Linear regression may show us the next bargain

Before we get started on today’s topic, some of you may be interested in how I plan on trading within the current market.

 

A brief market trading plan

Are we about to enter into a new bear trend? That’s the question of the day. S&P is set to open lower. That means today–if it stays below 1990 by the close,  is the first break of the neckline support level of 1990. I will give it 2 more days. Thus–if the S&P stays below 1990 by Fridays close, and does not open higher to indicate that this week’s action was a spike on Monday, I will begin to sell on Monday afternoon. The exception would be if stochastics, RSI etc are well oversold, I can wait for a bounce.

 

Statistically, there are greater odds for a turnaround in the coming week (seasonal data, and some data provided by sentimentrader that speaks to the coincidence of 1.5% drawdowns in the first trading day of January vs. a rally a week after that) –so the odds are good that this is a temporary blip–but I have to follow my rules. Thus, a break in support by 3 days = sell on the 4th day. Assuming no break back above support (1990) on or before that 4th day.

 

S&P timing

Using Linear Regression studies

In my last blog, I noted that linear regression studies can offer hints as to what sectors may be oversold from a long termed perspective. A linear regression study can tell us how far a security has deviated from its mean return over a period of time. When looking at linear regression lines, we track the volatility – or standard deviations, that surround the mean (average) line from point A to point B. Each standard deviation theoretically represents 68.27%, 95.45% and 99.73% of the prices that lie within one, two and three standard deviations of the mean. The standard used by technical analysts is for 2 standard deviations away from the mean. This means – over the period of study for each of the following charts, the upper and lower lines contain some 95% of the movements that have happened in the period of study. Should the trend continue (as you know—trends can change) – linear regression lines can offer a projection for the extent of movements in future.

 

On the following charts, I chose to look at a few sectors that typically have been either higher dividend paying (per my last blog) , or have been beat down significantly – or both. On the charts, I started the linear regression lines (blue lines) at the point where the trend changed. We’re assuming that the current trend is to continue until proven otherwise. The other thing I added – in green – is a “zig zag” line. This line simply stops and reverses at a set default percent movement. I set it for 30%. This means that the line follows the trend unless there is a 30% change in movement one way or the other. Obviously I’m looking for big moves within our linear regression study.

 

Below are charts for Canadian banks, US and Canadian Energy stocks, Canadian utilities, and BCE (which is Canada’s largest telecom company). I’ve made notes above each chart.

 

Canadian & US energy stocks

 

The linear regression line was set on the iShares Energy ETF (XEG-T) from the 2009 lows. Since that time, energy prices, and the producers, have been quite volatile. As you will see on the chart, our zig-zag setting shows plenty of 30% swings in that 6 year period. Yikes! You can also see that the sector is getting close to the 2-std deviation low line. That line lines up with the 2009 lows.

XEG

On the US Energy ETF (XLE-US) I drew the linear regression lines going back to 2003. That’s a longer study period, thus the XLE chart contains more data – hopefully making the lines more significant. It too shows a move back to the lower line.

xle

My take: the buying point may soon approach us on the energy sector. Wait for a consolidation before buying. I expect to blog on the sector if/as/when a consolidation may occur.

 

 

Canadian utilities

 

I drew the regression on this sector ETF (XUT-T) only back to 2011 given the short history of that ETF. The sector presented enough volatility to give us an indication of potential future volatility –assuming the trend doesn’t change aggressively. It appears that XUT is trading below its mean average line. Its bouncing off of the lower linear regression line. That’s probably a good sign for the sector. The zig-zag tool shows us it had a 30% drop since its last high, and hasn’t begun retracing that drop to a significant point. This sector should be one to consider for contrarian investors.

xlu

Canadian banks

 

The September market malaise brought the BMO Cdn. Bank ETF (ZEB-T) temporarily down to spitting distance of its lower linear regression line. Recently, the sector rallied – then stalled out. The fruit is not ripe for picking in this sector at this time.

ZEB

Telecom-BCE

 

Negative volatility has not been much of a factor in BCE’s stock since 2009. The stock seems to comfortably trade in the upper portion of its volatility range. That’s a good thing – especially for investors who want to own a stable, high dividend paying stock. It’s not oversold – at least according to a regression study. There is no signal for contrarian investors looking for a bargain here. But you can see on the green zig-zag line that the stock has rewarded investors holding it over the past 6 years with very few sleepless nights. We at ValueTrend hold this stock in our income model and our growth model.

bce

 

Conclusion

 

Of the securities covered in this study, only the utilities catch my eye right now. Energy, though, is looking pretty interesting. Let’s keep an eye on that one. At this time, I have no positions in utilities, beyond one pipeline. But that may change.

11 Comments

  • PART OF THE PROBLEM HAS TO DO WITH HIGH YIELD BONDS, EMERGING MARKET CURRENCY AND ENERGY. AS MOHAMED EL-RIAN SAYS (ALLIANZ):”WE HAVE NO MORE LIQUIDITY SUPPORT TO PUSH THE MARKET BACK UP AND NO STOCK BUY BACK FOR A WHILE”. YES THE ECONOMY SHOW SOME IMPROVING, BUT THE ACTUAL PRICES ARE DECOUPLED FROM THE FUNDAMENTALS BY LOTS OF PAST LIQUIDITY. THIS MARKET IS ALL ABOUT VOLATILITY AND I FORGOT MENTIONING ABOUT THE GERMAN INDEX MISFORTUNE ($DAX) AND THE U.S. MAIN INDEX ($SPX) WORKING IN TANDEM.

    GOOD DAY,

    J.P.

