Search

AbstractWe derive a new fractional Heston model with self-exciting jumps. We study volatility persistence and demonstrate that the quadratic variation necessarily exhibits less memory than the integrated variance, which preserves the degree of long-memory of the instantaneous volatility. Focusing on realized volatility measures, we find that traditional long-memory estimators are dramatically downward biased, in particular for low-frequency intraday sampling. Conveniently, our Monte Carlo experiments reveal that some noise-robust local Whittle-type estimators offer good finite sample properties. We apply our theoretical results in a risk forecasting study and show that our frequency-domain forecasting procedure outperforms the traditional benchmark models. [ABSTRACT FROM AUTHOR]

In this paper, we propose a novel approach to improving forecasts of stock market indexes by considering common stock prices as hierarchical time series, combining clustering with forecast reconciliation. We propose grouping the individual stock price series in various ways including via metadata and using unsupervised learning techniques. The proposed approach is applied to the Dow Jones Industrial Average Index and the Standard & Poor 500 Index and their component stocks, and the results obtained with different grouping approaches are compared. The results empirically demonstrate that the combined use of clustering and reconciliation improves the forecast accuracy of the stock market indexes and their constituents. [ABSTRACT FROM AUTHOR]

In this article, we use a deep quantile estimator, based on neural networks and their universal approximation property to examine a non-linear association between the conditional quantiles of a dependent variable and predictors. This methodology is versatile and allows both the use of different penalty functions, as well as high dimensional covariates. We present a Monte Carlo exercise where we examine the finite sample properties of the deep quantile estimator and show that it delivers good finite sample performance. We use the deep quantile estimator to forecast value-at-risk and find significant gains over linear quantile regression alternatives and other models, which are supported by various testing schemes. Further, we consider also an alternative architecture that allows the use of mixed frequency data in neural networks. This article also contributes to the interpretability of neural network output by making comparisons between the commonly used Shapley Additive Explanation values and an alternative method based on partial derivatives. [ABSTRACT FROM AUTHOR]

Graphical models are a powerful tool to estimate a high-dimensional inverse covariance (precision) matrix, which has been applied for a portfolio allocation problem. The assumption made by these models is a sparsity of the precision matrix. However, when stock returns are driven by common factors, such assumption does not hold. We address this limitation and develop a framework, Factor Graphical Lasso (FGL), which integrates graphical models with the factor structure in the context of portfolio allocation by decomposing a precision matrix into low-rank and sparse components. Our theoretical results and simulations show that FGL consistently estimates the portfolio weights and risk exposure and also that FGL is robust to heavy-tailed distributions which makes our method suitable for financial applications. FGL-based portfolios are shown to exhibit superior performance over several prominent competitors including equal-weighted and index portfolios in the empirical application for the S&P500 constituents. [ABSTRACT FROM AUTHOR]

Abstract The spread of the coronavirus has reduced the value of stock indexes, depressed energy and metals commodities prices including oil, and caused instability in financial markets around the world. Due to this situation, investors should consider investing in more secure assets, such as real estate property, cash, gold, and crypto assets. In recent years, among secure assets, cryptoassets are gaining more attention than traditional investments. This study compares the Bitcoin market, the gold market, and American stock indexes (S&P500, Nasdaq, and Dow Jones) before and during the COVID-19 pandemic. For this purpose, the dynamic conditional correlation exponential generalized autoregressive conditional heteroskedasticity model was used to estimate the DCC coefficient and compare this model with the artificial neural network approach to predict volatility of these markets. Our empirical findings showed a substantial dynamic conditional correlation between Bitcoin, gold, and stock markets. In particular, we observed that Bitcoin offered better diversification opportunities to reduce risks in key stock markets during the COVID-19 period. This paper provides practical impacts on risk management and portfolio diversification.

We consider inference for predictive regressions with multiple predictors. Extant tests for predictability (especially for joint predictability) may perform unsatisfactorily and tend to discover spurious predictability as the number of predictors increases. We propose a battery of new instrumental variables-based tests which involve enforcement or partial enforcement of the null hypothesis in variance estimation. A test based on the few-predictors-at-a-time parsimonious system approach is recommended. Empirical Monte Carlos demonstrates the remarkable finite-sample performance regardless of numerosity of predictors and their persistence properties. Empirical application to equity premium predictability is provided. [ABSTRACT FROM AUTHOR]

We analyze the decomposition of the conditional, rather than the unconditional, variance of market returns based on an extension of the standard Campbell–Shiller approach. The relative importance of cash flow and discount rate news in determining the conditional variance of market returns exhibits significant variation over time and relates to economic conditions. The components of the conditional market variance outperform several benchmark variables, including the conditional market variance itself, in forecasting future market returns and realized variance across different horizons. The forecasts based on the conditional market variance components also provide sizable economic benefits compared with benchmark forecasts in an out-of-sample portfolio exercise where a myopic investor allocates her wealth between the market portfolio and a risk-free asset across different holding periods. [ABSTRACT FROM AUTHOR]

