Financial Series Data Research Articles

ANALYZING FINANCIAL MARKETS EFFICIENCY: INSIGHTS FROM A BIBLIOMETRIC AND CONTENT REVIEW

The nonlinear nature of financial data series and the intricate incorporation of data into market prices necessitate a comprehensive exploration of key research findings, prevailing trends, intense debates, and subfields in the market behavior realm. Studies exploring the way in which technical analysis can exploit the deviation from market efficiency in stock markets, based on new prediction techniques (machine learning, deep learning, and artificial intelligence), are lacking. This study presents a comprehensive bibliometric assessment of market behavior using the Scopus database from 1972 to 2022. A thorough assessment process, which included keywords, filters, and data cleaning, was employed to narrow down the literature from 30,551 to 8,289 relevant papers. The research framework delineates seven primary themes that underpin this study: market efficiency, behavioral finance, technical analysis, volatility, fractals, asset pricing, and price discovery. For practitioners, investors, and policymakers, our study presents evidence regarding emerging themes, such as technical analysis, adaptive market hypothesis, and machine learning, which diverges from the findings of the proponents of equilibrium models based on investors' rationality. Moreover, an in-depth inquiry into the role of technical analysis in shaping portfolio investment presents a promising future research avenue.

An Empirical Study for Comparison of Estimation Methods for Value at Risk, Tail Value at Risk, and Adjusted Tail Value at Risk Using Extreme Value Theory Approach to Stock Market Index

Risk management helps the financial industry to manage and estimate the risks that may occur by using risk measures. Financial series data mostly have a heavy tail distribution which causes the probability of extreme values to occur. To overcome these extreme values, it is necessary to apply a mathematical model in calculating risk estimates in financial data combined with the Extreme Value Theory approach. The Adjusted-TVaR model is a measure of the risk of modification of the TVaR model to eliminate outliers in the tail of the distribution. The purpose of this study is to measure the accuracy of the Value at Risk, Tail Value at Risk, and Adjusted Tail Value at Risk using the Peak Over Thresholdapproach in Extreme Value Theory Models.The results of the risk estimation research using the POT approach method, show that the higher the level of confidence and the chosen constant, the higher the value of Adj-TVaR presented. This value represents that the potential loss will be higher. The estimation results obtained that the VaR value is smaller than Adj-TVaR and Adj-TVaR is smaller than TVaR. This shows that Adj-TVaR is more efficient to use in terms of predicting risk value when compared to TVaR with the Peak Over Threshold approach

Nonlinear Volatility Risk Prediction Algorithm of Financial Data Based on Improved Deep Learning

With the gradual integration of global economy and finance, the financial market presents many complex financial phenomena. To increase the prediction accuracy of financial data, a new nonlinear volatility risk prediction algorithm is proposed based on the improved deep learning algorithm. First, the financial data are taken as the research object and the closing price is set as the prediction target. Then, the nonlinear volatility risk prediction model of the financial data is established through the wavelet principal component analysis noise reduction module and the long and short‐term memory network (LSTM) module, and the nonlinear volatility trend is extracted from multiple financial data series to realize the nonlinear volatility risk prediction of the financial data. During the whole experiment, the time of the research method was less than 1.5 minutes. And for 1200 test samples, the average error of data risk prediction of the proposed method is 0.0217%. The average cost of the research method is 114.25 million yuan, which is significantly lower than other existing algorithms. Experimental results show that the research method can effectively predict the risk of financial data and is more suitable for the risk control early warning of Internet financial platform.

FORECASTING MODEL OF AGRICULTURE COMMODITY OF VALUE EXPORT OF COFFEE; APPLICATION OF ARIMA MODEL

Indonesia is currently one of the largest coffee producers in the world, and involved in exporting coffee countries. The financial series data such export value of coffee is highly volatile in both mean and variance. Thereby, the model of ARIMA with order p,d,q is one way to deal with this error. The aim of this study is to determine the best-fitted ARIMA(p,d,q) model to forecast the monthly series of export of coffee from January 2005 to April 2020. The findings suggest that ARIMA(1,3,1) is the best-selected model due to its very significant p-value (less than 0.0001), which showed that the model is applicable for forecasting. The model is then used to establish the prediction of ExCof monthly data for the next 12 months.

An ABC approach for CAViaR models with asymmetric kernels

ABSTRACTThe Value at Risk (VaR) is a risk measure that is widely used by financial institutions to allocate risk. Optimal conditional VaR estimates are typically generated using a likelihood function based on the check-loss function. However, issues such as VaR's bias estimation and asymmetric financial decision-making, based on the sign of the forecast error, can result in the use of combined losses or of intractable likelihood functions. In such cases, likelihood function intractability gives ground for Bayesian inference using likelihood-free methods such as the Approximate Bayesian Computation (ABC) one. This method generates posterior estimates when the likelihood function is analytically unavailable. Here, we introduce a novel ABC-MCMC algorithm based on asymmetric kernel density functions that allows for the asymmetric decision-making rule used in VaR estimation. We illustrate this method in CAViaR models where VaR is forecast using simulated and real financial data series.

