On Diversification

Diversification is the buzz word in investments. We see some funds that hold hundreds of stocks and others (especially value-oriented funds) only a few. Before we delve into the reasons let’s examine diversification and volatility in more detail.

The common misconception is that diversification is always good. It is not. I would argue that for most people it would only improve their portfolio if it doesn’t substantially decrease returns. It is much better to hold 10 stocks that could yield 100% rather than 100 stocks that could yield 10%. On the other hand, it would make sense to hold 100 stocks that could yield 99%  rather than 10 that could yield 100%.

If we look at volatility of the two potential portfolios (100 vs 10), it should be no surprise that the 10 stocks portfolio will be much more volatile (assuming similar value drivers). Volatility is the product of simple statistics. SE(volatility) of the mean(asset price)  decreases as the number of trials increase. Even if the true distribution of returns of the 100 stocks is the same as the 10, we may still observe higher volatility in the 10 stocks portfolio simply because of lower number of “trials” (stock holding).

If we assume all stocks earn similar returns (ie market is efficient) then we have a very compelling reason to diversify. But things get a bit more interesting when we introduce imbalances. What would happen if you could earn more return on the 10 stocks than the 100 stocks? To simplify the concept we can use coin flips as an analogy. Suppose we have $100, would it be wise to bet on heads A. once for all our money  for 10/1; or B. 100 times $1 bets for 2/1? The expected profit is 450 and 100, respectively. But the right choice isn’t so clear. The variance (volatility) is a lot higher for game A. You have a 50% chance of going bankrupt. Game B is relatively safe, but the return is mediocre. So we arrive at this trade off.

What’s great about the financial markets is that we can essentially play the above game repeatedly and we can adjust bet size accordingly. Does it make sense to split your money evenly between the two games? Surely not. You would only play game A, but bet only a portion of your bankroll every time. I hope we can all agree that no one should be playing game B.

Why doesn’t everyone play game A? Because not everyone knows where to play. Not everyone can correctly identify stocks that have significantly asymmetrical return profiles. EMH proponents think that no one can correctly identify such opportunities consistently. EMH is a whole another topic that we can explore at another time. But I do find it rather amusing that stock picking is so easily dismissed when there is no rigorous proof. And there can be none. No one can prove that every asset in the world is fairly priced.

And we are back at where we started. So why do value-oriented funds hold concentrated positions? Size aside, they also only want to pursue the best opportunities. Think about it this way, at any given time, there will only be 10 best ideas. And since returns get progressively worse as you move down the rank, why bother? One would only need to diversify enough to mitigate the risk of ruin. Using the coin analogy, it would be silly to play game A with all of your money only once. You could however profit handsomely if you play the game 10 times with $10 each time. The chance of you losing all 10 games (akin to being wrong with all of your investment ideas) is 0.09%. Maybe this isn’t enough for some people, and so people can hold 15, 20, or even 30 stocks. But the incremental “risk” that you mitigate will be negligible beyond a certain point.

MagicJack Opportunity

Summary

• 60% of value is backed by cash with no debt.
• Core customers will continue to provide stable cash flow.
• Optionality value in Mexican expansion.

Full writeup

Why I pick under-performing companies

One man’s trash another man’s treasure. The saying has never been more true to me. When one rejects the infallibility of the market, then it becomes quite clear why under-performers are sometimes the best bargains.

These securities are often “dirty” and undesirable, for whatever reason. Often negative news obfuscate their true value. A simple example. Imagine that you dropped a pound of gold in a bucket of manure when visiting Auntie’s farm. Would you retrieve it? Of course you would! Yet when securities are thrown away in the real world, seldom do people have the desire to pick them up. When you truly understand the value of the security, then you would be able to see past its undesirable characteristics.

Take the gold bar example. Many investors would only acknowledge the fact that the gold bar is covered in excrement, but not the fact that it is still gold. Clearly this presents an opportunity to those who recognize the true value of the gold bar. Often the bucket of manure is sold as if there is no gold, and it is up to you to uncover its true value.

