International Journal of Computational Economics and Econometrics (29 papers in press)
Multi-period Mean-variance Portfolio Selection with Practical Constraints Using Heuristic Genetic Algorithms
by Yao-Tsung Chen, Hao-Qun Yang
Abstract: Since Markowitz proposed the meanvariance (MV) formulation in 1952, it has been used to configure various portfolio selection problems. However Markowitzs solution is only for a single period. Multi-period portfolio selection problems have been studied for a long time but most solutions depend on various forms of utility function, which are unfamiliar to general investors. Some works have formulated the problems as MV models and solved them analytically in closed form subject to certain assumptions. Unlike analytical solutions, genetic algorithms (GA) are more flexible because they can solve problems without restrictive assumptions.
The purpose of this paper is to formulate multi-period portfolio selection problems as MV models and solve them by GA. To illustrate the generality of our algorithm, we implement a program by Microsoft Visual Studio to solve a multi-period portfolio selection problem for which there exists no general analytical solution.
Keywords: Multi-period portfolio selection; Mean-variance formulation; Genetic algorithm; Transaction costs.
Performance Evaluation of the Bayesian and classical Value at Risk models with circuit breakers set up
by Gholam Reza K. Haddad, Hadi Heidari
Abstract: Circuit breakers, like price limits and trading suspensions, are used to reduce price volatility in security markets. When returns hit price limits or missed, observed returns deviate from equilibrium returns. This creates a challenge for predicting stock returns and modeling Value at Risk (VaR). In Tehran Stock Exchange (TSE), the circuit breakers are applied to control for the excess price volatilities. We extend Weis (2002) model, in the framework of Bayesian Censored and Missing-GARCH approach, to estimate VaR for Iran Khodro Company (IKCO) share in TSE. Using daily data for the period of June 2006 to June 2016, we show that the Censored and missing- GARCH model with t-student distribution outperforms. Kullback-Leibler (KLIC), Kupic (1995) test and Lopez score (1998) outcomes show that estimated VaR by Censored and missing- GARCH model with t-student distribution is of the most accuracy among all other classical and Bayesian estimation models.
Keywords: Circuit Breakers; Censored and Missing–GARCH; Bayesian estimation; Value at Risk; Ranking Models.
Stages and determinants of European Union Small and Medium Sized firms failure process
by Alexios Makropoulos, Charlie Weir, Xin Zhang
Abstract: This paper uses a combination of Factor and Cluster analysis to identify and compare failure processes in Small and Medium sized firms from a number of European Union countries. Panel data analysis is then used to identify the determinants of the firms transition from financial health towards liquidation in the alternative failure processes. The results suggest that there are 4 different firm failure processes. We find that financial performance and director characteristics differ between firm failure processes. We also find that the economic environment, the legal tradition of countries and excessive firm growth are determinants of the transition of firms towards liquidation across most firm failure processes. These findings may be of practical use to policy makers, lenders and risk managers who will benefit from a better understanding of the differences between the alternative firm failure processes and from the determinants of a firms transition towards liquidation within these failure processes.
Keywords: SME failure; firm failure; firm failure process; factor/cluster analysis; ordered random effects regression; failure status transition.
Abnormal returns and systemic risk: evidence from a non-parametric bootstrap framework during the European sovereign debt crisis.
by Konstantinos Gkillas, Christos Floros, Christoforos Konstantatos, Dimitrios I. Vortelinos
Abstract: We investigate the impact of European Central Bank (ECB) interventions on major European and Turkish stock and credit default swap (CDS) markets highlighting the importance of abnormal to excess abnormal returns in the systemic risk. In particular, we examine the impact of ECB announcements (news) on major European and Turkish financial markets (stocks and CDSs indices) for a high and low-volatility period, i.e. from November 6th, 2008 to December 31st, 2015. We also examine the market efficiency by using both an event study methodology and the Capital Asset Pricing Model. Moreover, the impact of the ECB events is measured by an event study and a systemic risk analysis. The results show that investors exposed to Finland, Sweden, Austria and Spain tend to be more vulnerable to risk and volatility, when ECB announcements are published.
Keywords: abnormal returns; bootstrap; ECB events; financial crises;.
