Symmetric Buy-Sell Decision Process
Apply the same research discipline to exits that you apply to entries
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 5
- Confidence
- 97%
The symmetric buy-sell process begins by testing whether a manager's sales add value against simple counterfactuals, such as selling another randomly selected holding. If sales underperform, the team treats this as a process defect rather than assuming that awareness will fix it. Every proposed exit receives the same research depth, documented thesis, evidence standards, and review applied to a purchase. Independent reviewers, cooling-off periods, or mechanical decision aids can interrupt loss aversion, attention effects, and the desire to realize emotionally satisfying gains. The process closes with an audit comparing sold securities with retained alternatives. Its central mechanism is structural humility: professionals accept that expertise can be asymmetric and install guardrails where evidence shows their judgment is weakest.
Origin
Alex Imas described research showing that institutional investors added value when buying but lost substantial value when selling, even relative to random sales from their own portfolios.
Core principles
- 01Expertise in buying does not guarantee expertise in selling.
- 02Selling can activate laboratory-documented biases even among professionals.
- 03Awareness alone does not override psychology.
- 04Humility is a prerequisite for adopting effective decision aids.
How to run it
- 1
Build a counterfactual benchmark
Compare each historical sale with credible alternatives, including retaining the asset or selling another portfolio holding.
Pro tip Use several counterfactual rules so the conclusion does not depend on one benchmark.
Watch out Do not compare sales only with a broad index if the relevant choice was between existing holdings.
- 2
Diagnose asymmetric discipline
Map the research, documentation, and review used for purchases and sales. Identify where the selling process becomes more intuitive or hurried.
Pro tip Interview decision-makers before showing performance results to reduce hindsight rationalization.
Watch out A sophisticated title or large portfolio does not eliminate behavioral bias.
- 3
Standardize the sell thesis
Require every proposed sale to state why the original thesis changed, what evidence supports the exit, and what would invalidate the decision.
Pro tip Use the same analytical template and data quality thresholds as the buy decision.
Watch out Do not let the purchase price become an economically irrelevant reference point.
- 4
Add an external interruption
Use an independent reviewer, committee member, algorithmic screen, or cooling-off period to challenge the proposed sale.
Pro tip Assign a reviewer to argue for retaining the asset or selling an alternative.
Watch out A review that merely confirms the manager's preference will not counter the bias blind spot.
- 5
Audit outcomes
Track sold assets and relevant alternatives after each decision. Review patterns periodically and refine the guardrails when biases recur.
Pro tip Evaluate process quality as well as short-term price movement.
Watch out One lucky or unlucky outcome does not establish whether the process was sound.
In the wild
A fund discovers that its purchases outperform random alternatives while its sales underperform randomly selected sales from the same portfolio. It introduces a sell memo matching its buy memo, assigns an independent reviewer, and compares each sale with retaining the asset and selling another holding.
→ Exit decisions become more evidence-based and less exposed to loss aversion and attention-driven judgment.
Common mistakes
Relying on awareness
Knowing that selling biases exist does not stop them; the process needs an external aid or guardrail.
Assuming expertise is symmetric
A manager can demonstrate genuine skill in selecting purchases while still destroying value through sales.
Protecting professional identity
The bias blind spot encourages experts to see documented errors as problems affecting other people rather than themselves.
Is it for you?
Best for
It is best for portfolio managers whose purchase process is rigorous but whose selling process is comparatively informal.
Not ideal for
It is not ideal for purely mechanical strategies that already execute symmetric entry and exit rules without discretion.
From the transcript
“let's throw a dart at your portfolio and sell that instead”
“on the selling side we found literally the same biases that we found in the lab.”
“use the same sort of research technology for my selling as I'm doing for my buying.”
From the episode
How Investors Fall Into Bias Traps with Economists Richard Thaler & Alex Imas
Alex Imas