martes, 30 de mayo de 2017

martes, mayo 30, 2017

The Great Reset, Part Two

By John Mauldin


“Premature optimization is the root of all evil…”

– Donald Knuth, from his 1974 Turing Award lectura


This is the second of two letters that I think will be among the most important I’ve ever written.

These letters set out my philosophy about how we have to invest in the coming days and years.

They are the result of my years spent working with clients and money managers and thinking about the economic and particularly the macroeconomic world. Because of some of the developments I will be discussing, I think the future is likely to be extremely challenging for traditional portfolio allocation models. In these letters I also discuss some of the changes in my thinking about the new developments in markets that allow us to more quickly adapt to a changing environment – even when we don’t know in advance what that environment will be. I hope you today’s letter helpful. At the end I offer a link to a special report with more details.

Last week I discussed what I think will be the fallout from the Great Reset, when the massive amounts of global (and especially government) debt and the bubble in government promises will have to be dealt with. I think we’ll see a period of great volatility in the markets.

I offered a solution for dealing with this complexity and uncertainty in the markets by diversifying trading strategies. But that diversification must reflect a rethinking of Modern Portfolio Theory, including a significant reshaping of valuations in asset classes. We’ll deal with those topics today.

Modern Portfolio Theory 2.0

I think successful investing in the future will use a variation of Modern Portfolio Theory. MPT argues that you should diversify among noncorrelated asset classes to reduce overall portfolio volatility.

That strategy is wonderful when asset classes are truly noncorrelated – but we found out in 2009 that noncorrelation isn’t a reality anymore. Going forward, I think it will be more useful to diversify among noncorrelated trading strategies that are not committed to a buy-and-hold process for any particular asset classes. Call it MPT 2.0.

There’s a story here about how my thinking has changed on how we deal with Modern Portfolio Theory. About a decade ago, I gave the luncheon keynote speech at a major alternative investment (hedge fund) conference, on why I thought Modern Portfolio Theory no longer worked. My talk had to do with the rising correlation among asset classes and was an argument for active management and, yes, hedge funds (which of course the audience liked).

The next year the conference organizers invited Harry Markowitz, the Nobel Prize winner who developed the theory, to do the same luncheon keynote. That year, I was speaking at the conference later in the day.

Before Harry’s speech, I met him (for the first of what would be many times) out in the hall and began to question him, based on what I thought I understood about Modern Portfolio Theory. (Yes, there was hubris there – and worse.) Politely, with a smile as if he were lecturing a new student, Harry began to explain to me why I didn’t understand what he was saying, and he commenced drawing quadratic equations in the air with his fingers to explain his points.

What was so remarkable (I swear this is true) was that he was drawing the quadratic equations in reverse so that I could read them. Once I realized what he was doing, I was so stunned that someone could even do that I really didn’t hear much of anything else he said. We talked politely for about 10 minutes, and then he moved on. I don’t think I recovered for a week. But because I didn’t understand what he was saying, I still thought he was wrong.

A few years later, my friend Rob Arnott invited me to his annual Research Affiliates client program, where Harry was in attendance. I reminded him of our conversation and asked the same question I had before, and once more he began to try to get into my feeble brain what he was saying. I will admit he just wasn’t connecting. But Rob kept inviting me back; and as Harry was an advisor to his organization, we renewed our acquaintance annually and became what I like to think of as Friends.

Let me provide a little background on Harry. When his seminal paper was published in 1952, he had just left the University of Chicago to join the RAND Corporation, where he worked with George Dantzig on linear programming and the critical line algorithm that ultimately led to the concept of mean variance optimization. What I think is interesting is that the goal of linear programming at the time was to determine the best outcome in a model (i.e., to maximize profit subject to cost constraints or minimize costs subject to profit constraints/targets – typically applied to the allocation of resources in industrial companies or government agencies). In the 1940s, Dantzig had developed his ideas in work he did for the US Air Force, work he subsequently shared with John von Neumann, the father of game theory. Linear programming has been used to program models of transportation, energy, telecommunications, and manufacturing. The work Harry did in taking linear programming to the next level leads me to think of portfolio construction in terms of “engineering” a portfolio with whatever ingredients are available (stocks, bonds, asset classes, or trading strategies).

Interestingly, Markowitz’s work didn’t achieve importance until the early ’70s, when stocks and bonds got slammed at the same time. It had taken 20 years for his ideas to get a serious look. In addition to the movements in the stock and bond markets that were changing investors’ understanding of risk and its relationship to returns, the development of computing power and the founding of the Cowles Commission and CRSP (The Center for Research in Security Prices) at the University of Chicago gave Harry’s theories new life.

Back let’s turn back to engineering and the concept of utility. The math that is used to engineer a portfolio has been commoditized. We have computers that can do all the work no matter what asset classes we input. Pure robo-digital advisors are doing this task at low cost. The other important aspect of Markowitz’s work is the concept of utility and preference, which showed investors how to trade off risk and return on an “efficient frontier.” This is the act of determining one’s portfolio risk profile and deciding what level of risk is appropriate – where do I want to be on the efficient frontier?

