Page 13 of 24

Re: QSPIX - thoughts on interesting fund

Posted: Sun Aug 23, 2015 9:20 am
by lack_ey
afan wrote:Some of the discussion in this thread has conflated the behavior of a given factor with the effectiveness of a particular mix of factors. Two different issues. The latter always raises multiple testing concerns. The former might or might not depending on how the factor has been tested and supported in the past.

Two different issues, but related and both of interest. Also, as mentioned before, the 'mix' being touted and tested is not the optimal one over the data set (if it were, this would be a clear sign of data mining and would strongly require more stringent tests for significance of outperformance over other combinations). That doesn't mean multiple different combinations weren't at least considered, though, but if you treat all four styles individually as significant or close to it then it is hard to imagine a scenario where you'd be skeptical of any combination of them, especially with the correlations seen.

afan wrote:Some of the discussion has also ignored concerns about the definition of performance. If the goal is higher expected risk adjusted returns, then one must define risk. Some of the discussion has assumed risk is captured by variance alone. For other comments it is not clear what is meant by risk. I think it is clear that real world investors care about at least skewness and perhaps higher moments. Analysis of the effect on an overall portfolio should take this into consideration.

Standard deviation may be close enough. But there's some minor negative skewness in the dataset, and I would not at all be surprised by more in real life. With around a 10% standard deviation target, because of the unknowns I think a close enough mental model is somewhere between 60/40 stocks/bonds and 100% stocks kind of risk, probably on the latter side.

afan wrote:The concern about running a fund with a mix of strategies, a mix of factor bets, is that there may be covariances not accounted for when the factors are treated as independent. Even if one models some level of interaction, the risk remains. That was what happened to LTCM. They thought they had modeled the covariances, only to discover that in an extreme event, their bets were much more highly correlated than they imagined. Given the Nobel - level IQ's involved, these events say that getting the matrix right is hard. Really hard. Probably too hard for yhe smartest people in the world. Unless the collective knowledge is much greater now than then, a difficult argument to make, it implies that one should be skeptical of claims to have solved this problem.

Then it's clear you haven't skimmed any of the relevant papers, because they're claiming something worse than independence: negative correlation. They show about -0.60 between value and momentum, with a t-stat of -13.14 on the monthly data series. Spuriously significant? You tell me.

Re: QSPIX - thoughts on interesting fund

Posted: Mon Aug 24, 2015 10:54 am
by sharpjm
I have glanced at this thread a couple of times, but haven't paid much attention to the discussion for awhile.

My understanding is that this fund has little or no correlation with traditional asset classes, is that correct?

If so, it seems like a great time to "judge" this fund.

What is its 5 day return? 1 month return? How does it compare to VTI, BND, etc during this timeframe?

I guess we will have to wait and see what its closing price is today.

Re: QSPIX - thoughts on interesting fund

Posted: Tue Aug 25, 2015 8:44 am
by lack_ey
sharpjm wrote:I have glanced at this thread a couple of times, but haven't paid much attention to the discussion for awhile.

My understanding is that this fund has little or no correlation with traditional asset classes, is that correct?

If so, it seems like a great time to "judge" this fund.

What is its 5 day return? 1 month return? How does it compare to VTI, BND, etc during this timeframe?

I guess we will have to wait and see what its closing price is today.

Charts.
http://quotes.morningstar.com/chart/fun ... A%5B%5D%7D

QSPIX = AQR Style Premia Alternative
VTSAX = Vanguard Total (US) Stock
VTIAX = Vanguard Total International (ex-US) Stock
VBTLX = Vanguard Total (nominal, taxable, investment grade) Bond
VTABX = Vanguard Total International (ex-US, USD hedged) Bond
DBC = PowerShares DB Commodity Tracking

DBC excluded from "since inception" graph because value is down 43%, making other curves hard to distinguish.


