Skip to main content
main-content

Über dieses Buch

This book presents recent research on predictive econometrics and big data. Gathering edited papers presented at the 11th International Conference of the Thailand Econometric Society (TES2018), held in Chiang Mai, Thailand, on January 10-12, 2018, its main focus is on predictive techniques – which directly aim at predicting economic phenomena; and big data techniques – which enable us to handle the enormous amounts of data generated by modern computers in a reasonable time. The book also discusses the applications of more traditional statistical techniques to econometric problems.

Econometrics is a branch of economics that employs mathematical (especially statistical) methods to analyze economic systems, to forecast economic and financial dynamics, and to develop strategies for achieving desirable economic performance. It is therefore important to develop data processing techniques that explicitly focus on prediction. The more data we have, the better our predictions will be. As such, these techniques are essential to our ability to process huge amounts of available data.

Inhaltsverzeichnis

Frontmatter

Keynote Address

Frontmatter

Data in the 21st Century

The past couple of decades have witnessed exponential growth in data, due to the penetration of information technology across all aspects of science and society; the increasing ease with which we are able to collect more data; and the growth of Internet-scale, planet-wide Web-based and mobile services—leading to the notion of “big data”. While the emphasis so far has been on developing technologies to manage the volume, velocity, and variety of the data, and to exploit available data assets via machine learning techniques, going forward the emphasis must also be on translational data science and the responsible use of all of these data in real-world applications. Data science in the 21st century must provide trust in the data and provide responsible and trustworthy techniques and systems by supporting the notions of transparency, interpretability, and reproducibility. The future offers exciting opportunities for transdisciplinary research and convergence among disciplines—computer science, statistics, mathematics, and the full range of disciplines that impact all aspects of society. Econometrics and economics can find an important role in this convergence of ideas.

Chaitanya Baru

Fundamental Theory

Frontmatter

Model-Assisted Survey Estimation with Imperfectly Matched Auxiliary Data

Model-assisted survey regression estimators combine auxiliary information available at a population level with complex survey data to estimate finite population parameters. Many prediction methods, including linear and mixed models, nonparametric regression, and machine learning techniques, can be incorporated into such model-assisted estimators. These methods assume that observations obtained for the sample can be matched without error to the auxiliary data. We investigate properties of estimators that rely on matching algorithms that do not in general yield perfect matches. We focus on difference estimators, which are exactly unbiased under perfect matching but not under imperfect matching. The methods are investigated analytically and via simulation, using a study of recreational angling in South Carolina to build a simulation population. In this study, the survey data come from a stratified, two-stage sample and the auxiliary data from logbooks filed by boat captains. Extensions to regression estimators under imperfect matching are discussed.

F. Jay Breidt, Jean D. Opsomer, Chien-Min Huang

COBra: Copula-Based Portfolio Optimization

The meta-elliptical t copula with noncentral t GARCH univariate margins is studied as a model for asset allocation. A method of parameter estimation is deployed that is nearly instantaneous for large dimensions. The expected shortfall of the portfolio distribution is obtained by combining simulation with a parametric approximation for speed enhancement. A simulation-based method for mean-expected shortfall portfolio optimization is developed. An extensive out-of-sample backtest exercise is conducted and comparisons made with common asset allocation techniques.

Marc S. Paolella, Paweł Polak

Multiple Testing of One-Sided Hypotheses: Combining Bonferroni and the Bootstrap

In many multiple testing problems, the individual null hypotheses (i) concern univariate parameters and (ii) are one-sided. In such problems, power gains can be obtained for bootstrap multiple testing procedures in scenarios where some of the parameters are ‘deep in the null’ by making certain adjustments to the null distribution under which to resample. In this paper, we compare a Bonferroni adjustment that is based on finite-sample considerations with certain ‘asymptotic’ adjustments previously suggested in the literature.

Joseph P. Romano, Michael Wolf

Exploring Message Correlation in Crowd-Based Data Using Hyper Coordinates Visualization Technique

Analytical exploration for necessary information and insights from heterogeneous and multivariate dataset is challenging in visual analytics research due to the complexity of data and tasks. One of the data analytics target is to examine the relationship in the dataset, such as considering how the data elements and subsets are connected together. This work takes into account the direct and indirect connection relations: elements and subsets of elements might not only be directly linked together, but also possibly be indirectly associated via the relationships from other elements/subsets as well. Stream of messages instantly put on the cyberspace from the crowd is an example for such kind of dataset. In this paper, we present an approach to estimate the correlation between streaming messages collection in terms of large scale data processing, whilst the Hyper Coordinates visualization technique is designed to support those correlations exploration. The prototype tool is built to demonstrate the concepts for crowd-based data in the financial market domain.

Tien-Dung Cao, Dinh-Quyen Nguyen, Hien Duy Tran

Bayesian Forecasting for Tail Risk

This paper evaluates the performances of Value-at-Risk (VaR) and expected shortfall, as well as volatility forecasts in a class of risk models, specifically focusing on GARCH, integrated GARCH, and asymmetric GARCH models (GJR-GARCH, exponential GARCH, and smooth transition GARCH models). Most of the models incorporate four error probability distributions: Gaussian, Student’s t, skew Student’s t, and generalized error distribution (GED). We employ Bayesian Markov chain Monte Carlo sampling methods for estimation and forecasting. We further present backtesting measures for both VaR and expected shortfall forecasts and implement two loss functions to evaluate volatility forecasts. The empirical results are based on the S&P500 in the U.S. and Japan’s Nikkei 225. A VaR forecasting study reveals that at the 1% level the smooth transition model with a second-order logistic function and skew Student’s t error compares most favorably in terms of violation rates for both markets. For the volatility predictive abilities, the EGARCH model with GED error is the best model in both markets.