An analysis of major Moroccan domestic sectors interdependencies and volatility spillovers using Multivariate GARCH models.
by Ouael EL JEBARI, Abdelati HAKMAOUI
Abstract: This paper tries to give a thorough analysis of the mechanisms of volatility spillovers, as well as, a study of the time-varying interdependencies of volatilities of seven major sectors of the Moroccan stock exchange by proposing an empirical approach based on multivariate GARCH models. It uses daily data spanning the period between 02/07/2007 and 15/12/2016, covering seven principal sectors indices. The results of the study confirm the existence of multiple volatility transmissions in both ways and of both signs between sectors of our sample, along with, the quasi-abundance of positive correlations suggesting possible contagion effects. More importantly, our findings are in line with those discovered in the U.S financial market. The notoriety of this article resides in the fact that it broadens previously documented studies focusing mainly on external shocks by providing a study of internal shocks while applying two multivariate GARCH models.
Keywords: Volatility spillover; dynamic conditional correlations; interdependencies; domestic sectors; multivariate GARCH models.
A note on the use of the Box-Cox Transformation for Financial Data
by Dimitrios Kartsonakis Mademlis, Nikolaos Dritsakis
Abstract: This paper tests whether the Box-Cox transformation reduces the problem of non-normality in financial data.
Keywords: ARIMA models; Box-Cox transformation; Box-Jenkins methodology; normality; stock market; oil prices.
INFRASTRUCTURE DEVELOPMENT AND INCOME INEQUALITY IN INDIA: AN EMPIRICAL INVESTIGATION
by Varun Chotia
Abstract: The purpose of this paper is to investigate the relationship between infrastructure development and income inequality in India from 1991 to 2012, by using the auto regressive distributed lag (ARDL) bound testing approach. The co-integration test confirms the presence of a long run relationship between infrastructure development and income inequality. The ARDL test results indicate that infrastructure development does not help in reducing income inequality. Both inflation and economic growth amplify the income inequality both in the long run as well as the short run whereas trade openness comes out to be the indicator which is able to decrease the gap between rich and poor in India. The study calls for adopting economic policies and reforms which are aimed at developing and strengthening the infrastructure levels, bringing in more investment in order to achieve the much required inclusive growth, and ultimately reduce the income inequality currently prevailing in India.
Keywords: infrastructure development; income inequality; co-integration; auto regressive distributed lag; ARDL; India.
The Role of R&D in Economic Growth in Arab Countries
by Mohammed Shahateet
Abstract: This paper explores the impact of research and development (R&D) activities on economic growth in 12 Arab countries of the Middle East. We have conducted different pre-estimation tests in order to select the appropriate econometric model, including cross-sectional dependence, stationarity, causality, and cointegration. We perform post-estimation diagnostics to test for both long-run and short-run relationship by applying a panel Auto Regressive Distributed Lag (ARDL) model using data for the period 1996-2016. Long-run analysis confirms that R&D activities positively affect economic growth while in the short-run this relationship is insignificant.
Keywords: Economic growth;R&D; ARDL model; Panel data; Arab countries.
Economic and Business Cycle with Time Varying in India: Evidence from ICT Sector
by Chukiat Chaiboonsri
Abstract: Combining the theoretical concept of the Real Business Cycle (RBC) and computational econometric modeling for ICTs systems, the purpose of this paper is divided into two main sections. The first part is to study the relationship between Indian ICT industries and GDP by applying Bayesian inference, now referring to the modern statistics in this era. The section of data classifications, five-yearly predominant indexes collected during 2000 to 2015, including Indian GDP, fixed phone usages, mobile phone distributions, Internet servers, and broadband suppliers are analyzed by employing the Markov-Switching model (MS-model) and Bayesian Vector Autoregressive models (BVAR). The second section is the application of time-varying parameter VAR model with stochastic volatility (TVP-VAR). Based on Bayes statistics, this time-varying analysis can more clearly provide the extended perception regarding nature underlying structure in the economy in a flexible and robustness. In addition, the Bayesian regression model is used to investigate the ICT multiplier related to Indian economic growth. The empirical results indicate that IT sectors are becoming the major role of Indian economic expansion in the forthcoming future, compared with telecommunication sectors. Moreover, the result of the ICT multiplier confirms that high technological industrial zones should be systematically enhanced continuously, in particular, research and development in cyberspace. This additionally confirms that the high-tech industries are playing an important role to raise the levels of employment in India and lead it up to erase poverty in social-economic levels.