Since inception:
Image

Last month:
Image

Last several days:
Image

Last year's hiccup:
Image


Correlation matrix (daily return series):
Image
https://www.portfoliovisualizer.com/ass ... +VTABX+DBC


Based on just the data you can't really say much at this point except that
(1) volatility seems consistently higher than the broad bond markets
(2) correlations overall seem low for the period as a whole

Re: QSPIX - thoughts on interesting fund

Posted: Tue Aug 25, 2015 10:14 pm
by matjen
Ben Carlson is one of my favorite financial bloggers. I believe he is also a forum member actually. He had a nice post up today about volatility, correlations and corrections.

You can find it here: http://awealthofcommonsense.com/when-co ... go-to-one/

It included this chart regarding what he has termed the May to August Correction:

Image

And the following:
"I’m sure there are some select strategies that are holding up during this downturn, but from a traditional perspective, bonds are really the only assets treading water."


During this same period QSPIX is actually up. 10K invested on 5/1 would be worth $10,570 after today's close.

Re: QSPIX - thoughts on interesting fund

Posted: Wed Aug 26, 2015 11:08 am
by grap0013
matjen wrote:And the following:
"I’m sure there are some select strategies that are holding up during this downturn, but from a traditional perspective, bonds are really the only assets treading water."


During this same period QSPIX is actually up. 10K invested on 5/1 would be worth $10,570 after today's close.


I can unequivocally state I have more money with QSPIX than I would have had without it and it has reduced downside volatility. Plus it's fun to look at in the evening and compare to the rest of my portfolio. "My precioussssss......"

Re: QSPIX - thoughts on interesting fund

Posted: Wed Aug 26, 2015 12:08 pm
by afan
lack_ey wrote:
if you treat all four styles individually as significant or close to it then it is hard to imagine a scenario where you'd be skeptical of any combination of them, especially with the correlations seen.


Actually, this situation comes up all the time. This common misunderstanding is why some people are so enamored of approaches like stepwise regression, and why it can lead you so far wrong. Not likely to be addressed in an intro course in statistics, but certainly would be covered in a treatment of data analysis and regression...

Re: QSPIX - thoughts on interesting fund

Posted: Wed Aug 26, 2015 12:13 pm
by afan
needtosave wrote:If Bonferroni is so well known, then it's a mystery why it or some similar statistical adjustment isn't used more in investing literature.


This was addressed in several of the papers already cited in this thread.

needtosave wrote:Of all the fixes for multiple hypothesis testing, out of sample testing is probably the crudest tool because it doesn't take into account the number of hypotheses tested.

Just not true.

Re: QSPIX - thoughts on interesting fund

Posted: Wed Aug 26, 2015 12:51 pm
by lack_ey
afan wrote:
lack_ey wrote:
if you treat all four styles individually as significant or close to it then it is hard to imagine a scenario where you'd be skeptical of any combination of them, especially with the correlations seen.


Actually, this situation comes up all the time. This common misunderstanding is why some people are so enamored of approaches like stepwise regression, and why it can lead you so far wrong. Not likely to be addressed in an intro course in statistics, but certainly would be covered in a treatment of data analysis and regression...


Could you elaborate? I'm not seeing it in this context, though I certainly don't have the background and probably not the smarts for this.

But maybe we're not quite talking about the same thing. Let us assume that we think processes A, B, C, and D all each have positive mean (at a significant level), that the expected value is positive*. Also, they have roughly the same standard deviations and distributions that aren't basket cases. So you're saying that a combination 0.25*(A + B + C + D) won't have positive expected value? How? Or even any weighted combination of the four? Does this even require another statistical test?

*perhaps there are some loose assumptions of processes being stationary as well to some degree

Whether or not the combination is better than the individual parts is much harder to determine and would take a lot of convincing, but supposing the individual parts work, in which ways would some combinations not?

