CC BY
25
ORIGINAL PAPER
DOI: 10.26794/2587-5671-2023-27-4-66-79 JEL 017, 011, Ll, C32, Gl
(CO ]
Threshold Cointegration and Price Transmission in Commodity Marketsof India
A. Mishra, R. P. Kumar
National Institute of Food Technology Entrepreneurship and Management, Sonipat (Delhi NCR), India
ABSTRACT
The purpose of this research work is to examine the relationships and price dynamics between agricultural commodities in India, i.e.maize, wheat, barley and soybean. Our approach is to study the long-term relationship using the method of modelling the price transmission for both linear and threshold autoregressive (AR) models and vector error correction (VEC) models. Results revealed that all the price series are well integrated, and threshold error correction models prove that all price series move to restore the long-run relationship, whereas commodity stock prices respond slightly faster than market prices in the short-run. Conclusions from this study show that understanding the price transmission flow and its impact on pricing might help in making better trading strategies. It also regulates the public policy implications of the active participation of farmers in national-level commodity exchanges. Keywords: price transmission; market integration; threshold cointegration; pricing
For citation: Mishra A., Kumar R. P. Threshold cointegration and price transmission in commodity marketsof India. Theory and Practice. 2023;27(4):66-79. DOI: 10.26794/2587-5671-2023-27-4-66-79
ОРИГИНАЛЬНАЯ СТАТЬЯ
Пороговая коинтеграция и ценовая трансмиссия на сырьевых рынках Индии
BY 4.0
А. Мишра, Р. П. Кумар
Национальный институт предпринимательства и менеджмента пищевых технологий, Сонипат (Дели NCR), Индия
АННОТАЦИЯ
Целью данной исследовательской работы является определение взаимосвязей и динамики цен между сельскохозяйственными товарами в Индии: кукурузой, пшеницей, ячменем и соей. Наш подход заключается в изучении долгосрочных отношений с помощью метода моделирования ценовой трансмиссии как для линейных и пороговых моделей авторегрессии (AR), так и для векторных моделей коррекции ошибок (VEC). Результаты показали, что все ценовые ряды являются интегрироваными, а пороговые модели коррекции ошибок доказывают, что они движутся к восстановлению долгосрочной взаимосвязи, в то же время цены на сырьевые акции реагируют немного быстрее, чем рыночные цены в краткосрочной перспективе. Выводы из данного исследования показывают, что понимание потока ценовой трансмиссии и его влияние на ценообразование может помочь в разработке наилучших торговых стратегий. Кроме того, происходит регулирование последствий государственной политики при активном участии фермеров в товарообороте на национальном уровне.
Ключевые слова: ценовая трансмиссия; рыночная интеграция; пороговая коинтеграция; ценообразование
Для цитирования: Mishra A., Kumar R. P. Threshold cointegration and price transmission in commodity marketsof India. Finance: Theory and Practice. 2023;27(4):66-79. DOI: 10.26794/2587-5671-2023-27-4-66-79
introduction
Agricultural trade has been one of the most important aspects of the economies of developing countries for years. It is associated with exports and imports and how domestic prices are integrated with world markets. Many recent studies show that developing countries have relaxed their policies after the 2007-2008 food
© Mishra A., Kumar R. P., 2023
crisis; they can keep the markets integrated with world agri-commodity prices[1-3]. Researchers have used the Law of One Price (LOP) concept to examine the market linkages in the existence of arbitrage, transport, and other transaction costs. This article has studied the price integration of Indianagri-commodities, considering four commodities: maize, wheat, barley and
soybean. It is to establish market integration models between commodity stock prices (NCDEX1 data) and market prices (eNAM2/Agmarknet prices). This article contributes to and extends the limited literature specific to market integration and price transmission for agri-commodities in the Indian context. We considered fouragri-commodities: (i) maize — India is ranked 4th in maize cropland area and 7th in production among maize-producing countries. India's maize production is more than 27.8 million MT during FY 2018-20193;
(ii) wheat — India is the second-largest wheat producer with more than 103.6 million MT in FY 2018-20194;
(iii) barley — is another critical crop primarily used as feed grains and consumed commercially for animal feed, beer production, seed and human food applications; and (iv) soybean — which is the world'slargest produced seed legume and contributes more than 26% of the world's edible oil and about 65% of the global protein concentrate for farm animals' feeding. Soybean's share is more than 41% of the total seedoils and more than 25% of the edible oils [4]. The market price and commodity stock price trends are shown in the Appendix, Fig. A1 to Fig. A4 for these four commodities.
review of literature
Market integration has been studied using various models and statistical approaches and has a growing literature available. It gained more attention after the 2008 food crisis [1, 2, 5-9]. It occurs when prices of goods follow the same pattern in two spatially separated areas over a period of time [10]. It is believed that if markets are more integrated, they will yield lower price volatility [5]. Generally, market integration refers to the degree but not a specific relationship [9]. Market integration usually requires the existence of price transmission among the markets, which may be in the form of reintegrated prices.
As many researchers have studied market integration and price transmission, some studies confirm the existence of integration [1, 11, 12] but few, in contrast, conclude weak or partial integration (Table 1) [13]. Beginning with Fackler, who presented the three market integration
1 NCDEX — National Commodity & Derivatives Exchange Limited is an Indian online commodity and derivative exchange based in India. URL: https://www.ncdex.com/ (accessed on 21.05.2022).