Cathy W. S. Chen, Yu-Wen Sun

Smoothing Spline as a Guide to Elaborate Explanatory Modeling

Although there are substantial theoretical and empirical differences between explanatory modeling and predictive modeling, they should be considered as two dimensions. And predictive modeling can work as a “fact check” to propose improvements to existing explanatory modeling. In this paper, I use smoothing spline, a nonparametric calibration technique which is originally designed to intensify the predictive power, as a guide to revise explanatory modeling. It works for the housing value model of Harrison and Rubinfeld (1978) because the modified model is more meaningful and fits better to actual data.

Chon Van Le

Quantifying Predictive Uncertainty Using Belief Functions: Different Approaches and Practical Construction

We consider the problem of quantifying prediction uncertainty using the formalism of belief functions. Three requirements for predictive belief functions are reviewed, each one of them inducing a distinct interpretation: compatibility with Bayesian inference, approximation of the true distribution, and frequency calibration. Construction procedures allowing us to build belief functions meeting each of these three requirements are described and illustrated using simple examples.

Thierry Denœux

Kuznets Curve: A Simple Dynamical System-Based Explanation

In the 1950s, a future Nobelist Simon Kuznets discovered the following phenomenon: as a country’s economy improves, inequality first grows but then decreases. In this paper, we provide a simple dynamical system-based explanation for this empirical phenomenon.

Thongchai Dumrongpokaphan, Vladik Kreinovich

A Calibration-Based Method in Computing Bayesian Posterior Distributions with Applications in Stock Market

Finding effective methods to compute or estimate posterior distributions of model parameters is of paramount importance in Bayesian statistics. In fact, Bayesian inference has only been extraordinarily popular in applications after the births of efficient algorithms like the Monte Carlo Markov Chain. Practicality of posterior distributions depends heavily on the combination of likelihood functions and prior distributions. In certain cases, closed-form formulas for posterior distributions can be attained; in this paper, based on the theory of distortion functions, a calibration-like method to calculate explicitly the posterior distributions for three crucial models, namely the normal, Poisson and Bernoulli is introduced. The paper ends with some applications in stock market.

Dung Tien Nguyen, Son P. Nguyen, Uyen H. Pham, Thien Dinh Nguyen

How to Estimate Statistical Characteristics Based on a Sample: Nonparametric Maximum Likelihood Approach Leads to Sample Mean, Sample Variance, etc.

In many practical situations, we need to estimate different statistical characteristics based on a sample. In some cases, we know that the corresponding probability distribution belongs to a known finite-parametric family of distributions. In such cases, a reasonable idea is to use the Maximum Likelihood method to estimate the corresponding parameters, and then to compute the value of the desired statistical characteristic for the distribution with these parameters.In some practical situations, we do not know any family containing the unknown distribution. We show that in such nonparametric cases, the Maximum Likelihood approach leads to the use of sample mean, sample variance, etc.

Vladik Kreinovich, Thongchai Dumrongpokaphan

How to Gauge Accuracy of Processing Big Data: Teaching Machine Learning Techniques to Gauge Their Own Accuracy

When the amount of data is reasonably small, we can usually fit this data to a simple model and use the traditional statistical methods both to estimate the parameters of this model and to gauge this model’s accuracy. For big data, it is often no longer possible to fit them by a simple model. Thus, we need to use generic machine learning techniques to find the corresponding model. The current machine learning techniques estimate the values of the corresponding parameters, but they usually do not gauge the accuracy of the corresponding general non-linear model. In this paper, we show how to modify the existing machine learning methodology so that it will not only estimate the parameters, but also estimate the accuracy of the resulting model.

Vladik Kreinovich, Thongchai Dumrongpokaphan, Hung T. Nguyen, Olga Kosheleva

How Better Are Predictive Models: Analysis on the Practically Important Example of Robust Interval Uncertainty

One of the main applications of science and engineering is to predict future value of different quantities of interest. In the traditional statistical approach, we first use observations to estimate the parameters of an appropriate model, and then use the resulting estimates to make predictions. Recently, a relatively new predictive approach has been actively promoted, the approach where we make predictions directly from observations. It is known that in general, while the predictive approach requires more computations, it leads to more accurate predictions. In this paper, on the practically important example of robust interval uncertainty, we analyze how more accurate is the predictive approach. Our analysis shows that predictive models are indeed much more accurate: asymptotically, they lead to estimates which are $$\sqrt{n}$$ more accurate, where n is the number of estimated parameters.

Vladik Kreinovich, Hung T. Nguyen, Songsak Sriboonchitta, Olga Kosheleva

Quantitative Justification for the Gravity Model in Economics

The gravity model in economics describes the trade flow between two countries as a function of their Gross Domestic Products (GDPs) and the distance between them. This model is motivated by the qualitative similarity between the desired dependence and the dependence of the gravity force (or potential energy) between the two bodies on their masses and on the distance between them. In this paper, we provide a quantitative justification for this economic formula.

Vladik Kreinovich, Songsak Sriboonchitta

The Decomposition of Quadratic Forms Under Skew Normal Settings

In this paper, the decomposition properties of noncentral skew chi-square distribution is studied. A given random variable U having a noncentral skew chi-square distribution with $$k>1$$ degrees of freedom, can be partitioned into the sum of two independent random variables $$U_1$$ and $$U_2$$ such that $$U_1$$ has a noncentral skew chi-square distribution with 1 degree of freedom and $$U_2$$ has the noncentral chi-square distribution with $$k-1$$ degrees of freedom. Also if $$k>2$$, this partition can be modified into $$U=U_1+U_2$$, where $$U_1$$ has a noncentral skew chi-square distribution with 2 degrees of freedom and $$U_2$$ has a central chi-square distribution with $$k-2$$ degrees of freedom. The densities of noncentral skew chi-square distributions with 1 degree of freedom, 2 degrees of freedom, and $$k>2$$ degrees of freedom are derived, and their graphs are presented. For illustration of our main results, the linear regression model with skew normal errors is considered as an application.

