Skip to content

The bezzle

The still-undiscovered stock of misappropriated entrusted wealth whose owner continues to treat the corresponding claim as intact.

Version
v1 · 2026-10-17 · History
Domain-specific #
14067
Domain group
Social Sciences
Origin domain
Economics & Finance
Subdomain
Financial Fraud → Economics & Finance
Aliases
Bezzle

Core Idea

The bezzle, in John Kenneth Galbraith's original sense, is the current stock of undiscovered embezzlement: entrusted wealth has been diverted, but its owner has not yet recognized that the corresponding claim is impaired. During that interval the diverter has the benefit, while the owner still behaves as though the asset remains available. The two apparent positions overlap without two real assets existing.[1] The stock is defined by the concealed gap at the time of assessment; no later discovery is required for an amount to belong to it now.

Galbraith suggested that trust in prosperous periods and sharper scrutiny in downturns can change the aggregate rate at which this stock grows or is discovered.[1] That historical argument is a possible pattern of accumulation, not a clock built into every instance. A single concealed diversion need not cause an economic boom, raise national spending, or end in a crash. Later writers extend The bezzle to noncriminal asset overvaluation; this entry keeps Galbraith's narrower meaning and distinguishes the extension explicitly.[2]

Structural Signature

Sig role-phrases:

  • Entrusted asset and apparently intact owner claim. A bank, investment intermediary, exchange, or other holder owes an owner assets or a claim whose recorded or assured availability the owner still accepts. Galbraith's bank example uses records of money supposedly placed in New York; FTX assured customers that deposits remained safely in custody.[1][3]
  • Unauthorized diversion and diverter benefit. Someone entrusted with access redirects the assets to another use. Without such misappropriation, an asset's honestly disputed market value belongs to the later broader usage, not this narrower stock.[1][2]
  • Undetected apparent-real gap. The owner has not yet recognized that the asset is unavailable as represented. This lag makes the diverter's gain and the owner's still-apparent claim coexist in perceived wealth. A concealed transfer, false record, or false custody assurance can sustain the lag; no one concealment technique is essential.[1][3][4]

Discovery is a conditional exit, not a fourth required role. When the owner learns of the shortfall, that amount leaves the stock of undiscovered embezzlement even if recovery remains impossible.[1] A bezzle can remain undiscovered beyond the period for which an analyst has evidence.

What It Is Not

The bezzle is not a synonym for every fraud or theft. A diversion noticed as soon as it occurs creates a real loss, but not the interval in which the owner's apparent claim remains intact. Nor is every false account statement enough: the narrow identity needs an actual misappropriated entrusted asset, not just a deceptive valuation of property that still exists.[1][2]

It is not a measured amount merely because a scheme's eventual loss is known. The discovered deficit and the then-undiscovered stock can differ in date, scope and valuation. Neither the Flint bank passage nor the cited FTX records measures an exact time series of the bezzle.[1][3][4] The later, noncriminal asset-bubble extension is a related interpretation and should not be silently counted as an original-sense positive.[2]

Scope of Application

The mechanism applies where assets are entrusted to an actor who can divert them while the owner continues to rely on an apparent claim. The form of that claim can vary. Galbraith describes bank officers recalling deposited funds from a New York lending placement while records still showed them there.[1] In the FTX case, the Department of Justice describes diversion of customer deposits even as customers were assured that those assets were held safely in custody.[3]

The label must be used with the time and evidence boundary stated. A bank ledger and an exchange's custody assurance are different ways to preserve an appearance of availability. Neither establishes every victim's internal belief, the precise stock on a chosen day, or a quantified economy-wide wealth effect. The SEC complaint gives detailed alleged FTX mechanics; its allegations should remain attributed as such.[4]

Clarity

The concept separates three dates that a report may blur: diversion, recognition, and recovery. The bezzle begins when an entrusted asset is diverted without the owner's awareness. Recognition removes the discovered amount from the hidden stock; repayment or recovery is a separate question. A victim can know of a loss and still not have recovered a cent.[1]

It also separates two balances. The owner has an apparent claim, while actual assets available to honor it are lower. The discrepancy is not proof that the owner physically held the same money as the diverter. It is an overlap in perceived wealth, the point of Galbraith's formulation.[1]

Manages Complexity

Fraud files can contain many accounts, transfers, assurances and dates. The abstraction compresses them into entrusted asset → concealed diversion → still-apparent owner claim. That smaller relation helps an analyst identify the period that matters without treating every false document or market loss as the same mechanism. In Flint, the key gap was funds absent from the stated New York placement while the bank's record still showed them there.[1] At FTX, the key gap was customer deposits used through Alameda while customers were assured of custody.[3]

The compression has limits. It does not determine legal charges, restitution, individual knowledge, or aggregate economic impact. Detailed records remain necessary to establish those facts and any amount at a particular date.

