- Joined
- Apr 7, 2010
- Messages
- 5,033
- Likes
- 8,354
In an interview with ET Now, Roopa Kudva , MD & CEO, CRISIL , gives her views on the need for an accurate estimate of GDP. Excerpts:
You have mentioned in that report that the current methodology for GDP estimation does not accurately measure the contribution of the unorganised sector. Now considering that nearly half of our GDP comes out of the unorganised sector, would you say that our GDP estimates are seriously out of whack and also would you say that the GDP reading is actually much higher than what is being printed?
Indeed, about half of India's GDP does originate in the unorganised sector and the way this GDP is currently estimated is by surveys that are conducted by the Central Statistical Organisation once in every 5 years and for the intervening years in between, the data is simply extrapolated. Now in an economy whose structure is changing so rapidly as of the economy of India's is and also in a country like us where self employment is increasing so significantly, use of benchmarks that are refreshed only once in 5 years can get pretty outdated and this would therefore in our opinion lead to an underestimation of the GDP. In our view, therefore, what would be more appropriate is that such surveys to estimate the contribution of the unorganised sector to the GDP should ideally be carried out once every year.
So does that mean that the GDP numbers ought to be higher than what actually comes out? Net net, is that the assessment that you have?
Yes, that will be assessment. Because we are using outdated benchmarks and because the structure of the economy is changing so rapidly, significant increases in self employment would lead us to believe that were these surveys to be done at more frequent intervals, the GDP estimates would be higher.
You have also mentioned in your report that the GDP estimates do not account for the improvement in the quality of goods, could you elaborate on that?
There are a wide variety of goods that we use today whose features, quality and therefore the value delivered is far higher than it was, say, 10 years ago and yet the prices of these goods have not gone up commensurately. Let me give you a simple example. For Rs 15,000 today, you can get a mobile phone that offers you far better feature and therefore far better value than paying Rs 15,000 for a mobile phone got you 10 years ago in this country or indeed anywhere else in the world. And, therefore, an accurate estimate of GDP would actually correctly capture the increase in value provided to the consumer through the improvement in quality. How they adjust for this in other countries like the US, for example, is by making a downward adjustment to price indices which accurately therefore...
Focussing on the services sector, based on your assessment, the estimation of the services sectors you have mentioned is constrained by the absence of appropriate price and this is to convert nominal GDP into real GDP. Now to the extent that the services sector accounts for more than 50% of our GDP, what is the impact on GDP that we should expect once this issue does get addressed?
It is really not an easy job at this point to say to measure what the quantum of the impact is, but the key issue really is that nominal GDP is converted into real GDP using price indices and today in India , we do not have any robust price indices for the services sector, with the result therefore that the service sector GDP estimation is constrained to that extent and is inaccurate to that extent. The first step therefore is for the statistical organisation to construct and put in place robust indices that actually come up with accurate service sector estimates. And in the absence of that, it is tough to really say what the impact is, but I would imagine that this too is more likely to have led to an underestimate in service sector GDP as opposed to an overestimate.
And does that mean that the revision in the base year for IIP computation is expected to add 40 basis points to the GDP this year, but is not revising the GDP estimate on the basis of this from 8.5% to 8.9% a rather simplistic approach on their various other factors that could still mean that the GDP will come in at maybe even though 8.5%?
Sure there are and the purpose of this study was not really to come up with a revised estimate of GDP, but just to carry out an analysis of what the methodology related issues are which are impacting an accurate estimate of GDP. The simulation that we have done is indeed very simple. What it assumes is that agriculture GDP and service sector GDP do not change. All we have done is we have used the revised, the fact that the base year for IIP computation has been revised from 1993-1994 to 2004-2005 and therefore we believe it will result in an increase in the industrial GDP estimate for 2010-2011. And that factor alone will move the GDP estimates up by 40 basis points from 8.5 to 8.9.
Now it is important to highlight that the impact of these changes can be pretty significant. Let's take the example of what the impact of this base year change has been for two years. If you look at 2007-2008, this change in base year has actually resulted in almost a doubling of the IIP industrial growth rate from about 8.5 to about 15.5 and for 2 years after that 2009-2010, it has actually led to a halving of the industrial growth rate from about 10.5% to 5.5%. So the point really is to highlight how significant an impact of the use of the correct benchmarks, of timeliness of capturing data, frequency of capturing data, all of that can really have a pretty significant impact on GDP estimation. And that becomes important because it is the single most important economic indicator and has a vital role to play in policy formulation.
While the base year for IIP has been changed and the IIP basket has also been upgraded, everyone agrees that the data quality has improved, but would you still say that it is still far from being reliable given how erratic the reading still is.
There is no doubt about the fact that the policy quality of the basket has improved in that it has become more contemporary. So you have a situation where about 140 new items have come in, about 30 odd items have gone out from the old basket. And so to that extent, the basket is more fresh and more contemporary and more reflective of the reality of the industrial production in the country. However, even between 2004-2005 and currently, there have been significant changes which need to be captured for a more accurate estimation of industrial production. Take for example a whole range of consumer electronics, LED TVs for example. One example of a new item which has grown significantly in industrial production, but is not being captured in the new basket. We would therefore say that it would perhaps be more appropriate to move the base year every 5 years as far as IIP is concerned.