2 Department of Agriculture G of I. e-NAM Overview. URL: https://www.enam.gov.in/web/ (2022) (accessed on 21.05.2022).
3ICAR. India Maize Scenario. URL: https://iimr.icar.gov.in/ india-maze-scenario/ (2020) (accessed on 02.11.2021).
4 IBEF. Wheat production may cross 113 million tonnes: Skymet. URL: https://www.ibef.org/news/wheat-production-may-cross-113-million-tonnes-skymet (accessed on 02.11.2021).
measures with a more specific economic interpretation [14]. They have examined how spatial equilibrium behaves when an access demand shock from one market affects another. Ahmedhas proposed a new VAR-BEKK-GARCH model based on the Chinese stock market, international Oil market, and commodities study [3]. They found a one-directional relationship between stock prices and oil prices relative to commodity prices. They also established a shock spillover between oil and stock prices. Arnade has used an ECM model to study long- and short-run price transmission [8]. They also examine the impact of Chinese commodity markets on world commodity prices. They concluded that short-run price transmission is lower than long-run transmission, and the impact of price transmission highly depends upon the commodity. Mensi has examined the transmission between commodity prices and stock prices using a VAR-GARCH model [6]. They have concluded the existence of transmission for return and spillover. Rapsomanikis had extensively discussed market integration and price transmission among agri-commodities in several developing countries, including India [11]. They found that equilibrium exists for commodities like wheat, maize, and milk in the long run but not for meat. Also, domestic transmission among retail and wholesale prices was found to besignificant. Esposti has investigated the price transmission if the market is uncertain for Italy and world prices [15]. They have used the VECM framework and found that the impact of a price bubble is minimal on the price spread and can be controlled by trade policies. Rezitis has used a non-linear ARDL model to investigate vertical price transmission for the Finland dairy market [16]. The researcher established thata positive — long-run price asymmetry is present. Martin-Moreno has used TAR-ECM and Markov-switching approaches to study European oil prices and found short-term and long-term equilibrium [17]. Bonatoinvestigates price correlations and spillovers with the GARCH model for commodities and oil [1]. Svanidze examined the market integration for wheat among several markets using linear and threshold error correction models, which suggest that trade and transaction costs broadly impact the prices [12]. Boffa studied the market integration among wholesale and domestic markets and examined the vertical integration from wholesale to retail prices [13]. Interestingly, they found a perfect vertical integration for wheat only but not forother commodities. They also studied the impact of GST and additional costs on market integration.
Oinexamined the oil, commodity and financial prices using a threshold error-correction model for the US markets [18]. The researchers have found a short-term non-linear asymmetric price transmission pattern, whereas long-term equilibrium does not show asymmetry. Gannevalused Threshold Vector Error Correction Models
(TVECM) to study market reintegration and price volatility [10]. Garcia-Germán also used error correction models (ECM) to study the impact of international prices on the agri-commodities of European markets and observed a long-run relationship but lower price transmission elasticity [19]. Ceballosextensive work examined the price transmission and volatility of agri-commodities for 41 food products in 27 countries [5]. Primarily observed a lead-lag relationship among the market prices and price volatility for maize, rice and wheat. Abdulaialso observed a long-run price equilibrium among the significant maize markets in Ghana and concluded that markets are well integrated [20]. Elleby used the two-fold regression method based on estimated price transmission elasticities and domestic food price changes [2]. They concluded that middle-income countries broadly impact international food prices. Greb'sextensive work and conclusions are based on the VECM model using log and short-term price transmission coefficients [7]. Drabikhas studied the US maize price integration with emery market prices and observed an imperfect price transmission [21]. Lence used Brand — TVECM to conclude that transfer cost is underestimated and speed of price transmission is also biased [22]. Hatzenbuehler has studied the prices of seven agri-commodities in Nigerian food markets concerning world and neighboring countries [23]. The price transmission was observed to be high for rice and coarse grains. Hassounehexamined the wheat prices using threshold vector error correction and multivariate generalized autoregressive conditional heteroscedasticity models and found that price adjustments are in sync with retailers' marketing margins [24]. Also, there is a long-run equilibrium for Slovenian wheat market prices. Distefano examined the rise of arbitrariness in the price formation mechanism [25].
research gap
Market integration is a well-discussed topic and has a growing literature but most of the work has been done either considering spot — future prices or domestic — international prices. Here, we have observed a gap in which market integration and price transmission are not explored mainly from the same commodity-multiple market perspective. We have taken this opportunity to study the integration of prices among multiple commodities and multiple markets.Our study has incorporated two major markets — commodity (stock) market prices and market prices (eNAM) for fourcommodities — maize, wheat, barley, and soybean. Our approach is more holistic and has included all established models of access price transmission. We tried to find out the answers to the questions below:
1. Do market integration and price transmission exist between India's domestic agri-commodity markets?
2. If price transmission exists, then what are the transmission mechanisms?
Threshold cointegration is employed to answer two research questions: are the pairs of price series tied together by a long-run relationship, and which of the series moves to restore the long-run relationship? The findings of this study can be used to understand the price transmissionflow and its impact on pricing to make relative trading strategies; if a commodity is being traded-in multiple markets. The farmers are trading directly at eNAM, and how far they get fair prices in the context of other markets'pricesis a great concern for policy implications.5 It also regulates the public policy implications oftheactive participation of farmers in national-level commodity exchanges.