Abstract Reasoning

Galbraith defines the bezzle as an inventory of undiscovered embezzlement.[1] For analysis, let R(t) represent the amount still treated as available under owners' apparent claims at time t, and A(t) the assets actually available to satisfy those claims under the same scope. A positive gap may indicate a shortfall, but it counts toward this narrow bezzle only to the extent that the gap comes from unauthorized diversion and remains unrecognized by the relevant owners. This is an analytical test, not a measured series supplied by the two case records.

The test changes the investigative question. Instead of asking only how much was lost once the scheme collapsed, ask which assets were already diverted while owners still believed the claims intact, and when did that belief change? Earlier reconciliation can reduce the undiscovered interval without undoing a diversion already made. Galbraith's cycle argument concerns how creation and detection rates may vary across conditions.[1] The amount and timing in a specific case require their own evidence.

Knowledge Transfer

The pattern transfers literally between unlike financial settings when the same roles can be mapped: entrusted assets, a beneficiary of unauthorized diversion, and an owner who has not recognized the resulting shortfall. A bank's call-market placement and a cryptocurrency exchange's customer custody differ in instruments, records and actors, yet each can carry the concealed claim-versus-assets gap.[1][3]

The name should not be exported merely because another field has a stale record or hidden state. A warehouse's phantom inventory, for example, may overstate physical stock but need not involve entrusted financial assets misappropriated for someone's benefit. The portable idea of record and reality diverging belongs to a broader abstraction; the bezzle's finance and misappropriation conditions remain necessary here.

Examples

Canonical: Union Industrial Bank of Flint

Galbraith reports that bank officers recalled funds deposited for New York call-market lending while the records continued to show the money there. The officers used the funds in speculative positions. Their position failed with the 1929 crash, within the author's broader account of embezzlement coming to light.[1] The passage supports the hidden placement gap; it does not establish what each depositor thought or a specific auditor-discovery chain.

Mapped back: entrusted asset and apparent claim → funds deposited for lending and the bank record showing a New York placement; unauthorized diversion and benefit → officers' recall and speculative use; undetected gap → the record continued to show a placement that no longer reflected the money's use.

Applied: FTX customer deposits

The Department of Justice states that Bankman-Fried diverted billions in FTX customer funds to Alameda and other uses while repeatedly assuring customers that their deposits remained safely held in custody; he was convicted of fraud-related crimes.[3] The SEC's civil complaint alleges details of hidden liabilities, Alameda's special platform privileges, and the withdrawal crisis that exposed the shortfall in November 2022.[4]

Mapped back: entrusted asset and apparent claim → customer deposits and custody assurances; unauthorized diversion and benefit → customer assets routed to Alameda and other uses; undetected gap → custody assurances continued while diversion was concealed. The SEC complaint alleges the internal mechanics and describes withdrawal demands and the resulting revealed shortfall in November 2022.[4] The DOJ sentencing account documents diversion and false assurances.[3] Neither source establishes one recognition date or state of knowledge for every customer.

Structural Tensions

T1: Concealed reuse vs. independent verification. The diverter can continue benefiting from entrusted assets while the owner relies on an apparently intact claim only if the real shortfall stays hidden. An independent reconciliation that compares available assets with the claim reduces that interval, while a slow or absent check lets the overlap persist. More verification may reveal the loss earlier but cannot itself restore assets already diverted. The cases show different concealment carriers—bank placement records and exchange custody assurances—without proving a universal optimal audit schedule or monetary cost of checking. Diagnostic: What evidence, independent of the holder's representation, would establish that assets remain available to satisfy the owner's claim, and when would that comparison occur?

This is an adversarial tension between concealment and verification, not a claim that every bezzle has a boom-and-bust trajectory. Galbraith offers an aggregate cycle argument about embezzlement and discovery rates.[1] It does not establish a second mandatory tension in each instance.

Structural–Framed Character

On the structural–framed spectrum, the bezzle is mixed but finance-framed. Its relation—an actual shortfall coexisting with an apparently intact claim—can be described in general information terms. The named entry comes from Galbraith's account of entrusted money and embezzlement.[1] Its literal cases need financial ownership, custody and diversion.

