Wednesday, 17 May 2017

Follow up on Coiled Spring Experiment and Final Blog Post (New Blog Coming)

It's been a long while since I last posted. My regular profession as a Chartered Professional Accountant resulted in my being inundated during tax season (in Canada this runs from basically, end of February through end of April) with very little time to opine on anything. But there's been lots of time to think since my last post.

I wanted to debrief my last couple of posts on options trading with a number of irrefutable observational truths in no particular order:
  • Option trading at the best of times is a frustrating endeavour, enough to drive a participant absolutely nuts. 
  • Options are not assets, they are decaying derivative instruments. If a participant does not realize this up front, he/she has no business participating. How best to understand the concept of price change and decay? Start with understanding theoretical greeks. Theta measures time decay. Delta measures sensitivity to changes in the price of the underlying instrument.
  • The more time there is to expiry, the more expensive the time value component of the option price is. As time lapses, time value decays, and the closer to expiry, the faster the decay.
  • Options are priced by professionals, using professional models (Black Scholes, etc.). On top of theoretical pricing is the bid/offer spread. A market maker will use a combination of theoretical model output + the bid/offer spread to make a market. The more liquid the underlying, the more liquid the option will be, and vice versa. So, not only are participants up against a decaying derivative instrument, they are up against professional market makers who set the option spread on any given day. They are also up against their own particular biases. 
  • Not only do participants have to be right in terms of strike price and time to expiry, they have to be right about direction, all while not getting juiced on the spread. All of this begs the question, why bother in the first place?
  • It is not uncommon (in my experience) for market makers to price the tails surrounding a price range at greater than the near term volatility range (on the offer anyway). For example, say a stock is trading at $50, and the near term at-the-money straddle (i.e., the $50 call + put) is pricing in between a 5 & 10% move. This would make the premium of both strikes between $1.45-$2.50 each, and depending on how liquid the underlying is, the market maker could vary the bid offer spread according to his/her discretion. For example, he/she could make a market with a $.05 spread (i.e., $2.45/$2.50) or a $.50 spread (i.e., $2/$2.50). The $.50 spread is an invitation to lose money. The market maker can then set the range of prices in the chain up/down accordingly. For example, at say +/-10% x $50, he/she could set the $45 put at $0/$.50, and the $55 call at $0/$.50. I ask again, why bother?
So why am I going into all of this detail?

Well, firstly, the Coiled Spring Experiment was a cute idea, but it was stupid. So I identified a bunch of charts with triangles. Who cares. Is this research? No. On top of that, I suggested buying strangles midway between the expected break-out range. What does this particular trade setup achieve? It really achieves one thing, an almost certain transfer of wealth in terms of premium paid for the strangle between the participant and the option writer.

Secondly, while the strangles were a novel idea back in January, and while one leg of the strangle did appreciate noticeably in the direction of whatever breakout occurred, the other leg depreciated noticeably once direction was established, and if my call on the mid-point of the breakout range was too distant or the premium paid for both legs ended up being more than the intrinsic value of the winning leg by expiry (i.e., for a call, IV = stock price - strike price, and for a put, IV = strike price - stock price), then the overall experiment was a losing proposition. As theta burn kicked in during mid-March, the winning leg started to decay feverishly. At least if I was long the underlying in the direction of the move, I'd be right on direction with no time limit, and no decay, and I'd earn dividends while I waited.

Thirdly, none of this is a good use of time or effort. Looking back at the above, I believe that the Coiled Spring Experiment was an exercise in first level thinking, something along the lines of, "hey, there's a triangle, it's going to resolve and I'm going to be able to take money from a derivatives desk who's only job is to professionally set the odds (by the second) by paying up for premium six months out".

Sounds dumb in hindsight doesn't it? Where is the research to substantiate risking capital above and beyond observing a chart pattern?

Fourthly, as I mentioned above in my irrefutable observational truths list, I've observed (experientially) that the tails around a volatility range are often over-priced. This negates the ability to buy tails cheap, and I believe that the best way to make money in options is to pay nothing, and therefore lose nothing. This really means paying nothing for the tails, so in the event of a abnormal standard deviation move, the tails get valuable. But market makers are not stupid, and they don't easily get picked off, which explains why the tails often appear overpriced relative to the near term volatility range.

In the context of the $50 stock above,  it wouldn't surprise me at all to see the $50 calls, $52.50 calls, and the $55 calls priced as follows (assuming 25% implied volatility and assuming a $.50 spread):

$50, $.95/$1.45
$52.50,  $.05/$.55
$55, $0/$.50

There's no way to buy cheap options in this case.

