Rolling Window Regression Research Articles

Preface: Celebration of the Career of Robert Nathan Ginsburg

Preface: Celebration of the Career of Robert Nathan Ginsburg

The Relationship Between Tourism Income and Financial Development in The Mena Countries

Purpose – The purpose of this study is to determine the causality between financial development and tourism income for the MENA economies between year 1995 and 2016. Design/Methodology/Approach – To do this, we employ Dimitrescu and Hurlin (2012) panel causality test depending on rolling window regression and Emirmahmutoglu and Kose’s (2011) panel causality test. Findings – In this analysis for Algeria, Bahrain, Egypt, Arab Rep., Iran, Islamic Rep., Israel, Jordan, Kuwait, Morocco, Sudan, Tunisia and Turkish economies the relationship between financial development and tourism income calculated by the IMF for the period of 1995-2016. Panel data analysis is employed in order to test the hypothesis. In the empirical analysis, it was concluded that the countries that formed the panel affected each other socially and economically. Discussion – According to the results of empirical analysis, when the relationship between financial development and tourism income is examined for each country forming the panel, it was found that the supply - driven approach in Jordan and Tunisia on the other hand in Morocco and Sudan demand driven approach is valid. On the other hand for the frequency domain test it is offered that the neutrality approach was valid in 2009 and 2016 and demand driven approach is valid in 2005-2013 (except 2009).

Send in the Clones? Hedge Fund Replication Using Futures Contracts

Replication products strive to offer investors some of the benefits of hedge funds while avoiding their high fees, illiquidity, and opacity. We test whether a replication algorithm can deliver the diversification and high Sharpe ratio that investors seek. Our procedure constructs monthly clone returns out-of-sample using fully collateralized futures positions held for one-month, with position sizes determined using rolling window regressions. Clone returns have high correlation with their hedge fund targets, indicating replication is possible. Clones also have high correlation with a buy-and-hold investment in stocks, however, and neither the targets nor their clones demonstrate successful time variation in factor loadings.

A trend analysis approach for air quality network data

Trend analysis of air pollutant concentrations becomes problematic when applied to data from air quality monitoring networks containing time series of differing lengths. The average trend from such data can be misleading due to biases in the monitoring network. For example, if new monitoring sites located in more polluted locations are added to a network, the introduction of these time series can leverage the trend upwards. A method for resolving this problem was developed, using rolling window regression to recursively calculate the change in pollutant concentration as a function of time, which can be used as a proxy for the true trend. The efficacy of the method was established by conducting simulations with known trends. The rolling change trend was shown to more accurately reflect the true trend than simply averaging the time series. Application of the technique to estimate trends in NOx, NO2 and NO2/NOx concentrations at London roadside monitoring sites over the period 2000–2017 revealed clear differences from the simple average. In particular, a significant monotonic downward trend in NOx concentration was observed, in stark contrast to the average trend, which suggested little change in NOx concentration had occurred over the same period. By accurately representing trends using time series of different lengths, this method has the benefit of being able to describe changes in air quality for locations and time periods with otherwise insufficient data.

Tree cover shows strong sensitivity to precipitation variability across the global tropics

AbstractAimVegetation is sensitive to mean annual precipitation (MAP), but the sensitivity of vegetation to precipitation variability (PV) is less clear. Tropical ecosystems are likely to experience increased PV in the future. Here we assessed the importance, magnitude and mechanism of PV effects on tree cover in the context of covarying environmental drivers such as fire, temperature and soil properties.LocationTropical land.Time period2000–2010.Major taxa studiedTrees.MethodsWe compiled climate, soil and remotely‐sensed tree cover data over tropical land. We then comprehensively assessed the contribution of PV at different time‐scales to tropical tree cover variations and estimated the sensitivity of tree cover to PV changes by conducting rolling‐window regression and variance decomposition analyses. We further adopted a mechanistic modelling approach to test whether water competition between trees and grasses can explain the observed effect of PV.ResultsWe find that PV contributes 33–56% to the total explained spatial variation (65–79%) in tree cover. The contribution of PV depends on MAP and is highest under intermediate MAP (500–1,500 mm). Tree cover generally increases with rainy day frequency and wet season length but shows mixed responses to inter‐annual PV. Based on the estimated sensitivity, tropical tree cover can decrease by 3–5% overall and by up to 20% in Amazonia under a 20% decrease in rainy days. Mechanistic modelling analysis reproduced the continental differences in tree cover along an MAP gradient.Main conclusionsUnder intermediate rainfall regimes (500–1,500 mm), PV can be a more important determinant of tropical tree cover than conventionally proposed drivers such as MAP and fire. The effect of PV likely results from the sensitivity of tree–grass competition to the temporal distribution of water resources. These results show that climate variability can strongly shape the biosphere.

