TNAG-0531-FCO40-626-Application-of-International-Labour-Convention-to-Hong-Kong-1975 — Page 128

106 -

There are some incongruities between the two indicators for social security development. This only recalls the caution mentioned earlier about a country's economic development status. Social security development is a multi-dimensional affair which no single indicator can adequately represent. Table 1 merely suggests that many more factors must be considered in addition to those two indicators before is reasonably sure of different social security development statuses of various countries.

one

2. Social security program index (SSPI)

was

Henry Aaron proposed an ingenious index designed to show the degree of historical experience with social security on the basis of the length of time in number of decades (up to the 1950s) since the first social security legislation enacted in a given country. I It turned out that this was the most powerful explantory variable for inter-country differences in the proportion of GNP devcted tc social security among Aaron's sample of selected countries (mostly developed). Phillips Cutright devised an index of "social insurance program experience" by counting the number of years for which a social security scheme had existed for each of the five major contingencies.2 Cutright arbitrarily started counting at 1934, disregarding differences in social security experience among countries before that date. This was a serious weakness in procedure.3 Benefiting from Aaron's rather gross historical index and Cutright's peculiar mixture of rigour and laxity, Peter Kilby and I experimented with our own "social security program index" arrived at by counting the number of decades (à la Aaron) for each of the five programmes (à la Cutright) since its first social security law was enacted and summing up these numbers over all five programmes. This index was more closely correlated with the proportion of GNP devoted to social security than Aaron's or Cutright's index.

Social Social

The same procedure has been followed to calculate SSPIS for Asian and Oceanic countries for which descriptive data are available in the usual source: Security Programs Throughout the World, 1971, published by the United States Security Administration. Table 2 presents SSPIs together with the proportion of GNP devoted to social security for all countries for which both are available. The table also contains per capita national income for use in the next subsection. The cross-country correlation co-efficient between the two is 0.88, which is significant at the level of one-tenth of 1 per cent (meaning that a relationship between two variables as tight as one indicated by this correlation co-efficient could not have been just any random occurrence but should have been a product of certain systematic forces of inter-relations). This only confirms the earlier findings by many that a social security system does not emerge in a complete form but improves over time after a small beginning is made with respect to a certain contingency. What contingency ignites social security efforts is not random, either. I will return to this question later.

1 Henry Aaron, "Social security: International comparisons", in Studies in the Economics of Income Maintenance, ed. Otto Eckstein (Washington, D.C.: Brookings Institution, 1967).

2 Phillips Cutright, "Political structure, economic development and national social security programs", American Journal of Sociology, vol. 70 (March 1965),

the

3 I would like to recall with thanks that V. Rys enlightened ne greatly about pitfalls of Cutright's method by thoughtful observations expressed in personal correspondences, when the Taira-Kilby article mentioned above was under preparation.

E-1195-2F:5