|
on Econometric Time Series |
| By: | Karol Gellert (Finance Discipline Group, UTS Business School, University of Technology Sydney); Erik Schlögl (Finance Discipline Group, UTS Business School, University of Technology Sydney) |
| Abstract: | This paper presents the construction of a particle filter, which incorporates elements inspired by genetic algorithms, in order to achieve accelerated adaptation of the estimated posterior distribution to changes in model parameters. Specifically, the filter is designed for the situation where the subsequent data in online sequential filtering does not match the model posterior filtered based on data up to a current point in time. The examples considered encompass parameter regime shifts and stochastic volatility. The filter adapts to regime shifts extremely rapidly and delivers a clear heuristic for distinguishing between regime shifts and stochastic volatility, even though the model dynamics assumed by the filter exhibit neither of those features. |
| Date: | 2018–06–01 |
| URL: | http://d.repec.org/n?u=RePEc:uts:rpaper:392&r=ets |
| By: | Igor L. Kheifets; Pentti J. Saikkonen |
| Abstract: | Smooth transition autoregressive models are widely used to capture nonlinearities in univariate and multivariate time series. Existence of stationary solution is typically assumed, implicitly or explicitly. In this paper we describe conditions for stationarity and ergodicity of vector STAR models. The key condition is that the joint spectral radius of certain matrices is below 1, which is not guaranteed if only separate spectral radii are below 1. Our result allows to use recently introduced toolboxes from computational mathematics to verify the stationarity and ergodicity of vector STAR models. |
| Date: | 2018–05 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:1805.11311&r=ets |
| By: | Leschinski, Christian; Sibbertsen, Philipp |
| Abstract: | We derive the properties of the periodogram local to the zero frequency for a large class of spurious long-memory processes. The periodogram is of crucial importance in this context, since it forms the basis for most commonly used estimation methods for the memory parameter. The class considered nests a wide range of processes such as deterministic or stochastic structural breaks and smooth trends as special cases. Several previous results on these special cases are generalized and extended. All of the spurious long-memory processes considered share the property that their impact on the periodogram at the Fourier frequencies local to the origin is different than that of true long-memory processes. Both types of processes therefore exhibit clearly distinct empirical features. |
| Keywords: | Long Memory; Spurious Long Memory; Structural Change |
| JEL: | C18 C32 |
| Date: | 2018–06 |
| URL: | http://d.repec.org/n?u=RePEc:han:dpaper:dp-632&r=ets |
| By: | Fabrizio Cipollini (Dipartimento di Statistica “G. Parenti”, Università di Firenze, Italy); Giampiero M. Gallo (Sezione Regionale di Controllo per la Lombardia, Corte dei Conti, Italy; Rimini Centre for Economic Analysis) |
| Abstract: | Similar volatility patterns are observables in the Euro area across national indices, suggesting the possibility of an underlying common component as a consequence of financial and monetary integration. This peculiar interdependence across market volatilities is captured by an additive component vector Multiplicative Error Model (vMEM) where the volatility dynamics is split between a common and a vector of market–specific components. When extracted from five major market indices and used as additional regressors in a HAR specification for the Euro STOXX 50 (a Euro area wide index) volatility, these components replace the terms that mimic long memory in the HAR, providing an interesting interpretation for volatility dynamics. |
| Keywords: | Realized Volatility, (vector) Multiplicative Error Models, GMM, HAR, Common Component, Euro area |
| Date: | 2018–06 |
| URL: | http://d.repec.org/n?u=RePEc:rim:rimwps:18-26&r=ets |
| By: | Abdullah Tahir (State Bank of Pakistan); Jameel Ahmed (State Bank of Pakistan); Waqas Ahmed (State Bank of Pakistan) |
