A small sample test for non-nested regression models

Optimal redistributive tax-subsidy schemes are devised leading to maximum unweighted and class-weighted balanced growth per capita consumption levels in the two-class neoclassical model. A subsidy, and, if workers are numerically dominant, a tax are required in each respective case.     

A quantile regression approach for estimating panel data models using instrumental variables

We introduce a quantile regression approach to panel data models with endogenous variables and individual effects correlated with the independent variables. We find newly developed quantile regression methods can be easily adapted to estimate this class of models efficiently.     

Inference approaches for instrumental variable quantile regression

We consider asymptotic and finite sample confidence bounds in instrumental variables quantile regressions of wages on schooling with relatively weak instruments. We find practically important differences between the asymptotic and finite sample interval estimates.     

Normality test for multivariate conditional heteroskedastic dynamic regression models

In this paper, we study the Jarque-Bera test for the normality of the innovations of multivariate GARCH models. It is shown that the test is distribution free and its limiting null distribution is a chi-square distribution.     

A non-nested test of the new classical neutrality proposition for Canada

This paper uses a non-nested testing technique to test the neutrality proposition of the new classical economics. A new classical model and an alternative model are estimated and tested against each other. The test results indicate that the alternative model rejects the new classical model, and that the neutrality proposition cannot be supported empirically.     

Asymmetric inflation dynamics: Evidence from quantile regression analysis

This paper applies the regression quantile approach developed by Koenker and Xiao (2004) to investigate the dynamic behavior of inflation in 12 OECD countries. By analyzing the behavior in a wide range of quantiles, this method allows us to quantify the influence of various sizes of shocks that hit the inflation, and is able to capture possible asymmetric adjustment of the inflation towards to its long-run equilibrium. It therefore sheds new lights on the inflation dynamics compared with the conventional unit root methodologies. Our results suggest that generally, the inflation rates are not only mean-reverting but also exhibit asymmetries in their dynamic adjustments, in which large negative shocks tend to induce strong mean reversion, and on the contrary, large positive shocks do not. Policy implications related to the empirical findings are also provided.     

Revisiting growth empirics based on IV panel quantile regression

We analyse the well-known issue of economic growth convergence using quantile regression. Most previous studies have used a least squares (LS) method or variation, which focuses on the issue only at the mean of the growth rate. Therefore, such results cannot provide a satisfactory answer to what can happen if the growth rate is far from the conditional mean level. For example, we consider the following question: do we still have economic growth convergence or is the convergence speed changed in a low growth period such as the ‘Great Recession,’ that started in 2008? We propose using instrumental variable panel quantile regression to answer this question. Our empirical findings demonstrate that economic growth convergence occurs at all quantiles over the entire conditional distribution, but that the convergence speed does depend on quantiles; the convergence speed is much higher when the GDP growth rate is at either high or low quantiles.     

The CUSUM of squares test for the stability of regression models with non-stationary regressors

In this paper, we propose a modified CUSUM of squares test in time series regression models with a non-stationary regressor and show that the limiting distribution of this test is the sup of the absolute value of a Brownian bridge.     

Monetary environments and stock returns revisited: A quantile regression approach

We investigate the impact of monetary conditions on stock market returns at different points on the return distributions. Our results reveal no association between stock returns and monetary environments at the lower quantiles. At the upper quantiles, however, we find that expansive monetary conditions lead to significantly larger stock returns. The relationship between returns and monetary conditions at the upper quantiles is also found to be asymmetric, exhibiting a monotonic increase in responsiveness at successive quantiles.     

A quantile regression decomposition of urban–rural inequality in Vietnam

We use the Vietnam Living Standards Surveys from 1993 and 1998 to examine inequality in welfare between urban and rural areas in Vietnam. Real per capita household consumption expenditure (RPCE) is our measure of welfare. We apply a quantile regression decomposition technique to analyze the difference between the urban and rural distributions of log RPCE. In the earlier survey, the urban–rural gap is primarily due to differences in covariates such as education, ethnicity, and age. This is true across the entire distribution. In the later survey, this is true only for lowest quantiles. For the rest of the distribution, the gap is primarily due to differences in returns to covariates between the urban and rural sectors.     

The effect of foreign aid on corruption: A quantile regression approach

This paper investigates the effect of foreign aid on corruption using a quantile regression method. We show that foreign aid generally reduces corruption, and its reduction effect is greater in less corrupt countries. Moreover, this effect is different by different donor countries.     

