ExplainSpeaking: Why is India’s GDP facing another controversy

After the headline quarterly GDP growth numbers came out last evening and after the usual round of drum-beating over them, here is the harsh reality:

1. The headline numbers are overstated by a full percentage point because of the price deflators used.

2. Consumption growth is… https://t.co/76SMiWc3aH

— Jairam Ramesh (@Jairam_Ramesh) September 1, 2023

Second, some economists also pointed to the two main ways in which India calculates its GDP and argued that they did not match up. In other words, the allegation is that the government is favouring the higher estimate of GDP by using a statistical tool called “discrepancy”.

However, to understand these two criticisms as well as whether or not they are well-founded, readers need to understand the following things about how GDP is calculated in India.

What is GDP and how is it calculated?

The GDP is the most basic way to assess the performance of any economy — both from one year to another as well as across different countries. The idea behind the concept is to assess the size of an economy.

Ideally, the size of the economy should grow from one year to another. And this is what happens more often than not. Barring years when some terrible crisis happens — like in 2020 there was a global pandemic — all economies tend to grow from one year to another.

That’s because technically GDP is defined as the total “market value” of all final goods and services in an economy. Thanks to the bit about “market value”, the GDP can grow either because an economy actually produces more of goods and services or because the same level of goods and services are priced higher, or — as it happens almost always — a combination of the two factors.

Real GDP versus Nominal GDP

It is important here to distinguish between nominal GDP and real GDP. The overall GDP that one observes — by adding up the market value — is called the nominal GDP. But to arrive at the “real” GDP, statisticians remove the effect of price inflation from nominal GDP.

Imagine for a moment that the Indian economy produces only tractors, and that it produced 100 tractors in 2022, each valued at Rs 100. Then India’s GDP in 2022 will be Rs 10,000. Suppose further that in 2023, India’s GDP is Rs 11,000.

In nominal terms, India’s GDP has grown by 10%. But as a policymaker, one might be interested in knowing what caused the increase in GDP.

To be sure, it is entirely possible that all the increase in GDP — Rs 1,000 — happened because the price of each tractor went up by 10%. In other words, inflation went up by 10%. In such a case, the growth of “real” GDP — the increase in the actual number of tractors — is zero per cent.

There is another way in which the GDP could have grown by Rs 1,000 — that’s if India made 110 tractors in 2023, while all of them continued to be priced at Rs 100. In this case both the nominal GDP growth rate as well as the real GDP growth rate would be 10% while inflation would be zero per cent.

So real GDP growth rate is arrived at by taking the nominal GDP growth rate and stripping it off the effect of inflation. This is done by using what is called a GDP or price deflator — the term that Jairam Ramesh used in the quote above.

Now here is the crucial thing: The GDP deflator is not the same number as the retail inflation or the wholesale inflation rate; rather it is a mix of the two.

Since the real GDP growth rate — which is the growth rate one usually talks about — is a derived number, it crucially depends on the rate of GDP deflator that the official statisticians assume to be the rate of inflation in the country.

Two ways to calculate GDP

Beyond real and nominal GDP, there is an additional complication. This pertains to how one calculates the GDP per se.

To arrive at the size of the economy, should one look at all the money earned by everyone in the country or should one look at all the money spent by everyone in the country?

The former method is the income method and the latter method is called the expenditure method of calculating the GDP. As it turns out, India calculates the GDP through both the methods.

On paper, GDP should be the same no matter which method one uses. Indeed, how can it be different? But in reality many issues crop up.

For instance, timely data availability. Official statisticians can’t always know exactly how much money was spent or earned — at least not within the time frame of a quarter or a year. That is why national income data often gets revised.

As such, it is not hard to imagine that the two estimates of GDP aren’t always equal — especially not when quarterly data comes out. The difference between the two GDP estimates is called “discrepancy” — it is like that “miscellaneous” heading under which one puts small expenditures that one cannot immediately recall or verify.

So, are the allegations correct?

Let’s tackle them one by one.

Is India overstating its real GDP growth rate?

As explained, India’s real GDP is a derived number and depends entirely on what statisticians assume to be the GDP deflator (the proxy rate of inflation during a period).

What is known for sure is that India’s nominal GDP in Q1 of FY24 was 8%. On the face of it, a real GDP growth rate of 7.8% implies that inflation was just 0.2% in the three months — April, May and June.

CHART 1 shows how rare this is. It maps both nominal GDP growth rate (in orange colour) and GDP deflator (in blue) for each quarter (going back all the way to Q3 of 2008.

As consumers, most Indians might find a real GDP growth rate of 7.8% an obvious overstatement. That’s because the retail inflation rate (the inflation consumers face) in these three months was 4.7%, 4.3% and 4.9%, respectively.

If one was to “deflate” nominal GDP using consumer price inflation, the real GDP would fall to less than 4%.

But then if one goes by wholesale inflation, the real GDP will actually be much higher than 8%. That’s because wholesale inflation was negative in all the three months — -0.8%, -3.6%, and -4.2%, respectively.

So where does the truth lie?

The fact is that while either of these — retail or wholesale inflation rates — throw up a stark picture, the effective inflation rate is somewhere in between. Even so, experts had warned about the possibility of real GDP growth in Q1 appearing too rosy long before either the data came out or the Congress made allegations.

In a research note released on August 4 — almost a month before the official data was released — HSBC’s Pranjul Bhandari [Chief India and Indonesia Economist, Managing Director, Global Research] had alerted that something like this may happen.

“The 1Q (April-June) GDP growth data is likely to come in at an impressive 8%. Though some of this exuberance is justified, led by improved terms of trade as commodity prices fell, some of it could simply be statistical. We estimate that issues with the manufacturing and services deflator could overstate GDP growth by 1 ppt (percentage point). And a low base could overstate it further,” stated the note.

