Performance Review - Four Years ending July 20, 2018

Last week marked four years since the start of the Focused 15 Investing publication.  This report provides information about how the investment approach has worked over this period.  This report discusses the performance of two model portfolios for the 4-year period ending July 20, 2018 and compares them to various reference points.  An appendix also shows results for all portfolios currently listed in Focused 15 Investing publications.  

The two model portfolios are:

1. Diamond (sg131), which is the most widely used model portfolio. It holds three ETFs. The signal set that drives Diamond, which I call “D5”, is the primary loss-avoiding component used in all model portfolios. It is designed to provide high absolute returns and avoid losses. 
2. “US Ind, LC/SC, Multi-Sector Bonds” (sg129). It holds eight ETFs and places greater emphasis on maintaining its performance advantage in upward trending markets. It includes the D5 signal set but also includes relative leadership sleeves for 1) large company (LC) vs. small company (SC) stocks, and 2) various bond sectors. Loss avoidance is not its primary goal. Portfolio sg129 was in the original publications 4 years ago. It is in the publication for institutional investors, which focuses on model portfolios aiming to provide strong returns relative to a benchmark.

The longer-term performance highlights the impacts of the design differences. The table below shows performance[1] characteristics since January 7, 2000.

	Annualized Return %	Variability[2] %	Maximum Drawdown[3] %
	Higher is better	Lower is better	Smaller loss is better
Diamond (sg131)	25.4	14.1	-15.6
Sg129	15.1	11.0	-13.8

As one can see in the table above, Diamond has a much higher return than sg129 (25.4 % vs. 15.1%).  At the same time, sg129 had a loss (maximum drawdown) just about as great as Diamond’s.    

Before discussing the performance of these model portfolios over the last four years, I’ll review the market environment. 

 Market Environment

The recent four-year period has seen above average rates of return and below average return variability compared to the last roughly 100 years of the DJIA.  The chart below shows the Price of the DJIA (brown line) on the left-hand scale (log). 

The chart also shows the percentile rank of the rolling 4-year returns for the DJIA (green) across the 100-year period and the percentile rank of the variability of the 4-year returns (yellow), both on the right-hand scale (the horizontal black line is at the 50^th percentile on the right-hand scale ranging from 0.00 to 1.00).  The rightmost point of the green line indicates that the 4-year return for the DJIA is at the 64^th percentile – a moderately high return.  The rightmost point on the yellow line indicates that the variability of returns is at the 33^rd percentile – a moderately low level of variability. 

                                                Source:  Bloomberg LLC and CPM Investing LLC

Generally speaking, Diamond produces better returns than its default when the market’s variability (yellow line) is high.  Portfolio sg129 is designed to do better when the market’s returns (green) are high.  This information indicates that the environment was more favorable to model portfolio sg129 than to Diamond.  Diamond has just three ETFs and is designed to avoid losses during periods of high return variability. 

In contrast, model portfolio sg129 (eight ETFs) is designed to outperform its default in positively trending markets, such as what we have experienced the last 4 years.  That said, the absolute return of Diamond is higher, making it attractive to end users in many environments. 

Traded Portfolios Compared to their Defaults

The chart below shows the performance of Diamond and its default.  The blue line represents the performance of using the dynamic target weights in the weekly publication.  The tan line indicates the performance of the same ETFs held at default weights over time.  As indicated in the material describing the Focused 15 Investing approach, the traded portfolio (blue line) can underperform its default (tan line) in the later stages of an upward trending market.  We can see this below in the narrowing of the gap between the traded and default over time. In a strong upward trending market, the line for the traded model (blue) could even pass behind the line for the default (tan).   

 The table below shows performance statistics for this period.  The traded portfolio returned 18.8% (annualized) over the four-year period.  The default returned slightly less, at 18.4%.  

The traded portfolio had less volatility (13.2% vs. 17.2% for the default).  As a result, the traded portfolio had a ratio of return-to-variability (RoR/Var) of 1.42 compared to 1.07 for the default. The value 1.42 is good for the RoR/Var ratio (see section “Funds in Bloomberg with High Return-to-Variability Ratios” for more information about interpreting this statistic).

The graph below shows the same information for sg129.  As it was designed to do, the traded portfolio maintained its performance relative to the default in the later stages of the recent ascending market.  

The table below shows that the traded portfolio had higher returns (12.0% vs. 10.2%, annualized) and lower variability (9.0% vs. 11.5%, annualized) relative to its default.  Correspondingly, the RoR/Var compares favorably at 1.32 vs. 0.89 for the default.  Also, the traded portfolio had a smaller maximum drawdown over this period (down 7% vs. down 13%). 

Traded Portfolios Compared to Select Retirement-Focused Funds

This section shows the return of these two model portfolios compared to two alternatives, which I selected as comparisons years ago based on what one might obtain in a company-sponsored retirement plan.  The two retirement-oriented plans are:

• VBINX - Vanguard’s ETF with a 60% stock and 40% bond mix (Vanguard 60/40)
• Russell LifePoints Growth Fund, designated as “RLP – Growth” below. This fund is the most aggressive of the LifePoints funds

The chart below shows Diamond compared to these two funds: 

These funds are more conservative than Diamond; the returns are not as high.  But it is important to note that the RoR/Var ratio for Diamond is higher than for any of the others.  This means that Diamond provides higher returns for the level of variability that must be endured.  Also note that the maximum drawdowns over this period for Diamond and the Russell LifePoints were about the same (down 16%).

