Part 1 of this article looked at general biases in decision-making, Part 2 set forth biases that can affect perceptions of market and investment dynamics and Part 3 discussed biases that can be found in the process of investment selection and valuation. This final Part 4 of the article will discuss ways to reduce investment decision-making bias through the creation of decision-making algorithms.

Dynamic Decision-Making Algorithms

One way to reduce investment decision-making bias is by building decision-making algorithms. An algorithm is a set of rules or procedures to be followed when filtering information, making calculations or solving problems. Although when we think of algorithms we often think of computational operations, they can also be used by companies, teams or even individuals to review strategic and practical options and make choices.

Decision-making algorithms can:

● significantly reduce decision-making steps and time, which can allow corporate resources to be spent on other activities or decision execution steps;

● lead to breaking decision components into different parts and having the people most qualified to analyze those parts review them;

● reduce subjectivity in decision-making, which can significantly improve decision-making quality and results; and

● increase the likelihood that similar types of situations will be addressed in the same way, allowing for more efficient use of historical firm knowledge and back testing of decision-making rules and assumptions.

Decision-making algorithms in the investment context have to be carefully designed given the complexity and interrelationships of investment decision drivers. Investment decisions generally require:

● careful analysis of the objectives of one and often multiple investment stakeholders which are rarely either identical or fixed for long periods of time;

● the input of persons and often departments in a company or investment firm with different perspectives and organizational objectives;

● not only making a decision based on a single set of facts but rather on a process of continually gathering facts, questioning the accuracy of those facts and seeking out additional facts;

● putting evolving factual knowledge into a broader and shifting business and market context;

● comparing the relative weight of different investment factors.

Determining What Decisions Should be Taken

Starting with the premise that decisions tend to be inherently biased, the first step in building a decision-making algorithm is to reduce as much as possible the number of decisions to be taken in connection with an action path. If a large company began every day trying to figure out what each person should be doing for the next eight hours it would be very difficult for the company to run effectively.

Ensuring an appropriate balance between deliberation and action is also very important with respect to investment strategy.  If an investor spent every moment trying to determine what their target return should be, what country and sector they should invest in and what type of vehicle would be best it would very difficult to give sufficient attention to the specific issues related to an investment that often determine investment success or failure. There is a point at which the expenditure of decision-making resources can surpass decision-making value.

It is not possible and would not be wise from an analytical or risk management perspective to reduce many investment choices to a single one dimensional yes or no decision, so a target should be to identify for any action path the minimum number of decisions necessary (MDN) to go from an initial point of consideration to a final decision. 

This allows for appropriate allocation of corporate or individual resources along the decision path and focusing the maximum amount of decision-making energy on the most crucial decision drivers.  100 hours of focus on a valuation analysis may turn out to be much more useful than spending 25 hours on valuation and 75 hours on market or competitive analysis.

For this to be accomplished in the investment context, it is of course necessary to put a significant amount of thought and effort into designing the investment strategy and setting up mechanisms to ensure that a reasonable number of investment opportunities will be able to be reviewed.  Without a review of a sufficient number of potential deals, it is very difficult to have confidence that the opportunities that are being reviewed represent either a fair reflection of the opportunities that exist in the market or the least amount of risk for the targeted return.

Key Decision-Making Levels

Once a general investment strategy is established, the next step is to set up a process for moving investment information through a decision-making framework.

Similar to trying to design an investment analysis process to limit the decision-making what, it also is useful to narrow the who; if every person in an organization participated equally in every investment decision it would not be efficient.  This because every person in an organization is not equally qualified to opine on every firm matter and each part of a decision problem does not require the attention of persons who are competent to opine on those matters.

How decisions are structured depends on the size, organization and focus of the firm.  Generally speaking, however, for a private equity investment, decision-making levels could be structured as follows:

● Level 1.  This level could involve an initial screening to determine if potential investment opportunities fit with the agreed parameters set forth in the general investment strategy.  In accordance with the MDN principle, investment cutoff parameters could be based, in addition to the general investment criteria set forth above, on factors such as investment amount, investment horizon, investment structure and amount of proposed co-investment by the investee company.  Opportunities that met with Level 1 criteria would pass to Level 2.

