Sunday, June 22, 2014

Continuing the Portfolio Descriptions

.
Update on Shooting Stars

In my previous entry I described the Shooting Stars portfolio, which I've developed to test the theory that stocks coming out of Stage 1 and into Stage 2 are significant.  The key of that portfolio is to select quality stocks that have accelerating sales and earnings on a quarter-over-quarter basis.  I started the portfolio "officially" on May 2 and May 5, and added another variant this past weekend.  Here, "officially" means that I started an official tracking/audit of every transaction, inclusive of commissions.

There are three portfolios that are using the accelerating sales and earnings rules:  one that looks at the present universe of qualified stocks that meet the Shooting Stars criteria and selects those that are projected to grow at a CAGR of at least 15%, another that uses the entire universe of qualified stocks without the 15% CAGR requirement (but still meeting the Shooting Stars requirements), and the third uses high-beta components only of the shooting stars universe, e.g. calculated beta has to be higher than the market beta (which by definition is 1).

As of the close of June 20 we have:

Shooting Stars 15% CAGR:

  • Realized gains/(losses):  -2.1%
  • Unrealized gains/(losses):  +0.5%
  • 22% invested
Shooting Stars All:

  • Realized gains/(losses):  -1.9%
  • Unrealized gains/(losses):  +4.4%
  • 77% invested
Shooting Stars HiBeta:
  • Realized gains/(losses):  N/A
  • Unrealized gains/(losses):  N/A
  • 0% invested (orders will be executed starting Monday)
The GGT recommendation for individual stocks controls the amount of cash in the portfolio, and the 15% CAGR has a high number of stocks that are being held in cash.  Details are in the GGT stock file, which is in the shared Dropbox directory.  Send a note to pduncan [at] v t {dot} e  d  u  with "Dropbox" in the subject line if you want access to the files.

For the three portfolios, allocations are made by optimizing the Sharpe Ratio of ALL possible holdings and then only buying those that are in some form of GGT "long" status.  Note that the act of optimizing potentially removes some of the candidates from the pool, in that additional stocks either worsen gain and/or portfolio volatility.  Hence, the process of optimizing allocation removes stocks that have a really poor reward/risk ratio as measured over the past year.

Selling of individual stocks will occur based on my nightly scans.  Same goes for purchases that are signaled via a GGT "New Long" recommendation.  Reallocation, and additional scans for stocks that fit the universe criteria happen every couple of weeks or so, time permitting.

~~~~~~~~~~~~~~~~~~~~~~~~

Steady Eddies

Steady Eddies is my long-term, low-maintenance portfolio concept.  Unlike most of my portfolios, this one does not subscribe to the accelerating EPS and sales requirement, although it does require growth in both of these components over a much longer time period than consecutive (trailing) quarters.  In fact, it requires incredible stability on a 10 year time frame for EPS and sales, and out of the universe of stocks only 9 (nine) make the cut right now.  Here's the exact selection strategy:

  • Must be on a US exchange
  • A linear fit to the 10 year revenue growth using yearly numbers must have a R-squared of greater than 0.95
  • A linear fit to the 10 year EPS growth using yearly numbers must have a R-squared of greater than 0.95
  • The current price, when compared to the projected "high" price using the EPS projection determined above, must provide at least a 2:1 upside in the next 5 years.
I've been tracking this portfolio for almost a year and it lives up to it's name.  It just plods along.  Without the GGT influence on the individual positions the portfolio swings with the market as a whole -- up a few/down a few.  With GGT timing individual stocks it appears that there is between a 20 to 25% reduction in volatility compared to the always-long approach, which is why I'm interested in this portfolio strategy.  Although some positions are held in cash and this limits the overall gains of the always-long stance, the returns seem to be very stable on a longer term.  We'll see.

Although the stocks are in the Dropbox file I'll list them here for your consideration:

AAPL, BRLI, CERN, CTSH, FFIN, JKHY, LKQ, MWIV, PNRA

Note that CERN, LKQ, MWIV, and PNRA are currently in GGT "cash" status.

Like Shooting Stars, there are multiple portfolios here.  The basic portfolio (Fully Invested) invests in the full  list of 9 (nine) above, timed with GGT.  The second portfolio is the 15% CAGR methodology, and presently, the only stocks on that list are AAPL, BRLI, CTSH, and LKQ.  As I stated above, LKQ is in "cash".  I officially started logging these two portfolios on May 5th.

Here are the performance numbers:

Steady Eddies 15% CAGR:

  • Realized gains/(losses):  -2.2%
  • Unrealized gains/(losses):  +2.8%
  • 67% invested
Steady Eddies Fully Invested:
  • Realized gains/(losses):  -0.8%
  • Unrealized gains/(losses):  +2.1%
  • 75% invested
For both portfolios, allocations are made by optimizing the Sharpe Ratio of ALL possible holdings and then only buying those that are in some form of GGT "long" status.  The allocations are listed on the "Portfolio" tab of the Excel workbook in the Dropbox folder that I described above.

