Which Uncertainty Measure is Most Informative? A Time-varying Connectedness Perspective

This article investigates the existence of spillovers from stock prices onto consumption and the interest rate for South Africa using a time-varying parameter vector autoregressive (TVP-VAR) model with stochastic volatility. In this regard, we estimate a three-variable TVP-VAR model comprising real consumption growth rate, the nominal three-months Treasury bill rate and the growth rate of real stock prices. We find that the impact of a real stock price shocks on consumption is in general positive, with large and significant effects observed at the one-quarter-ahead horizon. However, there is also evidence of significant negative spillovers from the stock market to consumption during the financial crisis, at both short and long horizons. The monetary policy response to stock price shocks has been persistent, and strong especially post the financial liberalisation in 1985, but became weaker during the financial crisis. Overall, we provide evidence of significant time-varying spillovers on consumption and interest rate from the stock market. JEL Classification: C11, C15, C32, E31, E32, E44, E52

This paper investigates spillovers between electricity supply shocks and US growth, using monthly data from 48 US States, spanning the period January 2001-September 2016, while it employs a novel strategy for electricity supply shocks based on a time-varying Bayesian panel VAR model. It accounts for the decomposition of electricity supply per fuel mixture and links its possible interactions with the US macroeconomic conditions. In that sense, the methodology models the coefficients as a stochastic function of multiple structural characteristics. The findings document that GDP growth increases after a positive electricity supply shock, irrelevant to the source of energy that generates it. The absence of a sluggish adjustment mechanism, may reflect weak competition and significant market power by the incumbents in the electricity industry. Lastly, we argue that the rate of response of GDP growth per capita to electricity supply shocks, provides an indication that a market power effect prevails in the US electricity industry.

In this paper, the multi-task motor imagery EEG(electroencephalogram) signals are pretreated by principal component analysis and Fourier transform. By use of the methods of time series analysis and mathematic statistics, pretreated EEG signal series are separated into the deterministic part and stochastic part. Then, the stochastic part is analyzed by use of TVVAR (Time Varying Vector Auto-regressive) model to obtain the residuals. Therefore, the EEG signals are studied on the stochastic parts and the residuals of TVVAR model. EEG signals of 3 types of actions, 60 signals per action type, are sampled, from which a signal is in turn analyzed to be recognized. Experiments in this study indicate that the recognition rates of left, right, hold still are 93.33%, 98.33%, 96.67% respectively, and the average recognition rate is 96.11% through both the stochastic parts and the residuals of TVVAR model. It verifies the TVVAR model be useful to analyze autocovariance nonstationary vector process.

This paper aims to examine the spillovers from stock prices onto consumption and interest rate for Turkey by using a time-varying vector autoregressive model with stochastic volatility. A three-variable time-varying vector autoregressive model is estimated to capture the time-varying nature of the macroeconomic dynamics in the Turkish economy between real consumption, nominal interest rate and real stock prices. In order to obtain the macroeconomic dynamics in a small open economy, the data covers the period 1987:Q1 until 2013:Q3 in Turkey. The sample data is gathered from the official website of Central Bank of the Republic of Turkey. Overall, this study provides the evidence of significant time-varying spillovers on consumption and interest rate coming from the stock market during financial crises and implications of monetary policy in Turkey. In addition, a time-varying vector autoregressive model with stochastic volatility offers remarkable results about the impact of price shock on consumption levels in Turkey.

Abstract: We present an extension of the Time-Varying Vector Autoregressive (TV-VAR) model by incorporating static variables and Type I error controlling, to expand its utility to time series data involving multiple individuals and to decrease the number of spurious interactions. A simulation study with 324 scenarios varying sample sizes, number of time points, correlation coefficients, the probability of cross-lagged effects, and number of time-varying variables is presented. The correlation between estimated and data generating parameters, the sensitivity and the specificity were used to assess model performance. The extension model was applied to empirical data to detect the dynamic changes in depressive symptoms during the period of electroconvulsive therapy in patients with depression. Results show consistent improvements over the TV-VAR model, with higher correlation and specificity, except for slightly lower sensitivity. Notably, enhancements shine in larger samples (≥ 250) and larger time points (≥ 12), while a decline occurs with higher interaction probabilities (e.g., 0.85). Empirical analysis shows that depressive symptoms were generally predicted by themselves and the relationship between depressive symptoms became weaker over time. The extension of the TV-VAR model presents a valuable tool for studying mental health dynamics, particularly in scenarios with appropriate sample size and time points.

