The obtained results are combined in the framework of a generalized model. In order to test the stability of coefficients the explanatory regional variables are broken down by years:

(__)(5.1)

where

is the increase in the outstanding creditor indebtedness of industrial enterprises over the period t relatively to the volume of output;

is the rate of increase in the nominal interest rate on granted credits;

is the real interest rate on granted credits;

is the annualized excess of actual federal budgetary expenditures over targets in per cent of targets;

is the increment in the overdue debtor indebtedness of industrial enterprises in the i-th region over the period relatively to the volume of output;

is the outpacing rate of growth in prices of consumer goods produced by enterprises of the i-th region over the period t, accumulated since the price liberalization (January of 1992);

is the aggregate financial result of enterprises operating at a loss in the i-th region relatively to the volume of output;

is the aggregate financial result of enterprises operating at a profit in the i-th region relatively to the volume of output;

the amount of credits in the economy granted by banks of the i-th region over the period t;

are dummies for the period j,.

Different specifications of this model with and without dummies, as well as with and without lagged variables are presented in Annex 2 (Tables 23 – 24 and 25 – 26 respectively). Besides, dummies for the crisis 1998 year (third quarter) and the periods of election campaigns (4th quarters of 1995 and 1999, 2nd quarter of 1996 and 1st quarter of 2000) were introduced to the model.

According to the evaluation results, the model explains about 20 per cent of variance for the total sample and over 30 per cent with excluded outliers. Values and statistical significance of coefficients do not critically change with the introduction of logical variables.

The majority of coefficients are statistically significant. The statistically insignificant coefficients include the nominal interest rate on the granted credits (it shall be noted that the rate of growth in the nominal interest remains insignificant if there are introduced lags and / or the real interest is excluded) and profits, amount of crediting, losses (for some periods), see Tables 23 through 26, Annex 2. The real interest and the budgetary variable remain statistically significant with and without lag and have expected signs of coefficients. The index of regional price structure demonstrates the expected negative coefficient, although at a low level of statistical significance.

The detection of a positive relationship between payment arrears and losses at a high level of statistical significance may provide an evidence that there exists the problem of “transfer” of losses in non-payments to creditors (suppliers, the state, employees), what agrees with the assumptions for a theoretical model of the third type. On the contrary, positive financial results of economic operations (profit) do not demonstrate coefficients of stable positive sign, what is not in variance with this hypothesis. In most cases the profit is statistically insignificant in the model. For years 1995 and 1999 the coefficient of this indicator is negatively signed, what may be explained by the inverse logic of the third theoretical model, i.e. the repayment of previously accumulated indebtedness at the expense of profits. However, profitable enterprises may show more propensity to grant commercial credits, what also may result in a negative relationship between these indicators.

A positive coefficient of the profit was registered for year 1998, what can not be explained by the previous logic and indicates that in 1998 even profitable enterprises might present a source of payment arrears. It shall be noted that 1998 is the crisis year (forex crisis and default on internal public debt). The premeditated causes of non-payments may become especially widespread in the crisis period, characterized by higher uncertainty (theoretical model 1), while a surge in inflation rates registered in this period might significantly affect the accounting results of economic operations (inflationary profits).

It shall be noted that bank crediting had a significant impact on the generation of payment arrears (excluding 1999), what indicates the presence of problems described by the second theoretical model.

Logical variables are statistically significant in the periods of the State Duma election campaigns (IV-1995 and IV-1999) and are positively signed. In fact, it means that overdue indebtedness grew at faster rates over these periods. The factor behind these developments might be the high uncertainty observed in these periods, which facilitated the spreading of premeditated causes of non-payments (Model 1). It shall be emphasized that at the end of each year there are usually registered certain decreases in overdue indebtedness (see Fig. 2.4, Annex 1). A factor behind this phenomenon is that at these periods the authorities more actively carry out offsets in order to improve tax collection and bridge budgetary gaps. However, even taking into account this and other factors of the model, the growth in payment arrears in these periods was significantly above the mean value.

However, no significant increases in the growth of indebtedness relatively to the mean path (predicted by other factors of the model) were registered in the periods of Presidential election campaigns (II-1996 and I-2000). To the contrary, the beginning of year 2000 was characterized by relatively lower increments in overdue indebtedness. This might be attributed to inverse expectations emerging in this period in anticipation of the state to resort to tougher measures concerning payment arrears.

Stability of Coefficients

The following is the testing of the possibility to combine annualized variables.

The hypothesis about the equilibrium of coefficients of model :(5.1) was tested:

(_)(6.1)

Table 8. Results of the Wald Test for coefficient restrictions, model (5.1)__.

F-statistic

15.183289

Probability

0.000000

Chi-square

91.09735

Probability

0.000000

According to the test results (Table 8) this hypothesis is rejected. Indeed, the coefficient values significantly vary across years. A unique “turning point” was registered in 1998. While in preceding periods coefficients were in the neighborhood of one, after 1998 their values decreased (see Tables 23 though 26, Annex 2). Another set of problems related to the poor comparability of the data on financial results in 1999 and 2000 caused by the aggregation embracing different groups of sectors (Goskomstat has started to provide data across all sectors only since 1999, previously only industrial data were available) shall also be mentioned.