~~~~~~~~~~~~~~~~~~

In a future entry I'll describe the "Greenfield" portfolios, which are presently outperforming everything else I am running.  

Let me know if you have any questions!

Regards,

pgd

Wednesday, June 4, 2014

Various Portfolios in the Dropbox Stock File

.
I received a question today that most likely is on many of your minds, so I thought I'd post a generalized response.

" I see the titles of the various stock lists- Greenfield dividends, Greenfield Leaders, Steady Eddies, Shooting Stars,etc. I suppose I can tell from some of the titles , what types of stocks are within, but where does " Greenfield" come from.  Also, is the criteria you are using for the various lists specified somewhere?"

Background

2014 has seen me focused on "life" things, but one item I devote time to is improving GGT.  In fact, it's in process of being moved to "the cloud", and I'll reveal more of that once I start initially kicking the tires and seeing how it can be used to influence our decision making when it comes to stocks and ETFs.

Portfolios

Part of the effort of moving GGT to the cloud is the creation of portfolios, and these portfolios are what you see on the "Portfolio" tab of the Dropbox stock file that I update daily.  Here's a brief rundown of the different portfolios; aside from my personal portfolios, I'm only going to explain Shooting Stars, 15CAGR and Shooting Stars, Fully Invested in today's entry.

Sharebuilder -GB, Sharebuilder -SB:  these are automatic investments for my two boys and the plan costs $12 per month per child, and you get 12 investment transactions per account per month.  I invest weekly for them (right now $214 per kid per week), and you can see where the allocations are going.  Allocations are done in terms of optimizing the Sharpe Ratio for the three stocks per account, so it's not a 33%/33%/34% in each.  Presently, you can see the weights in the "CAP Weightings" column.  Anything in a "CASH" status means that that portion is going to cash, not to the investment indicated.  The data is listed in this file simply so that I can easily track my personal investments.

TS-PF, TS-790, TS-535:  these are actual holdings in TradeStation accounts for private individuals, including myself.  TradeStation permits 3rd party transactions on an account, e.g., these folks have opened a TradeStation account and have permitted me access to trade their account.  I cannot withdraw monies, but I can trade their account as it was one of my own.  This is a for-fee monthly service that I am providing to individuals.  If you have under $25K in the account margin rules allow only 3 transactions per day, which is a pain in the tail if you're trying to act on a new buy signal.  Nevertheless, you can see actual holdings in these accounts and the overall allocations.  Again, the allocation per stock is done using software that optimized the Sharpe Ratio for the portfolio, so it weights more those stocks with a higher Sharpe Ratio.  Presently, TS-PF and TS-535 are shadow accounts, meaning they should be holding the same stocks as I'm mirroring performance, and the TS-790 account is my personal account that is not using the same method as the TS-PF or TS-535 accounts.  The TS-PF and TS-535 accounts are using a blended Greenfield Leader's / Leader's strategy, which I'll describe at a late time.  The TS-790 account is using a Darvas Box breakout strategy, which I'm still developing.  I will discuss this strategy also at a later time.

