The Price Support Breach Count - A Very Handy Indicator
For the record, I'm not a technical analyst, at least not by this measure: About 65% of my research is based on fundamental factors, 20% is based on investor behavior patterns, and 15% is based on technical indicators like my price support breach counter. Technical indicators are useful, but not in isolation. That's one of the reasons I get such a kick out of the Elliot Wave tribe. A little too much subjectivity in interpreting the numbers for my taste. But I digress.
Many of you are familiar with what a price support breach is, but for those who don't, here is a simplified definition. A price support breach occurs when the most recent low price for the stock or the index is "taken out" by a new, lower price.
Another popular example of a price support breach is the event where the stock or index trades below its 50-day or 200-day moving average. When this happens, technical analysts will consider the event to be something meaningful, perhaps a warning of sorts. The trouble with this type of breach is that the support price, whether it's the 50-day or the 200-day or any other number of day moving average is arbitrary and often capricious. But technical analysts love to see these breaches because it gives them something to get excited about. Who can blame them? After all, it's their job.
But my version of a price support breach is different in some important ways. First, it's not based on an arbitrary moving average price. It's based, instead, on a hard price that was observed at some date in the past. Moving averages change almost every day, but a support price doesn't.
Second, my version is not a binary choice. Once the market breaches a key moving average price, the game is essentially over. The stock or the index in question is either above or below its moving average. With my version, the first breach means nothing, and it has virtually no information value by itself. It takes multiple breaches for my version of this indicator to begin raising the alarm that the price trend might be changing.
Today, I'm going to do something I normally don't do - reveal the "secret sauce" that goes into one of my key market indicators, my price support breach counter. Why would I do this, you might ask? Well, because I'm not really giving away the keys to the store. I'm just giving away one recipe among many that I use in my models.
Here is how I calculate support breaches
Warning: This is going to get pretty deep into the weeds, so if you're not into the whole math thing, go ahead and skip to the next two charts below. For the rest of you information junkies, here it is.
They say there's a thousand ways to skin a cat, and there's also a thousand ways to slice and dice the historical price behavior of a stock or an index. The way I do it is shown in the table below.
Table 1. A random sample of support breaches from 1980
The first 3 columns should be fairly intuitive to most of you. Date, closing price, and "is today's price a new high?" are standard fare. Column 4 needs some 'splaining. What do I mean by "current support?" It's the lowest price recorded since the most recent new high price.
Looking at the top row of numbers in the table shows that on 9/22/1980, the S&P 500 closed at 130.40. I'm using the S&P in this table because that's where I'm going to take you later on. But you could just as easily use any price of any publicly traded instrument.
Since 130.40 was a new high for the market, column 3 answers "yes." And because 130.40 is a new high, it automatically becomes the current support price, and it resets the price support breach counter (column 6) to zero.
Column 4 asks whether 130.40, the current price, represents a breach of the previous support price. Since this is a new high, the answer is "no," so the cell is blank.
The next day, 9/23/1980 the market closed lower. This triggered a couple of things. First, it established a new support price of 129.43. Second, it was a breach of the prior support price of 130.40. The breach counter in column 6 now changes from zero to 1. There has been one price support breach since the most recent new high price.
In row 3, the market rose but did not take out the previous new high price. Therefore, the breach count remains at 1. Same for rows 4 and 5. When we get to row 6, 9/29/1980, something happens. The price that day represents a breach of the previous support of 126.35. Why didn't the previous two days trigger a support breach? Because of a very important detail in this calculation.
A support breach can only happen after an uptick in the closing price from the day before. The prior two days were both downticks, but on 9/30/1980 there was an uptick, which established the previous closing price as a support breach. At this point, if your head is spinning a little, don't worry. It will all make sense to you when we look at the next two charts.
To finish the discussion of this table, the breach count never got above 2 in the 16-day period covered. And that's usually what happens in a normal, upward-trending market. The count can go to 3 or 4 without raising any flags. But once it reaches 5 or more, flags begin to go up. More on that later.
What is the price support breach count today?
Take a look at Chart 1 below. The gold line is the S&P 500 index from January 2017 through today (April 2018). You can see that the scale is on the right side of the chart. The grey bars represent the rolling count of support breaches, and it keeps adding to the total until eventually a new high price is reached. The scale for the running total of breaches is on the left side of the chart. Take a minute to let this calculation sink in, and I'll give more details as we proceed.
Chart 1. The closing prices for the S&P 500 (right scale) and the support breach count (left scale)
Chart 1 covers the period from January 2017 through today. You all know that the market has been on a tear ever since the election in 2016, and you also know that this post-election rally ran into trouble in early January 2018. Let's look at what's going on in this chart.
As of today, April 30, 2018, the current support price for the S&P 500 index is 2,581. The index hit that price in early February this year, and so far, there has not been a lower closing price for the index. If the index closes below 2,581 at some point in the future, before it makes a new all-time high, it will be a price support breach event.
The gold line is the daily closing price for the S&P, and the grey bars are the count of the number of support breaches since the prior high price. As you can see, the breach count remained constrained between zero and 5 during the run-up since the election. It has since settled down to a maximum reading of 3.
When the count reached 5 in April 2017, I began to pay attention to the possibility of market correction. There was no correction. That is, until January-February 2018. The market dropped fairly quickly by 10% in less than one month. But interestingly, the count has not risen above 3. This tells me that the correction we are in now will probably work itself out, and the market is likely to make a new high before it finally succumbs to the substantial forces of stretched valuations, geopolitical risks, turmoil in Washington, rising rates, and a reawakening of inflation.
But those things are in the future. They may or may not happen within the next 12 months. Therefore, I prefer to stay nearly fully invested in equities, and I will do so until the count reaches 5 or more and is corroborated by the other indicators in my models.
This is what the price support breach count did when the market tanked in 2008.
I wanted to draw a stark comparison between our current market and the nasty bear market of 2008. Please note that I am not saying that we are headed in this direction today. It's just a comparison for information purposes.
A look at Chart 2 shows the track of the S&P and the rapidly rising count of the support breaches. From the top of the market in October 2007, it only took about a month to reach a 5 in the breach count. The count went quickly to 7 after that and stayed there for more than a month. As you know, things didn't improve from there.
On March 10, 2008, the count reached 10, which is a "hair on fire" reading for this indicator. By then, I had gotten most of my clients out of the market, thanks in part to this indicator.
You can see that the count kept rising throughout 2008, reaching its ultimate peak of 24 on March 9, 2009. I will summarize the lessons learned from this event below.
Chart 2. The closing prices for the S&P 500 (right scale) and the support breach count (left scale)
What can we learn from this indicator
First and foremost, remember that this is one indicator among many that give us warnings about trouble ahead. Anyone who thinks they can time the market based on this indicator alone is going to get hurt.
Second, remember that this is my own, homegrown way of tracking price support breaches and making judgments about their significance. I'm not the only guy out there who does this kind of research, but my advantage is that I've been doing this for 30 years. I know a thing or two because I've seen a thing or two.
Third, remember that as you formulate your own opinion about how bad things are getting, always look for confirmation from multiple sources. As I've been saying for a long time, the best thing an investor can do to protect their life savings from the ravages of a bear market is to have a solid Plan B. And don't be fooled into thinking that your Plan B is in your head, and that your "gut" will tell you when to get out of harm's way. That kind of loosey-goosey approach to managing risk never, ever works.