Previous PageTable of ContentsNext Page

Edible waste at time of ingestion

(prepared by FAO)

Recommended requirements for calories or nutriments are calculated from the physiological standpoint, that is, they represent values that must be supplied by foods actually ingested.

It is therefore desirable to have data on food consumption as measured at the stage of ingestion. For various reasons this is not possible and the available data deal with foods available for consumption at the "retail" level, in other words, with foods reaching the kitchen before they are prepared for consumption.

Allowance must therefore be made for possible wastage between retail purchase and the ingestion of the foods. These wastes are of two kinds: those that can be allowed for by adjusting food composition tables (extraction rate, peelings, bones, etc.) and those produced at either the cooking or the ingestion stages (plate, dish and saucepan waste).

Edible Waste at Time of Ingestion

It is difficult to obtain a precise idea of this kind of loss since food consumption surveys are usually published with plate waste already excluded. We must therefore resort to logical reasoning, backed up by a few figures.

It is reasonable to consider that food wastage always increases with supplies and that various factors such as climate, economic level, household equipment and custom may play a part.

On this general assumption, we shall consider the likely losses in developing countries and in countries where on the average there is plenty of food.

Plate Waste in Developing Countries

With regard to low income countries, workers responsible for surveys have given similar replies that edible waste is very small. There is doubtless some risk of food loss through spoilage but, in general, populations are aware of climatic conditions and refrain from cooking too much to avoid keeping cooked foods under adverse conditions. Moreover, domestic animals who are often famished are usually given whatever foods are left on the plate.

Thus, in developing countries the custom is to adjust the quantity of food cooked to what can be consumed at a given time which will minimize the risk of spoilage. These conclusions are borne out by few data such as from the national survey (1) performed in Madagascar (5000 households). The absence of edible waste, in so far as foods of animal origin (fish, meat) are concerned, is noteworthy. For cereals and starohy roots and tubers, the following percentages by weight were reported:

Rice3.4 %  


Sweet potato1.2 %
Maize5.6% Potato3.2 %
Manioc1.6% Bananas0.3 %

Some percentages prove to be higher than others because of local customs. For example, potato and maize waste result from quantities of food cooked in the family pot giving allowance for part of them to be fed young pigs and other small animals 1. Rice waste accounts for plate and especially saucepan waste. The custom is to cook the rice in a small amount of water and to toast what remains at the bottom of the pot to obtain a healthy beverage by the addition of water.

For all other products, the survey shows negligible losses at national level. If cereals and other starchy foods are replaced in the food diet as a whole, total calorie and protein losses are found in reality to be low. For the province of Tananarive whose staple food is rice, daily per caput losses of calories are 2.8% and of proteins 2.3%; for the province of Tuléar where the diet is based on manioc, losses are 2.1% for calories and 1.8 % for proteins respectively.

Plate Waste in Developed Countries

For countries where food is plentiful, it is even more difficult to obtain an idea of food losses. Improved household equipment doubtless produces the risk of waste through spoilage but plate losses are probably higher. Moreover, meal leftovers kept in the refrigerator are not necessarily consumed and may be thrown away after a certain time.

Furthermore, in countries which are more used to industrially prepared foods, it is not impossible that food waste from tins must be added to the various losses recorded.

Food given to domestic animals may serve as an indicator on the extent of plate waste as suggested by the data in Annex I. It is suggested that the feeding of domestic animals amounts to 5% of the calorie intake of humans. A similar percentage may be assumed for protein in the absence of a more reliable data.

Information is fragmentary for food waste other than that given to domestic animals. The surveys are mostly old and their statistical quality leaves something to be desired. Here we shall consider only those sources that appear the most reliable.

A series of surveys carried out between 1932 and 1934 by Cathcart and Murray (6) shows fairly low calorie losses comparable with what seems likely in under-developed countries today.

Calorie losses in four towns - U.K.

  Number of familiesAverage calorie losses
St. Andrews1492.7 %
Cardiff531.1 %
Reading491.8 %
Glasgow121.5 %

In view of the enormous change in standards of living since the second world war, there is reason to think that these data are low compared with the present situation.

