A Trend Model update: Still cautious

My Trend Asset Allocation Model is a market timing model that has been running since 2013. While the model only issues buy, hold and sell signals for stocks, investors nevertheless need to make their own decisions on how much to buy and sell. Based on my out-of-sample signals, I created a model portfolio by varying the equity weight by 20% around a 60% SPY and 40% IEF benchmark. The turnover characteristics of the model portfolio is manageable, averaging 3.5 signals per year in the last five years.

The risk-adjusted returns of the model portfolio are strong. The model has beaten the 60/40 benchmark on 1, 2, 3 and 5-year time horizons, as well from inception for the period from December 31, 2013 to April 29, 2025. In addition, it was able to achieve these returns with controlled risk, equivalent to roughly an 85/15 stock/bond asset mix with 60/40 risk. As the dotted line in the chart depicting relative performance shows, the model mainly reached the superior risk-adjusted returns by sidestepping the really ugly bear markets over the study period.

• 1 year: Model 9.5% vs. 60/40 8.7%
• 2 years: Model 13.3% vs. 60/40 11.4%
• 3 years: Model 9.6% vs. 60/40 7.7%
• 5 years: Model 10.9% vs. 60/40 9.0%

Here is what it’s saying now.

The full post can be found here.

 

 

Special announcement: Humble Student of the Markets will cease publication on March 31, 2026. See this announcement for more details and updates.      

A Trend Asset Allocation Model review

Preface: Explaining our market timing models 

We maintain several market timing models, each with differing time horizons. The "Ultimate Market Timing Model" is a long-term market timing model based on the research outlined in our post, Building the ultimate market timing model. This model tends to generate only a handful of signals each decade.

The Trend Asset Allocation Model is an asset allocation model that applies trend-following principles based on the inputs of global stock and commodity prices. This model has a shorter time horizon and tends to turn over about 4-6 times a year. The performance and full details of a model portfolio based on the out-of-sample signals of the Trend Model can be found here.

My inner trader uses a trading model, which is a blend of price momentum (is the Trend Model becoming more bullish, or bearish?) and overbought/oversold extremes (don't buy if the trend is overbought, and vice versa). Subscribers receive real-time alerts of model changes, and a hypothetical trading record of the email alerts is updated weekly here. The hypothetical trading record of the trading model of the real-time alerts that began in March 2016 is shown below.

The latest signals of each model are as follows:

• Ultimate market timing model: Buy equities (Last changed from “sell” on 28-Jul-2023)*
• Trend Model signal: Bullish (Last changed from “neutral” on 28-Jul-2023)*
• Trading model: Bullish (Last changed from “neutral” on 10-May-2024)*

* The performance chart and model readings have been delayed by a week out of respect to our paying subscribers.

Update schedule: I generally update model readings on my site on weekends. I am also on X/Twitter at @humblestudent. Subscribers receive real-time alerts of trading model changes, and a hypothetical trading record of those email alerts is shown here.

Subscribers can access the latest signal in real time here.

 

A Trend Model review

Over the course of several discussions with readers, it was apparent that some didn’t understand the Trend Asset Allocation Model, otherwise known as the Trend Model. This is a model that applies trend-following principles to a variety of global markets and commodities to form a composite signal.

While a history of out-of-sample weekly signals are available dating back to 2013, there is no actual portfolio return track record. However, a simulated strategy of using the out-of-sample signals to either overweight or underweight the S&P 500 by 20% around a 60% S&P 500 ETF (SPY) and 40% 7-10 year Treasury ETF (IEF) would have yielded significantly better returns with 60/40 like risk.

 
This week I review the model’s internals to reveal why I am bullish on equities.

The full post can be found here.

A 2022 report card

Preface: Explaining our market timing models 

We maintain several market timing models, each with differing time horizons. The "Ultimate Market Timing Model" is a long-term market timing model based on the research outlined in our post, Building the ultimate market timing model. This model tends to generate only a handful of signals each decade.

The Trend Asset Allocation Model is an asset allocation model that applies trend-following principles based on the inputs of global stock and commodity prices. This model has a shorter time horizon and tends to turn over about 4-6 times a year. The performance and full details of a model portfolio based on the out-of-sample signals of the Trend Model can be found here.

My inner trader uses a trading model, which is a blend of price momentum (is the Trend Model becoming more bullish, or bearish?) and overbought/oversold extremes (don't buy if the trend is overbought, and vice versa). Subscribers receive real-time alerts of model changes, and a hypothetical trading record of the email alerts is updated weekly here. The hypothetical trading record of the trading model of the real-time alerts that began in March 2016 is shown below.

The latest signals of each model are as follows:

• Ultimate market timing model: Sell equities*
• Trend Model signal: Neutral*
• Trading model: Neutral*

* The performance chart and model readings have been delayed by a week out of respect to our paying subscribers.

Update schedule: I generally update model readings on my site on weekends. I am also on Twitter at @humblestudent and on Mastodon at @humblestudent@toot.community. Subscribers receive real-time alerts of trading model changes, and a hypothetical trading record of those email alerts is shown here.

Subscribers can access the latest signal in real time here.

Some hope for the future

Every market cycle is different, but all bear markets share some common elements, namely that asset prices fall. 2022 was unusual inasmuch as both stock and safe haven Treasury prices fell together. Nevertheless, there is some hope for the future.

Nine years ago, Jesse Livermore found that forward 10-year equity returns were inversely correlated to household equity positioning. The equity bear market has sent household equity exposure skidding. I would further argue that the normalized position is actually lower. Normally, bond prices rise during equity bear markets, which raise bond allocations and depress stock allocations. When stock and bond prices fell in tandem in 2022, the diversification effect was lost. Household equity allocations should have been lower in a "normal" bear market (see The hidden story of investor capitulation).

As investors bid goodbye to 2022, here is how my models performed during a difficult year.

The full post can be found here.

