We think that Emerging Markets debt, which together with some other asset classes have acted as de facto substitutes for developed market government bonds (US, Germany, and Japan) over the last eight years should still experience a much larger correction. In this note, we attempt to explain the three key sets of reasons for this view, namely: valuations, fundamentals, and positioning.
The market consensus and why we disagree
As the dark blue line in Figure 1 below shows, the weighted average current account balance as a share of GDP for countries comprising the EM local bond index (JPM GBI-EM GD) has now returned to balance, which is in marked contrast to the deficits of the 2012-2013 period just after Chinese growth had started to decelerate. At the same time, the valuations of currencies that are part of the EM local bond index vs the USD (light blue line) would appear to be as cheap as they were at the worst point of the global financial crisis, albeit while only fairly valued vs. the EUR. As such, the prevailing market consensus is that EM debt valuations and fundamentals remain attractive and continue to merit significant allocations.
Figure 1. Current Account / GDP and valuations for EM countries / currencies in local bond index
Source: Bloomberg and GLG database as of September 29, 2017.
However, we think there are some problems with the above arguments. Firstly, if EM currencies were indeed as cheap as the graph above would seem to suggest, then one would have expected the current account balance for these countries to continue improving, particularly due to the meaningful improvement in terms of trade experienced by the larger countries in the index. However, what we have seen over the last one and a half years is an initial stabilization of current accounts and, more recently, a rolling over in (the current accounts) of many individual countries, which we think points to a deterioration going forward.
Our thesis is that the tightening of monetary conditions that we have started to see in the US and we feel we are about to see in Europe, which we will elaborate further on later, may soon slow EM capital account inflows (in fact, we think that we could actually see outflows), which could require many of these countries to return to current account surpluses akin to the early 2000s. This could require weaker real exchange rates, which in an environment of accelerating inflationary pressures in many EM countries, may entail nominal depreciation.
US inflation heading to 3%?
Figure 2 below shows the lagged evolution of underlying US inflation (yellow line) and US core inflation on a twelve-month rolling basis. The underlying inflation gauge is an indicator developed by the US Federal Reserve that takes into account goods and services prices as well as other aggregates usually correlated to future inflation, such a money supply, credit spreads, and equity valuations. The underlying inflation gauge in Figure 2 also suggests that US core inflation could accelerate to 3% p.a. over the next eighteen months. With the US unemployment level at 4.1%1 and underemployment at levels close to or below the bottom in the two previous cycles, we feel it is difficult to believe that, with the current pace of payroll generation, the underemployment rate will not continue to decrease, and sooner or later US inflation could accelerate substantially.
Figure 2. Underlying US inflation gauge (full data set lagged 18 months) vs Core US CPI yoy
Figure 3. US unemployment rate, underemployment rate, and recession periods
Source: Bureau of Labor Statistics / Haver Analytics.
We think the impact of inflation and employment trends on US monetary conditions will be further complicated by Treasury issuance dynamics. As shown in Figure 4 below, since mid-2016, the US Treasury increasingly replaced the net issuance of longer term Treasury notes and bonds by T-bills, which reduced upward pressures on Treasury yields. In 2018 however, the US Treasury has given indications that it will significantly increase longer term Treasury note and bond issuance to help fund (at least partially) the expected increase in the fiscal deficit stemming from lower US corporate tax rates. Indeed, from approximately USD400bn of net new issuance of coupon-bearing instruments in the twelve months ended in the third week of January 20182, we think net new Treasury Notes and Bond issuance may grow by close to USD700bn for all of 2018, in order to help fund the fiscal deficit and the Fed’s disinvestment of Treasury coupons under its balance sheet normalization program. In addition, the US Treasury could also issue around a net USD450bn in Treasury bills, which would take net issuance to a new record level. The precise timing of the composition and acceleration in issuance is dependent on the approval of a new debt ceiling, which we believe could be resolved in February.
Figure 4. US Treasury debt issuance patterns
Source: BPD / H, SIFMA / Haver.
The final element, which we think may produce a large change in the demand for risk assets in 2018, is the change in the balance sheet positioning of the three largest developed market central banks. In particular, as Figure 5 below shows, the Fed, ECB, and BOJ balance sheets could go from an expansion of approximately USD2 trillion in 2017, to just USD126bn in 2018, followed by an outright contraction in 2019. It should be easy to see that the main driver of this would be the combination of the Fed’s balance sheet contraction, with the ECB ending its bond buying program (we expect it by September 2018), and we find it difficult to believe that when such a large non-economic buyer stops intervening there will be no market impact. If the ECB gets out of the picture, new marginal capital flows could start to gravitate back to the high quality European government bond curves, and stop going into substitute asset classes, thereby potentially increasing the pressure on risk assets.
Can the extended positioning endure?
