A simple wavelet-based test for serial correlation in panel data models
Working paper

View/ Open
Date
2014-03Metadata
Show full item recordCollections
- Discussion papers (FOR) [575]
Abstract
Hong and Kao (2004) proposed a panel data test for serial correlation of unknown
form. However, their test is computationally difficult to implement, and simulation studies
show the test to have bad small-sample properties. We extend Gencay’s (2011) time series
test for serial correlation to the panel data case in the framework proposed by Hong and Kao
(2004). Our new test maintains the advantages of the Hong and Kao (2004) test, and it is
simpler and easier to implement. Furthermore, simulation results show that our test has
quicker convergence and hence better small-sample properties.