Sentiment whipsaws are masking the bear trend

Preface: Explaining our market timing models 

We maintain several market timing models, each with differing time horizons. The "Ultimate Market Timing Model" is a long-term market timing model based on the research outlined in our post, Building the ultimate market timing model. This model tends to generate only a handful of signals each decade.

The Trend Asset Allocation Model is an asset allocation model that applies trend-following principles based on the inputs of global stock and commodity prices. This model has a shorter time horizon and tends to turn over about 4-6 times a year. The performance and full details of a model portfolio based on the out-of-sample signals of the Trend Model can be found here.

My inner trader uses a trading model, which is a blend of price momentum (is the Trend Model becoming more bullish, or bearish?) and overbought/oversold extremes (don't buy if the trend is overbought, and vice versa). Subscribers receive real-time alerts of model changes, and a hypothetical trading record of the email alerts is updated weekly here. The hypothetical trading record of the trading model of the real-time alerts that began in March 2016 is shown below.

The latest signals of each model are as follows:

• Ultimate market timing model: Sell equities*
• Trend Model signal: Neutral*
• Trading model: Bearish*

* The performance chart and model readings have been delayed by a week out of respect to our paying subscribers.

Update schedule: I generally update model readings on my site on weekends. I am also on Twitter at @humblestudent and on Mastodon at @humblestudent@toot.community. Subscribers receive real-time alerts of trading model changes, and a hypothetical trading record of those email alerts is shown here.

Subscribers can access the latest signal in real-time here.

A risk-on stampede?

I pointed out in the past that risk appetite in 2022 can largely be attributable to changes in the USD. The S&P 500 has shown a close inverse correlation to the greenback. Now that the USD has decisively violated trend line support, does that mean that it's time for investors to stampede into a risk-on trade?

What are the fundamentals that explain the technical breakdown in the USD? Has the Fed signaled that it is about to out-dove the European Central Bank and other major central banks, which would narrow interest rate differentials and weaken the dollar? Will other central banks out-hawk the Fed?

The full post can be found here.

An upcoming seismic market shift in factor returns

Preface: Explaining our market timing models
We maintain several market timing models, each with differing time horizons. The "Ultimate Market Timing Model" is a long-term market timing model based on the research outlined in our post, Building the ultimate market timing model. This model tends to generate only a handful of signals each decade.

The Trend Asset Allocation Model is an asset allocation model which applies trend following principles based on the inputs of global stock and commodity price. This model has a shorter time horizon and tends to turn over about 4-6 times a year. In essence, it seeks to answer the question, "Is the trend in the global economy expansion (bullish) or contraction (bearish)?"

My inner trader uses a trading model, which is a blend of price momentum (is the Trend Model becoming more bullish, or bearish?) and overbought/oversold extremes (don't buy if the trend is overbought, and vice versa). Subscribers receive real-time alerts of model changes, and a hypothetical trading record of the those email alerts are updated weekly here. The hypothetical trading record of the trading model of the real-time alerts that began in March 2016 is shown below.

The latest signals of each model are as follows:

• Ultimate market timing model: Buy equities*
• Trend Model signal: Neutral*
• Trading model: Bearish*

* The performance chart and model readings have been delayed by a week out of respect to our paying subscribers.

Update schedule: I generally update model readings on my site on weekends and tweet mid-week observations at @humblestudent. Subscribers receive real-time alerts of trading model changes, and a hypothetical trading record of the those email alerts is shown here.



A seismic shift ahead
Last week, I highlighted the rising bifurcation of US and non-US equity markets (see The stealth decoupling sneaking up on portfolios). Further factor analysis reveals a possible seismic shift in cross-asset and factor return patterns, beginning with a steepening yield curve that is signaling better economic growth expectations.

The full post can be found here.


Announcing our Thanksgiving $2 Sale!
We are proud of our Trend Asset Allocation Model, which showed a steady record of improving performance and reducing risk against a passive 60/40 benchmark in a simulation using actual signals.

We are announcing our Thanksgiving Sale $2 sale, where you can get 14 months for the price of a 12 month annual subscription, plus $2 off! Just use the coupon code Thanksgiving2019 when you sign up for an annual subscription (two month adjustment will be made within 24 hours after checkout). This offer expires at midnight, Pacific Time, on US Thanksgiving weekend (December 1, 2019). Subscribe here.

Existing subscribers who would like to extend their subscription for 14 months at the price of 12 at their current subscription price, please drop me a note by email.

A 5+ year report card of our asset allocation Trend Model

For years, I have been publishing the readings of my Trend Model on a weekly basis. As a reminder, the Trend Model is a composite model of trend following models as applied to global stock prices around the world, as well as commodity prices.

The model has three signals:

• Bullish: When there is a clear upwards, or reflationary, global trend
• Bearish: When there is a clear downwards, or deflationary, global trend
• Neutral: When the trend signals are not discernible

The first derivative of the Trend Model, i.e. whether the signal is getting stronger or weaker, and combined with some overbought/oversold indicators, has performed admirably as a high turnover trading model (see My Inner Trader and this ungated version for non-subscribers). However, I have never produced a full report card for the Trend Model. While the actual signal dates were always available on the website, I never got around to compiling the performance record because I was always tied up on other projects, and the task never got to the top of the pile.

After several repeated requests from readers, here is the report card of the Trend Model. I want to make clear that this study represents the real-time track record of actual out-of-sample signals. These are not backtested. The results were solid, and the analysis was also revealing about what an investor should expect when using this model for asset allocation.

The full post can be found here.

How Covel inadvertently exposed the chasm between investors and traders

As a rule, I don't do book reviews. However, regular readers know that I am a big fan of trend following models and I use them extensively in my asset allocation work. When a publicist offered a free review copy of Michael Covel's Trend Following, 5th Edition: How to Make a Fortune in Bull, Bear and Black Swan Markets, I jumped at the chance.

