The End of Asset Allocation

It’s a pretty solid rule of investing: good ideas tend to become broadly adopted. And once enough people are all doing the same thing, that thing is probably not a good idea any more but rather a bad idea.

Case in point: asset allocation. John Kay seems to have taken a time-travel machine back to 2006:

The basic principles of asset allocation are diversification and contrarianism. Choose securities with returns poorly correlated to each other.

There are two huge problems with Kay’s prescription. The first is measuring correlation — something which turns out, in practice, to be pretty much impossible. And the second is that correlation measures, by their very nature, are always backwards-looking, and that you can be pretty sure future correlation will be very different from past correlation. What’s more, if there’s a crisis in the future, correlations have a tendency to move to 1.
Tom Lauricella takes the opposite tack from John Kay. Asset allocation, he says, is dead: the crisis killed it. And one of the contributing factors to its demise was one of the very things which Kay extols: the ETF. Kay loves ETFs, because they have low fees. But when investors piled into commodity ETFs over the past few years, in the name of diversification, all they really did was massively increase correlations between commodities and other asset classes. As a result, when stocks tumbled, so did commodities:

At Pimco, the firm’s head of analytics, Vineer Bhansali, points to commodities as an example of how diversification strategies can break down. Even as stocks and bonds struggled in early 2008, commodity prices were in the midst of a historic rally. Wall Street rolled out research showing the lack of correlation between stocks and commodities.

But that history didn’t take an important point into consideration. Prior to this decade, investing in commodities was a complicated process due to the complexity of the futures markets. The advent of exchange-traded mutual funds, or ETFs, allowed investors to buy and sell commodities with the click of a mouse. By the summer of 2008, ETF investors had poured billions into commodities in just a few months.

As the financial crisis worsened and stock and bond prices collapsed, ETF investors who needed to raise money found it easy to bail out of commodities, too. That contributed to a 37% drop for 2008 in the Dow Jones AIG Commodities Index.

“When people start buying an asset, the act of them diversifying ultimately makes the asset less of a diversifier,” says Pimco’s Mr. Bhansali.

Rick Bookstaber has dozens of examples along these lines, many of them much less obvious than the link between stocks and commodities. The fact is that if you think you’re invested in an asset class which will zig when the rest of your portfolio zags, you’re probably wrong. Look at the performance of so-called “market-neutral” hedge funds: they all went up in the up market, and they pretty much all went down in the down market. If super-sophisticated hedge-fund managers can’t get correlation right, there’s probably no point in the rest of us even trying.

Ultimately, I suspect that any investment strategy more sophisticated than “buy low, sell high” is doomed to fail eventually. If a certain strategy worked for your grandparents, that’s probably a good reason that it won’t work for you. (And yes, Warren Buffett counts as Gramps for these purposes.) In investing, nothing lasts forever. And the era of asset allocation is in its waning years. The problem, of course, is that no one has a clue what might replace it.

Comments

[Clive Corcoran:]

The very same thing that promoted the kind of asset allocation that you discuss was also its Achilles heel.

Correlation had one great thing going for it - it is very easy to measure - what it measures and what you can do with that number is altogether another matter.

Despite all of the limitations stressed in elementary textbooks and the even more profound misuses touched on elsewhere in your piece, academics have won Nobel prizes in finance and portfolio theory for little more than theories based on simple statistics.

the reason why so many flawed statistical values will continue be misused in portfolio construction is that they provide the illusion of rigor and robustness, and this is especially true for those whose education in modern finance has never exposed them to the requirements of exercising discretionary (not algorithmic) judgment.

Jul 14 07:26 AM

[pickaroonwyo:]

As the liquidity from overleveraging chased ALL assets, these assets became much more correlated because ALL assets became inflated and frothy.

A global marketplace has brought greater market efficiency, hence it is harder to find undervalued assets. This also increases correlation.

Still, with financial center traders living from nanosecond to nanosecond, behaving like lemmings in search of cliff, the average trader has the advantages of perspective, rationality and common sense....

This perspective can be employed to anticipate structural markets. It can also short financial insanity.

