Markov-switching VAR Research Articles

Granger predictability of real oil prices by us money and inflation in Markov-switching regimes

Abstract This paper presents new evidence that US money supply growth and inflation rates Granger predict real oil prices in a two-regime Markov switching vector autoregression (MS-VAR) model. An asset pricing theory motivates the empirical work by showing how jumps in real oil prices approximately follow jumps in the discount factor to keep constant the competitive return to oil capital. Using monthly data from January 1978 to June 2024, we consider alternative data combinations of US money supply growth rates, US inflation rates, and real oil prices to establish volatility regimes through goodness of fit testing. We set baseline model as that model with the highest likelihood in explaining the real oil price, which combines M2, the CPI less energy prices (CPIE), and real oil prices. Robustness considers two M2 variants combined with the CPIE that have the next highest likelihoods, for two alternative models. In the high volatility regime, results show robust Granger predictability of real oil prices by the baseline M2 and the M2 variants. In the low volatility regime for the baseline model, the CPIE inflation rate Granger predicts real oil prices. The paper contributes these new MS-VAR results that combined with the theory provide nuanced non-conventional support that monetary factors contribute to heightened real oil price episodes in volatile times as well as in calmer periods.

Measuring the Effects of Fiscal Policy Shocks on U.S. Output in a Markov-Switching Bayesian VAR

Measuring the Effects of Fiscal Policy Shocks on U.S. Output in a Markov-Switching Bayesian VAR

Examining the behaviour of inflation to supply and demand shocks using an MS-VAR model

This paper examines how inflation reacts depending on whether a supply (cost) or demand (markup) shock occurs. Despite their importance, the behaviour of markups remains an open empirical question in the literature. We use data for the US over the 1948q1-2019q3 period, decompose the price index to markups and costs, and employ a small-scale DSGE model to extract identifying size conditions for the coefficient estimates. These are then used in a Markov-switching VAR (MS-VAR) with fixed transition probabilities using an updating step. The empirical exercise shows that three different regimes exist (expansionary, contractionary, supply shock), while the Generalized Impulse Response Functions document that markups appear to be countercyclical and marginal costs are procyclical across all regimes. As such, inflation's reaction to a shock can be less volatile than expected depending on the regime. In addition, larger shocks have a lower and less persistent effect on inflation, because they are more easily identifiable, allowing corrective action to be taken.

Forecasting Markov switching vector autoregressions: Evidence from simulation and application

AbstractWe derive the optimal forecasts for multivariate autoregressive time series processes subject to Markov switching in regime. Optimality means that the trace of the mean square forecast error matrix is minimized by using suitable weighting observations. Then we provide neat analytic expressions for the optimal weights in terms of the matrices involved in a state space representation of the considered process. Our matrix expressions in closed form improve computational performance since they are readily programmable. Numerical simulations and an empirical application illustrate the feasibility of the proposed approach. We provide evidence that the forecasts using optimal weights increase forecast precision and are more accurate than the traditional Markov switching alternatives.

Transmission of Inflation and Exchange Rate Effects: The Markov Switching Vector Autoregressive Methodology

The aim of this study is to delve into the intricate the mechanism through which alterations in currency exchange rates give rise to shifts in inflation rates, while taking into careful consideration the country’s economic cycle. In order to accomplish this objective, we used a dataset that spanned from 1 January 1999 to 1 July 2023, focusing our analytical lens on three specific geographic areas, namely the Eurozone, the United Kingdom, and Canada. In our pursuit of understanding this complex relationship, we employed the Markov Switching Vector Autoregressive model. Our research outcomes can be succinctly encapsulated as follows: in the initial stages, particularly during phases characterized by robust economic growth, the transmission of exchange rate effects onto inflation levels appeared to exhibit a partial impact across all geographic areas under examination. However, during periods marked by economic downturns, both the United Kingdom and Canada displayed a distinctly more comprehensive transmission of these effects. Moreover, the prevailing projections for the forthcoming time horizon, across all the countries encompassed by our study, strongly indicate the onset of an expansionary phase that is projected to extend over a span of 25 months. Lastly, concerning the implications of unexpected disturbances or shocks, it is noteworthy that the response of exchange rates to inflation induced shocks was neither immediate nor as pronounced as the corresponding reaction of inflation to sudden shifts in exchange rates.