The book also featured a forward by Barry Ritholz. Ritholz's partner Josh Brown recently wrote that they use trend following techniques for tactical asset allocation, which is a sensible decision that I wholeheartedly agree with:

At my firm, we use trend for tactical asset management. It takes everything above into account- not only the things, but people’s actual reaction to the things, a sort of realpolitik for markets. Will it always work? Doubtful. What is the downside when it doesn’t work? What is the expected benefit when it does? Is there a behavioral aspect to why it makes sense to include tactical in client portfolios? We think so. Not everyone would agree that this is worthwhile.

After all that buildup, the book left me vaguely disappointed. Covel's approach has been to be a cheerleader for traders who use trend following techniques without digging into the deeper issues that face investors. He treats trend following almost as a magic black box that everyone should use. He doesn't take the next step to discuss the characteristics of this class of strategies, which sophisticated investors think about when they consider the use of such techniques have a place in their portfolio.

The full post can be found at our new site here.

Revealing the secret behind trend following models

The blogger Jesse Livermore at Philosophical Economics recently wrote another brilliant post about the use of trend following models and market timing. He found that trend following models work very well on diversified stock indices, but didn't really understand the mechanism of how they worked. As I pride myself on being a left and right brained quant, I am going to try and explain why these classes of models work and why they perform poorly on individual securities.

The full post is at our new site here.



Site Notice
If you signed up for the notification of our Early Bird Special Offer but did not receive it, I have had some reports that they got classified as spam so please check your spam folder. Otherwise you can email me at cam at humblestudentofthemarkets dot com and I will send you the link.

As well, I would like to remind readers that we will cease to accept new subscribers as of January 15, 2016 as a way to better control the growth of our new community. Hurry, the deadline is coming up faster than you think.

How to make your first loss your best loss

I have always found that the time when I have learned the most about an investment process is when it does not perform well. After all, it is during periods of drawdown that the blemishes that appear in a model and no amount of backtesting will warn you of those shortcomings. If you are willing to learn from periods of underperformance, you can, in effect, make your first loss be your best loss. This is a case study of how I learned to diagnose a model`s shortcomings and learn from that experience.


Trend Model underperforming
My Trend Model has seen some stellar returns for the past 18 months (last report card here), but had experienced some spotty performance in the past few months. While the returns were not disastrous, they were not up to the levels seen in earlier periods. A look at buy and sell signals in 2015 show a pattern of market choppiness and signal whipsaw:

In a way, that`s not a surprise. Regular readers will know that the Trend Model is based on the application of trend following techniques to global stock and commodity prices. The price volatility experienced for most of 2015 has led to an environment that is unfriendly to trend following models.

Is this just a "feature" of these kinds of models that has to be endured?


A long and short term regime change
Maybe not. In a recent post (see 3 secrets from the Book of (Trend Following) Revelations), I highlighted a study by James Paulsen of Wells Capital Management showing that the long-term trend in equity prices were getting overdone. Paulsen found that stock prices had more or less gone up in a straight line in the last few years and with only minor corrections.

He went on to calculate the rolling three-year R-squared of stock prices and found that they follow cycles of high and low levels of price trends. We just happen to be at the top end of a trending period, which is likely to end soon. When it ends, it will signal an intermediate term top for the US stock market.

In a more recent post (see Calling an audible (for more choppy markets)), I reproduced the Paulsen study and further extended some of the conclusions of that research.

Instead of just looking at the R-squared of 36-month rolling regression, I also examined the R-squared of a 6-month rolling regression as a measure of the short-term trend, largely because my Trend Model uses a much shorter lookback period and it is therefore more correlated with short-term trends than long-term trends. In my last post, I reported that I found that we are in the period where the long term trend is strong (top panel), but the short-term trend is weak (bottom panel).

I further showed that stock prices are likely to roll over in the next few months, given how extended the long-term trend is. In addition, the rollover of the short-term trend is likely a leading indicator of the long trend.

Going back to 1900
I am grateful for all the comments, feedback and suggestions that I have received since I started writing on this topic. One of the comments that I received is that the sample size of this study is absurdly low. I could extend the lookback period of the study using data for the Dow, which goes back to 1900, instead of the SP 500, which only goes back to 1950.

With that suggestion in mind, I reproduced the study using DJIA going back to 1900. Here is the R-squared chart. While the readings are slightly different from the SP 500 study, the longer time horizon revealed some interesting insights. First, the incidence of high 36-month R-squared readings was higher in the pre-1950 period. Nevertheless, the current reading of 0.957 is comparable to the reading of 0.930 in pre-Crash 1987 and 0.911 in pre-Crash 1929. Similarly, the spread between the 36-month and 6-month R-squared readings (bottom panel) is also at a similar order of magnitude when compared to 1987 and 1929.

There is an important caveat to remember! This model measures the direction of the move and not the magnitude. Just because the trend is so extended today doesn't mean that a market crash is around the corner. Other episodes have resoled themselves in 10-15% corrections.

Nevertheless, based on the current 36-month to 6-month R-squared spread of 0.844, I looked at what the return pattern of the DJIA was during past episodes with similar characteristics. The sample size was a more reasonable 16, compared to the minuscule N=4 in the SP 500 study that went back to 1950. The market outperformed initially, but rolled over at between 3-6 months after the first time the spread went above 0.8 (which was March 2015).

And if the trend got even more extended and the 36 to 6 month spread went to 0.9? The results were more dramatic, as the market declined almost immediately.

Historical analysis from Dana Lyons found a number of narrowly range-bound markets that appeared to coincide with tops in 36-month R-squared readings, though the samples did not totally overlap. Lyons had some good news and bad news for stock investors:

First, the good news. All 9 of the prior ranges saw the SP 500 eventually go on to make new highs, though some initially broke the range to the downside first. 2 months later, 8 of the 9 instances saw the SP 500 not only higher but at a new 52-week high. Only the 2007 instance saw the index lower 2 months later, although it was at a new high 3 months afterward (and it had made a new high immediately following the range break).

