Market Timing: Should You Attempt It?

» Normxxx - How mutual funds stole your money

By Jon D. Markman

As new allegations of Wall Street wrongdoing surface virtually every day, it's hard to escape the conclusion that three years of manic-depressive stock prices, major corporate bankruptcies, and the prosecution of analysts, insider-trading and IPO frauds did nothing to inhibit a pervasive culture of corruption in much of the mutual fund industry.

It's unclear whether this came about because of pressure to produce results at any cost, a blurred line between aggressive and illegal behavior or an expectation of light punishment for out-of-bounds acts. Whatever the reason, an alarming number of fund managers and their salespeople at brokerages appear to have slipped into a habit of unlawful acts that breached their fiduciary responsibility to retail customers and cost investors billions of dollars.

At first, it seemed that just a handful of fund firms were involved. But new evidence suggests that large swaths of the fund industry have systematically leveraged legal loopholes and lax enforcement by federal regulators to become the white-collar equivalent of organized crime, shaking down a quarter here and a dollar there from a public that had become its patsies.

“Half the industry is probably implicated in one way or another,” says Mercer Bullard, a former Securities and Exchange Commission counsel who now teaches law at the University of Mississippi.

-- posted by Normxxx

» JIMMY62 - VEGas

There is a lot that book learning can teach about gambling that you could learn in a longer time in the school of hard knocks.

The same is true to stock market investing.

A major major difference is the odds in stock market investing keep changing. Whereas Vegas games don't change that much.
Even in poker where you are playing a player the number and denomination of cards is fixed so that near the end of each hand the all ways in which you could be win or loose can be counted. Books can help you learn to do those counts.

The range of the variables in the stock market is so very much greater that if there was not a history of an upward bias few would invest.

-- posted by JIMMY62

» Kirk - Market Timing and Time Dependence Redux

Author: bernardo
Date: December 19, 2003 4:16 PM
Subject: Market Timing and Time Dependence Redux

The purpose of this post is to distinguish between market timing and a variety of other investment strategies that are time-dependent, i.e., result in changes in portfolio allocation over time. In the process, we define and describe the various strategies and their implications.

Market Timing and Time Dependence
Now "market timing" is another one of those terms like "risk", "rationality", "efficiency" that tends to get thrown around haphazardly, carelessly and thoughtlessly, often with nary a proper definition, description, delineation, distinction, or qualification. Let's take a small step toward correcting that sorry state of affairs.

Investing is, by its very nature, time dependent. For example, investors' life situations change over time, thereby impacting their utility and risk. The amount of available assets for investment also varies over time, sometimes significantly. Also, since investors can't hold assets forever (‘in the long run we’re all dead’ to quote Keynes), they must address a problem with finite boundary conditions, standing rather than traveling waves with periodic boundary conditions if you will. Moreover, utility issues and constraints in general complicate the solution to the individual investment problem.

This is known in mathematics as optimizing a functional, where the constraints are often expressed as Legendre multipliers or the equivalent. The important point for our purposes is that in the real world finite boundary conditions and varying constraints and conditions invariably and unavoidably yield time dependent solutions to investment problems.

Having said that, it's important to note that finite time spans and changes in resources and utility are hardly the only factors that induce time dependence in investments. The time dependence can result from conformance to an initial criterion. Consider, for example, the "coward's" portfolio allocation, i.e., Nx(100/N). In order to maintain the initially established target distribution, the investor rebalances periodically, which results in a time varying allocation.

Granted, buying and holding any basket of stocks, whether TSM, multiple indices or individual stocks forever are extreme examples of nearly time independent strategies. But clearly, most investment prescriptions other than autopilot (designate the initial distribution but don't adjust thereafter) incur time dependence, but that does not mean by any stretch of the imagination that they are market timing.

So just what is market timing? It's actively changing portfolio allocation based on market externals or time varying subjective judgments about equities. In market timing the initial prescription, formula, and/or critreia are allowed to change over time. Examples of market externals include leading economic indicators, US & world economic conditions, political factors, Fed actions, consumer confidence, astrology, proclamations of gurus, and the like. Judgments about equities include buying and selling individual stocks (active management) based on FA or TA, and the same for market sectors or the entire market for that matter.

