Centralisation Trap
The Treasury's overconfidence cycle: excluding local knowledge produces bad decisions, then punishes locals for the failure.
The Centralisation Trap is Collier's name for the self-reinforcing cycle by which highly centralised governments concentrate decision-making power in ways that guarantee failure, then use that failure as justification for further concentration. The UK Treasury exemplifies the mechanism: a small group of academically elite, geographically homogeneous civil servants (almost entirely from the South East, overwhelmingly Oxbridge) hold near-total power over decisions affecting places they have never lived in or consulted.
The trap operates in stages. First, the Treasury is staffed by people whose social profile guarantees they lack the lived experience to make good decisions for left-behind places. Second, their intellectual confidence — justified by examination performance — leads them to believe they can plan bus routes in Sheffield without visiting Sheffield. Third, they veto or undermine local government proposals (90% of Levelling Up bids were rejected by the Treasury), preventing local knowledge from reaching decisions. Fourth, local governments — starved of money and power, with therefore nobody voting for them — perform poorly. Fifth, the Treasury cites this poor performance as evidence that it was right to override local governments. The cycle repeats.
Breaking the trap requires doing three things simultaneously: staffing central government with people who have lived experience outside the magic circle, devolving meaningful financial power (not just mandates) to local authorities, and making local governments democratically accountable — which only becomes possible when they have enough power to be worth voting on.
- Overconfidence scales with the narrowness of one's social background — the most socially exclusive decision-makers are the least aware of what they don't know.
- Local governments perform badly when given no power and no money — their poor performance cannot be used as evidence against devolution.
- Democratic accountability at local level only emerges when local governments have enough power and resources to be worth voting for.
- Social diversity in gender and ethnicity without geographic and experiential diversity produces the same centralisation failures with a more inclusive face.
- The Treasury protects taxpayers from local government failures that the Treasury itself creates by withholding the resources locals need to succeed.
- Diagnose the knowledge gapMap the lived-experience backgrounds of those making decisions against the populations affected. Collier's test: how many people in your decision-making body have lived in a deprived area north of the Watford Gap? If the answer is negligible, the knowledge gap is structural.Pro tipGender and ethnic diversity data is tracked; geographic and socioeconomic background data almost never is. Request it explicitly — its absence is itself diagnostic.
- Identify where local knowledge is being vetoedAudit how many locally-generated proposals or bids are being overridden by central bodies — and on what grounds. The 90% Levelling Up veto rate is an extreme version of a pattern found in most centralised systems.WarningThe veto is often procedural or budgetary, not explicit — Treasury 'scrutiny' is functionally a veto even when it presents itself as quality control.
- Measure the democratic accountability gapCheck local election turnout. Very low turnout in local elections almost always tracks with local governments having no meaningful powers — voters rationally don't engage with institutions that can't affect their lives.Pro tipAndy Burnham's Manchester bus devolution is Collier's model: real power (over routes, fares, procurement) drives real democratic engagement.
- Devolve money and power simultaneouslyMandates without money are not devolution — they are cost-shifting. Real devolution requires fiscal transfer (block grants, not project bids), decision rights over spending priorities, and genuine accountability to local electorates rather than to Whitehall.Pro tipUse East Germany's KfW as the structural model: a dedicated institution with 7,000 staff and €70B/year, devolved to city level, learning across cities.WarningBeware 'levelling up' funds that require competitive bids vetted by Whitehall — this is the trap in disguise.
- Mandate geographic and experiential diversity in central hiringTreasury and Cabinet Office diversity targets must include where people are from and what their lived experience includes — not just gender and ethnicity. The permanent secretary anecdote (Oxbridge-only pipeline making non-Oxbridge permanent secretaries feel like outsiders) quantifies the culture problem.WarningSurface-level hiring diversity without changing the selection exam and socialisation pipeline will not shift the overconfidence dynamic.
The UK government's flagship Levelling Up programme had a total budget of £4B over four years — already peanuts against East Germany's €70B/year. The Treasury vetoed 90% of local bids, resulting in ~£400M actually spent across 50M+ people.
A retiring permanent secretary — who had attended a good public school and a non-Oxbridge university — said that despite reaching the top of the civil service, he was made to feel an outsider throughout his career because he had not attended Oxford or Cambridge.
The German Basic Law — co-designed by British advisers post-WWII — requires equal per-capita public infrastructure spending across all regions. The UK has no equivalent constraint, allowing the Treasury to concentrate transport and R&D spending in London and the South East.
Collier developed this diagnosis across his advisory role to the UK Levelling Up programme and his Oxford policy teaching. The Levelling Up experience was the direct empirical grounding — watching 90% of local bids vetoed by a Treasury that then cited the absence of successful local spending as evidence of local government incapacity. The German constitution's basic law (equal per-capita infrastructure spending by region) crystallised for him what the structural alternative looks like.