India is one of the leading producers and consumers of vegetable oils in the world. The integration of ’India’s edible oils markets with international oil markets (Rotterdam market) is studied with the overall objective of establishing long-run relationship and direction of causality. Keeping in view of the quantum of arrivals, five major domestic wholesale markets and one international market each for groundnut, soybean, and sunflower were selected. Johansen’s cointegration test revealed the prevalence of long-run relationships across the markets. In the case of groundnut oil, Rotterdam market prices are influenced by only Delhi market, whereas all selected domestic markets influence the latter. The results of causality in soybean markets confirmed a unidirectional relationship between all the domestic markets with the international market except Jaipur market, which has a bidirectional relationship with the international market. Hyderabad and Vijayawada sunflower market prices influenced the international market. The suggested policy intervention is to strengthen market intelligence for farmers by establishing online market analysis and dissemination system. The development/strengthening of market infrastructure, including communication, transportation, and storage networks, is mandatory to fully integrate the markets.
Highlights
Market integration occurs when prices among different locations exhibit similar patterns over an extended period.
The integration of ’India’s edible oils markets with international oil markets (Rotterdam market) is studied with the overall objective of establishing long-run relationship and direction of causality.
The development/strengthening of market infrastructure, including communication, transportation, and storage networks, is mandatory to fully integrate the markets.
Vegetable oilcointegrationpricemarketscausality
Oilseeds sector is one of the sunrise segments of world production, consumption, and income earnings of farmers for the past four decades. These crops are considered the building blocks of rural economy. It is imperative to understand edible oilseeds market linkages to sustain the oilseeds production achievements attained during early ‘1990’s by way of ”Yellow Revolution“. Oilseeds have become one of the backbone crops of the agricultural economy of India since independence.
Demand projections of vegetable oils in India by the terminal year of XII Plan (2017) were made by different agencies/researchers in the recent past, which is likely to be at least 16 kg/year per capita. However, the actual per capita oil consumption has surpassed 19 kg per annum for the same year. India is one of the leading producers and consumers of vegetable oils in the world. Edible Oil consumption is somewhat higher in Western India and lesser in Southern India, albeit, it is more or less proportional to the population distribution. On a comparative basis, palm oil is not much favored by North India. In contrast, South India prefers sunflower oil and is less inclined towards soybean and mustard oils.
How to cite this article: Nayak, A., Lokesha, H., Gracy, C.P. and Mahadevaiah, G.S. (2021). Is a World Price Influencing Indian Vegetable Oil Market? Evidence from Historical Prices. Economic Affairs,66(2): 207-215.
Source of Support: None; Conflict of Interest: None
The country’s consumption has been increasing due to changed food habits, affordability, and the raising percentage of working middle-class urban population percentage. Solvent Extractors Association of India reports that import of vegetable oils during November-December 2020 at 2459 thousand tonnes was about 2 lakh tonnes more compared to the same period in 2019. Karnataka, Andhra Pradesh, Tamil Nadu, and Uttar Pradesh are the oilseed bowl of India. Among the nine significant oilseeds cultivated in India, Karnataka is a leading producer of sunflower, second position in safflower, third position for sesame, and fourth in groundnut crop. Karnataka is the ninth and sixth most extensive state in production and area of oilseeds crops in the country, respectively, with a productivity of 824 kg per hectare (Nayak et al. 2020). Although oilseed production is concentrated in a few states, consumers are distributed across urban and rural areas. Therefore, price dissemination and feedback are essential for market price discovery in the spot market. The market integration is researched for oilseed crops to understand the pattern of relationship prevailing in important leading markets.
Market integration occurs when prices among different locations exhibit similar patterns over an extended period. When markets are integrated, a given change in the price of one market could help predict prices of other markets. Thus, market integration explains how different markets are related to each other concerning price of a commodity or related commodity. If prices in two markets converge, it shows the degree of price transmission and the speed at which information travels between two markets. Well-integrated markets follow ‘Law of one price’ where in the difference in prices is equal to the commodity’s cost of transportation from one market to another (Nayak et al. 2020). The literature on cointegration techniques, which concerns the market integration of agricultural commodities, especially about oilseeds in India (Akshata et al. 2013; Gracy et al. 2013; Sundaramoorthy 2014; Sangeetha et al. 2017; Nayak et al. 2020) reveals presence of perfect market integration and price transmission are crucial for efficient management of marketing system.
Considering the above issues in view, an effort has been made in this paper to analyze the integration of edible oil markets at both national and international level with the overall objective to check the longrun relationship and short-run relationship and the direction of causality among selected vegetable oil markets. Considering India as a major importer and consumer of edible oils, the price behavior of domestic and international reference markets is worth researching. Such studies help in identifying lead markets for framing suitable edible oil import policy of India.
