Stefano Fachin

Bootstrap and Asymptotic Tests of Long‐run Relationships in Cointegrated Systems

Hypothesis testing on cointegrating vectors based on the asymptotic distributions of the test statistics are known to suffer from severe small sample size distortion. In this paper an alternative bootstrap procedure is proposed and evaluated through a Monte Carlo experiment, finding that the Type I errors are close to the nominal signficance levels but power might be not entirely adequate. It is then shown that a combined test based on the outcomes of both the asymptotic and the bootstrap tests will have both correct size and low Type II error, therefore improving the currently available procedures.

The Size and Power of Bootstrap and Bartlett-Corrected Tests of Hypotheses on the Cointegrating Vectors

In this paper we compare Bartlett-corrected, bootstrap, and fast double bootstrap tests on maximum likelihood estimates of cointegration parameters. The key result is that both the bootstrap and the Bartlett-corrected tests must be based on the unrestricted estimates of the cointegrating vectors: procedures based on the restricted estimates have almost no power. The small sample size bias of the asymptotic test appears so severe as to advise strongly against its use with the sample sizes commonly available; the fast double bootstrap test minimizes size bias, while the Bartlett-corrected test is somehow more powerful.

Asymptotic normal and bootstrap inference in structural VAR analysis

The aim of the paper is to examine the performance of bootstrap and asymptotic parametric inference methods in structural VAR analysis. The results obtained through a Monte Carlo experiment suggest that the two approaches are largely equivalent in most, but not all, cases. While the asymptotic method turns out to be surprisingly robust with respect to the distribution of the errors, the bootstrap does deliver results superior in terms of both length of the confidence interval and coverage when highly non-linear statistics (such as the components of the variance of the forecast error) are considered.

Testing for Breaks in Cointegrated Panels - with an Application to the Feldstein-Horioka Puzzle

Stability tests for cointegrating coefficients are known to have very low power with small to medium sample sizes. In this paper we propose to solve this problem by extending the tests to dependent cointegrated panels through the stationary bootstrap. Simulation evidence shows that the proposed panel tests improve considerably on asymptotic tests applied to individual series. As an empirical illustration we examined investment and saving for a panel of 14 European countries over the 1960-2002 period. While the individual stability tests, contrary to expectations and graphical evidence, in almost all cases do not reject the null of stability, the bootstrap panel tests lead to the more plausible conclusion that the long-run relationship between these two variables is likely to have undergone a break.

Bootstrapping unit root tests

Unit root tests have been known for a long time to suffer from a variety of problems. Considering that simulation methods are obvious candidates for solving some of these problems, the aim of this paper is to assess the performance of bootstrap tests of the unit root hypothesis of the Dickey-Fuller type. The results obtained show that boostrap tests have empirical sizes very close to the nominal ones and deliver rejection rates generally at least as high as those obtained using simulated critical points, and are therefore a promising alternative to the latter. The applications sto non-standard problems such as structural stability analysis appear to be especially promising.

Trends of labour productivity in Italy: a study with panel co‐integration methods

Purpose - The main aim of this paper is to examine labour productivity trends in Italy over the period 1981-2004. Design/methodology/approach - To this end, relying on recent developments in the analysis of non-stationary dependent panels, the paper develops a new method for estimating total factor productivity (TFP) trends. Findings - The conclusions confirm the view that the recent decline in Italian labour productivity growth is mostly due to a widespread fall in TFP growth. Research limitations/implications - The main assumption underlying the proposed TFP estimation method is that technology growth is driven by a single trend common to all units included in the panel (industries, regions or countries). Originality/value - The paper provides two distinct contributions: empirically, it provides robust evidence that TFP slow-down is the main cause of recent negative trends in labour productivity in Italy. Methodologically, the paper proposes an approach to estimating TFP that enjoys several advantages: only basic data for input and output flows are needed, the non-stationary nature of the data is explicitly taken into account, and confidence intervals for TFP growth can be computed. This method can thus be easily applied to many routinely available datasets, to either corroborate existing growth accounting estimates or to obtain previously unavailable estimates.

Grana Padano cheese : thermoanalytical techniques applied to the study of ripening

Abstract In the present work, modifications of water–matrix interactions in Grana Padano (GP) cheese were studied using thermogravimetric techniques in order to provide an index of ripening. These techniques represent a simple and rapid method for determining water content and for monitoring its behaviour in ripening GP cheese. It is possible to evaluate the quantity of water present in the matrix, as well as the different types of water bound to it both qualitatively and quantitatively. Samples of traditional GP cheese produced in the winter season (same day), and collected at different ripening stages (5,10,14,19 months), were investigated. When statistical models of the STARMAX (Space Time Autoregressive Moving Average eXogenous variables) class, augmented with surface trends, were fitted to thermoanalytical data, valuable insights into the chemical nature of the ripening were acheived with good predictive performances.

The Long‐Run Relationship between Savings and Investment in Oil‐Exporting Developing Countries: A Case Study of the Gulf Arab States

The relationship between national saving and investment over the long term is examined for six Gulf Arab oil-exporting developing countries -- Bahrain, Kuwait, Oman, Qatar, Saudi Arabia and the United Arab Emirates. We show that, provided some large outliers are properly accounted for, long-run equilibrium relationships between saving and investment (both total and fixed) exist in these countries. Since these countries have typically large current account surpluses such relationships cannot be explained by standard arguments. Our hypothesis is that the response of investment to saving largely depends on domestic absorptive capacity.

A Note on the Estimation of Long-Run Relationships in Panel Equations with Cross-Section Linkages

We address the issue of estimation and inference in dependent non-stationary panels of small cross-section dimensions. The main conclusion is that the best results are obtained applying bootstrap inference to single-equation estimators, such as FM-OLS and DOLS. SUR estimators perform badly, or are even unfeasible, when the time dimension is not very large compared to the cross-section dimension.

Exploring 3D datasets: a factorial matrices analysis of the US industry in the 1980s

In this paper a Factorial Matrices technique suitable for exploratory analysis of multivariate, disaggregated time series is presented and applied to a data set covering 19 US manufacturing industries over the years 1979 to 1990. The empirical analysis confirms that the technique is a powerful tool, allowing otherwise difficult extraction of stylized facts from multidimensional datasets. In this case: (i) there are no signs of deindustrialization induced by growing import penetration, and, (ii), employment decline has generally not been associated to substitution of capital to labour.