Ziwei Ma, Weizhong Tian, Baokun Li, Tonghui Wang

Joint Plausibility Regions for Parameters of Skew Normal Family

The estimation of parameters is a challenge issue for skew normal family. Based on inferential models, the plausibility regions for two parameters of skew normal family are investigated in two cases, when either the scale parameter $$\sigma $$ or the shape parameter $$\delta $$ is known. For illustration of our results, simulation studies are proceeded.

Ziwei Ma, Xiaonan Zhu, Tonghui Wang, Kittawit Autchariyapanitkul

On Parameter Change Test for ARMA Models with Martingale Difference Errors

This study considers the CUSUM test for ARMA models with stationary martingale difference errors. CUSUM tests are widely used for detecting abrupt changes in time series models. Although they perform adequately in general, their performance is occasionally unsatisfactory in ARMA models. This motivates us to design a new test that can simultaneously detect the ARMA parameter and variance changes. Its null limiting distribution is derived under regularity conditions. Monte Carlo simulations confirm the validity of the proposed test.

Haejune Oh, Sangyeol Lee

Agent-Based Modeling of Economic Instability

Networks increase interdependence, which creates challenges for managing economic risks. This is especially apparent in areas such as financial institutions and enterprise risk management, where the actions of a single agent (firm or bank) can impact all the other agents in interconnected networks. In this paper, we use agent-based modeling (ABM) in order to analyze how local defaults of supply chain participants propagate through the dynamic supply chain network and interbank networks and form avalanches of bankruptcy. We focus on the linkage dependence among agents at the micro-level and estimate the impact on the macro activities. Combining agent-based modeling with the network analysis can shed light on understanding the primary role of banks in lending to the wider real economy. Understanding the linkage dependency among firms and banks can help in the design of regulatory paradigms that rein in systemic risk while enhancing economic growth.

Akira Namatame

A Bad Plan Is Better Than No Plan: A Theoretical Justification of an Empirical Observation

In his 2014 book “Zero to One”, a software mogul Peter Thiel lists the lessons he learned from his business practice. Most of these lessons make intuitive sense, with one exception – his observation that “a bad plan is better than no plan” seems to be counterintuitive. In this paper, we provide a possible theoretical explanation for this somewhat counterintuitive empirical observation.

Songsak Sriboonchitta, Vladik Kreinovich

Shape Mixture Models Based on Multivariate Extended Skew Normal Distributions

In this paper, the class of the shape mixtures of extended skew normal distributions is introduced. The posterior distributions for the shaped parameters are obtained. The moment generating functions for the posterior distributions of the shaped parameters are discussed. Also Bayesian analysis for this shape mixture model is studied.

Weizhong Tian, Tonghui Wang, Fengrong Wei, Fang Dai

Plausibility Regions on Parameters of the Skew Normal Distribution Based on Inferential Models

In this paper, plausibility functions and $$100(1-\alpha )\%$$ plausibility regions on location parameter and scale parameter of skew normal distributions are obtained in several cases by using inferential models (IMs), which are new methods of statistical inference. Simulation studies and one real data example are given for illustration of our results.

Xiaonan Zhu, Baokun Li, Mixia Wu, Tonghui Wang

Measures of Mutually Complete Dependence for Discrete Random Vectors

In this paper, a marginal-free measure of mutually complete dependence for discrete random vectors through subcopulas is defined, which generalizes the corresponding results for discrete random variables. Properties of the measure are studied and an estimator of the measure is introduced. Several examples are given for illustration of our results.

Xiaonan Zhu, Tonghui Wang, S. T. Boris Choy, Kittawit Autchariyapanitkul

Applications

Frontmatter

To Compare the Key Successful Factors When Choosing a Medical Institutions Among Taiwan, China, and Thailand

What are the most important factors that increases customers’ (patients’) satisfaction with a general clinic, dental clinic and cosmetic surgery clinic respectively in Asia? Our paper tries to answer the question by conducting survey in Taiwan, Thailand, and China (in the order of the time the survey was conducted), with Grey Relational Analysis (GRA) methodology applied to identify key successful factors in three regions. By the research, our paper found specific and interesting phenomenon of medical institutions in each regions. ‘Doctor’s Skill’ is the general factor considered to be important across regions.

Tzong-Ru (Jiun-Shen) Lee, Yu-Ting Huang, Man-Yu Huang, Huan-Yu Chen

Forecasting Thailand’s Exports to ASEAN with Non-linear Models

This work focuses on forecasting Thailand’s exports to ASEAN. Thailand’s exports to ASEAN reveal an overall increasing trend with a fluctuation since Thailand’s exports are integrated in the global economy. However, the linear model might not be able to capture the behavior of Thailand’s exports to ASEAN. Linear model cannot be applied in some phenomena such as fluctuation and structural breaks in time series data. In this study, we find that the Thailand’s exports-to-ASEAN time series is non-linear via test of linearity, and find that there are two thresholds. Therefore, we forecast Thailand’s exports to ASEAN with non-linear models. We employ four non-linear models, SETAR, LSTAR, MSAR, and Kink AR model. The simple linear AR model is also applied to compare with the non-linear models. To evaluate the forecasting performance of five different models, we use RMSE and MAE as criteria. The forecasting results indicate that the SETAR model is better than the other models. However, it is still not clear cut to conclude that the non-linear models outperform linear model. However, we can conclude that the SETAR is the most suitable for forecasting Thailand’s exports to ASEAN compared with other non-linear models.

Petchaluck Boonyakunakorn, Pathairat Pastpipatkul, Songsak Sriboonchitta

Thailand in the Era of Digital Economy: How Does Digital Technology Promote Economic Growth?

