Misleading Inferences from Panel Unit Root Tests: a Comment

In their 2001 paper, Breuer, McNown, and Wallace argue that panel unit root tests are flawed in that the “all or nothing nature of these procedures may lead to serious misinterpretations of tests applied to mixed panels” (2001, p. 483). The authors propose an alternative test, the SURADF test, which allows for and estimates heterogeneous rates of convergence. We show that SURADF test results are highly sensitive to the selection of panel members. As such, application of the SURADF test seems to require a sound basis for panel member selection and a caveat that all results are panel‐specific.     

Structural Breaks and Convergence in the Regions of Peru: 1970–2010

The aim of this paper is the analysis of stochastic and β‐convergence in the relative regional per capita outputs using different unit root tests both with and without structural breaks and using a further test that is robust to the presence of I(0) or I(1) errors. It allows robust inference on the estimates of the initial per capita output (intercepts) and the respective growth rates (slopes). The results of the application of unit root tests without structural breaks show the absence of stochastic convergence. However, by incorporating the presence of endogenous breaks, the results are reversed for all regions. In the case of β‐convergence, the results of the robust test W_RQF show that all regions have a structural break at some point during the period 1970–2010. We find different behavior for different regions. There is a catching‐up process and a lagging‐behind process for different groups of regions towards more negative or more positive paths.     

IMMIGRATION UNEMPLOYMENT RELATIONSHIP: THE EVIDENCE FROM CANADA*

This paper examines the relationship between unemployment and immigration in Canada. The bi‐directional causality test finds no evidence of a significant effect of Canadian immigration on unemployment. Cointegration tests indicate that there is no observed increase in aggregate unemployment due to immigration in the long run. The results from the causality test based on the vector error correction model confirm that, in the short run, past unemployment does cause (less) immigration but not vice versa. There is also a long‐run positive relationship among per‐capita GDP, immigration rate and real wages. The results indicate that, in the short‐run, more immigration is possibly associated with attractive Canadian immigration policies, and in the long‐run, as the labour market adjusts, Canadian‐born workers are likely to benefit from increased migration.     

TESTING FOR STATIONARITY IN HETEROGENEOUS PANEL DATA IN THE CASE OF MODEL MISSPECIFICATION

This paper investigates the performance of the tests proposed by Hadri and by Hadri and Larsson for testing for stationarity in heterogeneous panel data under model misspecification. The panel tests are based on the well known KPSS test (cf. Kwiatkowski et al.) which considers two models: stationarity around a deterministic level and stationarity around a deterministic trend. There is no study, as far as we know, on the statistical properties of the test when the wrong model is used. We also consider the case of the simultaneous presence of the two types of models in a panel. We employ two asymptotics: joint asymptotic, T, N →∞ simultaneously, and T fixed and N allowed to grow indefinitely. We use Monte Carlo experiments to investigate the effects of misspecification in sample sizes usually used in practice. The results indicate that the assumption that T is fixed rather than asymptotic leads to tests that have less size distortions, particularly for relatively small T with large N panels (micro‐panels) than the tests derived under the joint asymptotics. We also find that choosing a deterministic trend when a deterministic level is true does not significantly affect the properties of the test. But, choosing a deterministic level when a deterministic trend is true leads to extreme over‐rejections. Therefore, when unsure about which model has generated the data, it is suggested to use the model with a trend. We also propose a new statistic for testing for stationarity in mixed panel data where the mixture is known. The performance of this new test is very good for both cases of T asymptotic and T fixed. The statistic for T asymptotic is slightly undersized when T is very small (≤10).     