methodology
In general, there are three types of price transmission;
(i) spatial transmission: prices reintegrated between two spatially separated markets for the same commodity;
(ii) vertical transmission: cointegrated prices between two points or stages of the value chain e.g.— the price of wheat and price of floor and (iii) cross-commodity: cointegrated prices between two commodities; primarily, they may have substitution effects. Fackler has defined market integration as a measure of the degree to which demand and supply shocks ascending in one market are transmitted to another market [9]. Market integrationis mainly measured by the "price ratio" (Rassociated with a market shock.
dPY
R = _/
XY ЭР
de X
(1)
Эе,
where PX and PY refer to the prices in the markets X and Y respectively, eX represents a hypothetical shock in market X and d is for the first-order derivative of the respective price to the market shock. Rapsomanikishas suggested three components to understand the price transmission (i) Co-movement and completeness of adjustments (ii) dynamics and speed of adjustments and (iii) asymmetry of response may be upward or downward [11]. The first completeness of price transmission is in sync with the Law of One Price (LoP). In contrast, the second primarily depends upon policies and market power (short-run impact), several marketing
5 Department of Agriculture G of I. e-NAM Overview. URL: https:// www.enam.gov.in/web/ (2022) (accessed on 21.05.2022).
Table 1
Summaries of the Studies on Price Transmission
Study Reference Methods Period Commodity type Summary
Spillover network of commodity uncertainties [26] VAR, DY 2014 20072016 Energy, precious and industrial metals, and agricultural Connectedness tends to increase during the period of crisis and the global economic situation influences the connectedness of commodity uncertainty indexes
Vertical price transmission in wheat and flour markets in Bangladesh [27] Threshold cointegration 20082016 Wheat, flour Evidence of threshold effects hasa significant impact on the speed of adjustment toward the long-run & short-run
Spatial Price Dynamics and Asymmetric Price Transmission [28] Threshold cointegration 20102016 Skim milk powder New Zealand's export prices arethe market leader as compared with China, and Ireland's export prices are well more aligned with those in international markets
Impacts of COVID-19 and price transmission in US meat markets [29] Threshold cointegration 20102020 Meat All meat markets are well integrated and unexpected & large price movements are visible during Covid-19
Investigating the Impact of Trade Disruptions on Price Transmission [30] Threshold cointegration 20142019 Commodity markets Trade disruptions between Canada and China impacted global price transmission and resulted in market fragmentation
Asymmetric price transmission in a changing food supply chain [31] ECM, threshold cointegration 20082018 Salmon Price transmission relationship exists between the markets for fresh salmon; but not for smoked salmon
Food security and the functioning of wheat markets in Eurasia [12] TECM 20062009 Wheat A stronginfluence oftrade costs on market integration in Central Asia
Threshold cointegration and spatial price transmission when expectations matter [22] TVECM, threshold cointegration 2018 Agri commodity Transfer costs are systematically underestimated and the speed of price transmission is biased in three regime models
Global relationships across crude oil benchmarks [32] Threshold cointegration 20022014 Crude oil All price series move to restore the long-run relationship is at least one regime
How integrated is the Indian wheat market? [33] Momentum-threshold autoregressive (M-TAR) model 19842003 Agri commodity Asymmetric adjustments of wheat prices indicate that price signals within states are transmitted in an asymmetric manner
Cointegration and threshold adjustment [34] Threshold cointegration 19641998 Interest rates Equilibrium exists between short and long-term interest rates but the adjustments from disequilibrium are asymmetric
Spatial price transmission and asymmetry in the Ghanaian maize market [20] TVECM, threshold cointegration 19801997 Maize (Agri commodity) All major maize markets in Ghana are well integrated
Source: Compiled by the authors.
Notes: VAR: vector autoregressive; DY 2014: Diebold and Yilmaz (2014) model; ECM: Error Correction Model; TECM: Threshold Error Correction Model; TVECM: Threshold Vector Error Correction model; M-TAR: Momentum Threshold autoregressive.
stages, contracts between agents, and transfer costs. As per [11], if P1t and P2t are the prices in spatially separated markets that are integrated in the same order and have stochastic trends, then
P -ePt .
(2)
The above equation is called reintegration regression, where - is a reintegration vector, and M is stationary. In other words, the long-run relationship is also termed the reintegrating regression.
As a first step, we have to consider the time-series properties of price data. For that, we have used stationarity and reintegration methods. As per the literature, most of the articles started with an assessment of stationarity in individual price series. We have used the Augmented Dickey-Fuller (ADF) and Phillips-Perron (PP) unit root tests to check the stationarity of data, and both tests have been conducted with two models: (a) with intercept and (b) with intercept and trend. Both the tests have been executed at the level, and the first difference and Akaike information criterion (AIC) have been used for optimal lag selection. If the series is not integrated in the same order, then, by definition, they are not reintegrated. After that, we employed thecointegration tests. Engle and Granger (1987) introduced the concept of reintegration, which occurswhen two or more variables are nonstationary but their linear relationship is stationary. Cointegration infers that the price variables move together in the long run but may diverge in the short-run [29]. We can present the standard cointegration relationship as equation (3) below; which shows two nonstationary variables that are linked by a long-run, stable relationship
У =a + pxt +v,
(3)
of innovations while ni to np are m x m coefficient matrices, which is called the impact matrix and determines the extent to which the system is reintegrated.