Active Short Selling

We’ve all heard of activists. But can we apply the same principal to the short side?

Some times the company may be overvalued, but for whatever reason, the market is just not budging. Perhaps hedge funds who hold short positions should take matter into their own hands. Here are a couple scenarios.

Suppose Company A competes with Company B and produce similar products. Company B seems to be financially vulnerable to an economic meltdown, but betting on that seems to be a long shot. What we can do instead is artificially create a crisis for Company B. We can strike a deal with Company A to significantly undercut Company B and subsidize the costs, while at the same time purchasing derivatives on Company B to create synthetic short positions (buy put options, CDS, etc.). In all likelihood, Company B goes bust and the fund profits.

Another scenario. Company A is financially distressed, but bankruptcy is unlikely due to a potential restructuring. CDS premium is low as a result. We can take advantage of this situation by purchasing a strategic block of debt while simultaneously purchasing the CDS. We can then block the restructuring and profit on the CDS.

*Disclaimer: I am not a legal expert.

Automodular Corporation Opportunity

Summary

• Cash balance is greater than current market cap.
• The company has no debt.
• Litigation and liquidation gains may boost value.

This company is not exactly an “investment” in the traditional sense. While equity value is fully backed by cash, a poor acquisition could still derail the stock. However, I believe that more things can go right than wrong. Just about any announcement short of “We donated the cash” would be interpreted as a positive signal for the company.

Full write-up: http://seekingalpha.com/article/2898726-automodular-corporation-opportunity

How not to use backtests

I recently saw a video of a young fellow showing his backtesting results on Dragons Den and somehow the Dragons were very interested…

I think there are some general misunderstanding on how data from backtests should be used.

I reckon a lot of people would be surprised if I proclaim that I can find a winning strategy given ANY data. As Kevin O’Leary said in the episode, “In 2008, would we have been wiped out on the equity?”

Of course not Kevin, the time has passed already and the kid wouldn’t have brought this chart with him if the strategy lost money.  When you are fitting a model over historical data, there are inherent assumptions that you make, but their existence is outside of the model.

Take this example. A child sees a tiny chick grow up. Day by day the chick grows bigger. One day at the dinner table the child proclaims that it will grow forever! Imagine his disappointment when he learned where that delicious chicken soup came from.

(Child’s model: Xt = Xt-1 + 1   ;       Actual model: Xt = Xt-1 + 1            X<big enough to cook)

That is precisely problem. The child never knew that the chick would be made into soup. In the financial markets no one can foresee these “black swan” events solely through backtesting. That is because they have never happened before! Of course it’s easy to come up with a new idea when the crisis has passed. Now that unfortunate event becomes a part of your model, just a data point, a point around which your model will tweak to make sure that you don’t lose money. And of course everyone should have seen that coming, who would have put all of their money in 2007? The signs were all there!

Until next time.

Black-Scholes Model for American Options

A key input of the model is volatility. Of course since no one really knows the exact volatility in the future we estimate it using historical data. This is the root of the problem Perhaps the day to day fluctuation can be reasonably modeled, but to suggest that one can accurately predict volatility at all times (even months/years into the future) is completely absurd. This is where opportunities may arise and those with a deeper understanding of the company will have an advantage.

Does a DCF model almost always overvalue a company?

A key assumption in the DCF model is that the FCFF will compound. But how can that be if not all FCFF is paid out (ie. sits on the balance sheet)? In a typical DCF we have years of cash flows and then we discount them back, let’s examine the mechanics.

Suppose a company has no debt and generates $100 in FCFE into perpetuity but does NOT pay dividends. Using a discount rate of 10% we get a company value of $1000. Obviously there are nuances that affect valuation (replacement value, liquidation value, etc.), but if we focus just on the cash flows we see a very significant flaw.

