Saturday, 26 September 2015

Jean Coutu and First Level Thinking

I recently read Howard Marks' letter to Oaktree shareholders, "It's Not Easy".  In this letter, Marks essentially describes how difficult it is to consistently generate abnormal (excess) returns.  Passive investing at regular pre-set intervals, will never generate abnormal (excess) returns.  At best, the result will be the equivalent to the market return (S&P 500).

Now don't get me wrong, there is absolutely nothing wrong with the market return.  If an investor holds for a long enough time, and continues to deploy equal dollar amounts into the market (or a basket of stocks representative of the market), he/she should earn the market return.  I have no idea what the market return will be from here (and neither does anyone else), but if history is an indication, the return over a long enough holding period (20+ years) should be somewhere between 6% and 8% (including dividends) compounded annually.

Marks goes into substantial detail describing the concepts underpinning First Level vs. Higher Level Thinking (as an aside, First Level Thinking is something I've likely been guilty of my entire investing career thus far).

In my opinion, First Level Thinking is symptomatic of the most dangerous kind of cursory analysis, i.e., a stock is cheap if it has a low enough P/E, and it's expensive if it has a high P/E, without asking why the stock's P/E is what it is in sufficient detail to develop more than a cursory analysis of the business.

Similarly, First Level Thinking is turning to consensus analyst estimates for validation of an investment thesis.  Investors for the most part, would be best served ignoring analyst estimates entirely, for the simple reason that the analysis is not their own (not to mention the potential conflict of interest inherent as a result of issuer/bank relationships).

As I digest Marks' letter, I'm beginning to wonder if my EPV model doesn't exhibit First Level Thinking.

Case in point, I recently read a Globe and Mail article on Jean Coutu, one of the largest remaining public pharmacy franchisors in Canada.  The majority of their franchised pharmacies are located in Quebec, with a smattering of pharmacies in Ontario and other Eastern Canadian provinces.

The article spoke to an impending Quebec rule change in the works which would essentially eliminate the current 15% cap paid to pharmacists for generic drug purchase allowances.  From the sounds of it, the removal of the cap will negatively impact Jean Coutu's operating results at its Prodoc subsidiary, as pharmacists would be free to negotiate supply of generics from the highest professional allowance bidder (I'm assuming professional allowances are similar to volume rebates in return for a long term supply commitment).

Currently, pharmacists appear to be on a level playing field in dealing with generic manufacturers in Quebec.  The current 15% cap on allowances creates no incentive for pharmacists to look for more juice in terms of allowance.  Compounding things, Prodoc has been the engine of Coutu's growth: the generic drug manufacturing business was expected to result in additional sustainable growth going forward as non-generic drugs come off patent protection and Prodoc goes to work.

If the cap goes away, pharmacists could shop the supply business elsewhere unless Prodoc agrees to match competing allowance terms.

After reading this article, I ran my EPV model, which, on a cursory basis, produced a very non-spectacular result.  The results of my initial valuation matrix are below.  As you can see, using just the published IFRS #s, as a starting point, Coutu looks extremely overvalued at a range of WACCs between 7.5% & 10%.

Cost of capital rates:EPVAdj for Debt & Excess CashAdj EPVO/S sharesEPV / shareBVPS / shareExcess EPVCurrent pricePrice / EPVPrem/Discount vs EPV
Upper bound - VC15%1,643.7598.21,741.951879.325.661.6519.912.14113.74%
Combined NI & Div growth12.90%1,910.9798.22,009.1718710.745.661.9019.911.8585.31%
LT equity return; 2014 - 192811.53%2,138.4598.22,236.6518711.965.662.1119.911.6666.46%
Arbitrary 2 <10% - 12 %>10.00%2,465.6398.22,563.8318713.715.662.4219.911.4545.22%
Per GF8.46%2,914.4698.23,012.6618716.115.662.8519.911.2423.58%
Arbitrary 17.50%3,287.5198.23,385.7118718.115.663.2019.911.109.97%

And I started thinking how this could be when the stock has come down from close to $30 since 2014 to about $20 currently (a loss of about $2B in market cap in less than a year).