Now the not-so-good news. On 5 of the 9 occasions, the new highs were very small and very short-lived. Following occurrences in 1965, 1976, 1983, December 1993 and 2007, the market’s upside “breakout” resulted in tops shortly afterward that would predominantly hold for the following year. Only the occurrences in 1951, 1964 and 1995 saw the market persist at new highs for an extended period of time. Furthermore, only 1995 saw the SP 500 continue on to double digit returns over the following year.

In other words, the market did perform well initially, just as my analysis shows, but they more often than not marked a significant intermediate term top.

Weakening long-term trend, choppy short-term trend
My research results using DJIA data going back to 1900 confirm the conclusions of past studies. We are in an environment where the long-term trend is strong but starting to weaken. The weakness is evidenced by a faltering and choppy short-term trend, or price momentum.

For someone using a trend following techniques with lookback periods similar to the shorter 6-month trend, this suggests that the current unfriendly period for this kind of trend following model is temporary. Under these circumstances, I can choose from four options going forward:

1. Status quo: Continue to run the Trend Model as is and accept the drawdowns as a "feature" of the model.
2. Wait for a friendly environment: Go to cash and wait for signs that the trending environment has re-established itself.
3. Focus on the long-term trend and ignore the short-term trend: Most trend following models use a long term moving average to define the trend, e.g. 200 days, and a short-term moving average for risk control, e.g. 50 days. This approach would throw away the shorter (50 dma) and focus on the longer (200 dma) for trading signals.
4. Focus more on counter-trend models. Examples would be contrarian sentiment models looking for crowded longs and shorts, as well as overbought-oversold trading models.

I rejected options 1 and 2 out of hand. Following the status quo is an example of not learning about the investment process and not allowing your first loss to teach you a lesson. Going to cash is a cop-out and detracts from learning. I did consider option 3 seriously, but rejected it as it would result in excessive return volatility because of it strips away the risk control element out of the investment process.

I wound up adopting a version of option 4. I would focus on one side of the counter-trend models by fading strength (but not buying weakness). As my analysis indicates that the long term trend is turning and the decline from an intermediate term top could only be weeks away, buying weakness is the equivalent of picking up pennies in front of a steamroller. Selling strength when the market is overbought, on the other hand, is likely to be a higher percentage play.

The acid test of learning
In conclusion, this is a case study of how I learned about a model from drawdowns. When I interviewed investment managers in the past, I have always asked the acid test question about how they have learned, "Under what circumstances would your investment approach fail?"

If the manager has thought sufficiently about his strategy and he has learned from past mistakes, he will give an intelligent answer. It also shows that he sufficiently understands the kinds of bets that he is making, why it works and when it might fail.

That`s how you learn to make your first loss your best loss.

Calling an audible (for more choppy markets)

Trend Model signal summary
Trend Model signal: Neutral
Trading model: Bearish

The Trend Model is an asset allocation model which applies trend following principles based on the inputs of global stock and commodity price. In essence, it seeks to answer the question, "Is the trend in the global economy expansion (bullish) or contraction (bearish)?"

My inner trader uses the trading model component of the Trend Model seeks to answer the question, "Is the trend getting better (bullish) or worse (bearish)?" The history of actual (not backtested) signals of the trading model are shown by the arrows in the chart below. In addition, I have a trading account which uses the signals of the Trend Model. The last report card of that account can be found here.

Update schedule: I generally update Trend Model readings on my blog on weekends and tweet any changes during the week at @humblestudent.


Market rally, or more chop ahead?
I suppose that with the rally in risky assets, the trading model should get upgraded to bullish from bearish. In the US, the SPX has recovered and exceeded its 50 day moving average (dma). European bourses have strengthened and Asian markets, led by Shanghai and Hong Kong, have been on fire. These are the classic signs of a global bull trend.

I remain skeptical. Recent experience has shown that US equity prices has shown the tendency to reverse itself. As the chart below shows, the market has done the stutter-step and crossed the 50 dma, which is a reasonable proxy for a medium term trend, a total of 13 times in 2015, compared to only four times in the same period in 2014. By last Friday, the market was either overbought or mildly overbought on the 5 and 14 day RSI (top two panels) and at levels where rallies have petered out. At the same time, the VIX Index, which tends to be inversely correlated to the market, is now testing a key support level where it has stopped declining in the past.

Will the bulls be successful in pushing price up to test the old highs, or will the market roll over as it has done for most of this year? Despite the bullish technical and fundamentals underpinning this latest move, I am leaning towards the latter scenario, where the market falls to the bottom of the trading range.


Possible regime change = Intermediate term top
Let me explain why. I wrote about a possible change in the technical market regime in my last post (see 3 secrets from the Book of (Trend Following) Revelations). I referenced a study by James Paulsen of Wells Capital Management, where he showed that the long-term trend was getting a little too good and therefore investors were getting a little too complacent. To recap, the Paulsen study showed that US stocks have been going up more or less in a straight line since 2012, with few drawdowns. By contrast, 2011 saw a reversal of the upward price trend from the 2009 bottom as it grappled with both an impasse in Washington and a eurozone crisis.

Paulsen went on to calculate the rolling 3-year R-squared, which measures the fit of a trend-line, of the market and showed how it had up and down cycles. When the fit is too good, it can breed investor complacency and the stock prices ultimately reverse themselves. In other words, we may be on the verge of an intermediate term top.

I was sufficiently intrigued by the Paulsen study that I downloaded month-end SPX prices going back to 1950 from Yahoo Finance to reproduced the results and add some of my own insights. As the top panel of the chart below shows, the rolling 36-month R-squared reached an astounding 0.976 at the end of March 2015. This was the second highest reading in its history, only to be exceeded by an episode in July 1956 with a 0.977 reading.

