We have proxied local competition with two alternative
measures of investment intensity in the host
economy. The assumption is that investment reflects
either new entrants into industry, or an upgrading of
the technological level of existing firms, both of
which increase competition and reduce the technology
gap between the affiliate and local firms. The variables
are INVIOVTPUT (the gross fixed capital formation/
gross output ratio) and INVIEMPL (gross
fixed capital formation per employee) and both cover
the host countries’ entire manufacturing sectors,
excluding the US affiliates. They are based on data
from various issues of the United Nations’ Industrial
Statistics Yearbook, and for the INVJEMPL variable,
capital formation figures have been converted from
local currency to US dollars and corrected for intemational
differences in capital goods prices using data
from Summers and Heston (1988). Investment by
multinationals from other countries than the United
States has not been subtracted, which means that
“local competitors” are defined as all non-US actors in
the host country market, including MNC affiliates
from other coMtries. INVIOUTPUT and INVIEMPL
are used interchangeably in the estimations: the variables
provide alternative, although related, measures
of competition, as seen by the simple correlation of
about OS (see Appendix, Table A2). The hypothesis
from the theoretical model is that local competition
reduces the technology gap and the demand for the
affiliates’ products, and increases the marginal revenue
of further technology transfer. Hence, we expect
the affiliates’ technology imports to be positively
related to our proxies for local competition.