In times of such statistical abnormality, experts advise that one must buttress one’s understanding of the economy by looking at other available variables.

“In such a scenario, looking at a host of other activity data may do a better job in gauging the actual growth-on-the-ground,” stated Bhandari in the early August note.

However, this whole controversy is neither without reason nor without a lesson.

The fact is that India’s inflation indices need to be updated. In particular, experts have been repeatedly arguing that the WPI — the wholesale price index — must be discarded altogether in favour of a new Producer Price Index.

Why India needs to rethink the use of WPI inflation

In a detailed conversation, N R Bhanumurthy, VC of Dr B R Ambedkar School of Economics, explained why India needs to rethink WPI.

1> Normally WPI and CPI, which are representative of price pressures in wholesale market and retail markets respectively, are expected to be co-integrated and the causality should run from WPI to CPI. In a sense that changes in wholesale market prices should transmit to retail markets with a lag (as per our estimates in early 2000s, it was one month lag). However, in the recent period such co-integration/causality is broken, thus suggesting that one of the indices is not reflecting the real price and both are not moving together.

2> In terms of coverage, WPI is largely the prices in a segment of the economy as it does not include services. On the other hand, services are covered in CPI. As we now know the share of services in the consumption basket is large and increasing, any price index that excludes the services may not be representative price economy wide.

3> Presently WPI is collected and processed by the Department of Commerce and Industry and largely follows the IIP (Index of Industrial Production) frame, which itself is subject to criticism.

4> Moreover, WPI does not provide information on rural and urban as well as the state level estimates, which are all relevant for public policy.

What about the issue of “discrepancy”? Do they overstate the GDP?

Under normal circumstances, one would assume that there will be some difference between the GDP estimated using income and expenditure methods.

However, if this difference is large, that would reflect poorly on the quality of data and the estimation methodology.

Again, one would expect that the discrepancy level (as a percentage of the total GDP) would be higher in quarterly GDP estimates and lower in annual GDP estimates because with time, more reliable data is available and estimates can be expected to come close to each other.

Some economists such as Ashoka Mody (Visiting Professor of International Economic Policy at Princeton University) who wrote about the “discrepancy” issue in a recent Project Syndicate piece, have raised a question mark on the influence of “discrepancy” in estimating GDP.

“Typically, this discrepancy does not matter for calculating growth rates, because income and expenditure, even if they differ somewhat, have similar trends. But every now and then, the two series follow very different paths, with hugely consequential implications for evaluating economic performance. The Indian National Statistical Office’s latest report is a case in point. It shows that while income from production increased at an annual 7.8% rate in April-June, expenditure rose by only 1.4%. Both measures clearly have many errors. The NSO nonetheless treats income as the right one and assumes (as implied by its “discrepancy” note) that expenditure must be identical to income earned. This is an obvious violation of international best practice. The entire point of the discrepancy line is to acknowledge statistical imperfections, not to make them disappear. The NSO is covering up the reality of anemic expenditure at a time when many Indians are hurting, and when foreigners are showing only a limited appetite for Indian goods,” he wrote in a piece published on September 6.

This is from the official data release and it shows how the two GDP estimates (highlighted in red) are arrived at and what the level of discrepancies (highlighted in green) is.

Since “real” GDP data itself is under question, let’s look at the nominal GDP data and the size of discrepancy. The SCREENSHOT alongside if from the official data release and it shows how the two GDP estimates (highlighted in red) are arrived at and what the level of discrepancies (highlighted in green) is.

As can be seen, the GDP from the income side — calculated by adding Gross Value Added and Net Indirect Taxes — was Rs 7,066,534 crore while the GDP from the expenditure side (without adding discrepancies) was Rs 6,954,807 crore.

In other words, it is true that total expenditures were less than total income. The gap — Rs 111,726 crore — was added to the latter estimate under the heading of “discrepancies”.

But absolute values do not tell anything because with each passing year GDP and discrepancy will keep going up in absolute terms. What matters is the share of discrepancies in total GDP and whether this share has gone up significantly. What also matters is whether there is a trend in discrepancies.

CHART 2 shows that discrepancy level for Q1 is not an outlier.

CHART 2 maps this by taking quarterly data that goes back all the way to Q1 of 2011.

What the chart brings out is that discrepancy level for Q1 is not an outlier. In fact, the impact of discrepancy in explaining quarterly GDP has fluctuated a fair bit in the past as well even though broadly speaking it has stayed within 2%-3% of the total quarterly GDP.

However, when one looks at the share of discrepancies in annual GDP data (SEE CHART 3), two things stand out:

One, the fluctuations from one year to another are smaller than the fluctuations from one quarter to another. This is understandable because better quality data is available over the whole year.

CHART 3

Two, there is a distinct trend. Before 2012-13, GDP from the income method was lower than the GDP from expenditure method — that’s why the discrepancy component was negative. After 2012-13, GDP from expenditure method is lower and that is why the discrepancy component is positive.

Upshot

This is not the first time that experts have raised a question mark on the credibility on India’s GDP estimates. Here’s detailed explainer from 2019 when former Chief Economic Advisor (under the current government itself) had argued that India was overstating its GDP.

The important thing to understand is that the credibility of India’s GDP estimates depends on the quality of underlying data be it the WPI inflation or consumer expenditure or Index of Industrial Production (IIP). Over the past decade many of these databases have not received the attention they deserved. If policymakers do not address the quality of Indian macroeconomic data, the credibility of India’s GDP estimates will continue to be questioned repeatedly.

Share you views and queries on udit.misra@expressindia.com

Until next time,

Udit