So in the order of final debriefing:

  • Options are a losing proposition from the outset, so why bother?
  • If a participant wants to trade options, the best way to participate is to buy tails for cheap with a lot of time left for the underlying to do whatever it's going to do. BUT, in my observational experience, tails don't often get cheap due to market makers setting wide bid/offer spreads. And the longer the time to expiry, the more expensive the tails get due to wide spreads. The best way to participate is to be the market maker yourself and just have silly bids in between the existing spread and hope they get hit, and the likelihood here is that the bid will just sit there and not get hit, which begs the original question, why bother?
  • If a participant is lucky enough to have a silly tail bid filled with enough time to expiry, he/she should look to immediately sell another strike against the premium paid in order to have a spread on for free. Case in point on the $50 stock, if the ATM straddle is pricing in 10% and the bid/offer spread on the $55 calls is $0/.$50, and and the bid/offer spread on the $45 puts is $0/$.60, have silly bids in for $.05 on both legs. If either of the bids happen to get filled, look to sell the $60 calls, same expiry, for $.05, or look to sell the $40 puts for $.05. This way, the participant has either the $5 call or the $5 put spread on for free. I believe this is the only way to successfully trade this stuff. Have enough premium on that if it erodes to zero, you don't care, and if you just happen to get lucky enough for the underlying to blow through your short strike before expiry, you paid nothing and you make something.
  • There's a hidden element of psychology embedded in this entire discussion, and that is, what does your own read of a chart tell you about your own psychology and biases, and what does it tell you about other participants' psychologies and biases'. If I see a triangle, you can bet other participants see a triangle, and participants will either try to front-run the resolution or wait for resolution either in the underlying or in the options. Going back to my original post, I identified a set up in Abbott Labs that hadn't resolved yet in January 2017. The May 2017 $44/$38 strangle was priced at $2.35 in January with the stock at $40, the May $44C was priced at $1.01, and the May $38 put was priced at $1.34. Abbott broke up, but moved nowhere near midway between the $13 breakout range I originally identified in January. My bias was to overlay my expectations on resolution vs. what price actually did, and my model paid too much for the privilege of trying to participate. In theory, a $13 range/move suggested a +/- 30% move in the underlying, and here's where the gut check comes in: what research was done to support a +/- 30% move beyond observing a nice looking chart pattern? NONE. How likely are +/- 30% moves without some fundamental change in the operating business?
I close off with the following final observation:

If there's no research tied to allocation of capital, don't bother. And if a participant wants to bother for the fun of it, don't EVER overpay for the privilege.

On the subject of a new blog

I'm going to start a new blog which does not have Stable, or Dividend, or Portfolio in the title. As I progress in terms of experience (what fools endearingly refer to as their mistakes over time), I want to write closer to home on a range of financial topics which may include or exclude any or all of the above aspects of my previous blog.

I'm thinking of either two titles:

1) Confessions of a retail muppet 

2) Diary of a retail muppet 

I will try and get this set up in the next few weeks.

Thanks all for reading.

Monday, 16 January 2017

The Real Time "Coiled Spring" Experiment: Nike, Discovery, Abbott, Lennar, and Viacom

All of this thinking over the weekend on pennants, triangles, and strangles got me thinking.

I have long been of the belief that consolidations inside pennants / triangles lead to resolution, either up or down once price reaches the apex of the pattern. The longer the pattern of consolidation, the more meaningful the resolution.

The issue I have always had in my own speculation is leaning to one side in terms of positioning myself in advance of any resolution, and this is a dangerous endeavour because I truly believe that flexibility is the epitome of a successful speculative trading operation, not stubbornness. Anticipatory trading is bad enough: anticipatory stubbornness is a sure path to losing money!

(On the flipside, stubbornness is most likely the epitome of a successful investing operation, as evidenced by great investors who have succeeded because they have had better and deeper anchors than their participant counterparts in evaluating and sticking with great businesses over time)

Well, it's time to put my theory to test, partly by way of a real money experiment, and partly by way of a hypothetical paper trade experiment.

I spent this past weekend scouring the S&P 500 and the DJIA for monthly or weekly consolidation triangle / pennant patterns and have found the following 6 charts out (of 500+, believe it or not).

First up, Nike (already mentioned last week):

Next, Discovery Class K:

Next, Abbott Labs:

Next, Lennar:

Next, Viacom:

And finally, Sealed Air:

There is a noticeable commonality amongst all six setups, and this appears to be characterized by a long term consolidation inside a pennant / triangle over the last two years in each case. I have indicated total distance inside the consolidation, and the option strikes to be considered on eventual breakout/resolution.

