What is happening to consumption in India?

On 4th June 2020, National Statistical Office (NSO) – India’s official Statistical agency, released Periodic Labour Force Survey (PLFS) Annual Report for the year 2018-19. PLFS gives us critical information such as worker population ratio, labour force participation rate, unemployment rate etc. This second PLFS release comes as a breather as it shows the unemployment situation in 2018-19 has improved vis-à-vis 2017-18. It has come down from 6.1% in 2017-18 to 5.8%. The Labour Force Participation Rate (LFPR) has also improved from 36.9 to 37.5 so also Worker Population Ratio.

This is quite encouraging especially given the fact that real GDP growth has declined from 7% (2nd revised estimate) in 2017-18 to 6.1% (1st revised estimate) in 2018-19. Decoupling of labour force participation rate and GDP growth rate is not unusual if the workforce participation rate increases in sector where output growth is slow. One needs to deep dive to understand the full dynamics.

Every survey has a primary purpose. PLFS’s primary purpose is to understand labour dynamics. Therefore, any inference on other aspects are incidental in nature.

I am also refraining from doing any comparison with NSSO Consumer expenditure survey (of 68th round, 2011-2012). NSSO consumption survey was much more granular. Sampling procedure was quite different. Therefore, any comparison would not yield much result.

It is important to note that we have no reference point to have a temporal comparison on what is happening to consumption. The only reference point we have is PLFS survey of 2017-18. In a span of one year, I do not foresee any major changes to happen.

Hence, we are left with only one dimension that is cross sectional dimension. Here cross-sections would mean inter-state comparison. One can go deep by looking at social strata, religion, or comparison with employment situation, sector etc. But that is for a longer piece. Hence, I am not also discussing that.

I will just focus on a few pictures so that you can have a broad sense on what is happening in consumption expenditure.

Consumption story:

There is little we know about what is happening to household consumption expenditure in India post 2011-12. The last round of consumer expenditure survey was conducted as part of the 68th round of the NSSO in 2011-12. Given its quinquennial nature, a survey was conducted between July,2017 to June 2018. However, due to certain data quality issues, it was not released as per PIB press release dated 15-Nov-2019.

There was a silver lining in the press release. It said – “In view of the data quality issues, the Ministry has decided not to release the Consumer Expenditure Survey results of 2017-2018. The Ministry is separately examining the feasibility of conducting the next Consumer Expenditure Survey in 2020-2021 and 2021-22 after incorporating all data quality refinements in the survey process.”

It is highly unlikely that a survey would be conducted in 2020-21 given the Covid-19 situation. We may have to wait for a longer period to have a sense of what is happening to household consumption expenditure.

Why it is so important to understand consumption expenditure? It is simply because there is no better yardstick to measure well-being and progress than consumption expenditure. The PIB press release put it nicely by saying:

“The NSS Consumer Expenditure Survey generates estimates of household Monthly Per Capita Consumer Expenditure (MPCE) and the distribution of households and persons over the MPCE classes. It is designed to collect information regarding expenditure on consumption of goods and services (food and non-food) consumed by households. The results, after release, are also used for rebasing of the GDP and other macro-economic indicators.”

Now let’s focus on PLFS survey that was released on 4th June 2020.

In the entire questionnaire of PLFS, in Block-3 of ‘Instructions to Field Staff-Vol II’, there is one point under ‘Household Characteristics’. It is “household’s usual consumer expenditure (Rs.) in a month”. You must refer to Page-A10 of Instructions to Field Staff, Vol. I of PLFS to understand how this question was administered.

I am reproducing the description ‘as it is’ from Page A-10 of Vol-I:

“Household’s usual consumer expenditure (in Rs.) in a month: This may be ascertained as follows. The question “What is your usual expenditure for household purposes in a month?” will be put to the informant. Suppose the answer is Rs. A. Next, the purchase value of any household durables (mobile phones, TV sets, fridge, fans, cooler, AC, vehicles, computers, furniture, kitchen equipment, etc.) purchased during the last one year will be ascertained and the expenditure per month obtained by dividing by 12. Let this be Rs. B. Further, it should be quickly ascertained whether there is (usually) any consumption from (a) wages in kind (b) home-grown stock (c) free collection. If so, the approximate monthly value of the amount usually consumed in a month will be imputed. Let this be Rs. C. Then the sum of A+B+C is to be entered as usual monthly consumer expenditure of the household in whole number of rupees.”

Let’s now look at the data. There are eight charts and I will tell the story using those eight charts. Rather you can make out your own story out of these eight charts.

I have made these charts using PLFS Unit level data available in MoSPI portal. For students - MoSPI gives the data in text format hence you can write your own dictionary file using any of the available software to make those text file usable. Use the multiplier suitably so that you are not talking about the sample, rather your inference holds good for entire population. Also, while working on household file, make sure you are using the household size to talk about things at the individual level.

Figure 1 shows the Monthly Percapita Consumption Expenditure (MPCE) for the year 2019-20 for 20 States.[1] It shows that Delhi has the maximum MPCE of Rs. 3,893 while Bihar has the lowest MPCE of Rs. 1346 amongst 20 states. While an ordinary person on an average is able to spend Rs. 3,893 in Delhi, his counterpart in Bihar could spend only Rs. 1346 per month!

