SECTOR Home | What's New | About SECTOR | Meetings | Publications | Discussion Forum | Contact Us | Links | Site Map

Structural adjustment and agriculture in Guyana:
From crisis to recovery

Appendix I

Past poverty lines for Guyana and a new suggestion

An examination of two poverty lines for Guyana for which details are available (which is not often the case) reveals some basic conceptual flaws in them. These poverty lines were derived respectively by the World Bank during a mission in October 1993 and by the Inter-American Institute for Cooperation on Agriculture as part of a survey in rural areas during 1993. Their frequent citation in the literature -- and therefore their acceptance -- demonstrates that there should be some merit in a detailed exposé on how poverty lines should be judged and how done.

Basically, a poverty line has to separate the poor from the not-poor. The first difficulty of course is that no definition is available of what a "poor" person is; the poverty line constitutes this definition: anyone having an income less than the poverty line should be considered poor. The income figure is derived from a basket of goods made up of food, clothing, shelter, etc. Those unable to consume such a basket -- either from purchases in the market or from their own production -- should be considered poor. Minimum standards in terms of calories, yards of clothing, and floor space are used to constitute the basket. Since food is the most basic of the basic needs and congruently takes up the bulk of a person's income at poverty thresholds, it is almost always the starting point in the derivation of all poverty lines. A food bundle is chosen yielding a certain number of calories as the benchmark. Physical standards may also be set for the other basic needs, but more often than not, the food bundle is topped up by a certain percentage to represent the expenditure necessary to purchase the non-food items. Thus the derivation of the food bundle is critical in the derivation of the poverty line and by the same token the best test of a poverty line is to look at how the food bundle is composed.

The derivation of the food bundle may be thought to be easy, since a well-known and accepted threshold is available for food -- around 2,100 calories per capita per day -- but the problem arises since those many calories may be obtained from any number of combinations of foodstuffs which differ hugely in terms of cost. A food bundle aspiring to be accepted as a benchmark for defining poverty should contain only the cheapest foods -- not in terms of price per kg but price per calories. Many poverty lines fail this simple test, and this happens to be the case with the two poverty lines under discussion here.

The procedure

The procedure requires arraying the most common foods in any given country in terms of their cost per calorie. Two pieces of data needed for this are: price per kg and calories per kg, the former being obtained from surveys routinely carried out by the statistics office in relation to the consumer price index -- or failing that (or for the sake of up-to-datedness), directly from the market, and the latter from FAO sources. From these two data we obtain the price in terms of $ per (1,000) calories. The poverty basket should then attempt to use the cheapest items of foods. It happens (and this can be verified for the case in question) that in all countries, rich as well as poor, the cheapest foods are the starchy staples -- cereals and roots and tubers -- and in all countries the poor attempt to satisfy their basic calories by using the maximum amount of these staples, along with a sauce made up of vegetables and meat, usually the cheapest cuts. Often the starchy staples may provide upwards of 80 per cent of the daily calories of poor people in poor countries, which proportion then gradually falls to half for the richer people. We should not expect to see shrimp in poor people's diet! Thus a quick test of any poverty line is to check the percentage of calories derived from the starches; anything less than 40 per cent should be considered prima facie suspect; the appearance of luxury items in the basket should be grounds for its summary dismissal as a poverty line! Strange though it may seem this is indeed what happens in one of the poverty lines -- that by the World Bank. The other poverty line -- by IICA -- is not so easy to discard prima facie, but a disaggregated analysis leads to the same conclusion based on the types of foods included in the poverty basket.

Table A.I.1 provides the groundwork for the detailed examination of the two poverty lines. It should be read column by column. In the first three columns are shown for each food item the amount consumed, its cost, and the number of calories derived. These three pieces of data are explicitly shown in the World Bank source. The next three columns show the underlying data on prices per kg, calories per kg, and price per calorie. the last set of figures being crucial in the calculation of the poverty line. The calories per kg figures obtained are of the sort of magnitude found in FAO food composition tables, hence there is nothing to dispute there. However, some conceptual problems with calorie figures should be noted. Foodstuffs come in various forms -- e.g. maize as on-the-cob, decorticated, dried, ground. Clearly the calorie values would differ. "Beef" can be even more deceptive. The calorie content would vary according to the cut -- and therefore the fat content. The problem arises since consumption surveys often fail to stipulate the exact form of any given food; certainly almost none ever differentiate between the type of "beef" consumed by different classes.