| Abstract: | Business cycle dating, macroeconomic analysis and ex-ante policy prescription based on macroeconomic variables at annual data frequency is inadequate; as high frequency information on the state of the economy, otherwise inherent in quarterly data is averaged out at such low frequency. We use a robust method of disaggregating quarterly series from annual data, such that the aspect and information about the intervening business cycles is preserved. Extracting an orthogonal factor, which encompasses common variation (co-movements) of leading indicators of economic activity at quarterly data frequency, we use seemingly unrelated time series equation (SUTSE) model to disaggregate the annual GDP data into quarterly frequency. Utility of the quarterly GDP estimates is illustrated by (i) determining business cycle dates using a non-parametric Bry-Boschan (1971) algorithm and (ii) estimating the potential GDP and output gap for each of the 11 International Growth Center (IGC) partner countries. |
| Keywords: | Temporal Disaggregation, Business Cycle Dates, Dynamic Linear Model |
| JEL: | C32 E32 E58 |
| Date: | 2018–05 |
| URL: | http://d.repec.org/n?u=RePEc:sbp:wpaper:97&r=ets |
| By: | Afees A. Salisu (Centre for Econometric and Allied Research, University of Ibadan); Aviral Kumar Tiwari (2300 avenue des moulins, 3400 Montpellier, France Montpellier Business School, Montpellier, France.); Ibrahim D. Raheem (School of Economics, University of Kent, Canterbury, UK) |
| Abstract: | This study exploits several conditional heteroskedasticity models with various supported distributions in order to find the best distribution as well as the best GARCH-type model that may be used to model volatility of Bitcoin returns. Innovatively, the study is able to establish that pre-testing the residuals of Bitcoin returns for the best distribution can help to identify the appropriate distribution when modelling with GARCH-type models regardless of the data frequency. |
| Keywords: | Bitcoin returns; GARCH-type models; Error distributions |
| Date: | 2018–06 |
| URL: | http://d.repec.org/n?u=RePEc:cui:wpaper:0058&r=ets |
| By: | Akpa, Emeka |
| Abstract: | This study examined the short and long-run effects of private remittances on household consumption in Ghana from 1980 to 2016, controlling for structural breaks. Autoregressive Distributed Lag (ARDL) technique was used to investigate the relationship. Results showed that remittances positively impacted household consumption in Ghana in both the short and long-run howbeit, statistically insignificant. Hence, it was recommended, among other things that the government of Ghana collaborates with financial institutions to make remittance less expensive to encourage more remitting from the diaspora. |
| Keywords: | Remittances, Household Consumption, ARDL, Structural Breaks |
| JEL: | C32 F24 |
| Date: | 2018–05–25 |
| URL: | http://d.repec.org/n?u=RePEc:pra:mprapa:87103&r=ets |
| By: | Mendy, David; Widodo, Tri |
| Abstract: | This paper aims to identify the Indonesia rupiah per US dollar turning points using a regime switching model. Firstly, to detect if nonlinear model suits the data at hand, the BDS test and CUSUM of square test was used and the results indicates that a nonlinear model suits the data. The paper then proceeds by using a univariate two state Markov Switching autoregressive model (MSAR) developed by Hamilton (1989), Engel and Hamilton (1990) to capture regime shifts behaviour in both the mean and the variance of the Indonesian rupiah per US dollar exchange rate between 2000 to 2015. The empirical evidence indicates strong transition probabilities suggesting that only extreme events can switch the series from an appreciation regime to a depreciation regime vice versa. Moreover, the results of the MSAR model was found to successfully capture the timing of the regime shifts of the Indonesian rupiah per US dollar exchange rate because of government intervention, Indonesian presidential elections, US financial crises of 2008, and the Indonesian current account deficit in 2013. Finally, the non-linear exchange rate dynamic of the Indonesian rupiah implied that regime-switching models have potential important implication for the macroeconomic literature documenting the effect of monetary policy shock and political environment on the economic aggregates. Furthermore, regime-switching models is well suited to capture the non-linearities in exchanges rate. |
| Keywords: | Exchange rates (Indonesian Rupiah per US Dollar), Nonlinearity, Markov switching model(MSAR) |
| JEL: | E3 E5 |
| Date: | 2018–05–05 |