As Thailand has undergone the reformation in both social and economic dimensions due to the digital economy, technologies are now becoming the new driving forces of economic growth. Therefore, an attempt of this study is to provide an empirical evidence on this issue, examining how increases in digital technologies impact the Thai economy. This study employs the stochastic frontier model estimated by entropy approach to model the production function. Because of a specific capability of this model, we are also able to find out how efficiently those technologies are utilized. The estimated results show that technologies can contribute positively to the Thai economy although the magnitudes are small. Moreover, our finding emphasizes that the digital technologies are not being used at the maximum capability, therefore, there is still a room for improvement in Thailand.

Noppasit Chakpitak, Paravee Maneejuk, Somsak Chanaim, Songsak Sriboonchitta

Comparing Linear and Nonlinear Models in Forecasting Telephone Subscriptions Using Likelihood Based Belief Functions

In this paper, we experiment with several different models with belief function to forecast Thai telephone subscribers. This approach will provide an uncertainty about predicted values and yield a predictive belief function that quantities the uncertainty about the future data. The proposed forecasting models include linear AR, Kink AR, Threshold AR, and Markov Switching AR models. Next, we compare the out-of-sample performance using RMSE and MAE. The results suggest that the out-of-sample belief function based KAR forecast is more accurate than other models. Finally, we find that the growth rate of Thai telephone subscription in 2016 will fall around 6.08%.

Noppasit Chakpitak, Woraphon Yamaka, Songsak Sriboonchitta

Volatility in Thailand Stock Market Using High-Frequency Data

The objective of this research is twofold: First, we aim to investigate the performance of conventional GARCH and GARCH-jump models when the data has high frequency. Second, the obtained conditional volatility from the best fit model is used to forecast and matched with the macroeconomic news announcement. We use GARCH and GARCH-jump models with high-frequency dataset of log return of Thailand stock market index (SET) from January, 2008 to December, 2015. We find that the volatility estimations by these two models have the same pattern but volatility estimation by GARCH-jump is higher than conventional GARCH model. However, the GARCH (1,1) and GARCH (1,1)-jump performances are non-stationary to estimate the volatility for 5 min interval return of SET but are stationary to estimate for 15 min, 30 min, 1 h, and 2 h returns of SET. Our results also show the matching jump point with macroeconomic news announcement. The empirical results support our assumption that macroeconomic news announcement may lead to volatility change in SET.

Saowaluk Duangin, Jirakom Sirisrisakulchai, Songsak Sriboonchitta

Technology Perception, Personality Traits and Online Purchase Intention of Taiwanese Consumers

The aim of the study is to examine the influences of personality characteristics and perception of technology on e-purchase intention. This study uses a questionnaire survey in collecting relevant data. The target sample is Taiwanese consumers. Multi Regression Analysis is used to test the model and hypotheses. For the measurement model, descriptive analyses and factor analysis are assessed to verify the validity and reliability of the data. As results, the impact of perceived ease of use is the strongest influence on online buying intention. In addition, perceived usefulness, perceived ease of use, and openness to experience have significant impacts on online purchase intention, thereby mediating the relationship between consciousness and online purchase intention. Providing guidelines for strategic plan, technological project, marketing program decision, and website design for online suppliers. This study also has significant implications for personalization, e-commerce, and marketing in online stores. Due to the limited knowledge of the impact of personality traits and perception of technology on customers online purchase intention, the current study appears to be a newly emerging topic in the field of marketing research.

Massoud Moslehpour, Ha Le Thi Thanh, Pham Van Kien

Analysis of Thailand’s Foreign Direct Investment in CLMV Countries Using SUR Model with Missing Data

Thai enterprises and companies have turned their attentions to CLMV countries, since the establishment of ASEAN Economic Community (AEC) in 2015, due to market and production opportunities. This study is conducted with an attempt to provide useful information helping the Thai investors make investment decision. In particular, this study examines the determinants of outward Thailand’s direct investment to the CLMV countries, and later estimates marginal effects. During the analysis, data unavailability in the CLMV and missing values in many available variables become destructive. This study handles this problem by using the bootstrap-based expectation maximization with bootstrapping (EMB). Once a complete data set is obtained, this study then employs the Seemingly Unrelated Regression (SUR) model to analyse the effects of the considered variables on Thailand’s direct investment in the CLMV group. The estimated results show distinct determinants for the countries, which can be useful to investors.

Chalerm Jaitang, Paravee Maneejuk, Aree Wiboonpongse, Songsak Sriboonchitta

The Role of Oil Price in the Forecasts of Agricultural Commodity Prices

The objective of this paper is to examine whether including oil price to the agricultural prices forecasting model can improve the forecasting performance. We employ linear Bayesian vector autoregressive (BVAR) and Markov switching Bayesian vector autoregressive (MS-BVAR) as innovation tools to generate the out-of-sample forecast for the agricultural prices as well as compare the performance of these two forecasting models. The results show that the model which includes the information of oil price and its shock outperforms other models. More importantly, linear model performs well in one- to three-step-ahead forecasting, while Markov switching model presents greater forecasting accuracy in the longer time horizon.

Rossarin Osathanunkul, Chatchai Khiewngamdee, Woraphon Yamaka, Songsak Sriboonchitta

Does Forecasting Benefit from Mixed-Frequency Data Sampling Model: The Evidence from Forecasting GDP Growth Using Financial Factor in Thailand

It is common for macroeconomic data to be observed at different frequencies. This gives a challenge to analysts when forecasting with multivariate model is concerned. The mixed-frequency data sampling (MIDAS) model has been developed to deal with such problem. However, there are several MIDAS model specifications and they can affect forecasting outcomes. Thus, we investigate the forecasting performance of MIDAS model under different specifications. Using financial variable to forecast quarterly GDP growth in Thailand, our results suggest that U-MIDAS model significantly outperforms the traditional time-aggregate model and MIDAS models with weighting schemes. Additionally, MIDAS model with Beta weighting scheme exhibits greater forecasting precision than the time-aggregate model. This implies that MIDAS model may not be able to surpass the traditional time-aggregate model if inappropriate weighting scheme is used.