Keywords: Information and Communication Technology (ICT);
Markov-Switching Model (MS-model);
Bayesian Vector Autoregressive model (BVAR);
Time-Varying Parameter VAR (TVP_VAR).
Determinants of risk sharing via exports: Trade openness and Specialization
by Faruk Balli, Eleonora Pierucci, Jian Gan
Abstract: Economic theory predicts that one of the main benefits of financial globalisation is the improvement of international risk sharing. In this paper, we provide an empirical evaluation of the determinants of risk sharing via exports. We conclude that risk sharing via exports is somehow important in emerging countries but not among OECD countries. More importantly, we find that trade openness and production/export specialisation generally have, with some exceptions, positive and statistically significant relationship with risk sharing. On the contrary, concentration on export destinations has been proved to be negatively correlated with risk sharing.
Keywords: risk sharing; production specialisation; export specialisation; trade openness; financial globalisation.
Value-added in high technology and industrial basic research: a weighted network observing the trade of high-tech goods
by Antonio Zinilli, Mario De Marchi
Abstract: Expenditure on research and development (R&D) is a key indicator of the innovative efforts of countries. In this paper, we want to examine through a new approach the relationship between basic research and economic benefits. Although we are aware that the topic has already been extensively addressed, this paper differs from the previous literature because it focuses on value added trade instead of gross trade flows. We use an Exponential Random Graph Model for weighted networks to study the impact of private investment in basic research on value-added of exports in high technology, namely the domestic value added absorbed abroad. We want to understand if this measure (without intermediate imports) is able to confirm the results of the previous literature, which used gross flows. Our results show that private investment in basic research has a positive influence on exports, giving a competitive advantage in international trade.
Keywords: Exponential random graph model; Weighted Networks; Industrial Basic Research; Trade in Value Added.
Persistent dynamics in (in)determinate equilibrium rational expectations models
by Marco Maria Sorge
Abstract: Equilibrium indeterminacy in rational expectations models is often claimed to produce higher time series persistence relative to determinacy. Proceeding by means of a simple linear stochastic model, I formally show that, for reasonable parameter configurations, there exists an uncountable (continuously infinite) set of indeterminate equilibria in low-order AR(MA) representation, which exhibit strictly lower persistence than their determinate counterpart. Implications for empirical studies concerned with e.g. testing for indeterminacy and macroeconomic forecasting are discussed.
Keywords: Rational expectations; Indeterminacy; Persistence.
Size distribution analysis in the study of urban systems: evidence from Greece
by Dimitrios Tsiotas, Labros SDROLIAS, Georgios ASPRIDIS, Dagmar SKODOVA-PARMOVA, Zuzana DVORAKOVA-LISKOVA
Abstract: This paper examines empirically the utility of size-distribution analysis in the study of urban systems, on data referring to every urban settlement recorded in 2011 national census of Greece. The study lowers the scale of the size-distribution analysis to the regional, instead of the national level, where is commonly being applied, examining two aspects of size-distributions, the rank-size and the city-size-distribution, in comparison with three well-established statistical dispersion indices, the coefficient of variation, the Theil index and the Gini coefficient. The major research question is to detect how capable are the size-distribution exponents to operate as measures of statistical dispersion and to include socioeconomic information. Overall, the analysis concludes that the size-distribution assessment is useful for the initialization of the study of urban systems, where the available information is restricted to population size, and is capable to provide structural information of an urban system and its socioeconomic framework, but not more effective than other measures of statistical dispersion.
Keywords: power law; Zipf’s law; Regional Economics; cities; Regional Science; Econophysics.
Factor decomposition of disaggregate inflation: the case of Greece
by Nikolaos Krimpas, Paraskevi Salamaliki, Ioannis Venetis
Abstract: We use static and dynamic factor models to decompose Greek inflation into common components. Static factor analysis suggests the need to develop comprehensive underlying inflation measures for Greece. Dynamic factor analysis decomposes inflation into three components: pure inflation and relative price inflation both driven by aggregate shocks and, an idiosyncratic component reflecting sector specific shocks. We verify the idiosyncratic component as the main source of inflation variability while pure inflation and its associated shocks are dominant compared to relative inflation. Based on pure inflation correlations, the relative weight of anticipated monetary shocks is large only for the spread between the 10-year government bond yield and a three-month short run rate and only in times of monetary stability.
Keywords: Disaggregate CPI; Dynamic Factor Model; Pure inflation; Relative prices.