I have added a link to a new model to the blog which assumes that I invest equally in six strangles positioned slightly above and slightly below the apex of each consolidation pattern, and this is where I ran into my first problem, in that the spreads on the Sealed Air options are too wide, so I dropped Sealed Air from consideration due to illiquidity in the spreads.

I used the closing prices of the strangles as quoted at optionsexpress for each of the five remaining setups being considered as of Friday January 13th, 2017, assuming I purchased the strangles at the market.

In my new model spreadsheet, I have added objective targets based on breakout / resolution resulting in a move up or down equal to the triangle / pennant width, and my assumption in updating the model once a week is that I will have GTC orders in the market at all times equal to a theoretical profit of 1/2 x the resolution distance to be conservative.  Once a pattern resolves, the 1/2 x resolution P/L GTC order will theoretically close out one half of the strangle, while the other half is left to expire at a loss.

By conducting this experiment, I hope to see how resolution unfolds in real time on a forward looking basis.

As disclosed on Friday, I am currently long two June Nike spreads ($57.50/$62.50 and $47.50/$42.50) and I have decided that being the long the spreads is incorrect as it locks me in and there is diminished flexibility inherent in being short the out of the money spread strikes. Therefore, I have an order in the market GTC to buy back the short legs of the spreads which I will execute at some time tomorrow, leaving me long one Nike $57.50C  / $47.50P strangle.

I also have a GTC order to buy an ABT May $44 C / $38P strangle outright for a debit of $1.80 based on my evaluation of potential resolution of ABT's current consolidation.

Once I have executed my closing NKE short legs and ABT long strangle, I will immediately place GTC orders in the market for each leg of the strangle I own based on my evaluation of potential resolution in the market.

Finally, I have compared the overall P/L of the "Coiled Spring" Experiment to equal dollar exposure to SPY as of the close on Friday January 13th, 2017.

This should be an interesting experiment overall.

Saturday, 14 January 2017

Follow up post on TA and the psychology behind setups; Cisco, Brown Forman, And Abbot Labs

Passive Income Pursuit asked a great question on learning about TA. My response is detailed in my previous post, reproduced below:

"Ok, on TA, it may be worth doing an entirely separate post on this subject. My layman's opinion is that it works, and at the same time, it doesn't work. It works by virtue of participants believing it works. I now look at TA entirely from a psychological perspective, but I never used to. There are loads of participants at any one time looking to latch onto breakouts in either direction. This type of trading fits into trend following. A successful breakout is usually followed by a trend. The problem here is that because so many participants are looking to trade breakouts, the event of the eventual breakout often occurs after a bunch of false breakouts, whipshawing traders. So I've come to believe that while trading breakouts can be a viable system, a participant must be mentally prepared to lose on the first n independent trials before the breakout actually occurs, and even then, the breakout may go in a direction which was completely unexpected by the majority of participants.

Case in point, SPY on the night of the election, financial MSM worked all participants into a frenzy expecting a crash if Trump ended up winning, and the majority must have been positioned for a crash. The night of the election, ES was down close to 10% at one point and anyone short the hole in the futures market that night got creamed on the open. 

So my overall perspective is that TA is a function of my analyzing long term time-frames (monthly charts mostly), and making an educated guess as to how the majority of participants are positioned, and then doing the opposite in the options market. I seem to have stumbled onto this almost by mistake. I literally study all of the Dow components, the QQQ components and the S&P500 at least once a week on a monthly basis to see whether any setups look enticing. And for the most part, I can't find many enticing looking setups right now. The only Dow component that looked interesting was Nike b/c it appears to have consolidated for almost a year inside a triangle and this type of consolidation is usually followed by resolution out of the triangle (I just don't know which way, nor do I care).

There are more elements to this, including studying setups and evaluating which setups appear more probable in terms of volatility resolution, looking at the liquidity in the options market to see whether it makes sense to play, and looking at cost of the options themselves.

For example, I looked at the monthly setup this week on Brown Forman (BF B). It looks like a nice bearish setup on the monthly, about to break down. The problem here is two-fold, 1) if I'm noticing a bearish setup, you can bet other participants have as well, so the smart play is to either play both sides (like I have done with Nike) in case the break down doesn't resolve, and/or wait for resolution and trade in that direction, and 2) there is no liquidity in the options as the spreads are too wide, therefore the answer to the BF B problem is, don't bother playing.

If you really want to study TA, I suggest approaching it from the perspective of learning basic patterns and then trying to figure out the psychology behind the pattern itself. If you notice that a stock is currently in a one year triangle, you can bet that everyone else out there notices the same pattern and is waiting for resolution, and you can probably bank on a good %ge of those waiting for resolution being wrong when the resolving event actually occurs. This is why trading is so difficult, because you are fighting against yourself first, and if you get married to your perspective and are inflexible, it's about as good as flushing money down the toilet."

