47 / 330
臺大管理論叢
第
27
卷第
2
期
47
expense item, the coefficient of IT related costs in Model 1.2 should be significantly
negative. As shown, the coefficient of IT related costs (IT
it
) is insignificant (coefficient =
0.157, while
p
= 0.427), consistent with the benefits of IT not being fully reflected in present
accounting profits. Taken together, the above empirical results support H1, which states that
IT expenditures are value relevant from the market valuation perspective. Meanwhile, such
evidence appears to be stark contrast with results from Beccalli (2007) and Martin-Oliver
and Salas-Fumás (2008) who have failed to find the value-increasing role of IT spending,
and reinforce the necessity of using market-based measures when examining the value
implications of such outlays.
The results based on regressions for testing the effect of IT intensity, are presented in
Columns (3) and (4) in Table 5. In Column (3), ITD
it
is not significant (coefficient = 0.002,
and
p
= 0.200), while in Column (4), the interaction term NI
it
×ITD
it
is positive and
significant (coefficient = 2.191, and
p
= 0.002).
14
Therefore, it appears that IT intensity does
not affect firms’ market values directly. Instead, the result is consistent with investing heavily
in IT improves banks’ ability to expand the market share and retain long-term customer
relationship, ultimately leading to more persistent earnings, supporting the prediction of H2.
Also, it can be interpreted that we provide more direct evidence regarding the benefits of
heavy IT investment by focusing on the net effect of such outlay on the ultimate market-
based financial measures, compared to prior findings in Kim and Davidson (2004) that
demonstrated only the increase in non-financial measures (e.g., market shares) or the gross
amount of performance (e.g., revenues).
14 The coefficient of ITD
it
turns out to be negative while NI
it
×ITD
it
is positive, suggesting that the benefit of
improved earnings persistence from investing heavily in IT may be accompanied by considerable
financial burden, and hence impairs the market value to some extent, similar to what has been claimed by
some prior studies (Loveman, 1994; Brynjofsson and Hitt, 1996; Clegg et al., 1997). However, because
the coefficient of NI
it
×ITD
it
is greater than that of ITDit (in absolute values), the net effect of investing
heavily in IT is more likely to be positive.