OIL PRICE SHOCKS AND FINANCIAL ACTIVITY: A TVP-VAR APPROACH

Abstract

1. LITERATURE Transmission channels between oil price shock and financial indicators have been analyzed by scholars with application of various empirical models such as Granger causality (Jones and Kaul, 1996; Arouri and Nguyen, 2010), OLS (Chen et al., 1986; Jones and Kaul, 1996; Faff and Brailsford, 1999; Basher and Sadorsky, 2006; Aloui et al., 2012), GARCH (Filis et al., 2011; Jammazi, 2012), VAR (Huang et al., 1996; Sadorsky, 1999; Ciner, 2001; Hammoudeh and Aleisa, 2005; Park and Ratti, 2008; Apergis and Miller, 2009; Cunado and de Gracia, 2014; Nazlioglu et al., 2015), SVAR (Wang et al., 2013; Chen et al., 2014; Kang et al., 2015a). Some recent studies directly focus on the spillover effects of oil price shocks on financial stress indexes (Chen et al., 2014; Nazlioglu et al., 2015), yet their number is scant. This paper aims to fill this gap by investigating the time-varying effects of oil price shocks on financial stress by employing the TVP-VAR model. 1.1. RESEARCH SUBJECT In this study, we analyze time-varying changes of the structural shocks in the global oil market and their effects on the systemic risk stemmed from the US financial system. 1.2. RESEARCH PURPOSE AND IMPORTANCE The TVP-VAR model consistently and robustly captures the time-varying nature of the structural oil market shocks to the financial activity of the US. 1.3. 2.1. 2.2. CONTRIBUTION of the ARTICLE to the LITERATURE DESIGN AND METHOD RESEARCH TYPE We contribute to the related literature by evaluating time-varying propagations between oil specific shocks and financial activity of the US by employing a seminal approach. 2. This is an empirical research study. RESEARCH PROBLEMS To detect time-varying effects of structural shocks transmitted by the global oil market on the financial stress of the US. 2.3. DATA COLLECTION METHOD Our data set consists of monthly West Texas Intermediate (WTI) spot crude oil prices, world crude oil production in millions of barrels per day averaged monthly and Kansas City Financial Stress Index (KCFSI) covering the 1990 February – 2018 September period. The oil price and KCFSI data have been collected from the FRED database of St. Louis Federal Reserve and the oil production data have been obtained from the U.S. Energy Information Administration (EIA) database. 2.4. QUANTITATIVE / QUALITATIVE ANALYSIS In this study, we implement the Bayesian estimation of time-varying parameter VAR (TVP-VAR) model of Del Negro and Primiceri (2015) in which the coefficients and variance-covariance matrix of the innovations can change over time. 2.5. RESEARCH MODEL In this study, we employ the TVP-VAR model to capture time-varying nature of the oil price shocks. 2.6. RESEARCH HYPOTHESES Structural oil price shocks can significantly affect the financial activity of the US. 3. 3.1. FINDINGS AND DISCUSSION FINDINGS as a RESULT of ANALYSIS Time-varying unconditional standard deviation for the financial stress properly captures wellknown financial stress incidents over the analyzed period. The index significantly surges in the late-1990s covering the Russian Debt Moratorium and the Long Term Capital Management (LTCM) crisis. The timevarying volatility of the KCFSI peaked during the Global Financial Crisis (GFC) shortly after the Lehman Brother’x collapse. The index dramatically plunges to its average levels during the post-GFC era, yet it notably increases around well-known financial stress events of the European Sovereing Debt Crisis (ESDC). 3.2. Yucel, 1999). Posterior coefficient of variables with lag 1 in the TVP-VAR are computed by carrying out the Monte Carlo Markov Chain (MCMC) algorithm for 50.000 times in the structural VAR model. Using the first five years (1990:2-1995:6) data as the training sample, it is detected that the posterior coefficients of the lag of the KCFSI and real oil prices on the financial stress are negative and have a decrasing trend. Likewise, it is found that the posterior coefficients of the lag of the global o

Özet

Bu çalışma Şubat 1990 ve Kasım 2019 döneminde petrol fiyat şoklarının küresel finansal aktiviteye olan zaman-değişimli etkilerini Zamana Göre Değişen Parametreli VAR (TVP-VAR) modeli uygulayarak incelemektedir. Bu bağlamda aylık spot WTI ham petrol fiyatları, dünya ham petrol üretimi ve Kansas Şehri Finansal Stres Endeksi (KCFSI) verileri ampirik analizde kullanılmıştır. Çalışmanın ampirik sonuçları petrol fiyatlarındaki kalıcı bir artışın finansal koşulları olumsuz olarak etkilediğini göstermektedir. Bununla birlikte, pozitif bir petrol arz şoku petrol fiyatlarında düşüşe neden olmaktadır. Çalışmanın bulguları literatürde elde edilen sonuçlarla uyumludur ve TVP-VAR modelinin yapısal petrol fiyat şoklarının zaman-değişimli yapısını yakalamadaki tutarlılığını göstermektedir.