The bottom panel of the chart shows the difference between the rolling 36-month R-Squared and the rolling 6-month R-squared. When the readings are high, it indicates that the long-term trend is much stronger than the short-term trend and when the reading is low, the short-term trend is much stronger than the long-term trend. We have an instance where the long-term trend is strong, but the short-term trend is weak, which can be seen by the choppy up-and-down markets this year.

I have marked with vertical red lines to show how circumstances today are very similar to the instance in July 1956, when the market was trending strongly long-term but choppy short-term. Then, stock prices soon turned down, but the correction was relatively mild as the decline was only about 10%.

Current conditions are setting up the same way. I took a look at what happens to stock prices when the 36-month R-Squared first exceed 0.95, which occurred last year in June 2014. While the sample size is small, it does give us a rough roadmap as to what happens next. Stock prices continued to rise initially because of the powerful momentum underlying the long-term trend. They then peak out at between 6-12 months, and the decline bottoms out at around 18 months afterwards. We are now 9 months into this period.

While the number of data points in this study is very small and therefore confidence is low, let me try to torture the data a little further so that it talks. As the 36-month R-squared is 0.976, or the second highest level in its history, I looked at what happened when R-squared was above 0.97. The data points only fell by one (N=3) and the results are instructive. The market decline started almost right away, followed by a rally and a fall into the ultimate low in 18 months.

Watch your time horizon
There is a couple of important caveats to this analysis. I would urge investor not to panic, as this approach seems to be good at forecasting direction, but not necessarily the magnitude of the market move. As there is no fundamental trigger for an all-out bear market in the form of a recession on the horizon, my base case scenario calls for a correction, much like the episode seen in 1956.

As well, traders need to keep in mind that the time horizon for this analysis is in months, while the time horizon for many traders is in days or even minutes. While stock prices are likely on the verge an intermediate term top, that doesn't necessarily mean that the market goes down right away. Bret Steenbarger wrote a series of excellent posts where he observed the conditions of strong long-term trend, but weak short-term momentum. Here is one example:

The recent rise in realized volatility and vol of vol helps to explain why short-term traders have had difficulty trading U.S. stocks: We've had movement, but not trend. It's been a choppy, volatile range trade. Buying strength or selling weakness--trading with a fear of missing out--has been a consistent way to lose money. Viewed on a longer time scale, however, this market appears to be one of narrowing breadth--even among those smaller cap sectors that have been strongest. That keeps me cautious

In subsequent posts, he adopted the trader's perspective to say that current breadth readings does not support an imminent bear attack. One example here:

As a whole, the signals are showing reduced breadth of market strength, but not net weakness--consistent with the waning breadth readings noted in yesterday's post.

And here:

While a market without weakness might appear "overbought", in fact it is healthy in the short run. In order for the broad stock market to roll over, we have to see some leading stocks and sectors display weakness. At the moment, we are indeed "overbought", but not weak.

In other words, be cautious, but don`t get too bearish.


A test next week for bulls and bears
If I am correct in my analysis, we are in for more choppy markets. The week ahead will be a key test for my thesis. On one hand, overseas markets have rallied, which is bullish from an inter-market analysis viewpoint. As well, the upcoming week is option expiry week, which has historically had a bullish bias. This study from Quantifiable Edge shows that April OpEx has been one of the most positive weeks in the year:

Earnings Seasons starts in earnest next week. The latest update from John Butters of Factset shows that Street forward 12 month EPS estimates are edging up again, which is a positive sign for equities:

On the other hand, the market is short-term overbought and it has paused at these kinds of levels in 2015. In addition, past analysis from Ryan Detrick showed that the reaction to Alcoa earnings can be an early portend of stock market action. (Recall that Alcoa beat expectations on EPS, but missed on revenue estimates and the stock tanked as a result).

No doubt there will be lots of headline driven volatility in the days ahead. But if stock prices can't manage to stage a meaningful advance next week in light of the bullish tailwinds, it will be a validation of my choppy market and intermediate term top thesis.

My inner investor is cautious, but he is setting his stock-bond mix at policy weights. My inner trader is cautiously short the market with tight stops in order to to scalp a possible move to the bottom of the recent trading band.


Disclosure: Long SPXU, SQQQ

3 secrets from the Book of (Trend Following) Revelations

Good investors know about the adage that diversification is the only free lunch in investing. But diversification can mean more than just asset class diversification or the number of stocks in a portfolio, it can also refer to model diversification, where you combine models with uncorrelated return streams.

Regular readers will also know about the investment results of my Trend Model (last report card here). The Trend Model is based on the signals of a trend following model as applied to global stock and commodity prices. One of most naturally diversifying class of models to trend following are counter-trend strategies, such as overbought-oversold models or contrarian sentiment models. In the course of studying the dynamics of trend following models and counter-trend strategies, I came to three important observations:

1. Bull and bear markets behave differently and therefore they should be traded differently. Bull markets tend to get overbought a lot more easily than bear markets get oversold, which is another way of saying that bear markets are far more emotional than bull markets.
2. We may be on the verge of an intermediate term top in US equities. Just as markets become "overbought" and "oversold", so can models. Trend following strategies, when applied to US equities, are becoming "overbought". Momentum is starting to fade and we may be seeing a top building in the US stock market. That process has led to a choppier market, which has created difficulty with trend following returns, but...
3. Trend following strategy returns are subject to overbought and oversold conditions too. As the market tops and rolls over, trend following returns will disappoint because of the whipsaws as price momentum flags. However, as the market transforms from a bull to a bear market, that`s when long-only investors need trend following the most as these strategies will go short and profit from falling prices - which is precisely when the diversifying effects of this class of strategy is most valuable. Hence, a good time to commit new funds to a trend following program is when it is suffering from a period of poor returns.

Let me explain.



Has it been too good for trend following?

The chart below shows the SPX for the last two years. I have imposed a standard 50 and 200 day moving average (dma) as a proof of concept of a trend following system. Supposing that we were to use a simple trading rule of:

• Buy: When the price is above the 50 and 200 dma
• Short: When the price is below the 50 and 200 dma
• Hold cash: All other times

You would have done very well during this period. As the stock market has more or less gone up in a straight line, you would have participated in the rallies while limiting downside drawdown risk during the past two years.

