Subir Gokarn: Necessities and virtues

One of the most striking things about poverty in India is how laggard we are in measuring it. Given the fact that every government, certainly since the late 1960s, has staked its reputation and claim to be re-elected on alleviating poverty, it is somewhat ironic that debates about poverty take place around an estimate of its incidence that may be as much as five years old.

That is certainly the case now, where the last large sample household expenditure survey, which is carried out approximately every five years and forms the basis of official poverty estimates for individual states and the country as a whole, was done in 2000.

We have no clue as to what has happened since then. The next survey is currently under way; the results will presumably become available for public debate towards the latter half of 2006.

There is an obvious message here for governments in the context of basic principles of management information. If you want to evaluate your performance on the basis of a specific parameter, you must have up-to-date information on it.

If this is not possible, for whatever reason, then do not use it as a measure of performance. Some other indicator that is visible with a minimal time lag and closely correlated to the ultimate objective""poverty alleviation or anything else""should then become the reference point.

Discovering meaningful correlations between "intermediate variables" and the "target variable" then becomes critical to good policy or strategy.

More so, if the correlations are established with variables over which the government has a reasonable degree of control. In other words, even if you don't know how much poverty there is in the economy at any point of time, as long as you know what actions are likely to reduce it over time, you have the basis for an effective anti-poverty programme.

Finding robust, long-term correlations between policy actions and poverty outcomes is the objective of a cross-country project sponsored by The World Bank and the Department for International Development (DFID) of the UK government, which seeks to bring the best technical capabilities to bear on the question of these correlations.

The India case study of the "Operationalizing Pro-Poor Growth" project, carried out by Timothy Besley, Robin Burgess and Berta Esteve-Volart of the London School of Economics, was recently presented at a workshop in Delhi.

The statistical findings are based on an analysis of changes in poverty incidence across 16 Indian states over the period 1958-2000. The results demonstrate in precise quantitative terms the disparity of outcomes across states.

While the overall correlation between economic growth and poverty reduction is validated by the data, what is striking is how states vary in terms of their ability to translate growth into declines in poverty.

The extremes here are Kerala and Bihar. While in Kerala, an increase of 1 per cent in per capita income reduces the number of people below the poverty line by 1.23 per cent, in Bihar, the reduction achieved is only 0.3 per cent.

These and other disparities in the long-term poverty reduction performance across states support a view that other factors can either reinforce or offset the impact of growth in incomes.

The report goes on to identify five such forces, based mostly on empirical findings from recent work done by the authors themselves, though not necessarily in the framework of this project.

They identify six key "action" variables, which, the evidence suggests, complement growth in reducing poverty.

Land reforms""specifically measures that formalise property rights over agricultural land""are significant. Then comes access to finance for the poor. Human capital formation, reflected by even very crude indicators like literacy, is closely correlated with poverty reduction.

This factor is reinforced by the gender dimension, which suggests that states with greater equality between genders in indicators like educational attainment are better at reducing poverty.

Labour regulations are significant, in that states that have tended towards "pro-employer" regulation, meaning more flexible labour markets, have tended to do better than states that have been "pro-labour", in that they have increased protection for workers.

Finally, political accountability""as reflected in the combined impact of media scrutiny, political competition and an increasing political assertiveness of previously marginal groups""makes a difference.

Do these findings collectively represent fresh, new insights which will help sharpen the focus of anti-poverty programmes in India? There is, unquestionably, an element of "obviousness" about them, with the possible exception of the labour regulations factor.

All of them will, for most people, fall into the category of policy measures with "intrinsic", as opposed to "instrumental", merit. They are things that all governments should be doing as a matter of principle, regardless of any direct economic consequences that flow.

However, the significant contribution that the study makes is to demonstrate that intrinsically good policies also have strong instrumental value.

This is where technique becomes critical. The methods used do not simply demonstrate correlation without the direction of influence. Given the relatively long time period over which the relationships have been explored, a reasonable cause-and-effect pattern can also be established.

The finding on labour regulations, on which I have commented before, is extremely important for a couple of reasons. First, it suggests that the tendency to persist with strict job security regulations (the epitome of a "pro-labour" stance) may offset the beneficial impact on poverty of other measures.

Second, it implies that the stimulus to job growth, which market flexibility provides, more than compensates for the job losses that may occur in a pro-employer environment.

Perhaps the most important contribution of the evidence presented by the study is that, even within a relatively centralised governance and resource allocation framework, there is enormous opportunity for the state and even local governments to make an impact on poverty.

The implication of this is that the Centre will play a more effective role if it is able to incentivise actions by state and local governments to do the right things.

The increasing use of conditionalities in transferring resources from the Centre to the states and further down is, therefore, a welcome sign.

Implementing these with a strong dose of credibility is clearly an immediate priority.

To conclude, it is always reassuring to hear that doing the right things, satisfying enough in and of itself, is also good for economic performance.

Virtue and necessity overlap. But, doing the right things right is neither automatic nor easy. Do we need to wait for another cross-country study or can we learn from our own experiences?

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