At $28 per share, investors (or traders) were tickled pink to pay close to 30 x ttm earnings (ex-gain on disposal of Rite Aid).  Today, investors (or traders) don't seem overly interested in paying 18x ttm earnings.  So what changed so significantly between late December 2014 & now?  I'd argue nothing but sentiment coupled with diminished expectations surrounding Prodoc.

Here are some comments on Prodoc's contribution to both 2015 and the last three year's quarterly rolling results:
  • Generic drugs accounted for 24% of OIBA (Operating income before amortization) in Q1 2016 compared to 28% in Q1 2015
  • Generic drugs accounted for 27% of OIBA in 2015 compared to 23% in 2014
  • Per the following rolling three year quarterly analysis, the contribution in terms of both revenue and OIBA from generics has gradually increased over the last three years (all data taken from the 2015 and 2014 annual reports)

Q4 2015Q3 2015Q2 2015Q1 2015Q4 2014Q3 2014Q2 2014Q1 2014Q4 2013Q3 2013Q2 2013Q1 2013
Generic drugs 464547404244413345423434
Total Revenues714737675689685713654682683717659682
Generic drugs as a % of Revenues6.44%6.11%6.96%5.81%6.13%6.17%6.27%4.84%6.59%5.86%5.16%4.99%
OIBA (Op income b4 Amortization)252427212327251522201616
Total OIBA848581828888778282857779
Generic drugs as a % of OIBA29.76%28.24%33.33%25.61%26.14%30.68%32.47%18.29%26.83%23.53%20.78%20.25%

Based on the above, it's not surprising that investors would have an expectation of a continuation in trend in terms of contribution from generic drugs (Prodoc) to OIBA going forward, despite management hinting at the potential deflationary pressures on generics from both a regulatory and a competitive standpoint, and, given the existing trend, it's not surprising that in response to the Quebec rule change announcement in July, investors who believed one version of an alternate reality, lopped 14% off Coutu's market cap in less than two weeks between July 3 and July 14th.

For my purposes, I need to determine, by way of stretching my thinking beyond First Level Thinking, where there is a theoretical margin of safety.

In order to do this, I have to do enough digging in order to analyze Coutu's overall business.

Here are my findings:

  • Coutu is actually three distinct businesses.  Franchising, manufacturing and sale of generic drugs via Prodoc, and leasing company owned pharmacies.
  • In order to value Coutu, it doesn't make much sense doing a simple cursory EPV analysis, as all three businesses are lumped in together in the consolidated results.  
  • There's a potential value proposition hidden here in that the leased stores constitute a seperable revenue stream, and there is value in the company owned real estate itself.
  • The EPV should adjust for the "implicit value" of both the lease revenue stream and the company owned real estate.  This implicit value could be deducted from the price per share in order to compare to the stand alone capitalized value of just the franchise fee revenue and the generic drugs.
Similar to my previous analysis of Garmin, I'm interested in seeing if I can theoretically reduce the contribution from generic drugs to zero, and still come up with a theoretical margin of safety.

First task, estimate the value of the real estate:

In order to do this, I had to do some digging and look to previous commercial deals, preferably involving retail space.  I found the following four deals completed over the last 4 years, and I came up with an estimated value of the real estate of about $2.83 per share (not including the value of the sub-leases in place or the value of the real estate held for development).

CompanyMkt value ($B's)Sq feet (m's)P/sf
Cdn Tire$3,500.0019$184.21
PJC?$528.862.35$224.76<181 owned stores, avg sq feet /store = 13K>
O/S shares187
Estimated Pps?$2.83

Next, estimate the value of the rental revenue stream:

Per Coutu's 2015 annual information form, of the 181 owned stores, 141 of them are leased.  Per the 2015 annual report, rental revenue was $97.1M, ($37M from owned stores, and $60M from sub-leases).  Rental revenue as a %ge of total other revenue in 2015 amounted to $97M / $281M = 34.5%.  Assuming a similar proportion applies to non-generic OIBA of $242, I get 34.5% x $242M = estimated NOI of $83.5M.  But, $83.5M is both owned store rental revenue and sub-lease revenue (which should wash vs. external leases paid).  Taking $37M/$97.1M x $83.5M, I get $31M in estimated NOI attributable to leased stores.