The chart below compares these two alternatives (VBINX and RLP-Growth) to sg129.

The variability of sg129 (9.0%, annualized) is less than that of the Russell LifePoints’ (9.2%), but sg129’s return is quite a bit higher (12.0% vs. 4.4%, annualized).  The RoR/Var ratio for sg129 (1.32) is higher than all others. 

Funds in Bloomberg With Returns Similar to Diamond’s

For an additional comparison of Diamond’s return, I searched Bloomberg’s database to find all US-focused mutual funds (including ETFs) that had similar annualized rates of returns (roughly 18-20%) over the same four-year period.

The table below shows that, on average, these 11 funds had a return of 18.9%, which coincidentally matches Diamond’s return for the period. The average variability of the group is 26.8%, which is
more than double Diamond’s 13.2%.

Some of Diamond’s low variability is because of its diversified nature – it holds an allocation to bonds. I had hoped this search of Bloomberg data would have surfaced aggressive balanced funds (holding both stocks and bonds) that may be more directly comparable to Diamond. None were found, and I confirmed this with Bloomberg. It appears that those funds don’t exist in Bloomberg.

Diamond’s RoR/Var ratio is high compared to the average for these investments (1.4 vs. 0.72). 

The ETF DDM, which is the main ETF in Diamond just misses being on this list.  It returned 22.8% (annualized) over this period.  Its variability was 24.9%, which gives it a ratio of 0.92. 

Funds in Bloomberg With High Return-to-Variability Ratios

In order to evaluate the four-year RoR/Var ratios of the two Focused 15 Investing portfolios, I estimated a corresponding statistic based on data available in the Bloomberg databases. I filtered for funds with a North American focus, in US dollars, that are domiciled in the US. In Bloomberg, “funds” include mutual funds, closed-end funds, ETFs, and a range of other fund types. The result was 8,507 funds.

Bloomberg statistics are slightly different than the ones I use[4]. Bloomberg provides three-year and five-year statistics, but not four-year. I estimated the four-year statistics as the average of the three-year and five-year statistics.

The table below shows sections A and B. A shows the percentage of the 8,507 funds estimated to be in the range indicated for RoR/Var, “1.1 to 1.2” for example. B shows the percentage exceeding a certain level, “Higher than 1.2”, for example.

Based on these reference points, Diamond’s RoR/Var ratio of 1.42 and sg129’s of 1.32 place them among the upper 3% and 6%, respectively, of the pool of 8,000+ funds tracked by Bloomberg.

Morningstar ETF Managed Portfolios – Three-Year Return through 12/29/2017

I also compared Diamond and sg129 to funds tracked by Morningstar that are in the class they call “ETF Managed Portfolios.” Like the Focused 15 Investing portfolios, these funds consist of multiple ETFs. Morningstar indicates that they ignore whether the ETFs are leveraged.

Morningstar produces a report showing performance, but the most recent report available is for fourth quarter 2017. The overlapping period with the Focused 15 Investing portfolios is the three years ending that date (12/29/2017).

In the report, Morningstar tracks 1,180 ETF-based funds across a range of different attributes and classifications, shown below.

Of those with a three-year history (the report does not say how many have that length of history), the top 10 performers across all classifications listed above have annualized returns ranging from 11.34% to 17.92%.

Over the same three-year period, Diamond returned 21.0%, and sg129 returned 12.8%. If I reduce these returns by, say, 1.5% for fees, we get Diamond at 19.5%, and sg129 at 11.3% – figures that still place them among the top ten performing funds shown below.

The table below shows the top ten funds in the Morningstar report. (The bottom-performing ones referred to in the title of Exhibit 6 are not relevant and therefore are not shown.)

Conclusion

While the US stock market, as measured by the Dow Jones Industrial Average, has seen strong returns and low variability over the last four years, Focused 15 Investing’s model portfolios have demonstrated better performance characteristics than their default mixes. 

The performance of both Diamond and sg129 over this period has been consistent with their designs. Diamond is designed to avoid losses and have high absolute returns rather than to outperform its default in an upward trending market. While Diamond had strong absolute returns for the period, it had a rate of return similar to its default mix, but with less variability. On the other hand, sg129 is designed to maintain its outperformance during upward trending markets.  This was reflected in the portfolio’s performance; while sg129 had lower absolute returns, it showed better performance relative to its default mix.

The two Focused 15 Investing portfolios have also performed better than select retirement-oriented funds. This is indicated in the comparison to Vanguard’s VBINX and Russell’s LifePoints Growth funds. 

While obtaining directly comparable performance information for a large number of funds is difficult, the data available from Bloomberg and Morningstar suggest that the Focused 15 Investing model portfolios are among the strong performers over the last four years.  

-----------------------------------------------------------------------------------

Appendix: Appendix: Model Portfolio Performance by Focused 15 Investing Publication

The following section shows performance statistics for the current roster of model portfolios by publication.  The publications are:

1. The main US Dollar (USD) publication. 
2. Institutional Investors. These model portfolios tend to have a higher number of ETFs and emphasize performing well relative to their default mixes. Please note: Model portfolio sg99 was in the original publication along with sg129. Portfolio sg99 is designed for low variability and to have fewer ETFs than sg129. 
3. Investors in Australia. These portfolios focus on improving the RoR/Var ratio and keeping trading frequency low.
4. Investors in Japan.