Since the cash merely piles up on the balance sheet it does not generate any return, hence the compounded growth of 10% cannot be achieved. The value of $1000 is the upper limit of valuation, which can be achieved if the management actively uses the FCFE to enhance shareholder value (reinvest into projects that can generate more than 10% CAGR, share repurchase, etc.) The actual return will decrease with every passing year in which the company does not pay dividends. For example, the company would have $300 in cash at the end of year 3, suppose the market is efficient and the company is still worth $1000. If you discount the cash flows we get: (1000+300)/1.1^3 = 976. A difference of 2.4%!

Does this technicality actually matter? I don’t think that this would a number one factor when making an investment decision, as there are many other more important considerations. I believe that this finding would impact the valuation of certain tech stocks. They seldom pay dividends or buyback shares, and most of the cash either sits on the balance sheet (eg. Apple) or is used on acquisitions. Whatever discount rate you pick will effectively become the hurdle rate for the said company. Unless acquisitions and cash can generate a return equal to your discount rate, a DCF model will always overvalue the company,

 

CONN research

• High Growth Potential.
• Bad debts expense can be easily controlled.
• Management will unlock shareholder value following the strategic review.

Full paper available at Seeking Alpha Pro.

The Euphoria Surrounding Uber Will be Short Lived

Uber, the ridesharing company that is taking over the taxi industry by storm, is all over the headlines these days. Founded in 2009, the company has since expanded to 36 countries and is doubling revenue every 6 months.[1] Backed by influential investors including Google Ventures, Goldman Sachs, and Jeff Bezos, Uber completed its latest round of funding in June with a valuation of $17 billion.

The taxi industry is usually highly regulated, restricting the number of taxis operable at any given time. Given Uber’s similarity to a standard taxi, it should be no surprise that the company is the target of numerous lawsuits. Demonstrations have been held and Ubers cars have even been attacked by taxi drivers during protests.[2] Legal issues aside, the company is welcomed by both taxi riders and Uber drivers thus far.

So why is Uber loved by most people? Along with cheaper fares, Uber also offers a review system where its drivers are rated according to customer satisfaction. In addition, Uber claimed that Uber drivers earn significantly higher than regular taxi drivers.[3] Combine all of these factors and you get a company that is seemingly working to better the taxi industry for both users and drivers.

But will the future be what its proponents think it will be? Will Uber continue to deliver? The answer is no. Uber’s high margins, increased revenue for drivers, and cheaper fares are mutually exclusive and will be short lived.

Uber’s rapid growth is the result of integrating GPS into its fleet, thereby increasing efficiency, and removing the cap on the supply of drivers. The bulk of economic value that Uber contributes to the industry is the increase in utilization (increase in productivity). As a result, we see a rise in average revenue per driver (ARPD) as waiting times are reduced. Uber then passes some of the savings resulting from this increase in efficiency to the consumers, decreasing the fare. This creates the illusion of a virtuous cycle, as illustrated below by Mr. David Sacks.[4]

 

Before we delve further, I would like to make two reasonable assumptions about the industry.

1. There is a large supply of drivers
2. Ridesharing companies will compete in the same city (we are seeing this already)[5]

As prices decrease, the optimum will eventually be reached, and further decreases in the price would be uneconomical. However, Uber will still be pressured to recruit more drivers when it reaches that point. Why? If I may borrow Mr. Bill Gurley’s (an Uber investor) article: the network effect is the name of the game. It improves pick-up times, increases coverage density, and ultimately increases utilization. So how can Uber increase its network effect? One key variable is the number of customers (demand), and since price is fixed at the optimum, the only other input would be the number of drivers.

Unfortunately for Uber, the low barriers to entry implies fierce competition, we are already seeing a multitude of ridesharing companies: Lyft, Sidecar, Summon, just to name a few. Due to competition, every company has an incentive to recruit more drivers to increase its own network effect. This is the beginning of a vicious cycle.