In the bottom two panels, I showed the Percent Price Oscillator (PPO), which measures how far the price of SPX is away from the 50 and 200 exponential moving average (ema) is in percentage terms. The 50 day PPO has varied between +5% and -5% during this period and the 200 day ema has been mostly positive during this time (largely because of the steady market uptrend), but has topped out at around 10%.

Based on these observations, the time to be hypersensitive to counter-trend strategies is when the PPO is "overbought" or "oversold", defined as +/- 5% on the 50 day PPO and +/- 10% on the 200 day PPO. There was one occasion during the last two years - and that was marked by the dotted vertical line on the chart.

Interesting research observation? Not so fast!


How bulls and bears are different
When I stretched my research window out from two to 15 years, the conclusions changed. The chart below shows the same analysis for 15 years. I have also added the 50-200 PPO in the top panel, which measures how far away the 50 ema and 200 ema are away from each other in percentage terms.

When I examine the PPOs, I noticed that the overbought and oversold levels are very different for bull and bear markets. Bull markets tend to get overbought much more quickly, while bear markets can decline a lot further before the price starts to revert. A more realistic overbought level for the 50 day PPO is 5%, compared to -10% for an oversold reading. Similarly, the threshold for the 200 day PPO is 10% and -20%. We can also observe the same effect in the 50/200 day PPO in the top panel - bull markets get overbought more easily than bear markets get oversold.

My first conclusion: The overbought and oversold thresholds should be asymmetric for bull and bear markets. Apply counter-trend strategies differently, depending on whether you're in a bull or bear phase.


Flagging momentum = Intermediate term top?
Another important observation from the chart above is that we are seeing the PPO starting to fade at a both 50 and 200 day level (marked by the arrows on the chart). Fading PPO means that short term price momentum is flagging. Even though the trend is positive, buying short-term momentum yields lower returns as the strength of momentum surges seem to fade with every rally episode. These conditions are consistent with a recent comment in a post by Brett Steenbarger:

Overall, chasing new highs and stopping out of long positions on expansions of new lows has brought subnormal returns. We have had a trending environment since 2012, but not a momentum environment. Understanding that distinction has been crucial to stock market returns.

In the past, such episodes have see either full-blown bear markets (2000-03, 2007-09) or corrections (2004, 2011), which are marked on the chart.

Research from James Paulsen of Wells Capital Management seem to support that conclusion that we may be on the verge of a market top. Paulsen observed that stock prices have more or less gone up in a straight line. The chart below fitted a trend line to stock prices for the past three years and found that the R-squired came to 97%. In other words, we could be setting up for a Minsky Moment for US stocks (my words, not his).

Given that kind of fit, is it any wonder that trend following models have performed so well! Paulsen showed that stock prices haven`t always moved in a straight line as it has recently. Here is the R-squired of rolling fitted trend lines for stock prices since 1900. Note how R-squared has oscillated between 0 and 1 in the last 115 years.

This is starting to look an overbought-oversold model! In that case, could trend following model returns be getting "overbought"? Maybe. Paulsen went on to show the returns and batting averages of returns by R-squared deciles (recall that the latest R-squared is 97%, or leftmost decile).

Based on Paulsen`s research and the observation that short-term price momentum is rolling over, we could be seeing early warning signs of an intermediate stock market top. However, I am not panicking just yet. The market tops in 2000 and 2007 were marked by well-defined fundamental bearish drivers, but I am not seeing similar bearish signs from the macro data. If the stock market were to see an intermediate term top, then the most likely outcome is a corrective episode like we saw in 2004 or 2011.


Bear phases are profitable for trend followers
Further research into the characteristics of trend following models led to a more subtle observation. Trend following models are excellent diversifying strategies during bear phases because these kinds of strategies offset long-only portfolio losses because they can and will short the market during bear phases.

Research from AQR Capital Management on what they call "momentum", which are really trend following strategies illustrate my point. The x-axis is the return of the SPX, while the y-axis represents an indicative return of a "price momentum" strategy. As the chart shows, as stock prices go negative, momentum relative returns rise to compensate to cushion the effects of the market decline.

Here is another way of thinking about current market conditions. This is an idealized scenario, but bear with me on this. Stock prices have more or less been going up in a straight line for the last few years (trending), but short-term price momentum is fading and price trends are getting choppy. As a result, trend following returns start to disappoint. If the intermediate term top scenario unfolds as predicted, then stock prices will start to fall. Trend following will see positive returns once more as the model shorts the market.

Such a scenario is consistent with past research which indicates that drawdown periods are excellent opportunities to increase allocation to trend following strategies. As an example, Attain Capital did a study of trend following programs and allocated capital according to the following three sets of rules:

We first tested the ‘normal’ performance, compounding each manager’s returns individually, and then summing the performance of each (you can’t sum the returns and compound the total, as each manager is only trading the portion of the total allocated to them, and isn’t increasing or decreasing positions based on the movements of the overall portfolio). This is the hold tight, or ‘call’ method, where you rely on the portfolio as a whole to protect you against drawdowns in individual managers.

We then tested the ‘fold’ method, whereby you set a line in the sand at 1.5 times max drawdown, exit the program if that level is hit, and then replace it with the program showing the highest past 12 month Sharpe ratio that is not already traded or has been traded. This method resulted in 7 different programs coming into the portfolio over the 8 year period, as three of the original programs hit their lines in the sand; and four of the replacement programs hit their stop points. We started this portfolio with even weightings for each manager, and then used that same initial weighting for a replacement program if it was replacing a program which had lost money; and used the current allocation of the program being replaced if it had made money since inception.