MATERIALS AND METHODS
The study uses time-series data on prices of groundnut, sunflower, and soybean oil in domestic and international markets. The markets selected for the study from India were Chennai, Delhi, Hyderabad, Mumbai, and Rajkot for groundnut, Hyderabad, Bengaluru, Jaipur, Mumbai, and Bhopal for soybean, Bengaluru, Chennai, Hyderabad, Nagpur, Vijaywada for sunflower and one international market for each crop. Monthly price data for selected domestic markets were collected from the website of NIC and the international prices from the Global Economic Monitor (GEM), popularly known as the pink data sheet of the World Bank for Jan- 2009 to Feb- 2020. Various statistical/time-series analytical techniques, namely ADF unit root test, Johansen’s cointegration test, and Granger causality test method, were employed to study the market integration.
Steps in Co-integration AnalysisCheck for stationarity
The static data is the one that has a basic statistical property of constant mean and finite constant variance. The stationarity test is based on the Dickey-Fuller value statistic of β1 given by the following equation:
ΔPt=β0+β1Pt−1+∑k=1NδkΔPt−k+ηt
Where, ∆Pt = P1t – P1t-1
The test statistic is simply the t statistic. The values obtained can be compared with critical values given by Dickey Fuller table. For example, in estimating equation (1) the null hypothesis is Ho: Pt is I (1), which is rejected [in favour of I (0)] if β1 is found to be negative and statistically significant, the above test can also be carried out for the first difference of the variables.
Δ2Pt=θ0+θ1ΔPt−1+∑k=1NΦkΔ2Pt−k+μt
Where the null hypothesis is Ho: Pt is I (2), which is rejected [in favor of I (1)] if θ1 is found to be negative and statistically significant.
If the data series is non-stationary, make it stationary
If the given data series is already stationary, i.e., if I(0) for both the series, then we say they are not co-integrated; if not, make the data stationary by differencing. Test the differenced series for stationarity by repeating the above step.
Determine the order of integration
A series, which becomes stationary after first differencing, is said to be integrated into order one and expressed as I(1). Generally, a series may have been differenced ‘d’ times to become stationary in which case it is termed as I(d). A major difference between I(0) and I(d) series is that the I(0) series has a finite mean and variance, while in the I(d) series, these magnitudes do not exist. Thus, a differenced series has properties such as mean, standard deviation, and co-variance invariant with time.
If the order of integration is the same for both the series i.e., Pt ~ I (d).for ex: if Pit (2) and Pij (2), then test for Co-integration. If the integration order is not the same for the two series, i.e., Pit(1), Pij(2) then it is concluded that the series is not co-integrated. Having established that the variables are stationary at level, we may then test for cointegration.
Test for cointegration
The Engle-Granger two-step method was used to test for co-integration between the variables. Johansen’s Co-integration technique was used to test the long-run relationship.
Engle-Granger methodology
This methodology is based on OLS regression. It is most suitable for bivariate settings where the choice of the dependent variable is not a question and can identify only one cointegration vector. This is a residual-based cointegration test. It seeks to determine whether the residuals of the equilibrium relationship are stationary i.e. β’xt = et. Is et stationary? This is established through the Augmented Dickey Fuller (ADF) test on residuals of the co-integrating regression results.
Procedure adopted for cointegration analysis
Step 1: Pre-test the variables for the presence of unit roots and order of integration. If price series in both markets are stationary, it is unnecessary to proceed since standard time series methods apply to static variables. On the other hand, if the variables are integrated of different orders, it is possible to conclude that they are not co-integrated in the usual sense of the term.
Step 2: Estimate the long-run relationship. If the results of step 1 indicate that both yt and zt are I(1), the next step is to estimate the long-run equilibrium relationship in the form,
yt=β0+β1zt+et
run OLS and save the residuals. When yt and zt are co-integrated OLS regression yields a consistent estimator of the cointegrating parameters β0 and β1. The OLS estimates of β0 and β1 converge faster than in OLS models using static variables (Stock, 1987).
Step 3: Test the residuals to determine if the series are co-integrated in a real sense. These residuals are the estimated values of the deviations from the longrun relationship. If these deviations are found to be stationary, the yt and zt sequences are co-integrated of order (1, 1). It would be convenient to perform Dickey-Fuller test on the residuals to determine the order of integration. Fit the model
Δet=α1et−1+ϵt
The null and alternate hypotheses are,
H0: α1 = 0
H1: α1 ≠ 0
The parameter of interest in equation (4) is α1. If the null hypothesis α1 = 0, is not rejected, it could be concluded that the residual series contains a unit root. Thus, the yt and zt sequences are not cointegrated. Instead, the rejection of null hypothesis implies that the residual sequence is stationary. Given that yt and zt are found to be I(1), and the residuals are stationary, it is concluded that the series are co-integrated of order (1, 1).
