Misuse of economic data

I have often wondered as to how much the debate in India would have been different if the Congress had not lost the 2014 election so overwhelmingly. A decline of 162 seats — from 206 to 44 — was the second-largest in Indian history. The largest Congress defeat was a decline of 218 seats in 1989, after an assassination-boosted 1984 victory, and 2014 is tied with the Emergency decline of 163 seats.

But 2014 was especially devastating — it reduced the Congress to 11 seats short of even being called a major opposition party. The “intended consequence” of this defeat — all data began to be questioned. It was as early as January 2015, less than eight months after the installation of the new regime, when the revised GDP data were released. Senior and seasoned commentators and experts began to question the “rightness” of the data.

They began to question the downgradig of GDP growth for the last three years of the UPA-II regime — that in real terms GDP growth for the UPA years was lowered by a cumulative 2 percentage points (ppt). That is 0.7 ppt a year. Greater than this magnitude of error has been serially (continuously) committed by experts at the RBI/MPC in their three-month-ahead inflation forecasts. As is well known, there is an inverse one-to-one correspondence between real growth and inflation estimates.

So, the error being questioned by the CSO is of a lower order of magnitude than that by the experts at the RBI (which, incidentally, also questioned the GDP growth data when they were first released).

Forgotten in the din of complaints was the fact that nominal GDP was equal in the revised and old estimates for 2012-13!! That double exclamation mark is not a typo and is meant to doubly emphasise. But the complaints continued. Such complaints are age-old and justified.

It is not data quality that I want to discuss today, but rather the deliberate misuse of data. Let me repeat: Deliberate, means knowingly, and misuse means not for good purpose. As some readers may have noted, my column is titled ‘No Proof Required’, and has been titled as such for more than a decade, and for good reason. Indian “experts” are generally experts at half-baked ideologies, rather than seeking “truth”. I am not being presumptuous but the purpose is to at least strive to seek the truth, even though truth itself might be elusive.

So, what is new or different about the debate over the last few years? The fact that the discussion about the quality of data has become openly political, that is, it is being deliberately misused by experts. The politicians always misuse — wouldn’t expect any different and quite honestly, it is part of their job. But what is new is that so-called experts are readily loaning their misuse “analysis” to the politicians. This is disturbing.

It is my tentative hypothesis — and one is free to criticise or agree — that this misuse is related to the magnitude of the Congress defeat in 2014. What I want to pose is the following counter-factual. As economists know, they only think in counter-factuals. Think what the debate on data would have been if the Congress had won 140ish seats in 2014, as they did in 1996, 1998, and 2004. Indeed, in five elections, 1996-2009, the Congress averaged 149 seats; the BJP average during these five elections — 156 seats. Hence, 2014 was a deep structural break from the past — a gap of 238 seats between 282 for BJP and 44 for INC. Average absolute gap between the two parties in the eight elections between 1989 and2014 — 86 seats.

Think about it — the Congress is the oldest, and most successful political party in India. Founded almost 150 years ago, it was the lead party in the battle for Independence, and had ruled India for all but 12 years at the time of its ignominious humiliation in 2014. And that too at the hands of a chaiwallah. Such an abrupt end would break most mortals, forms and organisations, but the Congress is obviously made of sterner stuff (I am just being truthful). It decided to fight back by employing the tactic of discrediting the Modi government on any issue, even before the issue had arisen. Political questioning of the GDP data in early 2015 was just the opening salvo.

At the outset, let me emphasise that there is only one rule for the political opposition worldwide — embarrass the ruling party. Unfortunately, for the BJP, it has not understood, or mishandled the optics on several of these “embarrassments”.

The latest opinion polls show that the number one concern of the youth is jobs. And well it might be. If the government were to release the latest NSSO data, we would be wiser. At present, we have to infer that there are too few jobs in the economy, as we must because the unemployment rate, according to the leaked NSSO report, has shot up to a 45-year high of 6.1 per cent. This must be “true” because it is also supported by the data put out by Mahesh Vyas of CMIE. His data now shows that the unemployment rate is now a high and higher 7.2 per cent. As Mohandas Pai documents, (‘Unemployment in India? Use reliable databases for jobs, not thin samples’, Financial Express, March 8) Vyas links the decline in GDP growth to the decline in investment proposals, and this decline to the increase in unemployment. The circle is complete. Logical? Not really.

There are so many flaws in the circle that it looks more like a zig-zag convoluted Z. First, GDP growth is not a function of approvals but rather of actual investment. Second, GDP growth is not a function of nominal investments but real investments. In 2013-2014, the year just preceding UPA defeat, nominal investment was 31.3 per cent of nominal GDP. The highest ever for this ratio was 39.6 per cent in 2011-12. The latest data for 2017-2018 shows that this share has declined to 28.6 per cent — a 11 percentage point decline from the peak and providing fuel to the Congress fire that investment, growth and jobs are way below the 2011-12 levels.

However, unlike politicians and armchair economists, economists like to use data on real investments adjusted for inflation. We don’t use nominal GDP growth, we use real GDP growth. Why should investments, one of the main contributors to growth, be in nominal terms? Prices of investment goods move differently than prices of GDP.

Hence, using the share of nominal investments to GDP, while infinitely better than using investment approvals (honestly, I don’t know what good that data is good for) is still not as kosher as using the share of real investments in real GDP. The peak of this series was 35.6 per cent in the growth boom year of 2007-08. In 2013-14, the year before the installation of the Modi government (what the government inherited) the real investment to real GDP ratio was 31.1 per cent; in 2017-18, the real ratio was 30 basis points higher at 31.4 per cent of GDP.

And only 400 basis points away from the all-time peak, not 1,100 basis points away as indicated by nominal investment/GDP ratio. Thus, a very different picture is painted by real investment rather than the flawed nominal share of investments. Yet it is the nominal investment share that is being used to show that GDP growth cannot be 7 per cent plus as stated by the CSO, but is likely to be lower; and that unemployment rate is, naturally, zooming up.
Over the next few weeks, several other misuse data flaws will be pointed out. And hopefully, the government would have released the PLFS employment-unemployment data for 2017-18.