Page 119 - Invited Paper Session (IPS) - Volume 2

IPS 188 G. P. Samanta
various statistical and Government agencies, academic institutes, policy
makers and regulators, etc. Each of these alternative resources has been
experimentally assessed by the researchers for different purposes. For
example, Choi and Varian (2009a, 2009b, 2012), Ettredge, et al. (2005), Guzmán,
2011, and Seabold and Coppola (2015) explored the usefulness of Google
search data on tracking/nowcasting various economic activities and
macroeconomic variables; Agarwal et al. (2011) attempted sentiment analyses
based on Twitter messages; Cavallo (2013, 2015, 2016 & 2017), and Cavallo
and Rigobon (2011, 2016) constructed price indices based on online prices and
analysed various aspects of prices; and Cavallo et al. (2015) have studied the
price impact of joining a currency union.
India has a long tradition of model building for analysing and predicting
the inflation rate. In addition to building macro models and simultaneous-
equation based system, researchers usually exploited multiple alternative
techniques with varying degree of success. First, univariate time series models
– both linear and non-linear – and both conditional homoscedastic and
heteroscedastic models. This approach models interrelationship between
current and past observations of time series data on a target variable. Second,
single equation models by regressing inflation on own past as well as present
and past observations of influential variables or determinants of inflation.
Third, multivariate time series models exploiting the interrelationships of
inflation and one or more related variables. These models could be either pure
datadriven, such as Vector Auto Regression (VAR) or model developed
following certain economic principles as could be done in structural-VAR or
Vector Error Correction Models (VECM) under the co-integration framework.
Fourth, the construction of composite leading indicators for tracking Inflation.
Fifth, estimating economic-theory based models, such as P-Star model,
different variants of Phillips-Curve or output-gap models. For past several
years, the Reserve Bank of India (RBI), India's Central Bank, has been
conducting many monetary policy surveys to gauge market expectations on
inflation, growth and other economic parameters. While some of these surveys
capture qualitative responses from the respondents, a few are capturing
quantitative forecasts. As regards inflation, two of these surveys, viz., ‘Inflation
Expectation Survey of Household' (IESH), and ‘Survey of Professional
Forecasters' (SPF) capture quantitative forecasts of inflation. While the target
group of respondents to IESH covers households, the SPF, as the name
2
suggests, captures responses from a select list of professional forecasters .
Though inflation in India has modelled by various approaches in the past,
hardly any attempt is made to assess the information content of online

The SPF captures forecasts for inflation along with many other macroeconomic variables, such
2
as growth, export, import, the exchange rate from professional forecasters.
106 | I S I W S C 2 0 1 9