Two likelihood ratio tests are used to determine the number of reintegrating vectors — (i) Trace test and (ii) Maximum Eigenvalue.
Once stationarity and cointegration tests are complete with confirmation of cointegration, we estimate whether the price transmission and correction of short-run disequilibria are characterized by non-linear, asymmetric behaviour. To test the non-linearity, we can apply the residuals of equation 3 to test whether the threshold cointegration exits. If the tests fail to reject linearity, we can model the residuals using an autoregressive (AR) method and model the reintegrated system as a VEC model. In the third step, we implement three tests to check the linear behaviour (linearity). These are (i) Terasvirta test — which relies on Taylor series expansion of the neural network. (ii) White test — which is also based on the theory of neural networks. (iii) Tsay test — which is Turkey'snon-additivity type test.
Non-linear behaviour in error correction terms suggests that they do not follow a linear Autoregressive process. In particular cases, it can be more appropriately characterized by a self-exciting threshold auto-regression (SETAR) model [29]. The SETAR approach allows for asymmetric adjustment to shocks with the error correction term now following
\qLvt-i + e, :vt-i <T
Iw-i+e v-i :
(5)
where yt and xt represents prices at different levels at the time t and error correction term as vt =9y-1. The behaviour of vt decides whether the variables are reintegrated. Following the Engle and Granger (1987) cointegration testing procedure we have tested residuals for stationarity (Appendix, Table A1). We also used the Johansen cointegration test to check the cointegration between two or more time series. It has the advantage over the Engle-Granger and the Phillips-Ouliaris methods, which can estimate more than one cointegration relationship, if the data set containstwo or more time series. If there is a time series with order p, then
y =nY-i+nYt-2+• +nu, (4)
where Yt is an n x 1 vector of time series that are integrated of order one, that is, I(1), ut is an n x 1 series
where Threshold value of the two-regime case with regimes L and H. Asymmetric adjustment occurs when 9L is not equal to 9H. If the VEC model can be given by
Ay ai [Pii Pi2 1 "Áy - i" e y"
Ax, T2 a 2 vt-i + |_P2i P22 J _Áxt-i j e x _ t _
. (6)
Áy, "Ti" aL "af" H "в» Pi2 " "Á-i 1
= + L af Vi + +
Áxt T2_ a L P2i P22 _ Áxti _ .eX _
This representation can also be extended to the threshold vector error correction model (TVECM) such as
, (7)
where L and H are two regimes and vL and vH denote the error correction terms for both regimes respectively. Threshold behaviour in cointegration can thus be described by either a SETAR model of the residuals from
vt =
the cointegrating regression or a TVECM (Appendix, Table A2). There could be four possible scenarios; (i) Cointegration and threshold effects — threshold cointegration case, (ii) Cointegration and no threshold effects — linear cointegration case, (iii) No cointegration and no threshold effects — no cointegration case, and (iv) Threshold effects and no cointegration.
We have used both AR and SETAR models and VEC and TVEC models to understand the dynamics of price adjustment. We applied standard and generalized impulse response functions to examine pricebehaviour. Impulse responses depend on the timing, size, and direction of the shocks [29]. The generalized impulse is given by
GIF(yM)=E(yt+k | yt + v,^.,yt_j)-E(yt+k |y„.„y^). (8)
We have used regime-specific impulse responses, which use parameters from each regime for the threshold models. Please referto Hansen[35] for a further methodological explanation.
data
We have used monthly prices for the period of2005 to 2019, and lastly, the criteria for selecting the commodity are:
1. Commodities should be listed in more than one market. We have taken four commodities — maize, wheat, barley, and soybean- listed in both NCDEX and thepan-India electronic trading portal (eNAM) and have IMF price data.
2. Volume or quantity of trade in the last five years for that commodity.
3. A foodgrain is being selected considering its importance in the food basket.
4. We have not considered the storable or non-storable categories of commodities.
5. Also, we are not categorizing based on "seasonal" and "non-seasonal"commodities.
The first data source forthe price series is NCDEXCommodity Index data — commodity market data from NCDEXfor 2005 to 2019. We will refer to this data as "Commodity StockPrice". Thesecond price series is Agmarknetdata — wholesale market data for the pan-India electronic trading portal (eNAM)or Agmarknetand we will refer to this data as "Market Price".
results and discussion
Before westart establishing the model for fourcommodities — maize, wheat, barley, and soybean, we have observed that, in general, market prices are higher than commodity stock prices; however, there are cases where this relationship is inverted. Such cases are approximately 12.6% for barley, 32.3% for maize & wheat, and only 8% for soybean.
As the first step, we tested all the time series for stationarity using the Augmented Dickey-Fuller (ADF) Unit root (stationary) test. Along with the Phillips Perron test to check the unit root, the results are listed in Table A3 of the Appendix. Results show that price series are not stationary at level but become stationary if we take the first difference. All econometric tests and estimations are conducted using the log prices of the commodities.