Finally, we tested the ‘raise’ method, where you don’t just sit tight and you don’t head for the exits; you actually stare down the drawdown and increase your allocation. For this method, we tested using 50% of the past historical drawdown and 100% of the past historical drawdown as trigger points for doubling the allocation to whichever manager hit those drawdown levels. We also assumed this double allocation was only done until the program returned to equity highs, and that it was done with no extra capital - only with an increase in the nominal trading amount.

As it turned out, the "raise" method, which doubles down on losing trend following programs, had the best returns. It beat the buy-and-hold benchmark, which outperformed the "fold" method, which panicked out of programs that saw excessive drawdowns.

Conceptually, these results are consistent with the idea that trend following strategies can have overbought and oversold conditions. The research from Wells Capital Management, which showed that 3-year R-squared can oscillate between very low (high drawdown) levels to very high (high returns) is another way of thinking about this issue.

Don't panic!
One last word.

Despite my conclusions about a possible top, I would reiterate my observation that investors shouldn't panic over the prospect of a possible intermediate term top in US stocks. The fundamental drivers of a major bear market are absent and if the market should top out, it will most likely be a correction.

It would be entirely appropriate for James Paulsen should have the last word on this matter:

Our advice is to stay overweighted equities but to diversify away from the U.S. toward offshore stock markets. Most international markets have underperformed U.S. stocks in the last few years and currently offer more attractive relative valuations. Moreover, investors can diversify away from increasingly hostile U.S. policy officials toward hospitable policies for the foreseeable future in both the eurozone and in Japan.

Don't panic, diversify - it's the only free lunch in investing.

A deceptively simple way to make money in the stock market

In a recent post, David Merkel had a warning about the speculative excesses of the stock market. He cited the tendency of Silicon Valley startups to offer downside protection to investors on their own paper:

Here’s my take. When companies try to offer protection on credit or market capitalization, the process usually works for a while and then fails. It works for a while, because companies look best immediately after they receive a dollop of cash, whether via debt or equity. Things may not look so good after the cash is used, and expectations give way to reality.

He also found the trend of finance professionals leaving for Silicon Valley disturbing:

We saw this behavior in the late ’90s — people jumping to work at startups. As I often say, the lure of free money brings out the worst in people. In this case, finance imitates baseball: those that swing for the long ball get a disproportionate amount of strikeouts. This also tends to happen later in a speculative cycle.

How worried should we about these signs?

Last week, I wrote about how valuation didn`t seem to matter to the stock market until it mattered (see Cheap or expensive? The one thing about equity valuations that few talk about). During the expansion phase of an economic cycle, the main driver of stock prices are economic and earnings momentum. It is only during the recessionary part of the cycle that valuation puts a floor on stock prices.

This framework is similar to the one voiced by David Rosenberg, who wrote that bear markets only occur because of recessions and Fed tightening cycles (via Business Insider). Well, sort of, I would generalize the cause of bear markets as recessions and Fed tightening causing recessions or creating recessionary fears.

Ed Yardeni more or less said the same thing as I did (emphasis addded):

A long expansion is a persuasive argument for buying stocks even though forward P/Es are historically high. Investors are likely to be willing to pay more for stocks if they perceive that the economic expansion could last, let’s say, another four years rather than another two years. The more time we have before the next recession, the more time that earnings can grow to justify currently high valuations.

If a recession is imminent, stocks should obviously be sold immediately, especially if they have historically high P/Es based on the erroneous assumption that the expansion’s longevity will be well above average. Bull markets don’t die of old age. They are killed by recessions.

A simple (and stupid) rule
In that case, we can create a some very simple rule for timing the stock market:

1. Buy stocks during economic expansions
2. Sell stocks as recessions approach

At this point, you may say, "Don't be an idiot, that's like saying buy stocks when they go up and sell them when they go down. Or 'Buy stocks when they go up, when they go down, don't buy them in the first place.'"

How can you forecast a recession?

The key issue is how we can tell if a recession is on the horizon.  Already, the likes of David Merkel are warning about the appearance of speculative froth, which is a sign that we are nearing the top of the cycle.

There are several deceptively simple approaches to watch for a recession. Doug Short has his Big Four Recessionary Indicators, which is a useful framework to use. New Deal democrat analyzes high frequency economic releases (his latest one is here). Neither is flashing recessionary signals at the moment.

Real-time forecast signals

While those kinds of approaches have great value and I use their output to supplement my model results, I find them unsatisfying. That's because you are using coincidental or lagging indicators to forecast a leading indicator. Economic statistics, which get released with a time lag (and may be a coincidental or lagging indicator). On the other hand, stock prices represent a leading indicator. 

This has been the Achilles Heel of macro forecasting, using economic statistics (which are released with a lag) to forecast stock and bond prices (which are forward looking) will inevitably cause you to miss the early warning signs of a recession.

There is a better way. There are continuous real-time signals that can give us clues about expansions and recessions. They are called market prices. Here are the three key assumptions to my model:

1. Economic signals about expansion and contraction are persistent. Once an economy start to lose steam and roll over into recession, it will continue to do so until the fiscal and monetary authorities react (with a long lag). That`s why Doug Short`s Big Four Recessionary Indicators work in forecasting recessions. These kinds of indicators just operate with a lag and you will be long stocks as the economy rolls over.
2. The US economy is an important part of the global economy. US recessions therefore will not occur in isolation, but we will see feedback loops with other parts of the global economy. 
3. There are three major trade blocs in the world: US, Europe and Asia (China). Watching what happens with these three major economies will give you more or less everything you need to know about where the global economy is going. Note that they may not all expand or contract in a synchronized way, however.

That`s why I am a technician. It`s not that I believe that there is anything magic about technical analysis. They just happen to be the right tool to use in the current situation.

From model to buy-sell signals

With those assumptions in mind, here is my deceptively simple way of forecasting the US (and global economy). Use trend following models (because of the persistent nature of economic expansions and contractions) on:

• US stock prices 
• Non-US stock prices
• Commodity prices. 

