Challenges in Developing Capital Market Forecasts

Inappropriate capital market expectations could be devastating to portfolio performance. Several pitfalls could lead to false expectations and a less-than-optimal asset allocation. Regardless, the analyst is tasked with following a disciplined process and understanding potential challenges in developing capital market expectations.

The major challenges to developing capital market expectations are as follows:

• Limitations to using economic data: Acceptable economic data includes data from reputable sources such as the International Monetary Fund and the US Federal Reserve. The first limitation of using economic data has to do with the lag time between the collection and distribution of the data. The second limitation has to do with definitions in methodology changing over time. In addition, rebasing can change the way indexes are calculated. For example, what might be included in the consumer price index has changed over the course of its existence.

• Data measurement errors and biases include the following: 

  • Transcription errors emanate from the incorrect recording of data.
  • Survivorship bias occurs when low-performing managers, funds, or securities are removed from a data set after becoming defunct. This inadvertently biases the returns of the index upward.
  • The use of appraisal data, also known as smoothed data, refers to data or assets that are infrequently valued. When creating a time-series analysis, the lack of data points can cause the standard deviation to be underestimated.

• Limitations of historical estimates: Regime changes, or marked deviations in the structure and function of markets or complex systems, means analysts must remain aware of data being used and judge its merit based on their knowledge and experience. An example of a regime change could be the tech bubble of the early 2000s. After going sky-high, valuations came back to earth, and those levels of risk and return should not be expected to continue indefinitely. Regime changes lead to non-stationarity or the idea that one time period is not statistically similar to another.

• Using ex-post risk and return: The use of ex-post (historical) data to justify ex-ante (future) metrics is essentially anchored upon the assumption that the past always repeats itself without fail. This is an inherently false assumption.

• Non-repeating data patterns: False relationships may appear to be driving capital market expectations. Analysts must understand what drives the movements to spot spurious relationships. In fact, spurious relationships are expected to be present in about 5% of all trials and can appear for an almost unlimited number of reasons (using a 95% confidence interval).

• Failing to account for conditioning: Closely related to non-repeating data patterns, data conditioning refers to a lack of proper interpretation of data, which may join inappropriate assumptions together. For example, using the average market boom and market bust returns for a particular firm could lead to faulty projections, given that a boom or boost could be reasonably expected to occur.

• Misinterpretation of correlations: This involves mislabeling the independent and dependent variables. An example could be the amount of corn grown and rainfall in a local region. While the two are certainly related variables, it would be incorrect to think the amount of corn grown would influence the amount of rainfall. Rainfall is the independent variable, and corn is the dependent variable.

• Psychological traps: Include both emotional and cognitive biases. Many of the following biases have previously been detailed in the reading on behavioral biases.

  Anchoring bias: The first information received is over-weighted, and updates are slow to change from their original form.

  Confirmation bias: Only information reinforcing the existing belief is considered, and such evidence may be actively sought while other evidence is ignored.

  Status quo bias: Predictions are highly influenced by the recent past. At the same time, unwillingness to re-work an approach leads to an erroneous extrapolation of the past into the future.

  Overconfidence bias: This leads to ignorance of past mistakes, and the resultant accuracy of the forecast is overestimated.

  Availability bias: What is easiest to remember is over-weighted.

  Prudence bias: Forecasts are overly conservative to avoid making mistakes.

• Model or input uncertainty: This is a scenario in which an analyst does not know the tool to use for a particular forecast. Input uncertainty refers to the analyst not knowing what data to put in a model. An example could include an analyst being unable to decide to either use a 2-year or a 10-year risk-free rate for a particular discounted cash flow model.

 

Question

The use of appraisal data most likely:

1. Overestimates Sharpe ratios.
2. Underestimates Sharpe ratios.
3. Does not apply to Sharpe ratios.

Solution

The correct answer is A.

Using appraisal data results in an infrequently recorded data set or less than the case under real-time asset pricing (as in the equities markets). This lack of data points produces underestimated standard deviations and variances, as not all of the data’s fluctuations will be captured. Rather than reflect every change in the data, a certain periodic amount is assumed to arrive at the current asset price. Underestimated standard deviation results in an overestimated Sharpe ratio, as standard deviation is the denominator in the formula.

B is incorrect. The use of appraisal data is most likely to result in overestimated Sharpe ratios, not underestimated Sharpe ratios. The Sharpe ratio measures risk-adjusted returns, which compares a fund’s historical or projected returns relative to an investment benchmark with the historical or expected variability of such returns. Using appraisal data can lead to an overestimation of the Sharpe ratio because it can result in an overestimation of the expected return and an underestimation of the risk.

C is incorrect.  The use of appraisal data applies to Sharpe ratios. The Sharpe ratio is a widely used method for measuring risk-adjusted relative returns, and the use of appraisal data can affect the calculation of the Sharpe ratio by influencing the expected return and risk estimates used in the calculation. Therefore, the use of appraisal data applies to Sharpe ratios.

Asset Allocation: Learning Module 1: Capital Market Expectations – Part 1 (Framework and Macro Considerations); Los 1 (b) Discuss challenges in developing capital market forecasts