An Insurance Company Invented 64,000 Fake Customers. Then It Started Killing Them Off.

Its real business sold mutual funds and life insurance in one package that was perpetually starved for cash, so executives minted fake policies to sell to reinsurers for upfront money, then minted more fakes to pay the reinsurers back. A computer wrote the policies and hid them from auditors, and the stock ran from $6 to $80 on profits that never existed. The plan was to keep growing until they could buy a real insurer and "go straight."

Stanley Goldblum was a Los Angeles insurance agent who had previously been a scrap dealer and a meat salesman. In 1960 he helped found Equity Funding Corporation of America, and it went public in December 1964 at $6 a share. Its product was clever: sell a customer mutual fund shares, lend the customer money against those shares to buy a life insurance policy, and hold the funds as collateral. The bet was that the fund's growth would cover the premiums and the loan interest, leaving the customer, after ten years, with a paid-up policy and some investments left over. The flaw was structural. Because customers paid in installments, Equity Funding was perennially short of cash, and short of cash is a hard place to run an insurer that has promised steady earnings growth to Wall Street.

The legitimate cure for an insurer's cash hunger is reinsurance: sell a batch of your policies to another insurer, who pays you roughly $1.80 for every $1 of first-year premium upfront and takes over the risk, while you keep servicing the policies and pass along about 90 cents of each later premium. Equity Funding did this, then did it again with policies that did not exist. A fake policy costs nothing to issue and pays real cash upfront. The catch is that a fake policy generates no future premiums, so every year Equity had to send the reinsurer money it wasn't collecting, which meant selling still more fake policies to raise that money. It was a pyramid that paid its old debts with new fabrications, and the only exit the executives could see was to keep inflating the stock, use it as currency to acquire real companies, and eventually wash the fake book out inside a big enough legitimate insurer.

The computer turned a hand-forgery problem into an assembly line. On November 2, 1970, an employee was instructed to write a program generating fictitious policies with $430 million of face value. The fakes were tagged "Class 99," meaning no billing, and by the end between 56,000 and 64,000 of them sat on the books, about two-thirds of the subsidiary's entire book of business. Of the $3.2 billion of insurance shown in force at the end of 1972, about $2.1 billion, 66 percent, was fictitious. A second program filtered the fakes out of any printout an auditor requested, and when paper files were needed, employees held overnight "forgery parties," one recalling he "had fun being the doctor and giving the guy's blood pressure." When auditors tried to confirm policies by mail, Equity addressed the letters to its own salesmen, who signed them back. Executive vice president Fred Levin forbade mailing confirmations to policyholders, claiming it would upset the salesmen. To close the loop they filed death claims on policyholders who had never lived, and one programmer was writing code to automate the "killing," carefully spacing the deaths so no reinsurer saw a pattern.

It broke the way these usually do: a fired employee, Ronald Secrist, walked out in March 1973 and told a securities analyst named Ray Dirks that billions in policies were fabricated. Dirks investigated, confirmed it with the company's own computer technicians, and told the SEC and his institutional clients. The clients sold, five investment advisers liquidated more than $16 million, and the price fell from $26 to under $15 over two weeks until the New York Stock Exchange halted trading on March 27. Equity Funding filed for bankruptcy on April 5, 1973, the second-largest in U.S. history at the time. Touche Ross found a company that claimed nearly $750 million in assets and had $489 million, with about 80 percent of the insurance subsidiary's assets nonexistent; the trustee wrote that the public Equity Funding "did not have the assets; it did not have the revenues; it did not have the sales; it did not have the net worth; and it had not made the profits." Twenty-two people were indicted on 105 counts; Goldblum pleaded guilty and served four years of an eight-year sentence. Dirks, censured by the SEC for tipping, took his case to the Supreme Court and won, establishing the personal-benefit test for insider trading that still governs whistleblowers today.

The takeaway for a modern operator. Equity Funding was never really an insurance fraud; the trustee called it a securities fraud, and the engine was a cash-advance pyramid dressed in reinsurance. The incentives were baked into the structure: a business built on installment cash flows that could not fund its own growth, a stock price that doubled as acquisition currency, and a reinsurance market that paid real money today for policies that didn't have to exist. The computer didn't cause the fraud, it scaled it, and the auditors missed it because they audited the printouts, not the program that produced them, and because management had quietly captured them. The thing that finally broke it wasn't a model or a regulator (the SEC had tips as early as 1971 and did nothing) but a single human with nothing left to lose. If a business is structurally starved for cash and its only growth story is "keep expanding so we can merge our way to legitimacy," read the cash-flow statement, not the earnings. And when you audit a system, audit the code that generates the numbers, not the numbers it hands you.