Contagion and linkages across international currencies

This paper examines dynamic linkages among international currencies in the face of several turmoil episodes including the global financial crisis (GFC), eurozone sovereign debt crisis (ESDC) and the recent COVID-19 pandemic. We utilize weekly data from August 2005 till December 2022. Our empirical strategy consists of identification of crisis periods using a Markov-switching VAR, followed by estimation of the dynamic conditional correlations using an ADCC-GARCH model and testing for the impact of the crises using a regression model with robust standard errors. We compare impact of the episodes on the dynamic correlations across the currency pairs. Our findings suggest that correlation between developed economy exchange rates increased significantly during these episodes implying existence of a contagion effect. While significant negative correlation coefficients were observed across all pairs with Japanese Yen during GFC and ESDC, indicating flight to quality. We analyse factors that may be responsible for changes in co-movement of the currencies during the study period and find that financial stress, emergence of the pandemic and geo-political risk have a significant role to play.

Switching spillovers and connectedness between Sukuk and international Islamic stock markets

This study examines the switching volatility spillovers and connectedness between the Dow Jones Sukuk and global and two regional Islamic stock markets (World, Asia, and Europe). Using the Markov switching-vector autoregression and spillover index models, we show that Islamic stock markets are significantly affected by their own shocks under the low-volatility regime (Regime 1), whereas the results are insignificant under the high-volatility regime (Regime 2), except for the World Islamic markets. Similarly, Sukuk is affected by its own shocks, and the magnitude of the spillover increases in a high-volatility regime. Sukuk has an insignificant impact on Islamic stocks in Regime 1 and the effect becomes significant in Regime 2. Moreover, the likelihood of transition is more persistent in Regime 1 than in Regime 2. In addition, the shifts from one regime to another are more persistent than those in the opposite sense. Furthermore, the World Islamic stock markets are the net contributors of spillover to Sukuk for all regimes, and their capacity for spillover increases in Regime 2, whereas Sukuk is the net receiver of spillover for both regimes. In contrast, the ability of the European and Asian Islamic markets as net contributors decreases in Regime 2, and the contribution of Sukuk spillovers increases. Portfolio risk analysis under low- and high-volatility regimes is also discussed.

A matrix unified framework for deriving various impulse responses in Markov switching VAR: Evidence from oil and gas markets

We propose a new method to compute various impulse response functions (IRF) for a Markov switching VAR model in terms of neat matrix expressions in closed form. The key is to derive a suitable closed form representation for Markov switching VAR models using a state-space representation. By this representation, the IRF analysis can be processed with respect to either an asymmetric discrete or a symmetric continuous shocks. A simulation study demonstrates the actual advantages of the proposed matrix methodology. To illustrate the feasibility and the usefulness of our approach, we present empirical applications to oil and natural gas markets showing the relevance of accommodating asymmetries in the relationship between their price shocks and economic activities.

The Asymmetric Effects of Real Estate Uncertainty Shock

In this paper, we explore the asymmetric impact of real estate uncertainty (REU) shock on the U.S. economy and housing market in high- and low-volatility states. To do this, we apply the Markov-switching vector autoregression model. Our empirical results show that the macroeconomic and housing markets react negatively to unexpected uncertainties, and the response to the REU shock is much greater than other housing-related variables in high-volatility states. We confirm the presence of the real-options channel based on the following results: both housing starts and construction employment decrease, and housing inventory increases, especially in low-volatility states. In the short run, we observe a delayed pattern in housing starts and construction employment variables. The effect of the channel is less significant in times of high volatility but confirmed in low volatility. The forecast error variance decomposition results show that the REU shock is greater under high REU volatility and plays a more significant role in explaining most housing-related variables.