So what patterns seem to have a high probability of working (in my opinion)? I think this is probably easiest illustrated by virtue of some current chart setups I have on watch. I've had the most success in trading pennants or triangles, and the least success trading patterns like H&S tops (more on this later).

Caveat, none of the following are recommendations, and in most if not all cases, I'm going to argue both for and against the pattern working, so I may end up leaving readers more confused than when we first started the discussion.

First up, Cisco monthly:

I'd characterize this as 15 years of nothing. So the obvious question becomes understanding the rationale behind the nothing over the last 15 years, and trying to understand what the catalyst might be for Cisco to get over multi-decade resistance at $35. This is where fundamental research comes into play, in order to tie everything together.

One way to play this setup is long $35 leaps. January 2018 $35 calls can be had for around $.70. The risk is the premium. The obvious pattern appears to be a rising channel of some sort. The problem I have with this set up is that it's too obvious, and I can see that the January 2018 $35 calls have the largest OI out of all of the strikes (42K in OI). This could be symptomatic of participants selling calls against the underlying in order to generate additional income, or it could be symptomatic of participants positioning for a breakout. The pattern is therefore ambiguous. It's a chop until it actually breaks out above $35 and holds above it. A better way to play this is to probably just buy Cisco and collect the dividend and let it do what it's going to do.

A less obvious point to put on the January 2018, or even better, the January 2019 $35 calls, is on a sharp move down to $25 on an overall market or Cisco specific correction. If Cisco moved down 15-20% on an earnings disappointment or on an overall market correction, the contrarian play would be to have orders in the market to buy the January 2018 $35 calls for between $.15 & $.20 and/or the January 2019 $35 calls for between $.35 & $.45 as there seems to be significant t/l support going back to 2012. To me, the most opportune time to add risk in terms of premium is when risk comes out of the market. For the same $.71 of risk today, I can have a GTC order in the market to buy 3 Jan 2018 $35's at $.20 or 2 Jan 2019 $35's at $.35.

Somewhat riskier would be putting on a risk reversal on a sharp move down to t/l support, i.e., sell the Jan 2018 $25 strike puts to finance the purchase of the Jan 2018 $35 calls for a credit. The risk here is that you are obligated to take delivery of shares at $25 should Cisco keep going right through $25 if t/l support doesn't hold. If you are only long premium, you simply lose premium.

Next up, Brown Forman monthly, here's the current setup:

I hate to admit it, but this looks like a giant H&S top. The problem with this setup is that if I see it, so too do most other participants, and my experience with H&S tops is that they seem almost as likely to fail to complete as they are to complete, so if the options were cheap enough, I'd look to buy strangles or spreads in both directions in the event that the H&S top fails and catches everyone short positioned incorrectly. The problem with the options market here is that the options are expensive, the spreads are too wide, and there is no bid on the June $40 puts to hit, so the answer here is, move on and don't play in this playground.

Finally, Abbot Labs:

Once again, an ambiguous setup which looks like a long consolidation inside a triangle/pennant. No one knows which way it's going to break, so certainly this could be a candidate for a long strangle or spreads both ways in case in breaks up or down. The May 2017 $43/$38 strangle can be purchased for around $2.10. The distance inside the triangle/pennant is around $13, so I expect some sort of resolution either up or down of between +/- $6.50 & $13. This would result in a retest of $50 at the highs, or $28 at the lows. On a $13 move, this works out to a reward:risk ratio of $13:$2.1 of greater than 6:1.

Alternately, a trader could reduce the cost of the strangle by selling the May 2017 $47/$34 strangle against the $43/$38 strangle for a credit of $.54, so the overall cost becomes $1.56, but the downside is a cap on total profit of +$4 each way (at expiry), or more likely 1/2 of $4, so +$2 in the interim time between now and May expiry. The risk here is that there is no resolution between now and May expiry, and you lose the net premium.

Friday, 13 January 2017

Revisiting Disney, and now Nike

Back in August, I put on a trade based on a chart setup in Disney, the link to the original post is here:

And here was the monthly setup I identified at the time:

The way I played this was long two spreads. I bought one Jan 2017 $105/$110 call spread and I bought one Jan 2017 $87.50/$82.50 put spread for a net debit of $1.57.

I closed out the call spread today for $2.95 (after commissions), while the $87.50/$82.50 put spread expired worthless. Not bad for taking a flyer. Here's what I liked about this trade:

  • I was able to profit despite being direction neutral (I was both long and short, I just didn't know which way was correct at the time, nor did I need to)
  • I was able to allow myself lots of time between trade identification and time to expiry
  • I let the market do whatever it wanted to do
Here's what I did not like about this trade:
  • I did not have an open order in the market to close out the put spread. But, this is not a big deal, as Disney never actually got down below $87.50 where the p/s would have been profitable. I think that going forward, it would make sense to at least put a GTC order in the market on both legs in case there's a spike up or down in the market.

