sentimenTrader Blog


2016-02-05 | Jason Goepfert | Comments

Assigning a reason to why a market is up or down on any particular day is typically a fool’s errand. That’s especially the case with stocks when so many inputs are coming together in a highly liquid market.

An initial knee-jerk reaction to an economic report or news event can often be pointed to as a catalyst, but the longer a day wears on, the more dubious that particular reason becomes.

Stocks are getting hit hard today, but whether that’s due to the below-expectations Nonfarm Payroll Report or the heavy selling in some tech names (or some other reason entirely) is anyone’s guess.

Let’s just pretend it has something to do with the Payroll report (NFP for short). We’e discussed behavior surrounding extreme reactions to this report in the past, and it has generally held up over time.

It works better if prices are at or near an extreme level, which isn’t really the case this time around. Even so, here is what has happened when the S&P lost at least -1% on NFP day:

20160205_nfp

These results are only for the next two sessions, which showed by far the most consistency. And it was usually negative, as the selling pressure carried through into early the next week. After that, the results improved more toward random.

Maybe it’s just when NFP misses estimates. To check, here is the same table, but only for those dates when the report was below expectations (the “less than zero” in the table refers to the amount that the payrolls were above/below expectations):

20160205_nfp_zero

Pretty much the same results, though they actually improved a bit.

Most of them occurred during the bear markets of 2001-2002 and 2007-2008, but even so, results over the next month weren’t that bad. This is assuming we buy on the close two days later (so next Tuesday):

20160205_nfp_zero1

The S&P was higher 71% of the time, with a decent average return. We prefer to see the max gain be at least twice as large as the max loss, which isn’t the case here. And there were some large losses sprinkled in there.

Curiously, bonds are also down on the day, using the TLT fund as a proxy. This is highly unusual. With a weak jobs number, bonds would usually be higher, especially if stocks are selling off, because bonds act as a type of safe haven.

There was only one other time that the jobs report was below expectations, the S&P was down at least -1% and bonds also were down on the day, which was July 3, 2003. For what it’s worth, stocks rallied in the short-term after that, then fell back again before staging a longer-term rally.

Investors and traders like to obsess over the jobs number, mainly as a potential leading indicator for recession, so reactions to the number can be telling. Based on previous behavior, there tended to be some short-term follow-through to the selling pressure after days like this, but no more than over the next couple of sessions.

 


2016-02-03 | Jason Goepfert | Comments

The US dollar is getting hammered so far today, with the Dollar Index dropping nearly -2%, among the largest one-day declines since 1975.

It is on track for more than a -2.5% loss over the past three sessions, which is of course also one of its largest losses in that time span. What’s notable about the decline is that three days ago the Dollar Index was within spitting distance of a new high.

20160203_usd_chart

Sudden, severe price changes from a price extreme often signal a larger trend change, or at the very least a period of back-and-forth trading. Rarely do we see a market just turn right around and resume its previous trend.

The table below shows every time since 1975 that the buck suffered a big 3-day drop from near a 52-week high along with its performance going forward.

20160203_usd

Its worst performance was over the next month, losing more ground 12 out of 21 times and with a significantly negative average return. After 1986, every time it did happen to gain over the next month, it subsequently gave those gains back.

To see if there was any impact on stocks, here are the same dates, but with the performance of the S&P 500 going forward:

20160203_spx

Not much there. Some weakness shorter-term (two weeks) but overall the returns were in line with random.

The sentiment figures for the dollar that we track on the site are showing optimism, but nothing like last spring when it was through the roof. Nothing much to conclude from that – sentiment is less of a help when not at an extreme, and not showing a clear trend.

Based solely on this price break, we’d consider the outlook to have a modest negative skew, especially over a multi-week time frame.

 


2016-02-01 | Jason Goepfert | Comments

Stocks are backing off after the best gain in months on Friday. Optimism had reached the highest level in months as well, and when we see big gains like that in the context of a downtrend, high optimism tends to lead to sub-par short-term returns.

Futures are pricing in a gap down of about -0.5%. When such a gap down comes on the heels of a large gain, it tends to follow through to the downside throughout the day, but then rebound at some point in the next several sessions:

20160201_gap

We can see from the table that there was a close above the open at some point in the next week in 11 out of 12 instances, typically within one trading day.

There is a similar tendency for any gap down of -0.5% or more on the first day of a month:

20160201_gap2

Again, open-to-close returns were mediocre, but 22 out of 24 instances saw SPY close above that open at some point in the next week, typically right away.

If we put Friday’s thrust into context, it was the largest gain in the S&P in more than 90 days, after the index had closed at a 52-week low in the past two weeks. Common wisdom is that this is a typical bear-market relief rally that’s bound to roll over to new lows.

The table below suggest that record is mixed:

20160201_rebound

On a longer-term basis, it was telling if buyers persisted. After these large one-day jumps off of a low, if the S&P was higher 3 trading days later, then it added to its gains over the next six months 64% of the time, averaging +6.2%. But if the S&P was lower 3 days later, then six months later it was higher only 31% of the time, averaging -2.3%.