Green bonds, conventional bonds and geopolitical risk

This paper analyses linkages between green, conventional (corporate and sovereign) bond markets and geopolitical risk in high and low volatility periods between 2014 and 2022 using a Markov-switching VAR (MS-VAR) framework. The results indicate that geopolitical risk significantly affects green bonds in periods of high volatility, but does not do so to conventional bond markets. Green bond markets are significantly affected by sovereign and corporate bonds in both regimes, with stronger effects from corporate bonds evident in high volatility periods. This suggests that green bonds behave differently to conventional bonds and may be more susceptible to geopolitical risk and contagion.

The asymmetric volatility spillover across Shanghai, Hong Kong and the U.S. stock markets: A regime weighted measure and its forecast inference

Based upon generalized spillover index method under Markov switching vector auto-regression (MS-VAR) framework, this paper aims at a regime dependent measure of the volatility spillover asymmetry across Shanghai, Hong Kong and the U.S. stock markets. For that, we apply a regime weighted measure of spillover and its asymmetry. The empirical results show that there existed evident negative asymmetric effect of spillover across the three markets, and it was regime dependent and time-varying: it got intensified under high volatility regime and turned out stronger after the Program of Shanghai-Hong Kong stock connect. The regime weighted measure of spillover asymmetry is found to be reliable; it could be explained appropriately with the economic policy uncertainty and investors' fear in the U.S., extreme shocks like crash of Shanghai stock market and COVID-19 as well as the capital flow through the Program. From in-sample evidence and out-of-sample forecast of Shanghai stock market, we infer that traditional forecast models could be improved if regime dependent spillover and its asymmetry are considered as additional factors.

Impulse response function analysis for Markov switching var models

We exactly derive the regime-dependent impulse response functions for a Markov switching vector autoregression (VAR) model in terms of neat matrix expressions in closed form. The key is to recognize that the latent first-order Markov switching process in the model has a VAR(1) representation, and that the model can be cast into a state-space form. Using such a representation, the regime-dependent impulse response function analysis can be processed with respect to either an asymmetric discrete shock or to a symmetric continuous shock. Our results extend and correct those obtained by Ehrmann et al. (2003) and coincide with those by Hamilton (1994) for the case of standard VAR models.

Understanding Systemic Risk Dynamics and Economic Growth: Evidence from the Turkish Banking System

The banking crisis experienced at the beginning of 2023 in the aftermath of the global 2008 crisis served as a stark reminder of the importance of systemic risk once again across the world. This study examines the dynamics of systemic risk in the Turkish banking system and its impact on sustainable economic growth between the period of 2007 and 2022. Through the Component Expected Shortfall (CES) method and quantile spillover analysis, private banks, such as Garanti Bank (GARAN), Akbank (AKBNK), İş Bank (ISCTR), and Yapı ve Kredi Bank (YKBNK), are identified as major sources of systemic risk. The analysis reveals a high level of interconnectedness among the banks during market downturns, with TSKB, Vakıfbank (VAKBNK), İş Bank (ISCTR), Halk Bank (HALKB), Akbank (AKBNK), Yapı ve Kredi Bank (YKBNK), and Garanti Bank (GARAN) serving as net risk transmitters, while QNB Finansbank (QNBFB), ICBC Turkey Bank (ICBCT), Şekerbank (SKBNK), GSD Holding (GSD), and Albaraka Türk (ALBRK) act as net risk receivers. Employing the Markov switching VAR (MS-VAR) model, the study finds that increased systemic risk significantly reduces economic growth during heightened financial periods. These findings underscore the importance of monitoring systemic risks and implementing proactive measures in the banking sector. The policy implications highlight the requirement for regulators and policymakers to prioritize systemic risk management. Close monitoring helps detect weaknesses and imbalances that could put financial stability at risk. Timely implementation of policies and rules is crucial in the prevention of the accumulation of systemic risks and in dealing with the existing hazards. Such measures protect the stability of the banking sector and mitigate potential negative effects on the broader economy.

Moments, shocks and spillovers in Markov-switching VAR models

To investigate how economies, financial markets or institutions can deal with stress, we often analyze the effects of shocks conditional on being in a recession or a bear market. MSVAR models are perfectly suited for such analyses because they combine gradual movements with sudden regime switches. In this paper, we develop a comprehensive methodology to conduct these analyses. We derive first and second moments conditional only on the regime distribution and propose impulse response functions for both moments. By formulating the MSVAR as an extended linear non-Gaussian VAR, all results are available in closed-form. We illustrate our methods with an application to stock and bond return predictability. We show how forecasts of means, volatilities and (auto-)correlations depend on the regimes. The effect of shocks becomes highly nonlinear, and they propagate via different channels. During bear markets, shocks have stronger effects on means and volatilities and die out more slowly.