The calorie prices confirm the starchy staples as the cheapest sources of calories, although significant differences exist, and are in fact quite revealing. Rice is altogether the cheapest source of calories, among the starches three times cheaper than the roots and tubers group (cassava, plantains, etc.). Shrimps, not surprisingly, is the most expensive source of calories in Guyana, but, vegetables, surprisingly, are not too far behind. The hierarchy of foodstuffs gives us ready clues on how to compose our poverty bundle -- indeed, how the poor would do it themselves. The idea of substitutability permeates the discussion. Rice should dominate the diet and unless there are compelling reasons for including the roots and tubers group (racial preference, variety), they should be disfavoured. Similarly between oil and margarine, the latter should be excluded: although tastier, the poor simply do not have the luxury to indulge their palates. Shrimps should be avoided altogether! The point is, if such expensive foods are included in the poverty basket, we would certainly find lots of "poverty" -- meaning in this case, people failing to meet the calorie threshold specified, but composed of a diversified diet. In trying for such diversification they fail to gain their basic calories: they go hungry. In real life, of course, such "poor" people would simply curtail their consumption of expensive foods and divert the proceeds to the cheaper foods, at much gain in terms of calories. In fact they would not in the first place include such items in their consumption -- and neither should a poverty line.

When we turn to column 8 we find this is exactly what the World Bank poverty line has done. Rice gets only 10 per cent of the calories, the same as flour and spaghetti which are considerably more expensive, starchy roots and tubers are accorded even more calories than rice although they are even more expensive, meat group is given as many calories as rice, although this group as a whole costs around nine times more than rice. In consequence meat "eats up" nearly a quarter of the poverty line (column 7) while delivering only 10 per cent of the calories. Finally, among the meat group figure the famous shrimps. They calorie cost is 18 times as high as for rice; 23 grams are included (the same amount as for all the other meats, about which more later), yielding 14 calories at the cost of G$2.75 -- i.e. almost 3 per cent of the poverty line, more than allocated to rice! It bears repeating that that amount of money would deliver almost 18 times as many rice calories.

Why do we get this sort of "poverty line"? Not because the Bank was not striving to establish a "minimum low-cost food basket" (p. 10). It happens because the World Bank quite evidently consulted the Guyana Agency for Health, Environment and Food to define its poverty basket. Nutrition authorities by their mandate are bound to recommend a diet in which proteins dominate, but this renders the proffered basket inappropriate for measuring poverty, which is not its purpose. No doubt GAHEF gave the World Bank a "basket that provided the most nutrition at the lowest cost" (p. 10), but "most nutrition at the lowest cost " is certainly not the same thing as "the required number of calories at the lowest cost". Proteins, vitamins, iron etc., would figure heavily in the recommendation of any nutrition agency and it is quite possible that the basket used by the Bank satisfies the minimum daily recommendations of GAHEF with respect to such nutrients (a detailed breakdown is given, table annex H, World Bank, 1994) as well as the minimum number of calories. While non-carbohydrate foods are certainly important, poverty analysts have to recognize that for practical purposes they are out of reach of poor people. Why even for the sake of nutritional variety shrimps could be suggested must remain a mystery. There is also the curiosity noted previously about the amount of shrimps specified -- the same as for all other meats. This is also noticed for the other foods, whereby different foods in the same group are allotted the same weight; there is no known rationale for this. The same thing happens, as we shall see, in the case of the IICA poverty line. A final curiosity to notice about the Bank's estimate is its use of 2,400 calories as the norm -- said to be "a commonly accepted standard". Most people go for something around 2,200 calories.

Table A.I.1.

Often researchers approach the trades unions to obtain their basket for use as a poverty line. Quite naturally the TU's calculations are there to be used in the advocacy for higher minimum wages and therefore likely to contain a great deal of nutritional variety; the basket may even have been done by a nutrition agency. This happens in the case of the IICA poverty line. The poverty line is said to have been prepared with the assistance of IFAD (IICA, p. 81), but the real source would seem to be "a recent effort by the Federation of Trade Unions in Guyana to prepare a basic food basket" (ibid.). At first sight the calorie composition (column 9) appears to be of the right order of magnitude, but this is only because rice has been given 38 per cent. However, looking at the substitutables within each food group, we soon find that the more expensive foods are not resisted. In the cereals group, noodles and bread are given nearly 2.5 times more weight in terms of cost (column 8) than rice at a greater outlay than rice; in the oils group margarine is included; and fruits and vegetables together take up more expenditure than rice -- 20.2 per cent compared to 10.5 per cent given to rice, from which 6.3 per cent of calories are derived compared to 38.4 per cent from rice. Such strategic choices push the IICA poverty line towards the same realm as the World Bank's. The cost of the 2,200 daily calorie basket (notice at the level usually stipulated) turns out to be G$83.12 -- or G$37.78 per 1,000 calories, which is just about the same as for the World Bank (G$40.46 per 1,000 calories). Daily calories in Guyana can be obtained for much less than this -- conceptually, as well as in real life by poor persons. There is also the same curiosity as found in the Bank's poverty line -- most items are calibrated in multiples of 22.7 grams, practically the same as for the World Bank!