Computational method for approximating the behaviour of a triopoly: an application to the mobile telecommunications sector in Greece
by Yiannis Bassiakos, Zacharoula Kalogiratou, Theodoros Monovasilis, Nicholas Tsounis
Abstract: Computational biology models of the Volterra-Lotka family, known as competing species models, are used for modelling a triopoly market, with application to the mobile telecommunications in Greece. Using a data sample for 1999-2016, parameter estimation with non-linear least squares is performed. The findings show that the proportional change in the market share of the two largest companies, Cosmote and Vodafone, depends negatively on the market share of each other. Further, the market share of the marker leader, Cosmote, depends positively on the market share of the smallest company, Wind. The proportional change in the market share of Wind, depends negatively on the market share of the largest company Cosmote but it depends positively by the change in the market share by the second company, Vodafone. In the long-run it was found that the market shares tend to the stable equilibrium point where all three companies will survive with Cosmote having a projected number after eleven years (in 2030) of approximately 7.3 million subscribers, Vodafone 4,9 and Wind 3.7, the total number of projected market size being approximately 16 million customers.
Keywords: Volterra-Lotka models; triopoly; mobile telecommunications sector.
Separating Yolk from White: A Filter based on Economic Properties of Trend and Cycle
by Peng Zhou
Abstract: This paper proposes a new filter technique to separate trend and cycle based on stylised economic properties of trend and cycle, rather than relying on ad hoc statistical properties such as frequency. Given the theoretical separation between economic growth and business cycle literature, it is necessary to make the measures of trend and cycle match what the respective theories intend to explain. The proposed filter is applied to the long macroeconomic data collected by the Bank of England (1700-2015).
Keywords: Filter; Trend; Cycle.
Efficiency of microfinance institutions of South Asia: A bootstrap DEA approach
by Asif Khan, Rachita Gulati
Abstract: The MFIs are special types of institutions operate with the dual goals; financial sustainability and social outreach. Therefore, the present paper aims to assess the efficiency levels on the attainment of dual mission of financial sustainability and social outreach of MFIs operating in the selected four South Asian countries (i.e., Bangladesh, India, Nepal and Pakistan) during the financial year 2010 to 2015. The conventional data envelopment analysis (DEA) models do not follow the statistical properties, consequently, may produce the biased efficiency estimates. Therefore, we incorporated the homogeneous bootstrap procedure in DEA framework suggested by Simar and Wilson (1998, 2000) to estimate bias-corrected efficiency scores of individual MFIs. In addition, we designed two separate DEA models to assess the MFIs performance from both the prospects; financial and social, simultaneously. We first detected and removed the outliers from the dataset by using the procedure suggested by Banker and Gifford (1988) based on the super-efficiency concept and then proceed further. The empirical results confirmed that on an average, the South Asian MFIs remained more financially efficient than socially during the study period. However, financial efficiency has decreased over time while the social efficiency has improved slightly. Further, among the peer nations, Indian MFIs outperform in terms of both financial and social aspects followed by Nepali and Bangladeshi MFIs, respectively. However, the Pakistani MFIs were found to be the least performers in both social outreach and financial sustainability.
Keywords: Bias-corrected efficiency; financial efficiency; social efficiency; bootstrap data envelopment analysis; DEA; bootstrap DEA; microfinance institutions; MFIs; microfinance; South Asia.
Bootstrapping the Log-periodogram Estimator of the Long-Memory Parameter: is it Worth Weighting?
by Saeed Heravi, Kerry Patterson
Abstract: Estimation of the long-memory parameter from the log-periodogram (LP) regression, due to Geweke and Porter-Hudak (GPH), is a simple and frequently used method of semi-parametric estimation. However, the simple LP estimator suffers from a finite sample bias that increases with the dependency in the short-run component of the spectral density. In a modification of the GPH estimator, Andrews and Guggenberger, AG (2003) suggested a bias-reduced estimator, but this comes at the cost of inflating the variance. To avoid variance inflation, Guggenberger and Sun (2004, 2006) suggested a weighted LP (WLP) estimator using bands of frequencies, which potentially improves upon the simple LP estimator. In all cases a key parameter in these methods is the need to choose a frequency bandwidth, m, which confines the chosen frequencies to be in the neighbourhood of zero. GPH suggested a square-root rule of thumb that has been widely used, but has no optimality characteristics. An alternative, due to Hurvich and Deo (1999), is to derive the root mean square error (rmse) optimising value of m, which depends upon an unknown parameter, although that can be consistently estimated to make the method feasible. More recently, Arteche and Orbe (2009a,b), in the context of the GPH estimator, suggested a promising bootstrap method, based on the frequency domain, to obtain the rmse value of m that avoids estimating the unknown parameter. We extend this bootstrap method to the AG and WLP estimators and to consideration of bootstrapping in the frequency domain (FD) and the time domain (TD) and, in each case, to blind and local versions. We undertake a comparative simulation analysis of these methods for relative performance on the dimensions of bias, rmse, confidence interval width and fidelity.
