficiency for the following reasons. First, the actual amounts of protein and calories needed to meet nutritional standards would be much higher than the per capita averages if one considers the inequality in the distribution of income and food. Mangahas [25] suggests that a safety margin of at least 25 percent is necessary to compensate for inequalities in the distribution of food. Second, the rate of improvement in aggregate nutrient availability has fluctuated through time. While average per capita food supply exhibited a general upward trend from 1953 to 1978, the rate of improvement has not been constant. There was an acceleration up to 1965, followed by a deceleration from 1966 to 1972. From 1972 to 1974 there was a sharp increase in total nutrient availability per capita, probably associated with the government's massive rice production programme, and then a deceleration. The absolute level, however, has increased from 1953 to 1978. Unfortunately, more recent data show a decrease in per capita nutrient supply from 1980 to 1981, although the latter is higher than the 1978 level. The recent decreases in food production and per capita nutrient supply lead one to question the Philippines' attainment of its food self-sufficiency objective.

Food consumption and nutrition indicators from the Ministry of Agriculture Special Studies Division (MA-SSD) surveys and the nutrition surveys of the Food and Nutrition Research Institute (FNRI) reveal that food consumption and nutrient intake have been inadequate, particularly for low-income households [1, 18, 36). Results from the 1982 FNRI survey, for example, indicate that the average adequacy of calories was 89 percent, with reference to an estimated average energy requirement (AER) of 2,032 kcal per day. Protein intake was almost 100 percent of the recommended daily allowance (RDA) of 51 grams per day. While the average figures reveal that food energy was not sufficient in the aggregate, the actual picture is much worse, since the average figures do not consider the unequal distribution of calorie and protein intakes between households and within households. In countries characterized by great inequalities in income, it is possible to have a high incidence of under-nutrition even though the average level of nutrient intake is almost adequate. For the 1982 data, for example, 34 percent of households had an energy intake that was less than 80 percent adequate, a 5 percentage-point decrease from 1978.

The MA-SSD surveys, although not primarily intended as nutrition surveys, provide us with a picture of consumption patterns through time. Quarterly average per capita rates of use from 1970 to 1980 for energy-rich commodities are shown in figure 3. This figure indicates substitutability of rice and corn, demonstrated where abrupt fluctuations were observed during the food crises from 1972 to 1973. However, average per capita consumption of wheat, sugar, and cooking oil generally declined from 1970 to 1980 [36]. The declines in sugar and oil consumption are a bit surprising at first glance, since these are Philippine export crops. However, since production is primarily for the export market, it may be that coconut and sugar trading institutions are more concerned with the external rather than the domestic market. For example, Lim notes that the coconut sector has not been very supportive of the scheme to implement domestic subsidies on coconut oil, which are aimed at increasing consumption of fats and oils by those in need [22].

Regalado also presented data trends in the consumption of body-building and regulating foods (fig. 4) [36]. The seasonality of fruit and vegetable supplies probably accounts for the unstable trends in fruit and vegetable consumption. What is disturbing is the downward trend in the consumption of body-building foods from 1970 to 1980. The decreasing trends in consumption may be attributed to the rising general price level (CPI) for all foods, which has been singled out by Bautista as accounting for more than one-half of the overall inflation rate in the post-war period [2]. Although the average annual rate of population growth was fairly stable from 1970 to 1980, the average growth rate of the CPI has exceeded the growth rate of income since 1970, except for 1975-1976 [36]. The erosion of purchasing power no doubt has implications for the nutritional status of the population.

In the next section, we present the framework of a model that can be used to estimate the effects of price and income policies on the nutritional status of different income groups.

FIG. 3. Average annual per capita rates of use, energy-rich food commodities, 31 surveys, Philippines, 1970-1980, after Regalado [36]


This study used a partial equilibrium, market equilibrium displacement model that extended the work of Perrin and Scobie, generalizing it to cover more income strata [30]. The purpose of the model is to estimate the nutritional effects of food market intervention policies using an income-stratum specific demand elasticity matrix. Extensions of the model are used to calculate cost equations for target-group-oriented policies; the derivations can be found in Quisumbing [33].

This section is divided into two parts. The first introduces the general framework of the model while the second presents the demand elasticity estimates to be used in the simulations.

The Model

Using figure 5, suppose that increasing the quantity of food consumed (Q) from Q to Q' is desired. Any of the following three kinds of policies might be suitable for this purpose.

