This paper is to employ a vector autoregressive model to investigate the impact of stock market and saving rate on GDP growth. The result indicates that the lagged values of both stock index and saving rate don’t have influence on the current value of GDP. However, we find that the lagged value of stock index does have impact on saving rate. We conclude that one of the most important reason lead to this result should due to small sample size and data of saving rate still remains non-stationary under the condition of I(1).
Key words: vector autoregressive, GDP, stationary, saving rate stock index
Methodology and variable description
The VAR approach that this paper utilizes is to examine the relationship among GDP growth, stock index and saving rate. The VAR model will take each of the variables in the system and relate its variation to its own past history and past values of all the other variables in the system. In this paper, we have a VAR(4) model which contains three variables: GDP stock index saving rate. The first step we have to take is to deal with the data to see whether it’s stationary. We employed the Augmented DF test here. We will see whether the three variables are stationary respectively.
We can see those three t-statistic are less negative than critical values and then we know that they are all non-stationary. Next, we do a first difference to the three variables and to see whether they are stationary:
We can conclude that under the first difference, GDP and stock index data are stationary comparing the t-statistic with critical value. However, saving rate is still non-stationary and we should have do the second difference. But we will ignore this problem here to keep them in the same pace. Next step we have to do is to choose the optimal lag length using a likelihood ratio test. We do it in the Eviews and the result can be shown as follows:
From this graph, we can know the optimal lag length is 4...
Please join StudyMode to read the full document