Algorithms dominating financial markets: neglected risk with (potentially) large impact
Today, financial markets are dominated by algorithms: computer programs buying and selling financial products in a fraction of a second. Next to very real risks to the financial markets such as flash crashes (relatively frequently covered in the financial media), this raises a more fundamental issue: do prices of financial products still make sense?
Or to put it differently: if 70%-80% of the shares of companies (think Shell, Unilever, Apple etc.) are bought and sold by computer programs that only react to the actions of other market participants (think: if someone buys x stocks, then algorithm buys y stocks; if price in share x raises n%, then buy stock y), aren't prices then not just the result of algorithms responding to each other instead of real market forces? Or more technically: are financial markets really more efficient nowadays than 10+ years ago, according to the definition of efficiency of the Efficient Market Hypothesis (stating: prices fully reflect all available information in the outside world), if algorithms don't (or hardly) look at the outside world?
This is surely a neglected issue, because I never read about it. Yet the issue is real, and potentially has large consequences - both for the health of financial markets and philosophically.