A survey of 14 households carried out in 1951 by members of the Scientific Adviser's Division, Ministry of Food, U.K., (7) shows plate waste of an average of 3% for calories, with a range of 1 to 7 %. However, it seems that a fair number of meals were taken outside and thus these percentages represent only part of the ingestion losses for the group considered.

An estimate of plate waste in the case of meals taken outside the home can be attempted on the basis of the results for army messes.

Calorie losses in army messes

SourceSurveyCalorie losses %
(8)US 27 messes (1917-1918)7 %
(9)US 455 messes (1941-1943)8 - 10 % 1
(9)US 99 messes (1943)7 - 9 % 1
(10)US 44 messes (1945)6 - 8 % 1
(11) UK (1949)  
 55 men2.4 %
  12 men1.6 %
  12 men6.2 %
  15 men3.7 %
  25 men2.8 %
  Average2.8 %

1 The highest percentages correspond to the assumption that half of the waste was fat.

Another survey performed in Australia in 1938 (12) covers 2 532 meals served over four days to 211 persons in a children's institution. Half the plate waste was estimated at 4.6% of the ingested calories.

In considering these three sources of information one notices deviations of 1.6 % to 10 %. If the idea is accepted that waste in U.S. messes represents a limiting case, the assumption can be made that losses in communities and restaurants must be between 4 and 5 % of the calories consumed.

According to the latest censuses in the United States and France, the population enumerated separately (institutions, communities, barracks, hospitals etc.) represents between 2.5 and 3.5 % of the total population. If an average plate waste figure of 5 % is allocated to persons living collectively, the effect on national food balance sheets remains very low, about 0.1 to 0.2 %.

The population of rich countries is generally 60 or 70 % urban (see the United Nations demographic yearbook). It should be considered that part of these city dwellers regularly take meals in restaurants or work canteens. Taking into account holiday and weekend habits, a maximum of 230 meals is taken outside.

If two main meals are assumed each day, these 230 meals are equivalent to 115 days with two meals. Taking into account persons not gainfully employed (55 % minimum of the urban population) and small towns where the workers eat at home, it appears realistic to accept a minimum drop of 50 % on meals taken outside with respect to the whole urban population.

On these assumptions and considering plate waste to be comparable to waste in messes, which is probably an overestimate, the effect of these losses on national food balance sheets is around 0.03 % (U.K. and France) to 0.06 % (United States). Even if urban households were assumed to visit the restaurant as often as workers eat in their canteens, the figure of 0.2 % of calorie consumption would not be exceeded.

Recapitulating all the estimates of losses which it has been possible to pinpoint, it emerges that a figure of 8 % is certainly realistic as an average for rich countries.

Losses at the Cooking Stage (13)

In addition to plate waste, allowance must be made for losses due to washing and cooking of food since such losses are not taken into account in the compilation of food composition tables. The few available documents show that washing and cooking have an essentially destructive action on vitamins such as thiamin, riboflavin, niacin etc. On the other hand, the biological value of protein in meat, fish and milk does not seem to be affected, any more than the energy value is. Taking into account that the losses of lysine (15 %), threonine (11 %) and methionine (10 %) observed during the roasting of groundnuts have little effect on the food balance sheets, we may conclude that calorie and protein balances probably do not have to allow for losses at the cooking stage, at least in the present state of our knowledge.

General Conclusion

The second committee on calorie requirements suggested in 1956 an overall increase of 10 % in calorie and protein requirements in order to allow for plate waste which, included as it was in the daily per caput ration given in the food balance sheets, overvalued the ration by an amount which was agreed to be 10 %.

The present study suggests:

1) that this percentage is too high for developing countries. A drop of 5 % for calories appears to be more realistic. For proteins, the use of the same percentage must be considered to be an overestimate;

2) that for rich countries the retention of a figure of 10 % is certainly more than sufficient.

