Soros Fund Places Big Bearish Bet on SPX 

According to its latest publicly available filings, Soros Fund Management has increased its put position on the SPX (or rather, the SPX ETF SPY) to the second-highest level ever. Note that the fund seems to have held puts on SPY since at least the summer of 2011, and the position's size is often adjusted, a sign that it has served as a hedge most of the time that is occasionally taken off and then increased again. What is noteworthy about the most recent filing (keep in mind that there is a 45 day delay involved as well, so this was the fund's position at year end 2013) is that the position amounts to $1.3 billion and represents more than 11% of the total portfolio's value. In other words, this was probably not merely a hedge, but rather an outright bet on a decline.

According to Marketwatch:

“Soros Fund Management has doubled up a bet that the S&P 500 is headed for a fall. Within Friday’s 13F filings news was the revelation that the firm, founded by legendary investor George Soros, increased a put position on the S&P 500 ETF SPY by a whopping 154% in the fourth quarter, compared with the third. (A put or short position basically gives the owner the right to sell a security at a set price for a limited time, and in making such a bet, an investor generally believes the security is going to decline.)

The value of that holding, the biggest position in the fund, has risen to $1.3 billion from around $470 million. It now makes up a 11.13% chunk of all reported holdings. It had been cut to 5.14% in the third quarter, from 13.54% in the second quarter, which itself marked another dramatic lift on the bearish call.”

Here is a table showing how the put position has been adjusted over time:

Soros Fund Management's holding of SPY puts over time.

Note that in spite of holding a very large position in SPY puts in the course of the bull market period, the fund has performed quite well – since 2011, its NAV has almost tripled. Here is a detailed list of its current holdings by sector.

Since the fund also added various long positions, it is almost certain that the large position in SPY puts is of a tactical nature. Given the recent decline, it is quite possible that it has been decreased in the meantime (just as it is possible that it has been increased again in the subsequent rebound). Still, it is a very large position and Soros is famed for at times taking outsized positions that turn into big winners.

Technical Backdrop

The comparison chart showing the similarity between the current chart pattern of the DJIA with that of the DJIA in 1929 that was first suggested by Tom DeMark many months ago, has been published so widely by now that it is almost certain that it won't continue to hold up. Numerous equity strategists have rightly pointed out that such pattern comparisons rarely 'work', especially as the market and economic conditions accompanying the patterns are usually entirely different.

However, we disagree on one important point: namely that conditions have to be just right to make an event like 1929 or 1987 something that is 'expected' by many equity strategists (there is usually only a small minority raising the alarm in advance).  In fact, it is in the nature of such panic sell-offs that they are never expected beforehand. Rationalizations for their occurrence are always provided after they have happened, never beforehand. In the case of the 1987 crash the debate as to what actually ultimately 'caused' it remains unresolved to this day, and this will always remain so. Similarly, the 1929 crash cannot be ascribed to a single event or cause. Rather, in both cases numerous factors went hand in hand that eventually led to a wave of panic selling. In one case, the crash meant absolutely nothing, since obviously, no depression followed in the wake of the 1987 sell-off. By contrast, in the other case the crash was the first warning that a major economic downturn was imminent.

As to the limits of pattern tracking, we would point out that famous trader and hedge fund manager Paul Tudor Jones made a sizable fortune by betting that the bull market of the 1980s was following the pattern of the 1920s bull market very closely and that there would be a 1929-style crash once the bull market ended. As we know today, this bet worked out perfectly (he reportedly tripled his money in the crash) and today Jones is one of the biggest hedge fund managers in the world.

Here is a documentary that was filmed prior to the crash, where the pattern similarity is discussed in some detail (as an aside, JPT reportedly hates this documentary nowadays).

Documentary on a famously successful 'crash pattern' trade

It is of course true that the little bit of 'QE tapering' the Fed has so far implemented cannot be compared to the tightening of monetary policy prior to the two famous crashes, and neither can it be compared to the tightening of monetary policy prior to the 2008 crisis. In all these cases, money supply growth was lower than it currently is and a considerably amount of monetary tightening had been underway for some time. On the other hand, we cannot say with certainty how vulnerable the market is today, and to what extent perceptions about a likely further tightening of monetary policy may influence the actions of traders.

