In his just released piece, Bob Janjuah's partner at Nomura, Kevin Gaynor, makes some quite profound and very contrarian observations on correlations, a topic discussed extensively on Zero Hedge in the past year. While the prevailing thought is that recently cross-asset correlations have actually dropped (in some cases to record levels), the truth is quite different: "While our colleagues in the Macro Strategy team have made a cogent case that price action in several markets reveals a more discerning behaviour and reduction in observed correlations, Bob and I have been coming around to a slightly different view. Many clients with whom we have spoken over recent weeks are becoming aware of the rather narrow sources of market drivers (two we would argue) and consequent similarities in terms of themes that have driven individual asset classes. It logically follows that anything changing the actual or expected state of those market drivers will have an impact on market returns across a range of risk types and geographies. More to the point, given the nature of those themes (mostly one way), we must be aware of the possibility for a non-linear response to linear changes in those market drivers. That's a fancy way of saying that correlation risk is actually rising in our opinion, as two themes appear to be dominating markets – western liquidity injections without leverage or EM FX appreciation and EM as the source of marginal final demand. These two potent forces have come to dominate the return environment. Consider leadership in DM equity indices since March 2009; it is basic materials and industrials and more lately oil and gas. Ex those sectors, western stock market returns would look rather more threadbare. But perhaps more the point in terms of the non-linearity issue is that the beta of the major indices to these sectors has naturally risen over the past 2 years. Whereas in the past, one had to broadly get financials correct to have a decent stab at calling equity returns (a gross oversimplification I know), it now seems to be the CRB sectors you need to get right." The follow through of all this, and it can be read below, is that the 30 year "great moderation" is rapidly ending and the inflationary threat is now very close, and would be EM driven. At that point none of the Fed's emergency tightening policies, no matter what Alan Blinder's textbook says, would have any impact whatsoever.
The Sceptical Strategist: Correlation Collision Part 2, From Kevin Gaynor of Nomura
Strategists love a good tale. The best ones are those that have a beginning, middle and end. Bob and I have been fairly convinced about our understanding of the basic framework of what has been and is driving returns (that's the beginning), felt we had a view on what would happen to those drivers (the middle), but had been missing a key component in the story (the end). We think we've put the final piece of the puzzle in place and as this piece's title hints, it has to do with high potential cross asset return correlations, rather reminiscent of the 2007 and 2008 (some may recall our publication around that time on this issue) period except this time without the high level of private sector leverage. We feel that the recent wobble is a cracking good example of how this works and it deserves serious thought and scrutiny.
My last note set out a framework for thinking about asset allocation from the perspective of EM vs DM growth in which I argued that the two “regions” are still intimately related from a dynamic point of view. Faster domestic demand in the West would, I argued, result in faster aggregate growth in EM, and slower EM capital spending would result in slower aggregate growth in DM.
The question as to whether we ended up in “slower/slower” or “faster/faster” would be answered by how EM policy makers decided to react to a worsening trade off between price increases and growth. Bob and I at the time expected EM policy makers to respond to higher inflation via tightening policy, with a soft landing being the target. This led us to think of the “slower/slower” space as our central case. While we thought this scenario would have generated a correction in markets, it would be a correction one should be looking to buy into reasonably early on, since we felt it would take some of the steam out of the global economy and set western consumption up for continued steady improvement and ease overheating pressures in EM.
Events moved on though during the spring which changed the calculus for politicians in EMs. If we crudely put the policy choice as either higher unemployment or higher food prices, it seemed that the lesson learned from the Maghreb and Gulf was: attempt to accommodate price increases through subsidies and direct price setting in order to minimise the severity of that unpleasant policy choice. In fact, some policy makers seemed to have been comfortable pointing the finger at the Federal Reserve as the primary source of the worsening trade off between growth and inflation, and to underplay domestically generated demand/supply imbalances. In the absence of an increasingly politicised move in currency regime, something had to give.
In the context of our “slower/slower? vs "faster/faster? debate, this put the dénouement back in time but amounted to the “faster/faster” outcome in terms of asset prices. To remind you, that outcome indicated higher risk assets, further quite substantial increases in commodity prices, higher BEI, curve flatteners and USD index weakness and extensions of carry trades, and volatility declines.
In other words that outcome favoured more of the same but at a faster pace. The attendant risk of a subsequent hard landing in markets and the economy would be that much higher in that scenario. We need to discuss why.