The top 5 states in terms of MPCE are – Delhi, Punjab, Kerala, Tamil Nadu, Haryana while the bottom 5 states are – Bihar, Chhattisgarh, Odisha, Jharkhand, and Uttar Pradesh.

Figure 2 shows the distance of consumption – How far the states stand if we take Punjab as the reference point? Punjab’s rank is 2nd out of these 20 states in MPCE . As you could see – if people are consuming Rs. 1 in Punjab, it is Rs. 0.43 in Bihar and Chhattisgarh, Rs. 0.48 in Odisha and Jharkhand, Rs. 0.50 in UP and Rs. 0.51 in MP. Delhi in similar scale has Rs. 1.24, Kerala – Rs. 0.98, Tamil Nadu – Rs. 0.92, Haryana – Rs. 0.87 and AP – Rs. 0.74. 

Is the distance between urban and rural area shrinking in terms of consumption expenditure? Figure 3 suggests that – not at this point. The orange shaded part is for Urban MPCE and the blue color part is rural MPCE. As you could see, there is a significant gap between urban and rural MPCE. Urban people are consuming much more than their rural counterpart. Unfortunately, we do not have item-wise details so we will not be able to make comments on which goods and services. I understand the price situation what urban people face is very different from rural people. Therefore, there would be a variation between rural and urban. Still question remains - why there is so much interstate variation? Why the economically weaker states have much larger gap than rich or median income states?

So, the next question is – How much more urban people consume vis-à-vis their rural counterpart? Or How much is the gap between urban and rural consumption? As you could see in Figure 4, in Jharkhand, the Urban MPCE is 2.27 times higher as compared to rural area. The top 4 states where the urban and rural disparity is maximum are Jharkhand, Odisha, Maharashtra, and Delhi. However, it is the least in Kerala, Punjab, Bihar, and Andhra Pradesh.

Figure 5 shows the relationship between per-capita income i.e., Per capita NSDP at current price and MPCE. I have taken per capita income for 2017-18 while I should have taken it for 2018-19. Since MoSPI data does not have figure for Maharashtra for 2018-19 and since I did not want to use projected figure or to collect data from other sources, I have used the previous year figure. However, my sense is that – relative ranking would not change dramatically over one year.

How to read the graph? This graph has plotted two ranks – rank in Per capita income and rank in MPCE. For example – Delhi has 1st rank in per capita income. Since the states are arranged by their rank in per capita income, Delhi appears as the first state with an adjacent MPCE rank bar. If the height of the two bars i.e., per capita income and MPCE are same, then that would convey that the State has same rank in case of both the indicators. But, as you could see, there are a lot many states whose MPCE rank is quite far away from their per capita income rank. What does it reflect? At a very macro level, it indicates the asymmetry in relative distribution of income and hence consumption. The rank correlation between per capita income rank and MPCE rank is 0.753 which is remarkably high. It suggests that state per capita income holds the power to explain a part of consumption ranking. 

Figure 6 shows the rank difference between per capita income and MPCE. For example – Delhi’s rank in per capita income and MPCE is 1. Therefore, the difference is zero. At the same time, Uttarakhand’s rank in per capita income is 3 while it’s rank in MPCE is 11. Therefore, 3 minus 11 is minus 8. If it is in the negative axis, it reflects that while the state may be doing good in terms of overall GSDP and hence per capita income, it is not doing so well in consumption expenditure ranking. Look at Tamil Nadu. It’s per capita income rank is 9 while it’s MPCE rank is 4. Therefore, it is in blue color and in positive axis. 

Figure 7 shows the ranks of the State as per ‘Business Reform Action Plan – 2017:18’ score. Since I have taken 20 states out of 36 states/UTs for which BARP statistics was available, I have recalculated the relative rank of the States. States are arranged by their BARP reform ranking score. For example – Andhra Pradesh has rank 1 in BARP score while it ranks 6th in terms of MPCE. At the same time, while Delhi ranks 20th in BARP, it’s rank in MPCE is 1.

As you could see, there are many states for which there is large asymmetry between their BARP rank and MPCE rank. This suggest business reform and MPCE somehow became disconnected. The rank correlation between MPCE rank and BARP rank is -0.194. It reflects that the business reform rank and consumption rank are not correlated at all. Will take some more time before it gets reflected in consumption ranking?

Figure 8 shows difference in rank between BARP score and MPCE. For example, in case of Delhi, while its rank in MPCE is 1, it ranks 20th based on BARP score, therefore the difference was 19 (20-1=19). At the same time, Jharkhand has BARP rank of 4 while its MPCE rank is 17. Therefore, the difference is -13 (4 -17=-13). There are many states for whom the business rank is far better than the consumption rank. When it will reflect in consumption ranking? Isn't that the ultimate goal? 

So let's end this with a question : Will the structural reform bring much desired distributional changes and hence a change in consumption level?

I hope so. When? How long it would take? Sooner the better.

_____________________

[1] I have not taken these states into account for this piece- Jammu & Kashmir, Himachal Pradesh, Chandigarh, Sikkim, Arunachal Pradesh, Nagaland, Manipur, Mizoram, Tripura, Meghalaya, Daman & Diu, D & N Haveli, Goa, Lakshadweep, Puducherry, A & N Island.