Re: QSPIX - thoughts on interesting fund

Posted: Wed Aug 26, 2015 1:51 pm
by sharpjm
lack_ey wrote:
sharpjm wrote:I have glanced at this thread a couple of times, but haven't paid much attention to the discussion for awhile.

My understanding is that this fund has little or no correlation with traditional asset classes, is that correct?

If so, it seems like a great time to "judge" this fund.

What is its 5 day return? 1 month return? How does it compare to VTI, BND, etc during this timeframe?

I guess we will have to wait and see what its closing price is today.

Based on just the data you can't really say much at this point except that
(1) volatility seems consistently higher than the broad bond markets
(2) correlations overall seem low for the period as a whole

Thank you for putting those charts together and posting. I certainly didn't expect a reply with that level of detail. I agree it is tough to draw an firm conclusions at this point.

Re: QSPIX - thoughts on interesting fund

Posted: Wed Aug 26, 2015 3:38 pm
by Random Walker
Grap0013,
I agree completely. The lack of correlation satisfies my gambling instinct every night as I compare how my funds have done relative to the S&P 500.

Dave

Re: QSPIX - thoughts on interesting fund

Posted: Wed Aug 26, 2015 4:02 pm
by afan
lack_ey wrote: Let us assume that we think processes A, B, C, and D all each have positive mean (at a significant level), that the expected value is positive*. Also, they have roughly the same standard deviations and distributions that aren't basket cases. So you're saying that a combination 0.25*(A + B + C + D) won't have positive expected value? How? Or even any weighted combination of the four? Does this even require another statistical test?


OK, I don't know what "distributions that aren't basket cases" means, so I cannot respond to that part.

As for the rest, the problem is you are making a couple of assumptions that basically determine the reliability of the expectation. I keep bringing up LTCM because they did exactly this. They had a set of bets, I think it is often described as 4 general categories, but it seems there were many individual bets within these. Each of these categories of bets was expected to have a positive return. The fund blew up because these 4, or whatever, bets that they thought were not highly correlated turned out to all go bad in response to the same event. By design, the managers thought, these factors were not highly correlated, and one event should not have trashed the portfolio.

I don't know whether they assumed the bets were independent, or they assumed correlations among them. In either case, they assumed they knew what the correlations were. But they were wrong.

If you mix several variables together, the outcome depends on how the variables covary. The simplest case is that the correlations are zero and the variables are independent. But in the case of financial markets and extreme events, you don't know whether the correlations are zero. You might have observed that the correlations were close to zero over certain periods of time. But what does that tell you about their behavior when something extreme happens? Apparently, not much. Remember, you can model the likelihood of any particular event as extremely low, but collectively unlikely things happen all the time.

To believe that your portfolio has controlled its risk through betting on a variety of factors, you have to believe that you know the entire variance covariance matrix of the bets. You have to believe that you know it from "normal times" based on a look back at available data. But you also have to believe that you know what these covariances do in the extreme. You don't know what these extreme events will be, how many there are that can affect your portfolio, or how your investments will behave when one of these extreme events happens.

That is why I think it will take a long time- perhaps decades, to find out what extreme events could affect this mix of bets. Of course, if you change the mix of bets, then you change the sensitivity to particular extreme events, making some less important and others more important.

These problems apply just as much to a 3-fund portfolio. But with a 3FP there are plenty of data, decades, from after people started investing that way. They are 3 extremely liquid markets and by having only 3 bets, you have a far smaller matrix to estimate. As you add bets, your covariance estimation task gets much more complex, it is harder to check the accuracy of so many predictions against actual data. Because there will be many extreme events that might affect any particular relationship, you then have to decide how much attention to pay to errors you detect. Was this some insignificant, once in a lifetime, coincidence that did not really challenge the reliability of your estimate? Was it proof that you were fundamentally wrong in your analysis? Was it somewhere in between? For the 3 fund portfolio you have a simpler job and a lot of data. That by no means exhausts the universe of possible extreme events, but it gives you much more information about extreme events than AQR has for QSPIX or the managers of LTCM had for their strategy.

