Published papers

A multivariate, non-Bayesian, regression-based, or feasible generalized least squares (GLS)-based approach is proposed to estimate time-varying VAR parameter models. Although it has been known that the Kalman-smoothed estimate can be alternatively estimated using GLS for univariate models, we assess the accuracy of the feasible GLS estimator compared with commonly used Bayesian estimators. Unlike the maximum likelihood estimator often used together with the Kalman filter, it is shown that the possibility of the pile-up problem occurring is negligible. In addition, this approach enables us to deal with stochastic volatility models, models with a time-dependent variance–covariance matrix, and models with non-Gaussian errors that allow us to deal with abrupt changes or structural breaks in time-varying parameters.

We propose to utilize the band spectral regression for out-of-sample forecasts of exchange rates. When one period ahead forecast is considered, there is some evidence that the band spectral regression improves its accuracy, especially when the Taylor rule fundamentals model is employed. However, when the forecasting horizon increases, the purchasing power parity (PPP) fundamentals model is found to be powerful, and we can improve the out-of-sample forecast by selecting appropriate frequency bands. Bayesian model averaging shows that placing a large weight on the business cycle frequency improves the accuracy of the out-of-sample forecasting of the PPP model (as well as the monetary fundamentals model) when a longer forecasting horizon is our focus. Without specifying the frequency bands prior to applying the regression, LASSO can provide better out-of-sample exchange rate forecasts for many cases – most patently for the PPP fundamentals model – and provide information about the dynamic relationship between forecasting variables and exchange rates. The frequency domain approach not only improves the accuracy of exchange rate forecast but provides insights for understanding why the PPP fundamentals act as a powerful predictor when the forecasting horizon increases and there is a possible improvement in the time domain regression forecast.

How strongly are foreign exchange markets linked in terms of their similarities in long-run fluctuations? Are they cointegrating? To analyze such “comovements,” we present a time-varying cointegration model for the foreign exchange rates of the currencies of Canada, Japan, and the UK vis-à-vis the U.S. dollar from May 1990 through July 2015. Unlike previous studies, we allow the loading matrix in the vector error-correction (VEC) model to be varying over time. Because the loading matrix in the VEC model is associated with the speed at which deviations from the long-run relationship disappear, we propose a new degree of market comovement based on the time-varying loading matrix to measure the strength or robustness of the long-run relationship over time. Since exchange rates are determined by macrovariables, cointegration among exchange rates implies these variables share common stochastic trends. Therefore, the proposed degree measures the degree of market comovement. Our main finding is that the market comovement has become stronger over the past quarter-century, but at a decreasing rate with two major turning points: one in 1995 and the other one in 2008.

This paper first generalizes the trend-cycle decomposition framework of Perron and Wada (2009) based on unobserved components models with innovations having a mixture of normals distribution, which is able to handle sudden level and slope changes to the trend function as well as outliers. We investigate how important are the differences in the implied trend and cycle compared to the popular decomposition based on the Hodrick and Prescott (HP) (1997) filter. Our results show important qualitative and quantitative differences in the implied cycles for both real GDP and consumption series for the G7 countries. Most of the differences can be ascribed to the fact that the HP filter does not handle well slope changes, level shifts and outliers, while our method does so. Then, we reassess how such different cycles affect some so-called "stylized facts" about the relative variability of consumption and output across countries.

A non-Bayesian time-varying model is developed by introducing the concept of the degree of market efficiency that varies over time. This model may be seen as a reflection of the idea that continuous technological progress alters the trading environment over time. With new methodologies and a new measure of the degree of market efficiency, we examine whether the US stock market evolves over time. In particular, a time-varying autoregressive (TV-AR) model is employed. Our main findings are: (i) the US stock market has evolved over time and the degree of market efficiency has cyclical fluctuations with a considerably long periodicity, from 30 to 40 years; and (ii) the US stock market has been efficient with the exception of four times in our sample period: during the long recession of 1873–1879; the recession of 1902–1904; the New Deal era; and the recession of 1957–1958 and soon after it. It is then shown that our results are partly consistent with the view of behavioural finance.

It has been common to assume that the relationship between economic activity and oil prices is asymmetric. Theoretical underpinnings for this asymmetry include costly sectoral reallocation, partial equilibrium models of irreversible investment, and some version of precautionary savings. Yet, recent studies that use new methodologies to test for asymmetries in U.S. data have cast some doubts on that premise. In this paper, we use state-of-the-art techniques to evaluate the presence of asymmetries for a set of OECD countries containing both oil exporters and oil importers. We find very little support for the hypothesis that the response of industrial production to oil price increases and decreases is asymmetric. Our results have important implications for theoretical models of the transmission of oil price shocks: they point towards the importance of direct-supply and direct-demand transmission channels, as well as indirect transmission channels that imply a symmetric response.