1. Focused 15 Investing - USD

Model Portfolio Performance Since 7/18/2014 as of 7/20/2018

Diamond - 3 ETFs (sg131)

           Ann RoR    Ann Var    RoR/Var

Traded       18.7      13.2       1.4

Default      18.4      17.2       1.1

Sapphire - 1 ETF (sg147)

           Ann RoR    Ann Var    RoR/Var

Traded       23.1      17.1       1.3

Default      23.0      22.6       1.0

Emerald - 2 ETFs (sg148)

           Ann RoR    Ann Var    RoR/Var

Traded       21.5      15.8       1.4

Default      20.4      20.0       1.0

Ultra Diamond - 3 ETFs (sg174)

           Ann RoR    Ann Var    RoR/Var

Traded       26.4      18.7       1.4

Default      25.4      25.1       1.0

Diamond-Onyx - 5 ETFs (sg176)

           Ann RoR    Ann Var    RoR/Var

Traded       12.7       8.2       1.5

Default      12.4      10.9       1.1

Mosaic - 8 ETFs (sg213)

           Ann RoR    Ann Var    RoR/Var

Traded       11.0       9.7       1.1

Default       6.0      13.6       0.4

2. Focused 15 Investing - Inst'l Pro USD

Model Portfolio Performance Since 7/18/2014 as of 7/20/2018

US Ind, LC/SC, MS Bonds - 8 ETFs (sg129)

           Ann RoR    Ann Var    RoR/Var

Traded       11.9       9.0       1.3

Default      10.2      11.5       0.9

US Ind, US Ind/Trans, MS Bonds - 7 ETFs (sg99)

           Ann RoR    Ann Var    RoR/Var

Traded        6.4       6.2       1.0

Default       8.2       9.9       0.8

DJIA x2, US Ind/Trans - 5 ETFs (sg150)

           Ann RoR    Ann Var    RoR/Var

Traded        9.7       7.4       1.3

Default      11.3      11.3       1.0

DJIA x2, EME/B, Com, L Vol Sctrs - 8 ETFs (sg113)

           Ann RoR    Ann Var    RoR/Var

Traded       13.0       9.0       1.4

Default      10.5      12.9       0.8

DJIA, EM E/B, LC/SC, MS Bonds - 10 ETFs (sg122)

           Ann RoR    Ann Var    RoR/Var

Traded       10.4       7.2       1.4

Default       8.0       8.9       0.9

DJIA x2, LC vs SC, Onyx - 7 ETFs (sg204)

           Ann RoR    Ann Var    RoR/Var

Traded       16.3      10.7       1.5

Default      14.6      13.9       1.1

Simulated performance since the inception of the newsletter July 18, 2014

3. Focused 15 Investing - Professional AUD

Model Portfolio Performance Since 7/18/2014 as of 7/20/2018

IEM, IVV - 4 ETFs (sg98)

           Ann RoR    Ann Var    RoR/Var

Traded       10.5       7.2       1.4

Default      10.4      10.2       1.0

IEM, IVV, VLC/VSO - 6 ETFs (sg172)

           Ann RoR    Ann Var    RoR/Var

Traded       10.7       7.7       1.4

Default      11.3      10.5       1.1

MXAU, IVV, VLC/VSO - 5 ETFs (sg71)

           Ann RoR    Ann Var    RoR/Var

Traded        3.7       6.1       0.6

Default       5.5       9.3       0.6

MXAU, VLC/VSO, EM E/B - 8 ETFs (sg183)

           Ann RoR    Ann Var    RoR/Var

Traded        6.5       6.9       1.0

Default       7.0       9.5       0.7

VLC/VSO, EM E/B - 5 ETFs (sg187)

           Ann RoR    Ann Var    RoR/Var

Traded       10.7       8.4       1.3

Default       9.2       9.0       1.0

Currencies - 3 ETFs (sg184)

           Ann RoR    Ann Var    RoR/Var

Traded        1.1       4.2       0.3

Default      -3.8       9.3      -0.4

Simulated performance since the inception of the newsletter July 18, 2014



4. Focused 15 Investing - Professional JPY

Model Portfolio Performance Since 7/18/2014

Diamond JPY - 4 ETFs (sg185)
           Ann RoR    Ann Var    RoR/Var
Traded       22.1      17.3       1.3
Default      18.1      23.0       0.8

Emerald Impact - 3 ETFs (sg186)
           Ann RoR    Ann Var    RoR/Var
Traded       23.3      20.7       1.1
Default      20.4      26.7       0.8



Simulated performance since the inception of the newsletter July 18, 2014

 

---

[1] All performance figures for Focused 15 Investing model portfolios are based on returns of the indexes that the ETFs track and the weekly target weights for the ETFs.  The performance figures do not reflect costs and fees, such as fees collected by the ETF providers, trading costs for implementing the model portfolio in accounts, and fees for the Focused 15 Investing publication. 

[2] Annualized standard deviation of weekly returns. 

[3] Maximum Drawdown indicates the greatest loss over the time period indicated.  It is determined by finding the greatest gap between a high price level and the next subsequent low-price level.

[4] I use weekly return data to create a simple ratio of annualized return to annualized standard deviation of returns.  Bloomberg reports annualized monthly statistics and reports only 3y and 5y Sharpe Ratios.  I do not subtract the risk-free rate in my calculations, a difference that is less meaningful during the current period of low short-term rates.   

