Notes — Patrick McKenzie on Navigating Complex Systems

Notes on Patrick McKenzie in conversation with Tyler Cowen — Conversations with Tyler(https://conversationswithtyler.com/episodes/patrick-mckenzie/), 10 January 2024.

Four questions [Adler frame]

Q1 — What is it about as a whole? A rapid-fire tour of how large, layered institutions — banking, payments, debt collection, government, labour markets — actually work beneath their visible surface, conducted by a writer whose career has been explaining exactly this class of system (‘systems thinking applied to businesses’). The conversation ranges across payment-system mechanics, fraud and ‘evil’ as an empirical category, McKenzie’s two decades in Japan, crypto scepticism, and what running a volunteer COVID-vaccine-finding operation taught him about talent.

Q2 — How is it argued? Almost entirely through worked micro-cases rather than abstract theory: the credit-card signature’s actual legal function, the queuing-theory logic behind empty bank parking lots, the FBI’s median bank-robbery haul, the CFTC’s Tether examination findings, and a personal spreadsheet used to choose Japan as a country to move to. McKenzie’s method throughout is to interrogate a specific institutional practice for its original rationale, then ask whether that rationale still holds — often finding an answer that persists by inertia rather than continued function [§ Money, security, and the persistence of security theatre].

Q3 — Is it true, in whole or part? Much of the specific evidentiary weight rests on McKenzie’s own trade experience (Stripe, VaccinateCA, his salary-negotiation blog post’s self-reported reader impact) rather than independently audited data, and he flags this himself — spurious decline rates are ‘trade secret’, turnover figures are estimates he distrusts making up on the spot [?]. His claims about the CFTC’s Tether findings and FBI bank-robbery statistics are the most externally verifiable and are stated with citation-like confidence. His broader interpretive claims — that people ‘overpredict evil and underpredict competence’, that nonspecialist talent is systematically underrated — are stated as personal Bayesian updates from a small number of vivid cases (maple syrup fraud, VaccinateCA), not as generalisable findings, and he is careful to flag this distinction himself.

Q4 — What of it? The recurring mechanism across sections — a costly practice persisting because no single actor is accountable for fixing a distributed, patchwork system — is the episode’s most portable idea and recurs from credit-card declines to Western Union fees to bureaucratic bank security theatre. It is a mechanism, not a rebrand of an existing wiki concept, but the transcript does not develop it into a page-worthy standalone construct (no name, no dedicated treatment) — see the episode page’s Key idea 1 for its clearest statement. McKenzie’s well-known term ‘the dangerous professional’ does not appear anywhere in this transcript; it belongs to his separate written work and is not sourced here.

Glossary

Solemnisation — the original legal function of the credit-card back-of-card signature: not identity verification, but a formal, binding acknowledgement that the cardholder promises to honour charges made with the card. [§ Money, security, and the persistence of security theatre]

Spurious (false) decline — a legitimate credit-card transaction rejected in error by some point of failure in the multi-decade patchwork of payment infrastructure; the true rate is treated as a trade secret across the industry. [§ Money, security, and the persistence of security theatre]

Know-your-customer (KYC) / anti-money-laundering (AML) — regulatory identity-verification requirements imposed on money transfers to prevent laundering, terrorism financing, and tax evasion by structuring; McKenzie treats their cost as real but their intent as broadly defensible, distinguishing his view from crypto advocates who prefer evading them technologically. [§ Twitter, money transfers, and the costs regulation imposes]

Structuring — breaking a large money transfer into many small, individually unregulated transfers to avoid KYC/AML detection thresholds; the rationale given for why even small transfers face compliance friction. [§ Twitter, money transfers, and the costs regulation imposes]

Key claims by section

Money, security, and the persistence of security theatre [§ Money, security, and the persistence of security theatre]

The credit-card signature’s original purpose was solemnisation (a binding promise to the bank), not clerk-verified authentication; it now persists as a legal vestige with no real security function.
Spurious decline rates are far higher than any rational tolerance level would suggest, and persist because the payment ecosystem is a patchwork of legacy systems with no single accountable owner — not because the problem is unsolvable, since targeted executive attention has moved the number before.
McKenzie updates his sense of which ‘intractable’ problems are actually tractable by analogy to SpaceX cutting launch costs by three orders of magnitude, a result he would have bet against as recently as 2005.