Shooting Stars, CAGR 15%, GGT Timed:  this is a strategy that requires a number of conditions at the time of entry:

1) 200d Uptrend Test:  the 200d MA of the stock must be in an uptrend for the past 30 days (minimum).
2) Span Test:  the 75% level of the 52-week high must be above the 125% level of the 52-week low.  As an example, let's assume that the 52-week high is $100 and that the 52-week low is $50.  75% of $100 is $75 and 125% of $50 is $62.50.  In this case the stock meets the condition because $62.50 < $75.    If the 52-week high were $75, 75% of $75 is $56.25 which is less than $62.50 so in this case the stock would fail and would not pass the Span Test.
3) Accelerating EPS Test:  The last 4 quarters of earnings must be higher than the prior 4 quarters of earnings, one reporting cycle ago.  As an example, I'm writing this on 6/4/2014.  There are 4 quarters in a year, so I can take each earnings for each of the quarters 6/2013, 9/2013, 12/2013, 3/2014 and add them together.  I can then take the series of 3/2013, 6/2013, 9/2013, 12/2013 and add them together.  If the first lot (most recent 4 quarters) is greater than the previous lot (4 quarters one quarter ago) then we have accelerating earnings.
4) Accelerating Sales Test:  Exactly the same process as the Accelerating EPS Test above, but for sales figures.

Sales and EPS drive a stock price and company valuation.  200d Uptrend Test ensures that prices are moving upward over the longer haul.  The Span Test ensures that the moves are significant and that we're not getting a stale stock.

This specific portfolio requires that the projected CAGR, using Better Investing's Stock Selection Guide (SSG) methodology, shows that the projected 5-year return will be higher than 15% CAGR per year.  Your crystal ball is as good as mine.

GGT is used to time this portfolio.  As I write 5 stocks are in some form of "CASH" status, e.g., money is held in reserve.  When the stock transitions to a "New Long" I'll submit an order for the allocation shown.

Allocations into each of the positions are shown in the "AO" column on the Portfolio tab.  Again, this is done using the Sharpe portfolio optimization tool.

Allocations may change going forward if there is cash in the account.  I periodically update the CAP weights and as new stocks meet the criteria, they are tested as a whole within a portfolio setting.  It is quite possible that a stock in "CASH" status will disappear and be replaced by another, possibly in cash status, or possibly in some form of "LONG".  I generally will only do this update on the weekends, as I don't have much time during the week to do the analysis (although it is automated it takes time to set up and verify).

The published portfolio has been running since the beginning of May 2014, although I've been playing with the concepts for many years.

The values presently listed in the yellow/blue bar are simply estimates of future performance and are not a guarantee of anything.  Here's the terminology:

TR:  projected annualized total return.  For this portfolio the requirement of TR is > 15%, and the projection is 21.2%
PAR:  projected annualized average return.  This specific portfolio, as fully held, (coincidentally) projects a PAR of 15.0%.
R/R:  risk/reward of the portfolio.  Ideally, want better than 3.0/1.  This portfolio is presently at 6.2/1, which indicates that there is a 6x upside possibility relative to a 1x downside movement.
RV:  relative value.  Uses the present EPS value relative to the past 5 years EPS average to determine stocks/portfolios that are overvalued or undervalued.  Ideally, we want values in the 80-110 range, with a preference around 90-100.  A value of 79.1 is good, but a bit low and if on an individual stock, would be considered suspect.  The fact that it's spread across all the holdings improves confidence in the value so we'll see.
Hoadley:  13.52%/21.73%/0.62.

  • The first number is the projected annual return of the portfolio, fully invested, with a market return of 12% and a risk-free rate of 0.03% (I use the 4-week T-bill as the RFR, http://www.federalreserve.gov/releases/h15/update/).  In this case the portfolio has a projected 1.52% alpha level, e.g., on average it should return 1.52% more than simply investing in a market index.
  • The second number is the projected annual volatility of the portfolio, again fully invested.  In this case a fully-invested portfolio may vary as much as 21.73% over the course of the year in order to return the 13.52% projection shown.
  • The third number is the Sharpe Ratio, and by definition, it is the ratio of 13.52%/21.73%.  Of course, dropping volatility and increasing return will improve the Sharpe Ratio, and while easy to do for individual stocks, it is not easy to do for portfolios due to covariance.  Out of the universe of stocks that meet the criteria for this portfolio, this is the greatest Sharpe Ratio for the stocks listed, and the lowest covariance for the portfolio.
For you data wonks, here is the correlation matrix for all the stocks meeting the criteria within the portfolio (but not necessarily positioned):



Right-click on the image to open in a new tab or window.

