6
two inferences can be drawn from this finding. First, FDI flowing from developed countries is more
labor-intensive, thus, the capital cost is less important. Secondly, FDI firms may finance their capital
from home or international sources rather than from domestic financial institutions. However, to
confirm these hypotheses, further investigation needs to be carried out. When comparing the
estimated coefficient’s value of this variable, the negative impact deriving from the rising cost of
capital to FDI is largest for the
USFDI
. This is because the US firms significantly invest in the
financial sector in which the interest rate plays a key role.
As expected, the positive connection between variable
LRER
and FDI is attained from
all regressions, suggesting that currency depreciation in Thailand induces more FDI inflows. The
largest influence of this variable on Japanese FDI is obtained. This result is in line with the
arguments made by Phongpaichit (1989) and Krongkaew (1998), which state that the yen
appreciation in the mid-1980s is the prominent factor forcing Japanese investors to relocate their
production to the developing countries.
To conclude, regarding the long-run, FDI from developed countries is induced by the large
market and the real baht depreciation. On the other hand, the rise in production costs in Thailand,
and especially, the rise in real wage rates, deters FDI inflows. With regard to the estimated results of
FDI inflows from different sources, we found that the determinants of Japanese FDI are closely
parallel to those of
EUFDI
, while the determinants of FDI from the US are significantly different.
When comparing estimated coefficients of the variables, it is observed that Japanese FDI is highly
affected by currency depreciation and real wage rates, the
EUFDI
by the real wage rates, and the US
by domestic demand in Thailand.
SHORT-RUN ESTIMATED RESULTS
Having obtained the long-run cointegration relationships between FDI inflows from
developed countries and their determinants, it is now possible to estimate the short-run dynamic of
the FDI process with the error correction terms (
1
ˆ
−
it
u
) explicitly included in the estimating model.
The major reason for conducting this exercise is to test how fast the FDI inflows adjust to the
changes in their underlying equilibrium.
The short-run estimated results shown in Table 3 are parsimonious models, which are derived
from using standard variable deletion techniques. The null hypothesis of no error correction is
rejected for all equations because the error correction term (
1
ˆ
−
it
u
) is statistically significant. The
speed of adjustment toward the equilibrium in the long-run is quite fast for all equations, indicating
that most of the discrepancy between the actual and the long-run equilibratory values of FDI inflows
are eliminated in the following year.