Here's how Disney turned out:

In my play account, I'm always looking for interesting setups and ideas, and I stumbled across Nike, see below, the setup looks eerily similar to Disney pre break out:

So, once again, I am simultaneously long the June 2017 $57.50/$62.50 call spread and the June 2017 $47.50/$42.50 put spread for a net debit of $1.54 in my IB account.

The difference now is that I have a GTC order in the market to sell the p/s for $2.95 and I may do the same with the call spread so whatever direction this breaks, I may end up making something on either a move down or a move up (or both).  I may end up adjusting the GTC orders to $2.50 to at least be halfway between the two strikes on either side.

The distance inside the pattern (call it what you will), is around $20, so conservatively, I expect a move either up or down of at least 1/2 of this distance, depending on how exuberant the Fast Money traders are on the day/week of the break out. The one thing I can count on is that the idiots on CNBC will work everyone else watching them into a frenzy chasing the breakout on their say.

My ideal scenario is a sharp move down to test $44, my GTC p/s gets bought by someone in a panic, and then a sharp retracement back up to $64, and my c/s gets bought by someone else in a panic, whipshawing everyone who was positioned incorrectly by listening to Fast Money in the first place along the way.

I know this all sounds a bit evil, but it really is a zero sum game of chess.  I really don't care which way it breaks, I just care that it breaks and hits my GTC orders on one side or the other, and the more volatile, the better.

The risk to me is that it doesn't move below the short strike in either case and I lose my premium.

Sunday, 25 December 2016

Thoughts on the week ending Dec 23, Trump Market Moving Tweets, Globe and Mail 2016 Dogs (and by association Coca Cola Femsa, and Trip Advisor)

A few quick thoughts on the week ended Dec 23, 2016 and on Mainstream Financial Media (take your pick, BNN, CNBC, etc.).

For the most part, BNN, CNBC, etc., usually aren't a source of anything newsworthy. I say usually, because on the odd occasion, these networks will have great guests on with compelling viewpoints, but I'd peg these occasions as infrequent, especially in Canada where we get to enjoy BNN, rife with regurgitated news, entire broadcasts devoted to marijuana grow-ops posing as going concerns, and "unbiased experts" pitching their books in the guise of top picks nightly (in their defense, if they didn't pitch their books to BNN viewers, they'd have nobody to unload positions onto).

My daily BNN/CNBC routine is as follows:

I check in at 7 AM to see if I've missed any headlines from the previous evening and then I switch off and watch the sportscentre highlights which are far more interesting. I also watch market call while working out in the evenings in order to draw inspiration for fighting through my workouts. I find that market call provides great fuel for pushing out extra reps, especially when the discussions are exceptionally stupid.

This week, CNBC was falling over itself on two themes. 1) Dow 20,000, and 2) Trump's Tweets.

First off, Dow 20,000

I'm probably not the first participant to state that Dow 20,000 is irrelevant, and I won't be the last. But to me, the Dow is entirely irrelevant. Why? Well, firstly, the Dow is one of the financial world's most closely watched barometers, and because of this, I find the barometer itself irrelevant. By virtue of everyone focusing on the Dow, I choose not to. Secondly, 20,000 means nothing. It's a price level derivation based on the price weightings of the underlying constituent companies. Here's the calculation courtesy of wikipedia:


So here we have an index construct moving higher (or lower) based entirely on price weightings of the constituent companies. For reference, here are the top price weighted companies in the Dow as of Friday Dec 23rd, courtesy of

Based on the above, it's not difficult to understand why the Dow has rallied 11% since Trump won the election on Nov 9. Goldman Sachs alone has rallied close to 40% in less than a month!

Being the cynical participant that I am, I've started to wonder about the fundamental underpinnings driving this move. What has changed so drastically as to support a 40% move in Goldman Sachs? The answer is something along the lines of: everything and nothing at the same time.

From my readings thus far, the move has been in anticipation of relaxation of financial regulatory and accompanying constraints and the potential repeal of the Frank-Dodd Act as a result of Trump's ascension to office. So, devoting an entire broadcast week (or month, or quarter) fixating on whether GS will rally another 10%-20% in order to move the Dow another 100 points is idiotic to say the least. I'd hazard a guess that GS is more likely to give up some portion of the 40% move in the next year vs. continuing up ad infinitum, and any pressure on GS will translate into pressure on JPM (and other financials) as well. These two constituents represented around an 11% weighting in the Dow as of Friday's close, so if Dow 20,000 indeed happens, it will happen when no one is watching (or cheering for it).