Table A.I.2 shows a suggested poverty line for Guyana based on the principles derived above, mostly that the line should indeed be a "minimum" threshold, but also that it should take account of local preferences. Rice dominates at 50 per cent of total calories; according to Guyana's food balance sheet rice provided 31 per cent of total calories on an average (Appendix II) so that for the poverty groups 50 per cent would be a reasonable figure. Roots and tubers are included to account for local preferences, as well as peas, which are a part of the local diet and moreover provide cheap sources of calories (and proteins to boot). Vegetables are included despite their price, to provide the basis of the soup with rice. The daily cost of the postulated poverty line would be G$66.29. It provides 2,200 calories, so its price is G$30.13 per 1,000 calories. In contrast, the World Bank's poverty line came to G$97.11 for 2,400 calories, i.e. G$40.46 per 1,000 calories, and IICA's G$83.12 for 2,200 calories, i.e. G$37.78 per 1,000 calories, so that the poverty line suggested here is 75-80 per cent of the existing ones. However, almost one-half of the basket goes in purchasing vegetables and meat, at the addition of only 10 per cent of the calories. These would be candidates for any further paring down of the poverty line.

Table A.I.2. A suggested poverty line for Guyana, 1992-93 (per person)

		Calories composition (%)	Equivalent calories per day	Calories/kg	Price/kg	Equivalent price/1,000 calories	Cost per day (G$)	Percentage breakdown
Rice		50	1 100	3 614	39.55	10.94	12.03	18.1
Roots and tubers		20	440	1 000	33.10	33.10	14.56	22.0
Peas		10	220	3 429	88.21	25.72	5.66	8.5
Vegetables		5	110	215	40.36	187.72	20.65	31.2
Meat		5	110	1 723	165.07	95.79	10.54	15.9
Oil		10	220	8 820	114.29	12.96	2.85	4.3
Total		100	2 200	-	-	30.13	66.29	100.0
Memo items World Bank IICA						40.46 37.78
Sources: Authors' estimates based on postulations in col. 1 regarding calorie composition and a threshold of 2,200 per day in col. 2. Calories per kg. (col. 3) from World Bank, 1994, table annex II; these figures are almost the same as in FAO's food balance sheets, discussed in Appendix II. Col. 4 from World Bank and the rest of the figures are derivations, including memo items from table A.I.1.

Total poverty line

It is possible in practical terms to compose a minimum non-food basket -- yards of cloth, square metres of space, bus fares, school fees, doctors' visits -- to arrive at a non-food poverty basket, which can then be added to the food poverty line to obtain a total figure. More often, the food component is "bumped" up by a defined percentage to arrive at the same figure. The percentage by which the food basket is elevated is decided on the basis of either specific data available or generally accepted notions of consumer behaviour. Regarding the latter, it is established that food being the most basic necessity of households, and poor households by definition having small incomes, a preponderant proportion of incomes has to be devoted to food -- 60-80 per cent being not uncommon. Specific data available from Guyana's household survey confirm this (table A.I.3).

Table A.I.3. Share of food and rice by quintiles, 1992-93

		Quintiles						Total
		I	II	III	IV	V
% of expenditure on:
Food		67.0	61.1	56.8	53.1	44.4		55.2
Rice		30.8	21.3	18.4	15.5	7.9		13.7
Beef		0.07						2.54
Source: HIES, using tabulations provided in World Bank, 1994, appendix table 12 and text table 1.10.

The poorest quintile had to devote two-thirds of their income to food, a proportion that fell to 44 per cent for the richest group. A similar trend is shown for rice, both these sets of figures confirming that as people get richer they have more freedom to purchase non-food items and, amongst the food, more non-staple items.

