Page 24 - Azerbaijan State University of Economics

Murad Y. Yusıfov: Econometrıc Assessment Of Optımal Interest Burden: Case Study For Azerbaıjan

Therefore, when determining the borrower's creditworthiness, it must be taken into

account that the difference between the borrower's income and the debt burden is

equal to or greater than the subsistence minimum for the country determined by the
relevant law of the Republic of Azerbaijan for the relevant year.

As mentioned above GDP equals to the sum of the gross value added at basic prices
plus all taxes on products, less all subsidies on products. Therefore, share of debt in
borrower's net income after tax assumes an economic importance. That is why this
ratio reliably indicates that particular borrower's ability to pay back its debts.

Determining the optimal level of the interest burden that maximizes the bank's profit
and tax revenues as a whole is important from the point of view of macroeconomic
analysis. From the point of view of statistical significance and reliability of the
obtained results, the lack of longer time series can be considered as a limitation of the
research.

DATA AND METHODOLOGY
Polynomial regression is a special case of multivariate regression involving only one
independent variable. Relationships that are nonlinear in terms of variables but linear
in terms of parameters can also be determined by OLS method. The polynomial
regression model, which represents a nonlinear relationship from one independent
variable point of view is expressed in the following form:

2

= + + + ⋯ + + (1)

1
0
2

The degree of polynomial is the order of that model. Here, N is the degree of
polynomial. In essence, it can be viewed the case with the multivariate regression
model wherein

3
= , = , = ,..., = .
2

1
3
2

Among non-linear polynomial equations, the simplest one is the equation with one
variable and the highest power of 2 (or quadratic equation):

2
= + + + (2)
0
2
1

In general polynomial models are an effective and flexible having curve fitting method
(Ostertagova, 2012). As mentioned above the most widely used regression analysis
method here is the ordinary least squares method. Polynomial models have the
following problems in this regard: The first problem is the difficulty in interpreting
the results of polynomial regression.

24