While our colleagues in the Macro Strategy team have made a cogent case that price action in several markets reveals a more discerning behaviour and reduction in observed correlations, Bob and I have been coming around to a slightly different view.
Many clients with whom we have spoken over recent weeks are becoming aware of the rather narrow sources of market drivers (two we would argue) and consequent similarities in terms of themes that have driven individual asset classes. It logically follows that anything changing the actual or expected state of those market drivers will have an impact on market returns across a range of risk types and geographies. More to the point, given the nature of those themes (mostly one way), we must be aware of the possibility for a non-linear response to linear changes in those market drivers.
That's a fancy way of saying that correlation risk is actually rising in our opinion, as two themes appear to be dominating markets – western liquidity injections without leverage or EM FX appreciation and EM as the source of marginal final demand. These two potent forces have come to dominate the return environment. Consider leadership in DM equity indices since March 2009; it is basic materials and industrials and more lately oil and gas. Ex those sectors, western stock market returns would look rather more threadbare. But perhaps more the point in terms of the non-linearity issue is that the beta of the major indices to these sectors has naturally risen over the past 2 years. Whereas in the past, one had to broadly get financials correct to have a decent stab at calling equity returns (a gross oversimplification I know), it now seems to be the CRB sectors you need to get right.
Naturally, to the extent this is being driven by the USD itself, the S&P has developed a rather worryingly strong negative correlation with the USD. A weaker Dollar drives up commodities which in turn drives up the sectors that have become the market leaders and hey presto we all feel better. Well it is sadly not that easy for two reasons; 1) higher commodity prices act as a transfer of income and 2) it now seems that the beta between the USD and commodity prices is RISING. This may be a by-product of a market desperately attempting to clear itself, and/or rank speculative herding as behaviours are adapting to language like “extended period” and extrapolating ultra-loose monetary policy into the far future. Indeed, this is probably a good point to segue for a moment, and relate the latest version of that thinking, which is that the Fed should a) not be fought (standard thinking of course for years), but b) that QEIII will be rapidly implemented if stock values fall by as little as 5-10%. Thus, the Fed put has a trailing strike 5% below the spot price and will, so long as unemployment remains above, say 5%. Therefore don't be so worried, just remain long risk assets in general and commodity assets in particular, and let the bond market go hang.
The logic is, as most fallacious arguments often are, dangerously compelling but sadly specious on three levels; first it assumes that there is no cost to be paid by anyone important enough to matter by pursuing this approach (and this from a team that cheered both QE I and QE II!), second is that the political climate allows the Fed the latitude to just do more, as and when required with impunity, and third is that QE III would be more of the same rather than something fundamentally different in terms of size, or more importantly asset classes bought.
Moving back to the original topic, that rising USD/CRB beta is a problem for EM because it sends the signal that if nothing changes, then not only will their growth/inflation trade off worsen, but it will worsen at an accelerating rate. In other words, the price of delay in terms of the lost output required to stabilise inflation will be larger if the structural coefficients do not return to something more helpful. Now this is clearly a particular problem in economies with wage and rent indexation, since some of that deterioration is being locked in.
The side effect is a relatively rapid real currency appreciation for those EM economies pegged to the USD, for the simple fact that the energy and food intensity of the US economy is lower than the EM economies. So in that regard, the Fed's policy is to some extent achieving the important adjustment in the USD that all seem to want to pin their growth hopes too – an increase in competitiveness as the route to grow GDP sufficiently in order to make debt concerns wane.
One might argue that the “policy” is demonstrably working since charts show that the bulk of US value added growth since 2009 has come from the goods sector. Well, that is true but probably not due to a weaker USD. First, trade balances worsen when you devalue (remember the J-curve?) and second, in most models of trade demand variables tend to trump competiveness variables in the short-term (i.e., up to 2 years).
Instead, the direction of travel of the real USD against these countries is no doubt correct. However, it is the speed of travel as of late that is making both of us nervous, since it appears to be aiding behaviour which is backing EM in to a policy corner at an accelerating rate when global growth is still highly dependent on these countries' fiscal and private sector demand. In addition, it is putting western low and fixed income consumers into a real income squeeze without the luxury of resorting to credit as a temporary boon or asset price appreciation in any way commensurate with commodity and stock price appreciation.
