Credit Cycles: Econometric Analysis and Evidence for Russia

• Main Authors: Nikita V. Artamonov, Dmitry V. Artamonov, Viacheslav A. Artamonov
• Format: Article
• Language: English
• Published: MGIMO University Press 2014-01-01
• Series: Vestnik MGIMO-Universiteta
• Subjects: кредитные циклы; циклы деловой активности; ряды с единичным корнем; коинтеграция; модель векторной авторегрессии; модель векторной коррекции ошибок; функция импульсного отклика
• Online Access: http://mgimoreview.elpub.ru/jour/article/view/69

One of the principal problem in contemporary macroeconomics is concerned with factors increasing or decreasing economic dynamics. The mainstream approach is based on neoclassical assumptions, but recently new approaches appear mostly based on new Keynesian concepts. In present time the influence of monetary market and credit instruments become more and more significant. Credit resources of banking and financial structures can affect and distort to reallocation of resources for national and even for global economic. In present paper an empiric and econometric analysis for some macroeconometric and monetary indices for Russian Federation is done. An econometrical models describing the influence of credit variables onto real GDP is estimated. It is shown that in short-term periods changes in credit variables do influence significantly onto GDP. It is shown that on short-term periods changes in money aggregate M2 brings influence (through credit variables) onto national output. As well it is shown that changes in short-term interest rate brings significant negative influence onto real output. Impulse response functions for GDP on shocks of credit variables, monetary base and short-term interest rate are evaluated. For the present study of credit cycles and their impact to real business cycles statistical data (quarterly time series) on the following factors for Russian Federation are collected: nominal and real GDP, monetary base M2, short-term interest rate, long-term interest rate (10-year treasuries bill rate), total debt outstanding. All time series are seasonally adjusted and collected for the period 2004 Q1 - 2013 Q2. All interest rates are adjusted for inflation (i.e. we deal with real interest rates). The investigation of long-term relationship for the factors under consideration are based on integration. It is important to note that in the present paper all econometric models are estimated on "pure" statistical data, while in many research papers on business and credit cycles all evaluations and inferences are based on "filtered" time series (mostly filtered by Hodrick-Prescott's method). In present paper "causality" always means "Granger causality". All estimations are made in gretl, an open-source multiplatform econometric software.