And all of the above ignores that that the LTCM and QSPIX were paid for running the fund. So they had/have ever reason to paint a rosy picture of how things were expected to turn out. Assume no dishonesty, but through simple human nature, it would be easy to conclude that errors that turn up are insignificant, and to pick data ranges, sources and analysis methods that would encourage people to invest.

Remember, it seemed like a great idea, and it worked spectacularly for a while. It worked right up to an extreme event. Then it did not work so well.

Re: QSPIX - thoughts on interesting fund

Posted: Wed Aug 26, 2015 4:34 pm
by lack_ey
Okay, forget the basket case distributions. I just meant they're not incredibly far from normal, so there couldn't be some kind of very contrived counterexample involving specific mathematical relations. Though all of that was an unnecessary condition anyway, if we're just talking about first-order statistics.

In general, yes, it matters what the correlations are. But if the expected value of all four things is positive, any linear combination of the four is going to have a positive expected value too, no matter the relationship and correlations between those four things. Even if A = B = C = D and they're 100% dependent. Expectation is a linear operation.

That was my point, anyhow. I said that if you assume all four are significant, it's hard to come up with a scenario where you don't think the combination is significant. So if you have a problem with the whole, it pretty much has to be because of problems with individual styles.* So the question is then which styles you have a problem with. (There are some reasonable questions to be asked here, but you are still light on specifics.)

*Or maybe the amount of leverage combined with interactions between components from sequence-of-returns kind of risk. I guess now it's clearer you're more concerned with tail risks than anything else? Again, the bets are a whole lot more diversified and the leverage several times less than what LTCM was using, so I don't see why everybody keeps bringing that up. There's a concern but can you quantify it or express it more clearly? Does the low volatility version, which is about half as levered, look markedly less risky to you? What would you think about a very low volatility version that's not levered at all? (Assume the fees are now low enough such that the expected return is still positive.) Or a long-only version that skips the shorts? Would you still have the same concerns, or am I misinterpreting?

Re: QSPIX - thoughts on interesting fund

Posted: Wed Aug 26, 2015 5:05 pm
by Johno
afan wrote:
lack_ey wrote: Let us assume that we think processes A, B, C, and D all each have positive mean (at a significant level), that the expected value is positive*. Also, they have roughly the same standard deviations and distributions that aren't basket cases. So you're saying that a combination 0.25*(A + B + C + D) won't have positive expected value? How? Or even any weighted combination of the four? Does this even require another statistical test?

As for the rest, the problem is you are making a couple of assumptions that basically determine the reliability of the expectation. I keep bringing up LTCM because they did exactly this. They had a set of bets, I think it is often described as 4 general categories, but it seems there were many individual bets within these. Each of these categories of bets was expected to have a positive return. The fund blew up because these 4, or whatever, bets that they thought were not highly correlated turned out to all go bad in response to the same event. By design, the managers thought, these factors were not highly correlated, and one event should not have trashed the portfolio.

I don't know whether they assumed the bets were independent, or they assumed correlations among them. In either case, they assumed they knew what the correlations were. But they were wrong.

The question seemed to once again, though more patiently than I, I know :D , to once more seek from you any real specifics of what you think are the flaws in either history of research into 'style premia' or particular construction of QSPIX. But again, you don't.

LTCM had two dominant positions at the time it imploded: short equity option implied volatility and 'long' interest rate swap spreads (meaning you receive fixed, pay floating and sell govt bonds to hedge, by market convention that's called 'long' spreads even though you lose money when spreads go up) at an overall leverage ratio said to have been ~25:1. And there was no reasonable doubt how either position had correlated big sells offs in risk markets generally: equity option vol and swap spreads would go up*. In general, you make money being short equity option vol (there's a consistent risk premium in implied vol compared to realized over time, especially in below-the-money strike options); there's a positive carry in being long swap spreads too but mainly the idea is to go 'long' when they are high hoping they go down. Nor did LTCM have on any other big position they could reasonably have thought would be offsets to those.

