Accounting for Threshold Non-Monetary Events in Hybrid Inflation Models

Abstract

The article shows that in recent decades the phenomenon of inflation has turned mainly from monetary phenomenon mainly into non-monetary. Applied calculations fully confirm this conclusion in relation to Russia. The dependence of inflation on non-monetary factors requires the development of new approaches to its modeling and forecasting. The calculations and analysis have shown the emerging transition from monoinstrumental model complexes to poly-instrumental analytical systems and the shift in potential inflation volatility into the area of short-term fluctuations. Thus, the focus in inflation modeling is shifting towards short-term forecasts. This is an independent instrumental problem. The authors propose a specialized system of inflation forecasting, which includes a decision-maker, an analytical core (a conjugate econometric model and a neural network) and an analytical interface (a system of accounting for threshold non-monetary events and a system of accounting for the volatility of inflation factors). A list of threshold events and a decisionmaking algorithm to adjust the model based on it with the help of the introduced indices of importance of events estimated the basis of an expert survey and the index of the inflationary potential of the environment are proposed in this article.

Short-Term Inflation Projection Based on Marker Models

“…Comparison of models ( 1) and (2) shows that the policy of the Bank of Russia in 2016-2017 was drastically different from the measures of the previous biennium, which was reflected in the consistent and intensive suppression of inflation; this predetermined a completely different composition of markers in model (2) (see Table 1). In particular, the regime change was directed to the "washing out" of external factors and markers.…”

“…The computational experiments for the period of 2016-2017 gave the following alternative inflation model: (2) where ∆W is the monthly financial and technological index of the ratio of labor productivity to wages; CSI is the consumer sentiment index; BCI is the business confidence index (among mining industry entrepreneurs); other indicators have not changed. Statistical characteristics of model ( 2) are satisfactory 2 .…”

“…In fact, it can be stated that the Bank of Russia has ensured that inflation in the country "became detached from" traditional external causes. At the same time, the key rate retained its value, although its influence "shifted" from one month in model (1) to three months in model (2). The change in the regulatory regime of the Bank of Russia has predetermined the limited length of retrospective series for model (2) (24 observations).…”

“…Third, the model calculations show that the phenomenon of inflation is changing from a purely monetary to a predominantly nonmonetary thing [2]. In applied models, these are the nonmonetary factors that usually serve as the inflation dominant.…”

“…Second, the potential volatility of inflation almost entirely goes to the field of short-term CPI fluctuations. The model calculations based on Russian data for 2017-2018 show that the average monthly error of approximation of simulated inflation is 6-7 times higher than the average annual error [2]. Thus, the volatility of the price growth process in Russia is distributed mainly within the years under consideration by months; therefore, it is the short-term projection that poses a particular instrumental problem.…”