A count data model with endogenous household specific censoring: the number of nights to stay

In this paper a count data regression model accounting for endogenous censoring with household specific censoring thresholds is presented. The presented modelling approach is utilized in an analysis of household choice of total number of nights to spend on monthly recreational trips. The empirical study reveals that the suggested approach is feasible and that accounting for endogenous censoring gives a better fit to the data. For helpful comments on an earlier version of the paper we thank two anonymous referees.     

Virtual integration of financial markets: a dynamic correlation analysis of the creation of the Latin American Integrated Market

This article investigates the role of virtual integration of financial markets on stock market return co-movements. In May of 2011, the Chilean, Colombian and Peruvian stock markets virtually integrated their stock exchanges and central securities depositories to form the Latin American Integrated Market (MILA). We utilize the dynamic conditional correlation model proposed by Engle (2002) to identify a statistically significant positive correlation between these markets. Moreover, we find strong evidence that the creation of the MILA increased the levels of dynamic correlation between stock returns. A higher correlation was also found during the dot-com bubble and the 2007 financial crises. Our results imply a decline in gains from international diversification by holding portfolios consisting of diverse stocks of these countries.     

Patents,R&;D and lag effects: evidence from flexible methods for count panel data on manufacturing firms

This paper investigates the relationship between patents and research and development expenditures using new longitudinal patent data at the firm level for the U.S. manufacturing sector from 1982 to 1992. The paper also develops a new class of count panel data models based on series expansion of the distribution of individual effects. Estimation results from various distributed lag and dynamic multiplicative panel count data models show that the contemporaneous relationship between patenting and R&D expenditures continues to be strong, accounting for over 60% of the total R&D elasticity. The lag effects are higher than have previously been found for the 1970s data. We would like to thank Chris Bollinger, Bronwyn Hall and Paula Stephan for useful comments on the previous version of the paper. Earlier versions were presented at the 11th International Conference on Panel Data, Texas A&M University, the Midwest Econometrics Group Meeting, and the Annual Conference of the Southern Economic Association.