Financial Bubbles: Two Identification Methods Applied in Colombia

Abstract

In the present article two different econometric approaches to date multiple bubblesare explored. The methodologies are applied to the prices of three important financialvariables in the Colombian economy: the exchange rate, the stock market index (IGBC)and the oil price WTI. Both approaches are based on the analysis of the integration orderof the series involved. The first one uses a robust method to detect random walks, basedon a sign test. The second one is built using variance ratios, frequently employed in theliterature to test market weak efficiency. The measures are estimated dynamically andtests of power and size are provided. Evidence in favor of bubbles is found in two seriesout of three, for recent periods of time.