In order to determine the value of the stand alone leases, I performed the following analysis using recently published Canadian retail REIT data, courtesy of Brookfield Financial's April 1, 2015 Realty Market Update.  I basically wanted to estimate a range of cap rates in order to value the rental stream in perpetuity, and in order to do so, I looked to recent P/AFFO multiples with the expectation that the inverse of P/AFFO could be a reasonable proxy for a suitable cap rate.

Here are the results:

2015 E2016 EAverage1 / P / AFFO
1 / P/AFFO as a proxy for cap rate?
Choice PropertiesP / AFFO14.814.414.66.85%
Plaza RetailP / AFFO15.614.515.056.64%
Crombie REITP / AFFO14.213.8147.14%
CT REITP / AFFO16.115.415.756.35%
Riocan P / AFFO18.817.918.355.45%
Slate REITP / AFFO1211.411.78.55%

Here is my cap rate analysis and determination of estimated price per share allocable to the rental revenue stream based on the lowest, average and highest cap rates per above:

Cap rates5.45%6.83%8.55%
NOI / Cap rate$569$454$363
O/S shares187187187
Estimated Pps?$3.04$2.43$1.94

Value of the real estate + value of the capitalized rental stream:

Overall, I've estimated that the total "implicit value" of both the lease revenue stream and the company owned real estate ranges between $4.77 and $5.87.  If I take the average of the range, I get $5.32

Feedback loop and recalculation of EPV using above results:

Now I'm ready to revisit my EPV analysis using the following adjustments:

  • Current price per share should be reduced from $19.91 to $14.61 to reflect the "implicit value" of the both the lease revenue stream and the company owned real estate 
  • EBIT should be reduced by the estimated rental revenue contribution to account for seperate valuation per above

 Here are my EPV results at $14.61:

Cost of capital rates:EPVAdj for Debt & Excess CashAdj EPVO/S sharesEPV / shareBVPS / shareExcess EPVCurrent pricePrice / EPVPrem/Discount vs EPV
Upper bound - VC15%1,629.7398.21,727.931879.245.661.6314.611.5858.11%
Combined NI & Div growth12.90%1,894.6798.21,992.8718710.665.661.8814.611.3737.09%
LT equity return; 2014 - 192811.53%2,120.2198.22,218.4118711.865.662.1014.611.2323.15%
Arbitrary 2 <10% - 12 %>10.00%2,444.6098.22,542.8018713.605.662.4014.611.077.44%
Per GF8.46%2,889.6098.22,987.8018715.985.662.8214.610.91-8.56%
Arbitrary 17.50%3,259.4798.23,357.6718717.965.663.1714.610.81-18.63%

A better result than before, but not compelling in terms of valuation or margin of safety.  I'd really like to see a discount to EPV of at least 25% to 30% at a 10% WACC, and I wonder how Coutu might get there.  For anyone interested, this would be equivalent to PPS of $10, per below:

Cost of capital rates:EPVAdj for Debt & Excess CashAdj EPVO/S sharesEPV / shareBVPS / shareExcess EPVCurrent pricePrice / EPVPrem/Discount vs EPV
Upper bound - VC15%1,629.7398.21,727.931879.245.661.63101.088.22%
Combined NI & Div growth12.90%1,894.6798.21,992.8718710.665.661.88100.94-6.17%
LT equity return; 2014 - 192811.53%2,120.2198.22,218.4118711.865.662.10100.84-15.71%
Arbitrary 2 <10% - 12 %>10.00%2,444.6098.22,542.8018713.605.662.40100.74-26.46%
Per GF8.46%2,889.6098.22,987.8018715.985.662.82100.63-37.41%
Arbitrary 17.50%3,259.4798.23,357.6718717.965.663.17100.56-44.31%

Concluding Thoughts

Overall, there doesn't appear to be a sufficient margin of safety reflected in the current price, even taking the estimated values of the real estate and the rental revenue streams into consideration (however, my estimates may be too low as I have not attempted to value the land held for development or the sub-leases).  My thoughts on how it could get there:

  • Quebec rule change is finalized and Coutu quantifies the likely short term impact as a result (synonymous to a short term profit warning)
  • As a result, Coutu loses a significant portion of its generic supply business to a competitor in the wake of the Quebec rule change 
  • Investors/traders flee the stock in response to the above

Finally the Long Term Chart

No comments:

Post a Comment