Natthaphat Kingnetr, Tanaporn Tungtrakul, Songsak Sriboonchitta

A Portfolio Optimization Between US Dollar Index and Some Asian Currencies with a Copula-EGARCH Approach

There is a strong correlation between the value of the US dollar and the Asian currencies. EGARCH-copula model, with the skewed student-t distribution and the skewed general error distribution, can be used to capture the dependence correlation between US dollar and an Asian currency from those seven currencies in this paper. Building a bivariate portfolio based on the fitted EGARCH-copula models can be used to make portfolio optimization with the methods of max return, min risk and max sharpe ratio, to obtain a positive and reasonable return.

Ji Ma, Jianxu Liu, Songsak Sriboonchitta

Technical Efficiency Analysis of China’s Agricultural Industry: A Stochastic Frontier Model with Panel Data

This paper imposed the translog stochastic frontier production model to analyze the China’s province-level agriculture productivity by using panel data during 2002–2012 on 31 provinces in China. The results show that China’s province-level agriculture productivity has been improved for over 11 years. Hunan, Bejing and Shanghai approached the agriculture technical efficiency frontier. The agriculture technical efficiencies in underdeveloped area such like Guizhou, Yunnan and Anhui increased sharply and approached to the national province-level mean, 60%, in terms of the technical efficiencies over 11 years which, however, still have 40% space to be improved. We recommend that the provinces with lower technical efficiency, such as Anhui, Yunnan and Guizhou, should learn experiences from those provinces that have high technical efficiency so that improving the agricultural productivities of themselves.

Ji Ma, Jianxu Liu, Songsak Sriboonchitta

Empirical Models of Herding Behaviour for Asian Countries with Confucian Culture

The purpose of this study is to investigate the insights of herding behavior in the Confucian markets by conducting a set of empirical tests. More specifically this study investigates a sample of 7 countries and 13 markets to gain a deeper understanding of the causes of herding behavior and the potential factors that cause investors to behave in a group manner. Following a comprehensive review of the existing methodologies on herding behavior this study employs return dispersion approaches and herding tests developed by Chang et al. (2000) and Tan et al. (2008). This study investigates a sample of 13 stock markets of seven Asian economies with three different hypotheses. Those economies, which are considered to have Confucian culture, are mainland China, Hong Kong, Japan, South Korea, Taiwan, Singapore, and Vietnam. The hypotheses of this study aim to investigate formation of herding behavior in different market and economic circumstances. In testing the empirical models, this study uses OLS regression for the main test as well as regression with Newey and West (1987) for the robustness test of each result from OLS regression analysis. Data of this study consists of 13 index returns (Shanghai A and B share, Shenzhen A and B share, Hang Seng index, NIKKEI225, TOPIX, KOSDAQ, KOSPI, Straits Times Index, TAIEX, and indices of Hanoi and Hochiminh city Stock Exchanges) and returns of their constituent stocks. The time period of the sample data is from January 01, 1999 to December 31, 2014. All data were collected from the Thomson Reuters Datastream database. According to the empirical findings, all hypotheses are accepted. The sample markets demonstrate significant herding behavior in general and significant herding behavior in different markets conditions, such as in rising-falling markets and high-low market volatility states. This study has some major contributions to the literature of herding behavior and the link between herding behavior and cultural aspects. First, this study uses dataset of 13 Confucian stock markets of seven Asian economies, with time range from 1999 to 2014. Second, this research developed and tested three different hypotheses, and all of them are accepted. Third, this study adds the new dimension of the cultural aspects in order try to explain the root causes to herding behavior among investors in the equity markets. Recognizing that the Confucian culture appears to be one of the most influential cultural aspects in management, this study examines herding behavior of Confucian culture in stock markets under the umbrella of one empirical study. According to the findings of this study, Confucian culture has a positive and significant effect on herding behavior among investors in equity markets.

Munkh-Ulzii, Massoud Moslehpour, Pham Van Kien

Forecasting the Growth of Total Debt Service Ratio with ARIMA and State Space Model

Since the global financial crisis erupted in September 2008, many recent economists have been worried about the health of financial institutions. Consequently, many recent researches have put great emphasis on study of total debt service ratio (TDS) as one of the early warning indicators for financial crises. Accurate TDS forecasting can have a huge impact on effective financial management as a country can monitor the signal of financial crisis from a TDS’s future trend. Therefore, the purpose of this paper is to find the modeling to forecast the growth of TDS. Autoregressive integrated moving average (ARIMA) models tends to be the most popular forecasting method with indispensable requirement of data stationarity. Meanwhile, State Space model (SSM) allows us to examine directly from original data without any data transformation for stationarity. Furthermore, it can model both structural changes or sudden jumps. The empirical result shows that the SSM expresses lower prediction errors with respect to RMSE and MAE in comparison with ARIMA.

Kobpongkit Navapan, Petchaluck Boonyakunakorn, Songsak Sriboonchitta

Effect of Macroeconomic Factors on Capital Structure of the Firms in Vietnam: Panel Vector Auto-regression Approach (PVAR)

The article examines the impact of macroeconomicfactors on capital structure during the period of economic recession and economic recovery. The authors collected data from the financial statements of 82 firms listed in Vietnam stock market during the Quarter 1/2007-Quarter 2/2016 and using PVAR. The results demonstrate that during economic recession, the economic growth, the bond market, credit market positively impacted the capital structure whereas the stock market showed negative impacts on the capital structure. During economic recovery, economic growth positively impacted on the capital structure and the remaining macroeconomic variables negatively impacted on the capital structure. In addition, capital structure was affected bymicroeconomic variables such as profitability, asset structure, size, growth and liquidity.

