Экономические науки/10. Экономика предприятия

Assistant. Natalia V. Krivovyaz.

National Research Tomsk Polytechnic University, Russia

The multiple-factor analysis of the criteria influencing size of insurance payments on the CMTPL on the example of Limited liability company IC «Komestra-Tom»

At calculation of a tariff rate numerous factors are analyzed. Any sign which has insignificant impact on risk implementation, can be rejected. Besides, some signs can be formed only in small groups, for example, the sign "floor" has only two groups.

Considering special complexity of an assessment of insurance of risks and calculations of insurance tariffs, the Russian Insurance Supervision recommends to use the techniques of 1 and 2 insurance tariffs developed and approved by it on risk types of insurance. In this research it would be desirable to pay attention that most often has impact on any economic indicator not one, and some factors.

For carrying out the analysis data of limited liability company of Insurance corporation «Komestra-Tom» formed in 1993 were taken. IC «Komestra-Tom» in the insurance market is about 20 years, dynamically develops and takes the first places in ratings of insurance companies of the Siberian region.

Monthly insurance payments on risk types of insurance make over 3 million rubles. More than 75 thousand contracts of insurance are annually signed. The most used and simplest of models of multiple regression – model of multiple linear regression.

In regression models as explaining variables often it is necessary to use not only quantitative, but also qualitative variables. In models influence of a qualitative factor is expressed in the form of a fictitious (artificial) variable which reflects two opposite conditions of a qualitative factor. In this case the fictitious variable can be expressed in a binary form:

Let's carry out the analysis of influence of factors at a size of insurance payments according to the CMTPL on the basis of statistical data of IC «Komestra-Tom». To IC «Komestra-Tom» I provided statistical data on 3517 contracts of insurance by the CMTPL concluded in the period from 01.01.2009 to 01.01.2011. Data include: sum of insurance payments, sum of collected awards, driving experience, sex of the driver, year of release of the car and make of the car.

First of all it is necessary to calculate under each contract of insurance the sum the earned company of an award. For this purpose for all contracts, since 02.01.2010, we multiply the sum of a collected award by the share of the year which has remained till 31.12.2010. The share of year can be calculated by means of built-in function of an Excel «DOLYAGODA» package. For the contracts signed till 02.01.2010, the sum of a collected award is already equal to the sum of the earned award.

Let's consider influence of various factors which in model will be explaining variables (X1 – the earned awards; X2 – a driving experience; X3 – a sex of the driver; X4 – year of release of the car; X5 – make of the car), on an explained variable Y – insurance payments.

The X3 and X5 variables are qualitative therefore in model they will be expressed in the form of fictitious variables:

The greatest difficulties in use of the device of multiple regression arise in the presence of a multicollinearity of factors when more than two factors are connected among themselves by linear dependence, that is cumulative influence of factors at each other takes place. The multicollinearity of factors is stronger, the assessment of distribution of the sum of the explained variation on separate factors by means of a method of the smallest squares is less reliable.

Factors included in multiple regression have to explain a variation of an independent variable. If the model with a set of factors is under construction, for it the determination indicator which fixes a share of the explained variation of a productive sign at the expense of factors considered in regression pays off. If the coefficient of determination doesn't increase at inclusion in model of the next factor and these indicators practically don't differ from each other, the factor included in the analysis doesn't improve model and practically is an excess factor.

The greatest impact on the size of insurance compensation (Y) is made by a driving experience (X2 factor). As value of coefficient at X2 factor negative, dependence the return, that is the more a driving experience, the less sum of payment, and vice versa.

Other factors influence the size of insurance compensation in much smaller degree. The size of an insurance tariff (X1 factor) has the weakest impact.

Results of the carried-out analysis showed that the driving experience rather strongly influences the amount of insurance payments, so, tariff differentiation depending on an experience of driving is a reasonable measure.

In the Russian union of autoinsurers offers of Minpromtorg on differentiation of tariffs of the CMTPL depending on age of the car arrived. However the carried-out analysis revealed rather weak dependence between year of release of the car and the amount of insurance compensations. So, such measure isn't reasonable.

Literature:

1 . The Federal law of the Russian Federation of April 25, 2002 N 40-FZ "About obligatory insurance of a civil liability of owners of vehicles" (in an edition. Federal laws of 01.07.2011 N 170-FZ (an edition 30.11.2011), of 11.07.2011 N 200-FZ, with the amendments brought by the Federal law of 24.12.2002 N 176-FZ, the Resolution of Constitutional the Vessels Russian Federation of 31.05.2005 N 6-P, the Federal law of 16.05.2008 N 73-FZ).

2 . Official site of Insurance company KOMESTRA-TOM//access Mode: http://www.komestra.tomsk.ru/

3 . Dougerti, Christopher. Introduction in econometrics: the textbook for higher education institutions: the lane with English / K.Dougerti. — 3rd prod. — M: Infra-M, 2010. — 465 pages: silt. — University textbook. — ISBN 978-5-16-003640-3.