Up until now (1981), the job of the Treasury was to lend, borrow and exchange currencies in order to ensure that everyone’s interests were protected: customers’ money was properly looked after, and the bank didn’t expose itself to undue risk. In fact, they had to end each day with a near-enough ‘flat’ position, which had to be reported to the Bank of England. Failure to fetch up within the statutory limits would meet with unspeakable retribution. (I never found out exactly what this was, and nobody ever asked; we amused ourselves imagining ritual button–snipping humiliations before a Star Chamber of hooded pinstripes, with loaded revolvers left discreetly behind the aspidistra, but I’m pretty sure it never actually happened.) In a phrase, the culture was one of avoiding danger rather than courting risk.
All that changed when in 1980, as one of her very first actions, Maggie abolished Exchange Controls and the markets blinked and realised they didn’t have to bother with all that anymore; they could do whatever they wanted. A very significant moment. Not only could the volume of transactions explode, but they could invent new ones willy-nilly: money had begun its inexorable path from being the means of exchanging commodities to becoming a commodity in its own right. (You know the rest.)
From where I and my fellow academics sat across the river, it became obvious that they’d need our help. In particular, they’d need the computers to tell them what they’d been up to; which meant capturing the information as early as possible. The days of filling in* a piece of paper, a ‘deal slip’, and passing it through a window to the back office to be processed sometime before close of play were gone – the deal slip had to be visible as it was written.
The trouble was, our computer systems weren’t up to this kind of challenge. When I suggested to John, the departmental guru past whom all proposals had to be run, that we might have a need for one transaction to be revamped into another, then to be queued up in the back office for completion, he frowned for a bit, then smiled. “Always one for pushing the boundaries,” he said. “You do realise we’ll have to rewrite the whole lot, don’t you?”**
So we did.
Once again, it was seat of the pants stuff. I knew how it worked by now, so I got myself a Test Team before I had a single programmer. The programs gradually started to trickle through, in more or less testable condition but by no means stable. It didn’t take long for one of the Team to uncover a trick which would infallibly crash the system, and would take at least two hours to recover. This would tend to happen once or twice a week at around five to twelve. As I was on both sides of the river, I didn’t always get down the pub with the Team: indeed, I sometimes heard the southern end of the conversation, which might have gone something like “….” “Oh, again?” “….” “Yeah. Couple of hours at least.” [Programmer rings off, glances around the room.] “Pint, anyone?”
Anyway, somehow it got done. One of my fondest memories is of a parallel run, where a cloned and converted day’s live work was run through the new overnight programs to make sure everything came out looking the same. It didn’t. At about two a.m., Malcolm, the guy in charge of this portion of the process, looked up and suggested a coffee break. He was deep in thought for a while, then delivered his diagnosis. “The trouble is,” he said, “we’ve either got too many debits or not enough credits.”
Such is the stuff of which banks are built. It did get done in the end***; actually, most of those systems were still doing the business when I made my final escape (eighteen years later to the day, and thirty years to the day after I’d turned up as a rookie on the doorstep of 67 Lombard Street) on 1 June 2000.
The remainder of my banking career was mostly spent attending and chairing various standards-making committees, which involved swanning around the world to meaningless meetings solving problems nobody else knew existed, whilst pretending to be important; I’m sure you don’t want to hear about that, so I’ll shut up now.
* If you were lucky – as often as not the trade would remain on the dealer’s pad for hours, until he remembered to do his paperwork just before pub time that evening.
** I know how much you like it when I talk dirty, so I’ll just say that this involved converting the entire system from a TCAM to a VTAM platform, with old Assembler code being rewritten in the industry-standard PL/1 language – a classic example of my favourite intellectual concept, the inverted pyramid.
*** I have no idea how much of a contribution this made towards today’s sleek, efficient financial trading mechanisms.