Friday, July 29, 2016

The Case for Hedge Funds / Creating an Ideal Liquid Alt

A hedge fund is simply a go anywhere investment vehicle that attempts to provide excess returns to cash with a low correlation to traditional asset classes (i.e. they are vehicles that attempt to provide alpha). Hedge funds and liquid alternatives have taken a lot of heat recently, much of it deserved, but in this post I'll outline the case they are being incorrectly evaluated. Specifically, this post will outline the benefit of a hedge fund that can provide excess performance to cash with low correlation to stocks / bonds even if it provides only a minimal level of excess return. In addition, I will share how an investor can effectively utilize a hedge fund (or liquid alternative) within a broader portfolio, which will touch upon why hedge funds should rarely be judged by their level of absolute performance.  

Why Have Even the Best Hedge Funds Disappointed Investors?

The below equity curve is for the actual returns of a hedge fund, which I'll label Hedge Fund X, that I pulled specifically because the since inception returns have:
  • Consistently been uncorrelated with those of stocks (0.13) and bonds (-0.10)
  • Been in excess of cash
  • Underperformed a 60/40 portfolio

In other words, Hedge Fund X has had very strong risk-adjusted performance, yet has provided only average absolute returns when compared with stocks and/or bonds. As a result, an investor that reallocated from stocks and bonds to fund the allocation to Hedge Fund X wouldn't have accomplished much. While the allocation did improve risk-adjusted returns of the overall portfolio, it came at a lower overall return and (as the chart shows below) it hardly moved the needle in terms of the return path. Given the amount of incremental due diligence required to make the allocation and the high fees paid (which can now make headlines for public investors), there may be buyer's remorse for the allocation (this despite the luck that was likely involved in finding a hedge fund that was able to produce such remarkable risk-adjusted performance ex-post).

The Case for Leverage: Hedge Funds as an Alpha Overlay

Rather than carving out an allocation to a hedge fund from stocks and/or bonds, an investor can make an allocation while maintaining their broader asset allocation through the use of leverage. The resulting economic exposure can be viewed as a traditional beta portfolio with the hedge fund as its alpha source. The structure could be created a number of ways, including selling out of stocks and overlaying stock futures over the existing bond portfolio (an example of a mutual fund that does that here). The end result is a strategy that provides an investor with a fully invested exposure to a stock / bond allocation using only 50 cents on the dollar, freeing up proceeds for Hedge Fund X with the remaining 50 cents.

Using a hedge fund as an alpha source is typically performance enhancing as long as the hedge fund outperforms the investors borrowing cost (i.e. cash rate, which is currently ~0%) net of fees. This may provide an investor the ability to reduce the risk of the underlying beta portfolio without reducing the expected return of the overall portfolio. In the example below, the stock allocation was reduced to a "less risky" 50% allocation, yet the combined portfolio was still able to outperform a 60/40 blend on both a risk-adjusted and absolute return basis due to the incremental return provided by the Hedge Fund X.

A performance comparison since the financial crisis may be more interesting as this has been a period in which the absolute return of Hedge Fund X has materially lagged. This has been driven by the strong performance of stocks and bonds (allocating away would have been a huge opportunity cost in hindsight) and due to the low level of yield on cash (see here for more details as to why that's a drag). Since the March 2009 bottom, Hedge Fund X returned only 4.4% annualized (underperforming the 18% return for the S&P 500 and the 4.8% return for aggregate bonds), yet the allocation utilizing Hedge Fund X as an alpha overlay continued to add value in the form of higher returns and lower risk when utilized with a 50/50 beta portfolio. 

Hoping for Innovation in Liquid Alternative Space

As outlined above, the real benefit of using uncorrelated return streams often comes if they are utilized as an alpha enhancer via an overlay, especially true in the current environment as financing costs have a low hurdle rate of ~0%. This benefit becomes a challenge for liquid alternatives that are structured as "cash plus" investment vehicles as record low cash rates are a drag to performance (rather than a low hurdle). As a result, my hope is the industry moves back towards an alpha / beta model that was popular pre-crisis where investors can gain access to cheap beta within the fund itself. Better yet, perhaps as an alpha overlay over a dynamic beta structure that has a historical track record of outperformance.

Friday, July 22, 2016

What Drives Momentum Performance?

Mar Vista Investment Partners has a really interesting research piece out The Price You Pay which has a great table outlining the benefit of an asymmetric return profile (i.e. having more market exposure during up markets than down markets).