DO REAL INTEREST RATES REALLY CONTAIN A UNIT ROOT? MORE EVIDENCE FROM A BOOTSTRAP COVARIATE UNIT ROOT TEST

This paper re‐examines the empirical finding that international real interest rates usually have a unit root. This conclusion is put forth in Rapach and Weber (2004 ), using the Ng and Perron (2001 ) tests. We use Rudebusch's (1993 ) approach to construct the small sample distributions of the Ng and Perron tests, and calculate their asymptotic sizes, size‐adjusted powers and rejection rates. These numbers show that the lack of power in the Ng and Perron tests might account for the findings of Rapach and Weber (2004 ): that the unit root null cannot be rejected for most OECD countries. Size distortions are mild in the case of Ng and Perron tests for two series, but are serious for the Phillips and Perron Z‐test on inflation rates. We then apply a powerful covariate augmented Dickey–Fuller unit root test to examine the series for which stationarity cannot be determined with the Ng and Perron tests. The bootstrap technique is also used to control possible size distortions. In contrast to the results of Rapach and Weber (2004 ), the bootstrap covariate augmented Dickey–Fuller test yields striking evidence that real interest rates are stationary for 14 of 16 OECD countries, because nominal interest rates are stationary for the 14 countries, while inflation rates are stationary for all countries.     

TESTING STRICT STATIONARITY WITH APPLICATIONS TO MACROECONOMIC TIME SERIES

We propose a model‐free test for strict stationarity. The idea is to estimate a nonparametric time‐varying characteristic function and compare it with the empirical characteristic function based on the whole sample. We also propose several derivative tests to check time‐invariant moments, weak stationarity, and pth order stationarity. Monte Carlo studies demonstrate excellent power of our tests. We apply our tests to various macroeconomic time series and find overwhelming evidence against strict and weak stationarity for both level and first‐differenced series. This suggests that the conventional time series econometric modeling strategies may have room to be improved by accommodating these time‐varying features.     

Optimising and Non‐optimising Balance of Trade Models: A Comparative Evidence from India

This study estimates and compares the competing optimising and non‐optimising balance of trade models using Indian data. The results obtained from the optimising model suggest that the prices relative to user cost of capital and the real wealth lead to a deterioration, while the real capital stock results in an improvement in trade balance. The estimates of conventional non‐optimising balance of trade model show the significant effect of domestic income and real exchange rate and the insignificant effect of world income on the balance of trade. The error correction models reinforce the long run estimates and show the significant effect of lagged equilibria on the balance of trade. The non‐nested hypothesis tests provide mixed evidence for the preference of one model over the other. The J test suggests that the optimising model outperforms the non‐optimising model, while the F test shows that both these models are acceptable in explaining the balance of trade.     

HYSTERESIS HYPOTHESIS IN UNEMPLOYMENT AND LABOUR FORCE PARTICIPATION RATES: EVIDENCE FROM AUSTRALIAN STATES AND TERRITORIES *

This article simultaneously investigates the hysteresis hypothesis of unemployment and labour force participation using regional level data in Australia. The conventional univariate and panel unit root tests indicate that the hysteresis hypothesis cannot be rejected for most of the regions in Australia. To further confirm if the hysteresis finding in unemployment and labour force participation results from not considering structural breaks. We employ a panel stationarity test recently developed by Carrion‐i‐Silvestre et al. (2005 ), which considers multiple structural breaks and cross‐sectional dependence. The test results lend support for unemployment hysteresis and participation regime‐wise stationarity. Accordingly, the findings imply that a temporary shock may have permanent effects on the unemployment rate but not on labour force participation in Australia and thus call for policies aimed at improving the adjustment mechanism in unemployment.     

Can individuals’ beliefs help us understand nonadherence to malaria test results? Evidence from rural Kenya

In malaria‐endemic countries about a quarter of test‐negative individuals take antimalarials (artemisinin‐based combination therapies [ACTs]). ACT overuse depletes scarce resources for subsidies and contributes to parasite resistance. As part of an experiment in Kenya that provided subsidies for rapid diagnostic test and/or for ACTs conditionally on being positive, we studied the association between beliefs on malaria status (prior and posterior the intervention) and decisions to get tested and to purchase ACTs. We find that prior beliefs do not explain the decision of getting tested (conditional on the price) and nonadherence to a negative test. However, test‐negative individuals who purchase ACTs report higher posterior beliefs than those who do not, consistent with a framework in which the formers revise beliefs upward, while the latters do not change or revise downward. We also do not find evidence that prior beliefs on ACT effectiveness and trust in test results play any major role in explaining testing or treatment behavior. Further research is needed to improve adherence to malaria‐negative test results.     

