119 /302
119
臺大管理論叢
第
28
卷第
2
期
undertaken by export-oriented firms (Chen, 1992). These firms ventured abroad ever since
the value of Taiwan’s currency (NT$) and the domestic wage rate skyrocketed
simultaneously, both undermining their competitiveness (van Hoesel, 1996). According to
information published by the Investment Commission in the Ministry of Economic
Affairs, Taiwan's government approved of only US$374.58 million in overseas investment
in 1987, while it surged to US$62.5 billion by the end of 2010. Hence, the dataset of
Taiwan’s manufacturing firms is appropriate to test the hypotheses, the relationship
between internationalization and performance.
Previous studies employed the maximum likelihood (ML) method or the ordinary
least squares (OLS) method to estimate the contribution of internationalization on
performance. However, they only provide location measures of mean, representing the
“averaging” behavior or “central” tendency of a conditional distribution; hence, they are
unable to explain the behaviors of firms with small and/or large operating scales (i.e., the
right-tail or left-tail behaviors of the distribution). Therefore, the entire conditional
distribution needs to be fully explored to investigate the relationship between
internationalization on performance. The quantile regression method offers multiple
vectors of parametric estimators corresponding to each conditional quantile of a firm’s
operating scale distribution. Hence, it can better completely describe the relationship
between internationalization and performance. Moreover, the corresponding estimators are
robust to outliers, skewed tails, or truncated distribution. Taiwan’s SMEs have the
competitive advantage of flexibility and adaptability to confront capital shortages, energy
crises, and depressions. After rapid economic growth since the 1960s, the SMEs still
constitute 97.68 percent of all businesses in Taiwan as of 2010. The share of sales and
exports by SMEs were 29.55 percent and 16.16 percent, but they accounted for 50.65
percent of exports and even 78.06 percent of employment. Hence, a quantile regression is
very appropriate for analyzing the relationship between internationalization and
performance of Taiwan’s manufacturing firms.
The dataset, collected from “Report on Foreign Investment Strategies of the
Manufactures”, conducted by Department of Statistics, Ministry of Economic Affairs,
consists of 1,151 manufacturing firms. Empirical results show that all hypotheses are
supported and that the marginal contribution of internationalization to different sizes of
firms exhibits a non-linear relationship. More specifically, the larger the firm size the
higher the marginal effects are for the depth of internationalization, while the contribution