Speaker: Frank Partnoy, UC Berkeley
ABSTRACT: We describe two problems – omitted variable bias and measurement error – that arise when a ratio is the dependent variable in a linear regression. First, we show how bias can arise from the omission of two variables based on a ratio’s denominator, and we describe tests for the degree of bias. As an example, we show that the familiar “inverse U” relationship between managerial ownership and Tobin’s Q is reversed when omitted variables are included. Second, we show how measurement error in the ratio denominator can lead to bias. We urge caution about using ratios as dependent variables.