Step 4: Estimate the Error Correction Model. If the null hypothesis is rejected in Step 3, the residuals from the equilibrium regression (yt = β0 + β1zt + et) can be used to estimate the Error Correction Model. If yt and zt sequences are co-integrated of order (1, 1), the variables have the error correction form,
Δyt=α1+αy(yt−1−βzt−1)+∑α11(i)Δyt−i+∑α12 (i) Δzt−i+εytΔzt=α2+αz(yt−1−βzt−1)+∑α21(i)Δyt−i+∑α22(i)Δzt−i+εyt
Where, βi = the parameters of the co-integrating vector given by equation (3).
εyt and εzt = White noise disturbances.
α1, α2′, αy′, αz′, α11(i), α12 (i), α21 (i) and α22 (i) are all parameters.
The items in parentheses are the error correction terms.
Establishing the long-run relationship
Johansen (1988) has developed a multivariate system of equations approach. The long- rum run relationship between the price series is estimated through Johansen co- integration model. The test shows whether the selected vegetable oil markets are integrated or not. This test allows for simultaneous adjustment of more than two variables. Only when two series are integrated can there be a feedback mechanism of price information and market price discovery.
RESULTS AND DISCUSSION
Before analysing cointegration, it is necessary to check the univariate time-series data generating process to examine whether the series under study exhibit a standard stochastic dynamic process. This was analyzed by employing the ADF test, and results are presented in Table 1. The null hypothesis of non-stationarity was tested based on the critical values given by MacKinnon. The result showed that all the market prices had unit roots and concluded that all the price series are non-stationary at their level, and the data became stationary after the first differencing. This implies that all the markets are integrated of the same order I(1) and thus, share a standard long-run dynamic process. These findings are supported by Sangeetha et al. (2017) pertaining to observation of groundnut markets integration.
Augmented Dickey-Fuller tests for selected oil markets
Johansen’s Maximum Likelihood Test (trace test) results are shown in Table 2. The trace test statistics value results showed that test statistic value 99.79 was greater than the MacKinnon table value (95.75), which indicated the presence of at least one co-integrating equation at five percent level of significance. This implies that there exists a long-run dynamic equilibrium relationship among the selected markets of groundnut. Therefore, any price shocks in these selected markets would be transmitted across the other markets.
Johansen’s Co-integration Test for Selected Groundnut Markets
Granger causality test for different groundnut markets
A Granger causality test was also performed across the groundnut markets; the results of which are given in Table 3 and Fig. 1. According to the Granger causality test, there were bidirectional causalities between Mumbai-Chennai, Hyderabad-Rajkot, Delhi-Hyderabad and the former market in each pair granger causes the wholesale price formation in the latter market, which in turn provides the feedback to the former market as well. The remaining market pairs showed unidirectional causalities, meaning that a price change in the former market in each pair granger causes the price formation in the latter market. In contrast, the price change in the latter market is not flowed back for the price change in the former market in each pair. Delhi market didn’t influence any domestic market but was influenced by all other selected markets at the same time influencing the international market (Rotterdam price). These findings agree with Venujayakanth et al. (2017), who observed the integration between three groundnut domestic markets.
Johansen’s Co-integration Test for Selected Soybean Markets
Granger causality test for different Soybean markets
Johansen’s Co-integration Test for Selected Sunflower Markets
Causal relationship among major groundnut markets under study
The, Johansen’s cointegration test has shown that even though the selected wholesale soybean markets are geographically and spatially isolated, they are well-connected in terms of prices of soybean, revealing the presence of long-run price linkages among the soybean markets.
Causal relationship among major soybean markets understudy
After finding cointegration among different soybean markets, Granger causality was also estimated between the selected pairs of soybean oil markets. The Granger causality shows the direction of price formation between two markets. The results are presented in Table 5 and visualized in Fig. 2. According to Granger causality test, there are unidirectional causalities between the selected soybean oil market pairs: Bengaluru market of Karnataka Granger cause price formation in Hyderabad and Bhopal markets in Madhya Pradesh whereas Hyderabad market Granger cause price formation in Bhopal market. Mumbai market Granger cause price formation in Hyderabad and Jaipur markets, whereas Jaipur market Granger cause price formation in Bhopal market. A unidirectional relationship between all the domestic markets with international markets except the Jaipur market, which has a bidirectional relationship with international market.