Outdated benchmarks lead to understimation of GDP: Roopa Kudva, CRISIL - The Economic Times
You have mentioned in that report that the current methodology for GDP estimation does not accurately measure the contribution of the unorganised sector. Now considering that nearly half of our GDP comes out of the unorganised sector, would you say that our GDP estimates are seriously out of whack and also would you say that the GDP reading is actually much higher than what is being printed?
Indeed, about half of India's GDP does originate in the unorganised sector and the way this GDP is currently estimated is by surveys that are conducted by the Central Statistical Organisation once in every 5 years and for the intervening years in between, the data is simply extrapolated. Now in an economy whose structure is changing so rapidly as of the economy of India's is and also in a country like us where self employment is increasing so significantly, use of benchmarks that are refreshed only once in 5 years can get pretty outdated and this would therefore in our opinion lead to an underestimation of the GDP. In our view, therefore, what would be more appropriate is that such surveys to estimate the contribution of the unorganised sector to the GDP should ideally be carried out once every year.
So does that mean that the GDP numbers ought to be higher than what actually comes out? Net net, is that the assessment that you have?
Yes, that will be assessment. Because we are using outdated benchmarks and because the structure of the economy is changing so rapidly, significant increases in self employment would lead us to believe that were these surveys to be done at more frequent intervals, the GDP estimates would be higher.
You have also mentioned in your report that the GDP estimates do not account for the improvement in the quality of goods, could you elaborate on that?
There are a wide variety of goods that we use today whose features, quality and therefore the value delivered is far higher than it was, say, 10 years ago and yet the prices of these goods have not gone up commensurately. Let me give you a simple example. For Rs 15,000 today, you can get a mobile phone that offers you far better feature and therefore far better value than paying Rs 15,000 for a mobile phone got you 10 years ago in this country or indeed anywhere else in the world. And, therefore, an accurate estimate of GDP would actually correctly capture the increase in value provided to the consumer through the improvement in quality. How they adjust for this in other countries like the US, for example, is by making a downward adjustment to price indices which accurately therefore...
Focussing on the services sector, based on your assessment, the estimation of the services sectors you have mentioned is constrained by the absence of appropriate price and this is to convert nominal GDP into real GDP. Now to the extent that the services sector accounts for more than 50% of our GDP, what is the impact on GDP that we should expect once this issue does get addressed?
It is really not an easy job at this point to say to measure what the quantum of the impact is, but the key issue really is that nominal GDP is converted into real GDP using price indices and today in India , we do not have any robust price indices for the services sector, with the result therefore that the service sector GDP estimation is constrained to that extent and is inaccurate to that extent. The first step therefore is for the statistical organisation to construct and put in place robust indices that actually come up with accurate service sector estimates. And in the absence of that, it is tough to really say what the impact is, but I would imagine that this too is more likely to have led to an underestimate in service sector GDP as opposed to an overestimate.
And does that mean that the revision in the base year for IIP computation is expected to add 40 basis points to the GDP this year, but is not revising the GDP estimate on the basis of this from 8.5% to 8.9% a rather simplistic approach on their various other factors that could still mean that the GDP will come in at maybe even though 8.5%?
Sure there are and the purpose of this study was not really to come up with a revised estimate of GDP, but just to carry out an analysis of what the methodology related issues are which are impacting an accurate estimate of GDP. The simulation that we have done is indeed very simple. What it assumes is that agriculture GDP and service sector GDP do not change. All we have done is we have used the revised, the fact that the base year for IIP computation has been revised from 1993-1994 to 2004-2005 and therefore we believe it will result in an increase in the industrial GDP estimate for 2010-2011. And that factor alone will move the GDP estimates up by 40 basis points from 8.5 to 8.9.
Now it is important to highlight that the impact of these changes can be pretty significant. Let's take the example of what the impact of this base year change has been for two years. If you look at 2007-2008, this change in base year has actually resulted in almost a doubling of the IIP industrial growth rate from about 8.5 to about 15.5 and for 2 years after that 2009-2010, it has actually led to a halving of the industrial growth rate from about 10.5% to 5.5%. So the point really is to highlight how significant an impact of the use of the correct benchmarks, of timeliness of capturing data, frequency of capturing data, all of that can really have a pretty significant impact on GDP estimation. And that becomes important because it is the single most important economic indicator and has a vital role to play in policy formulation.
While the base year for IIP has been changed and the IIP basket has also been upgraded, everyone agrees that the data quality has improved, but would you still say that it is still far from being reliable given how erratic the reading still is.
There is no doubt about the fact that the policy quality of the basket has improved in that it has become more contemporary. So you have a situation where about 140 new items have come in, about 30 odd items have gone out from the old basket. And so to that extent, the basket is more fresh and more contemporary and more reflective of the reality of the industrial production in the country. However, even between 2004-2005 and currently, there have been significant changes which need to be captured for a more accurate estimation of industrial production. Take for example a whole range of consumer electronics, LED TVs for example. One example of a new item which has grown significantly in industrial production, but is not being captured in the new basket. We would therefore say that it would perhaps be more appropriate to move the base year every 5 years as far as IIP is concerned.
Outdated benchmarks lead to understimation of GDP: Roopa Kudva, CRISIL - The Economic Times
