If accurate economic data was available in a timely and accessible way, India would benefit enormously. Bureaucratic lethargy and arrogance stand in the way.
We live in a world where ‘data’ is the buzzword. Data is considered by many to be the new Oil — a source of riches. Big data and data analytics are attracting tremendous attention from investors as well as researchers. And yet, when it comes to economic data, which is essential for effective policymaking, India is remarkably poor.
Our policies are often made in the dark – we don’t know whether the policies employed are the right ones, and we cannot judge the efficacy of those that we choose to implement. Worse, we often do not even know what is happening in the economy currently. And the problem is not just about the availability of data, though that is an important problem. As I see it, there are four distinct problems regarding economic data in the country.
1. Availability of data
A key problem, of course, is near absence of several important pieces of data. There is a vigorous debate on job creation in the country, but the necessary ingredient for that debate, data on the labour market, is absent. We do not have comprehensive labour market data. The NSSO used to conduct a sample survey every couple of years. The last such survey was six years ago.
While the government has started to release the provident fund data as a proxy for payroll data, given that organised sector jobs are a small fraction of total employment in the country, this data is meaningless in the overall scheme of things. We do not produce data on household income and expenditure, something which is common in most other countries, at even an annual frequency. Similarly, we do not have high-frequency data on the services sector of the economy – a sector that accounts for almost 60% of output. The list of important but missing data points on the economy is long.
2. Timeliness of data
The second problem as I see it is that several of the available data are not available in a timely enough way to be useful from a policy or analytical perspective. Take some high-frequency data – the current account data, which is important for the currency market, and by implication for monetary policy, is released by the RBI 2.5 to 3 months after the quarter has ended. The data for a quarter is thus released almost at the end of the following quarter, by which time it becomes stale. The industrial production data, which is an important gauge of short-term economic activity, is released more than 40-days after the month has ended. The data for August, the month in which we currently are, will be released in mid-October.
In contrast, inflation data is released with a lag of fewer than 2 weeks after the end of the month, and is thus highly useful as it is near ‘real-time’. The latest period for which the highly comprehensive data on organised manufacturing – the Annual Survey of Industries – is available is for the year 2015-16. There is no data yet for the year 2016-17. And we are almost halfway through the year 2018-19.
Take the whole debate over minimum support prices (MSP) or farm loan waivers. The last survey that went into detail over what farmer incomes were, what their debt levels are and of the reach and effectiveness of MSPs was in 2013 – 5 years ago. And the one prior to that was 10 years back. So we have just two data points in the last 15 years.
In an age of computing and automation, it is hard to understand why data processing should take this much time – it should be a matter of hours and days, not months, and certainly not years. A bigger problem here is that while policymakers are at least aware of the problem of lack of data, they do not appreciate the problem of making data available in a timely manner. If data is not available within a reasonable time, it loses its importance.
3. Accuracy of data
The next issue with the official data is the perceived lack of accuracy. This problem is being exacerbated due to standard bureaucratic responses. Take the latest revision to the GDP series a couple of years back. The data as per the new series was counter-intuitive to many, and questions were raised about its accuracy. The data might very well be accurate, but when honest questions are simply ignored and brushed over, this deepens suspicion. There has to be a willingness on the part of bureaucrats to engage with users of data to address their concerns.
Take the survey on household expenditure that NSSO conducts (or used to conduct). It gives very detailed data on household consumption across categories and regions. This is a widely used data set. However, the expenditure aggregated as per this survey is only 50% of the household expenditure reported by National Accounts. This clearly seems wrong prima facie. The Government knows this. But beyond attributing this to methodological issues, there is no attempt to explain the difference in such a way that the data could be then used keeping its limitations in mind. The monthly industrial production data is so volatile and subject to bizarre internal contributions that not many people take the data seriously. And this is not a new issue with this data. This issue is at least 10 years old, if not more. And yet, nothing has been done to fix the issue.
4. Dissemination of data
The website of the Ministry of Statistics and Program Implementation (why is program implementation clubbed with this ministry anyways?) cuts a sorry figure. Even when data is available, is released on time, and is perceived to be accurate, users have a hard time accessing it. There is no unambiguous place on that website where users can go and access all the data that CSO and NSSO publish. In fact, it is hard to figure out what data sets are available on its website.
Users have a hard time accessing historical data as well. Even current press releases are hard to find. There is no distribution list that automatically sends out data releases to users who want them on a real-time basis. For that users must head to the site of PIB. The RBI does a far better job in disseminating data. It has a scheduled publication schedule and a database on its website where users can access all the data it puts out. Most of RBI’s data releases allow the ability to import data directly into MS Excel, so it is easier for users to analyse the data.
None of what I have discussed above is complicated. None of what I have discussed above is political. None of this is about economic ideology. And yet, there has been little progress on this in the last few years. Indeed, it can be argued that things have gotten worse given that we have made no progress while the rest of the world has moved ahead. Fortunately, to date, there is not even a suggestion that official data in India is ‘managed’ for political considerations. The statistics ministry is perceived to be independent. However, it is bureaucratic lethargy and arrogance that is the hindrance.
Let me give one concrete example of where some of the factors discussed above imposed a tangible cost on the Indian economy. In the immediate aftermath of the Global Financial Crisis, the Indian economy saw a sharp slowdown in growth. Inflation, as measured by the WPI (which was the primary gauge back then) also came down. And the RBI responded with a massive monetary stimulus with the repo rate being cut by more than 400bps. By mid-2009, global financial markets had stabilised and even domestically the economy had stabilised. But the CSO was suggesting only a gradual uptick in activity. In early 2010, it projected that growth in the 3QFY10 (October-December 2010) would be around 6% and the full year FY10 growth would be just over 7% only a modest 50bps acceleration from the year before. This was one of the factors that allowed the RBI to continue with its highly accommodative monetary policy. However subsequent data revisions (and base year changes) suggest that the economy saw a V-shaped recovery – the final estimate of growth for 3QFY10 is 7.7% (170bps higher) and that for full year FY10 is 8.6% (140bps higher) and thus in hindsight the RBI monetary policy stance was clearly wrong.
One can only wonder, how much of the BoP crisis of 2013 and the subsequent phase of low growth due to sticky high inflation was due to highly accommodative policy which could have been averted if we had better data.
The private sector will and already is stepping in to fill these gaps, and this will only increase. But the problem is that this data, in most cases, resides behind paywalls, and is not accessible to all. Perhaps that might change. But for now, the availability of private sector data is limited and expensive. In any case, the government cannot absolve itself of its responsibility to generate and provide timely, useful data which is accessible to all. This is necessary for its own sake, for its policy implementation and it is necessary for an informed debate over government decisions or lack of them. It is time the government steps up to this.