The hypothesis was reformulated taking into account a possible change in coefficients after 1998:

(_)(6.2)

Table 9. Results of the Wald Test for coefficient restrictions, model (5.1)__.

F-statistic

1.6765921

Probability

0.137129

Chi-square

8.38079605

Probability

0.1376519

According to the test results, the hypothesis is not rejected. Similar testing of other coefficients demonstrated instability of coefficients of profits, insignificant change in coefficients of granted credits (in spite of the poor comparability of the data from different years), and statistical equilibrium of quarterly dummies (insignificance of the seasonal factor).

For evaluations of the model with excluded insignificant variables see Table 9.

Table 10. Results of the evaluation of model (5.1),WLS, White, IV/1994-IV/2000.eq22_3

In spite of a certain decrease in the coefficient of the multiple regression (0.31 as compared to 0.2); the loss of the explanatory power seems to be caused by combination of profit values from different years (the hypothesis about stability is rejected16), as well as by the aggregation of credits into a single variable and the exclusion of lagged variables). However, this did not affected the key inferences from the model:

losses are a significant factor behind the growth in payment arrears thus indicating that they may be financed at the expense of non-payments (theoretical model 3);

increase in bank crediting of the economy facilitate decrease in non-payments (theoretical model 2);

failures to meet state targets may result in the generation of payment arrears in the real sector (theoretical model 3);

high interest rates on credits facilitate growth in non-payments via higher price of credit resources, decrease in liquidity (theoretical model 2) and / or creating incentives for premeditated non-payments (theoretical model 1);

growth in the real Ruble exchange rate and changes in the price structure (slower rate of the rise in producer prices in comparison to consumer goods) facilitate increase in non-payments (deteriorating effectiveness, therefore theoretical model 3);

indebtedness may grow most intensively in the periods of uncertainty (Model 1).

Evaluation and Analysis of Offset transactions

The intensifying growth in outstanding payments has brought about a number of negative consequences. One of these consequences is the non-monetary administration of budgetary revenues and expenditures. It would suffice to mention that such administration predetermines the structure of budgetary expenditures thus deteriorating their effectiveness to mark this phenomenon as negative. Another negative aspect is that these operations set the precedent for further accumulation of budgetary indebtedness on the part of economic agents in anticipation of offsets. In other words, there arises the moral hazard problem (profitable enterprises delay payments due to the budget). As it was mentioned above, the behavioral aspect of the generation of payment arrears by enterprises anticipating offsets was studied in Perotti (1998), Nikitin (2000).

Offsets took place both at the federal and regional levels (for details on the development of offset transactions in the Russian Federation see Annex 4) and apparently had a certain impact of regional indicators, including payment arrears.

The obtaining of reliable empirical evidence about the impact of offsets on the behavior of economic agents presents certain difficulties due to the fact that, first, behavioral parameters are non-observable, and, second, there is no sufficient statistical data on offset operation (especially at the regional level).

Evaluation of the Share of Offset Transactions

No centralized statistics on regional offset transactions is available, what renders the offset-related testing of the hypotheses more difficult. Taking into account this fact, this paper elaborated a method to compute an indicator indirectly characterizing shares of offset transactions in regions. To be more precise, this indicator characterizes the (weighted) difference between federally conducted and local offsets. The methodology is based on the specifics of tax collection.

The tax legislation of the Russian Federation stipulates the collection of federal, regional, and local taxes. Alongside with the federal agencies, RF subjects are also vested with certain rights concerning the collection of federal taxes. The tax jurisdictions of RF subjects are limited to their respective shares in tax revenues, where they are free to introduce additional privileges and set taxation regimes. The budgetary regulations are stipulated by tax laws, special legislation, and the annually approved federal budget.

Profits of legal persons are taxable at 30 per cent (35 per cent prior to 1999), and the profit tax revenues are shared between the federal and regional as follows: 11 per cent to the federal budget and up to 19 per cent to regional budgets (13 and 22 per cent respectively prior to 1999). The federal budget receives 85 per cent of VAT revenues, regional budgets receive 15 per cent (75 and 25 per cent respectively before 1.4.1999). The authors limited their research to these two taxes, however, the budgets also share excises on certain goods and raw materials, personal income tax, payments for the use of natural resources, the road fund tax.

These arrangements result in the fact that there simultaneously arise tax obligations to both budgetary tiers. These obligations shall be settled basing on the aforementioned proportions, even in case the taxpayer does not meet these obligations in full. Therefore, the budgetary payment arrears are accumulated in accordance with these arrangements.

Of course, in practice these proportions are not strictly observed due to several factors, the first of them being offsets. There is a variety of monetary and barter offsets, both at the regional and federal levels (for details see Annex 4). Offsets are necessarily conducted with participation of either the federal, or a regional budget, who seek to repay the respective indebtedness and administer expenditures and revenues. Therefore the tax collection improves only with regard to the participating budget.

Материалы этого сайта размещены для ознакомления, все права принадлежат их авторам.
Если Вы не согласны с тем, что Ваш материал размещён на этом сайте, пожалуйста, напишите нам, мы в течении 1-2 рабочих дней удалим его.