Table 4 Unit Root test result
Variables t-statistic
Error of Model t-statistic
Producer Producer-Rice miller
Level 2 2.1271
u1 in level 0 -6.9459
Differences 1 -11.7015
Producer-Wholesaler Rice miller
u2 in level 0 -4.2549
Level 10 5.8841
Producer-Retailer
Differences 1 -12.2466
u3 in level 2 -2.9671
Wholesaler Rice miller-Wholesaler
Level 0 1.7797
u4 in level 0 -3.8677
Differences 0 -10.1124
Rice miller-Retailer Retailer
u5 in level 2 -2.4995
Level 2 4.6505
Wholesaler-Retailer
Differences 1 -7.7232
u6 in level 0 -3.8611
Note: the number in parentheses indicates the lag length. One , two , and three asterisks indicate rejection of unit root at 10, 5, and 1 level of significance, respectively.
Critical values for 10 = -1,62; 5 = -1,94; and 1 = -2,56. Reference: Davidson, R. and MacKinnon, J. 1993,Estimation and Inference in Econometrics p 708, table 20.1, Oxford
University Press, London
Beside the stationary test for series price data, we have to confirm the stationary of the errors in the pair wise models. They are the errors of Producer-
Rice miller, Producer-Wholesaler, Producer-Retailer, Rice miller-Wholesaler, Rice miller-Retailer, and Wholesaler-Retailer. Table 4 shows there are strong
evidences to reject the null hypothesis of non stationary for the errors on level for all models. It means all errors are I0. Therefore, we conclude that all variable are
stationary and valid to use. This also means that each pair wise models are cointegrated.
5.2.2 Cointegration and Error Correction Models 1
Producer Price-Rice Miller Price
First relationship is between producer price and rice miller price. Table 5 shows that there are strong evidences that producer price and rice miller price are
cointegrated. The Johansen Trace test rejects the null hypothesis of no cointegration, but fails to reject the null hypothesis of one cointegrating vector.
The cointegration indicates that producer price and rice miller price are integrated.
Table 5 Cointegration test for producer price-rice miller price 3 lags
Johansen Trace Test Level of significant
r0 LR
P-value 90
95 99
29,71 0,0014
17,98 20,16
24,69
1 6,36
0,1704 7,60
9,14 12,53
After we know that there is one cointegrating vector between producer and rice miller, we continue to examine the extent of their relationship. The Error
Correction Model result shows the short-run dynamics, the adjustment parameter, and the long-run equilibrium. In the short-run there are no significant differenced
lag variables. It means in the short run producer price and rice miller price move in the different way. The error correction as the adjustment parameter shows that
it adjusts very fast about 0,513 of the divergence from the long-run equilibrium being corrected each month. The long-run equilibrium will be achieved when
producer price increase about IDR 1.000, rice miller price will increase about IDR 1.025.
The rice price from producer is transmitted to rice price in rice miller. The factors which may influence this relationship are product linkage, transaction
costs, trade barrier, transparancy of information and market power. Producer price and rice miller price interact directly in the market, so there is strong linkage in
the product flows. This product linkage strongly influences the price determination for both producer price and rice miller price. So, in the long-run
they move together in the same way. Transaction costs including transportation cost to moving rice product from producer to rice miller is relatively low because