The following is an article (with some minor modifications and added content) that is presented in our November FX Monthly, hopefully to appear on these pages very soon.
An attempt at a Carry Trade Model
The USD index recently touched a new low for the year as global equities touched new highs for the year as well. Our fundamental view remains that risk markets are merely driven by the liquidity infusion from central banks and that the “real” recovery ahead will pale in comparison with current expectations. There is a circular logic/reflexivity at work here: easy money policy and stimulus does lead to economic stabilization and an appreciation of some assets due to shrinking risk spreads from the bailout mentality. The recovery in asset prices itself becomes a driver for more economic activity and round and round. The problem is that the price paid at the input to this circular model is a dangerous, unsustainable one: endless piles of fresh public debt that represent the creation of money out of thin air that must be paid by future generations of taxpayers.
Still, markets in a bubble-like, self-reinforcing upward spiral can continue their defiance of gravity for surprisingly long periods of time and long ignore all attempts to disprove the market’s fragile logic with rational contrarian indicators or theses. This leads us to ask the question: what signposts should we look for in the market that can warn us that the bubble is in danger of popping? In other words, are there any good “distant early warning” indicators that will help us put a finger on when the odds are beginning to shift in favour of a more rational appraisal of this upswing in asset markets and risk appetite in general? All cycles change eventually, after all.
We feel the answer to the above question is yes. And since the single greatest input into the force driving the markets at present is “risk willingness”, then creating a model that measures the strength of risk willingness relative to actual moves in the currency market, which has been dominated by the axis of risk over the last several years, should help us understand when the odds are shifting in favour of an unwinding of the USD carry trade bubble. Of course, risk willingness is driven by liquidity, and measures of liquidity and risk willingness are often interchangeable – as is evident from the inputs to the carry trade model we present below.
The Carry Trade Index
We created a model with multiple, evenly weighted inputs that is shown in the graph below. The model inputs include the following:
- Emerging Market to Government Spreads: this is a measure of the spread between emerging market debt and the US 10-year benchmark.
- Junk Bond to US Government Spread: a measure of the spread between long maturity US corporate bonds rated by Moody’s as BAA or worse and the US 30-year yield.
- FX Volatility: uses a measure of FX volatility created by Deutsche bank that is more or less weighted according to the liquidity in the major dollar pairs and the most common crosses in the G7 currencies.
- VIX: the most common measure of stock market volatility based on the implied volatility in the options market. It tends to rise the most in down markets due to fear.
- CB 1-year forward: this is an average of the 12-month forward expectations from the G10 central banks
- CDS X-over: a measure of CDS prices for riskier European companies, and one of the most widely followed indices of corporate credit.
Model methodology
The model simply measures day to day changes in the various components and the standard deviation of those changes relative to historical moves. We also smoothe the data. Taken together, the components offer a very broad measure of risk willingness in the market and the higher the risk willingness, the more it tends to support the carry trade, as high risk appetite means that market participant will gravitate to the fastest growing, highest yielding investments. In the world of currencies, this means the highest yielding currencies vs. the lowest yielding currencies. But carry trades only tend to succeed as long as risk willingness is not just high, but actually "expanding", and that is what we are trying to show here as the model asks whether risk appetite is expanding or contracting, not what its “level” is at any given time.
In the chart, the strong blue line shows the overall Carry Trade Index (right axis), essentially representing the average standard deviation of the component indicators. The strong red line is a sample carry trade created from a basket of 7 high yielding and EM currencies (AUD, NZD, BRL, PLN, TRY, MXN, IDR) vs. the USD and includes the carry from interest rate differentials in the calculation of its value. Generally, one would expect pressure on risk willingness whenever the carry trade index drops below zero for any length of time, and especially if it continues lower as this points to a contraction in risk appetite.
Carry trade conditions should be considered positive when the index is in expansive territory above zero. Note how the direction of the index leads the results of the sample carry trade at the key turning points and also note that the index crossing back above zero this spring coincided with the USD carry trade beginning its resolute march higher starting in March. In an early example of a contraction in risk appetite tipping its hat on the carry trade, note that as the USD carry trade was topping out the last time in the summer of 2008, the carry trade index barely nudged above 0 – far below the previous peak and therefore a divergent top - and began dropping persistently thereafter.
As the carry trade was bottoming temporarily in late November, on the other hand, the carry trade index was less negative than previously. This was true to an even more pronounced degree in early March. As equities were bottoming out and keeping the carry trade in the dumps, the risk inputs into our carry trade model were edging up above zero and providing a strong contrarian signal to the worry evident elsewhere. Another thing to note is that the performance of the index was not as closely correlated with the USD carry trade in early 2008 and even a bit later because there was still more focus on the JPY carry trade at that point. The carry trade is a constantly shifting thing and going back too far in history with any fixed carry trade basket is a bit dangerous. At present, however, the USD carry trade is certainly the one most in focus.
What is the model telling us now?
The model is telling us that risk appetite is still expansive, if less so than when it was peaking a few months ago in the early summer. This makes sense if we look at a few of the inputs to the model – indicators like FX volatility, the VIX and fixed income spreads have flattened out in recent months. Yet as long as we remain above zero, the conditions remain nominally favourable for carry trades. The only warning sign is a the divergence in the index vs. the actual performance of the carry trade, suggesting that gains in risk markets may be a bit harder and slower to come by. So for us to see a more sustained risk negative scenario, we would definitely like to see the Carry Trade Index dip into negative territory as a leading indicator.
One interesting development of note: sovereign and corporate CDS prices have risen lately, and credit was a cycle leader the last time around, so this could be a distant early warning of a turnaround in risk. The CB expectations component has also dropped sharply of late. Still, past history suggests that a full scale move of the index below zero is needed as a coincident indicator to suggest that any sell-off in risk is potentiall turning into a new trend, as opposed to a simple consolidation of the dominant trend higher in risk appetite.
We will continue to provide snapshots of the index in the future on this space.