Trends in urban and rural food consumption and
implications for food policies in Tunisia
J. Périssé and A. Kamoun
This study of the evolution of food consumption among
urban and rural populations in Tunisia is based on the
results of household consumption and budget surveys.
To make sure that any differences observed in such a
sutdy actually reflect the consequences of urbanization
and are not attributable to mere chance, it is essential that
the surveys should meet a number of criteria:
-- They should have a national sampling framework
that ensures a proper distribution of households,
taking into account both seasonal variations and
geographical differences.
-- Recording of household budgets, showing the
amounts of money spent on foods, goods, and
services, should be supplemented by actual weighing
of food commodities on various days, particularly
in those countries where home-produced, home-
consumed foods are an important part of the diet.
-- For any study of the evolution of consumption over
time, the methods of data collection, the concepts
followed, and the way of expressing the results
should be identical from one survey to the next,
lest distortions be introduced that might appear to
reflect behavioural differences.
TABLE 1. Urban and rural population and growth rates in
Tunisia,1975-1985
Population (1,000s) Average annual
growth rate (%)
1975 1980 1985 1975-80
Urban 2,656 3,325 3,880 4.6
Rural 2,932 3,044 3,274 0.8
Total 5,588 6,369 7,155 2.7
Tunisia is one of the very few countries to have carried
out four such national surveys between 1965 and 1985.
Of the four surveys (in 1965-1968, 1975, 1980, and 1985),
conducted under the auspices of the National Institute of
Statistics and funded by the government, the last three
met the criteria listed above. This article compares them
and comments on their results.
Food and Nutrition Bulletin, vol. 9, no. 4 © 1987, The United Nations University.
Population trends
For the purposes of the present study, we have divided
urban areas into two groups: one (U) comprising small
and medium-sized cities with their own town councils,
and another (C) comprising the major cities with over
90,000 inhabitants, such as Tunis, Bizerta, Sousse, Sfax,
and, since 1985, Gabès. The rural areas (R) consist of
villages, hamlets without any administrative structure,
and scattered dwellings.
The evolution of the Tunisian population from 1975
to 1985 is shown in table 1 (based on the 1966, 1975,
and 1984 censuses). The national population increased
by 28% between 1975 and 1985 and is now above 7.2
million, as against 5.6 million 10 years ago. During that
time the population of the urban centres, including both
the smaller and the major cities (U + C), rose by 46% and
that of the rural areas by only 12%.
For the first five years of the period under review, the
rural population showed only a slight yearly rate of growth
(0.8%) as the consequence of a marked exodus to the
cities (and to foreign countries). The cities, on the other
hand, had an annual growth rate of 4.6%. Tunisian cities
are thus becoming overpopulated, and in 1980 contained
the majority of the overall population: 52%, compared
with 47% in 1975. Because employment opportunities
in the cities have become scarcer, movement from the
rural areas has somewhat abated since 1980: the rural
population has tended to remain settled and increased
1980-85
twice as fast between 1980 and 1985 as during the five
3.1 previous years (1.5% versus 0.8%). As a result, the
1.5 growth of urban population slowed down (3.1 % versus
2.4 4.6% as previously).
Food-consumption patterns
The traditional Tunisian diet is dictated by the ecology
of the Mediterranean basin. it is characterized by a
high consumption of durum wheat and olive oil, with
a moderate and uneven intake of animal products
(mutton, eggs, milk), and is very diversified with regard
to vegetables and fruits. This pattern, however, is not
1
2 Trends in urban and rural food consumption