Annex I

Human population 1964-1966191,900,00054,700,00049,157,000
Daily per caput consumption      
Calories 315632233108
No. of dogs22,000,0003,500,0005,000,000
Average requirements      
(Average weight 10 kg)      
No. of cats30,000,0004,800,0006,800,000
Average requirements      
(Average weight 3.2 kg)      
Percentage of human ration      
I. Calories      
Dogs3.9 %2.2 %3.6 %
Cats1.3 %0.7 %1.1 %
Total5.2 %2.9 %4.7 %
II. Proteins      
Dogs10.4 %6.1 %8.8 %
Cats3.7 %2.2 %3.2 %
Total14.1 %8.3 %12.0 %

The figures given in this table are rather theoretical for the following three reasons:

1. Coverage of animal requirements from the human ration

Calculation of the food of domestic animals as being entirely taken from the human diet is obviously a limiting working assumption, especially for animal protein. If countries which waste meat because of over-supply (such as Argentina or Uruguay) are left aside, the plate waste of animal proteins is not such as to meet the protein requirements of domestic animals. A food complement is supplied either by scraps bought retail (and already left out of the food balance sheets at the carcass weight stage) or by special tinned foods1 which use the same offals, or again by hunting on the part of the animals concerned (rats, mice, etc.).

For calories, this is probably not the case and animal food is in fact often derived from the human diet. Hence, it may plausibly be suggested that the feeding of domestic animals amounts to 5% of the calorie consumption of human beings. A similar percentage could be taken for protein in the absence of highly reliable data.

2. Indeterminacy of statistics

For dogs the statistics are fiscal ones corrected for a certain estimated failure to declare. Data were taken from the following encyclopaedias:

Chambers Encyclopaedia 1955

Encyclopaedia Americana 1968

Grand Larousse encyclopédique 1960.

For cats, only an American estimate was available and the same dog/cat ratio has been kept for the other countries.

3. Approximate value of the average requirement

Requirements were calculated in kilogrammes of live-weight from standards used in laboratories (2, 3, 4). However, it is very difficult to pinpoint the average weight of a population of dogs or cats. For the latter the average adult weight was taken, while for dogs the figure of 10 kg of live-weight was adopted as a reference since it was believed that development of the dog population was becoming an urban phenomena which would favour the medium and small breeds in the countries considered here.


(1) Ministère des Finances et du Commerce, République Malgache. Budgets et alimentation des ménages ruraux. 1962.

(2) Falin, L. et Griess, D., L'Alimentation du chat. Cahiers de médecine vétérinaire, mars-avril 1970.

(3) Abrams, J.T., Animal Nutrition and Veterinary Dietetics. Ed. W. Green and Son Ltd., Edinburgh. 1969.

(4) Worden, A.N., Feeding of laboratory dogs and cats. Proceedings of a symposium organised by the British Small Animal Veterinary Association. Pergamon Press Ltd., London. 1964.

(5) Petfood market in Tidsskrift for Hermetikindustri, March 1970, Oslo.

(6) Cathcart and Murray. Medical Research Council Spec. Rep. Ser. No. 165, 1932; No. 218, 1936; No. 242, 1940 and the Journal of Hygiene 39, 1939.

(7) Unpublished records, Ministry of Food, UK, 1951.

(8) Murlin, J.R., Hildebrandt, F.M., Average food consumption in the training camps of the US Army. American Journal Physiol. 49, 1919.

(9) Howe, P.E., Berryman, G.H., Average food consumption in the training camps of the US Army. American Journal Physiol. 144, 1945.

(10) Schor, C.H., Swain, H.L., Simultaneous surveys of food consumption in various camps of the US Army. Journal Nutrition 38, 1949.

(11) Arneil, G.C., The losses of edible food due to plate waste in Army dining halls. Brit. Journ. Nutr. 2, 1949.

(12) Bourne, G., Report of the Biochemist. Appendix 1, 5th report of the Advisory Council on Nutrition. Canberra, 1938.

(13) Lang, K., Influence of cooking on foodstuffs. World Review of Nutrition and Dietetics. Vol. 12, 1970, S. Karger, Basle, Switzerland.

1 Emma Reh has observed in Guatemala that 7 % of maize cooked for the family might be given to animals (in allowance for waste FAO/56/4/3157).

1 Pet food market per year (5)
USA 1,400,000 metric tons -> 1 billion dollars.
UK 300,000 metric tons -> 115 millions dollars.

Previous Page Top of Page Next Page