The latest report from Morgan Stanley's Graham Secker can be summarized simply as follows: i) in January everything has disconnected as traditional linkages between asset classes have broken down, ii) also in January every major asset class (equities, treasurys, gold, oil) was up materially, iii) such a phenomenon has been seen only 5 times in the past 5 years, iv) a double digit decline followed 3 of the past 4 such surges. Then again, as Bob Janjuah lamented earlier, when a bunch of bespectacled economists who have never held a real job in their academic careers since transplanted with banker blessings to various central bank buildings, and who continue to plan the fate of the world in secrecy (a fate that can be summarized as follows: CTRL+P), as the only marginal decision makers, who really cares anymore?
Breadth of positive returns across asset classes is rare …
Reading the January version of our global cross-asset strategy team’s excellent ‘Global In The Flow’, it struck us just how unusual performance trends were last month. While we’re all aware that we’ve just witnessed the best start to the year in equities since 1994, what was more interesting to us was the sheer breadth of positive performance across a wide array of assets. Effectively, the only major asset to fall in value in January was the dollar, and the only other laggards we could see were corn, coffee, coal and natural gas.
… and has often preceded equity market corrections
Unfortunately, the report in question hasn’t been in existence long enough for us to see just how rare such a breadth of positive performance is. So we have screened the past five years to identify periods of coinciding monthly price appreciation in the S&P, Treasuries, Oil and Gold. As shown in Exhibit 1, January 2012 was only the fifth month in the past five years when all four of these major asset classes have risen in unison. More interesting, on three of these four prior occasions that month proved to be a significant peak in equities and was followed by a substantial double-digit decline.
The traditional relationship between equities, treasuries and gold has broken down in recent months
While this analysis doesn’t guarantee that the market is about to suffer a reversal, it probably does reflect an abundance of liquidity plus rising investor optimism that this liquidity can lift asset prices across the board. Exhibit 3 and Exhibit 4 chart the longer-term performance of equities relative to USTs and to gold, and both clearly show a breakdown in the relationship in recent months. Of course, it is possible this gap can close through either falls in stocks or declines in the other assets, but we think it is unlikely this disconnect will continue for very long.
We see the breadth of recent strong performance as a warning sign
While we believe Exhibit 1 is a powerful argument to position for a market pullback, investors should note that this rule-of-thumb was less compelling in prior years. For example, although it gave correct sell signals in June 2000 and August 2002, it also gave a number of false sell signals during 2003 and at the end of 2004, as shown in Exhibit 2. We believe the macro environment going forward is more akin to the last five years than the preceding decade, and hence consider this signal is an important warning sign; however, we acknowledge that others may take a different view. [yes, like ChairSatan]
Speculators are bullish on equities, bonds and oil …
In seeking corroborating evidence to support the rule-of-thumb suggested by Exhibit 1, we have analysed CFTC positioning across similar asset classes. Exhibit 5 plots CFTC net speculative longs as a % of open interest for the NASDAQ (historically this metric has been a good predicator of European equity performance), US treasuries and the oil price. Within the chart the grey shading indicates areas when investors were net long all three asset classes based on a rolling 3-month moving average basis. To illustrate its efficacy for stocks Exhibit 6 then shows the S&P and MSCI Europe with the same periods again shaded grey.
… which has provided strong sell signals over the last decade
If anything, we think Exhibit 6 suggests the CFTC analysis is even more powerful than that shown in Exhibit 1, as there do not appear to be any false sell signals (although it was a little early at the tail end of 2010) even when we take the analysis back to 1999. Further, Exhibit 7 details some standard performance analysis around this data – for example, since July 1999 the average 6-month return from MSCI Europe has been 0% and the probability the market rises (hit ratio) is 54%. However, when we measure performance from periods when net longs were present across the three asset classes, we find the average subsequent 6-month return was -6.6% with a hit ratio of just 19%.