By PC Mohanan
There are some new initiatives from the ministry of statistics and programme implementation to enhance survey-based estimates in socio-economic sectors and the delivery of statistical products to users. Base years over a decade old have been used for economic indicators like GDP, index of industrial production (IIP), and consumer price index (CPI), now they are being revised. Survey-based monthly labour force indicators and quarterly sectoral surveys have been instituted; and an improved website-based data dissemination is in place. Above all, the long overdue population census operations are, at last, underway. These are indeed welcome after years of near slumber and silence, especially when its data failed to support the political narrative of the day.
Sub-national GDP: While the steps to update the economic indicators are long overdue, the extent to which these will improve the sub-national indicators begs examination. In particular, the revision of gross state domestic product (GSDP) is critical for the states, as it is crucial in deciding the state borrowing limits and is used by the Finance Commissions in their criteria for devolution of funds. It is often a normalising denominator for fiscal and other indicators for inter-state comparisons. Presently, for estimating GSDP, states are dependent on the National Statistics Office for allocating/apportioning national figures in several sectors. The use of centralised databases like MCA-21 or the use of a commodity flow approach generally excludes state-wise information. Even the presentation of GSDP does not help reveal the structure of a state’s economy. Significant sectors are clubbed together as “other services” while several sub-sectors with negligible shares at the state levels are given following the national practice.
One of the major grievances with the last base year revision of GDP was the increased use of allocation principles for GSDP estimation. For example, the establishment-level data from the Annual Survey of Industries (ASI) was earlier used for the organised manufacturing sector. With the base year revision, data from annual filings of manufacturing companies are used at the national level and then allocated to states using certain indicators. Corporate manufacturing, despite forming only 30% of the units in ASI, accounts for over 80% of the gross value added from this sector. This is then allocated to states. For unorganised manufacturing, certain benchmarking/extrapolation is first used at the national level and then distributed to states on the basis of indicators. The new GDP series and other initiatives will likely rely more on data from e-commerce and other digital platforms. These centralised databases are unlikely to strengthen the sub-national economic indicators.
District domestic product: The estimation of district domestic product (DDP) at present is more of a mechanical exercise using very little district-level data (except for agriculture output) for most of the states.
Proportionate allocation based on outdated population or other means leads to near-identical growth rates for the districts of a state, defeating the very purpose of compiling DDP. The alternative is to build estimates from the bottom and aggregate upwards. GDP arises from economic activities within a defined geographic boundary. At the national level and to a certain degree at the state level, the boundaries of economy are well-defined. At the district level, the use of labour input from household labour force surveys has limitations as it covers households in the district. There is no way of accounting inter-district worker movements to estimate the actual district-level production. Currently, there is no tracking of non-agricultural activities within a district except for the rare economic censuses. The absence of state participation in the current Periodic Labour Force Surveys and the Annual Survey of Unincorporated Sector Enterprises also precludes district-level estimation possibilities using any direct survey data.
It is possibly in this context that states like Uttar Pradesh are attempting to build up DDP and hoping for possible aggregation of these to get GSDP. Given the current state of availability of district-level data, such exercises have to adopt indirect estimation for most sectors. GDP is an aggregation of value addition from every type of economic activity, and the lower we go the greater would be the role of assumptions and apportioning. The resulting aggregate GSDP would be distant from reality, leading to trends that conflict with the central data systems.
Statistical capacity: There is also a question mark on the states’ capacity in undertaking nuanced, data-based exercises. Various schemes and assistance for states to improve their statistical capacity has not led to any standardisation of the procedures or strengthening capacity at the state level. The National Sample Surveys also have a state sample component surveyed by state statistical staff. The pooling of both these data was expected to provide reliable district-level data. But most states do not tabulate the pooled data or even the state sample data. Similar is the case with efforts by states to replicate CPI, IIP, and the ASI that can possibly help produce better district-level indices.
There is also no uniformity in the structure or the functioning verticals within the state statistical organisations. In spite of being declared as the nodal statistical agency in the state, the statistics directorates play a very subordinate role in the administrative hierarchy. There is also an increasing centralisation of administrative data gathering through portals dedicated to various schemes. These data, mostly shown in dashboard formats, have uncertain definitions and coverage, limiting their usability for statistical purposes. Occasionally, rankings of states and even districts using the sustainable development goal indicators or multi-dimensional poverty index framework using data from such sources generate contradictory rankings.
The uneven capacity and importance attached to official statistics at the state level is a serious obstruction to bringing data standards and aggregation. Much-touted databases from goods and services tax, ministry of corporate affairs, or digital sources have issues of access and classification for sub-national use. Under these circumstances, the best option appears to be more extensive and regular survey data with district as a domain by agencies like the National Sample Survey Office.
(The author is the former acting chairperson of the National Statistical Commission.
Views are personal