That's completely different in specifics from QSPIX which besides only 5-ish leverage is seeking to harvest 4 different premia (value, momentum, carry, defensive) in 4 different markets (bond, short fwd rate, currency, stock) against a historical record of low correlation among those premium/market combinations. So there's no specific usefulness to the LTCM comparison, and we're left again with just a vague general statement, 'we don't know what the correlations of those 16 pairs will be to one another or the equity risk premium (important since the fund is intended to at least slightly diversify a portfolio from overdependence on ERP)'. No of course we don't know, along with not knowing anything else for sure about the future.

Like the argument about expected return of the premia an argument for high uncertainty about the correlation of the premia needs to be backed by specifics, not references to general principals, and not comparisons to completely different situations. And in contrast to the expected return debate, there's a strong common sense element in the correlation debate. Will carry in currencies plus 'defensive' in short fwd rates plus momentum in equity, etc. produce a consistently positive real return after expenses? IMO it's hard to say in common sense terms, though relatively compelling statistics indicate so in the world of the past at least. But will carry in currencies, defensive in short fwd rates, momentum in equity, etc. and the ERP all start to correlate closely? That has more of a common sense element: why would they? And again that's nothing to do with whether you get crushed if massively short equity option vol and long swap spreads and a major sovereign debt crisis erupts, where the answer is very easy: yes, you do.

*first is as true now, look at the VIX past week; USD swap spreads OTOH are distinctly lower and less volatile in bps in the post 2009 environment than previously; it seems to be in the 'world can change' category, though it could of course change back.

Re: QSPIX - thoughts on interesting fund

Posted: Wed Sep 02, 2015 12:06 pm
by oneleaf
Looking into this fund more, I feel like the momentum across different asset classes is the most attractive diversifier esp for my value tlted portfolio. I am considering adding a managed futures fund. Pimco has a share class PQTIX which has a 1.1 ER and invests in trendfollowing strategies in currencies, commodities, interest rate, and equities. Any thoughts on the diversification benefit?

Re: QSPIX - thoughts on interesting fund

Posted: Wed Sep 02, 2015 3:23 pm
by garlandwhizzer
There is no doubt that QSPIX has done well relative to US and INTL equity during its short lifetime in these volatile markets. Lets compare its results to another fund that invests globally with the aim of long term capital gains and lowering overall portfolio volatility, Vanguard's Global Minimum Volatility Fund, VMVFX. Like QSPIX, VMVFX has too short a lifespan to fully evaluate its potential in all market cycles. It uses only one factor, low volatility, an unpopular choice among factor enthusiasts relative to say V, MOM, and S. It employs no leverage and is actively managed. In spite of these shortcomings it has not only outperformed QSPIX consistently since its inception but it has done so with lower volatility. It's strategy is much simpler, using a single factor largely discounted by many factor investing proponents (Larry among them). Its expense ratio in Admiral shares is 0.20%, 1.30% lower than QSPIX. Actually I believe that the fact that many factor advisors and investors prefer factors like momentum, value, and size in preference to plain old boring low momentum is a huge plus for this fund. Larry has made the point that trading costs in low volatility portfolios substantially reduce its gain, citing the paper The Limits to Arbitrage and the Low Volatility Anomaly. Other papers however dispute this point and it seems to me that the question is not clearly settled. AQR discounts low vol, while RAFI is enthusiastic about it. The other point that low vol detractors make is that the only meaningful outperformance comes from the lottery effect which involves only the most volatile quintile or decile of stocks. Throw those extremely volatile stocks away and the premium largely disappears they argue. However, RAFI quintile charts from 1963 to 2008 demonstrate that the lowest beta quintile outperforms the second lowest beta quintile by 1% annually over that 45 year period, much less than the 4% annual outperformance over the highest beta quintile but still highly significant. So, if trading costs can be controlled which they can in an actively managed fund that doesn't have to sell according to a fixed formula, it is entirely possible that this fund can deliver not only lower volatility a higher Sharpe ratio, and lower overall portfolio volatility but also even outperformance.