How might the negative impact of higher education on growth be explained? The role of vertical qualification (mis)match in an MRW‐type model

This study explores the impact of higher education on growth in 11 new EU members over the period 2000–2016 using an augmented MRW‐type model which takes into account the quality of education and vertical mismatch. More precisely, it tests whether the accelerating increase of college graduates causes an oversupply of educated labour which, in turn, extends the qualification mismatch and adversely affects growth of per capita income. We find that an increase of the most educated labour force does not per se lead to a higher growth rate. The impact of human capital becomes positive and statistically significant only if graduates with occupations requiring tertiary education are considered in the model. In addition, the econometric outcome implies that the extent of skill mismatch is determined primarily by the rising female population with college education and the gross value added of trade‐related services.     

INTERACTIONS BETWEEN WELFARE CASELOADS AND LOCAL LABOR MARKETS

This paper provides an empirical examination of interactions between welfare caseloads and local labor markets using data on caseload stocks, entries, and exits. Granger‐causality tests show that unemployment rates Granger‐cause caseload activity but caseload activity does not Granger‐cause unemployment rates. The results also reveal differential dynamics between caseloads and labor market conditions for rural versus metropolitan markets. Several models of one‐way association between caseload activity and unemployment rates are presented. The results show that higher unemployment rates are positively associated with welfare caseloads and entries and negatively related to exits. (JEL I38, R23)     

THE BURDEN OF DEBT: AN EXPLORATION OF INTEREST RATE BEHAVIOR IN LATIN AMERICA

What determines the behavior of interest rates in Latin America? Is the recent sharp reduction in rates in the region just a transitory deviation from much higher long‐term rates? To answer these questions, this study raises two main hypotheses. First, external debt plays a central role in the sustainable behavior of domestic interest rates because it explains country risk. Second, country risk provides valuable information for predicting the behavior of exchange rate risk and not the other way around. Econometric tests confirm these hypotheses and lead to an important conclusion: unless important reforms (leading, for example, to improved tax‐collecting capacity or deeper local financial markets) are undertaken, highly indebted/high country‐risk economies will tend to be associated with high domestic interest rates. (JEL E43, G15, 016, 054, C22)     

NONPARAMETRIC BOOTSTRAP PROCEDURES FOR PREDICTIVE INFERENCE BASED ON RECURSIVE ESTIMATION SCHEMES*

We introduce block bootstrap techniques that are (first order) valid in recursive estimation frameworks. Thereafter, we present two examples where predictive accuracy tests are made operational using our new bootstrap procedures. In one application, we outline a consistent test for out‐of‐sample nonlinear Granger causality, and in the other we outline a test for selecting among multiple alternative forecasting models, all of which are possibly misspecified. In a Monte Carlo investigation, we compare the finite sample properties of our block bootstrap procedures with the parametric bootstrap due to Kilian (Journal of Applied Econometrics 14 (1999), 491–510), within the context of encompassing and predictive accuracy tests. In the empirical illustration, it is found that unemployment has nonlinear marginal predictive content for inflation.     

THE EFFECT OF WELFARE ASSET RULES ON AUTO OWNERSHIP,EMPLOYMENT, AND WELFARE PARTICIPATION: A LONGITUDINAL ANALYSIS

This paper investigates how asset tests for welfare eligibility affect auto ownership, employment, and welfare participation for single mothers without a college degree. We combine longitudinal data from the 1996 Survey of Income and Program Participation with data on state‐level welfare program rules from the Urban Institute and data on state‐level controls to test whether these single mothers were more likely to (1) own a car, (2) be employed, and (3) be off of welfare, depending on the welfare asset rules instituted in their state. We find evidence that, taken as a group, the asset rules have a statistically significant effect on the probability of car ownership. Ordinary least squares results and cross‐sectional two‐stage least squares (2SLS) results using the asset rules to instrument for car ownership show a large, positive, statistically significant effect of car ownership on employment. However, in 2SLS models controlling for prior car ownership and prior employment, the asset instruments are weaker and we do not find an effect of car ownership on employment. Of significance for policy makers, we find that the asset rules do not have a statistically significant joint effect on welfare participation, even after addressing possible endogeneity. (JEL I38, J68, J08)     