The Informational Content of the Term-Spread in Forecasting the U.S. Inflation Rate: A Nonlinear Approach

The difficulty in modelling inflation and the significance in discovering the underlying data generating process of inflation is expressed in an ample literature regarding inflation forecasting. In this paper we evaluate nonlinear machine learning and econometric methodologies in forecasting the U.S. inflation based on autoregressive and structural models of the term structure. We employ two nonlinear methodologies: the econometric Least Absolute Shrinkage and Selection Operator (LASSO) and the machine learning Support Vector Regression (SVR) method. The SVR has never been used before in inflation forecasting considering the term–spread as a regressor. In doing so, we use a long monthly dataset spanning the period 1871:1–2015:3 that covers the entire history of inflation in the U.S. economy. For comparison reasons we also use OLS regression models as benchmark. In order to evaluate the contribution of the term-spread in inflation forecasting in different time periods, we measure the out-of-sample forecasting performance of all models using rolling window regressions. Considering various forecasting horizons, the empirical evidence suggests that the structural models do not outperform the autoregressive ones, regardless of the model’s method. Thus we conclude that the term-spread models are not more accurate than autoregressive ones in inflation forecasting.

Human Capital-Economic Growth Relationship: Finding the Most Relevant Level of Education in Pakistan

The study employed annual dataset for the period 1981-2014 to find the most relevant level of education in terms of its contribution to the economic growth of Pakistan. The study used three models assuming gross enrolment rates at primary, secondary and tertiary level as human capital and found secondary education as the most relevant level of education for economic growth. The long-run coefficients obtained through fully modified ordinary least square (FMOLS) are used to reconfirm the findings of dynamic ordinary least square (DOLS). Rolling window regression found that the human capital elasticity of aggregate output oscillated around zero in the sample period whereas CUSUM test found that the parameters of equation with secondary education as human capital are stable. It is recommended therefore that secondary education should be given preference in resource allocation to ensure long run economic growth of Pakistan.

The Informational Content of the Term Spread in Forecasting the US Inflation Rate: A Nonlinear Approach

AbstractThe difficulty in modelling inflation and the significance in discovering the underlying data‐generating process of inflation is expressed in an extensive literature regarding inflation forecasting. In this paper we evaluate nonlinear machine learning and econometric methodologies in forecasting US inflation based on autoregressive and structural models of the term structure. We employ two nonlinear methodologies: the econometric least absolute shrinkage and selection operator (LASSO) and the machine‐learning support vector regression (SVR) method. The SVR has never been used before in inflation forecasting considering the term spread as a regressor. In doing so, we use a long monthly dataset spanning the period 1871:1–2015:3 that covers the entire history of inflation in the US economy. For comparison purposes we also use ordinary least squares regression models as a benchmark. In order to evaluate the contribution of the term spread in inflation forecasting in different time periods, we measure the out‐of‐sample forecasting performance of all models using rolling window regressions. Considering various forecasting horizons, the empirical evidence suggests that the structural models do not outperform the autoregressive ones, regardless of the model's method. Thus we conclude that the term spread models are not more accurate than autoregressive models in inflation forecasting. Copyright © 2016 John Wiley & Sons, Ltd.

Impact of Tourism on Long-Run Economic Growth of Vietnam

Tourism is a potential industry as a new impetus for economic growth of Vietnam, and is identified as a key sector in Vietnamese economy, as such, the Vietnam government has identified the tourism industry as a key sector in the national economy and one of the crucial importance in securing its long-term growth. The study examines the role of tourism in driving the economic growth of Vietnam over the period of 1995-2013 by using Tourism-Led Growth Hypothesis (TLGH). The stability of TLGH is analyzed using a Rolling Window regression technique. The analyses performed in this study are based on an econometric methodology consisting of the pooled Ordinary Least Squares test, unit root test, co-integration analysis, and Granger causality test. The findings show that the TLGH is supported in the case of Vietnam. In other words, tourism has indeed played a key role in driving economic growth in Vietnam in recent decades. Moreover, the Rolling Window regression results confirm the stability of the TLGH in the case of Vietnam.

