CSO decided to shift the basis of manufacturing estimates to businesses' balance sheets and income statements from exclusive use of indicators such as production and input use from the Index of Industrial Production (IIP) for first estimates and checked subsequently from the Annual Survey of Industries (ASI). The new numbers did not apparently pass consistency checks with production, inputs, or movements in the National Stock Exchange (NSE) index that should reflect the overall business and investment climate. Even if one ignored the NSE-listed companies, non-NSE firms could not account for the newfound spurt in manufacturing.
On the other hand, it would seem a safe principle to use what the productive sector itself reveals about their business outcomes. For one thing, CSO expanded its sample from ASI's coverage of 200,000 companies to the Ministry of Corporate Affairs' 500,000 accounts declared by companies themselves. Thus, not only should the confidence level for the new numbers be higher reflecting the much larger sample, but using the companies' own declarations to the government must bring the data closer to the true picture of the economy.
More From This Section
For example, to gauge tax evasion, advanced country tax administrations routinely check company accounts to gauge potential tax revenue and compare them with actual collection to estimate the tax gap. So why should CSO not adopt a similar approach? Indeed it has been found that what companies reveal in their balance sheets and profit-loss statements to shareholders differs from how they calculate them for the tax authorities where the numbers are less optimistic, thus yielding lower potential tax to be collected. No doubt there is some reason why tax administrations including India attempt to narrow the gap between the two sets of accounts.
Further, CSO's shift to using companies' income statements and away from production brings the estimation method closer to that of the services sector though, for the agricultural sector, a production method based on observed crop surveys, will continue to be used.
Analysing the old and new GDP reveals interesting differences in approach. First, CSO's new approach essentially implies that not only is value of production of a business included in its new manufacturing GDP, but also the cost of marketing and selling. In other words, distribution costs revealed in company accounts have been clubbed with manufacturing. Expectedly, GDP from manufacturing, mining and quarrying have gone from negative to significantly positive due to expanded scope of manufacturing. But this should have reduced the scope of corresponding services though, other than (1) real estate, financial and business services, and (2) electricity, gas and water, every other service has shot up including transport, storage and communication. The final outcome seems to indicate manufacturing gain has overwhelmed services loss.
Second, the new series uses the concept of value added or income or GDP directly rather than beginning with production and removing inputs. To explain, if the value of production of every economic activity is measured and aggregated, that yields gross "value" of the economy's production. Inputs that a sector has used from another sector's production have to be netted out across the economy to arrive at the economy's "value added" or GDP. Ideally, an economy's input use for each output can be represented in an input-output (I-O) table. It is challenging to produce one. I witnessed the tough process in Mexico in 1985-87 which had a UN-financed project to generate a new GDP series through an I-O method when I was attempting to measure potential tax revenue from Mexico's value added tax (VAT) using the I-O method. The intricacies of calculating sector by sector value added posed themselves on a daily basis in the deeper reaches of the tax office. I also recall Professor Sukhamoy Chakravarty expressing surprise that India did not possess an I-O table when he went to the Planning Commission in the early 1970s, resolving that he would generate one. In sum, the "production method" remains tough to adhere to even in the long run. CSO has opted to move from the production method to an "income method" of GDP derivation.
Third, after the derivation, indirect taxes need to be added in and production subsidies subtracted, so that only the exact value added generated by a producer in the production-distribution chain is taken and nothing else. Fourth, perhaps CSO's inflation adjustment has had the highest impact on new GDP. By assuming lower inflation for 2014-15, when CSO divided nominal GDP by inflation, it obtained a higher real GDP growth rate. This imparted an immediate boost to the GDP number that seemed to co-mingle the issue of change in methodology with a ceteris paribus revision in estimation.
Looking forward, reform of the GDP series has to continue in particular in the services sector. Government salaries directly comprise services sector GDP whose any increase adds to GDP. Lamentably, GDP has little to do with productivity. Since GDP reflects any output of the economy without consideration of its morality or legality, output from illegal or immoral activities get included provided they can be measured. Thus, trafficking of bonded or child labour adds to the services sector GDP if their transportation is measured and reported. Incredulously, measured activities related to murder or suicide would also get reflected. Escalating up, activities from social conflict comprise GDP.
Given that such condemnable activities generate GDP, it is conceivable that as an economy gets riddled with social or political conflict, GDP might keep creeping up as long as such activities somehow are measured and documented. Is it any wonder that the value added or GDP of India's services sector is burgeoning in the backdrop of the rapidly deteriorating social environment, or that anthropologists and sociologists view economists with derision? Finally, reform of the GDP series should address how it can measure the correct extent of economic activity in the black, or non-tax paid, economy and then include it comprehensively in the series.
To conclude, while it may sometimes be useful to confuse oneself about an issue before trying to understand or solve it, in the case of GDP, the confusion doesn't seem to dissipate. How come the world depends on it as an exclusive tool to measure growth and development? I strongly recommend its thorough re-examination to reduce its irrationality and perceived inconsistency. In the present, more limited event, CSO should report the old with the new series for five, if not 10, years so that their underlying differences are transparent on a continuing basis and they are able to serve the needs for research and statistical use. Otherwise, econometric work would have to continually use a dummy variable at the break point for analysis using GDP time series, certainly a potential source of frustration.