Browsing by Author "Nolan, James, P."

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Parametric and Non-Parametric Tests for Scale Economies in a Regulated Industry: The Case of Cable Television
    Law, Stephen, M.; Nolan, James, P.
    We examine a situation where parametric and non-parametric approaches to the study of production and optimal scale can be used as complements, rather than substitutes. We illustrate this concept with data from the cable television industry in an evaluation of the timeliness of deregulation. Although we begin with a large sample offering adequate degrees of freedom for parametric estimation, important regulatory policy issues and the structure of the industry lead us to consider parameter estimation over sub-samples. Some sub-samples are small enough that parametric models cannot guarantee reliable estimates. To deal with this problem, we estimate production parameters non-parametric ally using various forms of data envelopment analysis (DEA). Since it is not statistical in nature, the use of DEA is not constrained by degrees of freedom. Not only do the non-parametric estimates shed light on important characteristics of the industry sub-samples when considered in isolation, we also find that on aggregate they agree with parametric estimates when these can be computed.
  • Thumbnail Image
    Item
    Parametric and Non-Parametric Tests for Scale Economies in a Regulated Industry: The Case of Cable Television
    Law, Stephen, M.; Nolan, James, P.
    We examine a situation where parametric and non-parametric approaches to the study of production and optimal scale can be used as complements, rather than substitutes. We illustrate this concept with data from the cable television industry in an evaluation of the timeliness of deregulation. Although we begin with a large sample offering adequate degrees of freedom for parametric estimation, important regulatory policy issues and the structure of the industry lead us to consider parameter estimation over sub-samples. Some sub-samples are small enough that parametric models cannot guarantee reliable estimates. To deal with this problem, we estimate production parameters non-parametric ally using various forms of data envelopment analysis (DEA). Since it is not statistical in nature, the use of DEA is not constrained by degrees of freedom. Not only do the non-parametric estimates shed light on important characteristics of the industry sub-samples when considered in isolation, we also find that on aggregate they agree with parametric estimates when these can be computed.