Then, after chosen the best model is tested BLUE
(best linear unbiased estimation). If the escape is
interpreted the model, but if not pass will be made an
alternative model of GLS (Generalized Least
Square).
Chow test results show that prob-F = 0.0000 <0.05,
it means that the model selected is the FE model.
Hausman test shows prob-F value = 0.06> 0.05, it
means the model chosen is RE model. The LM test
shows the prob-F value = 0.0000 <0.05 which means
the model chosen is the RE model. Therefore, from
the model selection test it is evident that the suitable
or appropriate model is the RE model. The RE
model as the best model needs to be tested BLUE
via post-estimation test.
To find out whether the selected RE model meets
the BLUE criteria or not, a multicollinearity and
autocorrelation test is required. The multicollinearity
test results show a VIF (Variance Inflating Factor) of
32.65> 10; means there are indications of
multicollinearity. While the Autocorrelation test
shows prob-F = 0.0002 <0.05, there is an indication
of autocorrelation. So, RE does not meet the
qualification of BLUE test. Therefore, we need to
look for the alternative models. In this research, the
alternative chosen model is GLS (Generalized Least
Square).
Based on Table 5, the GLS model shows that the
mobilization of public funds by RDBs which is
represented by LDR is influenced by
macroeconomic factors, market structure, and
banking characteristics. This means the systemic
development of mobilization of public funds by
Indonesian RDBs is related to regional
macroeconomic conditions such as GDP-R, CPI-R
and ER-R; the structure of the national banking asset
market (CR4A) and the bank's internal conditions
such as NPL, NIM, ROA and TE/TA.
4.1 The Effect of NPL on LDR
The non-performing loan has a significant and
negative effect on LDR. This is a good condition
because a low NPL indicates smaller credit risk
which makes bank increase the allocation of credit.
This result is in accordance to research from Saryadi
(2013) which stated the higher NPL wouldresult in
the greater credit risk. Fitria dan Sari (2012)
statedthat a high NPL would make the bankmore
selective in distributing credit.
4.2 The Effect of NIM on LDR
The ratio of NII/TA variable has a significant
positive effect on LDR which means the greater net
interest margin will also make the greater LDR. This
indicates bank has managed to optimize the
difference between interest income and interest
expenses from total assets operated by the bank. The
optimal net interest margin drives to increase LDR.
The result isin accordance with the findings of
Rosadaria (2012) and Buchory (2014) which found
out that NIM had a significant positive effect on
bank liquidity.
4.3 The Effect of ROA on LDR
Return on assets has a significant negative effect on
LDR which means alow ROA will make a high
LDR. This does not mean if a low ROA will make
operating profit goes down. Mathematically, a lower
ROA occurs because the growth ofoperating profit is
smaller than the growth of asset. Asset growth
affects increasing market access which makes the
bank's ability to attract and distribute fund is getting
stronger. This study supports Myers's (1984) who
stated that a high level of profitability would make
firms use retained earnings as a source of funds
compared to outside sources of funds from debt (in
this case, third-party funds) and it results in a decline
of the intermediary function of banks especially in
lending.
4.4 The Effect of TETA on LDR
A high capital adequacy ratio can provide a large
space internally and externally for banks because the
adequacy of banks capital is a requirement of safety
regulations. The higher capital adequacy will
makean optimal an intermediary function of banks in
this case the credit distribution.
In this research, the ratio of TE/TA has a
significant and positive effect on LDR which means
the high capital will increase LDR. Increasing the
capital of banks to make the solvency of banks
increases, this berimpak on trust society sehinggan
ability of banks to collect public funds and
channelled back to the community in the form of
credit becomes increasingly rising.The study
supports the findings of Saryadi (2013) which states
that if CAR increases, it will increase LDR.