It is a mathematical truism that superior down capture in negative periods provides more capital for compounding in the ensuing positive periods. Using S&P 500® Index monthly total return data for the last thirty years, the chart below demonstrates the expansive value created by preserving capital in the down periods even with subpar returns in the positive periods. Each column shows the ending amount of capital with $100,000 invested in the S&P 500® Index over thirty and ten years with various combinations of monthly up and down capture.

As shown in the table above, even a 10% difference in up / down capture can provide a material impact to returns and the path of returns.

This also happens to explain why momentum works... with a traditional momentum model where you are in either stocks or cash, you will rarely capture 100% of the market's upside (the strategy is not always going to be in stocks when the market is up), but you will by the same math almost always improve the downside capture. The chart below shows the rolling three-year upside and downside capture figures for a basic momentum model and shows momentum often, but not always, provides favorably asymmetry (i.e. upside capture > downside capture). Over the longer-term (50 years), the upside has been 70% vs downside of 56%.

It is the low downside capture during market sell-offs that has driven the strong relative and absolute long-term performance of this momentum strategy. Again... while there have been many periods in which this momentum strategy has underperformed, over the longer-term it has slightly outperformed the S&P 500 since 1966 (10.2% vs. 9.7%) and with a much lower standard deviation (12.7% vs. 18.9%). The chart below shows the tight relationship between three-year excess performance and three-year upside minus downside capture, which points to why momentum strategies have largely underperformed since the 2008-09 correction when markets have generally moved higher.

The momentum investment opportunity becomes even more interesting if you can pair the above return stream with a strategy that may improve up capture during positive market environments, but I'll save that for another day.

Tuesday, July 19, 2016

Why Does Crime Feel Exponentially Higher, When It's Materially Lower?

Every day we hear about some new horrific event taking place globally. The common view is that crime is rampant and the United States is headed in the wrong direction. But that view is not supported by facts. In a Citylab article related to the recent level of crime, it was noted:

"The story is actually better than we all anticipated it would be," says John Roman, a senior fellow at the Urban Institute's Justice Policy Center. "Violence is down a little bit. Property crime is down a lot… and all of this suggests that crime in America is continuing to move in the right direction.”
Over the longer-term, the decline in violent crime is even more notable. Despite the United States 21% population increase since 1995, violent crimes were down -35%, translating into a decline in violent crime per capita of almost 50%. Rape, murder, and robbery were all down over this period in absolute terms and down a lot in per capita terms.

Source: FBI

Why Does it Feel Like Crime is Getting Worse?

My theory is the perceived increase is related to the way information flow has made the world a much smaller / more local place.

One exaggerated example: if 150 years ago you lived in a 500 person town and your local newspaper maybe covered news from a 10 town population of 5,000, the sample size for your local crimes was that 5,000 person population. With a violent crime rate of 1000 per 100,000 people, you might hear about 50 crimes per year (or about 1 a week). If the crime rate were to move lower, given the same population, the decline would be felt via the reduced number of headlines in the local paper (see example A below). However, if a new paper came to town sharing news from a much larger base population, despite a material decline in overall crime, the number of crimes being reported might spike to 1 per day (outlined in example B).

Our Local Community has Grown Exponentially

Recently, we haven't just experienced a small increase to the size of our local population (from a town to neighboring towns, to our state, or even to our country), but the entire world (or the entire universe if you really want to freak yourself out). In addition, when news flowed through print, it was limited by what could be reported by space (only so much paper) and timing (only news could be reported as of a certain time each day before it became old news). Now, the cycle is infinitely larger and continuous.

If I was scared of being attacked by a homeless man, there is news for that. If I was scared my children would be abducted, there is news for that. Hell... if I were scared of the dangers of Pokemon Go, there is plenty of news for that. When the world becomes your local neighborhood, a 50% decline in the violent crime rate over the last two decades still means there are tens or hundreds of thousands more opportunities each day to freak out (note the y-axis in the chart below is in log form because of the exponential nature).

This won't stop any time soon. People have evolved over thousands of years to seek out scary news and freak out about things we perceive can harm us. Combine this evolutionary characteristic with the for profit nature of news organizations (fear sells) and strawmen politicians build utilizing some randomly occuring event as "proof" it happens all the time and you have a recipe for disaster.

The key is to remember that if properly weighted, only a tiny fraction of news would be bad news. Just imagine the sentiment if it that was how the world was accurately perceived.