Fraud, evil, and where crime hides [§ Fraud, evil, and where crime hides]

Debt collection’s brutal working conditions and near-universal turnover are partly explained by a subset of workers who are drawn to inflicting suffering — McKenzie treats ‘evil’ as a real, load-bearing feature that an accurate model of the world must include.
Maple-syrup supply-chain fraud running into the millions of dollars surprised him, since he assumed heavy regulation and ‘real company’ buyers implied near-zero fraud; smaller boutique buyers skip the verification checks larger ones run.
He is more uncertain whether low detected rates of insider trading reflect genuine rarity or poor detection of anyone above the least competent offenders, given how unsophisticated most publicised SEC cases appear [?].

Twitter, money transfers, and the costs regulation imposes [§ Twitter, money transfers, and the costs regulation imposes]

McKenzie rates the odds of X succeeding as a universal payments app at far below 1%, because grafting payments onto an unrelated user habit does not itself create adoption.
Half of Western Union’s revenue funds physical points of presence in both sending and receiving countries — the primary driver of high transfer fees, more than any technology gap crypto could close.
AML/KYC compliance costs are real and, in his view, understate a legitimate concern (small-transfer laundering via structuring); his disagreement with crypto advocates is that the fix should be legal reform, not technological evasion.

Banks, parking lots, and why bank robbery barely pays [§ Banks, parking lots, and why bank robbery barely pays]

Bank parking lots are over-provisioned nearly always to avoid losing rare, high-value new-account-opening visits (30–60+ minutes of dwell time), which McKenzie treats as confirmation that ordinary deposits are ‘a favour one does for the bank’.
The FBI’s median bank-robbery haul is about $8,000; banks train staff to comply fully with robbers because an employee’s liability-adjusted value vastly exceeds any single robbery’s loss, and typical bank robbers are non-professional and driven by acute distress rather than rational profit-seeking.

McKenzie chose to move to Japan at twenty using a spreadsheet ranking countries — reasoning that felt rigorous but rested on a since-falsified premise (that the dot-com bust would end US engineering hiring); the decision proved life-shaping regardless of its faulty basis.
Japanese salaryman culture substitutes direct social-status allocation for the money a Western worker would earn and spend on status goods, illustrated by two office colleagues visibly renegotiating a physical seniority chart based on recent performance.
McKenzie regards his six years as a burnt-out salaryman as a clear regret, believing the useful signal could have been extracted in weeks rather than years had social pressure not held him there.

Crypto scepticism: Tether, Bitcoin, and epistemic humility [§ Crypto scepticism: Tether, Bitcoin, and epistemic humility]

McKenzie treats the CFTC’s finding that Tether was under-backed on more than 75 of an 18-month examination’s days as vindication of his long-held fraud claim, expecting eventual collapse via a bank-run dynamic similar to Madoff’s exposure in 2008.
He remains bearish on Bitcoin’s price relative to demonstrated utility even after a decade of the market voting against that view, and cannot fully explain why Bitcoin — which he regards as clearly not crypto’s best product — remains its most valuable one.

Talent, negotiation, and returning home [§ Talent, negotiation, and returning home]

Running VaccinateCA with largely non-credentialed volunteers confirmed McKenzie’s prior that internet-sourced talent can match trained institutional staff (a county health department) within weeks, while he still values credentialed institutions for legitimacy and coalition-building.
His 2012 salary-negotiation blog post produced a self-reported aggregate reader impact he stopped tracking after roughly $10 million a year — evidence, in his reading, of a near-universal cultural taboo against negotiating that leaves large, cheaply captured value on the table [?].