A beginner’s guide to accounting fraud (and how to get away with it)

Leo Perry was a fund manager at Ennismore Fund Management between 2003 and 2022. He is now an Open University student.

I spent most of the last two decades running a hedge fund. I had a bit of a thing for finding accounting anomalies — situations where management was saying one thing but the numbers told a different story. For a long time betting against these companies proved to be really profitable.

I doubt many people in my position would say they chose the wrong career. But now that I’ve worked out how to read accounts, and find it quite easy to spot signs of fraud, I also have some ideas for how we could run a good one of our own. And I’m not so sure there won’t be a lot more money in that line of business, if you can call it that.

The mechanics of making up sales are pretty simple. If we’re running a business and want to boost our top line, all we have to do is phone a friend. A good friend to be sure, who doesn’t ask questions. It’ll only take them a few hours to do the paperwork for setting up a shell company and then we have our customer, one we can invoice whatever we want. We have our fake sales (and I’m pretty sure our mate has done nothing wrong, in the eyes of the law).

You might think that sounds too easy, that someone would spot the problem and the con would quickly fall apart.

Well back in 2014 I explained what I thought looked like the most obvious fraud to an FT journalist. One of the things that caught my eye about Wirecard was the accounts of a company it bought in Singapore. Tucked away in the notes were references to specific customers — like Ashazi Services. [1] This was a Bahrain entity with no apparent operating business. A dormant shell that had never filed financials. Even the product Wirecard said it was licensing to it, the Elastic Platform, seemed to be a fiction (at least I never found any other mention of it by the company):

Dan McCrum, the FT journalist I met with, went to visit what there was of Ashazi as a part of a long-running series of Alphaville posts on Wirecard. And the whole scam did come crashing down . . . a mere six years later.

So yes, fake sales are only a start, but we’ve got plenty of time to worry about what comes next.

Even if some over-eager analyst does turn up at our sham customer, we can always move the goalposts. A few years ago I asked a Chinese-speaking colleague to visit some companies on the mainland. These were businesses that were reported to have signed purchase agreements with a western mining startup, which I was short. This startup had announced a deal to sell product a few years earlier. Then the contract was suddenly cancelled and simultaneously replaced with a similar agreement, but with a different Chinese entity — which we’ll call Tulip Industries. The deal equated to an outlay of approximately $150mn a year by the customer.

What we found at Tulip Industries was little more than a startup itself, with only a few field trials in progress. Even its most ambitious presentation forecasts involved a fraction of the product it had apparently agreed to buy. And its CEO was very clear that the deal wasn’t a firm commitment, only a loose framework. In fact he said he’d never spoken to the company that I was short, the deal was agreed through a friend in Hong Kong whose nephew worked for the miner. (He was much more committed to explaining to my colleague why China needed to invade Japan.)

A few months later that deal was mutually cancelled too, with a third Chinese entity now taking up the slack. Trying to tie down who was actually going to pay for this stuff was like a game of Whack-a-Mole.

I don’t mention this to allege any impropriety. My point is simply that investors in the mining startup didn’t seem to question much why the goalposts kept moving. That level of investor trust is useful if we want to disguise what our company will really be doing (which, in our case, will definitely be fraud).

If we don’t want to rely on a third party there’s also the DIY approach, using an entity that we control. A related party.

One of the first sets of financial statements I really struggled to reconcile with the story company management was telling was for Cupid, a UK-listed operator of dating apps and websites. I don’t know how many of its shareholders bothered to try out the sites it ran, even briefly, but I would guess not many. For anyone who did it seemed like they were too good to be true. Wherever you signed in from in the world, dozens of very keen and very attractive women would quickly get in touch. And they all happened to live nearby.

The Kyiv Post looked into how the company might be managing this back in 2013. Australian short seller John Hempton at Bronte Capital even took the trouble to log in from the most remote island in the UK (not in person, he used a virtual private network) and still found no shortage of admirers in the local area — even though the population there was small enough to all know each other. The fact that his profile stated he had syphilis apparently wasn’t a problem either. [2]

Cupid commissioned KPMG to investigate; its report found there was “no evidence of a company-organised practice” of staff using fake profiles to encourage subscriptions.

Cupid’s accounts were not as straightforward as its business model, and shareholders seemed to have even less time for them than they did for its services.

The annual report for 2011 had a chunky £2mn receivable from a company called Amorix, which was controlled by Cupid’s founders. Cupid said Amorix owed it this money because it had been collecting customer subscriptions on its behalf. But Amorix’s own accounts showed it only had about £80,000 in the bank, and no other assets to speak of. There was no trace of the money Cupid said was being collected for it.

Again, there is no suggestion here of impropriety. Mistakes happen. Cupid said the problem was quickly fixed and all the money was accounted for. What matters to us is the principle: this kind of irregularity wasn’t a dealbreaker for investors, which presents us with an opportunity.

The weighted average of shareholder due diligence on display here suggests that any company taking an inventive approach to sales generation will be given the benefit of the doubt for a long while. I mean, even the most sophisticated investor in the world apparently thinks it is enough to just ask management if there is any fraud going on.

I also think there’s a much better way we can do this, so we don’t ever have to answer that question.

The magic thing about fake sales is they are 100 per cent margin. All profit. You don’t need to go to the trouble of actually producing whatever it is you are pretending to sell, do you? So £100 in sales is £100 of profit. Hold that thought for a minute.

Now let’s think about what kind of business we want to start with to run our little fraud out of. Not a profitable one obviously. That would cost us good money to get control of in the first place, and we want nothing to lose. What we need is a business that has a lot of turnover but makes no money, but isn’t burning cash either. Something like a very low margin distribution business.

That might sound old-fashioned when, for the best part of the last five years, you could just make up your own money and name it after your dog. But it looks like the SEC has put down any chance of us selling Troycoin for now.

So let’s say we go into the fruit wholesale business. We buy boxes of bananas and sell them on at cost. Why? Well, while we’re only washing our face, if we turn over £100mn in bananas who’s going to notice when we add £1mn that’s lemons? That’s still less than 1 per cent of our sales after all. But if the £1mn is fake then it’s all profit. And as we make no money shipping bananas, the fake lemons are all of our profit.

The reported value in our business now all comes from made up sales to a fictitious customer; a customer set up by a mate that no one outside our office is ever going to know about. No one can pay them a visit if they don’t know its name. And they won’t, because at that size we wouldn’t even have to mention it exists. From the outside there’s just no way to spot anything wrong in our revenue numbers.

So what do you say? Who’s coming with me? I mean you didn’t get into this game to make the world a better place, did you? You’re here for a pound note and it’s there for the taking. The Man from Del Monte, he say “Yes!”

Further reading:
— Wirecard made this short seller right but not rich (FT)

1. I’ve always had a nagging suspicion about Ashazi. It was unusual, and very helpful, to have a dubious Wirecard client — and even some of the terms of its contract — disclosed in detail. Other shell companies that were customers were named too. But why? Wirecard management obviously didn’t want to have these relationships spelled out, so how was I able to find out about them in its accounts? Was the local auditor in Singapore uncomfortable with transactions that were so obviously questionable it wouldn’t sign off without the extra disclosure? If so that would be ironic, I think, because that auditor was Ernst & Young.

2. I need to apologise to John Hempton. Whenever I’ve told the Cupid story to people I always said (like I did here) that he said he had syphilis, but checking the facts for this article it seems that’s wrong. So let me be clear, when John made his dating profile he said he was suffering from an unspecified STI.