By Stephen Jen (via Itau Global Connections).
Bottom line: QE3 has been a ‘success,’ as it has achieved its objective of artificially inflating the values of both bonds and equities. While its timing and the pace are uncertain, in theory, the Fed’s prospective tapering may mark the beginning of the unwind in these asset price distortions, with simultaneous declines in both bonds and equities being a risk, if corporate earnings don’t re-accelerate to compensate for the tapering effects. We believe it may be useful to review the concept of the ‘Beta Convexity Puzzle’ we introduced in February 2012, as this framework may be useful in helping investors think about the likely impact of tapering. High-beta assets (such as equities, AUD and oil) should not in theory rally at the same time as low-beta assets (such as gold, CHF, and Treasuries). For much of the period 2009-2013, however, most of these high- and low-beta assets rallied against the dollar. Such dynamics violate a key assumption underpinning the standard CAPM (capital asset pricing model). We call this the Beta Convexity Puzzle because, according to the standard CAPM, the returns on assets should be a monotonic function of beta, i.e., the high- and low-beta assets should be a monotonic function of beta, that they perform differently through a business cycle. In practice, however, QE1, QE2, and QE3 have had powerful effects on asset prices, inflating both wings of the beta curve. Thus, in theory, tapering could lead to an unwind of this convex curve. How bonds and equities behave in the US and elsewhere will be a function of various considerations, including the trajectory of corporate earnings. But all else equal, tapering by the Fed should lead to a simultaneous sell-off in both bonds and equities. That’s the new ‘gravity’ the Fed and investors need to keep in mind.
The Beta Convexity Puzzle concept: a brief review. In an earlier note we wrote (‘QE3 and the Beta Convexity Puzzle,’ February 2012), we first proposed this concept. Essentially, we argued that the behavior of the global currency markets, and financial markets in general, in the previous four years or so have been peculiar. Not only has volatility in asset prices been low relative to the underlying macro data, it has been unusual in that both high- and low-beta currencies have appreciated against the dollar. This beta convexity puzzle also applied to other assets: gold (traditionally a low-beta or safe haven asset) and oil (historically a high-beta asset) were bought for most of the last four years. So were both AUD and CHF, until recently. Such non-linearity in the relationship between asset returns and beta was in stark contrast to a key assumption made in the standard CAPM.
The two charts below help explain what we had in mind.
These charts show the relationships between various types of assets and their beta, relative to the growth rate of the global economy. The chart on the left is the ‘what-should-have-happened’ chart and shows two lines, highlighting what should happen on average during typical business cycles. The upward-sloping line shows that, for high-beta assets such as equities, crude oil, and high yield FX, their returns should improve as the global economy accelerates, but should not perform well if the world is in a recession. In contrast, the downward-sloping line shows that, in a recession, the low-beta assets such as bonds, gold, and low-yield currencies should outperform.
For the global economy and the financial markets, the CAPM assumes a monotonic relationship between expected returns and beta (systemic risks) in the long term. However, a persistent violation of this linear relationship at the global level would turn the standard finance theory upside down.
The rounds of QE and the Beta Convexity Puzzle. The chart below shows that, during January 2009-March 2013, there was indeed such a violation of the standard finance theory. There was an odd convex relationship between high- and low-beta assets, in which, almost regardless of the global business cycle, most assets appreciated in value, measured in US dollars. While there are various possible explanations for these distortions, we believe the multiple rounds of QE by the Fed, as well as the QE operations by the other G7 central banks, have played a dominant role.
The chart below (please refer to the attached pdf) shows this stark convex relationship of the asset returns relative to beta, very similar to what we had suspected in the 2011 note (compare the chart below with the chart on the right side above).
During this period, six of the G7 central banks were printing money. The Fed’s balance sheet rose from 14% of GDP to 20%, and the balance sheet expansions were significant for other reserve currency-issuing central banks as well: 23% to 32% for the ECB; 13% to 26% for the BOE; and 35% to 84% for the SNB. Extraordinary policy aggression by these central banks was evident. Both Brent and gold – the odd couple in this beta space – registered significant gains; so did MSCI US and 10Y USTs. The Sharpe ratios of most portfolios were extremely high, because both bonds and equities performed (high numerator) while volatility was artificially suppressed (low denominator). What were supposed to be hedged portfolios were in fact levered plays on the Fed’s QE.
Tapering by the Fed. More recently, however, the FOMC members have begun talking about ‘tapering’ – scaling back the size of the monthly QE operations. Chairman Bernanke’s JEC testimony on May 22, 2013, marked the main inflection point in this discussion. However, this idea had already been suggested by some FOMC members as early as April. Later, on June 19, 2013, Chairman Bernanke officially announced the possibility that, if certain macro conditions were satisfied, the Fed could start tapering in the fall this year and fully terminate QE3 by mid-2014. While Chairman Bernanke seemed to flip-flop on the intention to start normalizing policies in his speech on July 10, 2013, it still seems likely that tapering will start later this year, in our view.
In any case, since the tapering discussions began, global bonds and equities have experienced a sharp rise in volatility, and some of the assets (gold and the Nikkei) have experienced extraordinary declines in value. The chart below shows the asset returns relative to beta since April 2013. What used to be a convex relationship (the previous chart) has turned into a concave relationship (the chart below). In other words, since the tapering discussion began, both bonds and equities have come under pressure.
We have these thoughts.
