Granger Causality Analysis

Granger causality is one of the most primitive methods developed to formally analyze the dynamic relationship between the spot and future prices from the time series observation, linear and nonlinear (nonparametric).This test is based on the notion that the cause occurs prior to its effect. This test is used in investigating whether changes in the spot price leads in the change of future

price or not or vice-versa or both. Due to the simplicity, extendibility and robustness this test is quite dominant and successful in comparison to other tests. Importantly, granger causality test is based on the two principles:

i. The cause happens prior to the effect.

ii. The causal series contains unique information about the effect series that is not available otherwise.

Through the granger causality test, this paper attempts to explain whether the spot and future prices of different commodities cause-effect each other or not.

The granger causality tests for some variables are very sensitive to the selected number of lags in the analysis so we should be careful enough. For the monthly data the suitable lag terms can be from 1-12 or 24; for the quarterly data 1-4 or 12 and for the annual data the reasonable lag terms should be less.

In this test, the stationary prices of the commodities are considered thus the original prices of the commodities are first difference and then further used for this test. Granger causality is used to determine if the variable in time series is useful in forecasting another or not.

Following is the null hypotheses for the Granger-Causality test:

H0: X does not granger-cause Y.

H1: X does granger-cause Y.

Table 17: Results of testing causality between the spot and future prices of different commodities, 2000-2012

COF CFS

CFS CFF

CFF CROS

CROS CROF

*: Significant at 5% level.

CROF GOS CROF GOF CROF NGS CROF NGF CROF SS CROF SF

2. E-08*

0.0246*

0.5424 0.1533 1. E-57*

5. E-06*

Bi-directional causality Bi-directional causality No causality

No causality

Bi-directional causality Bi-directional causality

SF CROF SF GOS SF GOF SF NGS SF NGF SF SS

0.0003*

2. E-38*

0.1635 0.8913 0.8281 0.000*

Bi-directional causality Bi-directional causality No causality

No causality No causality

Uni-directional causality

Looking at the result, it is quite different with other previous studies and suggests that not only future price leads to the determination of spot price for the commodities but also spot price determines the future prices in return. At certain point of time due to the arrival of new information in the market spot price may lead the future price and future price may lead the spot price as market participants filter the available information relevant to their positions.

Most of the researches have concluded that future prices lead the spot prices.

Among them Asche & Guttormsen (2002), Moosa & Al-Loughani (1995) are some of them. However, Sivapulle & Moosa (1999); Quan (1992); Moosa (1996); defended this one and concluded that spot prices lead future price.

Similarly, Garbade and Silber (1983) examined the price discovery role taking in account the spot and future prices of seven commodities and concluded that not only the future market dominate the spot market rather spot market too play an important role in the price discovery.

We find the bi-directional causal relationship between the spot and future prices for the agricultural commodities (cocoa and coffee). For instance, the spot and future prices of cocoa and coffee granger cause each other respectively. Bigman et al., (1983) had stated that it is difficult to justify on which way does the direction of causality flow; is it the spot to future or future to spot. Here, we have solved this problem and come with the conclusion that the direction of causality runs both from spot to future and future to spot respectively. However, we have a kind of mixed results with the precious metals and energy commodities. Gold and crude oil have the bi-directional relationship too. Quan (1992) had concluded that crude oil spot market always dominate the future market .However, in this case both crude oil spot and future prices move together and causes each other rather than playing dominant role. In case of the natural gas, the significant role of spot price in the price discovery role is observed. It’s surprising that the spot and future prices of other commodities are not affected by the future price of natural gas at all. It acts as an independent. The reason behind this might be that natural

gas is regarded as the secondary product of crude oil exploration. Similarly, due to the rise of the alternative source of energy shale gas, the demand for the natural gas has declined. Because of this, it is difficult to analyze the leading role of natural gas future price in the price discovery process in the spot market in the long-run. In contrast, Yohannes & Gebre (2001) stated that in the long-run future price plays a dominant role and exert higher impact on the spot prices. As Brajesh & Ajay (2011) suggested in their research that in case of the non-agricultural commodities mostly future prices lead the spot prices in the long-run. However, our result came up with the conclusion that not only future prices lead the spot prices rather spot prices too lead the future prices in the long-run which can be proved by the relationship of spot prices of natural gas, gold and crude oil with the future prices.