The absolute health income hypothesis revisited: a semiparametric quantile regression approach

This paper uses the 1998–1999 Canadian National Population Health Survey data to examine the health–income relationship that underlies the absolute income hypothesis. To allow for nonlinearity and data heterogeneity, we use a partially linear semiparametric quantile regression model. The “absolute income hypothesis” is partially true; the negative aging effects appear more pronounced for the ill-healthy population than for the healthy population and when annual income is below 40,000 Canadian dollars. We would like to thank two anonymous referees and Baldev Raj, the editor, for useful and constructive comments and suggestions. The views expressed in this article are those of the authors and do not necessarily reflect the views of Statistics Canada. Both authors would also like to acknowledge financial support from SSHRC of Canada.     

Measuring the incentives to learn in Colombia using new quantile regression approaches

This paper employs newly developed quantile regression techniques to investigate a policy that could differentially affect students' performance. The Colombian vouchers were assigned using lotteries, and were renewable as long as the students maintained satisfactory academic progress. This second aspect of the program may provide incentives for low attainment students to work harder. The evidence supports the hypothesis that incentives could account for the impact of the vouchers, including lower repetition rate. The effect of the vouchers is largest in the lower tail of the educational attainment distribution, a possibility that was conjectured by others, but has not yet been confirmed empirically. The evidence suggests that the incentive effect of the program increases weak students' test scores by at least 0.1 standard deviations, roughly the score gain associated to a half year of school learning.     

School composition effects in Denmark: quantile regression evidence from PISA 2000

Data from the first wave of the OECD PISA study are combined with register data for Denmark to estimate the effect of the socioeconomic mix of schools on students’ test scores. A major disadvantage of the PISA design for the analysis of school composition effects is the small students-per-school samples. Adding family background data from administrative registers for all same-aged schoolmates of the PISA students helps overcome this. To compensate for endogeneity in the school composition variable, the results are conditioned on a rich set of family and school variables from the PISA data. Quantile regression results suggest differential school composition effects across the conditional reading score distribution, with students in the lower quantiles achieving the largest test score gains. Mathematics results suggest that high- and low-ability students benefit equally from attending schools with a better student intake, and most results for science are only marginally significant. These results imply that mixing students of different home backgrounds could improve equity of achievement for both reading and mathematics; however, the average skill level would improve only for reading literacy. In mathematics, mixing students would not raise average outcomes, because the detrimental effect on students in the higher quantiles would offset positive effects on those in the lower quantiles. I thank Amelie Constant, Bernd Fitzenberger, Eskil Heinesen, Peter Jensen, Craig Riddell, Michael Rosholm, Nina Smith, Robert Wright and participants at the ESPE and EALE 2003 conferences and at AKF seminars, and two anonymous referees for helpful comments and suggestions. Financial support provided by the Danish Social Science Research Council is gratefully acknowledged.     

Testing for normality in linear regression models using regression and scale equivariant estimators

In this paper we provide a general solution to the problem of controlling the probability of a type I error in normality tests for the disturbances in linear regressions when using robust-regression residuals. We show that many classes of well-known robust regression estimators belong to the class of regression and scale equivariant estimators. It is these equivariance properties that are used to reduce the nuisance parameter space under the null, from which we develop Monte Carlo and Maximized Monte Carlo tests for the null of disturbance normality. Finally, we illustrate in a simulation experiment the potential power gains from using robust-regression residuals in testing this null hypothesis.     

Equity market information and credit risk signaling: A quantile cointegrating regression approach

Investigating linkages between credit and equity markets, we consider daily aggregate U.S. CDS spreads as well as well-chosen equity market and implied volatility indexes over ten years. We describe such robust (to spurious correlation) relationship with the quantile cointegrating regression approach. Such approach handles extreme quantiles/CDS values and their behavior with respect to the equity market's influence. Heteroskedastic patterns such as time-varying variance, but also autocorrelation, skewness and leptokurtosis are captured. Thus, the sensitivity of aggregate CDS spreads to equity market price and volatility channels is accurately measured across quantiles and spreads. Such quantile-dependent sensitivity exhibits asymmetric responses to equity market shocks. A sub-period analysis investigates potential regime shifts in estimated quantile cointegrating regressions. Quantile cointegrating coefficients vary over time and quantiles, and exhibit different magnitudes across sub-periods and spreads. Therefore, the relationship is unstable over time. We also propose a scenario analysis and risk signaling application for credit risk management prospects. Under specific risk levels, credit risky situations are described conditional on the equity market's information over time, and related expected aggregate CDS spreads are computed. Estimated conditional quantiles/CDS spreads act as credit alert triggers.     

The minimum error variance rule for non-linear regression models

In this note it is demontrated that Theil's (1961) minimum error variance criterion is asymptotically valid for choosing between non-nested non-linear regression models, as long as one of the models being considered is ‘true’.     

Nonlinear regression for unit root models with autoregressive errors

This paper shows that the nonlinear least squares estimator for unit root models has the limiting distribution free of nuisance parameters and is more efficient than the augmented Dickey–Fuller estimator when the sum of coefficients for lagged variables is negative.