The term carries evaluative weight because it identifies unauthorized benefit at another's expense, not a neutral bookkeeping variance. It depends on human practice: owners, entrusted intermediaries, representations and recognition determine whether the gap is still undiscovered. Its institutional origin lies in banking and economic accounts, though bank and exchange rules can differ. Its vocabulary travels readily as metaphor to asset bubbles and other illusory wealth, but Pettis's later extension changes a constitutive role by removing embezzlement.[2] It is imported rather than merely recognized in another setting only when that setting actually has entrusted financial assets and the hidden diversion relation; calling any stale record a bezzle is analogy.

Its character: a domain-specific financial mechanism with a partially portable information-gap skeleton. The name's literal extension stops where entrusted-asset misappropriation and an unrecognized owner shortfall stop.

Structural Core vs. Domain Accent

The broad skeleton is an apparent representation staying valid in use while the represented reality has changed. Record-Reality Divergence describes authoritative records, finite reconciliation cadence and downstream decisions that consume those records. The bezzle adds a financial claim, entrusted assets, unauthorized diversion, diverter benefit, and the time during which the owner has not recognized the loss. An exchange's custody assurance may preserve the apparent claim without satisfying the Prime's record, cadence and downstream-decision conditions. The overlap is real, but neither name is a substitute for the other's full signature.

The bank and exchange examples show literal recurrence across financial institutions, not that the entire bezzle identity survives in warehouses, health records or software caches. Such cases may instantiate the broader record-versus-reality skeleton but lack this named mechanism. Whether an unrecognized claimant loss paired with another party's gain travels outside entrusted finance is a separate Prime question; the available cases do not establish that extension.

Record-Reality Divergence is related because a bank record can overstate real placement, but its full signature requires an authoritative record, a finite reconciliation cadence and decisions made from that record; the bezzle's original definition does not require those conditions. Fraud is a related domain-specific wrong or act, whereas the bezzle names the current stock of hidden loss during an interval. Fraud can be detected immediately, and an authoritative record can be wrong without misappropriation. These distinctions preserve the bezzle's narrower economic identity without treating either neighboring concept as interchangeable with it.

Neighborhood in Abstraction Space

The bezzle sits in a moderately populated region (52nd percentile for distinctiveness): it has near-neighbors but no dense thicket of look-alikes.

Family — Financial Ratios & Instruments (22 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-17

Not to Be Confused With

General fraud names deceptive conduct and can be discovered immediately; the bezzle requires a still-unrecognized stock after diversion. A Ponzi scheme is one possible architecture for concealed claims, not the definition. Phantom inventory is a warehouse record overstatement and does not require the financial diversion and double-perception relation. Money illusion concerns nominal versus real purchasing power, not stolen entrusted assets. Pettis describes noncriminal asset overvaluation as a later broadened use of The bezzle.[2] It is outside the narrower Galbraith identity recorded here.

The FTX civil complaint[4] details platform allegations, while the DOJ sentencing account[3] describes the conviction and broad diversion. The two source grades should remain distinct.

References

[1] John Kenneth Galbraith, The Great Crash 1929 (Pelican, 1968), pp. 152–154 in the Aftermath I chapter (one-based PDF pp. 154–156). Consulted full-text scan is a nonpublisher mirror with noisy OCR. https://www.elib.uk/wp-content/uploads/2024/02/The-Great-Crash-1929-Pelican-1968.pdf registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n ↩o ↩p ↩q ↩r

[2] Michael Pettis, “Why the Bezzle Matters to the Economy,” Carnegie Endowment for International Peace (published 23 August 2021), especially “Bezzle Without the Embezzlement.” The URL directory is dated 2024 but the page displays 2021. https://carnegieendowment.org/china-financial-markets/2024/02/why-the-bezzle-matters-to-the-economy registry ↩a ↩b ↩c ↩d ↩e ↩f

[3] U.S. Attorney's Office, Southern District of New York, “Samuel Bankman-Fried Sentenced To 25 Years In Prison” (28 March 2024), adjudicated-case summary and fact statement on customer deposits, Alameda diversion and false custody assurances. https://www.justice.gov/usao-sdny/pr/samuel-bankman-fried-sentenced-25-years-prison registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i

[4] U.S. Securities and Exchange Commission, SEC v. Samuel Bankman-Fried, Complaint, No. 22-cv-10501 (filed 13 December 2022), ¶¶1–5, 31–54 and 78–84. Detailed civil allegations, not an adjudication. https://www.sec.gov/files/litigation/complaints/2023/comp25616.pdf registry ↩a ↩b ↩c ↩d ↩e ↩f