Be Patient with the Variability, Even with a High Macro MRI

The market is currently choppy.  I believe we are still seeing the effects of excessively high returns in the six weeks leading up to January 26, 2018.  Historically, it is very difficult to avoid or benefit from a choppy market using a weekly trading strategy.  Thus, we ride through this type of period as long as longer-term resilience (Macro resilience) is present. 

For the DJIA, the Micro MRI has been on the negative (vulnerable) leg of its cycle for several weeks and is, I believe, related to the choppiness.  As of last week (March 2), it was at a historically low level (~4^th percentile).  It should soon turn and provide higher short-term resilience to stock prices.  Absent unusually negative news, stock prices are likely to move higher over the next 6 to 13 weeks.

The Macro MRI, indicating the longer term trend, has been positive and providing resilience since early 2016.  It is at a high level and beginning to weaken.  It could cease to be positive over the coming weeks or months.  Thus, we have MRI that will soon present opposing resilience forces.  Short-term (Micro) resilience is likely to increase.  Long-term (Macro) resilience is likely to weaken/decrease. 

Based on historical precedents, prices can move higher as a result of a positive Micro - even with a deteriorating Macro. 

I created two analyses to describe what the next several weeks could hold.  First, I compared last week’s various MRI statistics and other return information to the same statistics for the 5000+ weeks of history for the DJIA.  The weeks that are most similar to last week’s conflicting statistics tended to have positive returns over the subsequent 6 and 13 weeks.  Eighty percent had positive returns over the subsequent 6 weeks, and the average return was 3.2%.  Eighty-two percent had positive returns over the subsequent 13 weeks, and the average return was 5.3%.  This should provide some comfort as we ride through the choppiness.  

In this analysis, the similar weeks with negative subsequent returns, many had very high prior returns.  To investigate this pattern more fully, I focused the second analysis on the weeks with high Macro levels and their prior two-year returns.  I compared those weeks with the recent week.    

Over the two years prior to last week, the DJIA returned about 44%.  While this seems like a high rate of return, most of the weeks with high Macro MRIs experienced much higher returns over the prior two years (104 weeks). 

The chart below shows the 147 weeks out of the 5000+ for the DJIA with similarly high Macro MRI levels.  The bars represent the returns over the 13 weeks following the high Macro MRI reading. 

These columns are sorted (lowest to highest) by the return over the 104 weeks prior to the week in question. The leftmost week (9/12/1986) had a return of 42% over the prior 104 weeks.  Over the following 13 weeks, the DJIA experienced an 8.7% return.  Last week, which capped a 104 week return of 44%, would have placed second from the left. 

To the right are the weeks with progressively higher 104-week prior returns.  The rightmost bar has a prior 104-week return of 93% (9/6/1929).  This week was followed by a return (loss) of -31%.      

One can see that the weeks on the left – with relatively low prior 104-week returns – were followed by positive returns.  Last week would be in this area of the chart. 

Approaching the midpoint of the chart, we begin to see the weeks followed by losses. 

The story behind this is that over the last 100 years, the DJIA has appreciated more by this stage in a Macro cycle.  Periods that had the highest increases over the prior two years had the biggest decreases over the subsequent 13 weeks.  Periods with moderate increases had few decreases over the subsequent period.  

These analyses support that idea that, while the Macro MRI is influential and at a high level, a high level can be consistent with positive returns over the following weeks.  This is what the algorithms have concluded as well.   

Please feel free to contact me to discuss any of this and how the shift to more volatile returns may influence the model portfolio you are tracking. 

Two Outlooks for 2018

Stock market gains for 2017 were strong, and it is reasonable to wonder: Where will the market go from here? Will it continue to go higher? Will it stumble and fall in 2018?

Two well-known and respected investors have come out with very different forecasts for 2018. Jeffrey Gundlach says (link) stocks are in an accelerating phase now but will ultimately post negative returns for 2018.

The S&P 500 “may go up 15 percent in the first part of the year, but I believe, when it falls, it will wipe out the entire gain of the first part of the year with a negative sign in front of it” – Jeffrey Gundlach

Or, are we poised for a jump in stock prices as global growth accelerates, as Bill Miller suggests (link)?  

I think we could have the kind of melt-up we had in 2013, where we had the market go up 30 percent," -- Bill Miller

The tone of these statements is different, but they could both be true. The market could go up 15 to 30% in the first part of the year and then decline.

The CPM investment approach does not make or use such forecasts. Instead, we evaluate the market’s current ability to recover quickly from bad news and events. This assessment is based on our Market Resilience Indexes (MRIs) and typically remains relevant for a three-to-six week horizon.

Let’s review what the Market Resilience Indexes are currently telling us, and identify past periods that are similar to the current environment. This will give us a sense of where we might be in a broader cycle of stock market returns.

As of January 5, 2018, the Macro MRI, which tells us about the longer-term price cycles and trends, is at the 97^th percentile of all weeks since 1918 — clearly a high level. Yet, the direction of the index’s change is still positive, and decidedly so. Based on past cycles, the Macro trend could remain positive for a few months.  

The Micro MRI, which tells us about the bursts of resilience lasting 6 to 13 weeks, indicates that we have already experienced a period of short-term vulnerability and that the market is likely to become more resilient in the coming weeks.

Considering all the MRI together for the Dow Jones Industrial Average, our assessment is that the market can recover quickly from any bad news and events. Thus, our portfolios continue to be fully invested. 

