Gross Deflator Problem

Confusion endures(live on,टिके रहना) on India’s growth performance last quarter. Did growth accelerate or decelerate? This is important because, on a quarterly basis, tracking the momentum of growth is perhaps more important than its level. High-frequency investment and asset allocations as well as monetary policy are made in real time, and influenced by the state of the business cycle. So is growth accelerating or slowing down?

What is the source of the confusion? In the first quarter of 2015 (1Q15; all references are to calendar year) GDP growth surprised sharply to the upside, printing at 7.5 per cent, even as activity appeared soft on the ground. Corporations had their worst quarterly earnings in six years, a sharp fiscal squeeze was applied to meet the deficit target and agriculture was experiencing a drought year. Yet, national income accounts showed that GDP growth had accelerated sharply from 6.6 to 7.5 per cent — raising several eyebrows.

In the next quarter, exactly the opposite happened. Earnings improved and consumption was palpably(clearly,स्पष्ठ रूप से) benefitting from the collapse in oil and food prices. Yet, against all expectations, GDP growth decelerated to 7 per cent. What explains this disconnect?

The GDP is the sum of gross value added (GVA; at basic prices) and net indirect taxes (NIT). The GDP can be thought of as the whole pie, whereas GVA is the slice appropriated by labour and capital, and NIT is the slice appropriated by the government. There appear to be no particular issues with the dynamics of the GVA in recent quarters, though questions remain about the level of growth thrown up by the new GDP methodology — and whether that syncs with events on the ground. But that’s another debate. GVA growth slowed sharply from 6.8 per cent in 4Q14 to 6.1 per cent in 1Q15 and then re-accelerated to 7.1 per cent in 2Q15, as expected.

Instead, the issue lies with NIT. Real NIT growth jumped from 3.7 per cent in 4Q14 to 18.9 per cent in 1Q15, but then fell off dramatically to 6.5 per cent 2Q15. This is hard to comprehend. NIT should rise if collections rise or subsidies fall. So why did real NIT fall so dramatically in 2Q15, if underlying growth (as captured by the GVA) and, therefore, tax buoyancy(cheerfulness,प्रफुल्लता), was higher and subsidies lower?

The problem lies with translating nominal into real quantities; what economists call the “deflator”. The idea is simple: we want to deflate out all price movements so as to capture “real” economic activity. In the case of NIT, we also have to deflate out changes in tax rates or bases. Herein lies the nub(center,मुख्य बिंदु) of the problem. In nominal terms, NIT growth increased, as we would expect, to 22 per cent (year-on-year) in 1Q15 and to 40 per cent in 2Q15 — excise and service tax rates were hiked, activity grew and subsidies collapsed. But in real terms, growth decelerated sharply. Why? Because the implicit deflators used were 2.6 per cent in 1Q15 and a whopping 31.4 per cent in 2Q. This completely skews real NIT and GDP growth.

Both numbers are hard to reconcile(harmonize,सामजंस्य). There was virtually no inflation (as measured by the GVA deflator) in 1Q and 2Q15, so the deflators should largely be reflecting tax rate changes. Consider this: The weighted average increase in excise duties for petrol and diesel was 130 per cent in 1Q15, compared to the same period the year before. Just these tax increases — given the share of oil excise in total tax collections — should have resulted in a deflator of 8-9 per cent. This would have resulted in 1Q15 GDP growth printing at 6.9 per cent, not 7.5 per cent. Why then a deflator of 2.6 per cent? And why the humongous(large,बड़ा) jump in the deflator to 31.4 per cent in 2Q15? True, service taxes were hiked by about 13 per cent (from 12.36 to 14 per cent), but this was only applicable for one month (June) of the quarter. The service tax base was also increased slightly. But how does this result in 2.6 per cent jumping to 31.4 per cent?

Here’s another way to see the disconnect. The government estimates NIT grew at nearly 14.5 per cent in 2Q15, after adjusting for tax rate changes. With subsidies contracting, and GVA inflation at zero, 14 per cent should be a lower bound for real NIT growth. But instead it printed below 7 per cent. This suggests an improbably large collapse in state taxes during that quarter.

The bottom line is that the 1Q15 deflator has likely been underestimated and that of 2Q has been overestimated. The result: 1Q growth gets an artificial boost and 2Q growth is artificially depressed. We are, therefore, left with the wrong impression that growth is slowing. Instead, if the deflators are properly calibrated(fine tune,मापा हुआ), we believe GDP growth should have been just shy of 7 per cent in 1Q15 and a tad above 8 per cent in 2Q. This comes through from looking at GVA dynamics, untainted(stainless,निष्कलंक) by this problem. GVA sequential momentum surged(rise rapidly,तेज़ी से बढ़ना) from less than 4 per cent in 1Q to over 16 per cent in 2Q. Net of agriculture and government spending, it accelerated from 12.2 per cent to 15.4 per cent (quarter-on-quarter, seasonally adjusted and annualised).

By all accounts, the economy was picking up steam in 2Q — as one would expect given the large, positive terms-of-trade shock that India is experiencing. But this does not come through in the headline GDP numbers.

I’m sure, at some point, the data will be appropriately revised. But will that be too late? Remember, in 2010, the RBI began a gradual(step by step,क्रमिक) hiking cycle, only for GDP growth numbers for 2010-11 to be revised upwards retrospectively by almost 100 basis points. If it had had the right numbers, the RBI would certainly have embarked on(start,शुरुवात) a more aggressive hiking cycle. Similarly, inflation was retrospectively revised upwards for 1Q12, which, if known at the time, would not have elicited a 50-basis point cut in April that year. Policymakers don’t have the luxury of time. Neither should our statistics.