How to dodge the next Sam Bankman-Fried

“Multibillion-dollar seed rounds in new start-ups that don’t have a product or technology or anything yet do seem a little bit unsustainable,” Demis Hassabis told the FT recently, referring to a tidal wave of investment crashing into newly founded AI companies.

The AI start-up funding market, as the Google DeepMind chief noted, is running very hot indeed. New research suggests this kind of frenzied investment environment significantly increases the chances of fraud cases further down the line.

Should venture investors worry about that? And should the rest of us?

CORPORATE GOVERNANCE

What’s behind the rise in fraud by start-up founders?

The curse of the Forbes 30 Under 30 claimed its newest victim a few weeks ago.

Gökçe Güven, the 26-year-old founder of customer rewards platform Kalder, was charged by US prosecutors with defrauding investors in a $7mn seed round by misrepresenting the financial details of her business.

Güven, the cover star of a Forbes list last year celebrating young stars in her sector, follows a string of other entrepreneurs who received similar accolades before facing fraud charges.

Sam Bankman-Fried and Caroline Ellison, of crypto catastrophe FTX, are the most notorious. Others include “pharma bro” Martin Shkreli; Charlie Javice, who pumped up her numbers to secure a sale of her fintech business to JPMorgan; Joanna Smith-Griffin, who is accused by prosecutors of doing a similar thing with the numbers of her education tech company; and whizz-kid real estate entrepreneur Nate Paul, who got criminal convictions for perjury and false financial statements. (Elizabeth Holmes, of medical tech disaster Theranos, made the cover of Forbes in 2014 just after passing the age limit for the Under 30 list.)

It’s worth stressing at this point that the vast majority of people named on these Forbes lists have not faced allegations of foul play (however offensive their LinkedIn humblebrags may have been). But these high-profile cases still prompt a compelling question: has fraud by start-up founders become more common? And if so, why?

It’s not your imagination

A group of academics at the University of Toronto has published the results of a two-year research project that gives important answers to that question.

The annual rate of serious alleged or proven fraud cases at US start-ups backed by venture capital money rose more than 10-fold between 2002 and 2021 — while fraud rates for other sorts of companies were plunging.

The rate of such cases fell over the same period by nearly 89 per cent for companies backed by private equity firms. The metric used here for publicly listed companies — annual fraud-related enforcement actions — had a similarly strong decline.

What’s been going on here?

In case you were wondering, given the number of female founders among the highly publicised cases mentioned above, this is not about gender. The study found no significant correlations on that front — nor on the age of start-up founders.

But other factors and trends were clearly apparent in this paper, which offers several important lessons for investors seeking to avoid embroilment in the next criminal start-up disaster.

Fragile systems

“Governance incentives explain much more than founder characteristics,” study co-author Ting Xu told me.

Fraud was far more likely at start-ups with “founder-friendly” governance structures — where founders retained the lion’s share of voting rights, or had effective control of the board. This kind of governance set-up has become much more common over the past couple of decades.

Funny business was also more common among companies that had raised money in a “hot” market, with hungry investors competing to buy stakes in promising start-ups. Those conditions can set the stage for investment on the “founder-friendly” terms mentioned above, while often leading to unhealthy pressure on founders to justify inflated valuations.

The paper offers a timely warning as investors rush to pour money into AI businesses. It also gives food for thought for pension funds and other non-VC investors who have been increasing their exposure to start-ups.

Fraud incidence was significantly greater at start-ups where “non-traditional” investors such as pension funds held stakes. Investors in FTX, for example, included Singaporean sovereign wealth fund Temasek and the Ontario Teachers’ Pension Plan.

One possible theory here is that non-specialist investors are less well equipped than seasoned Palo Alto VCs to hold founders to account. The more obvious problem, as Xu and his co-authors note, is fragmented investor influence. Capital tables have been getting more crowded, with a larger number of investors each wielding less clout.

Who bears the cost?

These findings may not seem alarming to many venture investors. Such funds famously invest in start-ups assuming that most of them will fail, but that a few roaring successes will more than counterbalance the flops.

Provided the venture firm itself isn’t legally implicated, it may not greatly matter to it whether a bust start-up has committed fraud or not. And despite the rising rate of serious fraud allegations or convictions, the chance of a given start-up facing such a case in a given year is still only 0.3 per cent, as of 2021.

But while the risks of start-up fraud might seem acceptable to VC firms, that’s largely because many of the negative impacts are “externalised” on to wider society, Xu and his colleagues argue. Consider Theranos, whose unreliable blood testing machines jeopardised the health of thousands of Walgreens customers, some of whom received false positive or negative results for serious conditions.

Meanwhile, a new sort of risk is emerging as governments remove barriers to private markets investing by individual investors, who may be less able to assess — or afford — the danger of putting money into a dodgy start-up.

“We’re not saying that the optimal level of fraud is zero,” Xu told me. Wiping out fraud entirely, he said, would require an oppressive level of monitoring and enforcement that could destroy more value than it saved. “But what’s optimal or expected from a VC perspective, might not be optimal for society.”

Smart reads