This paper investigates the evolving dynamics of the macroeconomy of India in the post reform years after 1991, based on time-varying parameters structural vector autoregression model with stochastic volatility. We find sharp reductions in estimated stochastic volatility during the post reform years for all shocks and variables. In terms of the stochastic volatility, the period 2001 to 2006 seems to have the lowest volatility in the whole sample and can be dubbed as the short ’Great Moderation’ period of India. The estimated stochastic volatility of supply shocks is found to be more than demand shocks. We also note that demand shocks rather seem to be persistent than supply shocks during the period from 2007 to 2014.

We investigate the evolution of the monetary policy transmission mechanism in the Czech Republic over the course of the 1996–2010 time period through the use of a time-varying parameters Bayesian vector autoregression model with stochastic volatility. We evaluate whether the response of GDP and the price level to exchange rate or interest rate shocks has changed over time, focusing on the period of the recent financial crisis. Our results suggest that prices have become increasingly responsive to monetary policy shocks. However, in terms of credible intervals, the stability of the monetary policy transmission mechanism in the Czech Republic cannot be rejected. Furthermore, it is demonstrated that the exchange rate pass-through has largely remained stable over time.

In the 1970s and 1980s, the economies of several developed and developing countries witnessed much volatility in output growth, inflation and other macroeconomic variables. Thereafter, these became considerably low and more stable. This latter development has been named the "great moderation" with first observation in the US and comparable evidence for other developed and developing economies including South Africa. However, the 2007/2009 economic and financial crisis resulted into heightened uncertainty and volatility. This brings to question the role of monetary policy and its effectiveness in the economy. This paper examines the transmission mechanism of shocks to monetary policy in South Africa using quarterly data from 1980:1 to 2012:4. We also in addition identify demand and supply shocks. Our analyses are based on a factor-augmented vector autoregression with time-varying coefficients and stochastic volatility (TVP-FAVAR), which allows us to simultaneously analyze the changing impulse responses of a set of 177 macroeconomic variables covering the inflation, real activity, asset prices and monetary series. We also have intangible variables, such as confidence indices, and survey variables. These data capture the broad trends in the South African economy. Our results based on the impulse response functions, are consistent with economic theory as we observe no price puzzle that is often associated with the standard VAR models. We find evidence of modest time variation in the transmission of shocks. Overall, the macroeconomic variables seemed to have responded slightly more to the monetary policy shocks in the post -2000 (inflation targeting) sub-period than the pre-2000 period, albeit the differences in the effects are statistically insignificant. The forecast error variance decomposition results show the changes in the macroeconomic variables are largely determined by the demand shock relative to the monetary policy shock although the contribution of the latter increased slightly over time. Our results suggest the need for a more efficient role of the monetary authority as this will both improve its credibility and greater economic stability. As inflation still remains within the upper portion of the 3–6% target range, appropriate mix of supply and demand side policies could be explored alongside monetary policy to reduce inflationary pressures.

