Page 226 - Contributed Paper Session (CPS) - Volume 8

CPS2256 Norzarita Samsudin
Table 7: Variance Decomposition of Retail Sales
Period S.E. X Y
1 3987.510 100.0000 0.000000
5 5904.051 97.54495 2.455054
10 6300.303 89.83461 10.16539
Source: Author’s computation
Note. X is sales value of retail trade; Y is real GDP.

4.6 Diagnostic Test
The Unrestricted VAR model has a significant R-square and F-statistics.
VAR Residual Serial Correlation LM Tests shows there is no serial correlation
(prob. 0.5642). VAR Residual Heteroskedasticity Tests demonstrates there is
no heteroskedasticity (prob. 0.4041). However, VAR Residual Normality Tests
Jarque-Bera illustrates the residual is not normal (prob. 0.01),

5. Conclusions
The objective of this study is to examine relationship between Malaysia
real GDP and retail trade sales where Vector Autoregressive model, Impulse
Response Function and Variance Decomposition statistical technique were
used. Result showed that there was no co-integration in the long run between
the studied variables. In the short run, retail trade sales lag one and Malaysia
real GDP lag one have positive relationship with Malaysia real GDP.
If there is a positive shock to Malaysia real GDP, retail trade sales will reacts
positively both in the short and long run. Conversely, if there is a positive shock
to retail trade sales, the Malaysia real GDP will react insignificantly in the short
and long run. Variance decomposition of Malaysia real GDP study showed that
in the short and long run, retail trade sales has no influence to Malaysia real
GDP. Meanwhile, in the short run, the retail trade sales influence on itself is
very strong, while in the long run, the contribution of Malaysia real GDP will
increase to almost ten per cent in the variance decomposition of retail trade
study. In this study, there are only two variables are examined. For other
research, other variables that can be included are investment, consumption or
credit.

215 | I S I W S C 2 0 1 9