saved articles

launched by the Kremlin after months of rising tensions, quickly generated a flood of casualties and several million refugees, as well as bringing the senseless destruction of cities and towns. A negotiated peace may yet bring it to an end. But amid the continued bombardment of Ukrainian cities by Russian artillery and the ramping up of Western military aid to Ukraine, the possibility remains that the war will continue. With that, the odds of a wider conflagration involving several nuclear-armed states would shorten alarmingly. While it is not yet clear how the war will unfold, the world stands at the threshold of a troubled new period. What follows is an attempt to sketch out the historical matrix from which the present conflict developed, and to identify the possible scenarios that lie ahead. 1 The Kremlin bears the responsibility for unleashing this war, and regardless of the outcome will carry a heavy moral burden for the destruction it has already caused. Amid a broad surge of sy...

This paper develops a formal framework to assess policies of learning algorithms in economic games. We investigate whether reinforcement- learning agents with collusive pricing policies can successfully extrapolate collusive behavior from training to the market. We find that in testing environments collusion consistently breaks down. Instead, we observe static Nash play. We then show that restricting algorithms’ strategy space can make algorithmic collusion robust, because it limits overfitting to rival strategies. Our findings suggest that policy-makers should focus on firm behavior aimed at coordinating algorithm design in order to make collusive policies robust. 1 Introduction Software systems that take over pricing decisions are becoming widespread. Pricing algorithms can allow firms to monitor and process large amounts of data and adjust prices quickly to changing circumstances. The ascent of such systems poses a potential challenge for the current regulatory landscape: pricing al...

middle of the 2000s, the scale and persistence of hedge funds’ success was transforming the structure of the industry. The first generation of hedge-fund titans had been seen as freakish geniuses, whose eye-popping returns were possibly lucky and certainly not reproducible. But by 2005 nobody could argue that hedge funds were exceptional in any way: More than eight thousand had sprouted, and the long track records of the established funds made it hard to dismiss their enviable returns as the products of good fortune. Bit by bit, the old talk of luck and genius faded and the new lingo took its place—at hedge-fund conferences from Phoenix to Monaco, a host of consultants and gurus held forth about the scientific product they called alpha. The great thing about alpha was that it could be explained: Strategies such as Tom Steyer’s merger arbitrage or D. E. Shaw’s statistical arbitrage delivered uncorrelated, market-beating profits in a way that could be understood, replicated, and manufact...