In particular, commodities can tell us a lot about how global demand is changing at the margin. In the last decade or so, they have been more about Chinese demand as China has become the major user  of commodities. 

Recognizing that the regional economies of the three major trade blocs are not always synchronized, give each of the components of the trend following model signals votes and let them tell you whether the world is in expansion or contraction. 

For investors, buy stocks when the global economy is in expansion, sell (or reduce) stock positions when they are moving into contraction.

For traders, watch for direction of the change of the signal. If the signal gets better, buy stocks. If the signal gets worse, sell stocks.

The proof in the pudding

In short, that has been the basis for my Trend Model. The proof is in the pudding and I have been running an account based on the trading signals of that model since September 2013. The latest report card can be found here. The chart below shows the history of the actual (not backtested) buy (dark blue arrows) and sell (red arrows) signals of the trading model.

The performance of the account has been extraordinarily good and well beyond my expectations. It wasn't just the levels of the returns, which was embarrassingly high, but a number of other qualities:

• The consistency of monthly returns, as the monthly batting average was over 70%;
• The skew in the distribution of monthly returns, which suggests that the strategy is limiting losses while allowing winners to run; which leads to...

• Better than expected risk characteristics in the form of drawdowns as the return to maximum drawdown ratio is an astonishing 5 to 1; and
• The superior diversification effects of the strategy, as returns were negatively correlated to both stocks and bonds.

I would caution, however, that the this strategy has been benefiting from a friendly market environment for trend following models. The US equity market has more or less gone up in a straight line in this period. 2015 is likely to see a choppier market and that`s when this model will get an acid test. In that environment, I would expect that drawdowns will occur more often and be larger, which will lead to deterioration in average returns and the monthly batting average.

I am highly encouraged by these results and I will be monitoring and reporting on the results in the future.


Disclaimer: This blog post is for discussion only and I am not trying to sell anyone anything. I am not currently in a position to manage anyone`s money based on the investment strategy that I am describing. No offering will be done without the proper regulatory filings.

Rebalancing your portfolio for fun and profit

This is part two of a two part post on portfolio rebalancing (see How to rebalance your portfolio (NAAIM maniac edition) for part one).

The standard practice among portfolio managers is to establish a rebalancing discipline for their portfolios. A typical process would involve the following steps:

1. Determine the target asset mix, which could change depending on market conditions.
2. Re-balance if:
• The asset mix weights moves more than a certain percentage, e.g. 10%, from the target weight; or
• Periodically, e.g. on an annual basis

These are all sensible rules that have long been practiced in the investment industry. In essence, the strategy involves taking profits on winning asset classes and averaging down on losers as a form of risk-control discipline.

Then I came upon an intriguing paper by Granger, Greenig, Harvey, Rattray and Zou entitled Rebalancing Risk. Here is the abstract:

While a routinely rebalanced portfolio such as a 60-40 equity-bond mix is commonly employed by many investors, most do not understand that the rebalancing strategy adds risk. Rebalancing is similar to starting with a buy and hold portfolio and adding a short straddle (selling both a call and a put option) on the relative value of the portfolio assets. The option-like payoff to rebalancing induces negative convexity by magnifying drawdowns when there are pronounced divergences in asset returns. The expected return from rebalancing is compensation for this extra risk. We show how a higher-frequency momentum overlay can reduce the risks induced by rebalancing by improving the timing of the rebalance. This smart rebalancing, which incorporates a momentum overlay, shows relatively stable portfolio weights and reduced drawdowns.

Monthly rebalancing does the worst
You would think that, for example, given the massive losses seen during the Lehman Crisis episode, that a rebalanced portfolio where the investor bought all the way down would see superior returns. Interestingly, that was not the case.

In the paper, the authors compare and contrast a simple drift weight strategy, i.e. not rebalancing at all, with a fixed weight monthly rebalancing strategy. The chart below shows how that the monthly rebalanced portfolio actually showed a higher risk profile than the passive drift portfolio. (There were other examples in the paper, but I will focus on this period for the purpose of this post).

By contrast, they advocate a partial momentum strategy. In essence, this amounts to the application of of a trend following system to rebalancing. The idea is, as the stock market goes down and the bond market goes up, you keep overweighting your winners (bonds) and don't rebalance the portfolio until momentum starts to turn. As the chart below shows, the portfolio with the partial momentum overlay performed better than either the monthly rebalanced or passive drift weight portfolio.

Talking their book?
These are intriguing results and a demonstration of the positive effects of using trend following techniques for portfolio construction. However, I would add a couple of caveats in my read of this paper.

First, three of the five authors work for MAN Group, which is known for using trend following techniques in their investing. While this paper does show the value of these kinds of techniques, I am always mindful that researchers may be "talking their own book". As regular readers are aware, I extensively use trend following models in my own work, but I am cognizant of the weaknesses of these models.

In particular, these models perform poorly in sideways choppy markets with no trends. As an example, consider this chart of sugar prices for the period from 1891 to 1938. Unless the trend following system is properly calibrated, the drawdowns using this class of model are potentially horrendous.

As another example, try wheat prices for the 1872 to 1944 period:

A second critique of the approach used by the paper is the use of monthly rebalancing as one of the benchmarks. In practice, no one rebalances their portfolio back to benchmark weight on a monthly basis. A more realistic rebalancing technique might be a rule based rebalancing approach of rebalancing either annually or if the portfolio weights drift too far from the policy benchmark.

To be fair, however, the monthly rebalancing approach is an extreme one that does differentiate between a passive drift weight and a more frequently rebalanced portfolio.


Value vs. Momentum
In summary, this is an intriguing paper that compares and contrasts the use of price momentum, or trend following, techniques of chasing winners to a value-based rebalancing strategy of buying assets when they are down.

Before going out and blindly implementing a trend-following based re-balancing program, I urge portfolio managers to further study these approach and adopt it to their own circumstances. Your mileage will vary.