Column: Lies, damned lies, and worse

Written by Surjit S Bhalla | Updated: May 10 2014, 09:40am hrs
The closer Narendra Modi gets to be a possible PM, the more intensified becomes the slur and troll campaign of intellectuals and academics opposed to Modi. I have written several articles on growth, poverty, and living standards for the different states of India and for different socio-economic groups. The footnote to the table is just a partial listing of this research. The broad conclusion of the research: on all three countsgrowth, poverty reduction and welfare improvement of MuslimsGujarat has done very well.

This conclusion has not been met with approval, at least from the anti-Modi brigade. There is nothing wrong with disagreement, and such disagreement, if it points to errors in analysis, is always welcome. But what Soz, and to a lesser extent Alagh and Ghosh-Sood commit is intellectual dishonesty, and in my book there isnt a larger crime that an intellectual/academic can commit. The crimes arise from ideology. We are all ideological animals, that is not a problem. Intellectual dishonesty is when one makes an error fully knowing that one is wrongi.e. one lies. Such an accusation needs to be backed up by evidence, and that is what this article is about.

Alagh: A former economist and former Planning Commission official, Alagh does not even bother to present any evidence for his quite dishonest interpretations about poverty levels in Gujarat. His conclusion: the richer a state is, the lower its poverty levelshence, Gujarat has lower poverty, whatever that means. Alagh should know better, and given that he does, he is being intellectually dishonest in making the above statement. Poverty levels are a function of several initial conditions, among which per capita income or consumption and its distribution are two of the more important. Delhi, for example, in FY12 had a per capita income level 65% higher than the second richest big state, Haryana, yet its poverty level was just 1.1 percentage points lower.

The table shows poverty levels for various socio-economic indicators for two comparable states Gujarat and Maharashtra. Note the difference in ranking of Gujarat according to CSO data (4th) and NSSO (12th). The poverty data needs to be interpreted with reference to this NSSO rank among 21 big states; if any indicator for Gujarat is less than 12, than Gujarat is performing better than expected. Poverty levels for the different groups are generally lower in Gujarat than the richer Maharashtra. This holds true for all groups except Narendra Modis own OBC casteperhaps now the Congress intellectual trolls will complain that since his own caste has relatively lost out, Modi is not fit to be PM!

There are other problems with Alaghs rant, most importantly that he accuses Ashok Gulati of publishing results because he was paid to do so; Gulati has responded to the depths to which Alaghs posturing has descended (

I have published three research based articles specifically on the poverty situation of Muslims in

Gujarat. Modi Metric-I, (Bhalla1), was based on the then latest available NSSO data for FY10. This article concluded that while Gujarat had delivered exemplary growth, it had performed very badly on poverty reduction for Muslimsamong the worst. Gujarat has delivered growth under Mr Modi; equally emphatically, growth in Gujarat has neither been equitable nor inclusive.

This result has been seized upon by the dishonest detractors. Dishonestly not known to them, in two subsequent articles, I document in detail what happened to poverty in each of the three years, FY00, FY10 and FY12. The 2011-12 survey was especially commissioned by the Government of India (normal lapsed time between surveys is 5 years) because FY10 was a problematic drought year. The data for this survey were released in mid 2012; in Bhalla2, on October 26, 2013, I concluded: The poverty ratio for Muslims, which had not shown much change between 1999-2000 and 2009-10, now collapses to only a 11.4% level from the high 37.6% level observed just two years earlier. I could have chosen to not report the FY10 data, and thereby hide the sharp two year change, but did not do so, because that would have been dishonest.

One of the two survey year data, FY10 or FY12, has to be an outlier; both cannot be right. The very next week, I examined data for six large NSSO surveys conducted since 1983 and concluded: If the 2009-10 data was freely and willingly accepted and endorsed why not the same acceptance for the 2011-12 data the large decline in poverty shown between 2009-10 and 2011-12 is statistically suspect and deserving of further investigation It appears that several statistical criteria favour rejecting the estimate provided by the 2009-10 NSS data. In other words, the FY12 data was deemed to be comparable to the other NSSO years, not FY10exactly the same conclusion reached by most researchers and the government of India.

Soz: Given this background of full transparency, let us examine what the intellectually dishonest Soz does. First, he does not mention Bhalla3 at all, the article where I directly compare results for FY10 and FY12. Second, he accuses me of not looking deeper into sample sizes! And instead cites the news portal Counterview which claimed that rural Muslim poor in the NSS survey comprised of a mere 5 households. This betrays both Counterviews and Sozs complete lack of understanding of statistics. The relevant sample size to be considered is not of the rural poor, but of the rural Muslim universe. I had not cited any rural or urban figures precisely because the sample size of Muslims for each region was too small to reach any conclusion. The NSS surveys are not designed to capture the consumption behaviour of a subset of population and in Gujarat, Muslims constitute less than 10% of the population

Ghosh-Sood: They complain about three fundamental inconsistencies in my approach. Essentially, they use me as a peg to plug Soods book. I am flattered. But their article is flawed analysis and obfuscation. First, they criticise me for changing my views on Gujarat, but dont bother to explain to the readers that I changed my view after examination of the FY12 data. To rephrase Keynes, an honest person changes his view when new evidence presents itself, a dishonest person does not!

Second, Ghosh-Sood talk a lot about inequality worsening in Gujarat. Their analysis is flawed. They look at inequality for all 32 states, but most researchers prefer, for reasons of sample size, only the 21 big states. Inequality is better measured as real inequality (accounting for price differences between regions and states). Maharashtra has one of the highest inequalities in India; Gujarat inequality much better at 7th lowest, in the top one-third, and much better than its consumption rank of 12 would indicate.

It is sad that intellectual dishonesty is being indulged in by so called intellectuals/academics. One can escape disrepute and ridicule on Twitter because replies are restricteda newspaper column exposes ones nakedness.

The author is chairman of Oxus Investments, an emerging market advisory firm, and a senior advisor to Zyfin, a leading financial information company. Twitter: @surjitbhalla