1. Policies designed to shift supplies to the right (from S to S'). Agricultural production policies fall into this category and include (a) public investment in agricultural research, which generates new information and techniques; (b) public investment in rural infrastructure; and (c) direct subsidies of agricultural inputs. Food importation, while augmenting domestic supply, may be used more often as a price stabilization mechanism, in which case it would fall into the third policy classification.

FIG. 5. Increased food consumption via a supply shift (S), a demand shift (D), or a market wedge (W)

2. Policies designed to shift demand to the right (from D to D'), such as direct income transfers, certain types of food stamp programmer, and nutrition-oriented consumer education programmer.

3. Policies that affect prices or that drive a subsidy or wedge between the producers' price and the consumers' price. For example, with a market wedge of W. producers would receive P1, while consumers would pay P2. These include simple food stamp plans, ration shops, and premiums paid to producers. To analyse food policies, the model requires the specification of supply shifts, income (or food-budget) transfers, and price subsidies as exogenous parameters in each simulation, while demand elasticity matrices from each of the income strata-four in the case illustrated in table 5-constitute the core of the model. The model simulates the price and quantity equilibrium displacement effects of supply shifts, income transfers, and price subsidies. Given the nutrient content of these commodities, we can estimate the effect of these policies on equilibrium nutrient intake. In particular, the model incorporates the possibility of differential response across income groups, and thus provides a means for estimating the distributional effects of food policies.

FIG. 4. Average annual per capita use, body-building and regulating food commodities, 31 surveys, Philippines, 1970-1980, after Regalado [36]

In order to evaluate the cost-effectiveness of food policies aimed at increasing calorie consumption of the lowest income group, treasury costs are expressed as a function of the desired calorie gain of the lowest income group. The estimated costs then provide a basis for ranking such policies on the basis of cost-effectiveness.

Demand Elasticity Estimates

Income-stratum-specific demand elasticity estimates were obtained for four income groups from the 1978 Nation wide Nutrition Survey data of the Food and Nutrition Research Institute [14]. The 1978 Nation wide Nutrition Survey covered 2,800 households in all regions except Regions IX and XII of Mindanao. The data from the Food Consumption Survey, consisting in part of the results of a one-day food-weighing conducted by trained nutritionists, contained information on the consumption and cost of 146 commodities, in the form of weight as purchased, weight of the edible portion, and weight of the net intake.5 Each commodity had its corresponding equivalent for calories, protein, vitamin A, iron, and other nutrients. This data set also provided information on socio-economic factors of the four income groups, such as education and per capita income, fertility and health practices, type of livelihood, and extent of home production.

For the purpose of obtaining income-group-specific estimates, the sample of 2,800 households was divided into four subgroups on the basis of per capita income quartiles. A summary of selected sample characteristics is presented in table 5. The lowest income quartile, with an average income of 190 Philippine pesos per capita, also had the lowest daily per capita intake of all four nutrients (calories, protein, iron, and vitamin A), accounting for only 78 percent of the estimated average energy requirement (AER) for calories, 85 percent of the RDA for protein, 80 percent of the RDA for iron, and 52 percent of the RDA for vitamin A. The low calorie intake of the lowest income group was a cause for concern. The Food and Nutrition Research Institute (FNRI) has set a minimum energy adequacy level at 80 percent of the AER, and the sample results indicated that, on average, at least 25 percent of the sample had a food energy level below this value. The estimate for those with a below-minimum energy level in this sample was 38 percent. The low food energy level implied that protein was not being utilized for bodybuilding requirements, but was being used to provide energy for the body's needs.

Nutritional status improved in the second quartile, with an average per capita income of 490 Philippine pesos. Calorie intake was 87 percent of the AER, protein 101 percent of the RDA, iron 94 percent of the RDA, and vitamin A 65 percent of the RDA. In the third and the fourth quartiles nutrient intakes were close to the AER and RDAs for all nutrients, except for vitamin A in the third quartile.

TABLE 5. Selected sample characteristics, 1978 Nationwide Nutrition Survey, FNRI

Annual per capita income range (pesos) 4-330 333-679 680-1,357 1,360-30,500
Average per capita income (pesos) 190 490 985 2,887
Daily per capita calorie intake (kcal) 1,589 1,769 1,882 2,155
Percentage of AER (2,036 kcal) 76 87 92 105
Daily per capita protein intake (g) 44 52 56 69
Percentage of RDA (51.5 g) 85 101 109 134
Daily per capita iron intake (mg) 10 11 12 14
Percentage of RDA (12 mg) 80 94 98 114
Daily per capita vitamin A intake (IU) 1,870 2,343 2,645 3,753
Percentage of RDA (3,618 IU) 52 65 73 104
Ratio of measured food expenditures        
to measured income (%)a 520 185 114 66
Number of households 682 715 702 700

a. The size of the measured food-budget proportion relative to income suggests a great degree of income understatement. Perhaps these results reveal not only marked differences in consumption patterns among income groups, but also a clustering of observations in the middle-income ranges.
Source: Reference 14.

Demand elasticities were estimated for each income group using individual double-log commodity-by-commodity demand functions of the form:

Qi(m) = quantity of food i consumed by the individual in income stratum m, in pesos per day.
FBi(m) = food budget of the individual in income stratum m, in pesos per day.
Pj(m) = price of food commodity i faced by an individual in stratum m, in pesos per kilogram.
Pj(m) = price of food commodity j faced by an individual in stratum m, in pesos per kilogram.
eii, ejj = price and cross-price elasticities of demand.
Ej = food-budget elasticity of demand.
ui = error term.

Food budget rather than income was used as an explanatory variable for two reasons. First, income data in the 1978 FNRI survey were severely understated, and thus income elasticities estimated from understated income data might not be reliable. In addition, we did not know whether the degree of income understatement differed across income classes. Since this survey was designed primarily for nutrition purposes, food expenditures were more accurately measured than incomes.

Second, it was possible to formulate the demand for food as a function of only two factors: the budget allocation to the food subgroup and food prices. This formulation was dependent upon the assumption of weak separability in the utility function, i.e. that the marginal rate of substitution within the food group was not affected by the value of consumption outside the food group. Having made this assumption, we have obtained price, cross-price, and food-budget elasticities. The food-budget elasticities were converted into income elasticities by multiplying the former with the elasticity of food expenditure with respect to income. A more detailed discussion appears in a later section.

The equations were first estimated using ordinary least squares (OLS) with all the food prices and total food cost as independent variables. Insignificant explanatory variables were dropped in the next round, and homogeneity restrictions were tested. Where restrictions held, i.e. if the sum of price, cross price, and food-budget elasticities was not significantly different from zero, the equation was estimated as a system, including empirically valid restrictions and dropping the equation for miscellaneous products to avoid singularity of the variance-co-variance matrix. Demand parameters for miscellaneous products were eventually computed using the OLS estimate for the food-budget elasticity and Cournot aggregation and homogeneity for the price and cross-price elasticities. The complete seemingly unrelated regressions (SUR) estimates are not presented in this article but are available upon request from the author.

TABLE 6. Food-budget elasticities, 1978 FNRI survey: seemingly unrelated regressions (SUR) resultsa

Rice and rice products 1.708 1.477 1.071 0.553
Corn and corn products - 0.898 - 1.418 - 0.220ns 0.046ns
Other cereal products 1.625 2.177 1.285 2.280
Starchy roots and tubers 0.627 1.047 0.983 1.235
Sugars and syrups 1.771 1.302 1.449 1.419
Dried beans, nuts, and seeds 1.657 1.808 1.944 1.465
Green leafy and yellow vegetables 1.115 0.638 0.916 0.406ns
Vitamin C rich foods 2.338 2.551 2.137 2.528
Other fruits and vegetables 2.014 2.527 1.506 1.435
Fish and sea-foods 2.066 1.001 0.905 0.557
Meat 1.754 2.802 3.244 4.171
Poultry 0.941 0.877 1.583 1.987
Eggs 1.854 2.209 2.691 2.269
Milk and milk products 1.145 2.547 2.115 1.908
Fats and oils 1.802 1.964 1.609 1.109

a. Estimates are statistically significant at alpha = 0.05, except for those marked "ns"
(not significant).


The food budget entered the food-demand function as a substitute for income, assuming a separable utility function. Preliminary results for the income elasticities indicated that, while these were close in value to previous Philippine estimates, the income elasticity estimates were mostly insignificant [35]. The food-budget elasticities, on the other hand, were mostly significant, the exceptions being corn in quartiles III and IV and green leafy and yellow vegetables in quartile IV (table 6).

As expected, food-budget elasticities varied across commodities. Starchy roots, rice, and corn had lower elasticities compared to the more expensive wheat-based cereal products. Likewise, fish and sea-foods were, in general, less food-budget-elastic than poultry, milk, eggs, and meat, in that order. A monotonic decline in the food-budget elasticities was shown for rice, sugars, and fish. An increase and then a decrease was exhibited by dried beans, other fruits and vegetables, eggs, milk, fats, and oils. The elasticities for corn (although negative), meat, and poultry showed a general upward trend, while the other commodities exhibited erratic behaviour.


Contents - Previous - Next