Two scenarios that are fun to consider:
- Either the bots will be programmed by humans to trade like humans (ie, the bots would take trades based on recognizable chart patterns). In this scenario the markets will still move with a high degree of "emotion"/irrationality. (ie. nothing really changes)
- The bots are programmed to take trades based on data/fundamentals. In this scenario, in an all-bot marketplace, market price discovery would be a highly efficient process. the bots would process all publicly available information (nearly instantaneously), and determine a fair price. There would be practically no volume until new data was discovered that caused the bots to adjust their valuations.
On a side note, the vast majority of money in the markets today is speculative (something close to 85 percent I've heard). This money is trying to capture value from pricing inefficiencies. If the inefficiencies disappear, will investors pull their money from the markets, compounding the low volume I mentioned in scenario 2? How much risk will this low volume create?
So all that said, I think an all-bot marketplace would force the capital markets to take on a much different (smaller) role in our economy... and would be boring :)
That's what I am thinking. Most people with bot/AI/algo trading POV think about building the best bot and are pretty myopic in their perspective. But what happens after the bot's change the actual landscape and the system adapts around them?
There's an an analogue to that in passive vs active mgmt. There's been huge inflows to the former and everyone touts it as superior due to lower fees. But how well does a passive fund perform in an all passive universe?