Themes > Features

16.11.1999

Services Sector Growth : What Does it Mean?

Services constitute a very heterogeneous economic category, and it is one which has become more difficult to define over time. Older definitions of services tended to rest on the fact that services were often difficult, if not impossible, to separate from the service-provider and recipient, so that people became crucial to the definition. More recent definitions have incorporated "business services" which also externalise part of R and D and management functions, and include activities like retailing, banking and insurance, and administration.
 
For Adam Smith, a service was defined in terms of its transient nature, thus they were seen to "perish in the very instant of their performance". The simultaneity of production and consumption in services provision led Smith to argue that they "seldom leave any trace or value behind them". Since this meant that they could neither be stored nor transacted again, they amounted to unproductive activities for most classical economists.
 
Similarly, until very recently the national accounts data of socialist or centrally planed economies tended to make a distinction between material and non-material production. Those services (such as trade and transport) that contributed directly to material production were also classified as such themselves, while other services remained in non-material production. Even such a distinction was not entirely clear-cut, because of the possibility of longer-term effects of some services on material production, but the basic distinction between productive and unproductive economic activities was maintained.
 
Modern sensibility has tended to stand such reasoning on its head, giving tremendous importance to a range of services in terms of their potential for raising productivity in other economic activities. One recent definition treats a service as " a transformation of the user or the user's goods, as a result of the voluntary intervention by the producer of services". [T. P. Hill, 1977] This does not imply an acquisition which is transferable, but rather a modification of the characteristics of the recipient, and it allows for a variety of different interpretations, incorporating both person-based services such as education or health care, and goods-based services such as in transport and R and D.
 
The expansion of the services sector has always been seen as a necessary concomitant of economic growth, to the point where it is often listed as one of the indicators of the economic development process. The theory underlying such an assumption (which was elaborated by economists like Kuznets, Clark and Fisher) was one which viewed development as a three-stage process, with the primary, secondary and then tertiary (or services) sectors becoming more dominant in successive phases. Typically, therefore, a rise in the share of services in national income was viewed as being positively associated with both economic growth and quality of life.
 
At a very broad level, such an assumption would appear to be confirmed by the evidence on the share of services in national income across a range of countries, as presented in Chart 1. With the highest share of services being found in the industrial countries, and the lowest in the least developed countries, the basic argument seem to be quite plausible.

This theory in turn was based on certain assumptions about both production and consumption. First, in terms of production, productivity gains - or changes in value added per worker - were assumed to be faster in material production, especially in industry, than in services. Second, in terms of consumption, it was assumed that the income elasticity of demand for services was greater than that of demand for goods.
 
The assumption of differences across material and non-material sectors in terms of productivity growth was taken as axiomatic for many years. However, it is now increasingly widely recognised that making any such statement is problematic, not only because of the sheer difficulty of measuring productivity in the service sector, but because of the many links between growth and technological change in industry and services.
 
Measuring productivity in any economic activity requires at the very least, some knowledge of the quantities produced. For some services, there are indeed some basic production indicators (the number of haircuts given, the number of cheques processed, the number of telephone calls made) but even these are not always comparable because of issues of quality variation. But for a whole range of other services, even such indicators are not available, and calculations of output are then made on the basis of input volumes, which can help us very little when it comes to productivity measures. Often, there are various kinds of ad hoc assumptions made in estimating output, and the estimates tend to be quite sensitive to the particular assumptions made.
 
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.
 
The Indian Scene
As is evident from Chart 2, since 1980 the growth of the services sector in India has outpaced aggregate GDP growth.This is quite marked from the mid-1980s onwards, and barring only two years, the growth rate has been higher in services than overall for every other year since 1984. As a consequence of this, the share of services in GDP has been constantly increasing over this period, as shown in Chart 3. However, it should be remembered that while 43 per cent (the ratio in 1996-97) is a high ratio, it is still relatively low by international standards, as expressed in Chart 1.

Chart 4 shows the rate of growth of the various services sub-sectors over the 1990s. Not unexpectedly, financial services show the fastest rate of growth, followed rather closely by trade, hotels and restaurants. Community and personal services show the lowest rate of growth within the services sectors, in a pattern not unlike that of aggregate GDP growth.