An obvious example of a subjective judgment is proclaiming tech stocks are the wave of the future (or not) and varying an allocation accordingly. Note that the criterion for market timing is not applying a subjective judgment per se, but allowing subjective judgments to actually change over time based on market externals or perceptions thereof. For example, an initial judgment electing a coward's allocation is of course subjective, but to the extent that judgment remains constant regardless of market conditions and externals, then the strategy is not market timing according to the above definition.

Time Dependence of Investment Strategies
In the real world, as discussed above, all investment strategies are necessarily time dependent to some degree. Arguably, the least time dependent is pure buy & hold, but technically speaking the act of buying occurs at specific times, and also investors may choose to vary their allocation between fixed and equities over their lifetime, so some elements of time dependence are almost always involved. Examples of time-dependent strategies include but are not limited to:

1. Buy & hold . Often implemented using TSM (total stock market) index mutual funds or ETFs, an ideal approach for passive investors who wish to incur exactly the same risk as the market as a whole.

2. Buy, rebalance & hold . Those wishing to deviate their risk profiles in from that of the market as a whole, may fashion an allocation from near-independent (weakly correlated) asset classes and rebalance periodically to maintain the desired risk profile. The tilt can be toward lower risk (L & G) or higher risk (S & V), as the investor prefers.

3. Agnostic or cowardly. A third passive approach leaves the risk profile to the gods by distributing equally among asset classes. The simplest version is 4x25%, i.e., 25% each in domestic LG, LV, SG & SV. Others extend this further to 10 classes or more by adding the foreign analogues plus REITS & EM, say. In all cases, periodic rebalancing is required to maintain the initial distribution.

Clearly, the act of rebalancing injects time dependence into an otherwise passive approach. We might reasonably expect some built-in protection against bubbles from this approach when the allocation is well diversified among risk categories.

BTW, whenever we refer to risks in this context we mean compensatable, non-diversifiable underlying equity risk factors (see Swedroe’s, Armstrong’s or Bill Bernstein’s books or go to Troutner’s TAM site for more detail and clarification of these terms).

4. Price-centric investing determines initial purchases of asset classes in the portfolio as well as the rebalancing of those asset classes based on valuation metrics. Price-centric differs from VMA (see below) in that the comparison with risk free is incidental and the metrics don't necessarily center on expected return alone, but favorable pricing from multiple standpoints, including but not limited to PE, PB, PS & PD.

The historical basis for favorable pricing is expounded in Cochrane's seminal paper. The spiritual basis borrows from Ben Graham, but applied to asset classes rather than individual stocks.

In any case, the price-centric investor decides on a target mean price (values for price metrics) and an acceptable distribution about that mean, skewed to the high side for the risk averse and low side for the risk tolerant. The means and distribution shapes will vary by investor but are selected on some rational historical basis. The distribution is invariably narrow enough to solidly protect against bubbles.

Personally I favor mean PEs below 20 and PBs below 1.5, but each investor has to develop their own criteria. Be aware, however, that these values generally incur higher risk than the rest of the market. Also, there will be times when only a few asset classes satisfy the goals of the price-centric investor and they may lighten up on equities in that case because the degree of diversification falls near or below threshold. Other times many asset classes will be acceptable prey, just depends.

Bottom line: The key to price-centric investing is establishing an objective set of metrics and guidelines for allocation to avert timing (subjective decisions) and maintaining strict discipline.

5. Valuation-modulated allocation (“VMA”). This approach is discussed in more detail below. To make a long story short, VMA (initially expounded by Larry Swedroe in a series of posts on Indexfunds.com) bases the decision of whether or how much to invest in particular equity asset classes on comparisons of expected returns between risk free and those equity classes.

While there are no hard and fast rules, the criterion for investing would usually be a few % points higher expected return from equities before making the plunge. When SV indices were at PEs ~ 25 (expected return ~ 4%) and 4 yr CDs were ~ 4%, the VMA buff would likely have sat on the sidelines. Certainly the expected returns from high flyers with PEs of 50 or above never even came close to competitive with risk free, thereby avoiding bubbles from the git-go for the VMA practitioner.

The metrics may be absolute (PEs below some historic mean for that asset class, say) or relative between asset classes. Relative values are often suggested for allocating between fixed and equities. In this case one sets a target value for the premium for expected equity returns (based on E/P, say) over risk-free (short term fixed). If the target is not met then the % allocated to equities is zero or small, if it is met handily the % may be large.