We believe it is difficult for positioning to get much more extended than at present. Moving from a macro to a more micro perspective, in Figure 5 we illustrate three points that we feel are important for investors to keep in mind. First, we believe it is easy to conclude that a significant part of the rally in equity prices, proxied by the green line representing the S&P 500, was linked to the expansion of the balance sheet of the three largest global central banks (the Fed, BoJ, and ECB) since 2008. Secondly, when looking at EM credit spreads, represented by the purple line where spreads are shown inverted (i.e. when the line falls this means that spreads are widening and vice versa), there appears to be a clear correlation between EM credit spreads and equities, particularly from 2008 onward. The quandary, for EM bond investors, is that theoretically equity valuations could continue to going up infinitely, but credit spreads cannot logically go beyond zero, which is the point where we now seem to be converging to. Finally, the same Figure also shows that these extreme valuations, for spreads at least, are occurring when margin debt / US GDP have reached levels that in the past have occurred just before significant market corrections. Even after the equity market sell-off in February 2nd and 5th, the adjustment experienced by the equity market is small in our view and financial conditions have not tightened meaningfully.
Figure 4. NYSE margin debt / GDP to combine with CB balance sheet reduction to bring down risk assets
When we consider positioning specifically in EM, then we see that this is also quite extended both in local and hard currency terms. As Figure 6 below shows, speculative positioning at the Chicago Mercantile Exchange in a basket of five currencies that has historically been highly correlated with currency performance for the local bond index is now at levels that were only seen before in 2007, June 2011, and early 2013, and in each of these cases very deep corrections followed soon after.
Figure 6. Speculative positioning at CME / currency performance in GBI-EM GD index
When we consider what has happened with EM real money managers we see similar worrying patterns both in local and hard currency terms. Figure 7 below shows that in the first week of 2016, when the index returned -1.8%, only 55% of the 20 largest EM local currency mutual funds were overweight beta even though EM currency valuations where at the lows that we had seen only in the early 2000s. A year later, in the first week of 2017 we can see that 65% of managers had higher beta than the benchmark, and finally in the first week of 2018 we see that 80% of the managers were running overweights.
Figure 7. Local Managers Universe: Out / Under performance in first week of the year across universe of local currency managers
|12/31/15 - 1/8/16||Relative to benchmark|
|Average of Outperformers||0.22%|
|Average of Out / Underperformers||-0.06%|
|12/30/16 - 1/6/17||Relative to benchmark|
|Average of Outperformers||0.18%|
|Average of Out / Underperformers||0.08%|
|12/29/17 - 1/8/18||Relative to benchmark|
|Average of Outperformers||0.30%|
|Average of Out / Underperformers||0.22%|
Even though the Figure above shows clearly pro-risk money manager positioning, we believe the more troubling picture is the one for EM hard currency managers show in Figure 8 below – this shows that notably by the first week of 2018, ALL managers were overweight beta.
Figure 8. External managers universe: Out / Under performance in first week of the year across universe of Hard Currency managers
|12/30/16 - 1/10/17||Relative to benchmark|
|Average of Outperformers||0.25%|
|Average of Out / Underperformers||-0.08%|
|12/29/17 - 1/10/18||Relative to benchmark|
|Average of Outperformers||0.31%|
|Average of Out / Underperformers||0.31%|
Source: Man Group database.
To further complicate matters, such extended market positioning is in evidence when valuations are arguably not adequately compensating for risk in our view. Indeed as Figure 9 below shows, current credit spreads (ex-Venezuela) are today at levels that we last saw right before US tapering in 2013. However, even back then, the credit quality of EM assets was substantially stronger than today as indicated by the investment grade component decreasing in size from 63% to 53% and the “NEXTGEN” or “Frontier” component more than doubling its share of the EMBI Global Index – a dynamic, which by the way, is even more marked in the case of the EMBIG Diversified index, which is quite widely used by institutional investors. Also worth pointing out here is that “NEXTGEN” or “Frontier” in the last few years has become synonymous with “issuer(s) for which the main argument in favour of holding the debt is that they do not have much coming due, and they may therefore be less likely to default on coupon payments.” In an environment with tighter monetary conditions these countries could become particularly stressed.
Figure 9. Normalized EMBIG z-spread ex Venezuela (bps)
Source: Bloomberg, J.P. Morgan.
In conclusion, we believe that the current constellation of extended valuations, crowded positioning, and weak EM debt fundamentals compared to historical averages, justify us continuing to manage our portfolios extremely defensively as we have done for the last six months – a period over which we have also become increasingly cautious. Our base expectation is that we should be able to redeploy capital by buying the healthier components of our universe if the market has adjusted and if the various factors outlined above have reached more normal levels.
1. Source: Haver Analytics.
2. Source: Haver Analytics.