Just as they tend to over-glamourize murders, so TV dramas over-glamourize fraud: remember the BBC's rather good Hustle? In the real world, though, both have something in common; they are squalid crude affairs committed mostly by inadequates. This is a message of Dan Davies' history of fraud, Lying For Money.

He shows that most frauds fall into a few simple types. There's the long firm - setting up a fake company to borrow money which you run off with; Leslie Payne, an associate of the Krays, was a master of this. There are pyramid schemes, most famously those run by Charles Ponzi but Dan treats us to the story of the Pigeon King too. Then are control frauds, whereby someone abuses a position of trust: rogue traders fall into this type. And then there are plain counterfeiters. My favourite was Alves dos Reis, who persuaded the printers of legitimate Portuguese banknotes to print even more of them - so many as to devalue the escudo.

All this is done with the wit and clarity of exposition for which we have long admired Dan. His footnotes are an especial delight, reminding me of William Donaldson.

Dan has also a theory of fraud. "The optimal level of fraud is unlikely to be zero" he says. If we were to take so many precautions to stop it, we would also strangle legitimate economic activity. He calls this the "Canadian paradox". Fraudsters have, if only briefly, thrived there - not least by floating mining stocks. This is because Canada is a high-trust society; fraudsters exploit others' trust and the fact that dishonesty is the exception. By contrast, in low-trust societies people do business only with those they know well. That means less fraud, but also less business and more poverty. Trust, says Dan, is the basis of a prosperous commercial society - but it is also what fraudsters exploit.

If I have a criticism, it's that Dan doesn't much discuss the psychology of fraud - of how we fall victim to it. He writes:

Fraudsters don't play on moral weaknesses, greed or fear: they play on weaknesses in the system of checks and balances, the audit processes which are meant to supplement an overall environment of trust.

But is this wholly true? Some people are more vulnerable to scammers than others. I'd have liked Dan's analysis of why. He doesn't, for example, discuss Werner Troesken's fantastic history (pdf) of snake oil salesmen.

Also, perhaps we can extend Dan's work. He discusses outright criminal fraud. But there are similar types of behaviour which are legal - such as selling snake oil in 19^th century America. Dan hints at this when he describes the justification for bans on insider trading and other egregious types of market manipulation:

If you stopped robbing people blind with stock pools and takeover rumours, you would attract more of them into the capital markets, to make more money overall by robbing them through trading commissions and management fees.

We can think of fund management fees as a form of legal control fraud; using a position of trust to enrich oneself whilst failing to provide the service implicitly promised. CEO pay also falls into this category.

Similarly, whilst the directors of failing companies such as BHS and Carillion who took millions of pounds in salaries and dividends were not guilty of a long firm fraud, their behaviour was not wholly dissimilar.

And cynics might add that newspaper columnists, think-tankers or politicians are similar to affinity fraudsters; they take money for misleading like-minded people.

The law of course makes a sharp distinction between criminal and non-criminal behaviour: that's what it's for. But I wonder whether the distinction is so clear in reality. It's not just criminals who lie for money.