This brings me to my next point on Trumps Tweets

In my opinion, we are faced with an increasingly risky investing environment going into 2017. While the markets have vaulted to all time highs, the vault seems to have been built on anticipation of policy change due to a changing of the guard in the oval office. And the scary thing here is, this has become pervasively accepted investment rhetoric as participants seem to now all agree that Trump = good. When everyone seems to be lined up on the same side of the ledger, well, only good things can happen, right? My guess is that as 2017 unfolds, maybe, just maybe, some discernible level of doubt will rear its ugly hide and work to unwind some (or all) of the initial jump in optimism based on, well nothing (yet). I'm going to draw from Seth Klarman's 2015 annual Baupost letter as Mr. Klarman gives voice to short term exuberance much better than I can:

"Value investors must be strong and resilient, as well as independent-minded and sometimes contrary. You don’t become a value investor for the group hugs. Indeed, one can go long stretches of time with no positive reinforcement whatsoever. Unlike some other fields of endeavor, in investing you can do the same thing as yesterday but achieve completely different reported results. In the long run, the research and analysis you perform should overcome market forces; the fundamentals ultimately matter. But in the short run, markets can trump effort and insight. They move in unpredictable cycles, with investors stampeding this way and that. Businesses quickly come in and out of favor, and the same business can be valued by the market very differently in a matter of days, sometimes on the basis of new facts, but often because of mercurial investor perceptions or simply money flows"

This week, Trump tweeted about Lockheed Martin's F-35 fighter being too expensive which, according to the MSFM, "roiled" LMT. The stock was down close to 3% in the premarket and CNBC had a piece on questioning whether this was a buyable dip.

Well, here's LMT monthly as of Friday's close. Is it time to buy the dip yet?

In my opinion, the fact that the most influential politician on the planet resorts to tweeting is ridiculous, and my prediction for 2017 is that Trump will continue tweeting, and will rattle markets horrendously at some point during 2017 by being a moron on twitter. Maybe a Trump tweet will be stupid enough as to actually cause a real buyable dip (as opposed to a CNBC or BNN broadcasted buyable dip)!

Globe and Mail 2016 Dogs

Well, we got our 2016 dogs courtesy of the Globe and Mail this weekend. Included in the dogs are two very interesting ideas which the Globe (and by extension everyone else) seems to have written off, link here

First up Coca Cola Femsa (by extension from the Globe's write-up on the Mexican peso).

I'll start with a monthly chart and elaborate on the story behind it:

Why KOF? Well, firstly, the Globe profiled the Mexican peso as a casualty of Trump's ascension to office, and as the Mexican peso goes, so too, by logical extension, will the Mexican economy go, maybe. Funnily enough though, balance of trade has swung to a surplus in Mexico recently as a lower peso has resulted in higher exports to the US, go figure. Now, recently is way too short a time period to draw meaningful conclusions, and a repeal of or significant modification to NAFTA could materially impact future trade flows. There is huge uncertainty here, but the largest casualties of the threats thus far have been anything Mexican (from the peso to Mexican Bolsa constituent companies) dependent on continued unmodified free trade with the US. KOF is impacted by virtue of its exposure to purchases in US dollars. As the peso weakens, it gets more costly for KOF to purchase formula from Coca Cola under the terms of its existing bottler agreements. KOF has also taken on a significant amount of US dollar denominated debt in order to build additional bottling capacity over the last three years. I've plotted a correlation study between KOF and EWW over time, and we can see that KOF pretty closely tracks the performance of the Mexican stock market in general.

The story behind the last three years of abysmal performance in KOF is a function of the imposition of a soda tax in 2014, which hammered Coca Cola Femsa's sales. Prior to the imposition of the soda tax, KOF was firing on all cylinders. I have no idea why it rallied at a 51% CAGR clip between 2009 & 2013, but I'm going to guess that highly mobile capital flows looking to deploy capital into levered bets out of the 08/09 crisis chased KOF up to unsustainable levels predicated on higher estimated per capita consumption of soft drinks in Mexico (and the rest of Latin America), extrapolated forever (this type of logic couldn't possibly apply to capital chasing Goldman Sachs +40% now could it?).

