Evaluating the information coefficient
The Information Coefficient (IC) is a fundamental metric in quantitative portfolio construction. It gauges the predictive power of a forecast relative to future returns. At its core, IC uses the Spearman rank correlation, a non-parametric measure that assesses how well the relationship between two variables can be described using a monotonic function. The IC’s value ranges from -1 to 1 with a positive IC indicating a forecast’s genuine predictive power, a value near zero indicating an absence of predictive capacity, and a negative value indicating an inverse relationship between the forecast and subsequent returns.
The origins of the IC can be traced back to the 1960s and 1970s. The preliminary idea of the metric was introduced by Jack L. Treynor in the early 1960s as he presented the topic of correlating investment decisions with corresponding outcomes in his writings on performance measurement. By the mid-1970s, Fischer Black...
