independent variables they use net assets instead of total assets to remove the impact
of cash on total assets. In this study, because NWC is also considered separately in
addition to cash, NetAssets is calculated as total assets minus cash and marketable
securities minus the asset portion of NWC (i.e. accounts receivables plus inventories).
Fama and French (1998) and Pinkowitz et al. (2006) include research and development
(R&D) expenses in their model because it is mandatory in the USA and this causes assets
to be understated. Although for the Malaysian case R&D expenditure is not mandatory,
it is included in this model due to the robustness of the model as indicated by Pinkowitz
et al. (2006). R&D expenditure (RnD) is measured as the value of the R&D expenses of the
firms. It is noted that out of the 192 firms selected, only around 20 of them actually invest
in R&D and that too inconsistently. Hence this variable is not expected to be significant
in the Malaysian setting. Interest expenses (interest), according to Fama and French
(1998), convey information about leverage policy (and changes in it) hence they use firm’s
interest payment as a measure of its leverage and this measurement is adopted in this
study. Following Pinkowitz et al. (2006), this study uses the value of common dividend
paid out by a firm as the measure of the dividend policy (Dividend). Cash is measured as
the value of cash and marketable securities. All variables are scaled by total book assets
at time t to solve for heteroscedasticity problem. Changes in the variables are included to
absorb changes in expectations (Pinkowitz et al., 2006). Measurement for each variable
can also be found in Appendix.
Next, to study the effect of financing constraints on the valuation of NWC, the firms are
divided into two groups based on size. Faulkender and Wang (2006) present various
proxies for financing constraints, which are dividend payout, long-term bond rating and
commercial paper rating and size. For the Malaysian scenario, long-term bond and
commercial paper ratings are not easily obtainable due to the inactive participation of
firms in the issue of these securities. It is found in the summary statistics (to be presented
later) that the dividend policy of firms in the sample do not vary much and so using
dividend payout as a proxy for financing constraints may not enable proper classification
of the firms. Hence, although size may have its limitations as a proxy for financing
constraints, it is deemed to be the most suitable to partition the firms in this study.
The partition is carried out by first calculating the size of all firms listed on the Main
Board (now Main Market) for the period of 1999-2008 including those not considered in
this study so that the median of the entire population can be calculated. Size is measured
by the natural logarithm of total assets and the median is then calculated for each year
over this period. Then, firms with size below the median in year t are assigned to the
“small firms” or “constrained” group while those which are above the median are assigned
to the “big firms” or “unconstrained” group. Most studies (such as by Faulkender and
Wang, 2006) use deciles for grouping firms but due to the limitation of sample size, median
is used in this study. Generally, most firms are found to belong to either of the categories
over the period of study but there are some firms which belong to the “unconstrained”
group in some years and the “constrained” group in other years. It is expected from the
arguments by Fazzari and Petersen (1993) and from the results of Faulkender and Wang
(2006) that smaller firms are likely to have higher financing constraints and hence a higher
value will be assigned to the NWC investment of these firms.
5. Data and summary statistics
The data set used in this study is from year 1999 to 2008 due to the requirements of the
model, i.e. lag and lead values, although the analysis is from year 2000 to 2007. The data
are collected from three different databases, i.e. the Datastream International Database,