static but is undergoing changes under the influence of A change that is clearly taking place is a strong
economic development and the gradual urbanization increase in the consumption of imported soft wheat
of households. flour in the form of bakery bread, whereas the use of the
The urbanization process entails deep changes in same imported flour at home is being discontinued. This
the composition of the diet. These changes become increased use of soft wheat is far from making up for the
significant with time (table 2) and influence the levels drop in consumption of durum, so that the level of cereal
of energy and nutrient intake (table 3). In 1975 the intake is diminishing notably as the population becomes
consumption of locally grown durum wheat products urbanized.
in the smaller cities was half as much as in rural areas, Other commodities such as sugar and blended oils
and in the larger cities it was only one-fourth as much. show some increase in consumption, although only
A sharp decrease has been noted in the consumption of to a moderate extent. Because these have already been
couscous, semolina, and traditional pancakes that are strongly integrated into traditional recipes and beverages
usually prepared at home in rural areas. Similarly, the (green tea) for some considerable time, their use is not
customary consumption of barley (that of maize and altered by increases in income. The consumption of
sorghum being negligible) has fallen rapidly as a result olive oil is higher in the smaller cities, where most of the
of urbanization, and the grain has become a cultural relic owner-producers live.
without any nutritional significance. Urbanization is, however, followed by sharp increases
TABLE 2. Food-consumption patterns (kilograms per in the consumption of meat (mutton and beef), poultry,
person per year) in relation to urbanization in Tunisia, 1975 fish, fresh or reconstituted milk, and yoghurt. Typically,
and 1985 when settling in a city, Tunisians diversify and improve
the palatability of their diet by eating fewer staple foods
1975 1985
and more animal products. With regard to legumes,
R U C R U C vegetables, and fruit, such changes are less pronounced,
Durum 129.0 65.6 33.7 126.4 49.7 29.9 and rather take the form of consumption of more exotic
wheat varieties to the detriment of home-produced vegetables
Barley 15.2 2.6 1.8 11.7 4.4 1.8 and fruit or the yield from gatherings.
Flour 10.9 5.4 1.6 6.6 4.4 3.0 The results given by the different surveys are
Bread 20.4 73.7 99.4 32.4 84.1 100.8 remarkably consistent. In general, the decreasing trend
noted in 1975 with regard to traditional staple foods and
Pulses 5.2 6.2 5.7 3.6 4.7 4.8
the increasing trends for complementary foods were
Potatoes 12.6 21.0 21.0 14.3 22.3 23.1 confirmed in 1980 and 1985. A minimum threshold of
Olive oil 6.0 8.5 3.6 4.8 8.7 4.8 about 30 kg per capita seems to have been reached for
Blended 11.8 11.6 14.8 14.0 13.3 17.9 durum wheat in the larger cities and a maximum peak
oil for bread of about 100 kg. For a few commodities, such
Sugar 13.9 13.5 16.7 15.3 15.6 19.7 as milk in urban (1975) and rural (1980) areas, and
poultry in all three areas (1985), a number of seemingly
Meat 9.8 13.5 18.5 8.2 15.6 24.0
accountable decreases have occurred. They are discussed
Poultry 2.2 2.1 2.1 3.4 3.9 6.9 later in relation to changes in income because they seem
Fish 2.1 7.5 9.5 1.4 6.5 10.2 to be the consequences of policy decisions and not just
Eggs 1.3 2.2 3.2 1.9 4.3 6.5 inconsistencies in the survey records.
Milk 36.3 19.8 54.8 26.4 37.7 55.3
Yoghurt 0.2 0.6 3.3 1.3 3.2 5.5
R = rural; U = small and medium-sized cities: C = major cities (see Text)

TABLE 3. Levels of energy and nutrient intake per person per day in relation to urbanization in Tunisia, 1975-1985.

1975 1980 1985

R U C R U C R U C
Calories 2,474 2,228 2,122 2,452 2,274 2,232 2,424 2,128 2,189
Carbohydrates (g) 398 336 307 401 348 325 411 311 323
Total net protein (g) - - - - - - 68.7 59.4 63.8
J. Périssé et al. 3

Vegetable protein (g) 67.7 52.6 44.8 68.2 54.0 47.5 - - -

Iron (mg) 21.2 14.9 13.2 14.1 13.8 13.4 15.1 12.9 14.2
Thiamine (mg) 1.73 1.33 1.28 1.69 1.36 1.23 1.58 1.31 1.32
Niacin (mg) 17.2 13.2 11.6 16.2 13.8 12.8 15.9 12.5 13.2
Animal protein (g) 7.0 8.9 14.1 8.8 15.3 18.9 - - -
Lipides (g) 64.9 71.1 72.9 68.1 73.2 79.3 61.4 66.4 74.2
Retinol (µg) 485 658 690 552 779 763 568 731 865
Vitamin C (mg) 61 83 90 66.0 86.8 88.8 74 93 117
Calcium (mg) 438 391 501 451 477 527 394 440 568
Riboflavin (mg) 0.84 0.75 0.86 0.82 0.86 0.91 0.81 0.80 0.98