The following infographic courtesy of Statista, link here, details growth in per capita consumption of soft drinks by country between 1991 and 2012 (just before the imposition of the soda tax):

Guess what! Mexico, Panama, and Argentina round out 3 of the top 5 countries on the list, they all ranked ahead of the US with the exception of Argentina, and all three of these countries are major stomping grounds for KOF. What looked like permanent strength in trend (price trend, not valuation trend) between 2010 and 2013 has given way to an economic reality adjustment over the ensuing three year period since the soda tax was introduced, and with reference to the company's most recent annual report, the soda tax had a material impact on the company in terms of volumes, especially in the initial years after introduction. And if a soda tax wasn't bad enough, now comes the depreciation in the peso vs. the USD! Talk about the consensus paying $160 per share back in 2013 and getting it entirely wrong!

Are we at the point of maximum pessimism yet? No idea, but I do note the following:
  • The universe of $160 per share cheerleaders through soda consumption rhetoric extrapolaters are all likely gone by now (good riddance)
  • The current investing paradigm surrounding Mexico is now bearish
  • There is huge uncertainty surrounding what changes to NAFTA Trump may make
In support of KOF, I performed my own attempt at analyzing the company and performed my own valuation, and I note the following:
  • Notwithstanding the +60% decline in the price of the stock since 2013, the company has a moat and a potential runway by way of future Latam soft drink consumption. While the imposition of a soda tax hit volumes in the first years after introduction, interestingly, consumption seems to have adjusted from the initial reaction, per WSJ, link here:

  • Despite the drop in sales, KOF seems to have been able to maintain operating margins at 15%. No easy task given the absolute % drop in sales since 2013, per below:

While sales are off -25% since the 2013 peak (just before the introduction of the soda tax), EBIT margins has remained constant at around 15%. Taking a closer look, gross margins have actually improved slightly vs. 2012/2013 (on a tttm basis), and have remained constant at around 53%-54%, so I take this to mean that even in the face of the soda tax, the company has maintained pricing power.

And the valuation metrics haven't been this low for some time (ignoring 2008 & 2012 as outliers):

For comparison's sake, here's Coke itself:

Caveat, this is a bit of an apples to oranges comparison because KOF is a bottler and KO sells the syrup, so very different business models with very different capital requirement constraints. Even so, whether KOF deserves a multiple 33% lower than KO on a ttm PE basis, 44% lower than KO on a ttm EV/EBIT basis, or 48% lower than KO on a ttm EV/EBITDA basis, is up to the individual investor to decide.

A more suitable comparison would be to compare KOF to other regional botllers, such as CCE (Coca Cola European Partners), reproduced below, courtesy of gurufocus:

While I don't know where KOF may bottom (somewhere between $63 and $0 in theory), on a free cash flow basis, I get the following matrix of potential valuations at different growth rates:

Arguably, if KOF got to $45 from here, and assuming 2.5% growth, we'd be looking at a very interesting situation, based on the following:


  • Soda tax doubles from here, further impairing sales
  • Peso continues to drop vs. the USD, impacting future margins
  • KOF can't renew it's license agreements with Coca Cola on favourable terms, impacting future margins
  • Material changes are made to NAFTA which could affect KOF's future purchase supply chain 

Finally Trip Advisor

Courtesy of the Globe and Mail, here's this week's write up on poor old TRIP, categorized as one of 2016's dogs:

Superficial write-up in my opinion, and my questions are as follows:
  • Was valuation really relatively high?
  • Why did marketing costs soar?
  • Why wasn't TRIP able to capture market share to competing alternatives?
  • Does TRIP have a moat?
Was valuation really relatively high?

Yes, and no. In order to answer this, I have to look at closest comparables. When I think of Trip Advisor, I think of online travel, and closest comparables might be Priceline, Expedia, and to a lesser extent, Travelzoo and C-Trip. But there are differences in the models between the comparables. Priceline & Expedia are OTA's (online travel agencies), and are customers of Trip Advisor. TRIP earns revenue from CPC's by virtue of referring (directing) traffic to OTA's and other direct customers (hotels).

Superficially, TRIP entered 2016 at some pretty lofty multiples relative to the above group. 34x EV/EBITDA and 47x EV/EBIT, based on 12/31/15 reported #'s.

This compares to 18x EV/EBITDA and 20x EV/EBIT for Priceline and 12x EV/EBITDA and 18x EV/EBIT for Expedia, based on 12/31/15 reported #'s.

So we know that valuation looked relatively, high, but the next question is why? Why were participants content to own TRIP at 34x EV/EBITDA and 47x EV/EBIT up to the end of 2015, and how have these multiples adjusted during 2016?

My first observation regarding 2015 is that included in operating expenses was a one time donation of TRIP shares to the Trip Advisor Charitable Foundation of $67M. Normalizing 2015 EBIT for this donation, I get EBIT of $300M. This puts 2015 adjusted EV/EBIT at 39x EV/EBIT vs. 47x as originally thought. Per review of the company's 2015 10K, the contribution was made in Q4 2015, so the ttm numbers also reflect the Q4 contribution.

