Debiasing decisions - Some simple rules to follow

We all face biases or bring them to decision-making, so it is critical to develop strategies to reduce the risk of bias. The bias problem applies to both quantitative and discretionary decisions. Just because you are using a model does ot mean that you are immune to biases. There should be a checklist to review the potential impact of biased thinking. The paper “A user’s guide to debiasing” does a good job of exploring the basics of reducing biases and improving the quality of judgment in decision-making. Debiasing reduces logical inconsistencies and misperceptions or misjudgments of reality. Debiasing is not the same as gathering more factual informaiton. The objective is to improve decisions given a specific set of information.

Categorizing debiasing methods distinguishes between the person and the task. Improving the person requires training to help the decision-maker overcome their limitations. The second approach is to modify the environment to match the thinking required.

Many of the sources of our biases come from the confusion between system 1, fast and automatic responses, and system 2, which requires slower and more deliberate thinking. It is important to distinguish between narrow thinking and shallow thinking. Narrow thinking focuses attention on a single category of objectives and crowds out the ability to identify other alternative objectives. Shallow thinking devotes too little effort to the required task. For any decision, there has to be a level of readiness to perform better decision-making. 

The person can be modified through better education that generates more alternatives, tempers optimism, focuses on improving judgmental accuracy, and assesses uncertainty. Decision-makers can use defaults to get closer to making better decisions through regular processing, nudges to induce reflection through prompts or planned interruptions, and the formation of a set of active choices. If this can be done for the individual, a similar process can be applied to the organization. 

AI - all the time for business

 

You have to love some of these charts that bundle ideas that seem to be unrelated. In this case, we find that the word "AI" is used more often on earnings calss than the word "earnings". We should expect that the word earnings is fairly static on earnings calls, but the explosion of the word AI suggests that this is the what management and investors have as their main focus. For management, using the the word AI convesy what they are doing new to address issues of tehcnology and effciency. Investors want to know how firms are using new technology. The question is whether AI will actually lead to the efficiency gains that many expect.

Decision focused learning and portfolio selection

One of the more interesting papers on portfolio management links prediction with optimization. Rather than a two-step process, the authors focus on how optimization should be managed alongside prediction. The paper, “Return Prediction for Mean-Variance Portfolio Selection: How Decision-Focused Learning Shapes Forecasting Models”, provides insights on how decision-focused learning (DFL) can be used to improve overall portfolio returns. 

The usual process for building a portfolio is through mean-variance optimization. This process is two-staged. It is a predict-then-optimize method. In the first stage, a set of expected returns is generated, and in the second stage, the optimization selects the set of assets that maximizes return subject to a set of constraints. The problem with MVO has been studied extensively. The issue is that if the expected returns are poorly defined, the MVO will choose the “best" returns, yet the portfolio may be optimized on the forecast errors. The classic answer from Markowitz is that estimating expected returns is the investor’s job, not the optimizer’s.

The DFL framework will integrate the prediction and optimization to improve the outcomes. The issue is whether the MSE of forecasts for each asset is treated independently and equally or integrated with asset correlations. With DFL, the optimization accounts for prediction errors when finding the weights via a loss or regret function.

It is found that DFL identifies fewer assets than a standard MVO and exhibits a bias toward positively returning assets, given the optimization for a long-only portfolio. Still, it offers a better way to optimize a portfolio.

Pric e action by Rhetoric: Trump and Oil

 

The FT chart provides an interesting insight into the drivers of the current oil market. Of course, oil is driven by the events of the Iran War, yet the impact on oil prices is more subtle. There seems to be a direct connection between President Trump's comments and the surge or decline in oil prices.  When Trump makes comments that escalate tensions, usually before or on a weekend, there is a surge in prices, followed by comments that de-escalate tensions. The uncertainty in prices is associated with rhetoric rather than specific action surrounding Iran. The sample is small yet there may be a link between comments and oil price action.