If you’re busy during the workday, it’s unlikely you are going to take much time choosing where you go to lunch. You just want to get something to fill you up. By contrast if you’re on holiday, then you might take a bit more time about selecting somewhere to eat. You’ve got more time, and being on holiday, you’re more likely to place a premium on choosing a restaurant, where you really enjoy the food. Basically, our objectives are different.
In financial markets, it’s similar, different groups of traders will behave in different ways, given they have different objectives. If we aggregate the flows from various groups of traders, we’ll see different patterns. I recently read a Twitter thread by @bennpeifert, discussing the retail flows, which I found pretty interesting, where he summarised a large number of papers on the topic. The general gist of the datasets discussed in these papers is that retail traders tend to lose money, across a number of different asset classes, including FX. In the past, when I’ve looked at retail FX data, I’ve also found it to be a contrarian indicator.
More broadly, as we might expect, there is information content in being able to decompose flows depending on trader type. We’ve already noted that retail flows tend to be contrarian. One dataset which contains trading flows for FX, decomposed by trader types, is that which is aggregated by CLS, and I’ve written a paper about this. Corporates are trading FX, because they have to, as part of everyday business, as opposed to doing it on a pure speculative basis. For a large pension fund investing in trading foreign bonds and equities, FX is going to be a secondary consideration. They have to trade FX because they have to. Central banks will be trading FX as part of their managing their reserves. By contrast we could argue that hedge funds are likely to be more price sensitive for FX.
Hence, not everyone trading FX is going be equally price sensitive. As a result, we might speculate (and hopefully you can excuse the pun), that if we are a speculator, there is value in tracking FX flows from different types of investors. Furthermore, we can go one step backwards from the actual FX flows, and track the portfolio flows that are likely to result in FX transactions (the firm Exante Data creates realtime estimates of certain capital flows). It is of course a tricky business, given that there are also FX hedging flows. For example, let’s say a pension fund based in USD, buys a significant amount of EUR denominated bonds. This would require selling USD and buying EUR, in order to purchase the EUR denominated asset. As the same time, they can put on a short EUR/USD spot trade to hedge the FX risk, depending on their hedging ratio.
FX is an aggregation of a lot of trading activity where the primary motivation may not be to generate P&L from trading FX. Speculators can use data based on portfolio flows, and also the FX flows, from different groups of traders to help in their decision making process.