More distinctively, we can see the following uni-directional causality that exists between some of the selected variables: cocoa future causes coffee spot, coffee future causes silver spot, coffee future causes cocoa spot, coffee future causes crude oil future, coffee future causes gold spot, coffee future causes silver spot, coffee future causes silver future, crude oil spot causes coffee spot, crude oil spot causes silver spot, crude oil spot causes silver future, crude oil future causes cocoa spot, crude oil future causes coffee spot, gold spot causes cocoa spot, gold spot causes crude oil spot, gold future causes silver future, natural gas spot causes natural gas future, silver spot causes cocoa spot, silver spot causes gold future and silver future causes cocoa spot.

No causality exists between cocoa spot and coffee future, cocoa spot and crude oil future, cocoa spot and gold spot, cocoa spot and gold future, cocoa spot and natural gas spot, cocoa spot and natural gas future, cocoa spot and silver spot, cocoa spot and silver future, gold spot and cocoa future, gold spot and coffee future, gold spot and natural gas spot, gold spot and natural gas future, gold future and cocoa spot, gold future and cocoa future, cocoa future and coffee future, cocoa future and crude oil spot, cocoa future and crude oil

future, cocoa future and gold spot, cocoa future and gold future, cocoa future and natural gas spot, cocoa future and natural gas future, cocoa future and silver future, gold future and coffee future, gold future and crude oil spot, gold future and natural gas spot, gold future and natural gas future, gold future and silver spot, natural gas spot and cocoa spot, natural gas and cocoa future, natural gas and coffee spot, coffee spot and crude oil spot, coffee spot and gold future, coffee spot and natural gas spot, coffee spot and natural gas future, natural gas spot and coffee future, natural gas spot and crude oil spot, natural gas spot and crude oil future, natural gas spot and gold spot, natural gas spot and gold future, natural gas spot and silver spot, natural gas spot and silver future, natural gas future and the spot and future prices of all the other commodities respectively, silver spot and cocoa future, silver spot and coffee future, silver spot and crude oil spot, silver spot and natural gas spot, silver spot and silver future, silver future and coffee future, crude oil spot and gold spot, crude oil spot and gold future, crude oil and natural gas spot, crude oil and natural gas future, crude oil future and cocoa future, crude oil future and coffee future, crude oil future and natural gas spot, crude oil future and natural gas future, silver future and gold future, silver future and natural gas spot and silver future and natural gas future.

Following table shows the spot and future price of commodities that have the bi-directional relationship with each other.

Table 18: Spot price and future prices with bi-directional relationship Direction of causality

1. COS↔CROS 2. COS↔CFS 3. COS↔COF 4. CFS↔CFF

5. CFS↔CROF 6. CFS↔GOS 7. CFS↔SS 8. CFS↔SF 9. CFF↔CROS 10.CROS↔GOS 11.CROS↔CROF 12.CROF↔GOF 13.CROF↔SS 14.CROF↔SF 15.GOS↔GOF 16.GOS↔SS 17.GOS↔SF 18.GOF↔CROF 19.SS↔SF 20.SF↔COF 21.SF↔CROS

As shown in the table16 spot and future prices of all the commodities except natural gas and silver have the bi-directional relationship. Cocoa spot price has the bi-directional relationship with crude oil spot price and coffee spot price. Coffee spot has the bi-directional relationship with crude oil future price, gold spot price, silver spot price and silver future price respectively. Similarly, coffee future price has the bi-directional relationship with crude oil spot price.

Crude oil spot price has the bi-directional relationship with the gold spot price.

Crude oil future price has the bi-directional relationship with the gold future price, silver spot price and silver future price respectively. Gold spot price has

the bi-directional relationship with the silver spot and future price respectively.

Gold future price has the bi-directional relationship with the crude oil future price. Similarly, gold future has the bi-directional relationship with the cocoa spot price and crude oil spot price respectively.