Keywords: Long memory; bootstrap; log-periodogram regression; variance inflation; weighted LP regression; time domain; frequency domain.
R&D cooperation performance inside innovation clusters
by B. G. Jean Jacques Iritié
Abstract: This paper theoretically analyzes the effects of innovation cluster on R&D cooperation outcome in terms of private R&D investments. We develop a two-stage strategic R&D game with a duopoly in an innovation cluster. The two firms first conduct cooperatively their R&D decisions in a research joint venture (RJV) before competing in Cournot fashion on the market product. The results shows that an innovation cluster improves private R&D investments and social welfare through informational incentives. Then, belonging to an innovation cluster strengthens the willingness to cooperate in R&D and partnership. However, the model also show that innovation clusters can lead to a risk of monopolization of the market that could be extended even beyond cooperation; this risk thus makes it possible to relativize the expected positive effects of innovation cluster policy on R&D cooperation.
Keywords: Innovation cluster; research joint venture ; localised knowledge spillovers ; informational incentives; R&D cooperation performance.
Two algorithms in sign restrictions: An exploration in an empirical SVAR
by Lance Fisher, Hyeon-seung Huh
Abstract: This paper investigates whether the key choices for the implementation of algorithms in sign restrictions are consequential for the range of accepted impulse responses in an empirical SVAR. Two algorithms in sign restrictions are considered. In one algorithm, the key choices concern the method by which an initial set of orthogonal shocks are formed and the method by which they are rotated. For this algorithm, the range of accepted responses are invariant to these choices. In the other algorithm, the key choice is the selection of the set of coefficients on the contemporaneous variables in the structural equations which are to be generated. Each selection corresponds to an ordering of the variables. For this algorithm, the range of accepted responses can be affected by the ordering of the variables in which case the algorithm is extended to loop over all variable orderings.
Keywords: structural vector autoregressions; sign restrictions; Givens rotations; QR decomposition; instrumental variables; impulse responses.
The impact of the CAP Health Check on the price relations of the EU food supply chain: A dynamic panel data cointegration analysis
by Anthony Rezitis, Andreas Rokopanos
Abstract: The CAP Health Check in 2008 attested to the liberalization change initiated with the 2003 CAP Reform. The changes agreed are expected to affect the causal relationships among the supply chain actors. This study investigates the price causality along the EU food supply chain, employing a panel cointegration and error correction vector autoregressive (EC-VAR) approach. We utilize monthly data from January 2005 to September 2012, from nineteen European countries. The sample is split in December 2008 to examine how the Health Check affected causality among agricultural commodity prices (ACPs), producer prices (PPs) and consumer prices (CPs). The results indicate the existence of a long-run equilibrium. Furthermore, before the Health Check, ACPs are exogenous while both PPs and CPs are endogenous. After it, all the prices become endogenous. Short-run bidirectional causality is shown in both periods. The results suggest that decreased support has rendered ACPs more respondent to market signals.
Keywords: panel cointegration; error correction vector autoregressive (EC-VAR) model; causality analysis.
ENSEMBLE MARGIN BASED RESAMPLING APPROACH FOR A COST SENSITIVE CREDIT SCORING PROBLEM
by Meryem Saidi, Nesma Settouti, Mostafa El Habib Daho, Mohammed El Amine Bechar
Abstract: In the past few years, a growing demand for credit compel banking
institution to contemplate machine learning techniques as an answer to obtain
decisions in a reduced time. Different decision support systems were used to
detect loans defaulters from good loans. Despite good results obtained by these
systems, they still face some problems such as imbalanced class and imbalanced
misclassification cost problems. In this work, we propose a cost sensitive credit
scoring system, based on a two-step process. The first is a resampling step which
handles the imbalance data problem followed by a cost sensitive classification
step that recognizes potential insolvent loans red in order to reduce financial
loss. A resampling algorithm called Ensemble Margin for Imbalanced Instance
(EM2I) is suggested to manage imbalanced datasets in cost sensitive learning.