Momentum doesn’t die easily, and we’ve seen some upside momentum recently. As much as we can rely on historical price patterns and seasonality, it argues that some very short-term weakness that burns off Friday’s surge in optimism should lead to another poke higher in the days ahead, and that would be a good sign longer-term as well.


2016-01-29 | Jason Goepfert | Comments

Stocks have been pushing consistently higher all day, a change in character from what we have been seeing this year. It is further confirmation of the medium-term positives we’ve been discussing in the Daily Report.

A big gain to end the month has often led to follow-through during the first few days of the next month, especially the first day. The beginning of a new month has a bullish tint to it anyway, but especially so when the previous month ended with a daily gain of 1.5% or more (Figure 1).

20160129_last_month

One caveat about this is that the S&P set a 52-week low during the month. When it did so, the results got noticeably less positive (Figure 2). Returns over the next month were especially negative.

20160129_last_month2

There are some strong cross-currents here, with the negative price trend and dangerous readings in some long-term indicators (equity vs cash assets, etc), but extreme pessimism and many signs of capitulation last week.

The bottom line with cross-currents like this and the tables above, is that two-sided volatility should be expected. We still expect an upward skew in prices in the coming week(s) but don’t see compelling enough reasons to be aggressive.


2016-01-28 | Jason Goepfert | Comments

The selling pressure in Chinese shares has been relentless in January and shows no signs of letting up as the Shanghai Composite dives to yet another yearly low.

There is not a consistent relationship between that index and U.S. stocks. It’s a fascination among many market-watchers but share performance in China has been a poor predictor for share performance in other markets.

Breadth figures for stocks in the Shanghai are starting to reach “puke” levels according to several measures. One holdout is the percentage of stocks in the index that are sitting at 52-week lows. It reached 29% on Wednesday while prior peaks in that measure were north of 50% – 60%.

But if we look at the shorter-term persistence of the selling pressure, this month stands out.

On an average day this month, 43% of stocks in the Shanghai closed at a 4-week low. That’s tied for a record for any month since 2002, the furthest we can go back. January 2002 had a similar stretch.

Also on an average day this month, 30% of the stocks closed at a 12-week low. That’s the third-highest of any month since 2002.

20160128_shanghai

Looking at both figures combined, there were only two months that approached our current level of persistent 4-week and 12-week lows, November 2002 and December 2011.

After November 2002, the Shanghai lost another -5% in December, but was 5% higher three months later and 10% higher six months later.

After December 2011, the Shanghai gained 4% the next month but that was about it as the gains petered out in the months ahead.

The relentless nature of this decline has reached climactic levels, at least on a shorter-term basis, but is not yet to a level that would suggest a long-term disgust like occurred at prior long-term lows.


2016-01-20 | Jason Goepfert | Comments

The breadth flush today is getting more notable.

If the figures stand as they currently are into the close, then since 1940, only October 19, 1987 and August 8, 2011 will have seen more lopsided Up Volume over 1-day and 10-day periods.

New lows are climbing continually higher and it will not improve before the close (new low figures are calculated as of intraday lows so the figure can only get worse as the day progresses). Currently, 1,353 stocks have hit a new 52-week low, more than 42% of total issues.

Figures 1 and 2 show every time since 1965, the furthest we can go back, when more than 40% of NYSE issues hit a 52-week low. With the exception of October 2008, even in the midst of bear markets these led to medium-term relief rallies.

20160120_new_lows_chart

20160120_new_lows_table


2016-01-20 | Jason Goepfert | Comments

Markets are now entering “puke” territory, at least in terms of market breadth.

There have been over 1,228 securities on the NYSE that traded to a 52-week low, and it is rising rapidly. It reached 1,342 on August 24 and is getting close to 40% of all issues, a rarely-achieved level.

Even more notable, the Up Volume Ratio on the NYSE is only 3%, meaning that 97% of all volume traded today is in declining stocks. The 10-day average of that figure stands at 26%, if we close as badly as we are now.

Since 1940, there have been only four days that saw an Up Volume Ratio at 5% or below with a 10-day average at 27% or below:

  • August 29, 1966
  • May 25, 1970
  • October 19, 1987
  • August 8, 2011

The average return over the next six months was +17.7%, with all of them being at least +14%.

20160120_breadth


2016-01-20 | Jason Goepfert | Comments

Another morning, another large opening loss.

The S&P 500 futures are indicated to open down more than 1.5% below Tuesday’s close, marking the 5th opening gap of -1.5% or more in 12 days. In the history of S&P futures, only three time periods had this tight of a cluster of large negative opening gaps:

  • Late October 1987 (aftermath of Black Monday)
  • October-December 2008 (aftermath of the financial crisis)
  • March 2009 (bottom of bear market)

There were several time periods that saw 4 gaps in 12 days, all medium-term bottoms:

  • November 7, 1997
  • September 21, 1998
  • June 4, 2010
  • August 19, 2011
  • September 19, 2011

Clearly, an unusual situation. What’s also notable about today’s indicated open is that it is occurring when sentiment is already panicked.

