Today's Main Column : Ila Patnaik
How to read tea leaves
Business Standard, January 09, 2002
The use of leading indicator analysis in India continues to be limited
Last week, the index of activity in American factories showed a sharp jump. At the same time, the index of non-manufacturing activity rose, non-farm payroll employment increased and retail sales of many items showed an upswing. This evidence is interpreted as an indication that the recession in the US is bottoming out. The reason being that they belong to a set of leading indicators for the US economy.
The usefulness of leading indicators for an economy that witnesses recurrent cycles is without question. While on the one hand they are able to warn businesses about the times to come, on the other they allow the government to set in motion policy changes that can help reduce the intensity and length of a recession.
Despite being subject to criticism for the lack of a theoretical basis and being purely data-driven, leading indicator analysis has not died down and the fact that the data continues to be collected and the indicators monitored, shows that they are useful.
The cyclical behaviour of the Indian economy in the 1990s, especially the downturn since 1996, has raised interest in business cycle research in India. One of the earliest studies on the subject was by Vikas Chitre who applied the business cycle indicator approach to India.
He analysed 94 monthly time series and tracked cycles in them. Other studies on the subject include those by Shubhashis Gangopadhyay and Wilima Wadhwa for the ministry of finance, O P Mall of the RBI, A Banjeree and Pami Dua of the Economic Cycle Research Centre, New York and the Delhi School of Economics and an RBI-funded project on business cycle indictors at the NCAER. A special issue of the Indian Economic Review was devoted to business cycle analysis recently, indicating the increasing importance of the theme as a research area in India.
But though leading indicator analysis has proved to be a useful tool in advanced economies like the US, its use in India continues to be limited. The reasons are simple. The first is the availability of data. Data for most of the series that are usually among leading indicators are not available at all or at a frequency not high enough to be useful.
For instance, the series included in the index of leading indicators for the US includes hours worked, average weekly new unemployment claims, new orders for consumption goods and materials, percentage of companies reporting slow deliveries, index of new business formation, contracts and new orders, plant and equipment, index of new private housing units and change in inventories on hand and order. For India, most of the data series have not been collected systematically by any agency — government or non-government.
This leads businesses to rely much more heavily on their own data such as sales and orders as well as some other series they choose to track. Among the conventional series are those like non-food credit and non-oil imports. Some of the unconventional series that are tracked include production of paints, sale of cigarettes and production of textiles — the logic being that these are items on which money is first spent when incomes increase and cut when employment and income fall.
The other reason why the usefulness of this approach to India is limited is related to the nature of the cycles. By definition: “Business cycles are a type of fluctuation found in the aggregate economic activity of nations that organize their work mainly in business enterprises: a cycle consists of expansions occurring at about the same time in many economic activities, followed by similarly general recessions, contractions and revivals which merge into the expansion phase of the next cycle; this sequence of changes is recurrent but not periodic” (Burns and Mitchell, NBER, 1946).
The cycles described here are generated by the internal dynamics of the structure of production and not caused by exogenous shocks to the system. It is in this sense that the usefulness of the approach is limited when analysing the downturns witnessed by the Indian economy.
A review of the studies on the Indian economy mentioned above suggests that at least five periods may be identified as recessionary phases. Though the methodologies used in these studies are different and the frequency of the data is also not the same, some being monthly and some quarterly, at the risk of some imprecision we can identify the years that have commonly been identified as those in which recessions took place.
These years were 1957-58, 1965-66, 1979-80, 1991-92 and 1997-98. In these years, GDP growth declined sharply. In the first three episodes, growth was actually negative. Interestingly, in all the five episodes the growth rate of GDP in agriculture was negative (see graph). Thus, almost any steep decline in GDP growth in India has coincided with a sharp fall in agricultural growth, caused mainly by bad monsoons. In some cases, the rainfall shock was also coincident with an oil price shock.
Since industry was supply-constrained, investment in capacity in the private sector was by licensing decisions, and that in industry and infrastructure in the pubic sector determined by five-year plans rather than in response to market signals, this behaviour makes sense. It is only in the 1990s that cycles may be identified as market-driven. Even the monsoon continued to play a role because GDP growth in agriculture was negative in both the years of recession in the 1990s as well.
The limited availability of data and the fact that cycles that have been observed were largely caused by exogenous shocks limit the usefulness of the analysis of historical patterns for forecasting or policy changes. However, this is not to say that it should be abandoned.
Considering that the economy is increasingly becoming a market economy, the usefulness of the approach will increase over time. Further, it will also help create awareness of the kind of data whose collection and monitoring must be undertaken to obtain early warning systems.
Also, this approach is useful not only in trying to forecast GDP but also other variables such as inflation, exports and financial variables. Indeed, it may be very useful, for instance, for the finance ministry to have a leading indicator for tax collection. As the finance minister recently admitted, tax collections this year are expected to fall short of the budgeted levels significantly.
If the finance minister could get a better idea of tax collections, then hopefully expenditure will be better contained when taxes are expected to fall short. Or, if fiscal deficits are large and fiscal policy is expansionary, it will be by design and not default.
What do you think, Mr Finance Minister, how about a leading indicator for tax collections?