Nguyen Ngoc Thach, Tran Thi Kim Oanh

Emissions, Trade Openness, Urbanisation, and Income in Thailand: An Empirical Analysis

This study investigates the relationship between emissions, income, energy consumption, trade openness, and urbanisation in Thailand over the period of 1971 to 2014. The ARDL cointegration technique is employed and CUSUM and CUSUMSQ tests are used to ensure the stability of the estimated results. Our findings indicate there is a long run relationship among variables for the case of CO2 emissions while there is none for the SO2. The results indicate an increase in income can cause significantly more CO2 emissions. Energy consumption also contributes to environmental degradation with slight impact, while there is no effect from trade openness. On the contrary, urbanisation greatly helps reduce CO2 emissions in the long run.

Rossarin Osathanunkul, Natthaphat Kingnetr, Songsak Sriboonchitta

Analysis of Risk, Rate of Return and Dependency of REITs in ASIA with Capital Asset Pricing Model

This study introduces an approach to fitting a copula based seemingly unrelated regression to an interval-valued data set. This approach consists of fitting a model on the appropriate point of the interval values assumed by the variables in the learning set. To find the appropriate point of the interval values, we assign weights in calculating the appropriate value between intervals by using convex combination method. We apply this methodology to quantify the risk and dependence of Real Estate Investment Trust (REITs) in Asia. Our results suggest that Hong Kong and Japan markets have a positive sign of the beta and both markets have less volatility than the global REITs market. On the other hand, we find that the estimated beta for Singapore market shows a negative relationship with global REITs market. We conclude that Singapore market can be viewed as a hedge against higher risk in Asian REITs.

Rungrapee Phadkantha, Woraphon Yamaka, Roengchai Tansuchat

Risk Valuation of Precious Metal Returns by Histogram Valued Time Series

The price of precious metals is highly volatile and it can bring both risk and fortune to traders and investors, and therefore should be examined. In this paper, we introduce an approach to fitting a Copula-GARCH to valued time series and apply this methodology to the daily histogram returns of precious metals consisting of gold, silver, and platinum. The study also conducts a simulation study to confirm the accuracy of the model and the result shows that our model performs well. In the empirical study, our results suggest investing on gold and platinum in high proportion while silver is not recommended for inclusion in the precious metal portfolio. Moreover, precious metal portfolio of the intraday 30-min returns gives lower risk when compared with portfolio of the intraday 60-min returns. Therefore, investors should not hold assets for long period of time because the long-term holding is likely to face a higher risk.

Pichayakone Rakpho, Woraphon Yamaka, Roengchai Tansuchat

Factors Affecting Consumer Visiting Spa Shop: A Case in Taiwan

The research aims to search for factors that influence the customer’s decision-making process regarding spa services of a case study spa in Taiwan based on a Count Data Model. The estimation is via Poisson regression analysis and negative binomial regression analysis. 167 questionnaires were collected from Taiwanese customers. The results of both Poisson regression and negative binomial regression are statistically significant at the conventional levels, which provides the predictions of the consumer’s decision-making. The study shows that the customer’s demography and customer satisfaction towards the case study spa have an impact on a consumer’s decision-making process when selecting spa services. Therefore, the spa can consider its marketing strategies based on the result.

Meng-Chun Susan Shen, I-Tien Chu, Wan-Tran Huang

The Understanding of Dependent Structure and Co-movement of World Stock Exchanges Under the Economic Cycle

This study was to focus on the patterns of economic booms (bull markets) and recessions (bear markets) among world stock exchanges such as Europe (Euro Stoxx), USA (S&P 500), Asia (SSE composite index and Nikkei 225 index) and ASEAN (FTSE ASEAN). Monthly data was collected during 2000 to 2016. Econometrically, we employed Markov Switching Bayesian Vector Autoregressive model (MSBVAR) to determine regional switches within these financial data sets as well as CD-Vine copula approaches was used to explore the contagions and patterns of structural dependences. To clarify the connectional details in each type of switching regimes, the results presented the Elliptical copula was chosen and it indicated these monthly collected data contained symmetrical dynamics co-movements. In addition, it implied the stock markets were assumed to have small fluctuations since the governments had stable policies to control the risk and asymmetric information in financial markets efficiently. Base on CD-Vine copula trees, the results indicated Asia and European stock markets had a strongly dependence in economic booms and recessions during the pre-crisis period (2000 to 2008). Conversely, in the post-crisis period, the US stock market and ASEAN stock market became the strong dependence with Europe. This meant that capital flows was mostly transferred between Europe and Asia financial markets during the pre-crisis periods (2009 to 2016). After that, the direction of capital flows were changed dramatically to the US stock market in the post-crisis periods. Predictively, this seems that the capital flows will return to European and US financial market, which these two continents have a strongly long-term financial dependence and deeply positive diplomacy.

Songsak Sriboonchitta, Chukiat Chaiboonsri, Jittima Singvejsakul

The Impacts of Macroeconomic Variables on Financials Sector and Property and Construction Sector Index Returns in Stock Exchange of Thailand Under Interdependence Scheme

This paper investigates the impacts of macroeconomic variables, namely consumer price index, exchange rate, minimum loan rate and oil price movement, on the financials and the property & construction stock index return in the Stock Exchange of Thailand (SET). The monthly data is collected from January 2004 to November 2016, covering 155 observations. We employ a copula based SUR regression as a tool for this study. Ten copula functions are considered in this regression and the best copula function is selected based on Akaike’s Information Criterion (AIC) and Bayesian Information Criterion (BIC). The estimated results show that Gumbel 270 copula is the most appropriate function for being the linkage between the marginal distributions of residuals of financials sector and property & construction sector equations. In addition, the marginal distribution is also tested, and the result shows that normal distribution is the best fit for the marginal distribution for both financials and property & construction equations. Our results suggest that the exchange rate can exert significant impact on both sectors. The dependency parameter also suggests that dependency between financials sector and property & construction sector is negative, and very low dependency, meaning when the impact of macroeconomic variables in one of these two sectors, it just has a little effect to another one sector.