The integration relation between the wholesale prices of selected sunflower markets and the relationship between wholesale prices of selected sunflower markets was examined and presented in Table 6. The results revealed the presence of four co-integrating equations at a five percent level of significance. Hence, it is concluded that longrun equilibrium exists among selected sunflower markets.
Causal relationship among major sunflower markets under study
As a part of cointegration analysis, Granger Causality test was conducted to examine whether there was a causal relationship between the cointegrated markets as revealed by Johansen’s test. The causal relationships among major sunflower oil market prices were approached through Granger’s causality technique. The results depicted in Table 7 and the direction of causality in Fig. 3 revealed bidirectional and unidirectional relationships among the selected sunflower oil markets. The results confirmed bidirectional causalities between the sunflower oil market pairs: Vijaywada-Chennai and Bengaluru-Hyderabad oil markets. The remaining market pairs showed unidirectional causalities. Hyderabad market influences only the Vijayawada market. However, it is influenced by the other three markets: Chennai, Nagpur, and Bengaluru oil markets. Vijayawada, Chennai, and Bengaluru oil influences the price formation in Nagpur markets. Bangalore market influences price formation in the Chennai market. On the other hand, only Hyderabad and Vijayawada markets are influencing the international market (Rotterdam). Similar results were obtained by Gracy et al. (2013) and Vasudeva and Sujatha (2012).
Granger causality test for different Sunflower markets
CONCLUSION
The study investigates the stationarity and cointegration in major groundnut, sunflower, and soybean oil markets of India and the international market. The study examines the market integration in five selected domestic markets and one international market for each selected crop using Johansen’s cointegration test and Granger Causality test. Unit root test showed non-stationary of price series at their levels, and it became stationary after the first differences. Johansen’s cointegration test has shown that even though the selected wholesale oil markets are geographically separated and spatially segmented, they are well-connected in terms of oil prices of all selected crops, demonstrating that the selected oil markets have long-run price linkages. The outcome of the Granger causality test, confirmed unidirectional and bidirectional causalities between the selected oilseed market pairs. In the case of groundnut, Rotterdam (International) market is influenced by only the Delhi market while all selected domestic markets influence the latter. The causality results in soybean markets affirmed the unidirectional influence of domestic markets on the international market except for Jaipur market, which has a bidirectional relationship with the international market. In the case of sunflower, only Hyderabad and Vijayawada markets are influencing the international market prices. The suggested policy intervention calls for faster movement of market information through strengthening market intelligence and establishing an online marketing system through networking. Development/ strengthening of market infrastructure, including communication, transportation, and storage facilities, is the need of the hour to integrate the market prices fully.
Akshata, N., Gracy, C.P., Nagashree, N., Jyoti, N.2013. Growth of sunflower production in India - a comparative analysis, Mysore J. Agric. Sci.,47(3):628-634.Gracy, C.P., Akshata, N., Nagashree, N., Girisha, K.2013. Sunflower market cointegration analysis - a market intelligence tool for marketing decisions. Mysore J. Agric. Sci., 47(4):835-839.Nayak, A., Lokesha, H., Gracy, C.P.2020. Co-integration of Groundnut Markets in India with Special Reference to Karnataka State, Curr. J. Appl. Sci. Technol., 39(18):14-22.Nayak, A., Lokesha, H., Gracy, C.P.2020. Market Integration of Major Oilseeds and Vegetable Oils in India- Evidence from Karnataka. Int. J. Agri. Env. Biotech.,13(4):01-08.Sangeetha, R., Ramanand, M.S., Menaka, S.2017. An Econometric Analysis on Groundnut Markets in India, Int. J. Curr. Microbiol. Appl. Sci., 6(8):2131-2142.Shruti, M., Singh, J., Kumar, S.2018. Cointegration among major cauliflower markets in Punjab. Indian J. Econ. Dev.,14(2):330-335.Sundaramoorthy, C., Girish, K.J., SureshPal and Mathur, V.C.2014. Market integration and volatility in edible oil sector in India, J. Indian Soc. Agric. Stat., 68(1):67-76.Suresh, B., Anilkumar B.Kote, 2012, Anempirical analysis of marketing of oilseeds in Haveri district of Karnataka state of India. Int. J. Curr. Res. Rev., 4(11):16-29Vasudeva, N.K., Sujata, S.2012, Integration of sunflower markets in South Karnataka. J. Progress. Agric.,3(2):97-98.Venujayakanth, B., Dudhat, A.S., Swaminathan, B., Ardeshana, N.J.2017. Price integration analysis of major groundnut domestic markets in India. Econ. Aff., 62(2):233-241.