| URL: | http://d.repec.org/n?u=RePEc:pra:mprapa:86728&r=ets |
| By: | Gosego Mothuti (Department of Economics, Nelson Mandela University); Andrew Phiri (Department of Economics, Nelson Mandela University) |
| Abstract: | Does inflation affect economic growth in Botswana over the short-run and long-run? In applying bounds procedure for modelling ARDL cointegation effects applied to empirical data collected between 1975 and 2016 we find that this hypothesis does not hold true for Botswana as inflation is found to be insignificantly related with economic growth over both the short and long-run. Our growth equation estimates point to increased exports and an appreciated Pula currency as having a significant positive effect on steady-state GDP growth. Further empirical exercises show that an appreciated Pula increases inflation whilst a depreciated Pula decreases inflation for the Botswana economy. Policymakers should be this aware that attainment of lower inflation rates which occurs through a depreciated Pula/Dollar currency will only retard economic growth. |
| Keywords: | Inflation, Economic growth, Exchange rates, Bank of Botswana, Nonlinear autoregressive distributive lag (N-ARDL) model. |
| JEL: | C13 C32 C52 E31 F43 |
| Date: | 2018–06 |
| URL: | http://d.repec.org/n?u=RePEc:mnd:wpaper:1824&r=ets |
| By: | Louisa Kammerer; Miguel D. Ramirez (Department of Economics, Trinity College) |
| Abstract: | This paper examines the challenges firms(and policymakers) encounter when confronted by a recession at the zero lower bound, when traditional monetary policy is ineffective in the face of deteriorated balance sheets and high costs of credit. Within the larger body of literature, this paper focuses on the cost of credit during a recession, which constrains smaller firms from borrowing and investing, thus magnifying the contraction. Extending and revising a model originally developed by Walker (2010) and estimated by Pandey and Ramirez (2012), this study uses a Vector Error Correction Model with structural breaks to analyze the effects of relevant economic and financial factors on the cost of credit intermediation for small and large firms. Specifically, it tests whether large firms have advantageous access to credit, especially during recessions. The findings suggest that during the Great Recession of 2007-09 the cost of credit rose for small firms while it decreased for large firms, ceteris paribus. From the results, the paper assesses alternative ways in which the central bank can respond to a recession facing the zero lower bound. |
| Keywords: | Cost of credit; General impulse response functions; Granger causality test; Great Recession; Gregory Hansen single-break cointegration test; Johansen cointegration test; KPSS unit root test; Monetary policy; Vector error correction model (VECM); Small and large firms; and Zero lower bound (ZLB). |
| JEL: | C22 E50 G01 |
| Date: | 2018–05 |
| URL: | http://d.repec.org/n?u=RePEc:tri:wpaper:1707&r=ets |
| By: | Nilofer, Nilofer; Qayyum, Abdul |
| Abstract: | Investment is vital ingredients of growth in an economy. Saving contributes to investment which contributes to physical and human capital formation both of which promote growth of Gross Domestic Product (GDP) of a country. This study aims at determining the role of the three types of investment i.e., public, private and foreign direct investment (FDI) in the growth of Pakistan economy with a special focus on the contribution of FDI in GDP growth of the Pakistan. Cointegration analysis of time series data was used to analyze model. Autoregressive Distributed Lag (ARDL) approach has been used to analyze the long run relationship between GDP growth, investment and government expenditure for Pakistan using data (1970-2015). The results indicate that while public and private investment and lending rate have a positive impact on growth, public consumption and FDI decelerate GDP growth. Also the investor confidence should be bolstered by improving the law and order and security situation of the country and introducing investment friendly policies to further harness the positive impact of investment on growth. |
| Keywords: | Investment, FDI, Growth, Cointegration, Autoregressive Distributed Lag Model, Bounds Testing, Pakistan |
| JEL: | E22 O4 |
| Date: | 2018 |
| URL: | http://d.repec.org/n?u=RePEc:pra:mprapa:86961&r=ets |