VMA is based on several concepts, including RTM, the tendency of equity premia and returns to revert over time to historical values, and more importantly, to the Cochrane arguments about the performance of low priced stocks (topics poorly treated by Ben Stein but better expressed in Swedroe’s new book). Lastly, depending on where the bars are set, VMA by its very nature will protect investors against the infrequent bubble in the market as a whole or in particular asset classes.

The above approaches, while admittedly time dependent to varying degrees, do not constitute market timing for several reasons, mainly:

(a) They are essentially passive and mechanical
(b) They do not make subjective judgments about the economy, direction of the markets, politics or any other external factors
(c) Buying and selling is determined by life cycle factors which dictate changes in underlying asset risk profiles over a lifetime, not by any assessment of where the market is going or whether it’s properly valued at a given time

I’d argue that when VMA or price-centric investing are implemented in a mechanistic and objective fashion, and do not rely on any subjective assessments of market externals, then despite being time dependent, they do not constitute market timing.

Jeff Troutner disagrees and claims any system that attempts to value the market, even using objective standards or metrics, is by definition active (often referred to as “dynamic asset allocation”). However, Bill Bernstein’s discussion of dynamic asset allocation, although somewhat evasive and tentative, appears on the whole to be quite tolerant of valuation based strategies. In addition,

I believe Larry’s position on the issue is that VMA is not market timing in the conventional sense, and constitutes a legitimate approach to passive equity investment as long as it is implemented objectively & mechanistically.

In contrast, true market timing is characterized by subjective assessments of where markets are headed, based largely on externals, and sometimes on market internals as well. In market timing time dependence is at the core of an active strategy, in contrast to passive investing, where time dependence is incidental, dictated by mechanistic factors bereft of any subjective judgments on the part of the investor.

Agnostic Equity Diversification
The primary decision in portfolio allocation is between risk free (ie, to capital) and risky (equities). This decision should not be based on diversification per se, but on an individual's desire to take on risk.

IMHO, all investors should place all the money they can't afford to lose in risk-free. Only put into equities what you can afford to lose in its entirety. I readily concede there are other approaches, preferences and opinions on how to allocate the risk free portion, so there's certainly some latitude.

The weakest argument IMHO is to allocate based on the global asset weighting of fixed & equities, which is obviously absolutely & totally irrelevant to any particular individual's situation. Anyway, the guiding principle here is building a safe core, we can argue how best to accomplish that goal, but in any case the diversification (lack of correlation) angle is incidental and secondary to that decision.

Second, if you think there is a good amount of efficiency in the market or that you're incapable of exploiting whatever inefficiencies there might be, then your guiding principle ought to be passive equity investing, and what's more agnostic composition and time dependence to allocation.

Agnosticism is best implemented by assigning some form of equality to all equities in the world which meet appropriate standards (so that only compensatable risks remain), and acquiring them at no one particular time but continuously (however, since we're all dead in the long run there are built in limits). The idea behind agnosticism is unbiased distribution of risk and reward across the full range of the spectrum without playing favorites or making any subjective judgments whatsoever.

The problem of implementing agnostic equity diversification is practical not philosophical. The best we can do with available investment vehicles is try to find parameters reflected by indices that are mimicked by index funds and ETFs. One basis set is all the equities in the world (local basis functions) equal weighted, clearly impossible or impractical for individuals to implement.

A more practical but clearly limited approach looks to (company-diversified) collections of equities with similar characteristics (modal basis functions). The difference between local and modal function sets is age old in mathematics and each has its benefits and limitations.

In any case, using a modal approach means we better ferret out the "best" (most nearly orthogonal & complete) basis set of quasi-independent, low correlation variables. FF have proposed some of the most revered variables ("factors") among the academic community, certainly not unique but quite credible nonetheless.

But the problem doesn't end there because even for small N, say N=3 (size, style, country of origin) the full matrix of 27 possibilities is difficult to implement or requires active vehicles or those from restricted sources (DFA). That's why us ordinary blokes settle for less, much less, not because we don't know what we want, just that we can't quite get there with available vehicles and mechanisms. So we make compromises, perhaps many of them, but the guiding principle remains firm and simple.

