(industrial prices would not adjust after a shock to the system), but a joint zero restriction of
the speed of adjustment vector is rejected (χ2(3) = 9.807, p = 0.02).
The coefficients of determination are similar to those obtain by other studies ranging between
0.29 and 0.57, thus the model explains a relatively high percent of change in the
macroeconomic variables. The Jarque-Bera statistics reject the normality null at 10% for 3
equations. However, non-normality - implies that the test results must be interpreted with
care, although asymptotic results do hold for a wider class of distributions (von Cramon-
Taubadel, 1998).
Table 8. Residual serial autocorrelation LM and LB tests
|
Lags |
LM-Stat |
Prob.a |
Lags |
LM-Stat |
Prob. |
|
1 |
19.18801 |
0.2590 |
7 |
8.600210 |
0.9290 |
|
2 |
16.41018 |
0.4247 |
8 |
15.34749 |
0.4994 |
|
3 |
11.53637 |
0.7752 |
9 |
21.08346 |
0.1753 |
|
4 |
16.56960 |
0.4140 |
10 |
10.37361 |
0.8464 |
|
5 |
21.45633 |
0.1616 |
∏ |
11.87551 |
0.7525 |
|
6 |
20.28460 |
0.2077 |
12 |
21.57624 |
0.1574 |
|
Ljung-Box |
χ2(244) =288.472 | ||||
a Probabilities from chi-square with 16 df.
Multivariate LM tests for serial autocorrelation do not reject the no-autocorrelation null
hypothesis for up to the 12th order, but the no-autocorrelation in the first 21 observations null
is rejected.
5. Conclusions
In this research, a theoretical model developed by Shagaian et al. (2002) was employed for a
small, open economy. As most post-communist economies, Hungary experienced numerous
monetary shocks during the transition period, many of them due to the less developed
monetary instruments and ad-hoc measures. Empirical evidence is presented that these shocks
quickly found their way into the agricultural sector causing significant though largely
15