Wilawan Srichaikul, Woraphon Yamaka, Roengchai Tansuchat

Generalize Weighted in Interval Data for Fitting a Vector Autoregressive Model

This paper employ VAR model to analyse and investigate the relationship among oil, gold, and rubber prices. A convex combination approach is proposed to obtain appropriate value of the interval data in VAR model. The construction of interval VAR model based on the convex combination method for the analysis of their forecast performance are also introduced and discussed via the simulation study, as well as comparing the performance with conventional center method. To illustrate the usefulness of the proposed model, an empirical application on a weekly sample of commodity price is provided. The results show the performance of our proposed model and also provide some relationship between commodity prices.

Teerawut Teetranont, Woraphon Yamaka, Songsak Sriboonchitta

Asymmetric Effect with Quantile Regression for Interval-Valued Variables

In this paper, we propose a quantile regression with interval valued data using a convex combination method. The model we propose generalizes series of existing models, say typically with the center method. Three estimation techniques consisting EM algorithm, Least squares, Lasso penalty are presented to estimate the unknown parameters of our model. A series of Monte Carlo experiments are conducted to assess the performance of our proposed model. The results support our theoretical properties. Finally, we apply our model to empirical data in order to show the usefulness of the proposed model. The results imply that the EM algorithm provides a best fit estimation for our data set and captures the effect of oil differently across various quantile levels.

Teerawut Teetranont, Woraphon Yamaka, Songsak Sriboonchitta

The Future of Global Rice Consumption: Evidence from Dynamic Panel Data Approach

This study investigates the future outlook of global rice consumption using dynamic panel data regression (DPD) with penalised fixed effect model. The three main factors affecting rice consumption include previous rice demand, GDP per capita, and world rice price. The data set covers 73 countries that is almost 80% of world rice consumption from 1960 to 2015. We separate these countries into 4 groups based on income levels classified by the World Bank including low income, lower middle-income, upper middle-income, and high income. The results show that, at the global scale, rice consumption is expected to be slightly higher. Such demand is driven by rising demand from the upper middle- and high income countries, while it is offset by the lower demand from lower middle- and low income countries.

Duangthip Sirikanchanarak, Tanaporn Tungtrakul, Songsak Sriboonchitta

The Analysis of the Effect of Monetary Policy on Consumption and Investment in Thailand

This study highlights on the analysis of Thai monetary policy transmission channels, i.e. interest rate, credit, exchange rate, and asset price channels, to private consumption and private investment. The analytical methods are Time Varying Parameter Vector Autoregressive (TVP-VAR) with stochastic volatility, and its impulse response function. The results showed that the credit channel contribute the greatest impact on private consumption and investment. We also found that the effect of monetary policy to private consumption and investment are vary over time.

Jirawan Suwannajak, Woraphon Yamaka, Songsak Sriboonchitta, Roengchai Tansuchat

Investigating Relationship Between Gold Price and Crude Oil Price Using Interval Data with Copula Based GARCH

This study investigates and compares the performance of center method, equal weighted convex combination and unequal-weighted convex combination methods through various GARCH and copula-based approaches for the analysis of relationship between gold and crude oil prices using interval data in Comex and Nymex tradings. The results of this study confirm that unequal-weighted convex combination method improves the estimation and it tends to perform better than both the center method and its equal-weighted variant. In addition, the marginal from the best fit GARCH model is used to measure dependence via copula function in the form of Student-t copula as selected according to the lowest AIC among all candidates. Finally, we can conclude that there exists the dependence between Comex and Nymex not only in the normal event, but also in the extreme event.

Teerawut Teetranont, Somsak Chanaim, Woraphon Yamaka, Songsak Sriboonchitta

Simultaneous Confidence Intervals for All Differences of Means of Normal Distributions with Unknown Coefficients of Variation

This paper presents a procedure for simultaneous confidence interval estimation for the differences of means of several normal populations with unknown coefficients of variation. The proposed approaches are a generalized confidence interval approach (GCI approach) and method of variance estimates recovery approach (MOVER approach). A Monte Carlo simulation was used to evaluate the performance in terms of coverage probability, average width and standard error. The simulation results indicated that the GCI and MOVER approaches are satisfactory in terms of the coverage probability, but the average widths of the MOVER approach are slightly shorter than the average widths of the GCI approach. The proposed approaches are illustrated by an example.

Warisa Thangjai, Sa-Aat Niwitpong, Suparat Niwitpong

Estimating the Value of Cultural Heritage Creativity from the Viewpoint of Tourists

Creativity in cultural heritage raises expectations of added value, and helps promote local economic development through new elements introduced to the original cultural industry. The purpose of this study is to discuss the willingness to pay (WTP) for commercialization and creativity of cultural heritage. However, no general market price for cultural heritage exists. Thus, the Contingent Valuation Method (CVM) in the non-market valuation method was applied to determine whether commercialization and creativity of cultural heritage through a personally inherent attitude and through different preferences could analyze the tourists’ willingness to pay, and to consider the factors that influence their WTP. In this study, 410 subjects were used for the CVM construction of the WTP based on three situations when cultural heritage and creativity are combined, as well as in the application of the double-bounded dichotomous choice model of survival analysis to estimate the WTP-influencing factors. The results showed that the higher the income of the subject, the higher of WTP for value-added services for preserving cultural heritage, participating in activities, and helping local development.