Common sense also allows the prudent investor to hold cash and wait for value during capitulations.

So while excess liquidity and overleverage increase correlation, their cyclicality provides an additional dimension to asset allocation. In this case, "buy low and sell high" still holds some truth.

Jul 14 08:10 AM

[La Marque:]

Felix: you have this one right. Anything more complicated than “buy low, sell high” will probably fail. That said, traders can do well by being momentum 'investors.'

As a value investor, I try to buy low ETFs, CEFs and some core stocks for my portfolios.

Jul 14 09:51 AM

[Analyste de...:]

Now we see the blame game.

I don't think "asset allocation" is dead, and I don't even agree with part of Kay's definition. CONTRARIANISM did not inform asset allocation models which largely all went down in equal measure (see all Target 2020 Funds for example.) Why did that happen? Similar losses across the board resulted from managers' largely inflexible mandates to stay fully invested in equities, neither truly "diversified" not "contrarian" in any meaningful way. The proof is in the same-same lousy performance!

You can asset allocate prudently, but you have to have strategic maneuverability to avoid the worst market carnage. Case-in-point: what percentage of mutual fund managers were moving to cash in 2007 or short selling to defensively offset portfolio losses in 2008? No, conventional thinking (industry groupthink) did NOT permit the kinds of strategies that truly contrarian, diversified money managers would have used.

So the argument against asset allocation above is misframed, although several points are certainly valid. Ignorance of RISK, not asset allocation, is entirely at fault here. If we must identify a single major culprit, it's more likely to be same-same broken black-box algorithms: even if your firm didn't use Monte Carlo simulators, you were modeling against someone who was. The circle closed!

Agreeing with the Comments above: solid, contrarian, conservative discretionary judgment was in short supply. There are certainly times to be out of (or shorting) the market: asset allocation doesn't dictate otherwise!

Jul 14 11:04 AM

[TradingHelp...:]

Asset Allocation is theory that incorrectly presumes different risky asset classes: equities, commodities, high-yield debt, commercial property when combined offers risk diversification and an enhancement to risk-adjusted returns. It fails to achieve it when it is most needed - in a recession and inparticular in a financial crisis.

There are not dozens of asset classes in a financial crisis recession, just two: cash and all risky assets. You could maybe claim gold was a 3rd asset class for obvious reasons - its behaviour in times of chaos.

The "leaders" of the banking and wealth management industry are rarely brillilant investment professionals. They can be better described as corporate politicians.

The heads of the big banks that failed here in the UK, HBOS and Northern Rock didn't have a single investment / banking qualification between them. One used to in charge of the clothing range at a supermarket, the other went to a nice Ivy League school.

But there were both rich in privileged contacts and are now semi-retired from the banking industry suffering with generous pay-offs and pensions for their "time" at the banks.

Jul 14 12:26 PM

[Mike McGinley:]

The real problem is not that asset allocation failed. It was that most did not get the allocation correct. Most investors had/have something like this:

70% stocks (domestic, foreign, REITS, etc)
30% bonds (with a large portion of this in lower quality bonds)

It only required common sense to realize that a portfolio like this was not really diversified. A better allocation would be one like this:

50% stocks and lower quality bonds
10% gold
40% high quality bonds (government) and cash

Unfortunately, so many investors are convinced their portfolios MUST have lots of stocks.

Jul 14 12:40 PM

[Larrysyr:]

The article and comments highlight the need for investors to control their own investment decisions.

Mutual/hedge fund managers are not necessarily going to make decisions that are best for the ultimate investor. From my perspective, for instance, their incentives tend to be tied to much too short a time frame relative to my objectives. I'm not claiming I can out-perform the pros, just that I'm more comfortable with results I've generated myself than with going over the cliff with the crowd.

Jul 14 12:59 PM

[Kinabalu:]

There is a lot of good stuff in this article and in the linked articles. However I don't think it supports the authors conclusion. the weaknesses in the correlation calculation are deadly for arbitrage strategies and rating agency evaluations because the calculations cannot be precise enough to ensure success. But asset allocation doesn't need that precision to be successful. In fact, many recent studies have shown that asset allocation based on historically low correllated assets would have greatly reduced losses in the recent market meltdown.

