The finance minister's recent proposal for a National Tax Tribunal to bring about a reduction in the fiscal deficit by freeing locked funds in tax disputes is well-meant.

However, a lot needs to be done at the Central and state government levels to tackle the problem of fiscal deficit.

Recently, a public sector company in the petroleum refining sector sought fiscal concessions, including sales tax deferment, from the Madhya Pradesh government to make one of its refinery projects economically viable, since it could save the company about Rs 400 crore (Rs 4 billion) annually.

With most of the states set to move towards value-added tax, we have to question the effects of fiscal concessions on states' fiscal position, and their benefits, if they are provided to specific companies and/or specific sectors.

In the standard framework, subsidy for petroleum refining results in a reduction in its price because investors can produce and supply higher output at the same cost as before, due to the savings induced by the abatement.

Other investors in the petroleum refining sector (non-incentive, for example, Reliance) continue to produce their good at a higher price, which includes the tax (there are no taxes abated/deferred for them).

It is likely that these non-incentive companies will attempt to increase efficiency in their supply of petroleum refining, so that they are able to sufficiently lower the price of their goods, on par with that of the goods produced by the incentive-receiving companies.

Whether or not the non-incentive companies are able to do this depends on the demand elasticity for the goods produced by these firms.

If the demand elasticity were to be greater than zero, the demand for petroleum refining by the incentive-receiving company increases in response to a decrease in its price.

This is especially so if the non-incentive receiving company is unable to cut its costs correspondingly with that of the incentive-receiving company, and there will be a shift in demand for the goods produced by the incentive firm to whom increased profits accrue.

So there is a distortion created within the petroleum refining sector, between companies receiving the incentives and those that do not.

Further, assume that the tax incentive is provided not to a single company within the sector, but to the entire petroleum refining sector.

Now, there is incentive for non-incentive firms' capital to flow into the petroleum refining sector that is receiving the incentive, so that returns to factors in both the sectors are equalised.

With increasing use of capital and increase in the capital-labour ratio in the incentive-receiving companies or sectors, labour productivity and industrial output rise.

Thus, the standard framework shows that a tax incentive, although sector or company-specific to begin with, would eventually have general equilibrium effects on other firms and sectors.

We have to question whether these effects, and if there are benefits, are worth the costs of fiscal concessions to the sector or a specific company.

Further, it is not clear what the effects of incentives to specific companies would be under VAT. It is quite possible that if various states resort to offering such concessions to individual companies, the spirit of VAT and Kelkar's recommendations could be lost and introduce many exemptions in the already complex indirect tax structure.

The standard framework also shows that if the initial company receiving the tax incentive were to be capital-intensive, we may not expect the company to create much employment. Rather they have to be labour-intensive (companies with a high labour-output ratio).

Already in the post-reform phase, the domestic employment base in the organised sector is shrinking. So we want to question whether a tax incentive is "good" if it is demanded by a public sector company that is capital-intensive.

Conversely, an incentive, if sought by a labour-intensive company, intending to locate in a low-income, high-unemployment or poor area, could be justified for its effects on employment and net benefits from employment.

Earlier, I have argued that, net benefits from employment are likely to be higher in poorer, high-unemployment areas, as the wage at which the unemployed accept jobs in these areas are lower than what they would be in richer areas, consistent with what the literature has argued.

Net benefits from employment are the actual wages earned, net of this wage at which they would be willing to accept a job.

This is not to argue that it is socially beneficial for unemployed to accept jobs at low wages, but it makes sense for governments to encourage policies that increase the demand for workers. It does not make sense if tax incentives were given to capital-intensive companies.

Whenever a tax incentive is sought, it is necessary for the state governments to make an analysis of whether the benefits are worth the costs.

For this, the governments may make a contractual agreement with every company that demands a tax incentive from the government, regarding various aspects of the company's effects on the local communities where it intends locating.

These could be investments to be made, employment to be generated, expected wages to be paid, proportion of local residents to be recruited and any exporting they plan.

This introduces some accountability on part of the companies that seek various tax incentives and impose strain on the state's exchequer, and on the part of many populist governments that are responsible for fiscal consolidation.