QSPIX is one approach to lowering overall portfolio volatility with options, factors, and leverage while maintaining positive returns, thus improving Sharpe ratio for the portfolio as a whole. It is not, however, the only way. There are less expensive and less complex approaches to consider as well.

Garland Whizzer

Re: QSPIX - thoughts on interesting fund

Posted: Wed Sep 02, 2015 3:55 pm
by larryswedroe
RE LTCM
LTCM blew up because they actually ignored the evidence and simply started making directional bets, like going long Brazilian debt and Russian debt, thinking they were uncorrelated perhaps. But this had nothing to do with the premise the fund was founded on, arbitraging anomalies. And then they applied 100x leverage which means even 1% moves against you and you can die. QSPIX is nothing like this. There is systematic exposures to factors that are well documented, not personal bets made by some managers and very minor leverage.
there is simply no logic to comparing these two IMO. NONE. They are very different animals
BTW-note that if LTCM actually could have waited out their positions (if they had used no or limited leverage) they would have been fine.But that's not all that relevant (except to show that it was leverage that killed them, not their positions), because what killed the fund was ACTIVE positions, not betting on factors or anomalies unwinding.
Larry

Re: QSPIX - thoughts on interesting fund

Posted: Wed Sep 02, 2015 4:02 pm
by larryswedroe
Here's piece I dug up that explains what really happened at LTCM

