10 Dec 2019 18:03 IST

Higher power output and GDP growth may not be interlinked

In this context, understanding the difference between correlation and causation is crucial

Even before the GDP numbers for July-September 2019 were released on November 29, many macro-indicators were flashing red. The Index of Industrial Production (IIP) was one such indicator that gave advance warning to economists that contraction in industrial activity between July and September would impact GDP growth. What the IIP numbers were pointing to was eventually confirmed.

One element of the IIP index is electricity generation, which has been tepid since early 2019. This was said to be a leading indicator of a slowing Indian economy.

A recent study by CARE Ratings looked into the causality of a drop in electricity production and the slowdown in economic growth. But before we get into the findings, let’s look at the power production data.

Electricity generation slows

Electricity generation increased marginally in April-October 2019 to 757.95 billion units, or by 1.2 per cent over the same period the previous year. This was also seen in the IIP numbers as electricity production fell under the new IIP (with 2011-12 as base) in August 2019. Many people linked it directly to the industrial slowdown, and called it a precursor to weak GDP numbers for the Q2 of FY20. IIP in August was -1.1 per cent.

Many auto companies shut down their plants due to excessive inventory with the dealers in the September 2019 quarter (Q2). Some of this might have been reflected in the IIP numbers. But the point to note here is that only 40 per cent of domestic electricity demand comes from the industrial or manufacturing sector. Also, bountiful rains did lead to a fall in thermal power generation which, being over 80 per cent of India’s power generation capacity, did reflect in the monthly electricity generated.

Power generation at listed power producers such as JSW Energy, NTPC and Adani Power also saw a dip in the quarter ended September. At NTPC, generation dipped 6.6 per cent YoY in the quarter ended September to 61.6 billion units. JSW Energy’s growth in the same quarter was flat, at 6.7 billion units, compared with 6.67 billion units. The minuscule rise in production was because the company sold more power from its Ratnagiri thermal plant to short-term customers. Adani Power’s total electricity sales volume was 13.6 billion units (14.6 billion units).

But is there a direct correlation between the dip in power output and economic activity? Let’s now look at the CARE Ratings study.

Weak correlation

The CARE Ratings study, authored by Madan Sabnavis and Manisha Sachdeva, found that when quarterly power production is compared to manufacturing activity, the link (or correlation) between the two data sets is very weak.

The study looked at the correlation between electricity production and GDP growth. It found that although there is a link between economic activity and power generation, the correlation is, again, weak. In addition, the study said that: “Factors such as improvement in production processes (in terms of new technologies) and energy efficiency makes comparison between electricity production and economic growth challenging.”

Then, there is the fact that manufacturing makes up only 18 per cent of the economy. And this sector consumes 40 per cent of the electricity generated in India. So, there is obviously an impact on the economy if electricity consumption goes down.

To put things in perspective, if consumption goes up, it is not because manufacturing activity has increased. The Centre has spent a lot of money to take power lines to the hinterland and claims to have achieved 100 per cent success in bringing power to all Indians. When these consumers, who were earlier off-grid, start consuming more electricity in the future, it would be wrong to say that GDP should go up.

In the future, if electricity generation in the country increases or decreases, as reflected in the monthly IIP number, it will not directly mean that GDP will rise or slow down. That’s almost like putting the cart before the horse. There is a correlation between the two measures, but that doesn’t mean that one causes the other.