We compare our algorithm with other techniques from the state of the art and
experimental results on German credit dataset demonstrate that EM2I leads to a
significant reduction of the misclassification cost."
Keywords: Cost sensitive learning; imbalanced problem; ensemble margin; credit scoring.
A nonparametric estimator for stochastic volatility density
by Soufiane Ouamaliche, Awatef Sayah
Abstract: This paper aims at improving the accuracy of stochastic volatility density estimation in a high frequency setting using a simple procedure involving a combination of kernel smoothing methods namely, kernel regression and kernel density estimation. The employed data, which are thirty years worth of hourly observations, are simulated through a Constant Elasticity of Variance-Stochastic Volatility (CEV-SV) model, namely the Heston Model, calibrated to fit the S&P500 index, in the form of a two-dimensional diffusion process (Yt,Vt) such that only (Yt) is an observable coordinate. Polynomials of different degrees are then adjusted using weighted least squares to filter the observations of the variance coordinate (Vt) from a convolution structure before applying a straightforward kernel density estimation. The obtained estimates did well when compared to previous results as they have displayed a certain improvement, linked to the degree of the fitted polynomial, by reducing the value of the Mean Integrated Squared Error (MISE) criterion computed with respect to a benchmark density suggested in the literature.
Keywords: nonparametric estimation; kernel smoothing; kernel regression; kernel density estimation; convolution structure; stochastic volatility; Monte Carlo simulations.
Special Issue on: ICOAE2017 Applied Economics
Demand Monotonicity of a Pavement Cost Function Used to Determine Aumann-Shapley Values in Highway Cost Allocation
by Dong-Ju Lee, Saurav Kumar Dubey, Chang-Yong Lee, Alberto Garcia-Diaz
Abstract: Pavement thickness and traffic lanes are two essential requirements affecting the cost of a highway design project. The traffic loadings on a pavement are typically measured in 18-kip Equivalent Single-Axle Loads (ESAL). In this paper both ESAls and lanes are treated as two types of players and a pavement cost function is developed to determine the average marginal cost for each type of players. These averages are known as the Aumann-Shapley (A-S) values and are used to allocate the highway cost among all vehicle classes. The proposed pavement cost function is proved to be monotonically increasing as the traffic loadings (ESALs) are increased, a necessary condition for the function to be acceptable for computing AumannShapley values. A severe limitation of the procedure to calculate marginal costs for the trafficloading players is the extremely large number of permutations since the number of players is enormously high. To overcome this limitation this article derives a compact form for the discrete A-S values of ESALs and lanes that allows the Aummann-Shapely values to be calculated in a computationally effective manner. rn
Keywords: Cooperative Game Theory; Discrete Aumann-Shapley Values; Linear Optimization; Demand Monotonicity; Transportation Economics; Computational Efficiency.
Special Issue on: MIC 2017 Managing the Global Economy
Stabilization Policies in a Small Euro Area Economy: Taxes or Expenditures? A Case Study for Slovenia
by Klaus Weyerstrass, Reinhard Neck, Dmitri Blueschke, Boris Majcen, Andrej Srakar, Miroslav Verbič
Abstract: In this paper we investigate how effective stabilization policies can be in a small open economy which is part of the Euro Area, namely Slovenia. In particular, we analyse whether tax policy or expenditure policy has stronger multiplier effects. Slovenia is an interesting case because it is a small open economy in Central Europe that was already in the Euro Area before the Great Recession. Using the SLOPOL10 model, an econometric model of the Slovenian economy, we analyse the effectiveness of some categories of taxes and public spending. Some of these instruments are targeted towards the demand side, while others primarily influence the supply side. Our results show that those public spending measures that entail both demand and supply side effects are more effective at stimulating real GDP than pure demand side measures. Measures that improve the education level of the labour force are very effective at stimulating potential GDP. Employment can be most effectively stimulated by cutting the income tax rate and the social security contribution rate, i.e. by reducing the tax wedge on labour income, thereby positively affecting Slovenias international competitiveness. On the other hand, simulations show that fiscal policy measures can only mitigate but not undo the adverse effects of a crisis like the Great Recession.