Herding, anti-herding behaviour in metal commodities futures: a novel portfolio-based approach

The purpose of this article is twofold. Motivated by the heated debate on the financialization of commodities, we examine the existence of herding behaviour in metal commodities futures. In order to identify any time-dependent properties reflected in time-varying parameters, we employ the overlapping rolling window regression technique. The empirical evidence confirms a time-varying anti-herding behaviour before the global financial crisis and the absence of herding or anti-herding behaviour during the crisis. Next we attempt to formally establish the link between the documented anti-herding behaviour and portfolio management with the use of dynamic conditional correlations via the DCC-GARCH family multivariate modelling. After specifying the correlations, an in-sample recursive dynamic Markowitz portfolio is constructed and monitored. By doing so, we attribute the anti-herding behaviour to different portfolio positioning and rebalancing. On the other hand, in the absence of herding or anti-herding behaviour, we document a shift in the correlations and covariances of the commodity futures especially during the crisis, resulting in a decrease of the portfolio weights together with a substantial cash flow towards the risk-free asset.

TRADE OPENNESS AND ECONOMIC GROWTH: EMPIRICAL EVIDENCE FROM INDIA

The main objective of this study is to develop first time trade openness index and use this index to examine the link between trade openness and economic growth in case of India. This study employs a new endogenous growth model for theoretical support, auto-regressive distributive lag model and rolling window regression method in order to determine long run and short run association between trade openness and economic growth. Further granger causality test is used to determine the long run and short run causal direction. The results reveal that human capital and physical capital are positively related to economic growth in the long run. On the other hand, trade openness index negatively impacts on economic growth in the long run. The new evidence is provided by the rolling window regression results i.e. the impact of trade openness index on economic growth is not stable throughout the sample. In the short run trade openness index is positively related to economic growth. The result of granger causality test confirms the validity of trade openness-led growth and human capital-led growth hypothesis in the short run and long run.

Send in the Clones? Hedge Fund Replication Using Futures Contracts

Replication products strive to offer investors some of the benefits of hedge funds while avoiding their high fees, illiquidity, and opacity. The authors of this article test whether a replication algorithm can deliver the diversification and high Sharpe ratio that investors seek. Their procedure constructs monthly clone returns out-of-sample using fully collateralized futures positions held for one-month, with position sizes determined using rolling window regressions. Clone returns have high correlation with their hedge fund targets, indicating replication is possible. Clones also have high correlation with a buy-and-hold investment in stocks, however, and neither the targets nor their clones demonstrate successful time variation in factor loadings.

Time Varying Herding and Anti-Herding Behavior and Dynamic Conditional Correlations in European Market Indices

We examine the existence of herding and anti-herding (positive herding) behavior in major European benchmarking stock market indices. Following the recent events that unfolded in the Eurozone sovereign debt crisis our analysis is further expanded on two subsamples namely north and south European countries. In order to identify any time-dependent properties reflected in time-varying parameters, we employ an overlapping rolling window regression of varying window sizes. The empirical evidence confirms a herding behavior during the global financial crisis for the South countries, and anti-herding behavior for both North and South subsets before and after the shock, independent of the window size. The herding and anti-herding behavior, are explained with the use of dynamic conditional correlation via DCC-GARCH family multivariate modeling. The results indicate that the correlations have recorded a significant decrease before and after the crisis for both North and South subsamples, accounting for the anti-herding behavior, whereas correlations for the South countries during the crisis have increased, thus justifying the herding behavior. Our results entail significant implications for international portfolio diversification strategies.

Now or Later? Trading Wind Power Closer to Real-Time and How Poorly Designed Subsidies Lead to Higher Balancing Costs

An important challenge facing many deregulated electricity markets is dealing with the increasing penetration of intermittent generation. Simulation studies have pointed to the advantages of trading closer to real-time with large amounts of intermittent generation. Using Danish data, I show that, as expected, shortfalls increase the probability of trade on the short-term market. But in the period between 2010 and 2012 surpluses are shown to decrease the probability of trade. This unexpected result is likely explained by wind power policies that discourages trading on Elbas and leads to unnecessarily high balancing costs. I use a rolling-window regression to support this claim.

Uporczywość inflacji i jej komponentów – badanie empiryczne dla Polski

The paper aims to evaluate inflation persistence at a disaggregated level. The measures of inflation persistence used in the exercise rely solely on time series methods of AR and long memory models (cf: Marques, 2004; Pivetta, Reis, 2007; Baillie, 1996) and are applied to Polish CPI and its 11 components. The choice between those two frameworks has been based on the results of the FDF test (Dolado, Gonzalo, Mayoral, 2006). The second part of the study consisted in investigation of the dynamics of persistence. An experiment of rolling window regressions revealed a decrease in the persistence of most of the price indices. A plausible source of the decline was a structural change, which occurred during the introduction of direct inflation targeting by the monetary authorities in Poland.