Nutrient intake (niacin); the final result depends on the predominance of

The above changes have obviously affected the energy,
nutrient, vitamin, and mineral content of the average
diet. Table 3 shows the changes in per capita consumption
of various nutrients in relationship to the level of
urbanization. Intake of energy, fat, vegetable protein, iron,
and thiamine is greatest in the rural areas and decreases
with urbanization. This decrease is attributable to lower
consumption of cereals, which, in rural areas, are by far
the principal source of calories, protein, calcium, iron,
and water soluble vitamins, and the second source of fats.
In addition, the rapid replacement of traditional durum
wheat by soft wheat (poorer in protein and industrially
milled with a high extraction rate) accelerates the loss of
nutrients even further. In the case of iron, however, it is
probable that urban intake, although lower in quantity,
is definitely more easily assimilated because of its higher
content of iron of animal origin.
Conversely, animal protein intake increases with
urbanization through the consumption of meat, fish,
milk, and eggs. Fat intake is also on the increase because
of a higher consumption of imported vegetable oils and
some animal fats. The intake of vitamin A is increasing
because of a higher consumption of leafy vegetables,
peppers (harissa), and, above all, offal, eggs, and milk
products. With respect to vitamin C, the gap is noticeable
between rural populations and the inhabitants of smaller
cities because of the higher consumption of vegetables
and citrus fruits in the latter.
In the case of calcium and riboflavin, and also thiamine
and niacin, the effects of urbanization are generally
positive, although not quite as markedly as for other
nutrients because they are the result of two contradictory
trends. On the one hand the drop in cereal consumption
lowers the levels of these elements significantly, and on
the other hand they are raised by higher consumption of
milk (calcium), eggs, milk, and offal (riboflavin), and meat
one trend over the other. Thus, in 1975 deficiencies in
calcium and riboflavin were most marked in the smaller
urban areas. The positive changes in nutrient content
observed in 1975 in relation to urbanization occurred to
a larger extent in 1980 and 1985.
Consumption and income
The budget servey makes it possible to determine total
cash expenditures and cash expenditures on food per
capita per year, to which should be added the value of
home-produced, home-consumed commodities, and
gifts at current market prices. The total expenditure may
be taken as an approximation of income, as income per se
is not determined. Table 4 and figure 1 show the changes
in total and food expenditures in diners per capita per
year from 1965 to 1985.
Over the past 20 years, income (in current dinars) has
increased more than 51/2 times in rural areas and almost
six times in the larger cities. While the gap between the
two sectors has not noticeably widened, food expenditure
in the major centres, which alone was equal to the overall
per capita income in the rural areas in 1965-1968, in
1985 was still very close to that in the smaller cities. As
in the case of the consumption levels shown in table 2,
the smallercity dwellers were in an intermediate position
from the point of view of income.
In 1965-1968 the average per capita income in the
small and medium-sized cities was barely higher than
in the rural areas. In 1975, however, those same urban
areas were in a particularly unfavourable position with
regard to intake of animal products (noticeably milk
products) in spite of improved and higher incomes
(table 2). This can be explained by the fact that, thanks
to home production and consumption in the rural areas,
and to the use of reconstituted milk in the major centres,
milk consumption was higher than in the smaller urban
4
areas, which probably experienced supply difficulties. In
nutritional terms, this results in a drop in calcium and
riboflavin intake (table 3).
This situation, which could have had serious
consequences for the vulnerable population, was
entirely resolved by 1980. This is explained by the fact
that during the oil boom of 1975-1980 a large number
of agricultural development projects were undertaken,
creating jobs in the rural areas, but this seems mainly to
have benefited the population of the smaller cities, where
landowners and traders are concentrated. Moreover,
on the suggestion of the World Bank, the government
decided in 1977 to liberalize public transportation and
to encourage its growth through investments in road
development and maintenance, which contributed to
better market organization and generated income.
Income having grown (in current dinars) at a higher
rate in the smaller urban areas than in the other sectors,
the income levels there became noticeably closer to those
in the major cities (figure 1). Food expenditure went up
in the same proportion; as a result the consumption
of meat increased by 20%, that of eggs and fresh milk
doubled, and that of poultry and yoghurt trebled (table
FIG. 1. Changes in total expenditures and food expenditures in relation to urbanization
Trends in urban and rural food consumption
2). This is illustrated in table 3 by an increase in intake of
riboflavin and calcium, as well as vitamin A, the intake of
which in retinol equivalent exceeds the levels recorded in
the major cities.
Between 1980 and 1985 the rate of household income
fell very slightly in the smaller urban areas, remained
unchanged in relation to the previous period in the major
centres, and went up sharply in the rural areas (figure
1). It should be noted that 1985 was a good crop year in
most agricultural areas. The increase in cash expenditure
was reflected in a sharp increase in the consumption of
bread and blended oils, prestige commodities for people
with modest budgets. No parallel increase was observed
in the consumption of animal products - status symbols
too, but far more expensive ones. Conversely, the
consumption of poultry, which had doubled during the
previous prioed, fell; that of fish, which was already low,
decreased still further, as was also the case for milk (table
2). (The number of households surveyed in April 1985
was unusually low. Because the supply and consumption
of home- produced milk usually reaches one of its peaks
at that time of year, it is likely that the average level may
have been underestimated, but this alone would not
seem to account for the overall drop in consumption.)
J. Périssé et al.
TABLE 4. Changes in total expenditures and food
expenditures (dinars per person per year) in relation to
urbanization
Total expenditure Food expenditure
R U C R U C
1965-1968 53 66 125 30 34 52
1975 104 148 208 50 62 85
1980 157 284 392 77 117 140
1985 294 501 748 134 196 258
The position was roughly the same in urban areas.
Intake of oils and sugar increased, but that of poultry,
meat, fish, and eggs dropped. Examination of the data
by region shows that nutritional levels deteriorated in
the north-eastern and particularly in the southern areas,
whereas they remained stable or even improved in the
other areas in comparison with 1980. The closing of the
Libyan border in 1984, followed in 1985 by the return
of expatriated workers, very probably accounts for this
regional degradation (less remittances from workers
abroad, less border trade, more workless members in
households).
Let us consider whether inequalities in the distribution
of income within each population group changed over the
time of the surveys. (A presentation of the data by deciles
of population could help, but these are not available yet.)
The 1975 survey showed that 33% of the rural population,
18% of the population in the smaller cities, and only 3%
in the larger cities fell in the lowest income classes (below
D 60 per year). Four per cent of the rural population,
9% in the smaller cities, and 21 % of the inhabitants of
the major centres were in the highest two income classes
(over D 300 per year). The limits of the income categories
were readjusted upward in successive surveys to take
into account both depreciation of-the dinar and the rise
in income resulting from economic development, but
the arbitrariness of the choice of those limits precludes
any valid comparison of income distribution patterns.
To find out whether income disparities increased or
decreased, Gini’s coefficient, which indicates the degree
of concentration (or dispersion) of overall income in the
various population groups, was used (table 5). All three
surveys showed that disparities in income distribution
increased with the degree of urbanization: incomes were
more evenly distributed in rural areas than in the cities.
But no clear trend emerged from one survey to the next,
save that Gini’s coefficient remained stable for the whole
of Tunisia between 1975 and 1985.
The level of food intake and per capita income are
related and depend on the age and sex of household
5
members. These in turn influence the level of per capita
income. The average number of persons per household is
highest in populations at the bottom of the income scale
and decreases steadily as per capita income increases.
Statistically, large families have a high probability of
having a low income, which is not surprising. When
we divide total income by the number of members,
households with few active individuals and many young
children are most likely to have a low total income and,
consequently, a modest per capita income. High-income
households are those with more working members, and
per capita income is proportionately higher as there are
fewer children and dependents.
Given equal income, the number of individuals per
household is slightly higher in the large cities than in
urban or rural areas, suggesting that the dependency
coefficient increases with population density.
Income, energy requirements, and energy
intake
Households in the same per capita income group are,
to some extent, similar with respect to age of members,
sex distribution, and anthropornetric features. This
affects the calculation of energy requirements. Indeed,
the 1975 data showed that the average per capita energy
requirements increased as household income rose [1).
The same pattern was shown, although less distinctly, by
the results of the 1980 and 1985 surveys.
For all urban and rural groups, the average per capita
energy intake grows faster than requirements when the
household income increases. In both affluent and poor
households, the gap is always wider in rural than in
urban areas. Similarly, when a deficit occurs, it is lower
in rural groups.
If one compares minimum energy requirements and
intake for each household in each income category of the
three strata, one can identify the families that are at risk
of suffering from an energy deficiency. The probability
of undernourishment is higher for urban families than
for rural ones. This analysis, already partly published
[1], is based on the results of the 1975 survey (table 6).
A comparison of household intake and requirement
distribution on the basis of the 1980 and 1985 surveys
gives the same results. That is, the percentage of
households suffering from energy deficiency is higher
in the two urban groups than in the rural ones. This
normally should be reflected in the physiological
condition of urban con- sumers. We will see later
whether this assumption is confirmed by facts.
6 Trends in urban and rural food consumption

