Above is proof of chai tapri economics when a high quality moron presents a blogpost of some consulting editor of ET vs Goldman sachs research report as "evidence" to buttress his ridiculous claims.
I rest my case.
Interesting rebuttal to the quoted blog post by a former chief statistician of India.
http://blogs.economictimes.indiatim...bottom-of-discrepancies-in-national-accounts/
I rest my case.
Interesting rebuttal to the quoted blog post by a former chief statistician of India.
http://blogs.economictimes.indiatim...bottom-of-discrepancies-in-national-accounts/
Over the years, much column space has been devoted to reporting and commenting on India’s national income estimates. However, I have never seen an article which overtly assails the professional integrity of the head of the national statistical system as in Abheek Barman‘s article.
Barman seems to harbour a deep visceral dislike towards Chief Statistician of India TCA Anant, which ranges from his appearance to his pedagogy to his personal hygiene. All this is par for the course for social media trolls. But to accuse Anant of fudging data to please “his boss Arun Jaitley or his boss’ boss Narendra Modi” takes it beyond mere personalities to an assault on the integrity of an institution.
Do join the dots, you first need to know your numbers
It really pains me when senior economic journalists display abysmal ignorance of national accounts. And even more so when they don’t bother to do even rudimentary research and cross-checking. Barman clearly bunked classes when Stone & Stone (Richard and Giovanna) was being taught, and I seriously doubt that he has ever seen a copy of Simon Kuznets’ seminal work, both of which he has cited. Neither does he seem to be aware of the UN System of National Accounts (SNA) on which all countries base their GDP estimates. Or the Sources and Methods of the Indian national accounts.
If this technical stuff was beyond Barman’s patience or competence, he should have simply called Anant. But perhaps I am asking for too much humility from economic journalists. On the other hand, I have greater faith in the good sense of the layperson who is willing to understand a new concept if properly explained.
All the four authoritative sources cited above — Stone & Stone, Kuznets, the SNA, and Sources and Methods of national accounts statistics — begin by explaining that GDP can be calculated using three different approaches: 1) Production 2) Income (3) Expenditure. With perfect and complete data, each three approach should give exactly the same figure for GDP, since production generates income, income generates expenditure and expenditure is spent on what is produced. Unfortunately, in real life, the data are nowhere as perfect or complete.
Production data is reasonably easy to get. It comes mainly from enterprises, from their accounts or through surveys. Income data has to be collected mainly from households, preferably broken up by type such as wages, rents, interest, dividends, etc. But would you honestly give such information to anyone, even under duress?
Discrepancies
Expenditure data is collected partly from households (for consumption) through surveys and partly from enterprises (for investment). Surprisingly, households are much more willing to divulge their expenditures than their incomes — as the National Sample Survey (NSS) has demonstrated repeatedly — although it tends to be understated. It should be clear, therefore, that the quality of the data available for each of the three approaches will always vary substantially.
In the Indian case, as with most other countries, the income approach is not feasible. So we don’t produce this estimate. Even the other two that we do produce — production and expenditure — the data quality varies significantly. We can measure certain things better than others.
As a result, when the two sets of GDP estimates are produced, the totals do not match up. The difference between the two, in accordance with the SNA, is shown as ‘discrepancies’ —the source of Barman’s angst. The convention is that discrepancies are shown under the approach for which the data is less reliable.
The data on the production approach is far more accurate than the data on the expenditure approach. This is especially true for the quarterly estimates where we have no hard data for the expenditure side, and all estimates (except net exports) are based on proxies. Thus, the discrepancies are shown under the expenditure side.
If the situation had been reversed — if we had more confidence in the expenditure estimates — the discrepancies would have been shown under the production estimates. In which case, the discrepancy figure in Q4 2015-16 would have been an equally large negative, and Anant would probably have been pilloried for being a desh drohi and running the country down by the same commentators trashing him for ‘fudging’. So damned if you do, damned if you don’t, and to hell with journalistic integrity.
P.S. Due to similar uninformed pressures around the world, the latest UN SNA recommends the creation of Supply-Use Tables (SUT), which map production estimates
(The writer is former Chief Statistician of India)
Barman seems to harbour a deep visceral dislike towards Chief Statistician of India TCA Anant, which ranges from his appearance to his pedagogy to his personal hygiene. All this is par for the course for social media trolls. But to accuse Anant of fudging data to please “his boss Arun Jaitley or his boss’ boss Narendra Modi” takes it beyond mere personalities to an assault on the integrity of an institution.
Do join the dots, you first need to know your numbers
It really pains me when senior economic journalists display abysmal ignorance of national accounts. And even more so when they don’t bother to do even rudimentary research and cross-checking. Barman clearly bunked classes when Stone & Stone (Richard and Giovanna) was being taught, and I seriously doubt that he has ever seen a copy of Simon Kuznets’ seminal work, both of which he has cited. Neither does he seem to be aware of the UN System of National Accounts (SNA) on which all countries base their GDP estimates. Or the Sources and Methods of the Indian national accounts.
If this technical stuff was beyond Barman’s patience or competence, he should have simply called Anant. But perhaps I am asking for too much humility from economic journalists. On the other hand, I have greater faith in the good sense of the layperson who is willing to understand a new concept if properly explained.
All the four authoritative sources cited above — Stone & Stone, Kuznets, the SNA, and Sources and Methods of national accounts statistics — begin by explaining that GDP can be calculated using three different approaches: 1) Production 2) Income (3) Expenditure. With perfect and complete data, each three approach should give exactly the same figure for GDP, since production generates income, income generates expenditure and expenditure is spent on what is produced. Unfortunately, in real life, the data are nowhere as perfect or complete.
Production data is reasonably easy to get. It comes mainly from enterprises, from their accounts or through surveys. Income data has to be collected mainly from households, preferably broken up by type such as wages, rents, interest, dividends, etc. But would you honestly give such information to anyone, even under duress?
Discrepancies
Expenditure data is collected partly from households (for consumption) through surveys and partly from enterprises (for investment). Surprisingly, households are much more willing to divulge their expenditures than their incomes — as the National Sample Survey (NSS) has demonstrated repeatedly — although it tends to be understated. It should be clear, therefore, that the quality of the data available for each of the three approaches will always vary substantially.
In the Indian case, as with most other countries, the income approach is not feasible. So we don’t produce this estimate. Even the other two that we do produce — production and expenditure — the data quality varies significantly. We can measure certain things better than others.
As a result, when the two sets of GDP estimates are produced, the totals do not match up. The difference between the two, in accordance with the SNA, is shown as ‘discrepancies’ —the source of Barman’s angst. The convention is that discrepancies are shown under the approach for which the data is less reliable.
The data on the production approach is far more accurate than the data on the expenditure approach. This is especially true for the quarterly estimates where we have no hard data for the expenditure side, and all estimates (except net exports) are based on proxies. Thus, the discrepancies are shown under the expenditure side.
If the situation had been reversed — if we had more confidence in the expenditure estimates — the discrepancies would have been shown under the production estimates. In which case, the discrepancy figure in Q4 2015-16 would have been an equally large negative, and Anant would probably have been pilloried for being a desh drohi and running the country down by the same commentators trashing him for ‘fudging’. So damned if you do, damned if you don’t, and to hell with journalistic integrity.
P.S. Due to similar uninformed pressures around the world, the latest UN SNA recommends the creation of Supply-Use Tables (SUT), which map production estimates
(The writer is former Chief Statistician of India)
Last edited:
