Tuesday, 13 December 2016

Thoughts on the week ended December 9, Financial Advisory, Portfolio Management and Morningstar, Francis Chou, Jean Coutu, and Non NVCC Insurance Preferreds

Before I get into the good stuff, a quick look at the week ending December 9:

New all time highs across the board because Trump is going to save the world.

On the subject of financial advisory and portfolio management

(Caveat, the content dealing with financial advisory was written around November 22nd, and since then, I have made some changes to the weightings in my portfolios)

I had a meeting with a highly recommended (big bank) advisor in late November, as part of me wants to outsource the management of my retirement assets which are now just over $151K in total. In his office there were two 60 inch screens on his wall, one with a portfolio of Canadian stocks updating in real time, and the other with an interface that reminded me of morningstar.com's website.

I prepared a summary information package illustrating the distribution of assets as follows (circa Friday November 18th):

During our discussion, the advisor made the following comments, in no particular order:
  • He never has more than 5% cash in his managed portfolios. 38% cash could lead an investor to make sub-optimal decisions. At my age (44) it's better to be fully invested and use downturns in the market to dollar cost average into existing holdings because my time horizon is so long. I can achieve this by making additional lump sum RRSP contributions during times when the markets move down.
  • With reference to my current common shareholdings, he didn't recognize many of the names I currently own (Acasta Enterprises common + warrants, Newfoundland Capital, Corby Distilleries, Becker's Milk, etc.).  His general comment was that I have to be careful holding smaller, unknown names due to liquidity.
  • He follows a Morningstar CPMS model tailored to screen for and set up model portfolios All of his current clients are invested in the same model. All of the advisors at his branch basically use a subset of the same model. The model consists of 20 Canadian and 20 US stocks, ranked and selected based on Morningstar's proprietary research (more on this later). I can tell you that the 20 Canadian names included nothing unexpected, most of the Canadian big banks, one of the Canadian railways, one of the major pipelines, one drugstore chain, one astronomically priced software consolidator, one non-bank financial services alt-mortgage lender, etc. These were the stocks I observed on one of the 60 inch screens on the wall when I first walked into his office.
  • If I were to become a client, I would need to relinquish current control and trust his ability to follow and adhere to the model process. I thought this comment was interesting because even though he expressed some concern over at least one of the current Canadian model constituents, he told me that his job is not to question the model, it is to adhere to the model process, which has yielded market beating returns (relative to the TSX total return) over time (even during 2008/2009).
He then proceeded to show me the Morningstar CPMS database and model selection process, which is an outsourced screener / stock selection model. Advisors who subscribe to CPMS get to choose one of Morningstar's themes (such as value, growth, momentum) in order to construct, monitor, and modify portfolios. The value proposition for advisors here is that resources do not have be wasted on constantly researching ideas. Instead, they use the CPMS model output which saves them time. I wouldn't categorize CPMS as 100% robo-advisory, as the underpinnings of the selection process are based on Morningstar's in house research, so I guess it's a hybrid between rob-advisory and human selection.

Maybe it's just me, but alarm bells started ringing in my head. The advisors at his branch all seem to follow Morningstar for research and stock selection, and they all pay a fee per year to subscribe to the Morningstar model. If Morningstar is that pervasive, why don't I own Morningstar? Maybe there's a runway in Morningstar's robo-advisory process I never knew existed? After all, research is expensive to perform in house as it's a product of human capital. If the industry is moving towards robo-advisory and Morningstar offers robust, fully integrated portfolio management solutions, maybe the answer is research Morningstar?

I believe the advisor is onto something in terms of putting cash to work and in terms of adhering to a process. I have been sitting in almost 50% cash for most of 2016 waiting for the perfect idea, and I believe I have been putting pressure on myself as a result of this. I have taken profits too quickly on ideas that showed me a profit this year and have redeployed the proceeds back into cash looking for the next set-up. The problem is, the set-ups never came! We never got a repeat of December/January/February, we just got a slow grind up without much backing and filling.

Perhaps my takeaway here is to work on articulating and defining my process explicitly. Once I have articulated my process, I can adhere to it. My process does not have to revolve around buying into something as exotic as a CPMS model and throwing caution to the wind. It can be as something along the lines of what I'm already doing, such as the following:

"Determine companies with the combinations of highest ROIC and lowest relative valuations drawing from any and all resources at my disposal (once a week/month/quarter) and prepare a short list for further research. Once researched, take initial positions. <Stage #1 - initial research>  Initial position size should be a function of total assets under management. Max position size should be no greater than 5% of AUM <Stage #2 - define initial position size>

Monitor once a month, and do not act on temporary adverse deviations from initial purchase price to exit. Similarly, do not exit after small initial advances. <Stage #3 - monitor subject to allowing holdings to breathe> Eventual exit should be based either on my estimate of valuation targets reached, or on my initial research being proven wrong <Stage #4 - exit rules subject to objectivity & flexibility>"

There should also be a feedback mechanism to evaluate ongoing performance. I manage my parents' retirement assets and each month, I send an email to my parents discussing performance drivers. It may sound silly, but I think I should also send an email to myself at the end of each month discussing performance as if I were a third party client of myself.

I also believe that objective expected hurdle rates have to be explicitly worked into the overall process. For example, I calculated my returns to the end of November, which worked out to around 4.8% annualized. Using my November 18th allocations, I made the following observations:

The main issue here is how to achieve a 10% expected ROR based on the asset mix above. There are three possible solutions, 1) increase expected ROR from common shares, 2) decrease cash and increase my equity allocation, or 3) a combination of 1) and 2).

Here's the problem with 1), expected ROR is already high at 20%, and if I use anything higher, I believe I'm setting myself up for undue pressure and disappointment. I am not a pro, I'm a retail "Joe" with a keen interest in continuing to evolve while not doing anything too stupid. If I expect a 30% ROR on equities, I will most likely be setting the bar too high, and even then, my weighted average return would still only be 5% with current weightings.

2) is more realistic. In theory, I can achieve a 10% weighted average return with a 40% weighting in equities and a reduction of cash to 8%. This would still leave me plenty of dry powder + the ability to continue to make cash contributions to take advantage of weakness to add to holdings. But, this then becomes a question of what to buy to take equities up to 40%. With respect to adding to equities, it can’t be done haphazardly, each purchase has to be based on a foundation of research.  I could go out tomorrow and throw darts at 10 names x $3K each, and I may end up lucky and get to my ROR!  But then again, I may not!

December 9 update

Since I wrote the original piece on advisory, I changed the portfolio weightings as follows, by decreasing my overall cash, fixed income and preferred weightings and increasing exposure to equities, without hopefully adding too much risk. In this case, my expected ROR is around 8%, but I have lots of dry powder to put to work in order to get to my target.  Around $10K of fixed income strip bonds are maturing at the end of December, so fixed income will drop to 37% from 44%, but cash will increase to 28%, so either I find replacement fixed income or I find additional equities, neither of which is as easy as it sounds.

Here's what I added in order to bring re-calibrate the weightings in the portfolio:
  • Added $5,000 of each of  the Chou Associates Fund and the Chou RRSP Fund
  • Added $3,675 of The Jean Coutu Group
  • Added $3,800 of Coty Inc.
  • Added $2,500 of Fairfax Financial
  • Added $3,000 of Gluskin Sheff
  • Added $3,800 of Linamar
So, in total, I increased equities by around $27K.

I'm going to provide a brief synopsis of why I added some of the above holdings.

Chou Associates and RRSP Funds

I have long been an avid reader of Francis Chou's writings. His annual letters can be found here.

He's an under the radar portfolio manager and I believe he's brilliant. I know that the dividend growth universe typically leans towards low cost do-it yourself investing (either by way of fractional share purchase auto-pilot investing or by buying ETF's, or a combination of both), and buying units of an actively managed fund (and paying management fees) is typically frowned upon, but, I believe there is a place in portfolios for intelligent active management. Why? Because if participants have embraced one approach, i.e., passive low-cost ETF indexing, to the detriment of active management, eventually, most, if not all participants will be positioned similarly, by mirroring the index. And when most, if not all participants all buy passive low-cost ETF index funds, in theory, the index should get more expensive, and active management should get cheap. So, what was initially viewed as a cheap way of owning the index becomes risky (over time).

I believe that Francis Chou offers an intelligent alternative to closet indexers believing they will generate sufficient alpha over time owning passive ETF's, and his record speaks for itself:

Here are the Chou Associates Fund returns between 1997 & 2006:

And here are the Chou Associates Fund returns between 2006 & 2015:

I split my analysis into two 10 year time-frames. Between 1997 & 2006, the Chou Associates Fund beat the S&P 500 total return in terms of average return, standard deviation, and risk adjusted excess return as measured by sharpe ratio. Between 2006 & 2015, the Chou Associates Fund underperformed the index, and I believe the majority of the underperformance is mostly attributable to three years, 2011, 2014, & 2015. The problem in 2011 is explained below:

So, Francis Chou was buying TARP warrants beginning in 2010 and well into 2011, which impacted returns. While the index rose on the back of Google, Amazon and Facebook, Chou ran concentrated bets in TARP warrants and other US financials that came to life last month.

Here are the current holdings as at June 30, 2016:

I'm willing to bet that the underperformances relative to the S&P 500 are outliers in 2011, 2014 and 2015, and the contrarian in me is willing to take the opposite side of Morningstar's 2 star rating by adding Mr. Chou's investment savvy to my retirement mix. My guess is that Mr. Chou will trounce the index over a sufficiently long period of time. I want to be buying Mr. Chou when Morningstar rates him 1 star over the last three years (too bad Morningstar didn't bother to analyze the reasons for underperformance), and everyone in the advisory world chases passive index ETF's or 5 star Morningstar rated funds hoping for linear extrapolation into the future.

The risks here are as follows:

  • Concentrated bets in some really nightmarish stuff (i.e., Valeant, Ascent Media, Resolute Forest, Sears Canada, Sears Holdings, Sears Hometown). At first glance, the weighting of the obscure bets was around 30% of total investments. The Valeant weighting was around 9%, and the total Sears weightings were around 11%. The counter argument here is that there is optionality in the event that these bets work out (but then again they may not)
  • Illiquidity: the funds have a minimum purchase amount of $5,000 and a 2% redemption fee charge if redeemed within one year. 
  • The fund has generated negative returns in previous years, -17% in 2011, -29% in 2008, -10% in 2007, -9% in 1999, -2.5% in 1994, & -11% in 1990. By my calculations, the standard deviation of the fund returns is not materially different than that of the overall index, so anyone invested in passive index ETF's already (implicitly) accepts the same volatility risk.

I'm currently researching the Chou Europe Fund as a possible candidate for consideration, and I'm honestly scared just looking at the holdings (almost 22% of the holdings are in Greece, and 14% of the holdings are in two Russian pharmaceutical companies). Maybe the fact that I'm feeling scared should be a reason to purchase units of the fund! The fund has performed horribly over the last few years, most likely due to the Greek bets.

The Jean Coutu Group

I wrote extensively about The Jean Coutu Group last year and I believe I came to a premature conclusion that the shares were worth around $10. This was first level thinking and I was wrong. I re-analyzed Jean Coutu recently, and while there still is concern over Pro Doc and the removal of the Quebec pharmacy cap, I believe that these concerns have been worked into the rhetoric for almost two years now, as the company has deliberately warned that future results will be adversely impacted, and during this period, the market has had plenty of time to digest the warnings and potentially reprice risk. One estimate from TD last year pegged the impact of an increase to 40-50% professional allowances at a negative $25M-30M hit to EBITDA, link here. 2015 EBITDA was $331M, so a $25M-$30M hit would reduce consolidated EBITDA to $301M at the low end.

Per the 2016 AIF, the company owned 184 stores, 146 of which were leased to franchisees, and the company sublet another 268 stores. This gets us pretty close to the 417 network stores in operation. In 2015, the company generated $100M in rental income ($39M from owned properties and $61M from subleases). At a minimum, the $39M rental income stream should be valued separately. Assuming NOI is about 2/3rds of gross rental income, a cap rate of between 6% & 7% suggests $39M x 2/3 / 6.5% = $402M. I'm going to ignore attributing any value to the sub-leases, as they should be a wash.

Enterprise value as of Friday was $3.86B, less say, $402M attributable to the value of owned property real estate, gets us to $3.46B for the operating business. At $3.46B, the company trades at 11.5x EV/EBITDA, as compared to say, Shoppers Drug Mart, purchased in 2012 for $12.4B plus assumption of debt and outstanding operating lease commitments, which at the time, amounted to around $4B (un-discounted). I estimate that Shoppers' 2013 EBITDA amounted to around $1.18B, so this would put the purchase price multiple at around 14.7x EBITDA.

The main investment risk here is that bill 81 levels the generic playing field in Quebec, further hampering (or worst case, eliminating) Pro Doc. With reference to the 2016 annual report, Pro Doc contributed 27% to consolidated EBITDA, and without Pro Doc, the company becomes a pure play franchisor with EBITDA of $240M. Estimated EV/EBITDA in this case jumps to 14.4x, in line with the Shoppers' purchase multiple. The investment thesis thus becomes predicated on how much of an impact bill 81 will have. If an investor believes that Pro Doc will still be relevant (although less of a contributor to overall profit vs. pre bill 81), then there is an investment case to be made.

Finally, Revisiting Non-NVCC Insurance Preferreds (and concentration risk)

I wrote extensively about non NVCC insurance preferreds last year. As at November 18th, I had close to 8% of assets invested in these preferreds (around $13K). While most of these had returned a slight profit + dividends since I purchased them last winter, I came to the conclusion (whether rightly or wrongly) that a basket of $13K worth of non-NVCC insurance preferreds is no different in terms of risk than buying one or two of these names. Obviously there are different characteristics between the fixed and floating names and different reset spreads amongst the group, but the end game is the same, an investor is buying these in the hope that OSFI eventually pushes NVCC rules down to insurance co's, resulting in eventual redemption at par prior to 2025 (James Hymas has discussed extending deemed maturity to 2027, which would further force YTW down).

So, rather than have $13K spread amongst MFC.PR.F, GWO.PR.N, SLF.PR.G, SLF.PR.J, IFC.PR.A, I did what I thought was logical and sold out of 1/2 of the holdings to bring the weighting down to 5% of assets, as I believe I have concentration risk in owning all of these preferreds as it's all the same bet.


  1. I do think its kind of funny this adviser tells you to be 95% adjusted and kind ride out the ups and downs in the market but can't own small caps because of liquidity. He needs to read this Howard Marks letter/essay on liquidity https://www.oaktreecapital.com/docs/default-source/memos/2015-03-25-liquidity.pdf?sfvrsn=2

    I would also be curious if he provided any audited returns that proved he was beating the market. Further, if he is split 50-50 between Canada and the US, it probably wasn't hard to beat the TSX if that was his benchmark. He should blend the S&P 500 (CAD) and the TSX to calculate his returns.

    I'm worried about all of these robo advisers / quant jockeys in terms of what it might mean for the average joe but its good news for stock pickers.

    1. How are you doing Safety? Yeah, the mandate is 95% all in and use the dips to add more cash. Not much different than BTFD rhetoric which has been in place for some time. My question is, if the majority of participants are all following a similar system, what happens when the system doesn't work? I calculated stdev of S&P 500 total return since 2010 (I excluded 2009 as this was a bounce back year after 2008). The stdev is around 11, which is about 2/3rds of the historical stdev since 1965. Has this fostered complacency? Maybe.

      What he showed me was the data imbedded in the Morningstar CPMS database which tracked historical returns vs. the benchmark total return over 15 years. CPMS did outperform the benchmark (I'm pretty sure this return data is audited too). You're right, if he's 50/50 US/Canada, he should be using a blended benchmark and I can't remember if the benchmark was blended (or not).