TABLE 5. Gini’s coefficient values in relation to income distribution

R U C Total
1985 0.37 0.40 0.42 0.43
1980 0.37 0.37 0.40 0.43
1975 0.39 0.40 0.43 0.44
TABLE 6. Food expenditure, energy intake, and cost of calories by per capita income level in relation to urbanization, 1975

Income(diners/ Food expenditure Energy intake Cost of energy

persons year) (dinars/person/year) (kcal/person/day) (millimesa/1,000 kcal)
R U C R U C R U C
>400 155.2 132.5 162.2 3,126 2,899 2,469 136 125 180
300-400 118.7 120.6 112.6 3,013 2,883 2,518 108 115 122
200-300 106.5 92.4 95.4 3,065 2,569 2,219 95 99 118
160-200 77.2 76.7 83.2 2,837 2,381 2,218 75 88 103
120-160 66.8 63.9 61.1 2,702 2,302 2,223 68 76 75
100-120 52.5 55.4 56.3 2,569 2,217 1,913 56 69 81
80- 100 47.2 48.2 48.0 2,437 2,200 1,854 53 60 71
60-80 39.7 38.4 38.2 2,431 1,981 1,829 45 53 57
30-60 27.6 28.7 29.0 2,168 1,809 1,429 35 44 56
<30 14.1 14.4 11.9 1,920 1,415 1,077 20 28 30
Totals 49.9 60.2 76.6 2,474 2,228 2,122 55 74 99
a. 1 millime = 0.001 dinar.

