correlation analysis
In the case of a large sample
The calculation results of correlation coefficient in the above sample interval are shown in the following table. The correlation coefficient between daily LME aluminum inventory and daily LME aluminum (March electronic disk) closing price is -0.333, which is significantly correlated at the level of 0.0 1. There is a negative correlation between them, but the absolute value of correlation coefficient is not high.
Select the latest 3 1 trading day as the sample.
Select the samples of the last 3 1 trading day, and the calculation results are as follows. As can be seen from the table, the correlation coefficient is -0.096, and the correlation is not significant.
Sample selection in the past 74 trading days
Select the samples of the last 74 trading days, and the calculation results are as follows. As can be seen from the table, the correlation coefficient is 0.677, and the correlation is significant.
Judging from the calculation results of the above correlation coefficient, there is a significant negative correlation in the case of large samples; When the sample is the last 3 1 trading day, the correlation is not significant; When the sample is the last 74 trading days, there is a significant positive correlation. From these results, LME inventory has no guiding significance for judging the daily closing price of LME aluminum futures!
granger causal relation test
In the same way as the above sample selection method, samples of different sizes are calculated respectively, and the Granger causality test results are obtained. X in the table represents the daily closing price of LME aluminum (electronic disk in March), and Y represents LME aluminum inventory.
Large sample situation
The calculation results are shown in the following table. It can be seen that there is no Granger causality between LME aluminum futures inventory and closing price in the case of large samples (there is no Granger causality when the lag period is 1 and 2).
Select the latest 3 1 trading day as the sample.
The calculation results are shown in the following table. It can be seen that there is no Granger causality between LME aluminum futures inventory and closing price (there is no Granger causality when the lag period is 2 or 3).
Sample selection in the past 74 trading days
The calculation results are as follows. It can be seen that the closing price is the Granger cause of LME aluminum futures inventory, but the inventory is not the closing price cause (the test results are the same when the lag period is 2 or 3).
From the Granger causality test above, inventory is not the Granger cause of closing price. Choose different samples, sometimes the closing price is the Granger reason of inventory, sometimes it is not. These results also show that it is meaningless to judge the futures price trend with LME aluminum inventory.
From the empirical analysis, there is no obvious correlation between LME aluminum daily inventory and LME aluminum futures daily closing price, and there is no obvious Granger causality. Therefore, it is not appropriate to use daily LME aluminum inventory as an indicator to help judge the trend of LME aluminum futures prices.