— Thought 1. Both bonds and equities can sell off simultaneously if the Fed tapers too fast. This is the new ‘gravitational pull’ both the Fed and the investors need to keep in mind when tapering starts. The preference of the Fed is presumably to allow the 10Y UST yield rise gradually, to reflect the expected economic recovery and tapering, but for equities to remain supported by improving corporate earnings. But it is not clear that they will get what they want. First, the Fed has argued that it is the ‘stock,’ not the ‘flows’ of their operations, that should support bond prices (keep the yields low). But this claim has not yet been tested. Tapering would be such a test: lowering flows while keeping the stock high. Recent market price action suggests bond prices may fall further with tapering. Second, the BOJ’s inability to stabilize the JGBs highlights that there is a threshold to how much markets can be ‘bullied’, that market expectations may at times be more powerful than the sheer size of the QE operations. Third, US corporations derive a great deal of their profits from outside the US. We have discussed global economic decoupling. But a faltering rest of the world will weigh on US equities.
— Thought 2. Gold and the Nikkei the leading indicators? The collapse of gold in April and of the Nikkei in May was extraordinary. Gold is a key safe haven asset, while the Nikkei is one of the highest-beta assets. For both to collapse so sharply, it does raise the question whether they are leading indicators of what could happen to other low- and high-beta assets. Further, not only were these sharp price collapses against the prevailing consensus expectations at the time, the fundamental arguments in favor of buying and holding gold and the Nikkei have not changed, even after the price corrections. This suggests that there might have been large gaps between the pricing of the assets and the underlying fundamentals. But the same could be said for most other financial assets, thanks to QE.
— Thought 3. Economic decoupling and asset-market decoupling: a negative outlook for EM. We believe the prospective decoupling between the US and China could add another dimension to this discussion, that EM assets, in particular, could be hurt in the months and quarters ahead, even if US equities hold up. We will not repeat the arguments we’ve made in the past, only to stress that the growth deceleration in China is likely to be both cyclical and structural: the former is healthy while the latter is inevitable. China has gone beyond ‘growing-at-all-costs’, mainly because the world has changed after the Global Financial Crisis. Some of the legacy issues (e.g., off-balance-sheet lending, local government debt, and low yields on infrastructure investments) will need to be dealt with. While China can take care of itself during this transition, the same could not be said about commodity-exporting economies and the economies in the manufacturing eco-system centered around China. We expect to see a strongly ‘concave’ relationship between EM bonds and equities and beta.
— Thought 4. The USD and the EUR. This Beta Convexity framework of course is not able to address specifically the current pricing of EURUSD. However, it does show that both the USD and the EUR have behaved in similar ways. In fact, over the last few years, there are more similarities between the USD and the EUR than differences: both the USD and the EUR had under-performed during 2009-2013; both the US and the EMU had major issues and conducted QE at the same time; both the USD and the EUR have been the backbone of the global financial system and sources of global capital. When both the USD and the EUR suffered at the same time, demand for gold, CHF, and others on the left side of the convex beta curve rallied, reflecting ‘global uncertainty aversion’. The assets on the right side of the convex beta curve had double reasons to rally. The EM countries have suffered a lack of asset supply. The global liquidity imparted by both the Fed and the ECB in turn helped fuel the credit and investment cycles in EM that were not sustainable. More recently, the Fed’s tapering talk also coincided with a hawkish turn at the ECB, when President Draghi refrained from adopting a negative deposit rate and instead re-adopted a rather optimistic view of the Euroland economy. In short, we think of the USD and the EUR as having played similar roles in fuelling the simultaneous rallies in global bonds and equities. Any curtailment in their policy aggression would lead to problems with the low- and high-beta assets.
— Thought 5. In search of balanced portfolios. One major problem in portfolio management is the lack of a balance between bonds and equities. As discussed above, if bonds and equities are positively correlated, as are gold and oil, it is very difficult for portfolio managers to construct hedged or balanced portfolios. If we are right, that, as the Fed tapers, there is a transition period where a concave relationship dominates in the beta space, then it is not clear what one would buy if he/she cuts back on both bonds and equities. Of course, when the Fed normalizes its policies, we should in theory see the normal beta relationships, and not just convex or concave relationships. But in getting from here to there, we fear that the Sharpe ratios in general will likely be inferior to what they were in the past four years.
Bottom line. The Fed’s QE operations artificially inflated the values of both bonds and equities. This was not a negative side effect of QE, but precisely the intended objective. But now that the Fed is contemplating tapering, we believe investors need to prepare for the possibility that both bonds and equities falter, at least this is the new ‘gravitational pull’, all else equal. If the Fed could somehow engineer this policy reversal in line with the recovery of the US economy and therefore of corporate earnings, US equities might just hold up okay, even if USTs sell off further. However, in most scenarios we can imagine, financial markets in the rest of the world (especially EM) will likely struggle much more, when the Fed tapers.
Stephen Jen is the managing partner at SLJ Macro Partners. Prior to establishing SLJ Macro Partners in April 2011, Stephen was a Managing Director at BlueGold Capital (since May 2009), working as the key risk-taker in currencies and as its macro strategist. Before BlueGold, Stephen was a Managing Director at Morgan Stanley and, from October 1996 to April 2009, held various roles, including the Global Head of Currency Research and the Chief Global Foreign Exchange and Emerging Markets Strategist. Prior to Morgan Stanley, Stephen spent four years as an economist with the International Monetary Fund (IMF) in Washington, D.C., covering economies in Eastern Europe and Asia. In addition, Stephen was actively involved in the design of the IMF’s framework for providing debt relief to highly indebted countries. Stephen holds a PhD in Economics from the Massachusetts Institute of Technology, with concentrations in International Economics and Monetary Economics. He also earned a BSc in Electrical Engineering (summa cum laude) from the University of California at Irvine. Stephen was born in Taipei, Taiwan, and now lives in London with his wife and two children.