TABLE 7. Cost of energy by food group and income level, in relation to urbanization, 1975

Income Millimes per 1,000 kcal

(dinars/ Cereals Sugar Oils and fats Animal Total diet
Person/ year) products
U &C R U &C R U &C R U &C R U &C R
>400 45.7 50.0 82.5 95.2 85.5 94.3 360 352 157 136
300-400 37.5 39.6 73.8 83.0 68.7 77.8 309 319 120 107
200-300 37.3 29.7 80.3 61.2 56.9 67.8 310 338 110 95.2
160-200 34.1 25.5 77.3 63.3 48.1 52.0 334 334 95.2 74.6
120-160 28.7 25.9 58.1 63.1 43.9 51.9 328 296 75.7 67.7
100-120 30.3 19.6 58.1 63.9 44.3 49.9 304 265 74.7 56.0
80-100 27.9 20.0 55.3 55.6 36.3 44.0 294 246 65.5 53.1
60-80 24.1 16.3 54.7 50.7 29.8 41.1 303 240 54.8 44.7
30-60 23.4 13.9 43.0 46.0 26.3 33.7 279 209 45.7 34.9
<30 14.6 9.3 51.0 39.4 19.5 23.8 269 175 27.9 20.1
Totals 31.0 19.9 64.2 56.1 48.1 47.2 309 271 85.7 55.2
1 millime = 0.001 diner.

