Likelihood Ratio Tests for Cointegration, Cobreaking and Cotrending

Project Details

Project Lead
Katherine Holcomb 
Project Manager
Katherine Holcomb 
Institution
University of Virginia, UVACSE  
Discipline
Economics (903) 

Abstract

Proposed a set of likelihood ratio tests for cointegration, cobreaking, and cotrending. Most existing tests are designed for only one of these common features with making strong assumptions on the others. Tests deal with all three common features at the same time in a unified framework. Consequently, tests control the size over a bigger parameter space while maintaining proper power.

Intellectual Merit

We propose a set of likelihood ratio tests for cointegration, cobreaking, and cotrending. Most existing tests are designed for only one of these common features with making strong assumptions on the others. Our tests deal with all three common features at the same time in a unified framework.

Broader Impacts

Tests control the size over a bigger parameter space while maintaining proper power.

Scale of Use

Thousands of cores for periods of days.