Currently, TRIP can be had for 26x ttm EV/EBITDA and 46x EV/EBIT. While EV has dropped by close to 50% since the end of 2015, EBIT hasn't improved much if at all, so on a superficial basis, TRIP is still expensive despite a 50% drop in pps. EBIT dropped from $232M at the end of 2015 to $131M on a ttm basis, but adding back the $67M Q4 charitable contribution, normalized EBIT on a ttm basis should be closer to $198M, and normalized EV/EBIT should be closer to 31x. A little better than initially thought.

This leads into the next point:

Why did marketing costs soar and why wasn't TRIP able to capture market share to competing alternatives?

Superficially, marketing costs soared as a result (or consequence) of erosion in CPC's in the company's hotel user base. This makes sense on the surface, as average # of hotel users dropped and CPC per hotel user dropped, the company theoretically played defense in order to win back customers by virtue of ramping up marketing spend, logical right?

Well there's more. Prior to 2014, the company identified a couple of problems with its business model, which were 1) failure to monetize (convert) unique site visitors into clicks, and 2) leakage. Leakage was a direct consequence of 1), and as a result, the company began spending heavily on transitioning its old user interface model with mutiple pop ups redirecting visitors to non TRIP / OTA sites, to an Instant Booking ("IB") model which allows visitors to book hotel travel directly within TRIP's interface without leaving the site.

The IB roll out commenced in 2014 and according to management's most recent quarterly 10Q, is now complete, so any analysis of prospective results must look to what IB may bring as opposed to focusing on the transitional time period during between 2014 and now. Management clearly explains the risks surrounding IB integration in its 2015 10K, including IB not resulting in closing the monetization gap, and/or OTA partners not buying into the new IB model.

There is a fantastic article on the potential of IB over at seeking alpha recently, link here, which outlines the case for IB in great detail. The author demonstrates that over the last few years, revenue per hotel shopper and EBITDA margin per hotel shopper have declined from close to $.60 per hotel shopper in 2014 to $.45 and from 42.7% per hotel shopper in 2014 to 32.4% most recently. The author argues that these declines are due in part to lower conversion rates on IB vs. metasearch, and low initial commission rates offered to OTA's as an inducement to adopt IB. The author makes a convincing argument that the pressure on margins is likely temporary, and goes on to explain that over time, IB adoption and conversion could very well eclipse metasearch conversion on a CPC per hotel user basis, resulting in CPC per hotel user getting back to pre 2014 levels. The missing clues were commitments from large OTA's, and interestingly enough, Priceline and Expedia have now signed agreements with TRIP under the new IB model.

Hypothetically speaking, what would TRIP's valuation look like in 2017/2018 if IB adoption results in CPC per hotel user returning to $.56 per visitor? For argument's sake, I modelled using EBIT margins at 30% as opposed to using EBITDA, despite the company being capital light. EBIT margins reflect operating income after stock based comp, R&D / marketing spend (which should drop post IB rollout) and non cash charges which I'm going to assume will approximate actual spend on ongoing capx / intangible.  Here are the results:

Assuming $.56 per hotel user, x 1.6B (ttm) annual hotel visitors (source: Trip Advisor Q2 2016 supplemental financials, link here) works out to $879M in IB hotel revenue. I modelled display based and other hotel revenue using TTM 2016 numbers, and used ttm non-hotel revenue of $338M.  The real key here is whether EBIT margins per hotel user will revert back to pre-IB rollout. Assuming 30% EBIT margins works out to an overall $393M EBIT contribution from hotel (net of negative contribution from non-hotel for now). The key to this assumption is the gearing in converting top line revenue into EBIT. The company seems to be a capital light compounder, so assuming cessation of 2014-2016 spend on IB going forward, any uptick as a result of conversion should flow directly to the bottom line.

Based on the above scenario, I get a current P/E of 25x and EV/EBIT of 15x, a lot different than conventional superficial analysis would have us believe. But that's our job isn't it? Looking beyond superficial summary reporting and asking how/what/why/where! If it wasn't for widespread groupthink like that published in this week's Globe, investors wouldn't have any work to do! This leads to my last question:

Does TRIP have a moat?

I'd say yes, but with a caveat. Borrowing from John Huber's writings on ROIC, TRIP seems to have characteristics of a two sided network. It is an accepted central hub for community ratings globally, but the problem facing the company has been converting these unique community visits into money! Therein lies the $1M question, is TRIP currently in the throes of a gradual move towards obsolescence as evidenced by falling revenue and EBITDA per user, or will it reinvent itself by virtue of rolling out IB? Only time will tell.

My bet is that IB will work.