Greater resilience at both the Macro and Micro levels is consistent with Miller’s “melt-up” statement and also with Gundlach’s view that prices will move higher in the early part of the year.

Historical Precedents

To see how the current market environment compares with past periods in terms of its particular mix of Macro, Exceptional Micro, and Micro MRI, we compared twelve different statistics related to MRI levels, direction of movement, and pace of change. Over 5000 weeks covering almost 100 years of weekly statistics were evaluated for the DJIA.

We determined the fifteen weeks most similar to the status as of January 5, 2018.  The fifteen weeks clustered into five different periods:

1996 – 6 similar weeks

1959 – 2

1955 – 4

1937 – 2

1926 – 1

The graph below shows these dates on a price chart of the DJIA. “Price” is on a log scale to appropriately show the magnitude of early price changes in percentage terms.

Based on the number of weeks identified as similar, the 1996 cluster of six weeks is most similar to the current environment. It is near the beginning of a strong move upward in prices that lasted for four years until 2000.

The next largest cluster occurred in 1955, with four similar weeks. This too was followed by a strong market that moved upward, with some volatility, for about 10 years until 1966.

The 1959 cluster of two similar weeks was followed by generally higher prices. It is within the upward trend following the 1955 cluster. 

Two similar weeks occurred in 1937. This cluster was followed by a sharp price decline and a negative trend in prices that lasted until 1942, which is discussed below.

The week in 1926 was also followed by a strong market for three years until the 1929 crash.

Across all the historical weeks most similar to last week, about 75% of the subsequent 13-week periods had positive returns. This pattern suggests a strong market in early 2018. However, as mentioned for the 1937 cluster, some of the similar weeks in the past were followed by declines. Two periods are noteworthy.

In the 1937 case, prices dropped by 11% over the following 13 weeks. This period became known as the “1937 Recession.” It occurred at the end of the Great Depression when an initial industrial recovery faltered. Unemployment jumped from 14.3% in 1937 to 19.0% in 1938.[1] However, our current unemployment statistics are quite different — they are low.  Furthermore, growth is accelerating globally. Thus, the 1937 period may not be similar to our own in an economic sense.

Another cluster that is somewhat similar to the current week is just prior to the decline of 1987. The week of September 4, 1987 is not shown above because it is outside of the top 15 shown, but it is in the top 30 of similar weeks. September 4^th was followed by a 30% decline over the next 13 weeks. While this is a chilling prospect and deserves consideration, a case can be made that the broader conditions are different. First, the market had appreciated 97% in the two years prior to September 4, 1987, whereas the market appreciated 56% in the two years prior to last week. While both are large numbers, the appreciation prior to the 1987 date is greater.

Second, in 1987, the market had been moving contrary to its resilience levels for about a year prior to the crash.  This is often seen in the late stages of bull markets, which are overly euphoric. This stage can last for several quarters.  Today, the market has been performing in a manner generally consistent with its MRI levels, with the exception of the last month or so (noted below).   

Thus, today’s conditions are different.  The positive aspect of the 1987 decline is that it was followed by 10 years of strong market returns.

While acknowledging the cautionary precedents of 1937 and 1987, the cases of 1996 and 1955 do have the highest number of weeks similar to the current environment.  These cases support a more optimistic view or where we are in a broader cycle. 

Recent stock market behavior relative to our assessments of resilience gives additional perspective.  During the recent two months of greater vulnerability, as measured by our MRI, the US stock market did not decline — it went up.  

This recent market behavior could be telling us that we may be moving into a more euphoric stage, where vulnerability does not result in price declines.  If so, this heightened euphoria is just beginning.

However, we may find that the last two months of defiance could more reasonably be attributed to the passage of the tax bill, which may boost corporate profits.  Either way, the current market environment does not yet appear to be excessively euphoric the way it was prior to September of 1987 and in other major declines of the last 100 years. 

This evaluation leans more toward the view expressed by Bill Miller — stock market prices will move higher.

That said, it is important to remember that our approach does not rely on forecasts.  The analysis driving our research and model portfolios is based on readings of the market done every week. We will evaluate the market another 51 times in 2018 and will change our model portfolios as needed. We can see the dynamics shift and the algorithms will shift the portfolios in response to the changing dynamics so that if the stock turns down as predicted by Gundlach, we will respond.

---

[1]                      Economic Fluctuations, Maurice W. Lee, Chairman of Economics Dept., Washington State College, published by R. D. Irwin Inc, Homewood, Illinois, 1955, page 236.

Increasing Vulnerability Before End of Year

The DJIA has MRI levels near the tops of their historical ranges. The Macro (long term) MRI continues to be positive but is at the 95th percentile of it levels since 1918. Its Micro (short term) MRI is at the 89th percentile. This market continues to be resilient but is nearing a natural transition to weaker resilience (greater vulnerability). A weakening dollar may contribute to higher Micro (short-term) resilience for this index. The more domestically-biased S&P and DJ industrial sectors lost their Micro MRI a week ago (November 3, 2017). As indicated in prior notes, with the increase in market vulnerability the character of the US large cap market is shifting toward having greater volatility.

The US 10y Treasury bond index is a mirror image of the DJIA. The Macro MRI continues to be negative but is at the 10th percentile level since 1983. Its Micro MRI is at the 11th percentile. This index is likely to shift to having higher resilience over the coming weeks.

The SPGS Commodity Index has a low Macro MRI (14th percentile since 1971) that is moving higher and a high Micro MRI (96th percentile). Importantly, the Exceptional Macro is present for this index. Overall, this index remains resilient.