Contrary to expectations, the budget data for those
same families showed that, for the same income level,
the proportion of income spent on food was virtually
the same in rural and urban areas. Yet expenditure on
goods and services other than food was greater in urban
than in rural environments, and one might expect rural
populations to use a larger part of their income to enjoy
a qualitatively better diet like their urban counterparts. It
would seem, however, that such is not the case.
J. Périssé et al.
As is shown in table 7, the cost of 1,000 calories
in each food group increases with the income of
the consumers. Cereals, sugar products, and animal
products are cheaper in rural areas; only oils and fats are
cheaper for the urbanized populations because of the
greater use of subsidized blended oil, Given the same
income, rural dwellers could try to diversify their diet
like their urban counterparts; however, they prefer a
FIG. 2. Composition of diets by food groups ( Percentage of total energy intake per person), and cost of calories, by income
level in urban areas (U and C), 1975
The cost of those 1,000 calories increases very quickly
with changes in consumption patterns from one income
category to the next, but the difference between urban
and rural areas remains unchanged all along the income
scale. One may therefore consider that for the same
income, urban life, with its better market supply, offers
greater possibilities for diet diversification because it
does not demand the same degree of energy expenditure.
For the same amount spent on food, the savings realized
in expenditure can be used by urban dwellers to diversify
their diet and increase its palatability. This, in our view, is
a specific effect of urbanization.
7
more plentiful and less inclusive diet (more cereals but
less oil, fewer animal products, and fewer vegetables
and fruits) that meets their energy requirements, which
is essential to their continuing productive activities. In
other words, for the same income, whether high or low,
the cost of 1,000 calories is lower in the rural areas, and
for the same amount of money rural dwellers can eat
more (figs. 2 and 3).
This fact, observed under conditions of constant
income levels, cannot be accounted for entirely by
economic considerations. For a large part, physiological
needs constitute the reason rural dwellers find themselves
compelled to adopt less expensive consumption patterns
for the same income than urban dwellers. To summarize,
changes in diet are imposed by income (in the present
case, the amount spent on food), and energy intake is
dictated by the degree of urbanization.
8
FIG. 3. Composition of diets by food groups, and cost of calories, by income level in rural areas, 1975
Figures 2 and 3 show that, for very low income levels
in both rural and urban areas, consumers have little if
any scope for changing consumption patterns to lower
the cost of their diet further. They cannot significantly
reduce their consumption of expensive calories (meat,
fish, milk) and fruit, because they are already eating
minimum amounts of these items compatible with a
palatable diet. They are therefore extremely vulnerable
to increases in the price of cereal products. A limited
increase may create a temporary inconvenience for
middle-class households, who can adjust their budgets by
cutting down on less essential items without any serious
effect on their nutrition. The same increase becomes a
nightmare for the poorer population, not only because
their budgets are weaker but also because it affects a
larger proportion of their total diet, and there is danger
of its causing energy and nutrient deficiencies.
TABLE 8. Minimum acceptable levels of income and
percentage of the population below those levels
Poverty line Population at
(diners/person/year) risk (%)
U&C R households-at risk.
1975 87 43 22
1980 120 60 13
1985 190 95 8
Trends in urban and rural food consumption
This leads us to the question of food subsidies. Table
8 gives estimates of the percentage of the population
too poor to meet their minimum energy needs, based
on the poverty-line method suggested by the World
Bank and data from the three surveys [2]. No critical
analysis of the method and the appropriateness of the
income level selected as the threshold point will be
attempted here. The decision made by the government
to help the disadvantaged population with incomes
below the selected threshold seems important enough
to be reported. In 1986 the government created a
Solidarity Fund to distribute a fixed amount of money
by household. The National Institute of Statistics (NIS)
was asked to determine from the survey data and
the socio-professional profile of the heads of families
the number of families in each administrative region
who were statistically at risk (profession, family size).
The government requested the provincial governors
to work with local authorities to identify and list the
critically poor. About 120,000 households had been
listed locally when the NIS calculated that there were
230,000 rural and 325,000 urban persons-about 100,000
J. Périssé et al.
Subsidies
A subsidy system was instituted in 1970 to stabilize the
prices of staple commodities and to keep the cost of
certain items down. Those subsidies, which rose from
D 1.3 million in 1970 to D 262 million in 1985, have
become a burden on the national budget. The General
Equalization Fund (Caisse générale de compensation)
intervenes at the production level by subsidizing the
prices of some items {fertilizers, fuel, animal feeds), and
at the consumption level by making up for the difference
between the cost price and the selling price of staple
goods (semolina, flour, bread, blended oils, milk and
milk products, sugar, tea, coffee), which are then sold at
fixed prices.
The results of the 1985 survey were used as a basis to