Once the stationarity is confirmed, we execute the Johansen cointegration testto understand the long-term association between the markets by examining the co-movement of price signals. The null hypothesis is that there areno cointegrating equations (r = 0) and at most one cointegrating equation (r < 1). Referring to Table 2 the null hypothesis of no cointegration was rejected at a 5% significance level that shows the existence of cointegration between market prices and commodity stock prices. Additionally, we also used the Phillips-Ouliaris Cointegration Test and the results are presented in Table A4 in the appendix. Both tests confirm that all the price series are reintegrated and hence VEC models are appropriate for modelling these price series.
Once the cointegration behaviour is confirmed for all the price series, we need to test the non-linear behaviour in the error correction term. Tests of the residuals from the cointegrating regressions are presented in Table A5 of the Appendix. The results confirm that linearity can't be rejected at 5% significance level for soybean and wheat. At the same time, the barley and maize series are found non-linear by all three tests. Non-linearity conditions have implications for the models considered below, namely differences in transmission and adjustment across the different regimes indicated by thresholds [29]. Next, it's necessary to test for the number of thresholds for barley and maize price series. Table A6 of the Appendix shows the results of SETAR model of the cointegration equation residuals. In the SETAR model, we have three null hypotheses: (i) no threshold vs. one threshold, (ii) no threshold vs. two thresholds, and (iii) one threshold vs. two thresholds. The results suggest one threshold for both barley and maize. Based on the linearity, we have anAR model for soybeans and wheat and the SETAR model for barley and maize. The estimated parameters are given in Table 3. We are interested in the autoregressive parameters, as the larger the autoregressive parameters, the slower will be the adjustment to shocks in the price equilibrium. All the autoregressive parameters are statistically significant, which means the time-related and sequential relationships among the prices. For barley and maize, the regimes are distinguished by the speed of adjustment over the two periods. For barley, the AR(1) term is very close to the high regime in the low regime and almost similar for AR(2).
Source: Author's analysis.
Note: Trace test indicates 1 cointegratingeqn(s) at the 0.05 level.Max-eigenvalue test indicates 1 cointegratingeqn(s) at the 0.05 level * denotes rejection of the hypothesis at the 0.05 level
Table 3
AR and SETAR Estimates
Table 2
Johansen's Cointegration Test Results
Number of Cointegrating Vectors
None At most one
Max. Eigenvalue Trace Max. Eigenvalue Trace
Barley 22.342" 23.338" 0.996 0.996
Maize 32.800" 36.151" 3.351 3.351
Soybean 65.275" 66.436" 1.161 1.161
Wheat 51.829" 53.612" 1.783 1.783
Intercept AR1 AR2
Estimate Std Error Estimate Std Error Estimate Std Error
Barley MP - CSP Low regime -0.004 0.001 0.552 0.019 0.363 0.020
High regime 0.002 0.001 0.578 0.027 0.368 0.029
Maize MP - CSP Low regime 0.001 0.001 0.527 0.018 0.401 0.019
High regime 0.015 0.004 0.699 0.033 -0.017 0.051
Soybean MP - CSP 0.007 0.000 0.640 0.018 0.295 0.018
Wheat MP - CSP 0.004 0.000 0.627 0.017 0.354 0.018
Source: Author's analysis.
Note: MP: market price, CSP: commodity stock price.
This means the speed of adjustment is the same for both regimes. Likewise, if we consider maize, the high regime parameter is higher than the low regime, but it is reversed in AR(2), which concludes the quicker adjustments in the low regime. Referring to Table A6 — of the Appendix, the coefficient of error correction term is larger and more significant for the market prices.
Orthogonalized impulse response functions are shown in Fig. Shocks to commodity stock prices are quick for barley, maize and soybean in the short-run. For wheat commodity stock prices, shocks are not visible until the first two periods but, after that, move to negative and show an asymmetric relationship in the long run. Barley and wheat market prices, trigger movements in the short-run and responses to shocks are mostly faster than responses in maize and soybean. The impulses indicate that both market and commodity stock
prices respond in the short-run andlong-run, whereas commodity stock prices respond slightly faster than market prices. To conclude, impulse responses indicate all the price series are well-integrated. The findings can be helpful for investors as well as policymakers. Since both markets are integrated, the shocks can be more prolonged during the crisis, which can be considered while preparing the policies.
conclusion
We investigated the price dynamics of agri-commodity prices between stock and market prices for India. We have considered four agricultural commodities — maize, wheat, barley and soybean. We used linear and threshold autoregressive (AR) models and vector error correction (VEC) modelsfor long- and short-term relationships. Prima-facie, all four commodity stock prices are
Fig. Impulse Response
Source: Author's analysis.
Note: Commodity name with suffix "E": market price andsuffix'N": commodity stock price.
Barley
Orthogonal Impulse Response from BarleyN Orthogonal Impulse Response from BarleyE
§ s ITs § BarleyN 0.004 0.006 0.008 k _ ...................., ...........
