The information coefficient (IC) is a metric used to assess the expertise of an investment analyst and active portfolio manager. An IC The information coefficient demonstrates how well financial projections made by the analyst correspond to real economic results. The information coefficient can be in the range of 1.0 to -1.0, with a value of -1 signifying that the analyst's predictions had no relationship to the actual outcomes and a value of 1 indicating that the analyst's predictions had exactly matched the financial results.

Information Coefficient (IC) Explained

The information coefficient, or IC, frequently used to evaluate the value of a financial analyst, describes the relationship between expected and actual stock returns. A precise linear relationship between expected and actual returns is indicated by an information coefficient of +1.0, whereas a linear relationship is not indicated by an information coefficient of 0.0. The information coefficient of -1.0 shows that the analyst never predicts the future correctly.

A forecasting financial analyst has an information coefficient score of close to +1.0. However, the probability of getting the prediction right is 50/50 if the definition of "accurate" is that the analyst's projection matched the trend (up or down) of actual financial returns. Therefore, even the most incompetent analyst could be anticipated to have an information coefficient of about 0, which denotes that half of the predictions were accurate and half were incorrect. An IC score near 0 indicates that the analyst's predictions are less accurate than outcomes that could be obtained by chance, indicating that ICs close to -1 are uncommon.

The information ratio (IR) differs from the information coefficient (IC). The IR measures an investment manager's skill, comparing the excess returns of a manager to the level of risk assumed.

The Fundamental Law of Active Asset Management indicates that an investment manager's effectiveness (IR) depends on their breadth and skill level (IC) or how frequently they apply it. This law includes both the information coefficient (IC) and information ratio (IR).

How Does Information Coefficient (IC) Work

The information coefficient (IC) is a popular tool for evaluating the stock-picking abilities of investment managers. The correlation between the expected and actual stock returns that the investment managers took into account during the asset management process is described statistically by this indicator. Since stock selection models, also known as "alpha models," are generally employed to anticipate returns, IC is frequently used to assess the efficacy of these models. The popular thinking holds that the more skilled the model developer or financial managers, and the best performing the fundamental stock selection model, the greater the IC. However, the metric's suitability and efficiency for assessing stock selection methods have yet to be thoroughly explored or examined.

However, the textbook definition of the information coefficient implies that the analyst uses the complete range of the correlation distribution to evaluate model performance and that a brilliant stock picker would have an information coefficient of 1. In contrast, a bad picker may end up with an information coefficient of -1. On the other hand, it is widely acknowledged that a practical stock selection model could barely have an IC that is considerably different from 0 due to the fundamental problems of projecting stock returns. For example, a stellar model may frequently have an information coefficient of 0.05 or 0.1 or otherwise "slightly" above zero; the IC need not be signed in the negative to become a very poor model.

Information Coefficient Formula

IC = (2 x Proportion Correct) - 1

Where proportion correct indicates the proportion of a prediction made correctly by the financial or investment analyst

Information Coefficient Example

For example, if a financial analyst made two forecasts and correctly identified two of them, the information coefficient (IC) would be as follows:

IC = (2×1.0) − 1 = +1.0

​

If an investment analyst's forecasts were only accurate 50% of the time, then:

IC = (2 x 0.5) - 1 = 0.0

However, if none of the predictions came true:

IC = (2×0.0) −1 = −1.0

Limitations of using Information Coefficient (IC)

There are several limitations when using information coefficient (IC) to assess stock selection models because of the divergence between the information coefficient distribution and its reality in the financial world.

First, it becomes more challenging to perform hypothesis testing on the data in a model since real information coefficients are rarely substantially different from zero. Weak information (materiality) is claimed to be present in a model with an information coefficient = 0.01 but may not be readily discernible (statistics). On the other hand, a relatively modest but positive IC, like 0.01, may not always imply the model contains actual information.

Second, a different paradigm is needed to examine small real information coefficient correlations. An investment analyst must devise additional criteria to distinguish between relevant models and "placebos" because "weak" correlations are frequently seen in the stock selection models.

Finally, ICs are typically a random variable because stock selection models have always been evaluated against unexpected returns. Thus, the behavior of this small, random variable over time ultimately determines the effectiveness of a stock selection model and the profitability of the investments backed by the model.