The United Progressive Alliance government has drawn a correct inference from the election results of last May relating to the employment situation in the country.

Unemployment was consistently showing up as a major concern in the pre-poll surveys and could well have been an important reason for the National Democratic Alliance's defeat.

Given the large number of people entering the workforce each year, the persistence of the gap between supply and demand will become an even more significant political factor when the next elections come around.

The last six months have seen the government take two initiatives to address the issue. The idea of providing an employment guarantee has now been concretised in the National Employment Guarantee Bill, which has gone through the Cabinet and awaits introduction in Parliament.

The idea of job reservation in the private sector is still being debated; the government has indicated that it will not legislate on this issue but is looking for meaningful voluntary participation from employers.

Since the legislative mandate for guaranteeing employment became a reality, much attention has focused on how much it would cost the government to abide by it.

Estimates range from Rs 9,000 crore (Rs 90 billion) per year (government estimates for the current version of the bill, as reported by this newspaper) to Rs 40,000 crore{Rs 400 billion (for the original National Advisory Council version)} to Rs 150,000 crore (Rs 1500 billion), assuming every poor household were to take advantage of the opportunity.

Apart from the potentially crippling fiscal burden (in a situation in which the government is committed to a strict regimen of fiscal discipline), there are concerns about the leakages and diversion of funds away from the intended beneficiaries.

Viewed as a whole, these reservations focus on the difficulties of implementation and the opportunity costs of the funds required. In doing so, they give rise to a fundamental question.

Nobody would challenge the basic premise that it is a government's responsibility to create conditions that maximise employment. But, is it incumbent on the government to be "employer of the last resort", providing jobs to people who have not been able to access other opportunities?

The key to answering this question is to understand the limitations inherent in the notion of "guarantee" or "last resort". As a concept, it is perhaps most often encountered in the financial sector.

The central bank, for instance, is a "lender of the last resort" to the banking system. Or, central and state governments often provide guarantees for loans taken by their enterprises.

The moral hazard problem is intrinsic to both these situations. Banks could well be cavalier in their lending behaviour, knowing that somebody is standing by ready to bail them out of a liquidity crisis.

Government-supported borrowers (and those who lend to them) could well pay less attention than they should be to the implementation and management of the project, increasing the probability of default.

Measures to deal with moral hazard are, therefore, absolutely critical if a last resort arrangement has to have any chance of working.

The moral hazard is inherent in traditional safety net mechanisms, such as unemployment insurance. People who can avail of unemployment benefits simply have less incentive to look for work.

Governments typically deal with this by keeping benefits time-bound; no one can claim payments indefinitely and certainly not without some proof of their looking for a job.

Given the size of the population of potential claimants to unemployment in India and the yet-to-be-fulfilled promise of universal identification, outright cash transfers suffer enormous moral hazard problems.

It makes sense, therefore, to impose the additional conditionality of having to work for the guaranteed income; the physical requirements of the work will deter anybody with the hit of a more attractive alternative.

However, even though guaranteeing employment rather than income appears to be a better approach in Indian conditions, other limitations of the last resort cannot be wished away.

In this context, the viability of the scheme, which can be measured by its fiscal sustainability will depend heavily on the performance of labour markets.

If the levels of employment -- and wages -- generated by market forces are relatively high, the number of people availing of the employment guarantee will be small. It is, then, truly a last resort.

The concerns expressed by several observers about the runaway costs of the scheme stem from the fact that it is being seen not as a complement to a generally dynamic labour market, but as a substitute.

How so? The macro-level picture of employment in India reveals an imbalance that does not suggest a particularly dynamic labour market.

Agriculture contributes less than a quarter of GDP but, according to the National Sample Survey 1999-00 round, is the primary employer of about 60 per cent of the 397-million-strong workforce.

Industry also accounts for slightly less than a quarter of GDP, while employing about 14 per cent of the workforce. Services account for over half of GDP and over a quarter of the workforce.

Over the past two decades, while the share of agriculture in GDP has declined from about 40 per cent to less than a quarter, its share of the workforce has declined by just a few percentage points.

The former is entirely consistent with the process of structural change that any economy goes through, at a greater or lesser speed. The latter reflects rigidities in the labour market, which prevent a similar transformation in the workforce and force a virtually constant share of the population to depend on a rapidly shrinking share of GDP for their livelihoods.

The obvious disconnect between economic structure and employment pattern is the fundamental challenge for employment policy. To attempt to address it with the instrument of an employment guarantee of the kind being contemplated is being, to put it mildly, unrealistic.

The disconnect is a fact. What causes it is a matter for debate. I believe job security regulations in industry are the most important reason. Other people downplay the significance of these and cite other factors.

Whatever they are, the UPA government should quickly converge on an acceptable set of explanations and begin to act on the policy implications of these. Because, if it doesn't, what is intended to be a last resort could well establish itself as the primary instrument of bringing about a change in the structure of employment.

This is perhaps where the greatest moral hazard of the programme lies. Having assured a sceptical electorate that it will guarantee jobs at minimum wages, has the government absolved itself of the responsibility to facilitate their upward mobility to more challenging, safer and better-paying jobs?