I ""'"'------------------/ 1 V—" ......\..... •
95% Bootstrap CI, 100 runs 2 4 6 8 10 95% Bootstrap CI, 100 runs
Maize
Orthogonal Impulse Response from MaizeN Orthogonal Impulse Response from MaizeE
| « 0.0010 0.0020 O.OC ^^---~-—
1 95% Bootstrap CI, 100 runs § 95% Bootstrap CI, 100 runs
Soybean
8 Orthogonal Impulse Response from SoybeanN 0.0045 0.0055 Orthogonal Impulse Response from SoybeanE
0.000 0.004 2 4 6 S 10 95% Bootstrap CI, 100 runs 0.0025 0.0035 2 4 6 8 10 95% Bootstrap a, 100 runs
Wheat
Orthogonal Impulse Response from WheatN Orthogonal Impulse Response from WheatE
8 V I | .005 0.010 0.015 0.020 _
-0.002 95% Bootstrap CI, 100 runs 95% Bootstrap CI, 100 runs
reintegrated with market prices. Results reveal that all the price series are well integrated, and threshold error correction models prove that all price series move to restore the short- and long-run relationship, whereas commodity stock prices respond slightly faster than market prices in the short-run. The findings of this study can be used to understand the price transmission flow and
its impact on pricing to make relative trading strategies if a commodity is being traded-in multiple markets. The farmers are trading directly at eNAM and how far they get fair prices in the context of other markets' prices is a great concern for policy implications. It also regulates the public policy implications of the active participation of farmers in national-level commodity exchanges.
references
1. Bonato M. Realized correlations, betas and volatility spillover in the agricultural commodity market: What has changed? Journal of International Financial Markets, Institutions and Money. 2019;62:184-202. DOI: 10.1016/j. intfin.2019.07.005
2. Elleby C., Jensen F. Food price transmission and economic development. The Journal of Development Studies. 2019;55(8):1708-1725. DOI: 10.1080/00220388.2018.1520216
3. Ahmed A. D., Huo R. Volatility transmissions across international oil market, commodity futures and stock markets: Empirical evidence from China. Energy Economics. 2020;93:104741. DOI: 10.1016/j.eneco.2020.104741
4. Agarwal D. K., Billore S. D., Sharma A. N., et al. Soybean: Introduction, improvement, and utilization in India. Agricultural Research. 2013;2(4):293-300. DOI: 10.1007/s40003-013-0088-0
5. Ceballos F., Hernandez M. A., Minot N. et al. Grain price and volatility transmission from international to domestic markets in developing countries. World Development. 2017;94:305-320. DOI: 10.1016/j.worlddev.2017.01.015
6. Mensi W., Beljid M., Boubaker A. et al. Correlations and volatility spillovers across commodity and stock markets: Linking energies, food, and gold. Economic Modelling. 2013;32:15-22. DOI: 10.1016/j.econmod.2013.01.023
7. Greb F., Jamora N., Mengel C. et al. Price transmission from international to domestic markets. Courant Research Center Discussion Papers. 2012;(125). URL: https://www.researchgate.net/publication/235944955_Price_ Transmission_From_International_To_Domestic_Markets
8. Arnade C., Cooke B., Gale F. Agricultural price transmission: China relationships with world commodity markets. Journal of Commodity Markets. 2017;7:28-40. DOI: 10.1016/j.jcomm.2017.07.001
9. Fackler P. L., Goodwin B. K. Spatial price analysis. In: Handbook of agricultural economics. Amsterdam: Elsevier Science; 2001;1(Pt.2):971-1024.
10. Ganneval S. Spatial price transmission on agricultural commodity markets under different volatility regimes. Economic Modelling. 2016;52(Pt.A): 173-185. DOI: 10.1016/j.econmod.2014.11.027
11. Rapsomanikis G., Hallam D., Conforti P. Market integration and price transmission in selected food and cash crop markets of developing countries: Review and applications. In: Agricultural commodity markets and trade: New approaches to analyzing market structure and instability. Rome: Food and Agriculture Organization of the United Nations (FAO); 2006:187-217.
12. Svanidze M., Götz L., Djuric I. et al. Food security and the functioning of wheat markets in Eurasia: A comparative price transmission analysis for the countries of Central Asia and the South Caucasus. Food Security. 2019;11(3):733-752. DOI: 10.1007/s12571-019-00933-y
13. Boffa M., Varela G. J. Integration and price transmission in key food commodity markets in India. World Bank Policy Research Working Paper. 2019;(8755). URL: https://documents1.worldbank.org/curated/en/896891551117861857/ pdf/WPS 8755.pdf
14. Fackler P. L., Tastan H. Estimating the degree of market integration. American Journal of Agricultural Economics. 2008;90(1):69-85. DOI: 10.1111/j.1467-8276.2007.01058.x
15. Esposti R., Listorti G. Agricultural price transmission across space and commodities during price bubbles. Agricultural Economics. 2013;44(1):125-139. DOI: 10.1111/j.1574-0862.2012.00636.x
16. Rezitis A. N. Investigating price transmission in the Finnish dairy sector: An asymmetric NARDL approach. Empirical Economics. 2019;57(3):861-900. DOI: 10.1007/s00181-018-1482-z
17. Martin-Moreno J.M., Pérez R., Ruiz J. Evidence about asymmetric price transmission in the main European fuel markets: From TAR-ECM to Markov-switching approach. Empirical Economics. 2019;56(1):1383-1412. DOI: 10.1007/s00181-017-1388-1