On the use of power transformations in CAViaR models

AbstractValue at risk (VaR) is a risk measure widely used by financial institutions in allocating risk. VaR forecast estimation involves the conditional evaluation of quantiles based on the currently available information. Recent advances in VaR evaluation incorporate a proxy for conditional variance, yielding the conditional autoregressive VaR (CAViaR) models. However, early work in finance literature has shown that the introduction of power transformations has resulted in improvements in volatility forecasting. Having a direct association between volatility and conditional VaR, we adopt power‐transformed CAViaR models. We investigate whether the flexible conditional VaR dynamics associated with power‐transformed CAViaR models can result in better forecasting results than those assumed by the nontransformed CAViaR models. Estimation in CAViaR models is based on an early‐rejection Markov chain Monte Carlo algorithm. We illustrate our forecasting evaluation results using simulated and financial daily return data series. The results demonstrate that there is strong evidence that supports the use of power‐transformed CAViaR models when forecasting VaR.

Consumer financial vulnerability: identifying transmission linkages that could give rise to higher levels of consumer financial vulnerability

Several macroeconomic indicators point to high consumer financial vulnerability in South Africa. These include, inter alia, a relatively high household debt-to-disposable income ratio, household consumption expenditure outstripping household disposable income and a declining real household net wealth-to-disposable income ratio. 12In a 2009 study, the first level of possible predictors of consumer financial vulnerability was identified. However, no study has been conducted in South Africa to establish the transmission path of consumer financial vulnerability. This paper attempts to identify such a transmission path by determining the order in which the four aspects of the consumer financial vulnerability index, namely consumer income, expenditure, savings and debt servicing vulnerability, impact on one another, making consumers more vulnerable. This was done by means of an econometric modelling technique called Vector Auto regression (VAR) using consumer financial vulnerability data series covering the period Q2 2009 to Q2 2012. 13The VAR results show that expenditure vulnerability received the highest coefficient of determination score. This indicates that expenditure problems are the Achilles’ heel of South African households, which activates the postulated consumer financial vulnerability index (CFVI) transmission path. To determine the extent to which other macroeconomic variables impact on the postulated CFVI transmission path, a consumer price index (CPI) time series was entered exogenously into the existing VAR equation. It appears from the results obtained that the exogenous 113 Consumer financial vulnerability inclusion of CPI in the model made a dramatic difference with respect to income and expenditure vulnerability. By including the prime lending rate variable exogenously in the CFVI transmission path, the strong impact of the prime rate on expenditure vulnerability became evident. Finally, by adding the expanded unemployment variable exogenously to the CFVI transmission path in addition to the CPI and prime rate variables, debt servicing vulnerability was strongly impacted. From the CFVI transmission path findings, it became evident that consumers are not able to afford their required necessities, which leads to their becoming expenditure vulnerable. If consumers cannot generate more income to compensate, they become income vulnerable. They draw on their savings to finance the excess expenditure and become savings vulnerable, and if they cannot afford the necessary credit they require to finance their expenditure and have no savings left, they become debt servicing vulnerable

The use of the Hurst exponent to investigate the quality of forecasting methods of ultra-high-frequency data of exchange rates

Over the last century a variety of methods have been used for forecasting financial time data series with different results. This article explains why most of them failed to provide reasonable results based on fractal theory using one day tick data series from the foreign exchange market. Forecasting AMAPE errors and forecasting accuracy ratios were calculated for statistical and machine learning methods for currency time series which were divided into sub-segments according to Hurst ratio. This research proves that the forecasting error decreases and the forecasting accuracy increases for all of the forecasting methods when the Hurt ratio increases. The approach which was used in the article can be successfully applied to time series forecasting by indicating periods with the optimal values of the Hurst exponent.

Which leading indicators have done better at signaling past recessions?

In this article, I analyze a broad range of leading indicators?economic or financial data series that change in advance of the rest of the economy?to see which ones have done better at signaling past U.S. recessions.1 I also use these leading indicators to form a new index that outperforms existing leading indexes and the Treasury yield curve at signaling historical downturns.