Sustainability, Natural Gas Consumption, and Environmental Pollution in the Period of Industry 4.0 in Turkey: MS-Granger Causality and Fourier Granger Causality Analysis

The effects of environmental pollution and Industry 4.0 on a sustainable environment are the main topic of this study, which may be regarded as a complement to the literature on energy and the environment. The paper aims to investigate the relation between Industry 4.0 (I4.0) and environmental sustainability, which is very important for policymakers, practitioners, and company executives in the period of Industry 4.0 in Turkey. To this end, natural gas consumption and technology patents as control variables of Industry 4.0, in addition to the variables of environmental pollution and economic growth, were selected during the period of 1988 to 2022 using Markov switching VAR (MS-VAR), Markov switching Granger causality (MS-GC), Fourier VAR (FVAR), and Granger causality (FGC) techniques. The reason for covering the period starting in 1988 is its recognition as the beginning of the Industry 4.0 era with AutoIDLab in 1988. According to the causality results, there was unidirectional causality running from technology patents to environmental pollution in the results of both MS-GC and FGC. However, the directions of causality between natural gas consumption and environmental pollution, and between economic growth and environmental pollution differed between regimes in the MS-GC model. Bidirectional causality was determined between economic growth and environmental pollution in the first MS-GC regime. However, in the second regime, unidirectional causality from economic growth to environmental pollution was determined. The causality direction determined by Fourier causality gave the same result with the second regime. A similar finding was observed in the direction of causality between natural gas consumption and CO2 emissions. While MS-GC determined unidirectional causality from natural gas consumption to environmental pollution in the first regime, a bidirectional causality result between GC and environmental pollution was determined in the second regime. The FGC result was similar to the second regime result. And lastly, the MS-GC and FGC methods determined unidirectional causality from Industry 4.0 to environmental pollution.

The Impact of Investor Sentiment on Housing Prices and the Property Stock Index Volatility in South Africa

Abstract While prior studies have examined the predictive effect of macroeconomic and country risk components on property stock index dynamics, limited explanations exist in the literature regarding the time-varying effect of investor sentiment on housing prices. Accordingly, this study assesses the impact of investor sentiment on housing properties’ returns and the effect of investor sentiment on the conditional volatility of housing price indices under different market conditions, using GARCH, GJR-GARCH, E-GARCH and Markov-switching VAR models. We found investor sentiment to significantly impact the risk premium of the property returns, where property returns increased with positive changes in investor sentiment, and conditional volatility of property returns decreased with the same changes in investor sentiment. Investor sentiment exerts positive predictive influences on the prices of small and medium houses, in both bullish and bearish market conditions but does not affect the large housing market segment. This makes the implementation of risk-related diversification across small and medium real estate portfolios more effective than large real estate portfolios. Our findings show that investor sentiment is a plausible driver of mass investor redemption actions under conditions of uncertainty.

A tool to simulate and visualize dyadic interaction dynamics.

ynamic models are becoming increasingly popular to study the dynamic processes of dyadic interactions. In this article, we present a Dyadic Interaction Dynamics (DID) Shiny app which provides simulations and visualizations of data from several models that have been proposed for the analysis of dyadic data. We propose data generation as a tool to inspire and guide theory development and elaborate on how to connect substantive ideas to specific features of these models. We begin by discussing the basics of dynamic models with dyadic interactions. Then we present several models and illustrate model-implied behavior through generated data, accompanied by the DID Shiny app which allows researchers to generate and visualize their own data. Specifically, we consider: (a) the first-order vector autoregressive (VAR(1)) model; (b) the latent VAR(1) model; (c) the time-varying VAR(1) model; (d) the threshold VAR(1) model; (e) the hidden Markov model; and (f) the Markov-switching VAR(1) model. Finally, we demonstrate these models using empirical examples. We aim to give researchers more insight into what dynamic modeling approach fits their research question and data best. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Interest Rates, Money, and Fed Monetary Policy in a Markov-Switching Bayesian VAR