Jul 14 01:29 PM

[Living4Divi...:]

Diversification is one of the few "free lunches" left in the investing world

I like to eat free lunches

Jul 14 01:32 PM

[xsellside:]

50+ years of empirical evidence wiped out by a single year? I’m not buying it. This is nothing new; we know correlations go to 1.0 in crashes. Look at the recent performance (YTD) differentials between growth and value, stocks and bonds, AAA vs. BB, sectors (tech up, telecom flat), developed vs. emerging, etc. and tell me diversification and asset allocation don’t matter anymore. What’s next? Investors capitulating on asset allocation at precisely the wrong time...

ps. If AA doesn't work in crashes, then consider long-dated deep out of the money put options as portfolio insurance.

Jul 14 02:10 PM

[walleke:]

A related idea concerns efficiency of markets and index funds. In an efficient market you "get what you pay for". Efficiency of markets depends on independent rational investors. When there are many such investors the market is efficient. A few can piggyback on an efficient market by buying index funds, ETFs and the like. This gives diversification at low cost. But index funds are not independent rational investors. As more and more people invest in index funds the market becomes less efficient, and the rationale for index funds disappears.

Jul 14 02:29 PM

[nobby73:]

Very well put. However, the concern is in the next financial crisis, or maybe I should say the next phase in this financial crisis, the risky asset may well be cash....


On Jul 14 12:26 PM TradingHelpDesk wrote:


> There are not dozens of asset classes in a financial crisis recession,
> just two: cash and all risky assets. You could maybe claim gold was
> a 3rd asset class for obvious reasons - its behaviour in times of
> chaos.

Jul 14 02:56 PM

[Krugger:]

That WSJ article is awful. first, AA was never a 'fail safe strategy'... and it's certainly not dead and second, what’s the alternative? tactical investing? Most people would get killed far worse if they tried that. It sounds to me like it's a bunch of advisors and investment houses saying AA is broken you need to pay us for our latest greatest thinking instead.

also, great point by Mike above
"The real problem is not that asset allocation failed. It was that most did not get the allocation correct."

Jul 14 03:28 PM

[Lawrence York:]

Asset Allocation is based upon statistics. It fails because the economy moves through business cycles and quants operate models they don't fully understand the assumptions behind. Foremost among those assumptions is regression to the mean. And although models now have become real-time adjusting for new inputs their failure rests on errors like chasing the market and staying diversified. In short they fail to correctly price both value and risk.

Jul 14 03:29 PM

[Carneades:]

Felix-
I like the premise of your article, however, correlation can very easily be measured. The errors occur when one either assumes causation, or that past correlation will always remain intact going forward. With financial time series, we simply have too few data points.

Jul 14 03:46 PM

[bcncv:]

Just like CAPM, asset allocation works perfect in principle, but breaks down when you try to quantify it in a meaningful manner.

The principles are still good. Pick a number of assets that you think have positive growth characteristics, but make sure there is diversification across industries, asset classes, and even investing styles. Don't be afraid to hold short positions (or puts). Companies that will outperform the market are hard to identify. Those that will under perform are pretty easy to find.

Invest some of your money in different investing styles. For example, put part of your portfolio in index ETF's and write covered calls against it. This part of your portfolio will profit over time, and benefits from non-volatile markets. Take another portion of your portfolio and invest it in a momentum strategy. This strategy does well in volatile markets, and should hold some short positions for diversification.

In the end, any investor can still diversify. Just don't fall into the trap of simplicity and put 70% in equities and 30% in bonds as too many advisers recommend. This makes their job easier, but it makes your portfolio weaker. Also remember, diversification doesn't mean zero risk. It is a tool for reducing risk, but it will never be 100% effective at all times in all markets.

Jul 14 04:13 PM

[Fred Swartz:]

If you allow short positions in asset allocation, then the problem of correlations is solved, but of course the hard part of making the right choices remains.

Jul 14 04:33 PM

[Jon G:]

managed futures, deep out of money puts, short managers all did well last year.

