The Hindu Business Line : Statistics as farce

Given the mismatch in their numbers, the methods of the two national survey organisations need to be fine-tuned to reflect accurate economic aggregates.

Statistics can be misleading if not absurd, particularly when they pertain to the Indian economy. Estimates of monthly per capita consumption expenditure (MPCE), drawn up by the National Sample Survey Organisation (NSSO), are a case in point. The average MPCE for rural India increased by 72 per cent between 1993 and 2000, but only by 13 per cent between 2000 and 2005. In urban India, the average MPCE rose by 86 per cent between 1993 and 2000 but just 23 per cent between 2000 and 2005. Besides, per capita rural and urban consumption was almost unchanged between 2003 and 2005. These figures collectively point to a recession between 1999-2000 and 2004-05, whereas the average annual rate of growth of the economy in the two five-year periods was 6.8 per cent and 6 per cent, respectively.

Economic aggregates — such as gross domestic product, fixed capital formation and consumption expenditure — are derived from National Accounts Statistics (NAS). NAS data tells us that private final consumption expenditure contributed to more than half the growth in gross domestic product in almost all the post-reform years. Clearly, there is a serious mismatch between NAS and NSSO estimates; the latter is estimated to be 25-30 per cent lower than the former. However, the accuracy of the NAS is not a given either, as approximations are made for unorganised and service sector output. The per capita MPCE for the whole country is about Rs 750, which is nowhere near the per capita income derived from NAS estimates of about Rs 2,750 per month, even after one accounts for savings and a degree of underestimation. Factors as crucial as the size of the economy and per capita income are grey areas.

Such discrepancies are too serious to be brushed under the carpet. Consumption is under-reported for two reasons: the NSSO questionnaire is unable to keep pace with the introduction of a range of goods and services and the recall period for many of these items is too small. Therefore, a 30-day recall may not be appropriate for clothing, footwear, mobile phones and other consumer durables. To address this shortcoming, the 55th Round introduced a ‘mixed recall period’ (MRP) for different sets of items. Even as some economists feared an overestimation of consumption, the average MPCE derived from uniform and mixed recall periods has turned out to be almost the same. Yet, MRP throws up lower poverty estimates, implying that the recall period is playing a crucial role. The URP underestimates expenditure on non-food items and, therefore, the shift from essentials to others with a rise in income. Both the NAS and NSSO methods need to be fine-tuned so there is no uncertainty over basic economic aggregates.

Related Stories:
Starvation threat looms as food, fuel prices shoot up
Industrial growth slumps to 3.8% in May
Inflation haunts consumers more in rural areas
Per capita consumption in rural India rises