For the dollar index, DXY, the long-term trend is still negative, as reflected in the absence of the Macro MRI. Its Micro is present at the 80th percentile and is moving higher. After it reaches its peak, DXY is likely to decline in the absence of any positive MRI.

The relative resilience of global inflation-linked bonds and world government bond index (WGBI), is currently favoring WGBI.

More broadly, the MSCI World Stock Index also has a high Micro MRI (89th percentile) and more moderate Macro MRI (80th percentile) compared to the US stock market. The US stock market is more advanced in the cycles than are the other major stock markets, which are likely to continue to exhibit resilience.

Japanese stock index, Topix, is at an inflection point in the Micro MRI. The Micro MRI is at the 99th percentile since 1972, and is likely to cease in the next few weeks. The Macro MRI for Topix is still moving higher and is at a moderate level (70th percentile since 1972), leaving room to move higher. Topix is likely to pause in its upward movement, but is, at this point likely to resume its upward movement after the pause.

The Russell 2000 lost its Micro MRI this week. It continues to have its Macro MRI, which is at the 82nd percentile since 1985. This index could experience some near-term declines.

All We Need is a Catalyst

Many of the stock indexes have Micro MRIs that are high and have recently shifted or will soon shift to their vulnerable phases. This alone can result in greater price volatility while the long term trend remains positive.  For some markets (e.g., the US stock market DJIA), the Macro MRIs are also at high levels.  Thus, at least on a short term basis, the markets are in a somewhat precarious state for shocks. Debt ceiling, tax reform jitters, North Korea or other issues could be catalysts for a more major shift in the markets.  Any of these could be seen as causing a shift to a more negative long term trend in stock prices.

Our portfolios are fully invested right now, despite this situation. Historical testing indicates that it is better to accept some early volatility and declines rather than to move assets in anticipation of greater volatility.

September 2017

Market Overview

As a quick reminder, I describe the Market Resilience Indexes in a footnote[1].

A. Stock markets…

Most major stock markets around the world have strong and positive Macro (long-term) MRIs. This indicates that the longer-term price trend is still positive. The DJIA has proceeded further along this positive trend than other major markets and is therefore closer to the inevitable end of the positive trend. However, for the time being, the trend for the DJIA continues to be positive. Most major stock markets around the world can recover quickly from declines related to the near-term events. If the positive Macro MRIs were absent, the potential for a quick recovery would be greatly diminished. 

Their Micro MRIs have been in the downward leg of their cycles for several weeks and are not providing any resilience to negative news.  This state indicates short-term vulnerability, and we have been seeing higher price volatility over the last several weeks as a result. Importantly, the Micro MRIs already descended to low levels by the end of August. As of last Friday (September 1^st), the Micro for the DJIA is at the 35^th percentile since 1918, which is a moderately low level – 65% of the weeks since 1918 have higher levels. And it is trending lower. Other markets have even lower levels, as shown below. The lower the percentile rank, the more the market has already descended to short-term lows.

a.      DJIA – 35^th percentile

b.     DJ and S&P Industrial Sectors – 11^th (the “industrial sectors” are more domestically focused than the DJIA listed above)

c.      Russell 2000 – 22^nd

d.     DJ Transports – 11^th

e.     NASDAQ – 24^th

f.       UK Stocks – 14^th

g.      Europe – 14^th

h.     Japan – 14^th

As you can see, the major stock markets were at relatively low levels as of last Friday. They are likely to shift to a positive reading and provide short-term resilience over the next few weeks. If the curse of September, the storms, and the North Korea tensions occurred when these values were higher, say at the 80^th percentile or higher, the downside would be greater and the condition described earlier would have been met.

The net effect of the current condition is that while the markets may decline with the news of the day, they are likely to recover quickly. This leaves little opportunity to exit and re-enter the markets.

B. US 10y Treasury Bond Futures (TY1)

The Macro MRI is trending lower and is at a low level (7^th percentile since 1986) and the Exceptional Macro appeared recently (8/18/2017).  Both suggest that the Macro may shift to a positive state in the coming weeks or months.  These longer-term indicators highlight the potential for a strong move higher – one that is beyond what might result from near-term fears linked to the news of the day.

The Micro MRI, which measures shorter-term resilience, has been trending positive for several weeks. It is at the 43^rd percentile since 1986 and can continue higher for several weeks. The net effect is more resilient bond prices over both shorter and longer horizons. 

C. Commodities

The SPGSCI continues to be rated 3 (highest resilience), but this rating has been borderline for several weeks. Its Macro has been trending positive but is very shallow; it could easily move negative and cease to provide resilience. The Micro MRI is at a high level (80^th percentile since 1973) and may cease over the next few weeks. Thus, the 3 rating continues to be borderline and at risk.  

D. Dollar (DXY)

DXY is experiencing only Micro (short-term) resilience. Its Macro and Exceptional Macro are not close to turning positive. The Micro MRI is at the 15^th percentile since 1970 and could continue to be positive for several weeks. However, DXY’s Micro has been erratic over the recent period (beginning in roughly April); it has shifted between being positive and negative over this time period. Furthermore, its Macro is at a relatively high level, which suggests that there is more downside to DXY than upside.

* An early version of this post indicated 67 years in which the conditions had been met.  Of those, 43 were for the September declines and 24 were for March declines.    