The market rebound has so far produced slightly more than a 61.8% retracement of the decline in the DJIA, approximately a 78% retracement in the NYA and a 100% retracement in the Nasdaq Composite. The SPX has managed to almost retrace 100% of the correction. Of all these indexes, the DJIA is the only one that can still be regarded as potentially fulfilling the conditions of George Lindsay's 'three peaks and a domed house' formation, but it won't take much for it to leave the pattern schematic behind.

Interestingly, sentiment has not recovered to the same extent as prices just yet (of course sentiment never went to an 'oversold' position in the first place, but a note of caution has crept in that has not been completely abandoned as of yet). Other than that, it is perhaps worth noting that volume has been quite a bit higher during the recent decline than in the subsequent rebound, but we have seen this happen several times in the course of the bull market. So far it has never meant much, but at some point it probably will.

The DJIA has made it to slightly above the 61.8% retracement mark of the preceding decline – click to enlarge.

The NYSE index (NYA) has managed approximately a 78% rebound - click to enlarge.

The Nasdaq has retraced the decline completely, however, there are notable momentum divergences – click to enlarge.

It obviously won't  take much to erase the divergences between the indexes these differently sized rebounds have created, but at the moment they still represent a caution sign.

As a matter of historical interest, below are charts of the DJIA in 1986-1987 as well as in 1998. In both cases the market succumbed to panic selling very shortly after reaching a new all time high. 1986-1987 was a 'three peaks and a domed house' formation that adhered very closely to Lindsay's pattern, with the 'domed house' slightly temporally extended compared to the original pattern, but nevertheless following the template more closely than the most recent example. The rebound from the initial decline managed a short-lived but powerful retracement that ended right between the 61.8% and 78% retracement marks.

In the 1998 decline there was only a very small rebound after the initial decline from the all time high before panic selling began (the peak was in turn put in following an initial brief correction period). The 1998 decline is also interesting because the 'trigger event' in this case was the culmination and end point of the Asian crisis in the form of Russia's default and the LTCM convergence trade debacle, in other words an emerging markets crisis.

The DJIA's 1987 pre-panic retracement rally was quite powerful. It included what was up to that time the biggest single day advance in the DJIA in terms of points (right at the beginning of the rally in the third week of September) – click to enlarge.

This chart shows the DJIA in both 1986 and 1987 – the formation followed Lindsay's 3P + DH pattern quite faithfully – click to enlarge.

In 1998, the market corrected when Russia's troubles began making headlines, but then rallied to one more new high before succumbing to panic selling. In this instance the retracement rally just prior to the panic sell-off was very weak – click to enlarge.

Conclusion:

Monetary conditions are probably not yet 'ripe' for a bigger sell-off and the current divergences between the indexes could easily be resolved by a little more short term strength. However, at the very least we would regard the recent swoon as an initial warning shot. In case readers are wondering why we keep making comparisons to previous panics: the reason is that we expect the bull market that began in 2009 to once again end with one.

The market is extremely extended, has recorded new sentiment extremes and historically extremely high valuations in concert with one of the biggest expansions of corporate profit margins in history (which have the tendency to mean-revert). Moreover it has been pushed higher by a huge Fed-engineered expansion in the money supply. This combination of factors has historically always ended with very large sell-offs in a very short period of time, in most cases not very long after the market attained a new high (there are exceptions to this, such as the 1973 peak or the 2007 peak – in these cases the market's peak and the beginning of the eventual panic were about 11 to 12 months apart). It therefore makes sense to examine past panic patterns and chart patterns resembling the current one, so as to get a sense of the technical warning signs that were in evidence at previous major peaks. This is especially so as situations like the 2007-2008 decline during which the growing troubles in the financial system were rather obvious are actually the exception. More often there is very little warning from the fundamental side of things, as the market in most cases manages to advance in the face of seemingly deteriorating or adverse data for some time, which makes it very difficult to pinpoint fundamental 'triggers' in advance.