Chesterton's Fence: Why New HR Leaders Delete the Policies That Kept the Company Alive

60-Second Summary

G.K. Chesterton (The Thing, 1929): if you find a fence in a field and can't see why it was built, do not tear it down until you can — because someone built it for a reason.
Applies most sharply to inherited HR policies: many were written after a specific incident (a discrimination suit, a comp-leak, an accidental IP loss) and appear irrational to a leader who wasn't there.
The reformer's fallacy: assuming that an unexplained policy is an unnecessary policy. Often the policy is what has prevented the incident from recurring for so long that the incident is no longer visible.
A Fence Audit — twelve questions to ask before removing any inherited HR policy — turns Chesterton's aphorism into an operational discipline.
This does not mean don't reform. It means: reform after diagnosis, not before. The policies genuinely worth removing survive the audit; the ones that don't reveal the crisis the removal would have caused.

There is a particular energy that a new Head of People brings in month three. The listening tour is over. The inefficiencies are visible. The Notion doc titled 'Stupid Policies to Kill' is 47 items long. Some of those items genuinely should die. Some of them will kill the company if they die, in ways the new leader cannot yet see — because the incident the policy was written to prevent has not happened in six years, and the reason it hasn't happened is the policy. Chesterton wrote about this in 1929 and it has not been improved on since.

What Chesterton actually wrote

“In the matter of reforming things, as distinct from deforming them, there is one plain and simple principle... There exists in such a case a certain institution or law; let us say, for the sake of simplicity, a fence or gate erected across a road. The more modern type of reformer goes gaily up to it and says, 'I don't see the use of this; let us clear it away.' To which the more intelligent type of reformer will do well to answer: 'If you don't see the use of it, I certainly won't let you clear it away. Go away and think. Then, when you can come back and tell me that you do see the use of it, I may allow you to destroy it.'”
— G. K. Chesterton, The Thing (1929)

Chesterton was writing about tradition in general, but the passage has become a durable operating principle in engineering, product design, and now HR. The point is not conservatism. Chesterton explicitly allows reform — after diagnosis. The point is that the burden of proof for removal sits with the reformer, not with the fence, because the fence has already earned its place by existing without visible failure.

The new-CHRO's reformer fallacy

There is a predictable arc. Month 1: listening tour. Month 3: the list of stupid policies. Month 6: aggressive removal. Month 12: three incidents that would have been prevented by policies the new leader removed. Month 18: quiet re-creation, under different names, of some of the removed policies. Month 24: the new leader publishes an essay called 'What I Wish I'd Known About Culture Change.'

The fallacy has a specific structure. The new leader sees the policy without seeing the incident. The people who remember the incident are often junior enough to not be consulted, or have left. The policy looks like bureaucratic residue. It is often the crystallised memory of a crisis. Removing it does not restore neutrality — it removes the memory, and the crisis returns because the underlying failure mode has not been addressed by anything else.

The signature of a fence you're about to wrongly tear down

The policy is oddly specific. Nobody can explain why it exists. Long-tenured employees shrug when asked. There is no obvious current business case for it. It costs almost nothing to maintain. Everyone new to the company finds it slightly annoying. These are exactly the properties of a fence that was built after an incident nobody remembers, that is quietly preventing the incident from recurring.

The twelve-question Fence Audit

The Fence Audit — twelve questions to ask before removing any inherited HR policy

1
1. When was it written?
Date pins to context. Post-2018? Post-#MeToo. Post-2008? Post-financial-crisis layoffs. Post-2001? Post-Enron. The era often names the fence.
2
2. Who wrote it?
General counsel? Compliance? Founder? Angry ex-CHRO? The author signals the failure mode addressed.
3
3. What specific incident preceded it?
Ask the longest-tenured HR person, longest-tenured Legal, and the CEO separately. If any of them remember an incident, you have your fence's origin.
4
4. What would happen if the policy were absent tomorrow?
First-order effect. Usually visible.
5
5. What is the second-order effect over 12 months?
The failure mode the policy prevents typically manifests on this horizon, not the next quarter.
6
6. Who currently benefits from the policy?
If nobody benefits and nobody remembers why it exists, it may be genuine residue. If a specific stakeholder (Legal, Finance, IT Security) actively values it, they can articulate the failure mode.
7
7. What is the closest similar policy at a peer company?
Institutional isomorphism cuts both ways: if every peer has it, the removal will produce a differential exposure worth understanding.
8
8. Is there a regulatory requirement upstream of it?
Many 'stupid policies' are the visible tip of a compliance obligation. Removing the surface without removing the obligation is expensive.
9
9. What replaces the policy if it goes?
'Trust' is not a replacement mechanism. 'Better manager judgement' is not a mechanism. Named process, named owner, named metric.
10
10. Can I run the removal as a reversible pilot?
Six-month sunset of the policy in one BU. Measure. If the failure mode manifests, restore. If not, expand.
11
11. What does the Legal / Risk / Compliance owner say?
Not for veto. For diagnosis. They often know exactly which incident produced the policy.
12
12. Am I removing this because it is wrong, or because I did not create it?
The most honest question. Fresh-broom reflex is a real bias. Naming it reduces it.

Real fences that new leaders wrongly tear down

'No comp discussion in Slack DMs' — often written after a leak that produced a class-action risk. Removal produces the next leak within 12 months.
'All offers must be countersigned by Legal' — usually written after a rescinded offer produced a promissory-estoppel claim. Removal produces the next claim.
'No 1:1 meetings between execs and ICs without the IC's manager knowing' — often written after a bypass that produced a management-integrity incident. Removal produces the next.
'All external speaking requires 30-day PR notice' — often written after an exec accidentally disclosed material non-public information. Removal produces the next disclosure incident.
'No performance-related conversation over email' — often written after an email trail became discovery in a wrongful-termination suit. Removal restores the discovery risk.
'Every leaver has an exit interview conducted by someone not their manager' — often written after a manager-caused pattern of regretted attrition was hidden. Removal hides the pattern again.

This isn't an argument against reform

Chesterton explicitly allows reform. The Fence Audit reveals which policies are residue (remove them cleanly) and which are crystallised memory (address the underlying failure mode before removal). Real reform is more surgical, slower, and more effective than the fresh-broom approach.

FAQ

Frequently asked questions

Isn't this just organisational conservatism?

No. Chesterton's rule is procedural, not substantive. It says: understand the fence before removing it. It does not say don't remove it. The audit exists to enable clean removal, not to prevent it.

What if the original incident was 10 years ago and clearly won't recur?

Then the audit will show it. Q1-3 pin the era; Q9 asks what replaces the mechanism; if the era's failure mode is genuinely no longer live and no replacement is needed, remove cleanly. The audit is a discipline, not a veto.

How does this relate to Loose Coupling and Institutional Isomorphism?

Loose Coupling: some policies exist on paper but don't operate. Isomorphism: some policies exist because peers have them. The Fence Audit reveals which policies fall into each bucket and which are load-bearing.

Takeaways

The reformer's fallacy is assuming that an unexplained policy is an unnecessary policy. Often it is crystallised memory of a crisis you can't see.
The Fence Audit — twelve questions — turns Chesterton's principle into an operational discipline that any new People leader can run in the first 90 days.
Reform is not banned. Reform after diagnosis is the point. The policies that survive the audit are stronger for having survived; the ones that don't reveal the crisis their removal would have caused.
The hardest question is Q12: 'Am I removing this because it is wrong, or because I did not create it?' Answering honestly reduces the failure rate of new-leader reforms dramatically.