Monday, November 28, 2016

Macro | How Sustainable is the US Dollar Rally

The dollar rally that started since the conclusion of US presidential election shows not much signs of abatement. 2014 was the year of crude oil, when the fantastic sell-off in crude set much of the moods prevailing in the world economy - from equities to rates. In 2015, this slowly turned over to dollar, partly through the surprise CNY depreciation, and later through Fed rate hike expectation. And without any doubt, despite all the noises around the Brexit and Italian referendum votes (early Dec), the dominating factor for risks is again US dollars. The figure below shows a quantitative look at the cross market risk drivers using Minimum Spanning Tree methodology (based on correlation). It shows clearly the dollar is in the center of the cross market driving force. A very similar situation to what we had back in 2013, but more concentrated role for the dollars to set the market sentiments. We already had a not-so-quite riot in rates following the dollar strength post election. The emerging market currencies and equities have taken significant beatings. And top houses are calling for this dollar strength to be one of the top trades in 2017. Naturally it begs the question how much leg is still left in this dollar rally.

To take a long-term look, the dollar rally is by no means extreme. In nominal terms the broad-based trade weighted dollar index is near its historical highs. However, when we compensate for the inflation differentials between the US and its trade partners, the rally is well within the historical range (still 13% off from the 2002 peak) - as the chart below shows. Note the rally in dollar since 2014 has been almost equal for the real and nominal exchange rate - 17% vs 20%.

But while this rally may not be extreme, it may not be sustainable either. A large part of the recent dollar strength has been on the back of expectation of US policy change, specifically a possible fiscal stimulus. The economic argument behind is that a fiscal stimulus, coupled with a budget deficit will increase interest rates and hence the exchange rate. In fact this is what was observed during early 1980s in the US (although it has not much support from observations in other non-US advanced economies). The actual mechanism is far from clear. There are extensive studies on budget deficit reduction and its impact on exchange rates, but reverse studies are rare. Theoretically, the direct impact (of increasing budget deficit) goes through the interest rate and asset return channel above and lead to a higher exchange rate (as demand for higher interest assets goes up among foreigners). On the other hand, increased budget deficit can increase the long term inflation expectation and hence expectation of future dollar depreciation. The second part of the policy is trade - which is basically a tightening pressure on the US current account deficit - if President-elect Trump follows though his promises. Typically for the US the current account in recent history has been driven to a large extent by demand for financials assets from overseas investors. This means a tightening of current account will have to be matched by reduced demand for US financial assets by foreign investors, resulting in a currency depreciation now (and possibly an appreciation later). In fact the post-crisis dollar weakness has resulted in a significant tightening of current account for the US already. A further sustained tightening in general may not be great for either the US or global economy. The other possible factors, i.e. the overall demand (or GDP differential) or real rate differential with major trading partners are relatively straightforward, an increase in both leading to a stronger dollar.

Looking in to the above set of arguments empirically, we run a quick vector auto-regression estimates with real dollar exchange rate, real rate differential, current account (% GDP), budget balance (% GDP) and GDP differential as endogenous variables (differential with GDP-weighted Euro area and Japan data representing rest of the world). The results are as shown in terms of impulse response - i.e. response of dollar real exchange rate for unit positive move in budget balance (FD), current account balance (ca), real rates differential (rates) and GDP differential (GDP). It seems at least for our data (spanning 1995 to 2014, quarterly), Trumps policy of budget deficit (negative fd) and tightening current account (positive ca) has off-setting impact for dollar real exchange rate. In fact there are good chances the current account tightening impact (negative for dollar in near term, positive long term) can overwhelm. An increase interest rates may make US assets attractive among foreign investors, but without matching trades the flows in to those assets will be difficult to sustain. Among other drivers, while inflation in US has been steady, including wage growth, we have seen some early signs of a come back of inflation in the Euro area. The main thing to look out there is the pick up in Euro area credit growth after a stall start of this year.

Given this ambiguous impact of policy, and hopefully a declining need for policy divergence and a head-room for trade-weighted dollars of only ~13% to reach all time high in real terms, it does not look like the dollar rally has much room left. one of the surprise trigger can come from ECB and/or BoJ in December, with QE in Europe still priced in. And the Dec Fed hike - which is almost a certainty now - will act to defuse this rally. 

