Indeed the level of per capita consumption can be taken as a first and approximate indicator of the level of economic development-of a country,
7l". Long-term and world-wide data of per capita cement consumption and
per capita GDP indicate a positive and direct relationship between the two variables. It could furthermore be noted from data that countries with low per capita levels of cement consumption display a higher rate of growth than countries with high per capita levels. The underlying reason for this tendency lies in the fact that as economic development proceeds construction-practices are bound to improves Thus wharfs in timber are replaced by concrete jetties^ timber pole mud-briok walls1 by concrete, or stone or briok walls § compacted earth airport run—ways and roads by concrete or bitumen runways and roads and so on. Economic development brings with it a preference for durable materials among which concrete constitutes a leading role072. It can be shown-'that the elasticity of per capita'cement consump
1/
tion with respect to per capita' GDP could be represented by the equation:
126,' + 0*8 where T = elasticity & X = per capita GDP.
A
1/ See the Building Materials Industry in Africa (H0U/WP/4/Add.2),
SCA, Addis Ababas
1964-e/ch.14/1NE/117
Page 36
The formula represents a hyperbolic curve which tends towards Infinity at "both axes. At low J^ea^oagita GDP levels, the elasticity has a steep slo-pe thus reflecting the high response of per capita cement consumption to a small increase in ^sr^capj^ta^ GDP.? while at higher per capita GDP levels, the slope of the elasticity curve tends to flatten out thus suggesting that per capita consumption "becomes increasingly inelastic.
73. The formula as such is a mathematical approximation of an observed correlation. No doubt, further.improvements and the rationalizing of inconsistencies would still be required to make the formula a reliable tool of projection. Nonetheless it is thought that in its present form, the formula could be acceptable for the limited aim of estimating broad orders of magnitude of future cement consumption.
(ii) Methodology based on the growth of the construction industry 74* Construction ia the single most important component of capital formation of which it accounts for about 5O~7O per cent.— The correla tion between these two variables is therefore a high one. The growth of one is directly related to the growth of the other as is clearly evident from Graph 1 in the annex.
75. The growth of construction in its turn affects the growth of cement consumption directly. The correlation between these two variables is ' again high as can be seen from Graph 2 in the annex. The constant relationship between the rate of growth of construction investments and
the rate of growth of cement consumption, however, seems-to apply mainly to developed countries. It can be shown that while the growth rate of cement consumption with respect to the growth rate, of construction is somewhat constant for countries with high ^gj^capj/fca cement consumption .levels it varies widely for countries with per capita levels of less
than 100 kg.
l/ The Construction Industry in Devslcpsignt Programmes. " A Techno^
Eoonomic Review in the TCas« African Sub-region. ECA? Addis Ababa, 1966.
Page 37
76. The existence of this variation3 therefores invalidates the direct use of Graph 2 for projecting the demand for cement of the countries of West Africa3 where3 it has been noted? current per Gap it a levels are far less than 100 kg. Graph 3 has thus been constructed to take account of the elasticity .of cement consumption with respect to the level of economic development attained as measured by the ratio of GDFCF to GDP, The correlation betweenthe two variables is high and the regression can be represented by the equations- X - -0.172 + 6?6O83X
Where T is the elasticity of cement consumption with respect to construction investments and X is the ratio of GDFCF to GDP attained in the country.
B. Estimate of Demand 1980
77* These two methods will now be used to estimate the demand for cement for the sub-region in 1980. A third method based directly on the estimate of construction activities in 1980 would have been useful in checking the reliability of the two methods. Unfortunately the estimation of the volume of construction activities cannot easily and reliably be undertaken. While the volume of certain types of housing and building can be estimateds on the other hand the estimation of the volume of other construction activities such as damss reservoirs;
sewerage^ irrigation and drainage works3 roads3 bridges etc 3becomes practically impossible.
78. The over-all expenditure on construction activities can be esti
mated because construction being a key economic sectors its .position and behaviour within the economic system can be fairly accurately defined at any given time. But how much can physically be accompli shed with a given investment outlay depends on the technical efficiency of the construction industry,, What volume of work is to be accomplished of each type of activity is settled \>y the priorities set within the given financial resources. What materials and in what proportions these
materials are to 'be used in each project depends on the technical
specifications that are drawn upa
Page 38
79, While from global observations of trends and relationships tools for the projection of cement demand could he established and could be justified for broad orders' of magnitude, a more ambitious method based on the estimation of the volume of construction activities is bound to' get bogged down in a maze of assumptions3 qualifications and explana tions. The method can only "be useful and accurate in circumstances in other words? the method is applicable for short-term estimates of demand and is one which is indispensable to plan annual production levels in a particular country. It would have little validity for long-term projections of demand such as the one at hand.
80. Table 3.1 summarises the procedure of estimating the demand for cement of the sub-region in 198Os using the .two methods established earlier. The two methods give results which differ by just over
3 million tons. Their arithmetic mean is about 10 per cent higher than the lower estimate and 10 per cent less than the higher estimate.
Considering the inherent difficulties involved in long-terra projections
based on meagre data this margin of error could be accepted as fairly
. reasonable. . . . _
TABLE 3-1 .