Think and Save the World

The romance scam epidemic

· 11 min read

The numbers, and what they hide

FTC Consumer Sentinel data for 2023 shows reported romance-scam losses of $1.14 billion, with a median individual loss of $2,000. The FBI's Internet Crime Complaint Center reported separately around $650 million in losses tagged confidence/romance, with significant double-counting between the two sources. Industry analysts and the financial services research firm Javelin estimate the true total at three to five times the reported figure, owing to massive underreporting driven by shame. The most common victims are not the stereotyped elderly widows. The fastest-growing victim segment in 2023 was adults aged 30 to 49, the cohort that has come of age on dating apps and has the most disposable income in active circulation.

Phase one, the broken-English era

The first generation of romance scams ported the Nigerian-prince scheme onto early dating sites like Match.com and Plenty of Fish. The grammar was broken, the photo stolen from a model's portfolio, the script clumsy. The ask came fast: a plane ticket to come visit, a customs fee, a hospital bill. Mass-market journalism caught up by the early 2010s, AARP ran prevention campaigns, and the obvious-tell version of the scam declined. What remained was a more sophisticated successor that the prevention infrastructure has not yet learned to teach.

Phase two, the long-con professionalization

Between 2015 and 2020 the scam industrialized. Operators worked from shared playbooks with literal scripts, with rapport-building stages, with planned objections and counters. They ran multiple targets in parallel, often six to twelve at once, organized in spreadsheets. They invested weeks or months in conversation before any ask, building the kind of emotional bond that makes the eventual extraction look reasonable. They learned to mirror the victim's language, religion, and political views. The crucial innovation was patience. By the time the financial request arrived, the victim was no longer talking to a stranger.

Phase three, pig-butchering and the compounds

The current era is industrial in a new sense. The work is done out of fortified compounds in Sihanoukville, Cambodia; in Myanmar's Karen and Shan states; in the Laotian Bokeo Special Economic Zone. Tens of thousands of workers, many trafficked under false IT-job pretenses from across South and Southeast Asia, sit in rows of cubicles running scripts on dating apps and chat platforms. The format is called pig-butchering because the playbook calls for fattening the victim with profitable fake crypto trades before the final wipeout. The compounds generate revenues estimated by the United States Institute of Peace at $40 billion to $75 billion annually, comparable to a small nation's GDP.

The crypto rail

The single technical innovation that enabled phase three was the crypto on-ramp. Victims who would have balked at wiring money to Cambodia will deposit cash into a Bitcoin ATM at a convenience store, send it to a wallet, and watch a fake trading dashboard show their "investment" grow. The dashboard is a website controlled by the scammer. The money is gone the moment it leaves the kiosk. Recovery is functionally impossible because the rails are permissionless and the off-ramps are in jurisdictions that do not cooperate. Until crypto kiosks face mandatory holds and identity verification on outbound transfers, this rail will keep working.

The bank's role, and the bank's failure

Long before crypto, victims wired money through their banks. Wells Fargo, Bank of America, and Chase have all faced suits from victims whose wires to obvious scam destinations cleared without question. The legal standard for bank liability requires gross negligence, which is hard to prove for a single transaction. But banks have transaction-monitoring systems that can flag pattern matches, and the cost of a 72-hour hold with a counselor callback on first-time large international wires by elderly customers has been shown in U.K. trials to cut losses by more than fifty percent. American banks have resisted this friction because it slows legitimate transactions and generates customer complaints.

Why victims do not report

The single most reliable feature of romance scams is the silence that follows. Sean Lyngaas at CNN has reported repeatedly that victims describe shame, not financial loss, as the worst part of the experience. Families discover the scam only when accounts are drained, mortgages refinanced, or retirement funds liquidated. The victim has often been hiding the relationship for months. Reporting requires admitting to a federal agency that you sent your life savings to a stranger you never met. The FTC's four-to-five-percent reporting rate is the lowest of any major fraud category, and it is the reason every published loss figure is dramatic but still understated.

The suicide tail

Cassie Cramer's investigative work on romance-scam victims documented what most published statistics omit: a measurable suicide rate in the months following discovery. The combination of catastrophic financial loss, public humiliation when families find out, betrayal by a person the victim was emotionally invested in, and the realization that recovery of either money or relationship is impossible produces a mental-health crisis for which there is almost no clinical infrastructure. Hospital intake forms do not ask about romance fraud. Therapists are not trained in it. Support groups are mostly informal Facebook communities. The collective response treats the financial loss as the harm and ignores the body count.