Phanee Thipwong, Chung-Te Ting, Yu-Sheng Huang, Yun-Zu Chen, Wan-Tran Huang

A Bibliometric Review of Global Econometrics Research: Characteristics and Trends

Using a bibliometric analysis, this research analyses on the Social Science Citation Index (SSCI) publications from the Institute for Scientific Information (ISI), Web of Science database, based on 12965 publications from 1497 journals during 1992 to 2016. The research was assessed the research’s characteristics and trends of most productive countries/regions and institutions, was pointed out the sharp increasing in China on econometrics research, Applied Economics (England) published the most econometrics articles. The research was also pointed out the temporal evolution of recent hot econometrics research issues. Global trends and characteristics was found throughout this research can give a general overview for further researches on econometrics.

Van-Chien Pham, Man-Ling Chang

Macro-Econometric Forecasting for During Periods of Economic Cycle Using Bayesian Extreme Value Optimization Algorithm

This paper aims to computationally analyze the extreme events which can be described as crises or unusual times-series trends among the macroeconomic variables. These data are statistically estimated by employing the optimally extreme point for supporting policy makers to specify the economic expansion target and economic warning level. The Nonstationary Extreme Value Analysis (NEVA) applying Bayesian inference and Newton-optimal method are employed to complete the researchs solutions and estimate the time-series variables such as GDP, CPI, FDI, and unemployment rate collected during 1980 to 2015. The results show there are extreme values in the trend of macroeconomic factors in Thailand economic system. This extreme estimation is presented as an interval. In addition, the empirical results from the optimization approach state that the exactly extreme points can be computationally found. Ultimately, it is clear that the computationally statistical approach, especially Bayesian statistics, is inevitably important for econometric researches in the recent era.

Satawat Wannapan, Chukiat Chaiboonsri, Songsak Sriboonchitta

Forecasting of VaR in Extreme Event Under Economic Cycle Phenomena for the ASEAN-4 Stock Exchange

This paper was proposed to computationally investigate the cycling details and risk management of the ASEAN-4 financial stock indexes, including Bangkok Bank (BBL), Development Bank of Singapore Limited (DBS), Commerce International Merchant Bankers (CIMB), and Bank Mandiri (Mandiri). These daily time-series data were observed during 2012 to 2017. Technically, this paper employed the econometric tool called Markov Switching Model (MS-model), the extreme value application called Generalized Pareto Distribution (GPD-model), and the risk management method called Value at Risk (VaR) to provide the estimated solutions and recommendations for investing in these financial stocks. Empirically, the switching regime estimation resulted that these four financial indexes obviously contain real business cycling movements, which were described as bull and bear regimes. Additionally, the results estimated by the GPD model confirmed that there were extreme events inside the trends of the four stock indexes. Ultimately, the outcomes calculated by the risk measurement for extreme cases, which were economic crises, stated that there was an enormously high risk to considerably invest only in short earnings within these four financial stock indexes. Consequently, long-run investment should be mentioned.

Satawat Wannapan, Pattaravadee Rakpuang, Chukiat Chaiboonsri

Interval Forecasting on Big Data Context

Purpose: The object of this research is to construct an optimal internal forecasting method in big data context.Design/methodology/approach: An intelligent model construction, including consumer behavior and market information, structural changes detection, nonlinear pattern recognition, spatial causality, semantic processing mode is presented.Findings: The major drawback in forecasting field is that the statistical forecasting result is derived from historical data but it often encounters non-realistic problem when people predict future trends or market changes in real world.Practical Applications: Construction of Big Data platform will be a new technique provides to solve the structured change and uncertain problems. According to the artificial intelligence evolution and on line improvement to the market conditions, it will do a better performance to prevailing future event. Originality: We efficiently integrate the idea of structure change, entropy and market behavior in the forecasting process.Conclusion: Since historical time series analysis has difficult to prove the relationship/causality with future events. Especially in the case of a structural change, the future is full of high uncertainty, ambiguity and unexpected.

Berlin Wu

Bayesian Empirical Likelihood Estimation for Kink Regression with Unknown Threshold

Bayesian inference provides a flexible way of combining data with prior information from our knowledge. However, Bayesian estimation is very sensitive to the likelihood. We need to evaluate the likelihood density, which is difficult to evaluate, in order to use MCMC. Thus, this study considers using the Bayesian empirical likelihood(BEL) approach to kink regression. By taking the empirical likelihood into a Bayesian framework, the simulation results show an acceptable bias and MSE values when compared with LS, MLE, and Bayesian when the errors are generated from both normal and non-normal distributions. In addition, BEL can outperform the competing methods with quite small sample sizes under various error distributions. Then, we apply our approach to address a question: Has the accumulation of foreign reserves effectively protected the Thai economy from the financial crisis? The results demonstrate that foreign reserves provide both positive and negative effects on economic growth for high and low growth regimes of foreign reserve, respectively. We also find that foreign reserves seem to have played a role in offsetting the effect of the crisis when the growth rate of foreign reserves is less than 2.48%.

Woraphon Yamaka, Pathairat Pastpipatkul, Songsak Sriboonchitta

Spatial Choice Modeling Using the Support Vector Machine (SVM): Characterization and Prediction

We take a cursory look at the support vector machine (SVM) as a useful and effective algorithm for characterizing and predicting spatial choice problems in economics. Beginning with a discussion of the SVM for the linearly separable case as well as the nonlinear non-separable case using the soft margin SVM and kernels, we then describe how the SVM can be used to characterize and predict spatial choice models, which can be seen as a special case of discrete choice models, using examples from a simple 1-D to the more complex multi-dimensional features space.

Yong Yoon

Backmatter

Weitere Informationen

Premium Partner

    Bildnachweise