The exact conditions under which volatility timing strategies yield value are documented. These conditions include: the ability to correctly forecast next period stochastic variance, and violation of a strict version of Merton's intertemporal capital asset pricing model (ICAPM). While the empirical evidence supports the first of these conditions, the latter remains open to debate. Our empirical results confirm the former, but demonstrate a significant violation of the (strict) ICAPM. It follows that volatility timing strategies appear to have value. However, using reasonable parameter values plugged into the derived formulae, the results also show that extreme leverage is often required for success. A method of tempering leverage is proposed, which is somewhat able to loosen the requirement of high leverage while still maintaining a good performance level. Given the likely variation in (strict) ICAPM violations across time and assets, it follows that volatility timing success (or failure) is very much sample dependent. [ABSTRACT FROM AUTHOR]

We perform a comparative analysis of machine learning methods for the canonical problem of empirical asset pricing: measuring asset risk premiums. We demonstrate large economic gains to investors using machine learning forecasts, in some cases doubling the performance of leading regression-based strategies from the literature. We identify the best-performing methods (trees and neural networks) and trace their predictive gains to allowing nonlinear predictor interactions missed by other methods. All methods agree on the same set of dominant predictive signals, a set that includes variations on momentum, liquidity, and volatility. [ABSTRACT FROM AUTHOR]

We investigate the effect of estimation error on backtests of expected shortfall (ES) forecasts. These backtests are based on first-order conditions of a recently introduced family of jointly consistent loss functions for value-at-risk (VaR) and ES. For both single and multiperiod horizons, we provide explicit expressions for the additional terms in the asymptotic covariance matrix that result from estimation error, and propose robust tests that account for it. Monte Carlo experiments show that the tests that ignore these terms suffer from size distortions, which are more pronounced for higher ratios of out-of-sample to in-sample observations. Robust versions of the backtests perform well with power against common alternatives. We also introduce a novel standardization of the conditional joint test statistic that removes the need to estimate higher-order moments and significantly improves its performance. In an application to VaR and ES forecasts for daily FTSE 100 index returns as generated by (GJR-)GARCH and HEAVY models, we find that estimation error substantially impacts the outcome of the backtests, and is not bound to particular subperiods such as the credit crisis. [ABSTRACT FROM AUTHOR]

At long horizons, multiplicative compounding induces strong-to-extreme positive skewness into stock returns; the magnitude of the effect is primarily determined by single-period volatility. Consequently, at horizons greater than 5 years, returns—individual or portfolio—will be positively skewed under reasonable parameterizations. From an investor perspective, the strong positive skewness implies that the mean compound return will serve as a poor guide for typical long-horizon outcomes. Moreover, the large effects of compounding on higher-order moments are shown to affect the validity of Taylor expansions used to approximate preferences for skewness, when applied to returns of annual or longer horizons. [ABSTRACT FROM AUTHOR]

We build a parsimonious agent-based model under the adaptive market hypothesis (AMH), which can explain the formation of equilibrium prices and market efficiency dynamics. Our model combines heterogeneous interacting agents, switching behavior, and investor feedback on past realized returns, with realistic assumptions on the market microstructure. Numerical simulations show that our model is empirically robust to the facts observed in returns of developed and emerging financial markets (leptokurtic distribution, excess volatility, time-varying linear and nonlinear autocorrelations in returns, and time-varying degree of market efficiency). These results reveal that the elements in the model are necessary for generating these empirical facts. They also confirm the AMH. The model can be used as an artificial laboratory to assess the effectiveness of regulatory policies and the profitability of trading strategies.

The Virtual Center for Paradigmatic Studies is a serious profession and, at the same time, a beneficial offer addressed to people who treat the process of their education and professional development both at home and in the academic environment as a unique life opportunity. This article presents material constituting the program and the basis for conducting theoretical and practical work at the Virtual Center for Paradigmatic Studies. The published material covers such issues as: the description of the Gorbatov's characterization principle, the theoretical correctness of the model of functioning, an example of the Semantic Model of 5 Projects, effective search for forbidden graph figures, the essence of functions that integrate semantic graphs, multi-object semantic sequences of events, the structure and purpose of Hasse diagrams.

In this paper the authors investigate the statistical properties of some cryptocurrencies by using three layers of analysis: alpha-stable distributions, Metcalfe’s law and the bubble behaviour through the LPPL modelling. The results show, in the medium to long-run, the validity of Metcalfe's law (the value of a network is proportional to the square of the number of connected users of the system) for the evaluation of cryptocurrencies; however, in the short-run, the validity of Metcalfe’s law for Bitcoin is questionable. According to the bidirectional causality between the price and the network size, the expected price increase is a driver for more investors to join the Bitcoin network, which may lead in the end to a super-exponential price growth, possibly due to a herding behaviour of investors. The authors then used LPPL models to capture the behaviour of cryptocurrencies exchange rates during an endogenous bubble and to predict the most probable time of the regime switching. The main conclusion of this paper is that Metcalfe’s law may be valid in the long-run, however in the short-run, on various data regimes, its validity is highly debatable.