Now we have an investment that tracks the VIX in an ETF which I begged Rydex for 3 years ago when the VIX was approaching single digits.

Asset allocation is not dead. As I say on the golf course, its not the tools in your bag, its how you use them.

Jul 14 05:33 PM

[E.D. Hart:]

Agreed, asset allocation is not dead, Treasuries and Gold did relatively well last year. The problem was in the implementation was wrong by most investors who werent not postioned for a 100 year storm.

Dead? Not by a long shot. I will continue to buy and hold stocks, commodities, and gold mining companies. One year does not wipe out 100 years of investing data points.

The global asset bubble burst in a global deleveraging process--something that arguably has never happened before. As the global economy gets back on track--at some point it will--re inflation will occur.

To say AA failed is like saying that roofs are obsolete and prone to failure because a 100 year hurricane tore my roof off. Well, for the next several decades my new roof will protect me from rain and storms just fine.

Jul 15 02:43 AM

[Living4Divi...:]

Felix Salmon - "If a certain strategy worked for your grandparents, that’s probably a good reason that it won’t work for you. (And yes, Warren Buffett counts as Gramps for these purposes.) "

Let me see gramps said: "save your money on regular basis. Be a buy and hold investor. Buy stocks of solid profitable companies and hold on to them rather than trading them, diversify your portfolio."

Warren Buffett said "buy undervalued companies with solid cash flows and hold on to them."

What new paradigm are you saying we should use instead of these old investing maxims? Credit Default Swaps? Mortgage Backed Securities? Triple-Leveraged ETFs? Double-Short ETFs?

Jul 15 06:27 AM

[twotraps:]

Mike has it right, the mix of allocation was not working. Hard to say its 'dead', its just a moving target. Everything is harder, or more interesting now with more money, traders, banks, individual investors, funds trading....with faster access to information and execution than ever before. When has their been so much cash cirulating so quickly in our history? New versions of asset allocation and dirversification will emerge as the mkt changes. Maybe the new strategies do more to account for the speed of changes, or allow provisions to make more freqent adjustments.

Jul 15 07:59 AM

[TLassen:]

Thanks for the article Felix. Asset class allocation of course is not dead, neither is buy and hold strategies. To paraphrase you” If a certain strategy worked for your grandparents, that’s probably a good reason that it won’t work for you”. With all due respect, that makes absolutely no sense. Influenced much by Wall Street Mantra? They claim the end of AA and B&H, no doubt to increase profits from transaction fees.

You make good points, but you misunderstand that is not the asset allocation strategy that is wrong. Investors are simply ignorant of risk and have forgotten the definition of asset classes.

The reason why you diversify by an AA method is to protect capital and to produce future cash flow from dividends and interest income. Not to outperform the S&P in percentage yield. Yes the last two years have been horrible but from an investment standpoint, it is not a reason to believe that we have to discard commonly acceptable investment methods. The more thing change the more they remain the same. To paraphrase Warren Buffet: “A pin lies in wait for every bubble and when the two eventually meet, a new wave of investors learns some very old lessons”.

I believe investors in the last 10-15 years are falling prey to the ‘quick buck mentality’. Capital appreciation is not the only way to make money.
The traditional definition of asset classes has been ignored by many. There are only 4 asset classes: Stock Market, Bond Market, Money Market and Tangible Asset market.

Commodity ETF’s, gold, REIT, mutual funds, Emerging markets etc are not asset classes but are sub categories of an asset class. So to be truly diversified, a better approach would be to divide your portfolio into these 4 broad classes first.

Risk (which is meant to be reduced by diversification) is backwards. Individual risk assessment must be secondary to Asset class risk. It doesn’t matter what tolerance I may have to risk, the fact is that the stock market as an asset class comes with a reward of 8% CAGR, BUT with a 16 Standard deviation. Meaning that in any given calendar year the asset class is as likely to drop -30% as it is to gain 50%

You see often hear Money Mangers tell us that a 25 year old with 30 years of earning potential should be 80% in stocks. Try the backwards approach, what if your 25 year old son won $1,000,000 in a lottery, would you still advise him to be 80% in stock? No instinctively you would advise him to diversify to preserve capital and to construct his investments to produce future cash flows (dividends and interest income)

Jul 15 09:31 AM

[David Van K...:]

I am always surprised, when asset allocation is discussed, that the concept of FLEXIBILITY is never mentioned. It's like there's a search for the "perfect" asset allocation, which, once discovered, will just keep working and working through all conditions and for all time.

That just doesn't make sense, because conditions change and correlations between or among assets can change back and forth over time. That's why you can hit a period where essentially everything's a loser. And the more granular you make it (i.e., the more asset classes you define, including hedges on some or all of them), it seems that the more likely there are to be mistakes, not to mention frictional costs to keep the allocations in line.

My asset allocation approach addds flexibility. Just to keep it simple, let's say I just define three categories: stocks, bonds, and cash. Let's say I determine that an ideal allocation target for me, at my age, with my level of risk tolerance, and in "normal" times, is 50%-40%-10% respectively. Then add flexibility to that, say +/- 30% of each "target", which now becomes the midpoint of a range. My stocks target then becomes 35%-65%, and my bonds target becomes 28%-52%. My cash target I treat as a special case, because its upper end needs to accomodate the lower ends of the other two categories. So the cash target range becomes 10%-37% (the latter would be when I pulled money out of stocks and bonds to the maximum extent).

The flexibility of target ranges allows me, without breaking from my basic allocation principles, maneuverability to move money around in response to changing conditions. I don't attempt to define, nor rely on, a rigid, "perfect" set of allocation targets that's going to work all of the time. I don't think that exists.

Jul 15 09:33 AM

[Ryan Barnes:]

Enjoyable read, both the article and (more importantly) the discussion that follows. I think the author went a bit too far in pinning down a dramatic title, and was really close to a much better idea of posing an open-ended question that deserves serious discussion. But the reports of asset allocation's death are greatly exaggerated .....

As others have pointed out, it's nothing new to say that in the midst of a financial panic, correlations tend to revert to 1 across the board. Asset allocation is about risk, and flexibility is being able to realize that there were higher levels of risk in real estate and stocks in 2007 than investors were being compensated for.

There's also a serious timing component here. It's quite unfair to quell 100+ years of thoughtful study and empirical evidence because of one market event. The whole point of historical analysis is to "smooth over the cracks". While today's cracks are certainly large in historical terms, the correlations of the past 18 months are the outlier, not the norm.

I think we need to take stock of the myriad of new financial products & vehicles to hit the street in the past 15 years and ask ourselves, "what are the new asset classes?".

Jul 15 10:49 AM

[Etsh:]

"Look at the performance of so-called “market-neutral” hedge funds: they all went up in the up market, and they pretty much all went down in the down market."

Yeah, down to the tune of -1.16% for the HRFX Equity Mkt Neutral Index, and -13.09% for the HFRX Absolute Return Index in 2008. Ill take that downside volatility anyday. Seems to me we should be talking more about correlation of strategies.

Jul 15 11:58 AM

[ERCaptain:]

Lawrence York makes a good point.

Much of the other posts I am reading here seem to indicate that people do not believe that 3 or 4 standard deviation events occur. They not only are possible mathematically, it just happened. IN other words, something that may only happen .3% of the time still happens.

I think that one who was applying modern portfolio theory dogmatically, wishes they had paid attention to some of their gut instincts about the amount of risk that had entered certain asset classes.

There will be a reversion to the mean, and once again it will look like AA worked. There is now a greater probability that the next measured variate will deviate less. In other words, this extreme event is likely to be followed by a less extreme event. Buyers will feel comfortable that it is safe to invest again, and the market will continue to cycle.

On Jul 14 03:29 PM Lawrence York wrote:

> Asset Allocation is based upon statistics. It fails because the economy
> moves through business cycles and quants operate models they don't
> fully understand the assumptions behind. Foremost among those assumptions
> is regression to the mean. And although models now have become real-time
> adjusting for new inputs their failure rests on errors like chasing
> the market and staying diversified. In short they fail to correctly
> price both value and risk.

Jul 15 01:55 PM