18. Oin X., Zhou C., Wu C. Revisiting asymmetric price transmission in the US oil-gasoline markets: A multiple threshold error-correction analysis. Economic Modelling. 2016;52(Pt.B):583-591. DOI: 10.1016/j. econmod.2015.10.003
19. Garcia-Germán S., Bardaji I., Garrido A. Evaluating price transmission between global agricultural markets and consumer food price indices in the European Union. Agricultural Economics. 2016;47(1):59-70. DOI: 10.1111/ agec.12209
20. Abdulai A. Spatial price transmission and asymmetry in the Ghanaian maize market. Journal of Development Economics. 2000;63(2):327-349. DOI: 10.1016/S 0304-3878(00)00115-2
21. Drabik D., Ciaian P., Pokrivcák J. The effect of ethanol policies on the vertical price transmission in corn and food markets. Energy Economics. 2016;55:189-199. DOI: 10.1016/j.eneco.2016.02.010
22. Lence S. H., Moschini G., Santeramo F. G. Threshold reintegration and spatial price transmission when expectations matter. Agricultural Economics. 2018;49(1):25-39. DOI: 10.1111/agec.12393
23. Hatzenbuehler P. L., Abbott P. C., Abdoulaye T. Price transmission in Nigerian food security crop markets. Journal of Agricultural Economics. 2017;68(1):143-163. DOI: 10.1111/1477-9552.12169
24. Hassouneh I., Serra T., Bojnec S. et al. Modelling price transmission and volatility spillover in the Slovenian wheat market. Applied Economics. 2017;49(41):4116-4126. DOI: 10.1080/00036846.2016.1276273
25. Distefano T., Chiarotti G., Laio F. et al. Spatial distribution of the international food prices: Unexpected heterogeneity and randomness. Ecological Economics. 2019;159:122-132. DOI: 10.1016/j.ecolecon.2019.01.010
26. Balli F., Naeem M. A., Shahzad S. J.H. et al. Spillover network of commodity uncertainties. Energy Economics. 2019;81:914-927. DOI: 10.1016/j.eneco.2019.06.001
27. Alam M. J., Jha R. Vertical price transmission in wheat and flour markets in Bangladesh: an application of asymmetric threshold model. Journal of the Asia Pacific Economy. 2021;26(3):574-596. DOI: 10.1080/13547860.2020.1790146
28. Xue H., Li C., Wang L. Spatial price dynamics and asymmetric price transmission in skim milk powder international trade: Evidence from export prices for New Zealand and Ireland. Agriculture. 2021;11(9):860. DOI: 10.3390/ agriculture11090860
29. Ramsey A. F., Goodwin B. K., Hahn W. F. et al. Impacts of COVID-19 and price transmission in US meat markets. Agricultural Economics. 2021;52(3):441-458. DOI: 10.1111/agec.12628
30. Mann J., Brewin D. Investigating the impact of trade disruptions on price transmission in commodity markets: An application of threshold cointegration. Journal of Risk and Financial Management. 2021;14(9):450. DOI: 10.3390/ jrfm14090450
31. Gizaw D., Myrland 0., Xie J. Asymmetric price transmission in a changing food supply chain. Aquaculture Economics & Management. 2021;25(1):89-105. DOI: 10.1080/13657305.2020.1810172
32. Mann J., Sephton P. Global relationships across crude oil benchmarks. Journal of Commodity Markets. 2016;2(1):1-5. DOI: 10.1016/j.jcomm.2016.04.002
33. Ghoshray A., Ghosh M. How integrated is the Indian wheat market? The Journal of Development Studies. 2011;47(10):1574-1594. DOI: 10.1080/00220388.2011.579108
34. Enders W., Siklos P. L. Cointegration and threshold adjustment. Journal of Business & Economic Statistics. 2001;19(2):166-176. DOI: 10.1198/073500101316970395
35. Hansen B. E. Threshold autoregression in economics. Statistics and its Interface. 2011;4(2):123-127. DOI: DOI: 10.4310/SII.2011.v4.n2.a4
about the authors / информация об авторах
Arunendra Mishra — Research Scholar, Department of Food Business Management and Entrepreneurship, National Institute of Food Technology Entrepreneurship and Management, Sonipat (Delhi NCR), India
Арунендра Мишра — научный сотрудник, департамент пищевой промышленности и предпринимательства, Национальный институт предпринимательства и управления пищевыми технологиями, Сонипат (Дели NCR), Индия https://orcid.org/0000-0002-6070-1373 Corresponding author / Автор для корреспонденции: [email protected]
Prasanth R. Kumar — PhD in Strategic Finance, Assist. Prof., Department of Food Business Management and Entrepreneurship, National Institute of Food Technology Entrepreneurship and Management, Sonipat (DelhiNCR), India
Прасантх Р. Кумар — PhD в области стратегических финансов, доцент, департамент пищевой промышленности и предпринимательства, Национальный институт предпринимательства и управления пищевыми технологиями, Сонипат (Дели NCR), Индия https://orcid.org/0000-0001-5299-7701 [email protected]
Conflicts of Interest Statement: The authors have no conflicts of interest to declare. Конфликт интересов: авторы заявляют об отсутствии конфликта интересов.
The article was submitted on 21.05.2022; revised on 25.07.2022 and accepted for publication on 06.02.2023. The authors read and approved the final version of the manuscript.
Статья поступила в редакцию 21.05.2022; после рецензирования 25.07.2022; принята к публикации 06.02.2023. Авторы прочитали и одобрили окончательный вариант рукописи.