The irony is that EM total factor productivity – labour and tech transfer and FDI sourced capital – has been so strong that it has absorbed the now 10 year boom in commodity prices that have gone hand in hand with the 10 year USD bear market and EM growth miracle. Commodity “shocks” slow productivity down and magnify the worsening growth/inflation trade off even more. For us in the west that have seen rapid margin expansion combined with P/E increases (driven, we argue, by policy), the sweet spot is probably behind in this regard as pass through pressures must be building. Meanwhile the legacy of that margin expansion is modest wage growth to which one would ordinarily be looking as the source of funding for the next round of expansion/consumption.
Naturally, some economies are more exposed to any change in both the key drivers of returns. Those with large manufacturing sectors will fare worse, those with policy flexibility better. This puts some European economies right at the forefront of risk from changes in US, ECB and BoE liquidity (pillar 2 of the global reflation) and tighter policy/slower growth in EM (pillar 1). We would both be very pleasantly surprised if Sweden, Norway, Germany, Austria et al would maintain consensus and official forecasted growth in 2011 and 12. While in the US, it is probably more the case that any aggregate growth slowdown shouldn't be too bad, rather that the composition would look a lot more service orientated than of late.
The reason we wanted to highlight the asymmetric GDP growth risk in Europe is of course in relation to funding maturing private sector commitments in CMBS etc. and some smaller government bond markets. It has been tragic to observe the workings of the actors when there isn't a script but chilling to realise that this play has been happening during what is being described as a German growth miracle. We'd both be intrigued to see how the actors perform when they are having to ad lib and tighten all the Götterdämmerung commando's belts at the same time! At least some euro area governments have long duration markets and strong domestic demand for debt – think Italy.
So where does that leave Bob and myself? To summarise our take:
- Global manufacturing is late cycle and has probably already peaked.
- EM policy makers are unlikely to be able to loosen with sufficiently speed to turn the cycle quickly. In any event, more capex really wouldn't be sensible.
- A soft landing in the EM world would be a fantastic outcome.
- But positioning has become extended in some markets, and when considered across markets it is probably overextended in the same underlying risk dimensions and themes. This could turn a modest slow-down into a correlation and liquidity rout with substantial knock effects on western demand.
- The addition of liquidity in the west is coming to an end and we think it is unlikely to return in the same guise unless risks of a hard landing materialise.
- For the west, two-way volatility in commodity prices could produce another bout of falling CPI – but that could turbo-charge equity market rotation for a while, at least as real consumption relatively benefits – but less so in the EU and UK than in the US.
- Sector rotation into the real economy related sectors makes sense, but defensive or not is up for debate.
In short, we believe that risk assets do not offer good risk reward characteristics now, whereas quality government bond markets do at least cyclically. And the clear losers would stand to be commodity prices and related assets while the winner initially at least would be the USD and similar currencies.
Looking back it now seems that a fundamental shift happened in mid 2001 to the commodity and currency world, a shift which has been ongoing since and that has affected the global supply side inflation picture around dramatically. This shift has not been analyzed before as far as I'm aware, but in fact, it appears to have dominated asset returns over the period. The US Dollar measured against its broad index shifted from being in a quasi permanent appreciation since the breakup of Bretton Woods, into a depreciating phase which is still going on today. The CRB index, which had been in a broad cycle since the 1960s, shifted into a turbo charged increase phase. Not surprisingly, the basic materials and oil and gas components of the global equity index shifted into a major bull phase at the same time and have together been the two best performing sectors over the period.
This shift coincided with China's membership in the WTO and the outsourcing of production post the equity bear market. This period has seen a substantial move toward a higher headline/core CPI index ratio in the US and other western economies as world commodity prices have trended higher. To the extent that the front end of the US bond is much more correlated to core rather than headline inflation (since the Fed is a core inflation targeter), short rates have not fully reflected this divergence. Considering the funding needs of the US, front end rates are more important than ever for the fiscal balance situation; which is analogous to the cost of funding for a debt stock that is being rolled rapidly.
Why hasn't core inflation tracked headline? We would argue that it is reflected in EM productivity growth and the growing export market share of China (see Figure 6). China's share of US imports is now approaching the peak levels achieved by Japan in the 1980s. As a result, we fell nervous about secular views that rely implicitly on this share rising forever. Truly then, we need to consider whether the twin pillars of market returns since 2009/2010 are not just cyclically extended, and now positionally extended owing to the Arab Spring inspired pause, but coming to a secular shift as well.