Credit Emissions and the Inflation Response

Emission of cheap credits to industrial enterprises and agricultural concerns is frequently identified as the root cause of Ukraine's slide into hyperinflation in 1993 (Hawrylyshyn, et. al., 1994; Babanin, 1995; Strauss, 1995; de Ménil, 1996). Here we examine the extent to which such credits drove the general increase in prices, and increases in the money supply during 1992-96. Sufficient data were available by the end of 1996 to permit econometric analysis. First, we examine the effects of credit emissions on the CPI Index growth rate. The quarterly exponential growth rate in the CPI Index was regressed over the quarterly exponential growth rates for broad money (M2), net banking system credits to state-owned enterprises, NBU and commercial bank credits to the government, and an index of energy price growth, which was derived from residential electricity price data from December 1991 to the end of 1996. [3] (Other models were tested which included growth in net domestic credit, and net domestic assets, but these variables were found to have no statistically significant effects on the CPI or M2 growth.) The energy price growth index was included in these, and in subsequent models, on the basis that it is reasonable to expect that at least a portion of Ukraine's general price rise over 1991-96 may be attributed to the general tendency for prices in transition economies to rise to world market levels. In fact, such energy price adjustments have been dramatic during this period. (See Table 6.)

Table 7 presents the results of three models, one of which includes all four independent variables; the others selectively exclude credits to government and the energy price index, in order to better discern their effects. It turns out that the growth of enterprise credits is a significant and powerful determinant of the rate of change in the CPI. In fact, growth of broad money (M2) is completely overshadowed by growth in enterprise credits, signifying its greater significance in the inflation transmission process. (Obviously, M2 growth remains an intermediate variable, albeit of lesser significance.) We may infer from these results that the main factor aggravating inflation in Ukraine has been the large quantities of cheap credits extended to the enterprise sector. These results are consistent with earlier tests by Babanin (1994), who examined the more limited period from the first quarter of 1992 to the second quarter of 1993, finding that enterprise credits were the primary culprit behind the 1992-93 price surge.

A refinement of this analysis examines the possibility of a lag in the effect of monetary growth on inflation rates, in the presence of enterprise credits. Lags in the inflation response may be due to such factors as the adaptation of the public's inflationary expectation; payment lags; and informational frictions that prevent instantaneous responses to price information by firms and households. Such lags have been observed in the case of several former Soviet and eastern European republics. In the case of Russia, Granville (1995) determined that the period of monetary transmission was approximately four months. Georgés de Ménil (1996) estimated that monetary emission led to inflation in Ukraine in 1992-93 following a three-month lag in response.[4] Strauss (1995) found that it took up to six months to complete price adjustments in response to Kyiv's monetary emissions, but with diminishing effects after 2 months had elapsed.

In order to test for the existence of a lag in monetary transmission, a model was developed which employed quarterly exponential M2 growth lagged 0 months, 1 month, and 2 months, together with the growth rate of credits to enterprises, and the energy price index growth rate. (Models were also tested with M2 quarterly exponential growth rates lagged 3, 4, and 5 months, but no statistically significant influence on inflation was present.) Results of the lagged M2 growth model were as follows:

GCPI = -2.995 - .367GM2 - 1.063GM2LAG1 + 1.560GM2LAG2 + .748GSOE + .334GENERGY
(-.160)(-.317)(-.872) (2.205)(2.352)(1.890)
R2 = .673(t-statistics in parentheses)
F = 4.932 (.011 significance level)
Durbin-Watson = 1.564Sample period: 4Q:1991 - 3Q:1996.
Durbin's h = (not defined)
Where:GCPI= growth rate of CPI index;
GM2= M2 growth rate;
GM2LAG1= M2 growth rate lagged one month;
GM2LAG2= M2 growth rate lagged two months;
GSOE= growth rate of state enterprise credits;
GENERGY= growth rate of energy price index.

In this case, the CPI rises with positive changes in both cheap credits to enterprises and a two-month "transmission belt" lag in monetary expansion. While the beta coefficient for the two-month lagged M2 growth variable is substantially larger than that for enterprise credits, it is best to keep in mind that these two really amount to the same thing: for it is well known that the National Bank of Ukraine (NBU) during this period routinely covered deficiencies in commercial bank accounts resulting from parliamentary-mandated extension of banking system credits to enterprises.

Consistent with these findings, Strauss (1995) found that in Ukraine, the price response to increases in the money supply takes a total of six months to run its course, with the bulk of the resulting increase in the price level coming within three months. It is significant that the period of greatest price adjustment in Strauss's model also occurs in the second month following monetary emission. Using data from January 1992 through December 1994, Strauss's model explains over 95 percent of the long-run price level rise as a consequence of monetary changes, primarily arising from credit emissions. Further, the effects of inflation were temporary, in that, after six months new credit issues dissipate, having no permanent effect on inflation. This suggests that there has been no "inflationary inertia" in the Ukrainian economy. It was the continuing injection of cheap, nearly zero-interest credits to industry that propelled increases in the CPI; the lagged effects have only given the appearance of momentum. Contrary to the arguments of some Ukrainian officials, there is no evidence that any such inflationary momentum exists in Ukraine. Graph 4 depicts quarterly growth rates of credits to enterprises, the government and the CPI from 1992-96. It is easy to see how closely growth in enterprise credits tracks with that of the CPI. Graph 5 depicts the growth in the CPI, M2 growth lagged 2 months, and energy price index growth, all of which tend to converge in 1995-96.

If inflation truly had no inertia in Ukraine, then we would expect that the rise in the CPI would have no effect on monetary growth rates. Strauss (1995) employed a Granger Causality Test to determine the statistical likelihood that money growth caused inflation, and not vice versa. He found that the growth of NBU credits was the primary determinant in the expansion of M2, and not the other way around. To confirm Strauss's result, both M2 and M3 growth rates were regressed over quarterly growth in the CPI Index, enterprise credits, net banking system credits to the government, and the energy price growth index.[5] Table 8 presents the result that CPI growth is not a statistically significant determinant of money supply growth. Rather, once again, the growth rate in credits to enterprises is the main determinant of money supply growth. Inflation thus truly had no "echo." Taken together with the preceding analysis, we may conclude that it was the government's monetization of enterprise credits which accounts for Ukraine's price surges during 1992-96.

A further test examines the rate of growth in broad money (M2) as a function of enterprise and government credits, and both the CPI and real GDP growth, lagged one quarter. The results are presented in Table 9. Again, the price rise is not a significant determinant of money growth; neither is the rate of change (i.e., collapse) in real GDP. Rather, enterprise credits remain a robust and significant determinant of money growth. Neither inflation growth nor the output decline help to predict money growth. It seems clear that other, more overtly political factors were in play.

Previous | Contents | Next

URI HomepageURI Working Papers Page