Friday, July 15, 2016

The Case for Avoiding Bonds During Disinflationary Environments

This post will highlight that what matters over the longer term is the level of inflation over the entire life of the bond (rather than the current inflation rate).

To drive this point home, the chart below outlines the ten year realized real yield of a ten year treasury (the nominal starting yield less the inflation rate over the forward ten years), as well as the realized real yield assuming the current 1.6% yield was the starting yield over each of these periods (going back 140 years).

It may be surprising to learn that the real realized yield was above 0% (i.e. treasuries provided a positive real return for an investor) in ~75% of these rolling ten year periods vs ~25% if each period had a starting yield of 1.6% (those numbers are ~80% and 0% - not once - assuming rates had started at 1.6% over the last 50 years). What may be more surprising was that the best time to have purchased bonds historically was when there were higher levels of inflation present. The chart below outlines the historical (i.e. backward looking) ten year inflation rate (x-axis) against the forward ten year realized real yield (y-axis).

Why have bonds been a better buy when historical inflation was high and worse when historical inflation was low or negative?

Similar to how the market is currently pricing in record low yields given the view of low inflation over the next ten years (likely based upon the low inflation we have experienced over the most recent ten years), the market has historically repriced the yield of Treasuries much higher when we've experienced high levels of recent inflation. Thus, when inflation mean reverted higher from low levels / lower from high levels, the rates (in hindsight) did not properly reflect the forward inflation environment. If we were to look at the predictive power of 30 or 40 year bonds, the predictive power would be even worse (good luck with that 0% yielding Japanese Government Bond).

Thus, despite the low level of historical inflation, current (historically) low levels of nominal yield suggest that bonds are in fact likely as wretched a long-term investment as they seem.

Thursday, July 7, 2016

Global Bonds are a Speculative Investment / The Case for Bond Speculation

Back in 2010 I had a post titled On the Value of Treasuries, where I outlined the total return for Treasuries was potentially greater than the (at the time) 2.71% yield given the very steep curve. The summary was that if the yield curve didn't change over a one year horizon, the total return for an investor was closer to 4.2% (the 2.7% yield plus and additional 1.5% coming from the ten year Treasury rolling down from a 2.71% yield to the 2.54% yield of a nine year Treasury). Not a bad trade at that time.

A Whole New World

Fast forward to today and you have a very different situation. Not only is the yield on the current ten year Treasury about half that 2.7% yield (1.39% as of yesterday), but the yield curve is very flat (the nine year Treasury yield is only 4 bps lower (1.35% as of yesterday) as the back-end has been smushed (a technical term).

As a result, the roll down math (assuming the yield curve does not change) is as follows:

Current yield + change in yield x duration = 1.39% + 4 bps x ~9 years = 1.75% total return

The Resulting Risk / Reward Profile for Treasuries is Horrific

As I'll outline with some simple bond math, a lower yield and a limited roll benefit makes the case for Treasuries much more speculative in nature... either speculative that rates in the United States will continue to move towards or through 0% (we've seen it in places such as Japan, Germany, and Switzerland) or speculative that cash will continue to yield less than the current 1.39% for years to come.

What we do know with certainty is that bonds will provide a nominal return roughly equal to their yield over a period of time roughly equal to their duration. In the case of Treasuries, that means with certainty Treasuries will provide a cumulative return close to:
  • 10 Year Treasury: (1 + 1.39%)^9-1 = 13.2% over the next 9 years
  • 30 Year Treasury: (1 + 2.15%)^22-1 = 59.7% over the next 22 years 
That doesn't mean Treasuries won't provide an outsized return over the next year, but any return in excess of its yield is like squeezing blood out of a turnip limiting its future return potential. The chart below outlines the potential returns for the ten year Treasury following various levels of 12-month performance. This short-term return could, in theory, be 15%, but that would just mean forward returns for the remainder of the duration will be negative (again... the cumulative return cannot move away from that 13.2% figure).

The Risk / Reward Profile for Global Bonds are Even Worse

Japanese Government Bonds (JGBs) show how much juice has already been squeezed out of some global bonds. There is a 40 year JGB that started 2016 yielding 1.42%, which now yields just 0.08%. The below chart uses the same framework as the chart above, but for the 40 year JGB at the start the year.

The result is a monster 50% return year-to-date (even more in US dollar terms given the yen rally), but this comes not at the expense of most of the returns for the next 29 years... ALL of the return for the next 29 years. In other words, to invest in 40 year JGBs effectively yielding 0% you are speculating that yields will move further negative or that cash rates will remain negative (on average) for a period that may be close to (or longer than) the rest of your life.