Forget couches, coffee houses or whatever. Strive to invest agnostically, and to the extent you falter, lament your deficiencies but never throw in the towel.

Dissecting VMA
First, some general comments: There are several criteria for evaluating a strategy, including consistency, correctness (no patently false assumptions or assertions), effectiveness, utility and ease of implementation. Now there may well be insufficient data to determine the effectiveness of VMA compared to other strategies, so we may not be able to judge the effectiveness quantitatively. When I say effectiveness, I’m talking about performance, average returns and SD, and incubation times.

So if we wish to consider adopting VMA, we have to bite the bullet and assess effectiveness based on “logic”, gut feel or common sense. Of course, when data are insufficient, each proponent of an argument believes that theirs and only theirs is the one sensible and logical alternative, and we’re unalterably reduced to using subjective judgments in the decision process.

Now as far as ease of implementation, there’s a complex behavioral issue, whether the average investor will in practice be able to implement and stick with the strategy. Investors who are able to do so might be called “sophisticated”. It’s important to note that intelligence is a necessary but not sufficient condition for sophistication, and the two should not be confused. Anyway, I’m going to omit the behavioral issues here and assume that an investor who chooses to do so can successfully implement the prescribed strategy.

What we will do here is attempt to define the main elements of VMA, focusing on the consistency of various elements of the strategy, but leaving any detailed consideration of effectiveness, utility and implementation aside. By consistent we mean all parts of the strategy fit together in a compatible and complementary fashion.
So, here we go. Seems to me VMA consists of 5 main steps (and perhaps several other small ones we omit here):

1. Determine a nominal allocation between equities and fixed based on risk tolerance and utility (obviously) a highly individual choice.

2. Select a nominal allocation highly diversified between asset classes by style, size, domestic vs. foreign, and additional components such as EM and RE. One possible choice is equal weighting all of the chosen classes (sometimes called “agnostic” or “coward’s portfolio”), but other allocations skewing toward historically higher performers are also possible choices.

3. Modulate the nominal allocation based on valuation, i.e., determine the likely return for each asset class and compare it to the current risk free return (the valuation test). If the return for that asset class is higher than the risk free rate then retain the asset class in the allocation, if not exclude it.

Perhaps the simplest yardstick for estimating equity returns is E/P, but DDM or other more detailed criteria can also be used. The funds earmarked to the asset classes that don’t pass the valuation test can go to fixed income or redistributed among the surviving asset classes, depending on an assessment of risk tolerance. (The redistribution to other equities would appear somewhat more consistent choice with the initial allocation between fixed and equities, but one might also argue for shifting temporarily to fixed until the excluded asset classes pass the valuation test.)

4. Implement the selected equity allocation using passive index mutual funds or ETFs.

5. Rebalance periodically (every 1 to 2 years). Exclude asset classes that don’t pass the valuation test, reincorporate those that do. Among the remaining components, redistribute (rebalance) to the initial allocation %s. This is easier with equal weighting, but more complicated for skewed allocations.

The above defines a consistent strategy, which says nothing about its effectiveness or ease of implementation. That apparently remains a matter of contention among the experts.

VMA falls into the category of a time dependent strategy but arguably, according to the definitions given above, not "market timing". In VMA an initial metric is established, namely, the returns of asset classes compared to risk free, and a mechanistic response such as "allocate equities as Nx(100/N), but if imputed returns on an any asset class falls below 75 basis pts compared to TIPs, then readjust the equity allocation among the surviving asset classes".

Of course, the latter prescription is just an example, there are many other possible variants. But because there is no change in the initial judgment or criterion, and market externals do not come into play in the investment decision at all, the strategy is not market timing according to the definitions stated above. The time variation is incidental to the strategy, not the direct result of an active decision process where the initial strategy, criteria and judgments are allowed to change over time as they are in "market timing".

