While running panel regressions, it is assumed that some

regression coefficients alter across individuals and/or over time in order to

account for individual differences. There are fixed parameters, even though the

regression coefficients are not exactly known. The model is classified as fixed

effect model’ when the coefficients are allowed to change in one or two

dimensions.

In the current model, the intercept is allowed to

change across individuals (households, firms or countries. However, it is

assumed that the parameters of the slope and error variances are constant in these

dimensions. The random effect model utilizes random quantities unlike the fixed

parameters. Random effects are associated with these unsystematic quantities. In

the model, the intercept and slope parameters are not different while there is

a variance in the components of error variances across individuals and/or

times. If the fixed effect model results in huge losses in degrees of freedom,

due to too many parameters in these types of models, then the random effect

model becomes more appropriate choice. As Judge et al. (1988) and Baltagi

(2001) explained in their studies, the random effect model is chosen when

individuals are randomly selected from a big population. There has been a

heated debate on the selection of the fixed effect and random effect model among

econometricians for so long. The choice of the appropriate model is based on

the assumptions on the interrelationship of the exogenous variables, both cross-sectional

and across time, the error term assumptions, and/or the researcher’s propensity

for increased efficiency and decreased bias in the estimators. Even though, fixed

effects model is mostly less efficient, the model is known to be more

consistent and less biased. Random effects model is more restricting than the fixed

effects models. The random effects model, which is a specific case of the fixed

effects model, needs additional assumptions. Fixed effect model delivers

control for all variables that do not vary across time. However, the coefficient estimates for

variables that do not vary across time can be calculated by using the random

effects model. RR

It is true that the random effects model has less sample

variability leading to more efficient estimators. However, if the assumptions

are not realized, the model can cause biased estimators. The researchers mostly

prefer fixed effects model due to its unbiased estimators and less restrictive nature.

If there is a necessity for the estimates of coefficients for time invariant

variables, then random effects model is more ideal.

It was important

to decide which of these two models can be more appropriate with missing or

unbalanced data, such as is the case in the current study. Both the fixed effects

and random effects models are sufficient to work with unbalanced designs of the

data, maintaining degrees of freedom compared to excluding observations in

order to generate a balanced data.

In this

study, the fixed effects model is expected to be the appropriate method since

we expect to see the effect of 2008 global crisis on most of the countries in

the sample. Additionally, our sample matches with the population of the study.

Finally, our data does not contain any time-invariant regressors. However, it

should be asserted that in the with the existing literature, the appropriate

model that fits the sample and the research’s objective must be used. Hausman

and Taylor (1981) test is conducted to decide on the utilization of the

appropriate model. In Hausman Test, the correlation between individual effects

and regressors is tested with the null hypothesis that there is no correlation

between individual effects and regressors. The Hausman Test results for each of

the models for two dependent variables are summarized in Table 3.4and Table 3.5

respectively.