DOES FINANCIAL DEVELOPMENT HAMPER ECONOMIC GROWTH: EMPIRICAL EVIDENCE FROM BANGLADESH

The objective of this study is twofold. (a) Construct the first ever financial development index (FDI) for Bangladesh using the principal component method (PCM). (b) Use the FDI to explore the existence of a long run relationship between FDI and economic growth. The Augmented Dickey Fuller and the Ng-Perron unit root tests have been applied to examine the stationarity properties of the series. To explore a long run relation, the Autoregressive Distributed Lag (ARDL) approach to cointegration; and to assess the stability of the parameters, the rolling window regression approach have been used. The results show that the impact of real interest rate (RIR) and FDI on economic growth is negative. Estimates from rolling window method show that FDI and RIR are negatively related to economic growth for the years 1987–1988, 1992–1999, 2002–2006, 2008 and 2009; and 1986–1998, 2006 and 2007, respectively. The results may help policymakers formulate effective financial sector policies as a tool to promote economic growth in Bangladesh.

Economics, finance and development in China

Purpose – The purpose of this paper is to construct a financial development index for China and to analyze the relationship between the financial sector development index and economic growth.Design/methodology/approach – This study uses Johansen‐Juselius cointegration approach to determine long run relationship between variables. To determine the strength of causal relationship variance decomposition is used. The stability of coefficient is evaluated through rolling window regression method.Findings – The results of Johansen‐Juselius cointegration approach confirm long run relationship between financial development index and economic growth. Normalized cointegrating vector indicates that financial development index, real interest rate, capital and labor force positively determine economic growth in China. The yearly coefficient is provided by the rolling regression and indicates that financial development index negatively link to economic growth in 1991, 1992, 1994, 1995, 1999, 2000, 2003‐2005. Interest r...

Export-led growth hypothesis: Multivariate rolling window analysis of Pakistan

Most of the empirical works have focused on the causal relationship between exports and economic growth in the case of Pakistan, whereas terms of trade have been neglected. This paper investigates the nature of relationship between the exports, terms of trade and economic growth by using the Autoregressive Distributed Lag (ARDL) approach and rolling window regression method. The empirical findings indicate that long run relationship exist when real Gross Domestic Product (GDP) and real exports are dependent variables. Key words: Pakistan, export-led growth, autoregressive distributed lag, rolling window.

Import Demand Function for Bangladesh: A Rolling Window Analysis

This study aims to estimate aggregate import demand function for Bangladesh economy by using the data of 1980 to 2008. Estimation evidence provided by using autoregressive distributed lag (ADRL) approach to cointegration and rolling window regression method to estimate the coefficient of each observation in the sample by fixing the window size. The estimation result confirms long run relationship between imports, relative price and economic activity, and long run economic growth elasticity is (0.93) positive and relative price elasticity in the long run (-0.29) is negative. In contrast regression results of rolling window method demonstrates that the long run elasticities of national income variable are vary in the range of 0.81 to 0.96 and the relative price elasticities are negative according to the theory except few years.

Actual Versus Perceived Taylor Rules - How Predictable is the European Central Bank?

This paper investigates the predictability of the European monetary policy through the eyes of the professional forecasters from a large investment bank. The analysis is based on forward-looking Actual and Perceived Taylor Rules for the European Central Bank which are estimated in real-time using a newly constructed database for the period April 2000-November 2009. The former policy rule is based on the actual refi rate set by the Governing Council, while the latter is estimated for the bank's economists using their main point forecast for the upcoming refi rate decision as a dependent variable. The empirical evidence shows that the pattern of the refi rate is broadly well predicted by the professional forecasters even though the latter have foreseen more accurately the increases rather than the policy rate cuts. Second, the results point to an increasing responsiveness of the ECB to macroeconomic fundamentals along the forecast horizon. Third, the rolling window regressions suggest that the estimated coefficients have changed after the bankruptcy of Lehman Brothers in October 2008: the ECB has responded less strongly to macroeconomic fundamentals and the degree of policy inertia has decreased. A sensitivity analysis shows that the baseline results are robust to applying a recursive window methodology and some of the findings are qualitatively unaltered from using Consensus Economics forecasts in the regressions.