The matter is further complicated by the possibility that, while the service itself may not gain in productivity terms over time, it may contribute to productivity gains in other sectors either immediately or - what is even more complicated to measure - over time. Thus, there is now near-universal acceptance of the idea that the provision of good health care and education improves the quality of the labour force in an economy. Similarly, it is accepted that certain services operate to change both the nature of production (in terms of its organisation and management) and the division of labour, which have both micro and macro effects on productivity in both goods producing and service producing sectors.
 
The consumption link between services growth and overall economic growth can be equally problematic. The idea that the income elasticity of demand for services is greater than the income elasticity of demand for goods emerges from Engel's Law. However, this basically suggests that such income elasticity of demand will be greater than one, which does not tell us much about the actual volumes. Empirical studies of industrial economies in the postwar period have not found substantive evidence to support the stronger proposition.
 
What has typically been found is that, while there is a trend of increasing expenditure on housing, health, education and leisure in general, within each such category, the purchase of goods has exceeded that of services. Even in pure services (such as in communication) the providers tend to use more and more goods - i.e., they are heavily capital-intensive in the production of their service.
 
All this makes the relationship between economic growth and services rather tenuous. One clear link that does seem to be substantiated by the empirical evidence is that of the association of services growth with the move from non-marketed to marketed activities. In other words, as the unpaid labour undertaken within households declines in favour of paid labour, there is an increase in the demand for certain kinds of services.
 
But then, it is no longer possible to make a clear association between a higher share for services in national income and higher economic growth. The slowdown in growth in the industrial economies has had no effect in terms of reducing the share of services. Similarly, in industrial countries fears are now being expressed that the very nature of much service sector work allows for downgrading or wages and work conditions, so that during periods of slow growth or even in some kinds of economic boom, the growth of a service economy can actually reinforce the emergence of a dual economy. This is especially evident in the US economy, for example.
 
For developing countries, the picture is even more muddy because quite often the problems of valuation tend to make the actual extent of services growth even more difficult to ascertain, while existing dualism in turn makes the services sector an easy "residual sector" for those unable to find productive employment opportunities in the goods-producing sectors. So growth of the services sector here, even more than in the industrial countries, need not be an indicator of healthy or buoyant overall economic growth. With this background in mind, we can turn to a consideration of the recent behaviour of the services sector in India.