Historical Inheritance Forecast
Predict a disruptor's risks by tracing the problems it inherits
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 6
- Confidence
- 95%
Feldman's forecasting method starts from the claim that people use the past as their first guide to what happens next. For a disruptive product or institution, identify the incumbent activity it replaces and examine the major crises, exclusions, or coordination failures in that activity's history. Then isolate the mechanism behind each problem and ask whether the new arrangement actually removes it or merely changes the interface. If the same people, incentives, or power relations remain, the problem is likely to be inherited. That prediction becomes actionable: protections can be designed before the failure scales. The method does not assume that history repeats exactly or guarantees an outcome. Instead, it uses historical mechanisms as risk hypotheses, especially where a technology claims to disrupt an industry but continues to perform the same underlying social function.
Origin
Feldman illustrated the method by showing how Airbnb inherited the hotel industry's long history of racial discrimination.
Core principles
- 01Disruption changes arrangements without erasing the industry's social problems
- 02Past crises reveal recurring mechanisms
- 03The right historical comparison produces actionable safeguards
- 04A new interface can inherit an old industry's failures
How to run it
- 1
Define the disrupted activity
Name the incumbent industry or institution whose function the new arrangement is taking over.
Pro tip Describe the underlying activity, not only the old product format.
- 2
Find the historical failure
Study the incumbent's most consequential past challenges, crises, and exclusions.
Pro tip Prioritize failures tied to the core transaction or relationship.
Watch out Recent market data alone may hide long-running social problems.
- 3
Extract the mechanism
Identify the incentives, human judgments, power relations, or coordination conditions that produced the failure.
Watch out Copying a historical label without its mechanism produces a weak analogy.
- 4
Test for inheritance
Ask whether the disruptor retains the mechanism even though its technology or interface is new.
Pro tip Trace who makes the equivalent decision in the new system.
Watch out Do not assume technological novelty removes human bias.
- 5
Forecast the recurrence
State where and how the inherited failure is likely to emerge in the new arrangement.
Pro tip Frame the prediction as a testable risk rather than a certainty.
- 6
Build a protection
Add a product, policy, or governance safeguard aimed at the inherited mechanism before it scales.
Pro tip Measure whether the safeguard changes the predicted failure mode.
In the wild
Airbnb set out to disrupt hotels. Feldman proposes looking at the hotel industry's twentieth-century history, where racial segregation was a central challenge. Because hosts and guests could still discriminate, the social problem did not disappear when lodging moved to a platform. That history could have prompted protections against racial discrimination at design time.
→ A historical industry failure becomes a specific, actionable platform-risk prediction.
A hiring platform studies the employment sector's history of exclusion, isolates discretionary screening and biased proxies as mechanisms, and checks whether its training data and employer controls preserve them. It predicts unequal candidate filtering and introduces audits and constrained decision rules before broad rollout.
→ The team treats inherited discrimination as a design risk rather than assuming automation eliminates it.
Common mistakes
Treating novelty as a clean slate
A new technical layer may continue the same activity with the same people and inherited social risks.
Using surface analogies
A forecast is only useful when it identifies the mechanism that can recur, not merely a resemblance between eras.
Is it for you?
Best for
It is best for teams replacing an established industry, institution, or transaction model with a new one.
Not ideal for
It is not ideal when the proposed activity has no meaningful historical analogue or inherited participants.
From the transcript
“the past is our first guide to figuring out what's happened”
“the social challenge that had faced that industry wasn't going to disappear when it passed into this new industry”
“looking at the past can actually help you make predictions about what the challenges are going to be in the future”
From the episode
#540: Harvard Polymath Noah Feldman — Deep Focus for Hyper-Productivity, Learning 10+ Languages, Predicting the Future with History, the Possibilities (and Limitations) of DAOs, Lessons from the Iraq Invasion, Designing the Supreme Court of Facebook, the Virtue of Scholarship, and the Wild Life of Oliver Wendell Holmes, Jr.
Noah Feldman