assess the redistribution effect of subsidies on household
budgets 13, 4]. Properly used and focused on the needy,
such subsidies ought to apply first and foremost to those
commodities whose consumption is reduced as income
increases. Thus, the consumption of barley, durum
wheat, and wheat flour for making bread decreases with
urbanization at the same time that household incomes
rise. Barley, which is essentially home grown and home
consumed, is only a negligible component of the diet.
The consumption of flour for bread making is on the
wane, but this is because commercially baked bread is
increasingly popular. Logically, subsidies should apply
only to processed durum wheat (semolina, noodles). As
income rises, the consumption of the other subsidized
commodities increases moderately (sugar, wheat), fairly
sharply (blended oils), or sharply (milk, eggs, meat},
with the result that subsidies on these commodities are
of greater benefit to the urban than the rural population.
The 1985 survey showed that wheat consumption
increased at the same rate as income for all population
groups, even in the larger centres. A saturation point was
reached only in the second- and third highest income
categories (D 350-500 and D 500-800), and consumption
decreased slightly for the highest income category only
(over D 800). It is therefore unavoidable that the richer
populations derive far greater benefit from this policy
than the poorer ones.
In 1980 “the least-favoured fourth of the population
only account[ed] for 12.5% of the total subsidies,
whereas the uppermost fourth enjoy[ed], three times that
amount, i.e., 37.3%” 141. In 1985 the least favoured one-
fourth received 14%, the most affluent one-fourth 35%.
“Between 1980 and 1985, the rate of subsidies increased
by 26% while annual private consumption only went up
by 1396 in current dinars” [3]. The cost to the national
budget, however, is higher than the social transfer thus
9
achieved, as subsidies are of greater benefit to urban than
to rural populations, and to middle- and upper-income
groups than to poorer ones.
Subsidies represent a not inconsiderable part of the
total budget of the poor: about 7%. If this bonus were
eliminated, it would lead to a shortage of about 150
calories per day for each individual concerned. At the
other end of the income scale, subsidies involve an
inadmissible waste of certain commodities such as bread.
Any sudden decision implying a return to more realistic
price levels could very well result in the kind of explosive
situation among urban populations that Tunisia and
other countries have already experienced, unless it were
supported by compensatory measures of public aid to
the target groups. To avail themselves of those subsidies,
the rural populations have given up a thousand-year-old
system of home production and home consumption in
the course of a decade (table 9).
TABLE 9. The importance of home-produced, home-
consumed durum wheat (kilograms per person per Year) in
rural areas
Bought from Home-produced, Total
market home-consumed
1975 67.8 61.2 129
1980 101.9 24.1 126
1985 121.7 4.3 126
This dramatic change has benefited the rural
population and partially contributed to the rapid
progression of the cereal subsidies (D 20 million, 34%
of the total, in 1975; D 158 million, 60% of the total, in
1985). Nevertheless, one of the brighter aspects of this
process so costly to the state has been the integration of
the rural population into a market economy. With less
food stored at home, however, they will, like the urban
populations, react more sharply to measures that might
be taken toward a return to more realistic price levels in
an effort to set the Equalization Fund on its feet again.
Body measurements
The 1980 and 1985 surveys recorded data on the weight,
height, and age of consumers to provide a means of
evaluating energy and protein requirements. These data
were collected from samples of 8,520 males and 8,890
females in 1980, and 7,940 males and 8,100 females in
1985, classified according to place of residence as rural
(R) or urban (U and C). So far, only the median height
and weight data according to age and sex are available
for 1980; they have been compared with those for 1985.
To the best of our knowledge, this is the first time it has
been possible both to study household consumption
10
levels and the body measurements of the individuals
concerned simultaneously and to follow their evolution
over a period of five years on the basis of a national
sample. The anthropometrical study is not a longitudinal
one, however.
In 1980 the average weights of males up to five years
old and females up to eight years old were not different
for rural and urban groups. The two curves then spread
out slightly but steadily for adolescents in favour of the
urban group and separated more distinctly for adults
(fig. 4). The maximum difference in weight is observed
in females 45 years old.
FIG. 4. Differences in body weights between urban and rural populations, by sex, 1980 (provisional)
Trends in urban and rural food consumption
The more favourable position of the urban population
contradicts what would be expected on the basis of the
consumption data, which consistently show a lower
energy intake for this group. It is quite unlikely that
urban consumption was underestimated and/or rural
consumption overestimated at all income levels. The
only explanation is that, in the absence of specific data on
time budgets, the energy requirements of the more active
occupations were underestimated. This underestimation
would lead to a mistaken conclusion that energy intake
levels are more satisfactory in the rural areas. It seems, in
fact, that the urban diet, which is lower in calories but far
more diversified and richer in nutrients, results in better
growth of body weight.
J. Périssé et al. 11

