Sunday, January 26, 2014

Weekly Market Summary

After opening Tuesday within $0.20 of its all-time high, SPY closed down 3% for the week, eviscerating the prior 10 weeks of gains (back to November 14).

So what's next? Here are our conclusions, with the details following below:

  1. Most importantly, as bad as the fall this week was, there's no lower low yet, in either the indices or sectors. Underestimating the strength of the uptrend was a mistake for most in 2013. The December lows (176.5 in SPY) are key to the uptrend. 
  2. The indices ended the week 'oversold'. Typically, a back test of at least the 50-dma follows after the first break below in more than a month. That seems likely. A second break below the 50-dma obviously signals more downside ahead. 
  3. A majority of the telltale signs of a durable bottom are not yet present. This suggests that any immediate bounce is sellable. 
  4. There is reason to suspect that a larger 7-10% correction is in store, to at least 165 to 171. That would match the corrections in April and September 2012 and May 2013. If that is the case, SPY will not see a new high for at least 3 months and probably more than 6 months.

Trend in Indices and Sectors

As bad as this week was, a majority of the indices in the US and the rest of the world held their December lows. There is, for the moment at least, a higher high and low. That's an uptrend.

The same is true in the US on a sector basis. Almost all are below their 20-dma and 50-dma, but those MAs are rising.

Regular readers know that we use the 13-ema for marking trend in SPY. Last weekend, we warned that the 13-ema had inflected downward and that the pattern was for a lower low to follow (circles). That happened this week. From a trend basis, we now await the 13-ema to be regained (blue lines).

Back-test of 50-dma

SPY ended the week below its 50-dma, the first time its touched the 50-dma since December. The two charts below look at similar cases since 2011: in every instance, the 50-dma (180.9) and even the 13-ema (182.8, but falling quickly) has been back tested, even if the trend continued lower. That seems likely here as well.

Going bank further, SPY fell hard through its 50-dma in February 2007. It did not form a back test, but the selling created a washout within a week and the index regained all its losses within three weeks.  

A bounce is therefore to be expected and it could well be this week. MLK week is typically bad, as it was again this year, but the end of January is seasonally strong.

Telltale Signals of a Durable Low

After a 3% fall, there are typically some telltale signs that the low is in. Those signs are not present at the moment. So, a durable low is not likely in, yet. That means that if the bounce happens before those signs appear, it's a sellable bounce.

Let's review each sign.

The RSI(5) on the hourly chart closed at a very low 8 on Friday. That indicates very strong downward momentum, but it's also 'oversold'. After making a new high and then an RSI this low, with only one exception, a bounce has followed but so has a lower low (green).

The McClellan oscillator (NYMO) closed Friday at minus 32. Since 2009, after SPY has made a new high, there has not been a durable low until NYMO closed under minus 50; 70% of the time, NYMO closed under minus 80 before a durable low was in place.

Total put/call (CPC) closed at 0.92. It has not exceeded 1.0 since the October 9 low in SPX, an exceedingly long time (as noted last week). Durable lows have always followed at CPC reading over 1.0; 75% of the time, CPC has closed over 1.2 before the durable low was in.

TRIN closed Friday at 1.75. After a 3% drop, lows have been marked by a TRIN greater than 2; 70% of the time, TRIN was over 3 before a durable low.

Friday was the first MDD (major distribution day; 90% down volume in the NYSE) since the August low in SPX. This indicates significant selling pressure. After a new high in SPX, MDD tend to come in clusters greater than one. This is especially true when the first MDD occurs within a few days of the high. A bounce is certainly likely in-between.

VIX jumped way above its upper Bollinger on Friday, and closed above 18. That normally leads to a near term retreat in VIX and bounce in SPY. But when its the first jump in VIX in more than a month (like now), most often a second jump follows. That has especially been true in the past year.

The ratio of VIX to VXV moved to a sell signal on Friday. Since 2009, with just one exception, a lower low has followed within a week, some minor, some major.

Finally, while seasonality is strong in the coming week, durable lows do not tend to come in January. February, March and April are far more likely (data from SentimenTrader).

Support Levels and Correction Size and Timing

So,  a durable low has not yet been signaled. Where is support?

SPY ended the week back in the middle of the 176.5 to 181 range that constrained SPY for a month in November and December before the FOMC meeting on 12/18 launched the index higher. 176.5 was also resistance earlier in October. That should mark very strong support now (green line). There is also a former trend line from June (dashed) near 178 that could be support. So downside might be 0.5% to 1.3% lower.

If 176.5 marks the low, the total correction would be 4.5%, smaller than the corrections in October (4.8%), August (4.6%) or June (7.4%).

There is reason to suspect that a bigger correction is in store.

Not just sentiment, but investors' positioning has become excessively aggressive equities. To take just one sample, investors' allocation to Rydex bull funds has reached an extreme greater than in April and September 2012 and in May 2013. The ensuing correction each time was 10.6%, 8.6% and 7.4%, respectively. That implies a correction to at least 165 to 171.

Moreover, the uptrend from the June low has now risen 19% without a 5% correction. That is above average, especially when you consider that it followed a 26% gain with just a one month correction in-between. Since 2000, corrections after such a long gain have been more than 7%-10% and new highs have not been formed for more than 3 months and as long as 9 months later (the width of the red boxes is 3 months).

The same was true during the 1990s. It's true that the rallies were larger than 20%, but the corrections ran deeper and lasted longer than in the 2000s. That makes sense: a longer and deeper shakeout sets up a longer and larger uptrend. Likewise, a longer and deeper correction now would be repaid with a better uptrend.

Note also in both of the above charts that when the weekly RSI (top panel) gets 'overbought' (above 70, like now), a durable low has not occurred until it closed the week under 50. It closed last week at 58, further support for more downside ahead.

If a larger correction is in store, one useful signal that a low is in will be the percentage of companies over their 50-dma. This has dropped to below 20% every year since 2000, except in 2013. It's overdue. When it happens, a significant low will be high odds. The first chart is for SPX, which closed at 38% on Friday:

The DJIA has been the weakest index and has led to the downside. Using the same measure as above, the percentage of its components above their 50-dma is now just 23%. Moreover, the weekly RSI, at 34, is close to being 'oversold' (less than 20). It's possible that the DJIA bottoms first and consolidates while the other indices catch up.


Underestimating the strength of the bull market was the biggest mistake of 2013, for us and most others. We suspect that the low is not in and that the bounce will be sellable. But we will let the direction of the 13-ema be our primary guide in the absence of other signals, detailed above, occurring first.

Our summary table follows below: