Systematic Relative Strength

Jeremy Grantham of GMO posits a tension between doing the right thing for the client and getting terminated as a manager. Much of this, he believes, is a function of the client’s patience. He writes in Advisor Perspectives:

Ridiculous as our market volatility might seem to an intelligent Martian, it is our reality and everyone loves to trot out the “quote” attributed to Keynes (but never documented): “The market can stay irrational longer than the investor can stay solvent.” For us agents, he might better have said “The market can stay irrational longer than the client can stay patient.” Over the years, our estimate of “standard client patience time,” to coin a phrase, has been 3.0 years in normal conditions. Patience can be up to a year shorter than that in extreme cases where relationships and the timing of their start-ups have proven to be unfortunate. For example, 2.5 years of bad performance after 5 good ones is usually tolerable, but 2.5 bad years from start-up, even though your previous 5 good years are well known but helped someone else, is absolutely not the same thing! With good luck on starting time, good personal relationships, and decent relative performance, a client’s patience can be a year longer than 3.0 years, or even 2 years longer in exceptional cases. I like to say that good client management is about earning your firm an incremental year of patience.

On the one hand, this is kind of funny from a manager’s point of view because it is something we can all relate to. DALBAR has documented that a client’s average holding period is about three years, and that is exactly the conclusion that Mr. Grantham comes to also. (The bold is mine; I think Mr. Grantham’s twist on Keynes may become a classic.)

On another level, this is very sad. It’s sad that good client management is required to earn an extra year of patience. As an industry, we apparently do a poor job of educating our clients about realistic expectations. If we start a relationship promising sunshine and rainbows, of course the client will be disappointed when the first dark clouds appear. On the other hand, if we warn clients about the inevitability of rain, and the possibility or likelihood of hail, tornadoes, and earthquakes, they are likely to sign up with our sunshine-and-rainbows competitor.

And, honestly, part of the blame may lie with the clients. Many clients want to hear about sunshine and rainbows, not rain and hail. If both are mentioned, they tend to remember the sunshine and rainbows and have only a hazy recollection of anything else.

Here’s the problem: return factors, even historically reliable ones like relative strength or value, tend to play out over periods longer than three years. This is why there is such a disconnect between manager returns (NAV returns) and client returns (dollar-weighted returns). I guess if return factors were so uber-reliable that they worked every year, there would be no patience problem. Clients would be happy to sit on their hands and collect the premium.

Unfortunately, collecting on return premiums is a lumpy business. In extreme cases, you can have situations where there are a number of years that go nowhere, followed by all of the excess return in a six-month period. Clients ideally would like to be invested just for those six months, but no one ever knows at what point in the cycle the excess return will occur. This makes it really tough for clients, as they essentially have to make a leap of faith.

The ideal client is one whose “standard client patience time” is infinite. We have a few very long-term clients here that have been with us since 1994, almost twenty years. They’ve moved from capital accumulation mode when they first joined us to distribution mode some years ago. In a couple of cases, they’ve already withdrawn more money than they started with—and still have balances in excess of their original deposit. Every money manager would clone clients like these if they could.

Here’s an interesting thing that Mr. Grantham doesn’t mention: if you talk to any number of advisors, you will find that, inevitably, the clients with infinite patience tend to be the clients with the best performance! I don’t know what twist of karma makes it so, but advisors all know this phenomenon. Clients who can make that courageous leap of faith tend to be rewarded. It’s our job as advisors to allow the clients to feel comfortable doing something that is inherently uncomfortable for them.

From MarketSci, a article with a graphic representation of all of the corrections!

Some time after 2008, Lou Harvey, the founder of DALBAR, did an interview with Barron’s. He discussed what had been learned in the 2008 bear market. It’s worth thinking about, especially his advice on what seemed to work. I’ve got some nice excerpts, but you can read the whole interview here.

In your study you point out that in spite of the “catastrophic” losses in 2008 “belief in modern portfolio theory has inexplicably remained strong.” MPT is grounded in the belief that asset classes are “predictably uncorrelated.” Because MPT is no longer good for all seasons you relegate it to one of several things investors need to consider.

Modern portfolio theory was pioneered by Harry Markowitz in a 1952 paper published in the Journal of Finance. It posited the construction of an “efficient frontier” of optimal portfolios offering the maximum possible expected return for a given level of risk. MPT is a principle-based investment strategy whose basic idea is that better returns can be produced if a portfolio’s holdings are diversified among different asset classes. The idea: to take advantage of the varying directions as each asset class fluctuates.

Nothing’s wrong with MPT. It’s how people use it. What is needed is a back-up plan to protect investors when the theory fails. And it will most likely happen again.

The pushback to our latest report has been strong. Who are we to criticize MPT, devised by a Nobel Prize winner? But this is our research and we stand by what we’ve found. I have lots of scars on my back to prove it. And I’m sure that after talking to you I’m going to get more. Investment results in 2008 showed clearly that correlation of asset classes varied unpredictably and with no warning. This brings into question the very basis for MPT and its ability to forecast an efficient frontier. MPT simply cannot be used in isolation. Instead it should be thought of as only one reference point for modeling the behavior of a potential portfolio. It is only one dimension of a more comprehensive investment-management process.

What’s the most important finding in your recent research?

For me the biggest recent issue stems from the 2008 market meltdown that defied many of the core beliefs in the financial community—the core belief that asset classes are not correlated. When stocks go down, bonds go up. So might real estate. By holding a little bit in each basket, the investor will make steadier returns and avoid losses. We found out that all of the methods based on modern portfolio theory worked within a certain range. Outside of that range, they all failed.

I added the bold. I’m not sure that a theory can be said to have worked if it fails repeatedly under extreme conditions, so I differ with Mr. Harvey on that. As another pundit said recently, it’s like having a seatbelt that only fails when you have a crash. Of more practical interest is his observation of what did work for investors.

We surveyed investors who outperformed in the crisis and tried to glean from them points to ensure success.

One I particularly like is to take your portfolio out and parse it out into smaller, purpose-driven components, and treat each component separately. Money you can’t afford to lose should not be put into the stock market, but rather in something else that guarantees repayment. It could be an annuity. It could be high-quality bonds. The approach functions on the different ways investors look at their holdings. You look differently at money you have set aside for a gay old time on the Riviera than you do at the money you use for breakfast tomorrow. Each basket should define the risk one should take in the market.

DALBAR’s always done interesting work and their research on investor returns versus NAV returns is now legendary. Here, Mr. Harvey suggests that investors who outperformed in the crisis were those who divided their money into buckets of differing volatilities. This is a psychological trick gleaned from behavioral finance—just a different way of looking at the same allocation. We’ve written about this before, but his observation suggests that it works in practice, not just in theory.

Best practices seem to suggest a belief in supply and demand rather than modern portfolio theory, and ratify the idea of using volatility buckets for clients.

After the 2008 market debacle, a couple of professors devised a Financial Trust Index to gauge how Americans were feeling about things. The idea was that a large sample would be surveyed every quarter and then the data tabulated. (You can see their website here.) My specific interest here is an article on the Financial Trust Index that appeared in the Wall Street Journal on July 20, 2010. This was a little more than one year from the market bottom in 2009. Prices had already advanced nicely, but this is what the article had to say:

About 45% of people think the stock market will drop by more than 30% in the next year, according to a survey by economists at the University of Chicago and Northwestern University. That’s even worse than the 42% a year ago who thought such a sharp decline in stocks was likely. (In December 2008, the share stood at 56%.)

And they’re not counting on much of a payback, either. The average expected return on investment in the next year was just 1.4%, versus 3.5% three months ago.

A huge percentage of investors were expecting another market drop of severe proportions. Why? Well, a big drop had happened in the not-too-distant past and like most forecasters, they were extrapolating more of the same into the near future. We’ve written a lot about the folly of forecasting before, so poor forecasting technique isn’t really surprising. Given the well-documented dismal performance of retail investors, why would the Wall Street Journal even run an article highlighting their forecasts?

Media is a business. Like all businesses, they are trying to maximize their revenues. Especially in this digital age, they can see which stories get the most clicks. Some websites even have a “most popular articles” list. The most popular stories are often stories that create fear. Fear is a much more powerful emotion than satisfaction. In fact, prospect theory shows that people react twice as strongly to negative outcomes as they do to positive outcomes. Therefore, many stories in the media are going to have a negative slant. And, of course, stories with a negative slant are most believable and appealing to us when things currently are going badly.

Ignoring forecasts in the media is a necessity for good investment performance. You’ve got to have a thoughtful investment process—and stick to it, especially when conditions are terrible. That was a good policy also when this article came out in 2010. Rather than the forecasted investment return of 1.4% for the following year, the S&P 500 was up more than 22%. That’s more than double the average annual return: it wasn’t just a decent year, it was a great year. Reflect on that before you deviate from a reasonable investment policy.

…comes from Ed Yardeni and his excellent blog:

Over the past 36 months through February, net inflows into bond mutual funds totaled $1.0 trillion, while net inflows into equity funds were close to zero. Unfortunately for bond investors, the equity funds enjoyed capital gains of $2.7 trillion over this period, while the bond funds had gains of only $437 billion. Now that bond yields are starting to move higher, those gains are likely to decline. That might convince individual investors to move back into equities.

His chart of bond flows is actually pretty amazing. Net flows to stock funds, in contrast, have been pretty flat for the last five years.

Source: Dr. Ed’s Blog/ICI (click on image to enlarge)

There’s no doubt that good equity returns provide some competition for risk-off assets like bonds. However, stock returns have been pretty good for several years and it hasn’t resulted in significant behavior change on the part of bond buyers. In my opinion, that’s because bond investors are still making money on an absolute basis. Sure, they aren’t making what the stock market is making—but they are not losing money. If and when bond owners start to actually lose money on a nominal basis, we might see a change in the dynamic. (Many are already underwater in terms of real returns.) That change could provide fuel for the stock market, but maybe only if stocks are doing well at the time.

Although seems logical that money would flow immediately when relative performance changes, there is obviously a big emotional component to investor behavior too. Emotions and good investing don’t go well together.