This article develops a non-Bayesian methodology to analyse the time-varying structure of international linkages and market efficiency in G7 countries. We consider a non-Bayesian time-varying vector autoregressive (TV-VAR) model, and apply it to estimate the joint degree of market efficiency in the sense of Fama (1970, 1991). Our empirical results provide a new perspective that the international linkages and market efficiency change over time and that their behaviours correspond well to historical events of the international financial system.

We propose an alternative model and method to reconcile the puzzling feature in the relationship between the real exchange rate and real interest rate differentials. Our simple two-country model with preset prices, along with firms' misperception about the future exchange rate, implies that the real exchange rate follows an ARIMA(0,1,p) process. This allows us to compute the exact Beveridge-Nelson decomposition, which is a model-consistent decomposition. In accordance with our model, unit roots in the real exchange rates are found; and statistical inference is partially found to be affirmative regarding the link between the real exchange rate detrended by the Beveridge-Nelson decomposition and corresponding real interest differentials.

This note provides explanations for an unexpected result, namely, the estimated parameter of the correlation coefficient of the trend shock and cycle shock in the state-space model is almost always (positive or negative) unity, even when the true variance of the trend shock is zero. It is shown that the set of the true parameter values lies on the restriction that requires the variance-covariance matrix of the errors to be nonsingular, therefore, almost always the likelihood function has its (constrained) global maximum on the boundary where the correlation coefficient implies perfect correlation.

This paper tests the three leading specifications of asymmetric and possibly nonlinear feedback from the real price of oil to U.S. industrial production and its sectoral components. We show that the evidence for such feedback is sensitive to the estimation period. Support for a nonlinear model is strongest for samples starting before 1973. Instead, using post-1973 data only, the evidence against symmetry becomes considerably weaker. For example, at the aggregate level, there is no evidence against the hypothesis of symmetric responses to oil price innovations of typical magnitude, consistent with results of Kilian and Vigfusson [Quantitative Economics, 2(3), 419-453 (2011)] for U.S. real GDP. There is strong evidence of asymmetries at the disaggregate level, however, especially for industries that are energy-intensive in production (such as chemicals) or that produce goods that are energy-intensive in use (such as transportation equipment). Our analysis suggests that these asymmetries may be obscured in the aggregate data and highlights the importance of developing multisector models of the transmission of oil price shocks.

This paper examines the effects of immigration on a specific occupation, registered nurses (RNs). To learn whether immigrant nurses reduced the earnings of RNs, we applied techniques developed by Goldin (1994) and Borjas, Freeman, and Katz (1996), but found the effect of immigrant penetration either positive or insignificant. We also found that the supply of immigrant RNs was far more elastic than the supply coming from natives.It is often argued that it will be hard to detect negative effects on wages and employment of natives in local markets, because natives will avoid a market which many immigrants have entered. This study finds no support for this hypothesis in this market, based on data that measures the rate of entry of RNs exactly. We find no adverse effect of immigration on native workers in this occupation.

Trend-cycle decompositions for US real GDP such as the unobserved components models, the Beveridge-Nelson decomposition, the Hodrick-Prescott filter and others yield very different cycles which bear little resemblance to the NBER chronology, ascribes much movements to the trend leaving little to the cycle, and some imply a negative correlation between the noise to the cycle and the trend. We argue that these features are artifacts created by the neglect of a change in the slope of the trend function. Once this is accounted for, all methods yield the same cycle with a trend that is non-stochastic except for a few periods around 1973. The cycle is more important in magnitude than previously reported and it accords well with the NBER chronology. Our results are corroborated using an alternative trend-cycle decomposition based on a generalized unobserved components models with errors having a mixture of normals distribution for both the slope of the trend function and the cyclical component.

We study a two-country model with changes in the technological growth rate. Such changes are attributed to transitory and persistent shocks in the growth rate of technology. Cases are considered in which agents in two countries do not have enough information to distinguish between the two types of shocks; gradually, however, the persistence of the shock is recognized through the learning process. Utilizing a set of parameters obtained from U.S. and European productivity growth rates, it is then shown that (i) when persistent shocks affect the two countries identically, there is no consumption-correlation puzzle, and the international comovement puzzle becomes imperceptible; and (ii) even when persistent shocks affect the two countries differently, imperfect information plays an important role in explaining both the consumption- correlation puzzle and the international comovement puzzle (provided transitory shocks are strongly internationally correlated and are relatively larger than persistent shocks).