From the above, two facts may be retained for our analysis: (i) the proportion of expenditure devoted to rice in our poverty line (18.1 per cent) is about right, being at the consumption pattern of the third quintile (18.4 per cent on rice); (ii) using a coefficient of 67 per cent for food in the total poverty line would be a reasonable assumption. The World Bank used 75 per cent -- i.e. the food basket was topped up by one-third to arrive at the total basket,⁽²⁾ whereas with the 67 per cent assumption it has to increase by one-half. The two respective total poverty lines would thus diverge by less than the food poverty line. The recapitulations are as follows:

Poverty estimates

The normal procedure is to apply the estimated poverty lines to an income (or expenditure) distribution table. This is captured in table A.I.4 where the poverty lines are juxtaposed with incomes data. The table notes explain the procedure in deriving the poverty estimates. For example, the food poverty line of this study was put at around G$2,000 per capita per month. From the first column of the table we see that 3.76 + 4.26 + 5.61 = 13.6 per cent of households had incomes below this threshold and therefore should be considered to be in food poverty. Similarly, 27.5 per cent of households fell below the total poverty threshold of around G$3,000 per capita per month. As opposed to this, the Bank's estimate (as reported) was 27.6 per cent of households in food poverty and 41.9 per cent in total poverty. These figures are significantly different from ours. The only way to interpret the difference is to remember what we are measuring -- a notion of poverty -- and the only judgement of that is what may or may not be expected to be in the poverty bundle.

Table A.I.4. Percentage distribution of households and persons by monthly income

(in G$'000) per capita juxtaposed with poverty lines; and ensuing poverty estimates, 1992-93

Appendix II

Food balance sheet, 1996

A food balance sheet shows succinctly the types of foods produced and consumed in a country. Data on food production are routinely collected by the FAO from national sources and then converted through the FBS procedure to consumed food and calories. To food produced are added net changes in stock and net imports, and subtracted output used for seed and feed and output wasted (SFW). For most foods the harvested product has to be rendered into its processed form -- paddy to rice, sugar cane to sugar, maize on the cob to maize grains -- and extraction rates have to be used for this purpose based on local conditions. Coefficients used for SFW are also based on local conditions.

The first part of table A.II.1 shows as an example the essential steps involved in converting rice to rice available for domestic utilization, and finally calories per day. Rice being the most important food in Guyana is an appropriate example to use. Conversion of paddy in the field to rice has already been done, without being shown, in the FAO source. The same steps would be used in converting other food items to calories and the ensuing calorie figures are shown after the paddy calculations. Some notable features of the food profile of Guyana may be highlighted:

• At 2,392 calories per capita per day, food availability is 13 per cent above the norm of 2,100 calories.
• Rice dominates the diet at around one-third of total calories.
• The second placing of wheat (17.4 per cent of total daily intake) is indicative of the high urbanization of the Guyanese population where bread has become a convenience food.
• With rice exports dominating, but also sugar exports, the country is a net exporter of calories. Imports, however, are important for specific food categories. All the domestic consumption of wheat is imported, 82 per cent of pulses, 31 per cent of vegetable oils, 23 per cent of meat, 67 per cent of milk, and 6 per cent of eggs. The country produces roughly 2,470 calories per capita per day, exports 570 calories in the form of rice and sugar, and imports 500 calories from the above foods, mostly in the form of wheat.
• Adding other cereals and roots and tubers shows that the diet is heavily concentrated on starchy staples, with 54 per cent of total calories.
• Meat, eggs and milk provide just 9.7 per cent of total calories, fish 3 per cent, and seafood practically none.

The above indicators reinforce the picture of general poverty in Guyana: food supply is at bare sufficiency level and diets are concentrated on cheap sources of calories. The known great variations in regional food production implies particular deprivation for the more remote regions.

Table A.II.1. Food balance sheet for rice, and calories from other foods, 1996

('000 MT per year first part and then calories per capita per day)

				Per cent
	Per year ('000 MT) Riceb	362.8
	Minus: Stock change Export	+10.9 -276.7
	Equals: Domestic utilization	97.1
	Minus: Seed Feed Waste	9.3 21.1 -
	Equals: Available for food ('000 MT p.a.) Equivalent: Grams per capita per daya Calories per capita per dayb	66.6 218 750		31.4
	From other foods: (Cals./capita/day) Wheat Starchy roots Sugar Pulses Oil crops and vegetable oil Vegetables Fruits Meat, offal, animal fats Eggs Milk Fish Crustaceans	417 115 289 61 240 8 99 124 25 82 72 1		17.4 4.8 12.1 2.6 10.0 0.3 4.1 5.2 1.0 3.4 3.0 0.0
	Total	2 392		100.0
a Population put at 838,000 in the source. b Calories per kg put at 3,445 in the source. Source: FAO, Food balance sheet for Guyana.

International Labour Organization (ILO): Contact us | Site map |

Copyright and Permissions 1996-2008 International Labour Organization (ILO) - Disclaimer