M Sc

The study entitled “Time series modeling and forecasting of tea price in India”
was conducted to study the components of time series data on prices of tea in India, to
develop time series forecast models for the prices, to develop statistical models for
price volatility and to study the integration between international and Indian tea prices.
Monthly auction prices of tea for North India, South India and All India for the period
from January 1980 to December 2020 collected from the Tea Board formed the main
database for the present study. International price of tea for Colombo (Sri Lanka) and
Mombasa (Kenya) for the period from January 1980 to December 2020 were collected.
To have an idea about the trend in A- Pr- Pd of tea in India, annual data on A- Pr- Pd
of tea from 1970 to 2019 in North India, South India and All India were also collected.
To have a general idea about trend in A- Pr- Pd of tea in North India, South
India and All India, models like exponential, quadratic, cubic etc were fitted. From
among several models tried, quadratic model was found to be the best fit for area under
tea in North India, while, cubic model was found to be the appropriate fit for production
and productivity of tea in North India and, A- Pr- Pd of tea in South India as well as
All India. North India and South India tea price data was decomposed to time series
components like trend, seasonal variation, cyclic variation and irregular variation.
North India and South India showed an overall increasing trend and a prominent
seasonal variation. Cyclic variations showed that South India exhibited more cycle of
price volatility compared to North India. All India tea price was found to be the simple
average of North India and South India tea prices. Compound Annual Growth Rate
(CAGR) was estimated for A- Pr- Pd of tea in North India and South India for the
period from 1970 to 2019. For North India, growth rate in production was more during
1996-2019 compared to period 1970-1995. For South India, a decline in production
was observed during 1970 to 1995.
Price forecast models like exponential Smoothing models and ARIMA models
were fitted to forecast the tea prices in North India and South India from January 2021
to April 2021. For North India tea price, SARIMA (0,1,3)(0,1,1)12 was identified as the
best forecast model whereas for tea price of South India SARIMA (0,1,1)(1,0,1)12 was
selected to forecast tea prices.
For tea prices in North India and South India, volatility in prices were estimated
using intra and inter annual volatility and its significance was tested by fitting suitable
ARCH model. Intra annual volatility indices of monthly tea prices in both regions were
varying irregularly. In most of the years, North India showed large variation in tea
price compared to South India. ARCH (1) model was fitted to check the significance
of tea prices and the estimate of ARCH parameter showed high volatility for tea prices
for North India and South India.
Cointegration analysis was carried out for tea prices to study the integration
between international and domestic Indian tea markets. One cointegrating relationship
exists between the market pairs, North India - South India, North India – Mombasa
and South India – Colombo. No cointegration exist between the market pairs, All India
-Mombasa and All India - Colombo. Unidirectional causality was observed between
South India and Colombo whereas, bidirectional causality was observed between
market pairs, North India - Mombasa and North India - South India.