Bibek Debroy: Outsource the Survey

Most people probably read only one chapter of the Economic Survey, the first one. The other chapters aren't fixed or constant. They evolve and there is chopping and changing.

For instance, Chapter 4 on securities markets is a relatively recent addition. Who knows, we will probably have an independent chapter on the services sector soon.

If you track the Survey's evolution in the reform decade of the 1990s, you will find that chapters that have changed substantially are public finance (Chapter 2), monetary and banking developments (Chapter 3), the external sector (Chapter 6) and infrastructure (Chapter 9).

Other than the introductory first chapter. In contrast, chapters on prices and food management (Chapter 5) and agriculture (Chapter 8) have changed little.

This may be a reflection of reform priorities. It may also be a reflection of the interests chief economic advisers (CEAs) have had. With agriculture, rural development and social sectors being the reform priorities of the Common Minimum Programme (CMP), CEAs might have to change their interests.

Indeed, in the Economic Survey 2004-05, the chapter on social sectors (Chapter 10) seems longer. It isn't actually that much longer in terms of number of pages. It is just that the social sector chapter has many more sections now.

And a section on law reform has crept in through contract enforcement into the infrastructure chapter.

This year's Survey has fewer outright typos. Proof-reading is more careful. But only in the actual text. No one bothered to proof-read the list of tables and boxes.

In these (pp. iv + v), you discover that the Survey still can't spell central, expiry, international, influenza, electricity, Jawaharlal and types. Rather perversely, the other typos are in the social sector chapter "" like an inability to correctly spell drought (page 226), per (page 237) and matric (page 242).

But the typos are certainly less this year. Once you have eliminated typos, you can begin to look for consistency across different parts of the Survey. That's always been a problem.

Because different advisers write different chapters and no one undertakes a simple test of internal consistency. What percentage of the Indian population earns a living from agriculture? The agriculture chapter tells you, "More than 65 per cent of Indians derive their livelihood from agricultural activities."

The social sector chapter tells you, "As a result, the share of agriculture in total employment dropped from 60 per cent in 1993-94 to 57 per cent in 1999-2000."

Clearly, the agriculture author doesn't believe the National Sample Survey.

How many polio cases are there? "During 2004, facilitated by a high quality of pulse immunisation programme, the number of polio cases has declined drastically to 133 cases, as on 4th February 2005, from 28,257 in 1987 [page 239]."

Rather oddly, on page 234 we have Table 10.7, described as time-trends in healthcare and the text tells us this table shows the large healthcare infrastructure in the country.

The table has figures on health centres, dispensaries and hospitals, beds, nursing homes and doctors. What is odd is that this table also has figures on malaria, leprosy and polio.

These can hardly be described as healthcare infrastructure. If anything, they demonstrate the reverse. Odder still, this table has a polio figure of 29,709 in 1951, 225 in 1981 and 214 in 2003.

True, no figure is given in the table for 1987 and it is indeed possible that polio cases dropped to 225 in 1981, shot up again to 28,257 in 1987 and declined again to 214 in 2003 and 133 in 2005. But this seems unlikely.

On page 237, we are given life expectancy at birth figures for males and females and the note to the table tells us, "The dates in the brackets indicate years for which latest information is available."

The figures happen to be 63.9 years for males and 66.9 for females. But the dates within brackets are 2001-06 and this isn't an ordinary typo because Appendix Table 9.1 repeats similar figures state-wise.

Are these figures therefore actuals or projections? Since no mention is made of projections, the government must be doing a remarkably good job of getting latest information before the due date. The next example concerns the literacy rate.

Page 231 tells us the literacy rate has increased to 64.8 per cent in 2001. This is a Census figure. On the same page, there is a table from Human Development Report (HDR) and for the year 2002, this gives an adult literacy rate of 61.3 per cent.

One can deduce the reason for the apparent anomaly. The first is for population aged seven and more, while the second is for population aged 15 and more.

But surely, these figures shouldn't have been quoted (that too on the same page) without some explanation. Actually, I am not very sure the Survey knows the difference between adult literacy and overall literacy. On page 233, there is a section that discusses adult literacy and the National Literacy Mission.

But the figures given (on male-female literacy gaps) are for overall literacy rather than adult literacy. In the issues and priorities section of the first chapter, a case is made for private sector participation in physical infrastructure, government production being inefficient.

The argument also holds for producing the Survey.