Thu Feb 17, 2005

Beat the Market

My friend Pierre-Olivier Gourinchas has co-authored a very interesting paper with Hélène Rey called “International Financial Adjustment.” (Here’s the PDF version.) You might think that’s not a title to set the world on fire, but don’t be fooled. A more appealing—though perhaps less responsible—alternative would be something like “Dude! We can predict exchange rates!” Here’s the abstract:

The paper proposes a unified framework to study the dynamics of net foreign assets and exchange rate movements. We show that deteriorations in a country’s net exports or net foreign asset position have to be matched either by future net export growth (trade adjustment channel) or by future increases in the returns of the net foreign asset portfolio (hitherto unexplored financial adjustment channel). Using a newly constructed data set on US gross foreign positions, we find that stabilizing valuation effects contribute as much as 31% of the external adjustment. Our theory also has asset pricing implications. Deviations from trend of the ratio of net exports to net foreign assets predict net foreign asset portfolio returns one quarter to two years ahead and net exports at longer horizons. The exchange rate affects the trade balance and the valuation of net foreign assets. It is forecastable in and out of sample at one quarter and beyond. A one standard deviation decrease of the ratio of net exports to net foreign assets predicts an annualized 4% depreciation of the exchange rate over the next quarter.

Now, I am not a macroeconomist so I should leave further discussion to Daniel and John. The guts of the paper are really beyond my competence to evaluate. But this is a blog, so naturally I will carry on regardless and make three points anyway.

The first point is that the paper is interesting because their result is a real surprise—especially that bit about “forecastable in and out of sample at one quarter and beyond.” You shouldn’t be able to reliably predict rates better than a random walk, and certainly not over relatively long periods. How do Gourinchas and Rey do it? They begin with the observation that if a country’s current account balance is in deficit, it must eventually get readjusted through increased exports or by depreciation of the currency. They argue that the “sharp increase in gross cross-holdings of foreign assets and liabilities” between countries that we’ve seen over the past twenty years introduces a significant pathway by which rebalancing can happen without changes in trade: “Put simply, a fall in today’s net exports or in today’s net external asset position has to be matched either by future net export growth or by future increases in the returns of the net foreign asset portfolio … The budget constraint implies that today’s current external imbalances must predict, either future export growth or future movements in returns of the net foreign asset portfolio, or both.”

They go on to derive a quantity—”the deviation from trend of the ratio of net exports to net foreign assets”—that captures movements in the external asset position. They decompose it into the bit explained by trade and the bit explained by financial adjustment. Then they collect a longitudinal dataset for the U.S. case including a measurement of this quantity and show that the financial adjustment component accounts for about a third of total readjustment, mainly in the relatively short-term. An implication of the theory is that this quantity should predict future returns on a country’s portfolio of foreign assets, and this turns out to true also.

Now, the rate of return on assets like this is determined partly what they earn abroad and partly by what the exchange rate is. The fact that the financial component of the rebalancing itself works through the mechanism of depreciation leads Gourinchas and Rey to “raise an obvious and tantalizing question: could it be that the predictability in the dollar return on gross assets arises from predictability in the exchange rate?” Again, the answer is yes:

Overall, these results are striking. Traditional models of exchange rate determination fare particularly badly at the quarterly-yearly frequencies. Our approach … finds predictability at these horizons. … a large ratio of net exports to net foreign assets predicts a subsequent appreciation of the dollar, which generates a capital loss on foreign assets.

Et voila, a theoretically grounded, empirically applicable technique for predicting movements in exchange rates.

My second point is about what happens next, assuming that Gourinchas and Rey are right. As they say themselves in their conclusion,

The challenge consists in constructing models with fully-fledged optimizing behavior compatible with the patterns we have uncovered in the data. A natural question arises as to why the rest of the world would finance the US current account deficit and hold US assets, knowing that those assets will under-perform. In the absence of such [a] model, one should be cautious about any policy seeking to exploit the valuation channel since to operate, it requires that foreigners be willing to accumulate further holdings of (depreciating) dollar denominated assets.

If rational agents knew about the relationship between external imbalances, foreign asset return and exchange rates that the paper describes, then we wouldn’t expect to find the patterns in the data that we do. If the rest of the world knew their assets were going to under-perform, they wouldn’t have made that particular investment. Now that this relationship has been uncovered, we might expect this gap to close as the knowledge is incorporated into investment decisions. It’s in the nature of economics to prove itself both true and false, in this way. True, because the specific bit of pricing technology uncovered by the paper is now grist for the decision models and pricing mechanisms used to make investment decisions (“Our theory also has asset pricing implications”, as the abstract says). So the rationality of market agents will more closely approximate the ideal state that theory says it ought to. But false, too, in the sense that the empirical relationship demonstrated in the paper might disappear as a result, and everything gravitate back to an efficient-market random walk again.

My last point follows naturally from the previous one: If Pierre is my friend, then why didn’t he let me know about all of this, oh, say, 18 months or so ago. Then I could have formed a CT investment consortium managed (for a small fee) by John and Daniel. This could have taken care of CT’s hosting bills for ever and ever, and I could be writing this from a private island in the Pacific ocean instead of a Starbucks in SeaTac airport.