Surjit Bhalla writes: GDP data critics protest too much

As it happens, the data turns out to support the government rather than the knee-jerk opposition.

Start with the deflator, where the critics have expectedly gathered. For a decade, the standard charge against Indian GDP data was single deflation. India applied one price index to output without separately deflating what firms paid for inputs — a squeeze on margins was recorded as extra production. The new 2022-23 national accounts series abolished single deflation outright, and manufacturing and agriculture are now double-deflated. That correction has not reached the critics, and certainly not the quick AI summaries doing the rounds. Further improvements are in the pipeline. We had only wholesale prices; producer price indices, which the IMF had asked for, were published for the first time in June.

Now to the devils/angels in the details. Prices of what India actually bought — consumption, government and investment together — rose 4.4 per cent, above consumer inflation of 3.9. Two facts pull the headline down to 2.3 per cent. Import prices rose 32 per cent, on imports worth 28 per cent of GDP, and imports enter GDP with a minus sign. And net indirect taxes fell 0.4 per cent in rupee terms even as the volume taxed rose 3.9. Recall that the government surrendered 10 rupees a litre of excise on petrol and diesel in March; fertiliser subsidies are heading for twice the budgeted allocation. Taxes forgone and subsidies paid show up as a lower deflator.

Further evidence comes from the GST cut itself. In October-December, the first full quarter after the September reduction, Indians spent 8.5 per cent more in rupees than a year earlier — down from 10.8 per cent the year before. Yet real consumption grew 8.2 per cent, up from 6.0. The consumption deflator had collapsed to 0.4 per cent. Households parted with fewer additional rupees and took home more goods — exactly what a tax cut passed through to consumers looks like. The real GDP growth number is real.

But… double deflation is only as good as the input prices you feed into it, and with imported input costs up 32 per cent this quarter, MoSPI should publish the input deflator series it used. A reform this consequential should be complete.

A request also to India’s GDP critics. Look at other, possibly more legitimate, targets of possible GDP data manipulation — the ones left alone because it is politically correct to do so.

Start with Bangladesh data, which yields the fashionable conclusion that its per capita income exceeds India’s — a claim that sells newspapers and makes careers, and is not so correct. Since 1980, four nominal base-year changes have cumulatively added 79 per cent to Bangladesh’s measured GDP, most of it through services never counted before. Take Vietnam, which added 25 per cent in a single GDP revision in 2019. Or take China, which has added 36 per cent through five economic censuses, and whose continued mercantilist export miracle succeeds, in part, because of the re-routing of exports via Vietnam, Mexico and elsewhere: On August 14, the White House published a report on what it called the Great Transshipment Scam, naming more than 40 countries through which Chinese goods are relabelled to dodge American tariffs.

And India? Its last two GDP data revisions, in 2011-12 and 2022-23, were downward. A statistical office that revises its own economy smaller, twice, is close to unheard of in the developing world. Perhaps India had bloated itself earlier and was merely correcting? Not so fast, and not so. India’s cumulative GDP addition since 1980 is only 11 per cent, against Bangladesh’s 79, Pakistan’s 51, China’s 36 and Vietnam’s 25 per cent.

Now to the all-important data on investment. Fixed investment grew close to 12 per cent, and its share of the economy rose almost three percentage points in a year, to above 34 per cent of GDP. Nor is this just one good quarter: Investment growth has risen in every quarter for the past five, from about 5 per cent in April-June last year to nearly 12 now. And there is supporting evidence. Bank credit growth has nearly doubled over the year, to 19 per cent, and with the repo rate held at 5.25 while inflation climbed towards 4, the real policy rate is down to about 1.3 per cent, near ideal.

Consider what a government intent on flattering itself would do. It would flatter consumption — 56 per cent of GDP, survey-based, much the hardest component for an outsider to check. Consumption is not where the strength is. It grew 7.1 per cent, slower than the economy, and its share fell. Investment, three-fifths its size, contributed exactly as much to growth. Nobody manufactures a boom where it is easiest to catch them.

The best evidence that the GDP growth data is real — it provides continuing evidence against the much-touted Viksit Bharat 2047 ambition. High income is defined in dollars. India’s dollar income per head grew about 5 per cent a year between 2012 and 2025. Over the past two years, it has grown by roughly nothing — last year it fell, from about $2,750 to $2,650, and we slipped from being the fourth-largest economy to sixth. Reaching 2047 requires close to 10 per cent a year for two decades. A good GDP quarter, perhaps even a year or two, does not even begin to dent that arithmetic. What can dent it is genuine structural reform — on trade, on factor markets, on the investment regime — which, tax cuts apart, has been in short supply.

Indian GDP data may be the best example we have of a government declining to flatter itself. We have succumbed to plenty else — the 2015 Model BIT, among the most self-defeating investment policies any country has written. But maybe, just maybe, a statistical system that has stopped cutting corners heralds a government that will stop cutting them elsewhere too.

The writer is chairperson of the Technical Expert Group for the first official Household Income Survey for India. Views are personal