It is interesting to assess the share of services in the GDP of public and private sectors separately. Chart 5 describes this. Within the public sector, the share of services has been broadly stable at around 60 per cent. This reflects the basic division of government expenditure, which has changed very little as between goods and services over this period, as well as the pattern of public service provision which has also changed relatively little.

However, in the private sector, the share of services has gone up, from around 29 per cent at the start of the 1980s, to as much as 35 per cent or more by the middle of the 1990s. Related to this, there has been a rise and then a fall in the share of services GDP which is accounted for by the public sector, as shown in Chart 6.
 

It is frequently argued that the services sector has become more capital intensive, and that it has in fact done so at a faster rate than the other sectors in the economy. The latter proposition is not borne out by the data presented in Chart 7, which suggests that the share of capital stock in the economy which is accounted for by the services sectors has actually been falling continuously since 1980, albeit relatively gently. It has gone from nearly 50 per cent in 1980 to just under 45 per cent in 1996-97.

In fact, it turns out that the capital-output ratio in the services sector has actually been falling over this period, contrary to the general perception. Chart 8 plots the movement of capital-output ratios in the services sectors, as well as in all other goods producing sectors taken together. In the early 1980s, the services sector tended to be much more capital intensive than other sectors, by a factor of around one and a half. However, over time the capital-output ratio in services has gradually declined - indeed quite substantially so - while that in the other sectors taken together has if anything increased slightly. In services, the capital-output ratio fell from 3.8 in 1980-81 to just 2.7 in 1996-97. Meanwhile, in the other sectors the average ratio increased from 2.2 to 2.5 over the same period. This has brought the two ratios closer together, to the point where there is hardly any difference between the two by 1996-97.

The fact that capital-output ratios in services have tended to fall can be interpreted in a variety of ways. In one sense, this can be interpreted as a more efficient use of the capital in the services sector. Thus, the same infrastructure is presumably being used to meet a greater range and quantity of services provision, whether in transport, banking, public administration or other services. In a capital-scarce economy, this is to be welcomed.
 
However, it can also be seen as a way of accentuating the dualism inherent in the economy, with a proliferation of low-wage low-productivity service jobs leading the growth in this sector, rather than more capital-intensive "modern" service activities. In itself, the information on capital-output ratios does not allow for the choice of any particular interpretation; both are possible. However, if there were information on employment and wages in the services sector, there would be some way of reading more into this data.
 
As it turns out, there are no data on wages in the service sector as a whole. However, we do have information on the share of wages in the organised sector of services. As Chart 9  shows, the organised sector constitutes a large and growing part of total services, increasing its share of total services from 46 per cent in 1989-90 to more than 50 per cent in 1995-96. It is likely that the movement of the capital-output ratio in all services as a group would be mirrored in that of the organised sector of services as well.

The share of wages (or rather, the broader category "compensation of employees") in the organised sector of services is indicated in Chart 10. This turns out to have fallen very substantially over the same period, from 81 per cent in 1989-90 to less than 68 per cent in 1995-96. This in turn suggests that the profit share in the organised services sector must have gone up.

This could have happened in several ways, including through worker competition bidding down wage payments in the service sector. But in a macro sense what this suggests is a process whereby the private organised service sector is taking over many of the activities of both the public sector and the unorganised sector. In the first case, it is able to pay lower wages and thus ensure higher profits for the provision of equivalent services, while in the second case it absorbs some of the low-wage employment that was already in the unorganised sector.
 
This sounds suspiciously close to the scenario of the proliferation of low paid service jobs that was described for some industrial countries above. In countries like India, the process can be intensified by the fact that many private service activities emerge or grow when infrastructure - especially public infrastructure - is poor. Thus, private courier services come up when public postal systems are found wanting, and so on.
 
Of course, the broad macro data described here do not allow for more than a relatively cursory speculation on the possible forces at work, and certainly do not make for a definitive statement on the nature of services sector growth. However, they do mean that the interpretation of such growth should be made with much more caution than is usually the case, and that the possibility of some less than desirable forces at work should be taken seriously.