Aspects of long-term socio-economic forecasting in the Altai region
There are several traditional forecasting methods. First, it is a trend extrapolation method. Trend extrapolation requires relatively low costs and can reveal seasonality of dynamics and extreme situations. A trend model is a mathematical model that only describes changes in forecast or analytical indicators over time: y = f (t). The function describes the tendency for a fairly stable socio-economic system to change over time t, especially such complex factors as gross national product, inflation, unemployment rate, etc. The passive forecasting method is called naive forecasting, because it assumes a strict inertia of development and represents the transfer of past trends to the future, but development indicators are not influenced by any factors. But it is impossible to completely transfer the trends formed in the past to the future only on the basis of the time factor on a limited data interval. The limitations of the extrapolation method are as follows:
1. In short-term forecasting, extrapolation of past averages leads to ignoring abnormal deviations in the direction of the trend.
2. In the long-term forecast, a high level of aggregation is used without taking into account structural changes in manufactured products, changes in the product itself, production technology and market characteristics.
Let consider the dynamics of monetary incomes on average per capita in the Altai region, Russia (Fig. 1). Average per capita cash income (monthly) of the population is calculated by dividing the annual amount of cash income by the average annual population, and then dividing by 12.
In this case, the dependence is significant at the 1 % level with the significance of the model coefficients. So, we expect by the end of 2021 an average per capita income of the Altai in the amount of 24 802 rubles. If the linear trend did not fit, we would look for another form in the form of a power dependence, logarithmic, exponential, or another functional form. So, often resorted to smoothing time series based on a moving average or exponential smoothing.
The econometric modeling method is one of the most important tools for analyzing and forecasting the socio-economic system. The dependent system can consist of a regression equation with one factor. For example, the Keynesian model, where consumer demand is the dependent variable; independent variable: disposable income. This method provides insight into the relationship between exogenous variables and regression variables.
Let us consider the dependence of consumer spending on average per capita on average income per capita in the Altai region. Having constructed a linear dependence of consumer spending y on cash income x per month per capita, we obtain an equation of the form (1).
y=68,85+0,76×x, (1) when x= -4892,65+ 1099,79×t .
By the end of 2021, the level of cash income will be 24,802 rubles / month, and the level of consumer spending y will amount to 18,876 rubles / month. Thus, linear regression equations can be viewed in conjunction with a trend model.
You can also use an autoregressive equation (ARIMAX type models), where the value of the result index y at any given time is a function of the value of the same index in previous (2) years xt_2. Let us construct an autoregression for the variable cash income x per month per capita in the Altai region(2).
xt=1427+1,08×xt_2, (2)
So, we can predict the value of cash income x by 2021 in the amount of 27 301 rubles per month per capita in the Altai region.
In addition to the regression equation, a so-called governing equation is given (public and private investments are predicted by two independent regression equations, and the third equation allows to calculate the predicted value of total investments). A balance equation is used, the form of which is similar to the equation of identity. For example, an equation that expresses the equilibrium conditions of the commodity market: total demand is equal to total supply.
Thus, the difference between the trend extrapolation method and the econometric method is that the econometric method allows for a meaningful analysis of the dependence of the forecast index on a specific index, while trend extrapolation reflects only the change in the index under study over time.
We construct a multiple regression for consumer spending y from cash income x per month per capita and the time factor, adding one more exogenous variable x ‘- the value of the subsistence minimum in Altai (3).
y=458+0,2×x×lnt+x’, (3)
We found that the regression equation (3) is significant by 5 % . Then, with an increase in cash income per month by 1 %, on average, consumer spending grows by 0,52 %, taking into account the time correction. With the growth of the living wage per month by 1 %, on average consumer spending grows by 0,43 %. In general, the influence of these factors is less than 1 %, which means that you need to look for more significant factors to study consumer spending.
The Gini coefficient is most often used to study income inequality (can vary from 0 to 1). The higher the indicator value corresponds the higher the degree of income inequality. Economists believe that the Gini coefficient should not exceed 0,4. When the number is higher, there will be huge inequality in the region: this will slow the pace of economic development and create a “poverty trap”. See fig.2 Gini coefficient in Altai region.
Then we can generate three forecast options for the Gini coefficient using the method of average chain increments (fig.3):
1. Positive in the case of a decrease in the growth rate of the Gini coefficient, when the average chain growth rate is — 0,1 %.
2. Standard with a moderate growth of the Gini coefficient growth rate, when the average chain growth rate is 0,3 %.
3. Negative with a strong growth of the Gini coefficient growth rate, when the average chain growth rate is 0,5 %.
Then, at the end of 2021, the Gini coefficient will be, in the framework of the positive scenario, 0,305.
Another method is an expert one, aimed at identifying and summarizing opinions based on information provided by experts using systematic procedures. Consequently, these methods require specialists to have deep theoretical knowledge and practical skills in collecting and summarizing all available information about the predicted object. Thus, according to the Rosstat system, the following expert methods can be distinguished: assessment by the population of the socio-economic situation, changes in the economic situation; assessment of the competitiveness of the organization and others.
In particular, for the Altai region, the population’s assessment of the socio-economic situation was carried out in urban districts and settlements. Negative tendencies among the population of the Altai Territory are observed in Aleisk area. The situation is much better in the settlement of Stepnoozersky and Novoaltaisk areas.