FIG. 5. Changes in male body weights between 1980 and 1985 (provisional)

FIG. 6. Changes in female body weights between 1980 and 1985 (provisional)
12
In the period between the two surveys, the weight
curves of the population groups changed (figs. 5 and 6).
By 1985 the weight of young rural males and females up
to 11 years old had caught up with, and even exceeded,
that of their 1980 urban counterparts, coinciding with
the 1985 curve of urban dwellers. The latter gained 2
to 3 kg at each age level between the ages of 4 and 11.
Rural adolescent males (between 12 and 18 years old)
and females (between 12 and 16 years old) were heavier
than those surveyed in 1980, and were on the 1980
weight curve of their urban counterparts. They did not,
however, catch up with the latter, who during the same
period gained 2 kg (males) and 4 kg (females) at each age
level. Adult males, both rural and urban, lost 2 kg at age
25 in relation to those of 1980, regained them at age 30,
and at age 45 were on the 1980 curve.
They gained additional weight beyond age 55 in
relation to their elders, but the difference between urban
and rural populations persisted.
Between the ages of 20 and 80 years, rural females
kept just above ( + 1 kg) their 1980 curve. Among all
groups, they were the only ones not to have increased in
weight. On the other hand, urban women from age 45,
gained 2, then 4 kg in relation to those surveyed in 1980.
In five years, except for males between 20 and 45
years of age, all age and sex groups gained from 1 to 4
kg. Unless an analysis of weight distribution for each
age group shows a shartp increase in the numbers of
overweight people, it may be assumed for the time being
that the body growth patterns of all three groups changed
for the better in Tunisia. The change was greatest in the
urban population.
The contradiction noted between consumption data
and the body-weight curve of the 1980 survey reappeared
in 1985. Indeed, food intake did not change significantly
between the surveys. The higher weight increase is
attributable to the urban diet, which is poorer in energy,
more diversified in nutrients, and richer in fats and high-
quality proteins. These considerations fail to explain why
the rural groups went beyond the 1980 median curve
on a diet that, except for retinol, was slightly less rich in
nutrients. Obviously, the 1985 survey is only one point
on the trend and cannot provide all the information on
what was happening during the period under review.
Conclusions
Technical considerations
The survey on household consumption and budgets is a
very useful tool for the decision-makers responsible for
food supply and nutritional planning. It is effective for
several reasons.
Trends in urban and rural food consumption
First, appropriate data are more convincing than
qualitative judgements. This was so well understood by
the Tunisian government that in 1975 the Council of
Ministers decided to repeat the survey every five years.
The results are consistent and comprehensive, and are
sufficiently accurate to permit an assessment of the
changes that occurred during that period. Such surveys
have shown that it is possible to evaluate ex post facto
the size of the population below the poverty line and to
define their socio-professional profile. They have been
used to quantify the economic and social impact of
food subsidy policies, and the nutritional and economic
effects on the diet of regional development projects.
Furthermore, survey results enabled the Service of
Agricultural Planning to project a model of food demand
divided into two distinct types of diets, urban and rural,
with specific income elasticities for each. This model is
far better and more sensitive than the previous one based
on food balance sheets.
Second, the data on heights and body weights cast
new light on the results of the household-consumption
survey. On the condition that such data are not limited
to children between birth and six years of age, as is
now the case in many countries, weight and height
distribution patterns may provide inexpensive and
objective indicators of changes in the food situation.
Such surveys are inexpensive, provided they are not
conducted independently but are combined with food-
consumption and budget surveys.
Finally, the use of energy-requirement data as a
basis for qualitative assessment of nutritional intake
leads to inconsistencies in the results and does not give
researchers the expected outcomes. More sophisticated
algorithms using variables depicting the profile of
each consumer, household by household, need to be
applied. In this way the variability of requirements
could be increased and brought closer to actual living
conditions. The variables now available (age, height, and
weight) in relation to basal metabolism are adequate for
evaluating maintenance requirements, as suggested in
the 1985 report.
Food policies
Paraphrasing a famous remark, one might say that
urbanization is far too serious to be entrusted to
urbanists. In fact, the future of the cities is in the hands of
the rural dwellers. Urban over-population is the result of
their feeling of frustration. Because small and medium-
sized towns constitute an intermediate stage in the
process of rural exodus, it seems appropriate to maintain
the distinction between their populations and those of
large cities in future surveys. It is on those intermediate-
J. Périssé et al.
stage cities that effort should be focused in order to stem
migration toward the larger centres. Results of the 1980
survey showed that the measures taken to liberalize
transportation services and develop exchanges in and
with the rural areas had a favourable impact on the
diet of these medium-sized towns. This policy should
be encouraged.
Possible alternatives to a policy of food subsidies are
discussed elsewhere. Any policy, however, that accelerates
the trend toward the substitution of soft wheat for durum
wheat would run counter to the nutritional interests of
the poor and of the rural producers. If this substitution
proves to be unavoidable and is justified by economic
exigencies, it should be accompanied by measures to
revise the extraction rate of industrial cereal products.
Data on physical activity are dramatically missing.
New surveys of the Tunisian type will need to include a
component on the time budgets of household members if
the complexity of the various family conditions and their
working capacities are to be reflected with any realism,
and if hasty and unfounded conclusions are to be avoided
in comparing requirements with consumption data. The
experience gained with surveys in Côte d’Ivoire and
Rwanda is very useful in this respect. Such data could
be used to improve our understanding of whether some
specific age and sex groups have additional working
capacities before new development projects are proposed.
Instead of subsidizing all types of bread, it might be
possible to enforce the production and sale at a controlled
price of another type of bread that alone would be taken
into account for the calculation of the cost-of-living index,
as is done in Italy for bread rolls (ciriolo). In Tunisia,
such a bread might be made from subsidized flour. To
preclude this flour’s being mixed with bleached flour
and used for other purposes, its extraction rate would be
lower; it would then be less white and therefore richer in
proteins, vitamins, and minerals. The end product would
be a typically shaped loaf, but less attractive because of its
brownish colour. Affluent households would neglect this
humble loaf for the baguette, made from whiter flour and
therefore more enjoyable, but more expensive because
it is non-subsidized. In five years, a new survey should
make it possible to determine the degree of acceptance
of this brown bread in the various population groups,
as well as the income categories of persons by whom
it is consumed.
If some animal products have to be deleted from the
list of subsidized commodities, milk products [except
butter) should be maintained. Eggs can be bought
one by one and therefore are accessible to persons
on a small budget. Milk and acidified milk products
13
(easier to store) are essential for the lowest-income
groups, who cannot afford to buy meat. At the lower
end of the income distribution it is not uncommon to
find households with many children, often headed by
widows or divorced women with jobless or handicapped
dependents. Such individuals are nutritionally at risk.
Subsidized commodities provide help, but a food-stamp
programme would be far more efficient. The recent effort
made by local authorities to identify those living below
the poverty level is a good start in that direction.
One of the good features of the Mediterranean diet is
its moderate fat content from good-quality oil. The calorie
content of the blended oil, however, is very close to that
of cereal products, so that its use results in high calorie
intake in the urban low-income groups. It is not desirable
to encourage consumption of this oil beyond the present
limits, lest it should lead, as in Malta and probably also
in Libya, and for the same reasons, to a proliferation of
nutritional overweight with a concomitant upsurge of
cardiovascular diseases.
Each family has its own unique characteristics and acts
as best it can to reconcile two contradictory requirements:
securing a maximum calorie intake at a minimum cost,
while diversifying the diet to enhance its palatability
within the limits of available income. In doing so, the
people depend on the instinct for survival rather than
a knowledge of physiology; solidarity takes the place of
economic accounting. For the time being, Tunisians are
doing well, but the patterns of consumption of urban
dwellers are accelerating behavioural changes, and such
changes are exposing these persons to nutritional risks.
This has to be watched. Tunis has proved here to be well
equipped for that purpose.
References
1. Périssé J. Kamoun A. The price of satiety: a study of
household consumption and budgets in Tunisia. FAD Food
and Nutrition 1981;7(2).
2. Fourati H. Identification des populations vulnérables:
méthodologie et étude de cas de la Tunisie. In: Les
consommation et les politiques alimentaires au Maghreb.
Séminaire international FAO-IMA. Tunis, 1986.
3. Kamoun A. L’incidence de la Caisse générale de
compensation sur la distribution des revenue et
l’équilibre de la ration alimentaire. In: Les consummation
et les politiques alimentaires au Meghreb. Séminaire
international FAO-IAM. Tunis, 1986.
4. Belhaouane K. Compensation, prix et revenue en Tunisie.
In: Les consummation et les politiques alimentaires au
Maghreb. Séminaire international FAO-IAM. Tunis, 1986.