Engineering a Fair Market
In class, we’ve talked about markets and matching buyers and sellers, and when doing so, it’s hard to overlook the stock exchange, perhaps one of the most prominent examples. Many buyers and traders are all executing simultaneously, and it’s the job of the stock exchange to find a perfect matching between the two group. However, one element that was not discussed in class was the timing. We’ve performed analyses of auctions knowing all the information and all the bids, but in reality, collecting all of that data would be too slow for all the stock that is being traded. This is why the stock exchange sets a price for the stock that is meant to encapsulate the entire market for that stock, and represent middle-ground between what the seller wants and what the buyer is willing to give. This way, instead of needing to collect bids, we can execute transactions at this price (matching buyer to seller), and adjust the price based on how many people are willing to sell or buy the stock (supply and demand). The participants have upped their game too; instead of humans on a trading floor, transactions are now executed by computers, capable of taking in information and executing transactions much faster and with greater frequency than humans, and now accounting for 50-60% of trades on average (1). Given that these machines can change the state of the market so quickly, it is key to have accurate, up-to-date information on the state of the marker for these “high-frequency traders”.
To this end, many companies have invested in high-speed connections to stock exchanges, particularly the New York Stock Exchange (NYSE). Given that information about the exchange is so vital, even the delay of transmission of the signal in the network can pose a lost opportunity to make money; whoever can act the fastest on new information will have the upper hand. When high-frequency traders are able to take advantage of their speed, they can significantly out-perform human traders, and generate billions in profits. These algorithms can sometimes even know how the market will change before the market does, allowing it to take advantage of rising/falling stock prices by buying/selling before others, respectively (2). To best enable algorithms to act quickly, infrastructure has been set up to allow for better connections to the NYSE. From laser networks (3) to new antennas (4), even the recent development of using experimental cables to shave ever-smaller fractions of seconds off of transmission times (5). To relate this to class material, we could think about the transmission time from a firm to the NYSE as an edge between those two nodes in the complex network that is the stock exchange market. Whoever has the shortest path to the NYSE in this network will have a distinct advantage. While this is great for high-frequency traders, it poses a significant threat to everyday human traders. If algorithms can act quicker and steal all the profits to be had, then the market is functionally limited to just those traders, and is no longer “free”. Because of this, many have advocated for more regulations on high-speed trading, including some who helped found it (6). One company, IEX, has even implemented this regulation on their exchange with a physical “speed bump”, where many miles of coiled cable ensure that the market will have time to update before the signal from high-speed trades can reach them, ensuring that transactions are executed at a fair price (7).
While the initial analysis of the issue talks about matchings in the market and the network of communications, all of which are applicable to the class, I wanted to bring one last insight from a relatively low-level place; the Prisoner’s Dilemma and basic game theory. We can analyze all of these traders as players (although I will only consider two, for the sake of simplicity) trying to get the best payoff. Let’s imagine a very simplified version of the game, where both traders either implement infrastructure to get a better connection (F, for fast), or stick with the connection they have (S, for slow). Let’s say that traders operating at slow speeds get a base payoff of 5 for trading at all. Having a higher speed than the other player adds 5 to your payoff (and subtracts the potential profit of 5 from the others, as you would be taking their potential profit), and the cost of setting up the infrastructure to do so subtracts 1 from that player’s payoff. Based on this, our game would look like this:
Notice how each player has a dominant strategy for acting fast, and having that high-speed connection (they will always get a payoff for doing so, no matter what the other is doing). Therefore, the Nash Equilibrium is when both players act fast (occurring when both act according to their dominant strategy, with is always a best response). However, both players would’ve had a higher payoff if they simply both acted slow, making this a Prisoner’s Dilemma. When both choose to act quickly, they both end up with the same access to the market (and therefore get the same amount as from trading before, assuming that the market is only traders that act like them), but both have to pay the cost of setting up the high-speed infrastructure, resulting in a worse outcome for both traders, as well as depriving the world of those resources spent towards what amounted to no net gain.
Based on this, I would favor some oversight and regulations about how quickly the market can act. The NYSE has circuit breakers in place to halt the market if prices drop too much, in order to prevent panic/mass losses (8), but I would propose circuit breakers for speed of transactions; transactions should only be allowed to happen at a certain rate. This not only removes the incentive for high-speed connections (preventing the Prisoner’s Dilemma from occurring), but also gives everyday human traders more of a chance, shifting the stock exchange towards a more free market where anyone can successfully participate. In preventing only those with access to high-frequency traders from being successful, we can decrease the wealth gap and provide more opportunities for everyone to have financial success.
References:
1) https://money.cnn.com/2018/02/06/investing/wall-street-computers-program-trading/index.html
2) https://www.nytimes.com/2009/07/24/business/24trading.html
3) https://www.wsj.com/articles/highspeed-stock-traders-turn-to-laser-beams-1392175358
4) https://www.wsj.com/articles/nyse-antennas-spark-high-speed-trader-backlash-11565272102
7) https://exchange.iex.io/about/speed-bump/
8) https://www.nyse.com/publicdocs/nyse/NYSE_MWCB_FAQ.pdf