This paper examines the spillover effects of global shocks to domestic output in Nigeria. The study becomes important now due to the unprecedented global events that shaped several economies and severely affected most countries in the world within the last two decades, including lower-middle income countries like Nigeria. To achieve the objectives of the study, we employ a simple Bayesian Time-Varying Parameter Structural Vector Autoregressive Model (B-TVP-SVAR) with Stochastic Volatility, using monthly data series from 2000 to 2022. The aim is to assess the ex-ante and ex-post of the spillover effects of global shocks to the domestic economy. Thus, we consider two distinct episodes and their respective impacts and deduce policy measures on how to moderate the impacts of similar shocks in the future. The episodes are the 2007/2008 Global Financial Crisis (GFC) and its spillover effects to Nigeria and the Global COVID-19 pandemic of 2020 as well. Subsequently, we establish that the impact of COVID-19 pandemic was more severe to the Nigerian economy relative to the 2007/2008 GFC. Finally, in addition to the measures proffered in this study on how to navigate Nigeria’s economy toward economic development and prosperity, we recommend both monetary and fiscal policy options that would serve as buffers to moderate the impacts of future unanticipated global shocks when they occur.

This note corrects a mistake in the estimation algorithm of the time-varying structural vector autoregression model of Primiceri (2005) and proposes a new algorithm that correctly applies the procedure proposed by Kim, Shephard, and Chib (1998) to the estimation of VAR or DSGE models with stochastic volatility. Relative to Primiceri (2005), the correct algorithm involves a different ordering of the various Markov Chain Monte Carlo steps.

This paper presents an empirical investigation of the nonlinear endogenous linkage between economic policy uncertainty (EPU), investor sentiment, and exchange rate volatility, utilizing the Time-Varying Parameter Structural Vector Autoregressive Model with Stochastic Volatility (TVP-SV-VAR). The study demonstrates that the effects of economic policy uncertainty and investor sentiment on exchange rate volatility are time-varying. Economic policy uncertainty not only directly influences the exchange rate but also exerts an indirect “superimposed” effect by influencing investor sentiment. These findings offer valuable insights for policymakers aiming to stabilize investor sentiment and maintain fluctuations in the renminbi (RMB) exchange rate within a reasonably balanced range.

Time-varying price shock transmission and volatility spillover in foreign exchange, bond, equity, and commodity markets: Evidence from the United States

Dynamic connectedness of inflation around the world: A time-varying approach from G7 and E7 countries

This study investigates the impact of environmental variables, such as carbon emissions and temperature anomalies, on cryptocurrency returns. While existing research has primarily focused on economic and financial determinants, the influence of environmental factors remains underexplored. Using Dynamic Conditional Correlation GARCH (DCC-GARCH) and Time-Varying Coefficients Vector Autoregression (TVC-VAR) models, this study provides empirical evidence that environmental variables significantly affect the volatility and returns of Bitcoin, Ethereum, and Tether. The results show that Bitcoin and Ethereum are highly sensitive to CO2 emissions and temperature fluctuations, while Tether demonstrates a more moderate response. Moreover, the impact of these environmental factors evolves over time, underscoring their dynamic nature in cryptocurrency valuation. These findings highlight the importance of incorporating environmental variables into forecasting models to enhance risk management and investment strategies. This study contributes to the literature by bridging the gap between environmental concerns and cryptocurrency market behavior, offering valuable insights for investors, regulators, and policymakers.

Time series models are widely used to examine temporal dynamics and uncover patterns across diverse fields. A commonly employed approach for modeling such data is the (Vector) Autoregressive (AR/VAR) model, in which each variable is represented as a linear combination of its own and others’ lagged values. However, the traditional (V)AR framework relies on the key assumption of stationarity, that autoregressive coefficients remain constant over time, which is often violated in practice, especially in systems affected by structural breaks, seasonal fluctuations, or evolving causal mechanisms. To overcome this limitation, Time-Varying (Vector) Autoregressive (TV-AR/TV-VAR) models have been developed, enabling model parameters to evolve over time and thus better capturing non-stationary behavior. Conventional approaches to estimating such models, including generalized additive modeling and kernel smoothing techniques, often require strong assumptions about basis functions, which can restrict their flexibility and applicability. To address these challenges, we introduce a novel framework that leverages physics-informed neural networks (PINN) to model TV-AR/TV-VAR processes. The proposed method extends the PINN framework to time series analysis by reducing reliance on explicitly defined physical structures, thereby broadening its applicability. Its effectiveness is validated through simulations on synthetic data and an empirical study of real-world health-related time series.