TESTING FOR A UNIT ROOT AGAINST TRANSITIONAL AUTOREGRESSIVE MODELS

This article develops a novel test for a unit root in general transitional autoregressive models, which is based on the infimum of t‐ratios for the coefficient of a parametrized transition function. Our test allows for very flexible specifications of the transition function and short‐run dynamics and is significantly more powerful than all the other existing tests. Moreover, we develop a large sample theory general enough to deal with randomly drifting parameter spaces, which is essential to properly test for a unit root against stationary transitional models. An empirical application of our test to the exchange rate data is also provided.     

Testing for the Presence of a Tremble in Economic Experiments

The classical trinity of tests is used to check for the presence of a tremble in economic experiments in which the response variable is binary. A tremble is said to occur when an agent makes a decision completely at random, without regard to the values taken by the explanatory variables. The properties of the tests are discussed, and an extension of the methodology is used to test for the presence of a tremble in binary panel data from a well-known economic experiment.     

Can Small Class Policy Close the Gap? An Empirical Analysis of Class Size Effects in Japan

Can smaller classes lead to better educational outcomes and greater equality in achievement? We estimate the causal effects of class size on achievement tests by using discontinuous changes in class size under the Japanese public compulsory education system. We employ a value‐added model that uses achievement tests conducted at two different times during the same school year. Our results show that a reduction in class size has significantly positive effects on Japanese language test scores in the sixth grade, especially at schools in wealthy areas. However, we find no evidence that a universal small class policy closes the achievement gap among schools.     

Evidence on the Convergence of Per Capita Income: A Comparison of Founder Members of the Association of South East Asian Nations and the South Asian Association of Regional Cooperation

This paper investigates the per capita income convergence patterns of a set of Association of South East Asian Nations (ASEAN) and South Asian Association of Regional Cooperation (SAARC) countries. We obtained a time‐series analysis for stochastic convergence by applying unit‐root tests in the presence of two endogenously‐determined structural breaks. We then supplemented the results by tests that produced evidence for β convergence. The evidence shows that the relative per capita income series of ASEAN‐5 countries were consistent with stochastic convergence and β convergence, but this was not found for SAARC‐5 countries. For the ASEAN‐5 countries, the structural breaks associated with the world oil crisis and the Asian crisis impacted heavily on the convergence/divergence process.     

Price Squeezes in a Regulatory Environment

This paper analyzes the relevance and the scope of price squeeze tests in a regulatory environment. We distinguish between three types of ex ante regulation (full, partial or no regulation) to identify regulatory squeezes, predatory squeezes, and squeezes as foreclosure. We argue that policy makers should (i) use price squeeze tests as an ex post instrument, compatible with the ex ante regulatory environment, and (ii) focus on predatory price squeezes. We propose a predatory squeeze test that respects regulatory choices that have been made previously, in contrast with the earlier proposed tests. We extend our framework to ask at which aggregation level predatory price squeeze tests ought to be applied, a much-debated issue in telecommunications.JEL Classification: L40, L51, L97Jan Bouckaert acknowledges the financial support from the University of Antwerp (NOI2003). Frank Verboven acknowledges financial support from the Belgian Science Foundation-Flanders. We are grateful to two anonymous referees and the Editor for their helpful remarks. We also thank conference and seminar participants at CEFIR (Moscow), CPB (The Hague), the University of Antwerp, and Massimo Motta, Walter Nonneman, Wilfried Pauwels, and Tomasso Valletti for very useful comments.