Premature Optimization

The full quote at the beginning of this letter is from the renowned computer scientist Donald Knuth (Stanford) and is very applicable here: Programmers have spent far too much time worrying about efficiency in the wrong places and at the wrong times. Premature optimization is the root of all evil (or at least most of it) in programming.

Many investment advisors use Harry’s concept of the efficient frontier and diversification among asset classes to actually “overengineer” their client’s portfolios, giving them a false sense of security: “Look, here is what this cool program tells us your portfolio should look like. It’s all in the math, and that’s why you can trust it.”

This premature optimization leads people into accepting volatility in their portfolios because they think it’s required. A truer understanding of the efficient frontier is that the frontier is always moving; it’s not constant. So you can’t sit down and plan out your investment portfolio for the next 10 years in one afternoon and expect it to give you efficient, optimized results for years into the future – especially when that optimization is based on past performance and market history that is not going to be replicated in the future. Just my two cents.

This point brings to mind another great Donald Knuth quote: “Beware of bugs in the above code; I have only proved it correct, not tried it.” I can almost guarantee you that the software most investment advisors will use to show you how your portfolio should be allocated will be absolutely mathematically correct. But you will discover the bugs only as the future plays out.

Now back to my story about Harry and me.

Last year, I had an opportunity to sit outside with Harry on a fabulous California spring day, and I began again (hubris alert) to tell him why I thought Modern Portfolio Theory was going to lead investors astray, and I opined that what we needed to do was to diversify trading strategies among the various asset classes. He questioned me about how I wanted to go about doing that, and then he said, “But you are using Modern Portfolio Theory in the formulation of your strategy.” I was puzzled and was determined to figure out what he meant. He had said the same thing to me for several years, and I clearly wasn’t getting it. Our conversation continued, with me as the student and he as the very gracious and patient professor. And then the light dawned.

This is the key: I had clung to the simplistic understanding that Modern Portfolio Theory is about diversifying among noncorrelated asset classes. And it is. But I had a preconceived notion about the importance of particular asset classes. Moreover, with the correlation of all the asset classes that I understood to be in the toolbox “going to one” in times of crisis, it seemed to me that using MPT was simply diversifying your losses, not smoothing out your returns.

Harry patiently explained yet again that the key to MPT is in the words diversification and noncorrelation. The asset classes are just tools. They are interchangeable. Which was precisely what he was telling me when he was drawing those quadratic equations in the air 10 years earlier. I was just too dumb to understand. I hope that if I took his graduate course today, I could pass.

I think much of the investment industry shares my preconceived notion. Rather than opening our minds to a larger world with more potential, we assume we are limited to the asset classes most easily traded (stocks, bonds, real estate, international bonds, international real estate, large-cap, small-cap, etc.) If all you have is a hammer, everything looks like a nail.

I walked away from that conversation realizing – and have come to more firmly believe – that diversifying trading strategies is just another variant use of MPT. Call it MPT 2.0. I can still use all of the asset classes mentioned above (and, as we will see below, hundreds more), but I just don’t have to use all of them at the same time. Back in the early ’50s when Harry was writing his paper, there were numerous asset classes that didn’t correlate. International stocks and US stocks showed significantly different correlation structures and trends. Not so much anymore. Harry’s answer would be to simply change the asset classes in your toolbox and to continue to look for and find noncorrelating classes – or strategies.

Further, most “correlation studies” use past performance to predict future correlation. The sad truth is, that’s pretty much all we have available to us to determine correlation. In my study of correlation, I’ve come to understand that more is required than simply comparing historical return streams. You have to understand the underlying structure of the strategies and asset classes involved.

And that brings us to the third and final problem that will define future investing.

If You Don’t Have an Edge, There Is No Alpha

For investors, alpha is true north on the investing compass. It’s the direction you want to go.

Alpha is the positive return you get through some form of active investing, above and beyond what you would get with simple passive index investing. The theory behind active management, in most asset classes, is that managers can make a difference by using their superior analysis and systems and then putting only the best stocks (or whatever asset class they use) in their portfolios and possibly even shorting the bad ones. The theory says that the better stocks, whose earnings are rising, should go up in price more than the less profitable stocks go down.

The manager’s edge is the ability to differentiate between good companies with positive profit performance and those companies that have problems. And – this is key – for that expertise the manager gets to charge higher fees. If you were particularly good, beginning in the 1980s and through the first decade of the 2000s, you created a “long-short hedge fund” where you went “long” the stocks you thought were the better ones and “short” those you thought would fall in value. There were many different variations on this theme, but they all depended on the manager having an “edge” – some insight into true value differentiation.

But in the past few years that edge seems to have disappeared, and money has been flying out of many funds, and not just long-short funds. Active managers in the long-only space have been underperforming just as badly as their hedge fund brethren. Only about 10% of large-company mutual funds outperformed the Vanguard 500 Index Fund in the last five years.