● Level 2: This level could involve a detailed analysis of investment opportunities that passed the Level 1 threshold to see if they warrant further firm and/or external analysis. This can be a very time consuming process as it involves reviewing often extensive investment information, meeting with company representatives and visiting company facilities.  Opportunities that met with Level 2 criteria would pass to Level 3.

● Level 3.  This Level could involve seeking the opinions of other firm departments, such as the legal and risk management departments, regarding the proposed investment opportunity, receiving their feedback and addressing any concerns raised.  Investment opportunities that passed this level of analysis would move to Level 4.

● Level 4 involves making a final investment decision and funding. 

Decision-Making Levels and Anti-Bias Measures

From the perspective of building an a decision-making algorithm that reduces bias, as decisions move farther along the decision-making path they should be increasingly subject to tougher scrutiny. This is important because at times decision items can gather their own momentum as they travel through corporate decision-making procedures which can make it progressively easier to overlook decision weak points.

In Level 1, the principal anti-bias measure should be to ensure that there are clear investment filters and that they are systematically applied by the people screening investment opportunities.

Level 2 involves analysis of potential investment opportunities that meet Level 1 screening thresholds.  At this level, there should be a considerable amount of criticism of the reasoning behind proposed investment positions to identify potential biases or weaknesses in investment review.

This critical process can involve:

● checking the investment opportunity against past firm experience to see if there is information in the firm’s data base which can be useful in pointing out strengths or risks related to the opportunity that may not immediately be apparent. 

● setting forth arguments for advancing or rejecting an investment opportunity.  This analysis should set forth drivers as well as pros and cons of an investment view so other team members can analyze whether or not these points can be fairly analyzed by the investment team or if outside opinions are required.

For example if a principal driver of an investment view is that “China is a great market” but there are no people on the team who understand the Chinese market well, this should be a very strong signal that the firm’s ability to review the opportunity has significant limitations.  Similarly, if the opportunity involves a technical innovation but the team is not in the position to analyze the technical merits or compare them with other technologies in the market again this should be an important yellow flag in investment review.

Level 3.  In this level, the key anti-bias measure is to have investment opportunities reviewed by other firm departments.  While the number of departments can depend on the investment strategy, the nature of the investment opportunity and the size of the firm, this can involve review by legal, compliance and risk management.  Ideally this review would be carried out by people whose compensation is not directly based on whether investments are approved or rejected.  

Level 4.  In this level, the key anti-bias measure is to seek opinions of third-parties regarding technical issues such as specialized legal or tax opinions regarding investment documentation or other matters.

Converting Decision-Making Results into Firm Assets

The results of each step of the decision-making process, even with respect to decisions that have been rejected, are important firm intellectual assets that can be used to strengthen decision-making in the future.

Regarding Level 1, careful records should be kept of how many opportunities were reviewed, how many passed to Level 2 and the reasons why opportunities that did not pass to Level 2 were rejected.  This can be very helpful in analyzing if investment intake procedures should be broadened, narrowed or restructured.

With respect to Level 2, an analysis of decisions that reached later levels and ultimately did not lead to expected investment results or investment opportunities that were rejected can help understand, among other things, if a firm’s investment strategy is appropriately focused or if different resources are required.

Regarding Level 3, if the same types of risk, legal or compliance issues are systematically causing deals to be blocked, this can lead to firm-level discussion as to the criteria that other departments are applying to proposed investment decisions and their own screening mechanisms.  In the event that there are sound reasons for deals being rejected, the investment team should work to change the types of deals that are being forwarded to other departments for Level 3 review.

Finally, with respect to Level 4, the reasons for approving or rejecting decisions should be documented and recorded.  For investments that were unsuccessful, it is important to try to analyze as much as possible the reasons why a decision failed.  Was it because of a lack of information?  Was it because that people reviewing the information did not have sufficient expertise?  Was it because of the business team?  What is because of market conditions that could reasonably have been predicted?

This entire review approach should not be carried out with a view to blame anyone for a poor decision that may have been made but rather to analyze the efficiency and effectiveness of a decision-making process and to try to improve it.  Good decisions often come from good decision-making processes.

Conclusion

One way to reduce this threat of bias in decision-making is to build a decision-making algorithm. In addition to reducing bias with regard to specific decisions it also leads to the creation of a better decision data base that can help strengthen investment analysis in the future.

The photo for this article was taken from Unsplash.  The photographer is David Clode.