A value of 1.00 means that two securities are perfectly aligned with each other, e.g., they move exactly together.  A value of (-1.00) means that they move exactly opposite of each other.  A value of 0.00 means no correlation to the target market.

As you can see above, relative to ^RUT, which is the Russell 2000 (R2K) index, WDR has the highest correlation with the ^RUT at 0.70, meaning that 70% of it's movement correlates to the movement of the Russell 2000, but 30% of the time it's movement is not explained by the movement of the R2K.  Contrasting,  EDGW only has a value of 0.17, so only 17% of it's behavior can be attributed to the R2K market.

Interestingly enough, not all the stocks make it into the preferred portfolio.  Here's the list and weightings as of 5/30/2014 (subject to change going forward):

ANIK 12.22%
BMA 1.86%
BWLD 4.98%
CTSH 13.15%
FFIV 12.15%
FSS 11.23%
HURN 8.11%
LKQ 5.11%
LVS 13.26%
OLED 6.07%
WDR 11.86%


Of significance is that WDR is listed and EDGW is not.  It's all in the math of finding the optimum Sharpe Ratio for a portfolio.  Suffice to say, WDR is more stable than EDGW and EDGW simply causes the portfolio to move away from the optimum point.  You can see this clearly on the Efficient Frontier chart:



The target portfolio is shown where the blue curved line and the straight line intersect (at the purple triangle). Note that the projected return for EDGW is far below the target AND it has extremely high volatility, hence, it's not a great candidate in the portfolio relative to other securities.

The lone gold diamond, just above WDR, is the present holdings portfolio estimate.  This is with the 5 securities that are in "CASH" mode NOT included in the optimum results.  While return is higher, volatility is much higher, so the trade-off is higher risk for a disproportionate amount of return.  It's not an ideal situation and I'm hopeful that the remaining 5 stocks transition to "New Long" soon so I can commit the monies.

Shooting Stars, Fully Invested:  This is a sister-portfolio to the one described above and the only major difference is that ALL stocks meeting the 200d slope test, the span test, and the EPS/sales tests are considered.  I do not screen on the 15% CAGR estimate requirement.

The impact here is significant -- the universe of stocks grows dramatically.  Consequently, there is more ability to fine-tune the Sharpe Ratio (it is presently estimated at 0.67 compared to 0.62 in the 15% CAGR portfolio), but estimated returns drop dramatically because more securities are involved.

In fact, if you look at the Hoadley numbers in the yellow/blue title bar on the dropbox sheet, you'll see that we have  10.91% / 16.34% / 0.67.  This means that instead of having a alpha greater than the 12% market baseline, we're actually working harder for less return than simply investing in the ^RUT using the ETF "IWM" and being done with it.  The only benefit of adding securities is that volatility drops dramatically (estimated to be 16.34% versus 21.73% in the 15 CAGR portfolio), so the Sharpe Ratio moves up from 0.62 to 0.67.  This is significant.

Here's the Efficient Frontier for the Shooting Stars, Fully-Invested portfolio:



You can see that there is an improvement in potential gains with a slight add of volatility as more positions are added using this portfolio.  This is simply due to GGT timing.

All of the stocks in the Shooting Stars Fully Invested portfolio meet the 4 critical criteria as outlined above so they are good stocks, BUT, they also are not necessarily projected to do as well going forward, according to SSG methods.

The purpose of this portfolio is to compare/contrast the 15 CAGR screen and to see how the SSG worksheet process of future projections work.

~~~~~~~~~~~~~~~~~~~~~~

Enough for now.

The next blog entry will consider "Steady Eddies", which I think you will find quite interesting if you have a long-term view of the markets.

Your feedback is solicited and welcomed.

If you are not a dropbox member, please send me an email to pduncan [ at] v t [dot] e d u (no spaces, trying to beat the bots) with the word "Dropbox" in the subject line and I'll add you.  Note I will also force add you to the Yahoo! GGT group, as it's the only way I can communicate with you.

Regards,

pgd