So it’s no surprise that money is flying out of actively managed retail funds. According to CNBC, passive funds added a record $504.8 billion in 2016:

When it came to funds that focused on U.S. stocks, there was nearly a dollar-for-dollar switch:

Passive funds brought in a record $236.7 billion in investor cash, while their active counterparts saw $263.8 billion go out the door, worse even than the $208.4 billion in outflows during the height of the financial crisis in 2008. That doesn’t even count the more than $100 billion that left hedge funds during the year.

So why would that tectonic shift create problems in the valuation world? It’s simple when you think about it. Let’s take the small-cap world of the Russell 2000 as an example. My friend Paul Lyons at Tectonics went to his Bloomberg terminal and found that 30.7% of the 2000 stocks in that small-cap index had less than zero earnings for the previous 12 months, as of 3/22/17. A chart in the Wall Street Journal shows that the price-to-earnings ratio for the Russell 2000 was 81.46 as of May 26. That is not a typo.

There is $2.26 trillion in US small-cap stocks. Almost exactly 30% of that is in ETFs and mutual funds. My guess is that another 20% is in large pensions and in institutions that simply replicate the index. (Note: There are numerous small-cap indexes to choose from.)

When you buy a small index fund like the Russell 2000, even if the fees are cheap, you are buying stocks that aren’t making any money and are possibly shrinking in company size right along with those that are profitable and growing. In short, you’re buying the good and the bad indiscriminately.

And when everybody is buying every stock in the index in a massive way, there is no way for value-oriented active managers to fight that kind of buying action. They simply have no edge, or very little.

With the massive moves into passive index funds that we have seen in the past few years, shorting small-cap stocks is a prescription for pain. Yes, if a stock has seriously bad performance, it’s going to go down as stock pickers and investors sell; but finding enough such stocks to make a difference in an active fund is apparently difficult. And in the large-cap space?

Forget about it. (Note: If you have a highly concentrated portfolio, with just a few st ocks, it should be possible to outperform. But most people don’t want to take the risk of working with a manager with highly concentrated portfolios.)

So the Trump rally and the massive move into passive investing has pushed up US stocks in general (and to some extent global stocks as well). But what happens when we hit the next recession or loss of confidence? When investors start selling those passive funds, they’ll sell the good and the bad at the same time. In the case of the Russell 2000, they’ll sell all 2000 stocks. In the case of the S&P 500, all 500 stocks. And the move down has historically been much more precipitous than “climbing the wall of worry.”

How Should We Then Invest?

So let’s sum up. In my opinion, the entire world is getting ready to enter a period that I call the Great Reset, a period of enormous and unpredictable volatility in all asset classes. I believe that diversifying among asset classes will simply be diversifying your losses during the next global recession. And yet active management does not seem to be the answer because of the move by investors into passive investing. So what do we do?

I think that the answer lies in diversifying among noncorrelated trading strategies with managers who have a mandate to invest in any asset class their models tell them to. For a reasonably sophisticated investment professional, there are any number of ways to diversify trading strategies.

Up to this point in this letter, I’ve been talking philosophy rather than telling you how I actually intend to go about investing. In the coming paragraphs I’ll tell you how I’m going to diversify trading strategies and give you a link to the actual strategies, performance history, and managers that I will be using. Some of you will not agree with the philosophy I outlined above; some of you will think you can do a better job (or at least a different one) of diversifying trading strategies and managing money.

But, putting on my entrepreneurial business hat, my hope is that some of you will join me.

Back in the day, I allocated money to asset managers who traded mutual funds, before the rules were changed to make active trading of mutual funds either illegal or extremely difficult. But with the growth of money in exchange-traded funds (ETFs), that has changed. Globally, there is about $3.8 trillion in ETFs today. There are almost 2000 ETFs in the US alone, and according to ETFGI there are 4,874 ETFs globally, whose assets have skyrocketed from $807 billion in 2007 to $4 trillion today.

You know how somebody will talk about getting a time-consuming task done and then the next person says, “There’s an app for that”? No matter what asset you want, there’s now an ETF for that. I am constantly amazed how narrowly focused ETFs can be. There is now an ETF that focuses its investments just on companies in the ETF industry. It’s not all large-cap-index ETFs anymore. Some really small, niche-market ETFs have attracted significant capital.

Not surprisingly, a growing number of asset managers actively trade ETFs using their own proprietary systems. I began searching for the best of them some three years ago. I soon realized, for reasons I will explain in a white paper, that a combination of several managers is much better than just one. I have assembled a portfolio of four active ETF asset managers/traders with radically different styles. That approach theoretically gives me the potential for much less volatility than each manager’s system would face individually. The combination I’ve put together has been less volatile historically than the markets, over a full cycle.

Part of my edge is that I have been in the “manager of managers” business for more than 25 years, looking at hundreds of investment managers and strategies. That has actually been my day job when I’m not writing. So when a manager explains his system to me, I can “see” how it fits with those of other managers, understand whether it is truly different, and finally, determine whether it would add any benefit to my total mix. I’ve also learned that having more than the optimal number of managers doesn’t necessarily improve overall performance, but it does add complexity and increase trading costs.

You’re dealing with a few new puzzle pieces analyst,

John Mauldin

0 comments:

Publicar un comentario