In April 2020, amid the initial wave of the COVID-19 pandemic, New Jersey’s unemployment system broke down.
Countless recently unemployed individuals were attempting to submit claims. The system, originating from the 1970s and never completely overhauled, was unable to cope with the surge. Applications were failing. Payments were delayed. The state’s emergency response was inundated.
Governor Phil Murphy appeared on television and did something nearly no one anticipated. He made a public appeal for volunteers. Specifically, he sought retired COBOL programmers who could emerge from retirement to assist in fixing the system. He required individuals who could work with a programming language that was, at that time, sixty-one years old and that very few people under fifty were proficient enough to manage.
Ultimately, a combination of retirees, contractors, and a gradual yet effective patchwork of solutions salvaged the state. However, the image lingered. A governor of a contemporary US state, in the twenty-first century, going live on television to request help from individuals old enough to remember Kennedy—because the software powering his state’s economy was written in a language they were among the last to understand fluently.
This is the future the global banking sector is now facing on a much grander scale.
What COBOL still accomplishes
COBOL was developed in 1959 during a gathering of American computer experts collaborating with the US Department of Defense. Its co-creator was Rear Admiral Grace Hopper, the same programmer who infamously discovered the physical moth in a Harvard Mark II relay in 1947 and documented it in her logbook, coinage of the term computer bug in the process. Hopper aimed for a language that ordinary business people could comprehend — a language where a line of code resembled “MOVE ACCOUNT-BALANCE TO DISPLAY-AREA” rather than a jumble of symbols.
She succeeded. COBOL — Common Business-Oriented Language — became the foundational language for the entire enterprise computing realm throughout the 1960s, 70s, and 80s. Banks embraced it. Insurance firms adopted it. Government bodies used it. Payroll systems, inventory management, banking ledgers, tax processing, benefits administration — for forty years, if you required a computer to reliably move money at scale, you nearly certainly coded it in COBOL.
Then, during the 1990s, the world progressed. New programming languages emerged. Object-oriented programming took precedence. COBOL fell into the “legacy code” category — technology that remained operational but was largely neglected in terms of new writing.
However, the existing COBOL code did not vanish. It continued to operate. It continued processing transactions. It continued transferring money.
Today, after sixty years, the figures are genuinely astonishing. Approximately $3 trillion in daily commerce is believed to flow through COBOL systems globally. Around 95 percent of ATM transactions interact with COBOL code at some point in their processing chain. Roughly 80 percent of in-person banking transactions follow suit. Close to 40 percent of US banking infrastructure relies on it. The IRS utilizes it. The Social Security Administration depends on it. The Department of Veterans Affairs uses it.
The most commonly repeated line regarding COBOL is that it predates the moon landing and still underlies the modern financial system. This statement is true and also downplays the enormity. If COBOL infrastructure were to fail tomorrow, significant portions of the American economy would plunge into darkness within hours.
The demographic concern
The individuals who created COBOL and who grasp the particular idiosyncrasies of specific decades-old codebases in various banks are predominantly in their late sixties or seventies now.
They wrote much of this code in their twenties and thirties during the 1970s and 1980s. They dedicated their careers to its maintenance. Many of them have been the sole keepers of particular systems for so long that the institutional knowledge of how their code functions resides solely in their minds.
They are retiring. Some are retiring voluntarily, while others have already left the workforce and are lured back with contract rates that would have seemed outrageous a decade ago. Senior COBOL contractors today can command daily fees in the low thousands of US dollars — not because COBOL itself is difficult, but due to the rarity of knowing how to manipulate a specific bank’s specific decades-old COBOL system without causing issues.
The informal statistic often cited in the industry is that many of the last generation of true COBOL experts have already passed away. Those who are left feel, by their own admission, that time is running out to pass on their knowledge.
The reverse migration
The response has been peculiar, and quietly intriguing.
A small group of younger developers — programmers in their twenties and thirties, who could easily be earning six figures in newer, more appealing fields — have begun intentionally learning COBOL. Not out of a love for the language, but because they’ve realized that COB…