4.3 Results of the regression analyses
The results of the OLS and fixed-effects models presented in Table IV are statistically
significant for the two dependent variables at the 1 per cent level. The results suggest a
positive relationship between size and STD and a negative relationship between size and
LTD, implying that the larger the SME, the more STD and the less LTD it will use. A possible
reason for these findings, which are in contrast to H1a and H1b, is that the sampled SMEs,
even the larger ones, generally rely more on STD than LTD. In line with H2a and H2b, the
OLS model suggests a significant and negative relationship between age and STD and a
significant and positive relationship between age and LTD. However, the fixed-effects model
suggests a significant and negative relationship between age and the two debt variables,
which is in line with H2a but in contrast to H2b. The results of the fixed-effects model
suggest that older SMEs use less debt, as they can rely more on internally generated capital.
The results of the two models suggest a positive and significant association between growth
rate and the dependent variables. In accordance with H3a and H3b, SMEs with a relatively high
growth rate tend to use more external financing. The results further indicate that profitability is
negatively and significantly related to the debt variables, more profitable SMEs being less likely
to use external financing. These findings, generated through the OLS and fixed-effects models,
are consistent with H4a and H4b. Similar results can be observed concerning the relationship
between liquidity and the debt variables. In line with H5a and H5b, high-liquidity SMEs are less
likely than low-liquidity SMEs to use external financing.
Supporting H6a and H6b, the relationship between asset tangibility and STD is significantly
negative, whereas the relationship between asset tangibility and LTD is significantly positive.
This means that SMEs with high levels of tangibility tend to use less STD and more LTD than do
other firms. In agreement with H7a and H7b, the fixed-effects model indicates that use of tax
shields is significantly and negatively related to both STD and LTD. However, according to the
OLS model, and in contrast to H7a, use of tax shields is significantly and positively related to
STD. Finally, the OLS results suggest an industry effect on the debt ratios.
Overall, the OLS and fixed-effects models provide evidence at the 1 per cent level of a positive
relationship between size and growth, respectively, and STD, and a negative relationship
between age, profitability, liquidity and asset tangibility, respectively, and STD. Regarding tax
shields, theOLS model indicates a positive association with STD, whereas the fixed-effects model
indicates a negative association with STD. Moreover, the OLS and fixed-effects models provide
evidence of a positive relationship between growth and asset tangibility, respectively, and LTD,
and a negative relationship between size, profitability, liquidity and tax shields, respectively, and
LTD. Regarding age, the OLS model indicates a positive association with LTD, whereas the
fixed-effects model indicates a negative association with LTD. The industry variable is
significantly associated with both SDT and LTD.
The results of the hypothesis tests are summarized in Table V. Note that for a hypothesis
to be rejected, it is sufficient if just one of the two models indicates an insignificant result.
When comparing how the determinants affect STD and LTD, respectively, Table V shows
that profitability and liquidity are negatively related to both STD and LTD, whereas growth
is positively related to the two dependent variables. Industry sector affects STD and LTD.
The remaining determinants affect STD and LTD, not only differently but also in four
different ways. Size is positively related to STD, but negatively related to LTD. Age is
negatively related to STD, but unrelated to LTD. Asset tangibility is negatively related to