    Reply
  • I accept the regression concept and I can see the utilities being a sector where this concept could apply. So I will be watching for an entry point.
    However…. for energy I would suggest that unless the price of oil, which must drive the energy sector rises substantially, I cannot see how the energy sector could possibly rise. Statistics and data will show one outcome but “fundamental analysis” must come into the equation. So your take of “Wait for a consolidation before buying”, I would expect you also to suggest “wait for the oil fundamentals to improve” in short meaning wait for oil to show a positive upward trend before buying. Am I wrong in saying the fundamental for this sector must be considered along with data and stattistics?

    Thanks for posting the study as it gives us another factor to consider going forward.
    Daddyo

    Reply
    • Daddyo–the fundamentals will show in the charts. If oil consolidates (it is certainly NOT doing that yet!) then the implication is that something fundamentally is changing–ie supply/demand is leveling to a price where no more supply will come on the market. Once supply starts to pull back, yet demand remains the same or continues to grow, at some point the price will rise. Given there is a lag in supply to get back on line–you will see the price of oil rally / breakout from that consolidation as a leading indicator. The market tends to look “forward”.
      I continue to only watch the charts–its different for a commodity than it is for a company. For example—AAPL earnings can be “estimated” by their producing a great new product that is expected to add to revenue. That is where fundamentals can be a valuable tool. Commodities have many moving parts–and those parts as and on the whole will influence the patterns. Its harder to do estimations on supply demand–as you have to look at total world usage etc–not an easy job for the professional economists–impossible for you and I

      Reply
  • Hi Keith, thanks for sharing this linear regression work.

    Hearing how you are approaching this current near-term market situation is quite helpful to my learning. Can I ask: in this kind of potentially major turning situation, when you discuss watching the $spx neckline level, are you using that as a signal for your $spx holdings only? Specifically, would you use that $spx signal for other developed ETFs, Euro or Japan, or would you be watching those charts for their own signals?

    Many thanks, Alex

    Reply
    • Alex I tend to use the SPX as my macro indicator. I don’t hold any foreign exposure beyond Canada / USA stocks now. The US will influence the stocks I hold in Canada–given none of my exposure is directly in oil. So I will sell much or all of my US holdings and some of my CDN holdings.

      Reply
  • Hi Keith:
    Once again thanks for teaching me how to fish (and not giving me the fish). Two questions: Should the outlined conditions materialize you will begin to sell but how far do you go to raise cash under that scenario, 20%, 50%, 80%….
    Secondly, many stocks have already broken below support before the SPX. You indicate the SPX is your macro indicator. Have you continued to hold stocks that have broken below support waiting for the macro indicator to break support? Which takes precedence or both?

    Reply
    • We sold one stock that broke support last week, and have another one that broke, started to rally (we were planning on selling it at its old support line = new Resistance) –but then, the market cracked and down it went again. Ahh! This is the nature of trading.
      Most of our other stocks are selling off but on their longer term trendlines–they will be keepers. So my goal is to sell some of the seasonally sensitive stocks – technology specifically. From there, I am going to sell the individual stocks that broke support as you note. I anticipate somewhere around 30% will be sold. We hold about 10% cash now in the equity model. I expect that cash level will be 30-40% in the next week or so.

      Reply
  • Laurentian Bank had a pullback to $45.45. The 52 week high on the stock is $55.82. I know it is not one of the big 5 but would you recommend it at this level? Thanks.

    Reply
    • John you know I don’t tend to answer individual stock picks on the blog beyond what has been covered in the subject–but just this once…LB does look to be testing support. But you know the drill–wait for a confirming bounce for 3 days. Then yes, it may be OK to buy. Keep in mind banks are “out of season” this time of the year, but your downside, should support hold, should be limited.

      Reply
  • Hi Keith,
    Thanks a lot for sharing this study. I appreciate your postings very much.
    I have one question re: using stockcharts as a charting tool in your analysis.
    For example, you used XUT.TO in the chart for utilities sector, and it shows an uptrend channel for the past 5 years.
    On the other hand, if I use symbol _XUT.TO (unadjusted), the channel on a 5 year chart will look differently, i.e. more sideways rather than uptrending. What would be your recommendation on what symbol to use: adjusted or unadjusted for the purposes of TA and regression analysis?
    Thanks,
    Eugene

    Reply
    • Eugene–that is a great question. I have complained to stockcharts.com twice about their decision to include dividends in stock pricing for their charts–it distorts things. They insist its part of the equation–I disagree, as we should only be looking at price. The better chart is the non-reinvested dividend type, I agree–and I do look at those on my Thomson One trading screen. It, as you say, shows a sideways channel. However, I think you could apply the same logic to the sideways channel – but I cant draw a regression line on that system –for some reason Thomson doesn’t have that tool (are you reading this, Thomson Reuters?) .
      Thanks for the comment–good observation.

      Reply

Leave a Reply

Your email address will not be published. Required fields are marked *

1 × five =

Never miss another blog post!

Get the SmartBounce blog posts delivered directly to your inbox.

Topics

Topics

Recent Posts

bitcoin

Bitcoin & Dirty Harry

S&P

The NASDAQ has the greatest risk for correction at this time. 

spx vs 200 day

One sign that the market is overbought

ac

Airlines: A value play?

WTI

Past picks and looking forward

vix vs xlp

Consumer staples should be on your radar

cta-bg

Never Miss an Opportunity

Sign up for our newsletter to receive valuable insights that are available only to subscribers.   Beyond the blog – beyond the videos – get the inside scoop.

Scroll to Top