Macro Factors in Corporate Bond Credit and Liquidity Spreads

This paper studies the macroeconomic determinants of the term structures of Treasury yields, corporate bond credit spreads, and corporate bond liquidity spreads in a unified no-arbitrage framework. Four economic factors, monetary conditions, inflation, real output, and financial market volatility, are extracted from a set of macroeconomic and financial data series. During the pre-crisis period, volatility shocks decrease Treasury yields and widen both credit spreads and liquidity spreads for all rating classes, and credit spreads widen as monetary conditions tighten, but the effects of inflation and real output are insignificant. In times of stress, financial market volatility has a similar impact and the impacts of inflation and real output become significant as well. Ignoring the liquidity component of corporate yield spreads is shown to lead to inaccurate estimation of the impacts of economic factors on corporate credit spreads. The paper also provides evidence of ”flight-to-liquidity” behavior which strengthens in bad times and sheds light on the negative correlation between the risk-free rate and corporate yield spreads as well as on the positive correlation between credit spreads and liquidity spreads.

A Bayesian encompassing test using combined value-at-risk estimates

The Value at Risk (VaR) is a risk measure that is widely used by financial institutions in allocating risk. VaR forecast estimation involves the conditional evaluation of quantiles based on the currently available information. Recent advances in VaR evaluation incorporate conditional variance into the quantile estimation, yielding the Conditional Autoregressive VaR (CAViaR) models. However, the large number of alternative CAViaR models raises the issue of identifying the optimal quantile predictor. To resolve this uncertainty, we propose a Bayesian encompassing test that evaluates various CAViaR models predictions against a combined CAViaR model based on the encompassing principle. This test provides a basis for forecasting combined conditional VaR estimates when there are evidences against the encompassing principle. We illustrate this test using simulated and financial daily return data series. The results demonstrate that there are evidences for using combined conditional VaR estimates when forecasting quantile risk.

Mixed-Frequency Macro-Financial Spillovers

We develop a new methodology to analyse spillovers between the real and financial sides of the economy that employs a mixed-frequency modelling approach. This enables high-frequency financial and low-frequency macroeconomic data series to be employed directly, avoiding the data aggregation and information loss incurred when using common-frequency methods. In a detailed analysis of macro-financial spillovers for the US economy, we find that the additional high-frequency information preserved by our mixed-frequency approach results in estimated spillovers that are typically substantially higher than those from an analogous common-frequency approach and are more consistent with known in-sample events. We also show that financial markets are typically net transmitters of shocks to the real side of the economy, particularly during turbulent market conditions, but that the bond and equity markets act heterogeneously in both transmitting and receiving shocks to the non-financial sector. We observe substantial short and medium-run variation in macro-financial spillovers that is statistically associated with key variables related to financial and macroeconomic fundamentals; the values of the term spread, VIX and unemployment rate in particular appear to be important determinants of macro- financial spillovers.

Adjusting the errors of the GM(1, 2) grey model in the financial data series using an adaptive fuzzy controller

Purpose The purpose of this paper is to focus on the adjustment of the GM(1, 2) errors for financial data series that measures changes in the public sector financial indicators, taking into account that the errors in grey models remain a key problem in reconstructing the original data series. Design/methodology/approach Adjusting the errors in grey models must follow some rules that most often cannot be determined based on the chaotic trends they register in reconstructing data series. In order to ensure the adjustment of these errors, for improving the robustness of GM(1, 2), was constructed an adaptive fuzzy controller which is based on two input variables and one output variable. The input variables in the adaptive fuzzy controller are: the absolute error ε i 0 ( k ) [ % ] of GM(1, 2), and the distance between two values x i 0 ( k ) [ % ] , while the output variable is the error adjustment A ε i 0 ( k ) [ % ] determined with the help of the above-mentioned input variables. Findings The adaptive fuzzy controller has the advantage that sets the values for error adjustments by the intensity (size) of the errors, in this way being possible to determine the value adjustments for each element of the reconstructed financial data series. Originality/value To ensure a robust process of planning the financial resources, the available financial data are used for long periods of time, in order to notice the trend of the financial indicators that need to be planned. In this context, the financial data series could be reconstituted using grey models that are based on sequences of financial data that best describe the status of the analyzed indicators and the status of the relevant factors of influence. In this context, the present study proposes the construction of a fuzzy adaptive controller that with the help of the output variable will ensure the error’s adjustment in the reconstituted data series with GM(1, 2).

Variational Bayes for assessment of dynamic quantile forecasts

Recently, various Bayes factor analogues of frequentist tests for the accuracy of dynamic quantile forecasts have been developed. However, in evaluating the marginal likelihoods involved, either inappropriate assumptions have been made, or pre-packaged multivariate adaptive quadrature methods have been employed, without an accuracy assessment. This paper develops variational Bayes methods for estimating lower bounds for these marginal likelihoods efficiently. This facilitates a more accurate version of one existing Bayesian test, and allows for the development of a new test based on the probit regression model. The size and power properties of the proposed methods are examined via a simulation study, illustrating favourable comparisons with existing testing methods. The accuracy and speed of the VB methods are also assessed. An empirical study illustrates the sensible performance and applicability of the proposed methods, relative to standard tests, for assessing the adequacy of a range of forecast models for the Value at Risk (VaR) in several financial market data series.