APPENDIX
Table A1
Granger Causality Tests Statistics for Selected Agricultural Commodities
Null Hypothesis F-Statistic Prob. Direction Relationship
Barley CSP does not Granger Cause MP 17.2323 7.00E-17 Bi-directional MP ^ CSP
MP does not Granger Cause CSP 16.5151 4.00E-16 Bi-directional
Maize CSP does not Granger Cause MP 16.9235 1.00E-16 Bi-directional MP ^ CSP
MP does not Granger Cause CSP 2.52123 0.0276 Bi-directional
Soybean CSP does not Granger Cause MP 102.032 9E-101 Bi-directional MP ^ CSP
MP does not Granger Cause CSP 5.46742 5.00E-05 Bi-directional
Wheat CSP does not Granger Cause MP 9.83574 2.00E-09 Bi-directional MP ^ CSP
MP does not Granger Cause CSP 10.0031 2.00E-09 Bi-directional
Source: Author's analysis.
Note: MP: market price, CSP: commodity stock price.
Table A2
VECM Model
VECM Model ECT Intercept Cointegrating vector
Barley: MP -0.0272(0.0060)""" -0.0432(0.0097)""" 1
Barley: CSP 0.0026(0.0036) 0.0046(0.0058) -1.209868
Maize: MP -0.0198(0.0036)""" -0.0089(0.0018)""" 1
Maize: CSP 0.0040(0.0021). 0.0022(0.0010)" -1.061873
Soybean: MP -0.0602(0.0073)""" -0.0045(0.0007)""" 1
Soybean: CSP -0.0139(0.0043)"" -0.0008(0.0004)" -1.001355
Wheat: MP -0.0640(0.0099)""" 0.0165(0.0025)""" 1
Wheat: CSP -0.0216(0.0105)" 0.0063(0.0026)" -0.9640472
Source: Author's analysis.
Note: MP: market price, CSP: commodity stock price. ***, ** and * indicate the significance of t-statistics at 1%, 5% and 10% level of significance, respectively.
Table A3
Unit-Root Test Results
ADF (first difference) PP (first difference)
Commodity Intercept Intercept & trend Intercept Intercept & trend
Barley MP -2.060 -3.700"" -2.790 -14.283"""
Barley CSP -1.902 -4.267""" -1.783 -5.843"""
Maize MP -1.642 -44.215""" -2.124 -4.179"""
Maize CSP -1.466 -54.039""" -2.004 -54.962"""
Soybean MP -1.786 -44.262""" -2.094 -81.267"""
Soybean CSP -2.768 -55.693""" -1.991 -56.037"""
Wheat MP -0.997 -29.792""" -2.354 -39.818"""
Wheat CSP -0.668 -15.768""" -2.293 -19.674"""
Critical values
1% level -3.960635
5% level -3.411076
10% level -3.127359
Source: Author's analysis.
Note: MP: market price, CSP: commodity stock price. The table contains the t-statistics of the ADF & PP tests results, where *** and ** indicate the significance of t-statistics at 1% and 5% level of significance, respectively.
Table A4
Phillips-OuliarisCointegration Test for Selected Agricultural Commodities
demeaned p-value
Barley -404.7""" 0.01
Maize -119««« 0.01
Soybean -631.35""" 0.01
Wheat -1037.2""" 0.01
Source: Author's analysis.
Note: MP: market price, CSP: commodity stock price. *** indicate the significance of t-statistics at 1% level of significance.
Table A5
Linearity Tests of Price Differences
Terasvirta White Tsay
Statistic P-Value Statistic P-Value Statistic P-Value
Barley 1001""" 2.2E-16 11.722""" 0.002848 3.022""" 2.03E-24
Maize 693.75""" 2.20E-16 27.275""" 1.20E-06 9.153""" 2.29E-21
Soybean 1126.4""" 2.20E-16 4.2145 0.1216 8.588""" 2.10E-52
Wheat 1678.7""" 2.20E-16 2.8488 0.2407 6.929""" 1.88E-281
Source: Author's analysis.
Note: MP: market price, CSP: commodity stock price. *** indicate the significance of t-statistics at 1% level of significanc.
Table A6
SETAR Specification Tests
1vs2: Linear AR TAF (setar /s 1 threshold (2)) 1vs3: Linear AR vs 2 threshold2 TAR (setar(3)) 2vs3: 1 threshold TAR vs 2 thresholds TAR
Series Test P Value Test P Value Test P Value
BarLeyE - BarLeyN
Low regime 31.3 0.02 70.6 0.81 29.3 0.51
High regime 28.3 0.03 67.6 0.92 24.7 0.67
MaizeE - MaizeN
Low regime 69.5 0.05 142.6 0.76 63.9 0.70
High regime 77.5 0.02 150.6 0.81 71.5 0.42
Source: Author's analysis.
7.8
Fig. A1. Barley Price Series
Source: Author's analysis.
Note: Blue represent market price (MP) and red commodity stock price (CSP).
7 8
7.6 _
7.4 _
7.2-
7.0-
6.8-
6.6-
6.4 _
6 2
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 06 07 08 09 10 11 12 13 14 15 16 17 18 19
Fig. A2. Maize Price Series
Source: Author's analysis.
Note: Blue represent market price (MP) and red commodity stock price (CSP).
Fig. A3. Soybean Price Series
Source: Author's analysis.
Note: Blue represent market price (MP) and red commodity stock price (CSP).
Fig. A4. Wheat Price Series.
Source: Author's analysis.
Note: Blue represent market price (MP) and red commodity stock price (CSP).