Just for clarification, my use of VMA or coward's allocations as examples above are not necessarily endorsements of these strategies. The objective is to distinguish between true "market timing" strategies and those that happen to be time dependent but don't rely on market externals or changing judgments or criteria over time. In fact:

1. VMA need not imply any change in the split betwen equities & fixed, rather, just realignments within the existing equity allocation

2. VMA does not depend on any externals, only market internals; the data ideally come from the asset classes themselves, not Wall St, talking heads or astrologers

3. The formula for applying VMA can have a range of deltas, that's OK, but arguably too large a range incurs the danger of subjective rather than mechanistic judgment, ie, bad habits

4. VMA is one among several valid (and self-consistent) approaches to portfolio risk management

5. In the case of VMA investors reap the benefits of consciously addressing several pivotal issues:
(a) Why am I investing in equities at all?
(b) How should I manage risk? Is the potential benefit of investing in a particular equity class worth the risk? How doe the risk fit my utility profile? Does that risk change over time?
(c) How should I distribute risk among equity classes?

So whether or nor the investor decides to actually invoke VMA, he or she will already have benefited substantially from seriously and honestly tackling the fundamental issues which underlie VMA.

Time Dependence due to Changes in Risk Profile
Larry discusses the issues of need or desire to take on risk at great length in his books, so I’ll be very brief here. The main point is that virtually all of us change our desire to incur risk at various points of our life.

Recall according to MPT & FF the primary risk factor is equity risk, so the main determinant of portfolio allocation is the distribution between risky assets (such as equities and LT bonds) and nearly risk free assets (such as cash, high quality ST individual bonds, TIPS and I-bonds. Our desire to incur risk is determined by many factors, such as age, temperament, financial condition, health, family matters and the like.

One important determinant is the progress in wealth accumulation. When one has accumulated sufficient resources, such as after a large market run up, then the need to take risk diminishes significantly (however, the desire may still be there for inveterate gamblers).

Investors who shift their allocations due to changes in risk profile alone, not their perception of where the market is headed, are not at all practicing market timing. Larry and others explain the whole subject very well in their books, so elaborating further here is not necessary.

Dialogue
I very likely have already overstayed my welcome in this post. Yet, these subjects are too vast for anyone to cover in a single post or even a series of posts of virtually any length. So let’s leave it like this. Anyone who would seriously like further discussion, information, links to source material, citations, or explanations on any of the topics above, or would like to challenge particular concepts or statements or offer alternatives, please feel free to post so I and others can indulge any such requests.

-- posted by Kirk

» Normxxx - Positive thinking pays off?

The top performing newsletters are 31/2 times more bullish than the bottom performers.

In fact, there is not a single out-and-out bearish newsletter among the five best newsletters of the past decade. Instead, three are outright bullish, a fourth is on the bullish side of neutral, while the fifth is on the fence.

In contrast, four of the five bottom-ranked newsletters are bearish, some aggressively so. There is only one bull.

This contrast is noteworthy, particularly since the ranking on which this contrast is based focuses on risk-adjusted performance over a 10-year period.

Those 10 years contain both a strong bull market as well as a devastating bear market, during which the stock market on balance performed no better nor no worse than its long-term historical average.

In other words, there should be no bullish or bearish bias to focusing on the consensus opinions of the best and worst performers over the past decade. That's why it behooves us to pay attention that the top performers are so much more bullish than the bottom performers.

  more. . .

-- posted by Normxxx

» mdorsey - Here is how market timing works.

December 19, 2003
.
Vanguard 500 Index (VFINX) 
YTD total return  +25.76% .
Since 12/31/99 return  -21.68%

.
Vanguard Prime MM Inst  (VMRXX) 
YTD total return  +0.87% . performance vs 500 Index –24.89%
Since 12/31/99 return  +12.56%…..…………………………+34.24%

.
Vanguard GNMA (VFIIX)
YTD total return +2.72%.  performance vs 500 Index –23.04% 
Since 12/31/99 return  +35.33%…………………………+57.01%

Vanguard Total Bond Mkt Index Inst (VBTIX)
YTD total return +4.32%. performance vs 500 Index –21.44%
Since 12/31/99 return +36.59%…………………………+58.27%

Vanguard Short-Term Bond Index (VBISX)
YTD total return +3.32%. performance vs 500 Index –22.44%
Since 12/31/99 return +31.64%…………………………+53.32%

Vanguard Prime MM Inst  (VMRXX) Since 7-21-2003
Vanguard Total Bond Mkt Index Inst (VBTIX)
YTD total return +2.23%. performance vs 500 Index –23.53%
Since 12/31/99 return 33.85%…………………………+55.53%