Each company will face constant pressure to recruit more drivers. In doing so, the overall utilization and ARPD of the industry will decrease. As more drivers sign up for Uber, the inevitable equilibrium is one where the ARPD will be at its minimum. As drivers’ revenue decreases, riders will also suffer from poor service.

Can’t Uber just fend off competition with its established fleet already? It is possible if Uber can somehow establish brand loyalty. However, it is highly unlikely considering that its service acts like a commodity, much like the commercial airline industry.

 

The $17 Billion Valuation

People rave about how fast Uber is growing, how Uber is creating jobs, and how Uber will expand into alternative markets. The exact market share that Uber will have is highly debatable. Mr. Aswath Damodaran, a renowned professor at NYU Stern, recently wrote an article questioning the potential market for Uber. Mr. Bill Gurley then wrote the aforementioned article in response. While the total market size is important, I believe that Uber’s operation should be scrutinized to determine its true growth potential and profitability in the long run.

Uber essentially places a GPS on a car, and voila, its owner becomes an Uber driver. The information from the GPS is then relayed to the rider’s app, where the rider can hail an Uber car. During the ride, the GPS is again used to calculate the total fare, set by Uber.

First of all, the technology used isn’t new. Without patents to protect itself, how can Uber sustain the current margin? Secondly, Uber’s continued operation is controversial because it violates the intended goals of taxi regulations around the world. In fact, the only reason Uber expanded so quickly was because its deep pockets allowed it to waltz into any city and start operating, regardless of any laws. When being investigated, Uber can throw cash at lawyers until a favourable verdict is reached.[6] This isn’t exactly what I would call good business practice.

We established previously that Uber will continue to face hiring pressure. Well that can’t go on forever, because the ARPD will decrease to the point where no driver will work for you. Uber cannot increase the price either because that would decrease the total revenue (deviation from the optimum). Uber can try to remedy this problem only by decreasing its own cut of the fare. Thus we can see that the current commission of 20% will not be feasible in the future. In fact, the price war between ridesharing companies has already prompted Uber to take a loss on rides following this update.

But is Uber worth $17B? On one hand, Uber is still bringing in huge amounts of cash at little to no variable cost. Without any need for capex, Uber may continue to expand at a rapid pace. However, the same holds true for other ridesharing companies. In the long run, ride sharing companies will operate with razor thin margins, as each company struggles to keep fares low and drivers happy. In addition, the fact that governments around the world are starting to ban ridesharing companies isn’t helping Uber’s case either. [7]^,[8],[9]

I do not know how the VCs valued Uber, but I believe that any valuation for Uber should place heavy emphasis on cash flows in the coming years. If the $17 Billion valuation is based on Uber’s cash flows down the line while assuming the current margin and growth rate, then investors should be in for a tough ride.

 

 

[1] http://blogs.wsj.com/digits/2014/06/06/uber-ceo-travis-kalanick-were-doubling-revenue-every-six-months/

[2] http://techcrunch.com/2014/01/13/an-uber-car-was-attacked-near-paris-as-taxi-drivers-protest-against-urban-transportation-startups/

[3] http://abovethecrowd.com/wp-content/uploads/2014/07/Screen-Shot-2014-07-07-at-6.41.45-PM.png

[4] https://twitter.com/DavidSacks/status/475073311383105536

[5] http://www.forbes.com/sites/quora/2014/06/10/whos-winning-right-now-in-the-competition-between-lyft-and-uber/

[6] http://techcrunch.com/2013/01/31/a-day-after-cutting-a-deal-with-lyft-california-regulator-reaches-an-agreement-with-uber-as-well/

[7] http://mashable.com/2014/07/21/seoul-korea-bans-uber/

[8] http://www.telegraph.co.uk/news/worldnews/europe/germany/10991089/Hamburg-breaks-ranks-and-bans-Uber-app.html

[9] http://www.engadget.com/2014/04/15/belgian-uber-ban-10k-fines/