A lot has been said and written about the great Indian middle class. The sheer numbers of people who can be so categorised promises a large and growing market for many products and services for years to come. This enormous domestic source of demand is a major reason why so many people believe that India can sustain its current growth performance for years to come, despite anything the government may or may not do.

In order to characterise and predict the behaviour of a supposedly homogeneous middle class, it is important to understand how it emerged. Sociologists and historians presumably have a lot to say on this, but from an economics perspective, the issue of origins seems to have been taken for granted.

The use of income as a defining criterion, often exclusively, has so far dominated the economic and business approaches to the middle-class phenomenon.

This may be all right as it goes and may even be enough to provide reasonable forecasts of the spending and saving behaviour of the class. But, what if origins matter?

Are spending and saving patterns dependent on how a particular household achieved middle-class status? Therefore, from a business perspective, are we in danger of increasing our forecast errors if we do not take origins into account?

First, let's explore the origins of the contemporary middle class and how they may be changing. They lie predominantly in the public sector. A government position, with its permanent employment contract, a guaranteed pension and health care and a host of other fringe benefits (including low-interest housing loans long before the housing finance industry entered the scene) provided exactly the kind of security and predictability that generates and shapes the behaviour and attitudes of a middle class.

It is important to emphasise that this process was not confined to the higher-echelon government services or public enterprise management cadres.

In fact, the great strength of the public sector as a force of socio-economic change lay in its provision of this security blanket to a very large number of people in the lower echelons of the hierarchy. The lifetime income and health security of parents were fundamental to the next generation's staying in school and perhaps going on to earn higher qualifications.

While a college degree or its equivalent is neither a necessary nor a sufficient condition for entry into the middle class even in India, it would be difficult to argue that it doesn't help a great deal. Professional qualifications have unquestionably been a great force of social mobility and equalisation over these decades.

Employment in the organised private sector would also have provided a similar, though not as comprehensive, pathway for second-generation entry into the middle class, but, in terms of numbers, it was always small relative to the public sector.

But, when we look at changing origins, it is the balance between the two that will be critical to sustainability.

From the mid 1990s onwards, the public sector has seen an absolute decline in the number of people employed. Apart from the central and state governments, public enterprises and banks have significantly reduced their annual intake of people at all levels.

Since the lower echelons make up the large majority of public sector employees, the largest decline in the number of opportunities would, logically, have occurred at these levels. The implication is that a significant instrument of socio-economic mobility is weakening and certainly cannot be expected to continue the same role in the future.

Is anything substituting for it? If the numbers are anything to go by, blue-collar private sector employment, which could have been expected to play the same role, has not grown very significantly, either. Growth in private sector employment has certainly provided an alternative outlet to the large cohorts whose parents worked for the public sector in a wide variety of capacities.

The second-generation dividends are clearly being earned. However, it does not appear that any significant substitute channel has emerged as far as the first generation is concerned.

This does not mean that the process has come to an end; there would be disastrous socio-political consequences if it had. What is happening, instead, is there is a variety of channels of socio-economic mobility emerging, none as large and as streamlined as public sector employment was in the past but cumulatively providing enough momentum to sustain the process.

But, the diversity of these channels, combined with the absence of a single dominant process, suggests that the emerging middle class is going to be a lot more heterogeneous in its behaviour than the current one has been. Large numbers of people at various income levels will not necessarily translate into homogeneous spending, saving and other economic decisions.

Extrapolations based on behaviour observed in the recent past, driven as it was by an essentially homogenising process, may prove to be quite misleading.

We need to begin exploring the sources of diversity and their economic implications. If the private sector is going to play a predominant role in socio-economic mobility, the nature of the employment contract is going to be fundamentally different.

A lot of the non-pecuniary benefits of public sector employment will be substituted for with cash. Most significantly, the security of permanence will make way for the risks of unemployment associated with both business cycles and individual non-performance.

The implication of the first difference is that the private sector will have to provide many of the safety net services that the public sector used to. But, unlike the "one size fits all" approach that the latter could afford to take, the former will have to take account of all the heterogeneity in tastes, attitudes towards risk and willingness to pay that it will face. The implication of the second is somewhat more difficult to predict.

Another source of diversity is the increasing importance of the entrepreneurial as opposed to the salaried channel of entry. Liberalisation has opened the doors for anybody to start up a business, which has been taken advantage of most visibly in the services sector. This seems to work both ways; entrepreneurship as entry into the middle class and the middle class as a source of entrepreneurship.

To sum up, the growth of the middle class is going to be fundamental to sustaining economic performance. But, to ensure that the virtuous circle relationship between the two variables, we need to recognise and respond to the changing nature of the class itself.

The author is chief economist, Crisil. The views here are personal.