Keywords: Macroeconomics; stabilization policy; fiscal policy; tax policy; public expenditures; Slovenia; public debt; econometric model; simulation.
Special Issue on: Machine Learning, Artificial Intelligence, and Big Data Methods and New Perspectives
A SAS Macro for Examining Stationarity Under the Presence of Endogenous Structural Breaks
by Dimitrios Dadakas, Scott Fargher
Abstract: The endogenous structural break literature present numerous computationally intensive procedures for the examination of stationarity under the presence of single or multiple structural breaks. Application of these grid-search procedures is rather complicated and not many researchers have access to code that can easily be applied. In this article, we present a SAS macro, that allows the examination of stationarity under the assumption of either one or two, endogenously determined, structural breaks using the Zivot and Andrews (1992) and the Lumsdaine and Pappel (1997) methodologies. We demonstrate the macro using the Nelson-Plosser (1982) data, that was also used by Zivot and Andrews (1992) and Lumsdaine and Pappel (1997), to highlight differences and similarities of the macro command with the original results published.
Keywords: Endogenous Structural Breaks; Stationarity; Time Series; SAS; Macro.
Automated detection of entry and exit nodes in traffic networks of irregular shape
by Simon Plakolb, Christian Hofer, Georg Jäger, Manfred Füllsack
Abstract: We devise an algorithm that can automatically identify entry and exit
nodes of an arbitrary traffic network. It is applicable even if the network is of
irregular shape, which is the case for many cities. Additionally, the method can
calculate the nodes attractiveness to commuters. This technique is then used to
improve a traffic model, so that it is no longer dependent on expert knowledge
and manual steps and can thus be used to analyse arbitrary traffic systems.
Evaluation of the algorithm is performed twofold: The positions of the identified
entry nodes are compared to existing traffic data. A more in-depth analysis uses
the traffic model to simulate a city in two ways: Once with hand-picked entry
nodes and once with automatically detected ones. The evaluation shows that the
simulation yields a good match to the real world data, substantiating the claim
that the algorithm can fully replace a manual identification process.
Keywords: traffic modelling; network analysis; commuting; automated detection; entry nodes; exit nodes; traffic simulation; mobility behaviour; agent-based model; road usage; congestion.
Depth based support vector classifiers to detect data nests of rare events
by Rainer Dyckerhoff, Hartmut Jakob Stenz
Abstract: The aim of this project is to combine data depth with support vector machines (svm) for binary classification. To this end, we introduce data depth functions and svm and discuss why a combination of the two is assumed to work better in some cases than using svm alone. For two classes X and Y, one investigates whether an individual data point should be assigned to one of these classes. In this context, our focus lies on the detection of rare events, which are structured in data nests: Class X contains much more data points than class Y and Y has less dispersion than X. This form of classification problem is akin to finding the proverbial needle in a haystack. Data structures like these are important in churn prediction analyses which will serve as a motivation for possible applications. Beyond the analytical investigations, comprehensive Simulation studies will also be carried out.
Keywords: Data depth; DD-Plot; Mahalanobis depth function; support vector machines,rnbinary classification; hybrid methods; rare events; data nest; churn prediction; big data.
Does time-frequency-scale analysis predict inflation? Evidence from Tunisia
by AMMOURI BILEL, Fakhri Issaoui, Habib ZITOUNA
Abstract: Forecasting macroeconomic indicators has always been an issue for economic policymakers. Different models are available in the literature; for example univariate and/or multivariate models, linear and/or non-linear models. This diversity requires a multiplicity of the used techniques.. They can be classified as pre and post-time series. However, this multiplicity allows the improvement of a better forecast of the macroeconomic indicators during unrest (be it political, economic, and/or social). In this paper, we deal with the problem of the performance of the macroeconomic models for predicting Tunisias inflation during instability following the 2011-revolution. To achieve this goal, the time-frequency-scale analysis (Fourier transform, Wavelet transform, and Stockwell transform) is used. In fact, we are interested in the ability of these techniques to improve predictor performances. We suggest the performance of the adopted approach (time-frequency-scale analysis). This performance is not quite absolute because their performance is less than multivariate model (Dynamic Factor Model) during economic instability.
Keywords: Inflation forecast; uni-varied model; multi-varied model; time-frequency analysis; Fourier transform; wavelet transform; Stockwell transform.