Экономические науки/8. Математические методы в экономике

docent Mukhanbetova S.M.

Narxoz University, Kazakhstan

Development trend of communication industry in Kazakhstan

Communication is a vital need for both community and economy. As a result of modern globalization and development of science and technology the industry of communication services is facing crucial changes. Economic condition of a country is also influencing to develop this industry. The economy with a steady development is determined with the balance of cash and goods movements. In such cases the industry with dynamic development attracts more resources and ensures optimal use of these resources. As a result of these processes the gross domestic product (GDP) of a country increases.

On the other hand the output product of communication industry depends not only on the macroeconomic factors, but also on effective demand, ability of population to use new technologies and internet penetration. In the last years the share of communication industry in Republic of Kazakhstan is on average 1.9% of GDP.

In the table below activity results of communication industry of Kazakhstan is shown (table 1).

Table 1

Production account of Kazakhstan communication industry, mln KZT

Parameters	2011	2012	2013	2014	2015
Output	835,7	857,7	1 005,8	961,2	1 066,9
Intermediate consumption	303,8	249,8	306,7	223,0	300,1
Gross value added	531,9	607,9	699,1	738,2	766,8

According to this table between 2011 and 2015 the output product of communication industry was increased on average by 6.3%, value added increase 9.6% and intermediate consumption costs decreased by 0.3%. Within this period the share of gross value added in output has increased from 63.6% to 71.9%. This positive trend was due to modernization of communication infrastructure, investments in automatisation projects, industry efficiency and improvement of labour productivity.

The next table presents significant changes in different types of communication services between 2011 and 2015 (table 2).

Table 2

Volume of communication services, mln KZT

	2011	2015
Total volume of Communication services:	582 740,4	702 148,0
thereof: intercity and international telephone services	44 434,9	33 981,1
Local telephone services	41 467,2	47 826,4
Data transmission	14 322,7	23 798,5
Internet services	96 324,4	190 437,5
Wireless, satellite and etc. connections	12 221,4	28 026,3
Mobile communication services	294 721,1	257 460,5
Other telecommunication services	79 248,7	120 617,8

Table 2 shows that main part of communication services belong to mobile communication services. On the other hand share of mobile services declined from 50.6% in 2011 to 36.7% in 2015. Such structural changes are explained by increase of share of internet services from 16.5% to 27.1% and development of data transmission and other communication services.

In order to run correlation-regression analysis of factors, that influence communication industry, the volume of communication services during 1995-2016 was selected as dependent variable. Following parameters were picked as independent variables: average nominal income per capita, consumer price index (CPI), foreign exchange rates and number of unemployed. With the help of GRETL statistical package the relation between these parameters was estimated. By using ordinary least squares method below linear regression equation was formed:

(1),

Here: – Theoretical value of communication industry output product; – CPI; – average nominal income per capita; – KZT/USD exchange rate; – number of unemployed.

In this example determination coefficient was equal to 99.4%. The closer determination coefficient to 1 the better example is.

During the testing with GRETL for multicollinearity the absence of strong correlation between independent variables of (1) model was shown. Hence they can be included in regression model.

Now we check how this model meets Gauss-Markov rules. If Gauss-Markov rules are fully met, then we can conclude that linear model of multiple regression is a classical model. Mentioned above rules are:

- Mathematical expectation of residuals should be equal to zero;

- Dispersion if residuals should be constant;

- Autocorrelation between residuals should not exist;

- Residuals are not dependent on x and y.

To check if mathematical expectation of residuals equals zero, we derive residual from 1 example as independent variable and find its statistical description. In this case expected value of residuals equals to , which is approximately close to 0.

If the assumption of homoscedasticity of residuals is not met then it would be heteroskedacity, which can be defined with White’s test. During the test for homoscedasticity value of p – 0,579, which means that residuals meet the homoscedasticity conditions because value of p is higher than 0,01.

To check for autocorrelation between residuals we add new variable (uhat3). This variable is one lag different than uhat2. We then create correlation matrix for these two variables that define residuals of the example. As a result we see that correlation coefficient equals to 0.14, hence correlation between uhat2 and uhat3 is insignificant. We can conclude that autocorrelation between residuals of this example doesn’t exist.

The independency of residuals from y and x variables can be checked with by “residuals’ graph” function of GRETL. The result showed that there is no dependency between these parameters.

Therefore our (1) model fully meets Gauss-Markov rules. Hence it can be treated as classical model.

To make a forecast based on given time series data we use ARIMA examples. Since partial autocorrelation function (PACF) exceeded confidence interval only at first lag, ARIMA example contains only one autoregressive component. Hence, in our case adequacy of the sample estimated only by one autoregressive ARIMA (1,1,0) parameter, one moving average ARIMA (0,1,1) parameter and mixed ARIMA (1,1,1) examples.

Since qualitative characteristics of stated examples have no big differences, we use all three of them to make a forecast (table 3).

Table 3

Forecasted volume of communication services of Kazakhstan, mln KZT

FC period	FC based on examples			General Forecast
	ARIMA(1,1,0)	ARIMA(0,1,1)	ARIMA(1,1,1)
2017	682999,8	693874,2	684962,6	687433,94
2018	701414,9	724486,1	696817,9	708018,78
2019	725605,7	755097,9	712122,9	731613,47

As a result we can conclude that positive trend will be kept until the end of predicted period and the volume of communication industry in the last predicted year will be 731613,47 million tenge.

References:

1. Razvitie svjazi i informacionno-kommunikacionnyh tehnologij v Respublike Kazahstan/Statisticheskij sbornik/Astana 2016

2. http://stat.gov.kz

3. Kufel Tadeusz. Ekonometrika. Reshenie zadach s primeneniem paketa programm GRETL: Per. s pol'sk. I. D. Rudinskogo. – M.: Gorjachaja linija–Telekom, 2007. – 200 c.