---

[1] I use proprietary metrics to determine where current price levels are relative to their historical norms. The Market Resilience Indexes indicate where prices are relative to their long-term (Macro) and short-term (Micro) historical cycles and whether they are moving toward their historical peaks or troughs.  Cycles for the Macro MRI last several quarters to multiple years.  Cycles for the Micro MRI last four to six months.  The presence of the Exceptional Macro MRI indicates that the Macro MRI is likely to move from a negative trend to a more positive one.

Managing The September Curse

You may see a few articles such as this one about the poor stock market returns that tend to occur in September.  As the article points out, September can be a trying month that ends with negative returns. Add in tropical storms and North Korea tensions and we can see why professional investors become nervous.  While I don’t try to forecast the weather or geopolitical events, I do assess the market’s ability to recover quickly from whatever events do take place by virtue of the market’s inherent resilience.

The investment process used to create the Focused 15 Investing model portfolios systematically evaluates whether it makes sense to reduce exposure to the stock market for the month of September. At the end of August, it will reduce exposure to stocks if the Micro MRI is higher than a specific threshold. If this condition is met, I expect price movements to be negative in September and for prices to remain low for some time. The process then takes some money out of the market. The rationale is that if the Micro is at a high level and is likely to turn negative in September and cause assets to be taken out anyway, do not wait for the turn negative event ― just reduce assets at the end of August.  This decision rule would not be in the process if it didn’t add value over the last 100 years.  Indeed, since 1918, these specific conditions have been met in 43* of the 100 years. This “conditional price movement” technique for September adds about 1.67% on an annualized basis over 100 years, which is a respectable amount.  This type of conditional technique working with the MRI information is important to the investment approach. So important, in fact, that I named my firm CPM Investing, for conditional price movement.

During the evaluation this recent August, the conditions were not met, and there was no change in the target weights. The Micro MRI for stocks was already quite low, which means that the market is closer to making a short-term move higher. After any decline in September, the market will most likely follow the trend indicated by the Macro MRI, which is positive at this time. While September may be bumpy, no change in target weights is justified.  

Current Market Environment

As a quick reminder, I describe the Market Resilience Indexes in a footnote[1].

A. Stock markets…

Most major stock markets around the world have strong and positive Macro (long-term) MRIs. This indicates that the longer-term price trend is still positive. The DJIA has proceeded further along this positive trend than other major markets and is therefore closer to the inevitable end of the positive trend. However, for the time being, the trend for the DJIA continues to be positive. Most major stock markets around the world can recover quickly from declines related to the near-term events. If the positive Macro MRIs were absent, the potential for a quick recovery would be greatly diminished. 

Their Micro MRIs have been in the downward leg of their cycles for several weeks and are not providing any resilience to negative news.  This state indicates short-term vulnerability, and we have been seeing higher price volatility over the last several weeks as a result. Importantly, the Micro MRIs already descended to low levels by the end of August. As of last Friday (September 1^st), the Micro for the DJIA is at the 35^th percentile since 1918, which is a moderately low level – 65% of the weeks since 1918 have higher levels. And it is trending lower. Other markets have even lower levels, as shown below. The lower the percentile rank, the more the market has already descended to short-term lows.

a.      DJIA – 35^th percentile

b.     DJ and S&P Industrial Sectors – 11^th (the “industrial sectors” are more domestically focused than the DJIA listed above)

c.      Russell 2000 – 22^nd

d.     DJ Transports – 11^th

e.     NASDAQ – 24^th

f.       UK Stocks – 14^th

g.      Europe – 14^th

h.     Japan – 14^th

As you can see, the major stock markets were at relatively low levels as of last Friday. They are likely to shift to a positive reading and provide short-term resilience over the next few weeks. If the curse of September, the storms, and the North Korea tensions occurred when these values were higher, say at the 80^th percentile or higher, the downside would be greater and the condition described earlier would have been met.

The net effect of the current condition is that while the markets may decline with the news of the day, they are likely to recover quickly. This leaves little opportunity to exit and re-enter the markets.

B. US 10y Treasury Bond Futures (TY1)

The Macro MRI is trending lower and is at a low level (7^th percentile since 1986) and the Exceptional Macro appeared recently (8/18/2017).  Both suggest that the Macro may shift to a positive state in the coming weeks or months.  These longer-term indicators highlight the potential for a strong move higher – one that is beyond what might result from near-term fears linked to the news of the day.

The Micro MRI, which measures shorter-term resilience, has been trending positive for several weeks. It is at the 43^rd percentile since 1986 and can continue higher for several weeks. The net effect is more resilient bond prices over both shorter and longer horizons. 

C. Commodities

The SPGSCI continues to be rated 3 (highest resilience), but this rating has been borderline for several weeks. Its Macro has been trending positive but is very shallow; it could easily move negative and cease to provide resilience. The Micro MRI is at a high level (80^th percentile since 1973) and may cease over the next few weeks. Thus, the 3 rating continues to be borderline and at risk.  

D. Dollar (DXY)

DXY is experiencing only Micro (short-term) resilience. Its Macro and Exceptional Macro are not close to turning positive. The Micro MRI is at the 15^th percentile since 1970 and could continue to be positive for several weeks. However, DXY’s Micro has been erratic over the recent period (beginning in roughly April); it has shifted between being positive and negative over this time period. Furthermore, its Macro is at a relatively high level, which suggests that there is more downside to DXY than upside.

* An early version of this post indicated 67 years in which the conditions had been met.  Of those, 43 were for the September declines and 24 were for March declines.    

