Using Pi Lambda and the Bonferroni Inequality to Extend Time Series Tests to Panel Data

In the financial field much of the data available is in panel form. Nevertheless, some researchers convert these data into indices in order to facilitate its use. Today, because of the availability of software programs, it is easier than in the past to handle panel data in those instances where its use is preferable to the use of indices. The objective of this paper is to analyze the long run equilibrium relationship between prices and fundamentals while proposing a very simple method of extending time series models to panel data. Although this method has multiple applications, here we are solely concerned with investigating the cointegrating relationship between prices and fundamentals with data expressed in panel form. This method has several characteristics that make it appealing. First, it is simple to implement (efficient). Second, it is general in scope (can be applied to many tests). Third, it takes into account arbitrary correlations (for example, among firms). Fourth, it does not require making unrealistic assumptions (such as the assumption of independence among firms within the same market). Our results are supportive of Han’s (1996) in that we do not find cointegration between fundamentals and prices.