The Case for Speculation / Low Rates May Be a Self-Fulfilling Prophecy

All of that said, the challenge in this current market is that while the risk / reward profile of global bonds is horrific, speculation of even lower rates may turn out to be self-fulfilling if ZIRP is not partnered with increased fiscal spending.

Why? Because the juice has effectively already been squeezed out of the turnip, meaning future returns in global bonds are gone. This "should" have provided a bump to the global economy and global inflation by taking all the returns (and consumption) and pulling it forward, but it has not.

Why? Perhaps these gains have accrued to institutions that own these bonds (central banks, insurance companies, pensions, etc...) that won't spend it / are already on the hook for liabilities in excess of their assets, while investors who already lack the belief forward stock returns will be near the 8-10% they have historically provided, now project their future interest income from their cash or bonds to be at or below 0%.

The result is the need for individuals to save MORE, not less, for retirement, for a house, for a car. With no institution (corporate, government, etc...) willing to spend this increased savings in the form of investment (which seems like a smart thing to do as the hurdle for a positive NPV project is effectively 0%), deflation becomes a much larger possibility... in turn creating the possibility of even lower rates.

Tuesday, July 5, 2016

How the NBA Destroyed League Parity by Increasing Compensation Parity

The Economist has an interesting article If You Can't Beat 'Em, Join 'Em on the NBA's disaster of a collective bargaining agreement "CBA" that has effectively ruined the chance of parity in the NBA, something it was meant to foster (hat tip Lawrence for the article) and set the stage for Kevin Durant joining the 73 win Golden State Warriors (note: I am a huge Warriors fan and still not happy that he left Oklahoma City from a basketball standpoint... though I am excited to watch Durant play more often).

Two of the main issues outlined:

1) Players are entitled to 51% of all basketball related revenue

Under the current CBA, the players are entitled to 51% of all basketball-related income. So when new money comes in, the team salary cap goes up—and when a lot of new money comes in, the cap goes up a lot.
2) The max any player can receive is capped (at roughly $24 million, depending on service time which I'll ignore)
That gave every team an extra $24m of cap space to sign whoever they pleased—an amount that just happens to be nearly as large as the individual maximum. 

What Does this Mean to the Compensation Structure of the NBA?

In simple terms... each team NEEDS to spend a certain amount of money (I say "needs" as the maximum amount each team can pay is likely below the clearing value if it were a free market) and the amount of money that can be paid to any individual player is capped. My guess is the result is players in aggregate make less, but that non-stars make much more (with the decrease being more than 100% compensated by the stars).

The below is a framework for my guess as to how players would be compensated with no collective bargaining agreement (i.e. a completely free market), with a cap of the current $94 million per team with no player cap, and the current cap of $94 million per team with each player being capped at ~$25 million / year (in the real world there are a lot of nuances in these numbers and they aren't meant to be exact).
  • No CBA: in a free market, teams will simply pay more for talent as a whole (which is what we saw before there were huge penalties invoked in the latest CBA to team's that paid more than their allotment), which would also allow rookies to make more than the limits the league has imposed. 
  • Team's (not players) capped: there has been a lot of research already pointing to stars being underpaid in the current structure due to the max contract each player can receive. Given only 5 players can play at any time and the fact that one player can turn a team from mediocre to the world champs (unlike in a sport such as football), true stars should make multiples more than anyone else and would likely receive the marginal dollar if an owner had to choose.
  • Team's capped + max player contract: if a player would have otherwise made more in % of a team's payroll in a capped structure, then by definition limiting these stars to a certain $$ amount would increase the amount all others could / need to receive. 
Throw in the additional detail that team salary levels spiked when current stars were already locked into below market rate contracts and you have all the ingredients you need for massive overpaying of below average players.

Back to the Economist with the impact...
The real lesson from Mr Durant’s decision is the same as the one provided by Mr James’s original move to Miami: that the NBA’s CBA is a train wreck. At the very least, the maximum contract needs to go—it reduces the entire art of team-building into a sycophantic exercise of courting superstars who cannot be paid what they are worth.
It should be no surprise that if you can't compensate a player much closer to what they are worth if they stay (or in other words... penalize a player more for leaving), they will be more likely to leave an existing team for other forms of compensation (including improved odds of a championship, a more ideal living location, friendship, other forms of compensation, etc...), all of which are more likely only at a handful of teams in the league... including Golden State.