We apply seven alternative t-distributions to estimate the market risk measures Value at Risk (VaR) and its extension Expected Shortfall (ES). Of these seven, the twin t-distribution (TT) of Baker and Jackson (in Twin t distribution, University of Salford Manchester. https://arxiv.org/abs/1408.3237, 2014) and generalized asymmetric distribution (GAT) of Baker (in A new asymmetric generalization of the t-distribution, University of Salford Manchester. https://arxiv.org/abs/1606.05203, 2016) are applied for the first time to estimate market risk. We analytically estimate VaR and ES over 1-day horizon and extend this to multi-day horizon using Monte Carlo simulation. We find that taken together TT and GAT distributions provide the best back-testing results across individual confidence levels and horizons for majority of scenarios. Moreover, we find that with the lengthening of time horizon, TT and GAT models performs well, such that at the 10-day horizon, GAT provides the best back-testing results for all of the five indices and the TT model provides the second best results, irrespective period of study and confidence level. [ABSTRACT FROM AUTHOR]

The modelling of the hard-to-model risks factors is one of the topics of great interest to the financial industry. The industry is spending lots of resources on efforts to account for the hard-to-model risks in their risk management frameworks. Currently, the concept describing these risks is the Risk Not in VaR. In its turn, the newly composed Fundamental Review of the Trading Book text similarly prescribes to classify risk factors that do not have a history of continuously available real prices as non-modellable risk factors. Both entities and financial regulatory authorities have shown great concern in the search for efficient techniques and models that allow for a more accurate estimation of the risks factors linked to the derivatives. An accurate modelling of these risk factors can lead to considerable optimization in the capital charges, but any model assumption must be duly justified and supported by the entities. In this paper, the (Multichannel) Singular Spectrum Analysis for modelling these risk factors is analysed. [ABSTRACT FROM AUTHOR]

Enhanced machine learning methods provide an encouraging alternative to forecast asset prices by extending or generalizing the possible model specifications compared to conventional linear regression methods. Even if enhanced methods of machine learning in the literature often lead to better forecasting quality, this is not clear for small asset classes, because in small asset classes enhanced machine learning methods may potentially over-fit the in-sample data. Against this background, we compare the forecasting performance of linear regression models and enhanced machine learning methods in the market for catastrophe (CAT) bonds. We use linear regression with variable selection, penalization methods, random forests and neural networks to forecast CAT bond premia. Among the considered models, random forests exhibit the highest forecasting performance, followed by linear regression models and neural networks. [ABSTRACT FROM AUTHOR]

This study investigates the effect of periodic events, such as the stock index futures and options expiration days and the Morgan Stanley Capital International (MSCI) quarterly index reviews, on the trading volume in the pan-European equity markets. The motivation of this study stems from anecdotal evidence of increased trading volume in the equity markets during the run-up to the index options and futures expiration days and MSCI rebalances. This study investigates this phenomenon in more detail and analyses the trading volumes of seven European stock indices and the MSCI International Pan-Euro Price Index. The analysis features a multi-step ahead volume forecast, which is important for practitioners in order to plan multi-day trades while looking to minimise the market impact. The results confirm higher trading activity on the futures and options expiration days, as well as on the MSCI rebalance day. We report a clear futures and options expiration day effect, which accounts for the Friday effect in terms of larger trading volumes. The MSCI rebalance trading volume is significantly different from the volume of the adjacent months with no MSCI reviews, but they cannot explain the end-of-month effect entirely. [ABSTRACT FROM AUTHOR]

This paper addresses the modeling and forecasting of daily high and low asset prices in the Brazilian stock market using a fractionally cointegrated vector autoregressive model (FCVAR). Forecasts are then used in a simple trading strategy to evaluate the application of technical analysis (TA) for equity shares traded at the B3. As a flexible framework, FCVAR is able to account for two fundamental patterns of high and low asset prices: their cointegrating relationship and the long-memory of their difference (i.e., the range), a measure of realized volatility. The analysis comprises the twenty most traded stocks at the B3 during the period from January 2010 to May 2017. Empirical findings indicate a significant cointegration relationship between daily high and low prices, which are integrated of an order close to the unity, as well as the range displays long memory and is in the stationary region in most of the cases. Based on historical data, results support that the high and low prices of equity shares are largely predictable and their forecasts can improve TA trading strategies applied on Brazilian equity shares. Further, the fractionally cointegrated approach appears as a potential forecasting tool for market practitioners on their investment strategies. [ABSTRACT FROM AUTHOR]

This paper evaluates the forecasting performance of a Brownian semi-stationary (BSS) process in modelling the volatility of 21 equity indices. We implement a hybrid scheme to simulate BSS processes with high efficiency and precision. These simulations are useful to price derivatives, accounting for rough volatility. We then calibrate the BSS parameters for the realised kernel of 21 equity indices, using data from the Oxford-Man Institute. Finally, we conduct one-step and ten-step ahead forecasts on six indices and find that the BSS outperforms benchmarks, including a Log-HAR specification, in most cases. [ABSTRACT FROM AUTHOR]