---

[1] I use proprietary metrics to determine where current price levels are relative to their historical norms. The Market Resilience Indexes indicate where prices are relative to their long-term (Macro) and short-term (Micro) historical cycles and whether they are moving toward their historical peaks or troughs.  Cycles for the Macro MRI last several quarters to multiple years.  Cycles for the Micro MRI last four to six months.  The presence of the Exceptional Macro MRI indicates that the Macro MRI is likely to move from a negative trend to a more positive one.

Are Stocks at a Peak?

A slightly pessimistic tone has developed in the markets. The geopolitical risks involving potential clashes with North Korea and Venezuela seem to have influenced markets. We also have words of caution from well-known investors about near-term market declines. Jeff Gundlach, a well-respected bond manager, has said that those holding risky bonds should start heading for the exits; he mentioned high-yield bonds and emerging market bonds. Others have echoed his comments.

This is a good time for us to assess the stock market’s ability to rebound from geopolitical shocks, its maturity, and how this bull market might proceed from here.

In a letter to clients a few weeks ago, I recommended reducing the equity exposure of their portfolios to 80% from 100% by July 28. Our research indicated that the stock market was losing resilience, and a shift in allocation was appropriate at that time.

While our portfolios became more defensive, many of our indicators suggest that the market will be able to rebound from any near-term declines that may be caused by moderate shocks. The long-term trend of the stock market is still positive. Near-term vulnerability will be replaced by higher resilience in a few weeks.

Of course, the magnitude of a geopolitical event is important ― a sizable nuclear conflict may overwhelm any inherent market resilience. Nonetheless, the market is still relatively resilient and should be able to recover quickly from market declines that fall short of that unthinkable event.

Regarding the cautious tone from well-known investors, we do not see the typical signs of the stock market reaching a peak and then experiencing protracted declines. Again, the longer-term positive trend in the DJIA is still intact.  

What History Tells Us About Market Declines

Our research on the last 100 years of stock market price movements indicates that protracted market declines are typically foreshadowed by a weakening of our proprietary Macro Resilience Index. Weakening resilience suggests greater vulnerability, and then some catalyst occurs to initiate the protracted decline. The progression from resilience to vulnerability to decline typically takes place over a period of a few quarters, giving advanced notice to reduce exposure to the stock market. This is generally true for the prominent declines of 1929, the 1970s, 2000, and 2008.  We are not yet in that type of environment.

However, the sharp decline of 1987 stands out as unusual. This decline was not foreshadowed by a smooth reduction in our long-term resilience measure. If we are indeed coming to a major decline while our long-term trend measure is still positive, it would be generally similar to 1987. Let’s take a closer look at what led up to the sharp 20% decline in August and September of 1987, exactly 30 years ago.

                                                                                                                                               

Our measure of the long-term trend (the Macro MRI) did shift to a vulnerable reading in November of 1986, almost a year prior to the decline, but it switched back to positive in May of 1987. And it made these changes after a long upward trend – almost 3 years after a decline in the Macro MRI and the stock market in late 1983 and early 1984. The Macro MRI had moved to a high level that placed it at the upper extreme of the range it has traversed over market cycles since 1918.

Today, the Macro MRI is positive, steady, and well short of its historic extremes.  Furthermore, it is moving higher in a recovery from the phantom bear market we experienced in 2015/6 ― just over a year ago.  In short, measures suggest this market is simply less mature and less extreme than the situation in 1987.

Let’s take a look at how prices actually moved in 1987 compared to today. The chart below shows the DJIA in 1987 with recent price movement as an overlay. These series start August 9^th of 1984 and 2014. 

The first thing that catches one’s eye is that the periods of strong upward movements and declines line up fairly well. Based on the apparent synchronization of peaks and troughs, one might conclude that we are on the precipice of a major decline.

However, as mentioned earlier, the level of maturity of the bull market is quite different. We estimate that our current level of maturity (corresponding to the current price level shown by the green triangle) is where it was in June of 1986 (white white), well over a year before the big decline.  We define maturity as the level of our Macro MRI, which can be thought of an indicator of investor euphoria. 

The chart above focused on price patterns, not actual returns. The chart below shows the same price movements in terms of returns.

The returns since the beginning of the chart for each time period are quite different. From August 9, 1984 to the top of the market in 1987, the DJIA rose 113%. Over the period from August 9, 2014 to August 9, 2017, it has risen 32%. The difference is big.

This information does not suggest there is no risk at this time. Thirty-two percent in today’s environment may somehow be equivalent to 113% in the 1980s environment of rapidly decreasing interest rates.  Historically, the best course of action has been to respond to the dynamics of the market as they present themselves at the time rather than to focus on repeating overall price patterns.  Our view could change in a matter of weeks should the underlying dynamics change. 

Near-term Outlook

I believe we will have a few weeks of soft prices and perhaps moderate declines. When those pass, we can expect a continuation of high resilience for stock prices.

Other Markets

Emerging market bonds, as Jeff Gundlach suggests, appear to be at an inflection point. This asset has been resilient for several quarters but is quickly shifting to being vulnerable. There is still short-term resilience, but that is likely to fade in a few weeks. Historically, this asset has not decline dramatically during the early stages of vulnerable periods.  

The SPGSCI (commodity index) is generally resilient. It has been borderline with a potential shift to vulnerable, but the shift has not taken place. Short-term resilience is developing.