Two studies published in June reached a conclusion rarely said out loud in Silicon Valley: firms that take venture capital end up facing fraud charges more often than those that do not. Not because their founders are worse people, but because the system funding them pushes them in exactly that direction.

The first study comes from Britain's Imperial College and France's Emlyon Business School. The researchers assembled a database of tech founders and companies against which American regulators brought civil and criminal securities fraud proceedings between 2000 and 2023. The second, from the University of Toronto, examined 654 fraud cases in venture-backed American startups over the same period.

The most telling figure comes from the Toronto study: startups founded in overheated markets, with weak oversight and weak investor due diligence, are 19 percent more likely to later commit fraud. The measure, then, is not the founder's character - the measure is the climate at the moment the money came in.

Fraud has phases, and they have a name. Tim Weiss, one of the authors of the first study, describes it as „facading“ in three degrees. The surface facade is when a founder lies about how successful the firm is - typically at an early stage, when a vision is still what is being sold. The second phase is a reinforced facade: evidence is manufactured to back up the lie. The study cites the example of a testing app that invented client contracts and invoices, booked non-existent revenue and used those documents to convince investors to value it at a billion dollars.

The third phase is the deepest. The lie spreads to the technology itself - demonstrations that do not work, capabilities that do not exist. Weiss calls that building „parallel realities“.

The most uncomfortable part of both studies does not concern the founders. Investors, the authors say, sometimes „co-create the fraud“ - and not only through the unrealistic growth expectations they impose, but by continuing to fund the very people who already faced accusations. The Toronto study found little evidence that prior fraud stops anyone raising money for a new firm, even when the case was widely covered in the media. „New investors and the wider venture capital market do not punish past misconduct“, the report states - which, according to the authors, fits a culture that embraces failure without asking what caused it.

One further finding is worth remembering: startups where the board was controlled by the founders were twice as likely to commit fraud as those with an investor-controlled board or split control. Control without a counterweight, it turns out, is not an advantage - it is a risk.

Weiss proposes that the regulator automatically audit startups once they pass a certain threshold of raised investment, rather than waiting for a lawsuit or an employee complaint. And he asks for something harder: that investors be held accountable when they pushed for growth no company can deliver honestly. „Founders have no professional body that would set rules about what growth expectations are reasonable“, he says.

The current euphoria around artificial intelligence, the authors warn, is exactly the climate these cases are born from. The Balkans know the scene in different clothing - a firm growing on paper, an audit running late, everyone knowing and nobody asking. The question is not whether the numbers will be checked, but who carries the weight when they are.