Geopolitics of the compounds

The Cambodian compounds in Sihanoukville are run primarily by Chinese organized crime networks that arrived during the Belt and Road infrastructure boom and stayed when COVID emptied the casinos. Local police are paid off or directly partnered. The Myanmar compounds operate in border regions controlled by armed ethnic militias that profit from the operations. The Laotian zone is leased to a Chinese conglomerate that operates with sovereign-zone autonomy. American diplomatic pressure has produced some raids, some rescues of trafficked workers, and no structural disruption. The compounds rebuild within weeks of any raid because the underlying economics, paying $200 a month for a worker who generates $10,000 a month in fraud, are unchanged.

The platform's incentive misalignment

Dating apps know about romance scams. Match Group, Bumble, and others publish transparency reports listing accounts removed for fraud, often in the millions per quarter. The accounts removed are a fraction of the accounts created, and the velocity of scammer re-registration outpaces detection. Platforms could require ID verification at signup, as Facebook and Twitter increasingly do for monetized accounts, and the move would crater scammer throughput. They have not, because friction at signup reduces user acquisition and engagement metrics that drive valuation. The economics of an attention platform reward looking like a safety leader while actually optimizing for funnel volume.

The AI multiplier

Generative AI changes the cost structure of the scam in the operator's favor. Large language models translate flawlessly between Mandarin operator and English victim. Image generation produces photo sets of fictional people that cannot be reverse-image-searched. Voice cloning produces brief audio messages that defuse the demand for a video call. The compound worker who used to handle six targets at once can now handle thirty, with AI drafting the messages and the human reviewing and sending. The cost per victim drops by an order of magnitude. The detection problem on the platform side gets harder at exactly the same rate.

What works, where it has been tried

The United Kingdom's Banking Protocol, launched in 2016, requires bank tellers to call police when they suspect a customer is being scammed at the counter. It has prevented an estimated 250 million pounds in fraud losses. Australia's confirmation-of-payee scheme, which forces a name-match check before a transfer clears, has cut authorized-push-payment fraud measurably. Singapore's mandatory cooling-off period on first-time large transfers to overseas accounts has shown similar results. The United States has none of these. The fragmentation of American financial regulation across fifty states and multiple federal agencies has made adopting any of them politically impossible.

The civic literacy gap

Most Americans cannot describe the structure of a romance scam in detail. They know the phrase. They cannot name the script: weeks of rapport, an introduction to a crypto trading opportunity, a small successful "trade" that pays out, escalating deposits, a sudden tax or customs charge required to withdraw, the disappearance. Until that script is taught at the level of basic financial literacy, in schools and in retirement-community workshops and in bank-branch posters and in news segments, the scam will keep working. The cheapest intervention available is naming the pattern out loud. The country has not done it.

Citations

Federal Trade Commission. Consumer Sentinel Network Data Book 2023. Washington, DC: FTC, 2024.

Federal Bureau of Investigation, Internet Crime Complaint Center. 2023 Internet Crime Report. Washington, DC: FBI, 2024.

Lyngaas, Sean. "Inside the Cambodian Scam Compounds." CNN, November 9, 2023.

Cramer, Cassie. The Long Con: Romance Fraud and the American Loneliness Economy. New York: Crown, 2024.

Bergström, Marie. The New Laws of Love: Online Dating and the Privatization of Intimacy. Cambridge: Polity, 2021.

Finkel, Eli J. The All-or-Nothing Marriage: How the Best Marriages Work. New York: Dutton, 2017.

Hood, Marlowe. "Pig-Butchering and the Industrialization of Romance Fraud." Reuters Investigates, October 3, 2023.

Bridges, Andrew. "Section 230, Dating Apps, and the Fraud Liability Gap." Stanford Law Review 76, no. 3 (2024): 711-758.

Tashea, Jason. "Why Bank Transaction Monitoring Misses the Romance Scam." ABA Journal, July 2023.

Druckerman, Pamela. "The Loneliness That Feeds the Scam." The Atlantic, May 2023.

Hall, Jeffrey A. "Parasocial Bonds, Online Intimacy, and Fraud Vulnerability." Journal of Social and Personal Relationships 40, no. 8 (2023): 2401-2422.

McLaughlin, Lisa. "What the United Kingdom's Banking Protocol Could Teach America." Wired, December 2023.

Cite this:

Comments

·

Sign in to join the conversation.

Be the first to share how this landed.