Effects of government policies on direct private foreign investment in the manufacturing industries
This proj ect was aimed speci fically at determining empirically
and quantitatively the impact of the 'host government policies to
promote and control Manufacturing Direct Foreign Investment (DFI) on
the inflows of Manufacturing DFI. To achieve this, an econometric
multivariate model of the country-related determinants of DFI - a
'Basic Model' - Was developed in order to control for the countryeffects
on the dependent variable. The statistical analysis of data
Was carried out by means of the Ordinary Least-Squares regression
technique on a cross-sectional basis. The basic model was the equation
linking flows of Manufacturing DFI with market size, environmental risk
and proximi ty (of the investor to the foreign market).. Two government
policy variables were added to this basic model, namely, controls 'on
DFI and effective tax burden.
Controls on DFI was an indicator obtained by determining the
score on a three-point scale of two poltcy areas: policy area A
'degree to which economic activity is closed to foreign control or
participation' and policy area B 'degree of restrictions on the
repatriation of profits and capital'. Each policy area consisted,
in its turn, of two policy items.
The effective tax burden was measured as the percentage of taxes
per unit of ,revenue pro rata for the foreign parent which would apply
to a firm' wi th certain parameters after considering all taxes on
income, dividends, capital, capitalization, tax incentives and tax
treaties between the investor and host countries in the sample.
A comprehensive survey of the previous quantitative and non~
.. quantitative research on the determinants of DFI was carried out.
Likewise, a comprehensive review of research on the measurement of
such determinants was carried out.
The quantitative analysis included tests to determine: a) the
validity of using proxy variables, b) the most suitable indicator
for the variabl es 0 f the basic model, and c) the form 0 f the
relationship - selecting the linear form. The individual impact of
the variables of the basic model was tested. Three variants of the
basic model were developed to eliminate the observed multicollinearity
and the dominant role of market size. subsidiary tests were carried
out to explore the relationships between the independent variables.
The relationships hypothesized in the basic model were tested in an
ex ante fashion with a fUrther set of data.
We obtained a model - variant one - which explained nearly twothirds
0 f the variance 0 f flows 0 f fvIanufacturing DFI with all ratios
highly significant and no outlier residuals in terms of an inveetment
potential variable - obtained by adjusting the size of the market by
the environmental risk - and a nroximity variable - obtained by weighting
a proxy for the psychic distance between the investor and host countries
with its related geographic distance measured on a three-point scale.
A statistically significant negative relationship was established
between controls on DFI and the dependent variable in variants two and
. t·hree. This relationship was established also at a higher level of
s~gnificance and with a larger contribution of controls on DFI to the
R for a subsample of the non-OECD countries in the original sample
with the basic model and variant 3.
A statistically significant relationship was not established
between the effective tax burden and flows of Manufacturing DFI.
Results which were consistent with these were obtained in a replication
of this test carri.ed out with a further set of data. Subsidiary tests
Were carried out with a corporate tax rates variable and with two
proxies for the level of development and wealth of a country, to
explain these results.