Like recession indicators, we could pick and choose among hundreds of indicators to confirm whatever we want to suggest, but we always try to most heavily weight model-based readings. These are objective, not subject to cherry-picking and proven in real-time through market cycles.

As of Tuesday’s close, the spread between the Smart Money Confidence and Dumb Money Confidence was +49%, down a bit from the previous day.

Figures 1 and 2 show the performance of the S&P futures from the opening print on any day that they opened down at least -1.5% and the Confidence Spread was +45% or greater.

20160120_gap_chart 20160120_gap

Same-day performance was mixed, but improved dramatically as we moved out the time frame. The only real failures were in June 2002 and during the midst of the October 2008 collapse.

The highest-probability risk at the moment is if the major indexes, primarily the S&P 500, move below its August lows. That could trigger a round of stop-loss orders during the day and bleeding into the next 1-2 days. Margin call selling is a small possibility. A quick loss of an additional 5%+ would not be out of line in that case and if we saw anything like that, then there would be an exceptionally high probability of higher prices over the next several weeks to several months.

The safer bet in kinda-sorta panic situations like this would be to wait for the S&P to exceed the high from the previous day, about 1900 in our current case. That is not foolproof by any means, but it does tend to capture most of the upside of these rebounds while preventing one from buying into a possible collapse. The risk in those cases is that one is buying into rising prices in a downtrend, subject to reversal, so (wide) stop losses are recommended.


2016-01-16 | Jason Goepfert | Comments

A big concern among technically-oriented traders is what looks to be a pending breakdown from a rounded top formation. Many also point out a potential head-and-shoulders formation.

20160116_tops

Technical analysis has never proven to be consistently predictive, partly because it is subjective. There are a lot of “it looks like…” types of statements when discussing formations, making it difficult to quantify and test.

That doesn’t mean it isn’t without merit. The whole point of that branch of analysis is a focus on what is happening right now, without the assumptions and forecasts that fundamental analysts make. So in terms of stripping away a lot of B.S., technical analysis serves a purpose.

Should we be concerned if stocks break down? After all, we saw a similar pattern play out in 2000 and 2008. Stocks rose for years, suffered a volatile back-and-forth few months, then broke down and fell into bear markets.

We are conditioned to put the most weight on recent occurrences, which is valid up to a point. Past that point, we become blinded to other possibilities.

Let’s go back as far as we can, to 1896, and look objectively for other instances like we’re seeing now, and in 2000 and 2008. We’re looking for rising prices for a year or more, followed by a rounded-top, head-and-shoulders type of formation.

This is inherently subjective, but I’ve been studying chart formations for 25 years. Even so, anyone with even 5 minutes of instruction on what to look for should point out most of the same time periods.

Figure 2 shows every similar occurrence in the past 120 years. There is a “Before” snapshot, and try to look at that first. Look at the far right edge of the chart and think about how scary it must have felt if you were experiencing it in real-time.

The “After” snapshot extends the chart out over the next 2-3 years, so we can see the implication if we had sold when stocks broke down.

20160116_tops_history

We would have felt really good about ourselves in 1930, 2001 and 2008 as stocks collapsed. We would have temporarily felt good in 1903, 1957, 1962, 1966 and 1984. We would have felt pretty stupid about panicking in 1949 and 1992 as stocks bottomed almost immediately and roared to new highs.

The “temporarily smart” instances, which comprised most of them, fell an additional 10%-20% from the breakdown point over the next 2-6 months (generally).

The most consistent factor among the worst losses was persistent selling pressure from the breakdown level. That’s the key to watch for – if stocks do break down below the August lows, and buyers show no interest in stepping in, then it makes sense to expect even more pain in the months ahead.

But a multi-year bear market that erases 40%-60% of market value? Not likely, especially if relying on some picture of a chart that happens to look like a couple of recent occurrences.


2016-01-15 | Jason Goepfert | Comments

Another morning, another large gap open.

For 8 out of the past 10 days, investors have started the session with prices far removed from the prior day’s close. All but two sessions began with the S&P 500 futures showing a pre-market gain or loss of more than 0.5%.

When this occurs with an open that’s at the lowest point of the past three months, it has signaled either the middle of a waterfall decline (2008) or the end of an exhaustive move (every other date).

It’s also extremely rare. Since the futures began trading in 1982, it hadn’t happened until 2001. All occurrences were clustered between 2001 and 2011 (Figure 1).

20160115_gaps_chart

Figure 2 provides the dates along with the S&P’s returns going forward. The returns are from the open of the 8th gap out of 10 days, when that open was at a three-month low in the futures.

20160115_gaps_table

This is nothing different from other studies in prior days. The kind of oversold reading and price jumpiness we’re seeing has led to binary outcomes – either we’re in the midst of a crash, or this is the end-game volatility as stocks are putting in a multi-week to multi-month low.

Given the recession probability noted in yesterday’s Daily Report, and the fact that (so far) the S&P is holding above its  August lows, we’re still giving more weight to the latter scenario.


← previous page · next page →