Interestingly, while from the emerging market point of view, the recent dollar rally was kind of risk-off, it was hardly so for Euro. Euro - which has lately became a funding currency like the Yen, sold off steeply. Arguably there was not much positioning to blame either, so this makes it a very interesting move. The Euro area as a whole has accumulated a huge current account surplus in its glut for savings in the post-crisis period. A substantial change in trade relationship with the US may start to unravel that. If you are positioning for the consensus Euro dollar parity, think again. 2017 may see a major reversal in Euro instead dollar.

Note: all data from the St Louis Fed FRED database.

Saturday, November 5, 2016

Markets | The Trump Trade?

Keynesian beauty contest is an interesting concept that shows a group of perfectly rational agents trying to predict the outcome of an event may not converge on the most expected case, provided their risk and reward depends on what most others think. I think something similar happens in the markets around a big event. Rarely it is clear what are the implications of different possible outcomes of such events. In such a scenario, a trader's immediate pay-off depends on how good he is at predicting market reaction (as opposed to the actual implications). As a result collectively the market ends up reacting in some ways that very few people may actually believe.

Next week's presidential election is such an event. There are strong evidences that economy has significant impact on election outcomes. however the reverse result is very weak if any. Performance of a large globally connected economy depends on more things beyond the control of the Oval Office than we give credit for. However the markets seem to have already formed an opinion and trading according to the poll results in recent week. This is not only the US market but across the globe. The common denominator is an expectation of underperformance with a republican win.

The consensus is more or less a status quo with a democratic victory and large uncertain changes with the republic candidate in office. Honestly, I think it is too early to say what will be the policy changes as we hardly have any clue on specific policies apart from election promises. For example it is usually considered republican victory will be good for defense stocks. However if Mr Trump carries out his promise on cutting down on NATO, will that necessarily be the case? He promised to unwind trade agreements. But sure there will be something to replace it, will that be very different than the existing one, and will that have really any significant impact on trades, prices and job? Or may be you should buy Apple? - he is sure to threaten EU to withdraw the taxation case and make America great again! My personal take is Mr Trump promised things, but post election (if he wins) it will be hard to deliver on many of them except in a much run-down version. Hence in the base case, sooner than later, we focus back on things like earnings and economy and inflation once the initial reaction is over.

However, the market is close to pricing in a crash scenario for an outcome favoring the republican candidate. The VIX (and volatility of VIX) are tad shy of last August peaks. The implied skew and (near-month) implied correlation in S&P 500 are racing sky-wards (and interestingly with a quite flat vol convexity, i.e. high skew and very low smile). There is a high amount of uncertainty. 

And if you are planning to take decision (being flat is one of them), I have already written about how to generally think about positioning under uncertainties before. If you are a hedger, you know what you need to do - that's quite it. And if you are a speculator, after all the analyses and mumbo-jumbo, basically you have to choose a side (rally or sell-off) and stick to it. And the only things that matter are:
  1. what is your expectation and how that looks from risk-reward already priced in the markets and 
  2. How to optimize your responses in case you are wrong.

The first one is commonly understood. At present the markets are definitely pricing a large sell-off. This is in the background of decent economic news and improving global PMIs. Technically most markets across the globe has or on the verge of confirming a bearish signal (see chart above). The asymmetric pricing in the downside suggests there are large price move expected, but at the same time it makes the risk-reward unattractive compared to the upside. And based on the past history in S&P which has breached a technical support recently, the distribution of near term returns favors the upside statistically (albeit with a rather large uncertainty spread around that). The chart shows the historical price distribution after such technical breaches (categorized in to three types of technical formation - megaphone, triangle and channel, and whether the existing trend was ascending or descending, and also if the breach is of resistance (up) or support (down))1. We are in a down breach within an ascending megaphone (see the figure above).

As far as the second point is concerned, if you are positioning for downside and it turns out wrong, your responses are limited if you assume it will be a relief rally, (not a sustained one). Alternatively, if you are positioning for the upside and if you are wrong, you will have plenty of opportunities to react. We will sure enter a period of high volatility and there will be plenty of trading opportunities.

So it appears purely based on the second criteria, a long risk positioning is preferred2. Of course this assumes the outcomes are fairly priced from criteria one and you do not have any strong view on either outcome.

Note: 1) This is based on systematic technical analysis, for details see here, for code page go here. You can select or de-select series on this interactive chart
2) this is not an investment or trading advice, do your own due diligence, form your own opinion. See the disclaimer page.