Let me start with an anecdote from many years ago. Back in 2008, I was a speaker on a panel at the Milken conference in Los Angeles. The topic was the dollar. I argued that one of the forces driving the dollar was a shift in capital flows. But while I was making the statement, I could see one of the other panelists looking down into the wooden planks of the stage, with an emotion of disagreement and disgust. He started his next set of remarks with a disclaimer, he firmly believed that there is no relationship between asset prices and capital flows.
This is a philosophical debate. You can imagine situations where flows are driving currencies. A big M&A transaction into Canada may drive the Canadian dollar stronger. This would be an example of flow driving price with a positive correlation. But you can also imagine situations where currency movements drive flows. If the Yen is weakening, Japanese investors may be selling foreign assets (buying Yen) to rebalance their portfolios. The correlation and direction of causality would be flipped in this example. This two-way causality in the relationship between flows and prices make it very difficult to establish stable and significant statistical relationships.
Nevertheless, we would argue that certain flows can reasonably be viewed as exogenous relative to a given price. And if you can predict those flows, you also have the ability to predict one price-driving factor.
Linked to this, we would argue that there are times when the relationship between certain asset prices and certain fundamentals (other than flows) changes, because there are (potentially unobserved) flow factors at play. The Euro is an example of that currently. Fundamental models for EURUSD, based on rate differentials, would have predicted a relatively stable EURUSD rate in recent months. But in reality, we have seen a significant Euro appreciation. In our analysis, this ‘Euro residual’ is tied to certain flow forces, which have turned Euro bullish lately.
The bottom line is this. Certain flows should matter conceptually, and even if it is hard to establish simple static statistical links between flows and prices, the effect of the flows may still be working in the background, and show their power in certain circumstances.
We spend a lot of time on tracking capital flows in real time, projecting them forward, and then filtering the signals to focus on those that can reasonably be argued to be exogenous and relevant in a given regime. Our goal is to have a sophisticated framework to cross-border flow analytics that has predictive power for currency rates (and other asset prices). This is easier said than done. But if you do not do this analysis thoroughly and creatively you will miss out on one factor that moves asset prices.
As always, we value a dialogue; with investors who would like to embed flow analysis in their process, as well as with vendors who may have unique flow data available that we can embed in our analytical infrastructure.
Jens NordvigFounder and CEO
Exante Data LLC