Abstract This paper evaluates the roles of short- and long-term private and government interest rates and inside and outside money in the monetary transmission mechanism. With money and credit markets present, changes in monetary policy set off a chain of relative price and portfolio adjustments affecting output and prices. I study interest rate and money supply rules within this monetary transmission mechanism by estimating several Markov-switching Bayesian vector autoregressions (MS-BVARs) on a quarterly U.S. sample from 1960 to 2018. The best-fit MS-BVAR restricts MS to the impact and lag coefficients of the monetary policy and money demand regressions as well as to the stochastic volatilities (SVs) of the structural shocks. Estimates of this MS-BVAR yield evidence of a SV regime which coincides with NBER-dated recessions. This MS-BVAR also identifies a regime switch in the Fed’s interest rate rule and banks’ demand for outside money around the dot-com bust of 2000 and again from the 2007–2009 recession and financial crisis to the end of the sample. Counterfactual simulations show the 2007–2009 recession and financial crisis would have not been as deep and long-lasting if the fed funds rate had been as low as −8 % in 2009 and remained negative from 2010 through 2016.

Analytical Perspective of Trade Volatility Spillover Analysis During COVID Pandemic Times

No one in the world, particularly the economic community, could have predicted the manner in which the 2019 COVID pandemic erupted or shook the economies. Developing countries, particularly South Asian economies and the lower middle income (LMI) countries (India, Pakistan, Sri Lanka, Bangladesh, Bhutan, and Nepal), had also experienced similar turmoil. The COVID outbreak and trade-related risks were closely associated with exports, imports, international liquidity, and SDR exchange rates, posing time-varying risks. In this sphere, an analytical approach to generate meaningful trade-related time-varying volatility predictions with respect to India’s position is critical while considering the above-listed variables. The Markov-switching (endogenous switching) VAR model using the exponential moving average volatility model (EWMA) of the four macroeconomic variables was used. The author tried to use the EWMA volatility model to witness the impact of spillovers and shock persistence. The present study earmarked the behaviour of time-varying trade-related risks (for Regime 1, i.e., the Early COVID period versus Regime-2 during the COVID period) in terms of the top six economies (as per December 2021 data) when endogenous factors like international liquidity and exchange rate shocks are implemented.

Investigating price fluctuations in copper futures: Based on EEMD and Markov-switching VAR model

Predictable global copper prices are crucial to the transition to a green economy. This paper examines the nonlinear characteristics between the international copper futures prices and their drivers, using combined Empirical Mode Decomposition (EEMD) and the Markov-switching VAR (MSVAR) models. We decompose the monthly international copper futures prices from January 2015 to October 2021 into its low-frequency and high-frequency components by using the EEMD method. In the next step, we employ the MSVAR to examine the nonlinear fluctuation of the international copper futures prices under different regimes. The results indicate that fluctuations of the international copper futures prices exhibit a dynamic regime switching patterns that is characterized as “expansion”, “plateau” and “contraction”. The long-term stability of the international copper futures prices is determined by factors associated with demand, especially demand for strategic metals. Both the London Metal Exchange (LME) copper stocks and the LME copper stock futures have a significant impact on the international copper futures prices in each regime. In several regimes, the global refined copper consumption has no significant effect on the international copper futures prices. The increase in copper turnover is the primary driver of the international copper futures prices during the contraction regime. Regarding the supply factor, an increase in global refined copper capacity would result in a rise in the international copper futures prices during the expansion regime. While a slump would occur during a plateau or contraction regime. The financial factor, reflected by non-commercial traders affects the international copper futures prices volatility differently under different regimes. The increase in speculation reduces the market volatility during regimes of expansion and contraction. In contrast, in plateau regime, speculation increases market volatility and activates the market. The broad dollar index has little impact on the international copper futures prices during each regime. The above conclusions indicate that we should focus on the fluctuations in the international copper futures prices that are caused by the demand for strategic metals and financial factors.