Lies, damn lies and statistics?

The transparency and accuracy of the official data brought out by the government of India and the states have always being questioned, especially in recent years as the demand for information soars up with the economy moving even faster to the free market mode. While expert after expert has pointed to the poor quality of the data supplied across different sectors the official view, at least from the ministry of statistics, has been that India has one of the best statistical systems in the world. How does one explain this strange dichotomy? An answer to this question can be found from the assessments of India?s statistical systems made by independent organisations.

The Statistical Capacity Indicator (SCI) provided by the World Bank that provides information on the statistical systems in 140 developing countries is a good yardstick for making such comparisons. The SCI is a composite score of the three major dimensions of the quality of data, namely statistical practice, data collection and indicator availability and rates each country on a scale of 0-100.

The Statistical Capacity Indicator (SCI) gave India a composite score of 86 in 2008 which is much higher than the 65 score of all the developing countries. And India?s score is certainly much higher than that of most other important developing countries like Russia (78), Brazil (77), South Africa (77), Turkey (77), Mexico (76) and China (59).

But the achievements in the three different dimensions of statistical capacity vary. The highest score of 90 is for statistical practice or ability to adhere to internationally recommended standards and methods. India also scores a respectable 87 in indicator availability, which is much higher than the all country average of 77. India?s worst record is in the area of data collection or the frequency of surveys, censuses and completeness of registrations where the score is 80, though it is still considerably better than the overall average of 62. All these indicate that the official view that India has one of the best statistical systems is not completely out of place.

However, this is only one side of the story. Assessments made by other organisations point out that despite the good capabilities, the actual quality of the data is far below expectations. A survey of the data users done by the World Bank a few years ago pointed to innumerable drawbacks. These include insensitivity to the needs of the changing economy, poor coverage of areas like services and the agriculture sector, large time gaps in reconstruction of production and price indices, absence of an overall consumer price index and service sector price index, lack of accuracy of data, exclusion of off budget activities from fiscal statistics, large errors and omissions in the BOP statistics, inadequate details on FDI flows, non-availability of current indicators on capital expenditure, inventory accumulation and housing.

Other drawbacks pointed out include the lack of timeliness in the supply of data, frequent revisions, poor flagging of large revisions and their inadequate explanation which leads to credibility problems. And when data supply is timely, like in the case of the WPI data, their coverage is unclear with large gaps between the provisional and final figures. On the whole the general refrain is that the official data is becoming more inaccurate and that the revisions are becoming more volatile.

There were also complaints of pressure to hide bad news by stretching definitions and using methodological tricks. Similarly inappropriate timing of revisions prior to important policy measures also raised credibility questions. It has also been pointed out that accountability is spread far too wide for any one agency to control quality.

The extensive nature of the complaints indicates that the scenario is far from satisfactory and that the poor quality of data and practices can have serious repercussions on policy. The recent monetary policy changes based on inaccurate price indicators are one such example.

But then the real question is what accounts for the growing gap between the capabilities of the statistical system and the quality of data.

An answer to this question can be found in the allocation of resources for the official statistics collection. Figures show that the total budget allocation for the two major statistical agencies, the Central statistical Organisation (CSO) and National Sample Survey Organisation (NSSO) was only Rs 22 crore and Rs 120 crore in 2007-08, a measly sum for collecting data in a trillion dollar economy. These small allocations have curtailed the growth of the human resource capabilities of the statistical establishment. Though no break up is available of the strength of the individual agencies, the numbers of the ministry of statistics and programme implementation show that the total number of personnel has declined from 6,692 in March 2000 to 6,483 in March 2008, even as the data needs and coverage have increased.

Apart from inadequate financial and human resources, another major problem is that the Collection of Statistics Act of 1953 covered only a part of the industrial sector and most of the official surveys were done on a voluntary basis. Consequently there were large gaps in collection of information with actual reporting coming down to about a fifth of the sample as in the case of the WPI. This handicap will hopefully be addressed by the newly enacted Collection of Statistics Act of 2008, which provides the various designated statistical agencies the legal backing for collecting information and also empowers them to enforce penalties on the defaulters.