Bayesian Assessment of Dynamic Quantile Forecasts

Bayesian methods for assessing the accuracy of dynamic financial value‐at‐risk (VaR) forecasts have not been considered in the literature. Such methods are proposed in this paper. Specifically, Bayes factor analogues of popular frequentist tests for independence of violations from, and for correct coverage of a time series of, dynamic quantile forecasts are developed. To evaluate the relevant marginal likelihoods, analytic integration methods are utilized when possible; otherwise multivariate adaptive quadrature methods are employed to estimate the required quantities. The usual Bayesian interval estimate for a proportion is also examined in this context. The size and power properties of the proposed methods are examined via a simulation study, illustrating favourable comparisons both overall and with their frequentist counterparts. An empirical study employs the proposed methods, in comparison with standard tests, to assess the adequacy of a range of forecasting models for VaR in several financial market data series. Copyright © 2016 John Wiley & Sons, Ltd.

Biclustering Learning of Trading Rules.

Technical analysis with numerous indicators and patterns has been regarded as important evidence for making trading decisions in financial markets. However, it is extremely difficult for investors to find useful trading rules based on numerous technical indicators. This paper innovatively proposes the use of biclustering mining to discover effective technical trading patterns that contain a combination of indicators from historical financial data series. This is the first attempt to use biclustering algorithm on trading data. The mined patterns are regarded as trading rules and can be classified as three trading actions (i.e., the buy, the sell, and no-action signals) with respect to the maximum support. A modified K nearest neighborhood ( K -NN) method is applied to classification of trading days in the testing period. The proposed method [called biclustering algorithm and the K nearest neighbor (BIC- K -NN)] was implemented on four historical datasets and the average performance was compared with the conventional buy-and-hold strategy and three previously reported intelligent trading systems. Experimental results demonstrate that the proposed trading system outperforms its counterparts and will be useful for investment in various financial markets.

An index of financial market stress for the United Kingdom

We construct and develop a new financial market stress index using twenty-three headline U.K. financial data series. A logistic regression framework provides a parsimonious representation of financial market stress in the U.K. based on the market dynamics around the time of Bank of England crisis-alleviating economic interventions. Our results present clear evidence that the Bank of England’s swift and decisive actions stemmed financial market stress as measured by the stress index.

An Index of Financial Market Stress for the United Kingdom

We construct and develop a new financial market stress index using twenty-three headline U.K. financial data series. A logistic regression framework provides a parsimonious representation of financial market stress in the U.K. based on the market dynamics around the time of Bank of England crisis-alleviating economic interventions. Our results present clear evidence that the Bank of England’s swift and decisive actions stemmed financial market stress as measured by the stress index.

A quasi-Bayesian model averaging approach for conditional quantile models

The value at risk (VaR) is a risk measure that is widely used by financial institutions to allocate risk. VaR forecast estimation involves the evaluation of conditional quantiles based on the currently available information. Recent advances in VaR evaluation incorporate conditional variance into the quantile estimation, which yields the conditional autoregressive VaR (CAViaR) models. However, uncertainty with regard to model selection in CAViaR model estimators raises the issue of identifying the better quantile predictor via averaging. In this study, we propose a quasi-Bayesian model averaging method that generates combinations of conditional VaR estimators based on single CAViaR models. This approach provides us a basis for comparing single CAViaR models against averaged ones for their ability to forecast VaR. We illustrate this method using simulated and financial daily return data series. The results demonstrate significant findings with regard to the use of averaged conditional VaR estimates when forecasting quantile risk.

A fractional perspective to financial indices

The application of mathematical methods and computer algorithms in the analysis of economic and financial data series aims to give empirical descriptions of the hidden relations between many complex or unknown variables and systems. This strategy overcomes the requirement for building models based on a set of ‘fundamental laws’, which is the paradigm for studying phenomena usual in physics and engineering. In spite of this shortcut, the fact is that financial series demonstrate to be hard to tackle, involving complex memory effects and a apparently chaotic behaviour. Several measures for describing these objects were adopted by market agents, but, due to their simplicity, they are not capable to cope with the diversity and complexity embedded in the data. Therefore, it is important to propose new measures that, on one hand, are highly interpretable by standard personal but, on the other hand, are capable of capturing a significant part of the dynamical effects.