When Genius Failed

Many individual investors have mistakenly viewed the disastrous failure of the highly publicized hedge fund Long Term Capital Management (LTCM) as the failure of modern financial theory. In fact, nothing could be further from the truth. The mistake is made because two of the most important members of LTCM were Nobel Prize winners in the field of financial economics, Robert Merton and Myron Scholes. The fund also had on its staff many financial economists and mathematicians as they sought to exploit “market inefficiencies.” The combination of the confidence (which proved to be misplaced) that investors and bankers placed in the fund’s management and the overconfidence and hubris of the company’s management eventually led to a near calamity for the financial markets, one that threatened to severely damage the banking system. The Federal Reserve came in at the eleventh hour to negotiate a bailout that, while killing the hedge fund, prevented serious damage to the capital markets. Roger Lowenstein’s When Genius Failed, The Rise and Fall of Long Term Capital, is a wonderful account of the failure not of modern financial theory, but is instead an account of both genius gone astray and just how efficient the markets really are.
There are many lessons to be learned from the failure of LTCM. The firm made many mistakes that were easily identified before the fact. It is not necessary to use Monday morning quarterbacking as all the mistakes could, and should, have been foreseen. Let’s review some of the major issues and the errors that led to the downfall of one of the great (though very brief) success stories of our times.
1. Markets are efficient. This was one of the major principles that was held by the firm and certainly held by its Nobel Prize winners. Not only did they believe that markets were efficient, but they also knew that markets tend to become more efficient over time. The reason is that as inefficiencies are uncovered, the very act of exploiting them will make the market more efficient. Given the huge profit opportunities available, the speed with which the financial markets eliminate an inefficiency is not surprising. Another way to think about this is that in free and efficient markets success is punished. If a firm can make “abnormal” profits by exploiting an inefficiency, competitors will quickly “reverse engineer” the trade to figure out how the abnormal profits were made. Competition will then enter the market and eliminate the inefficiency.
LTCM identified several market anomalies (or inefficiencies). For example, they noted that the newest 30-year treasury bond traded much richer (higher price) than the previous issue (until recently the treasury had held bi-annual auctions). This didn’t make sense. The 30-year was not only very similar an instrument, but it was also theoretically even slightly riskier given its longer maturity (by 6 months). The reason for the premium was that the most recent issue is always the most liquid (most highly traded), and investors are willing to pay a premium for liquidity. Whenever the spread widened a bit beyond its historical range the firm would go long the bond with the higher yield (the older less expensive bond) and sell short the more expensive lower yielding newer issue. Note that the firm took no bet on interest rates as it was both going long and short instruments of the same credit quality and basically the same maturity. Eventually the spreads would narrow (due to market efficiency) and the firm would unwind the trade at a profit. Given that the spreads never widened that much the firm had to both take huge positions and use lots of leverage to achieve great returns to their investors. The problem was that the very efficiency of the market they counted on came back to haunt them. Competitors quickly copied their success. Billions of dollars from competing firms began to chase the same spread opportunities—and the size of the spreads that they had been exploiting began to narrow. Lured by the scent of huge profits, by the late 1990s almost every investment banking firm had its own version of LTCM. And, of course, the profit opportunities diminished. In order to continue to earn the same returns for their investors the firm would have to take on ever-greater sized positions and use more and more leverage. More on this later.
2. You can reduce portfolio risk by diversifying across non- or low-correlating asset classes. This was Harry Markowitz’s greatest contribution to Modern Portfolio Theory. LTCM owned many positions across the globe. They had deployed their hedging strategy across the globe. Thus they believed that they had built a portfolio with many low correlating positions. For example, what did positions in Italy or Russia have to do with a bet on U.S. treasury spreads? Unfortunately for LTCM, they made one huge error—most of their trades, despite being geographically diverse, were all bets on credit/liquidity spreads narrowing. If there were a global crisis (like the one in not too long ago 1987) that would lead to a global flight to quality, all of their bets would go in the wrong direction at the same time. Thus if ever there was a global flight to quality all of their trades would be highly correlated and in the wrong direction. This is exactly what happened when the Asian crisis of the summer/autumn of 1998 began. While each trader was taking a position that seemed independent (and low correlating) of the others, the firm failed to exercise good risk management at the portfolio level.
3. The use of leverage means that you might be right in the long-term, but you might also be dead in the short-term. Leverage works two ways. It magnifies both gains and losses. And, unfortunately, if there are losses in the short term, you may be forced to meet margin calls as the value of your collateral (on which the margin loan is based) shrinks making the lenders nervous about getting repaid. Any experienced investor knows that the danger of using leverage is that even if you are right in the long term, if you are wrong in the short term and can’t meet your margin calls you will not survive to reap the profits. This lesson was one that LTCM simply forgot. As profit opportunities began to shrink, the firm took on more and more leverage. Eventually the markets went against them. Previously, when their positions were smaller, they could simply wait out the period by coming up with more collateral to meet the margin call. Unfortunately both the size of the market’s move and the amount of leverage deployed made meeting margin calls impossible and the firm was forced to liquidate positions at the worst possible time, driving prices against themselves even further as they unwound their positions (they had to buy back securities they had previously shorted, thus driving the prices even higher). Eventually, the losses overwhelmed their ability to raise collateral and the banks were calling in their loans. One other important piece of information that LTCM forgot. The financial markets are probably the most competitive of all markets. When competitors smell blood they will rush in profit from the kill. Wall Street is a relatively small community with a great grapevine. The other firms knew LTCM was in a desperate situation. With this knowledge traders began to drive the market against LTCM’s positions. They did so in the almost certain knowledge that LTCM would then be forced to turn back to the market and buy their positions out at a profit. In the end the positions that LTCM took turned out to be correct as spreads eventually narrowed. Unfortunately for LTCM they were basically dead and the “profits” went to the group that bailed them out. Right in the long term, but dead in the short term.
4. 100-year floods sometimes happen with much greater frequency (just ask any Missouri resident). The firm relied very heavily on their financial models that made the assumption that returns were normally distributed (like the bell curve). Unfortunately for LTCM that assumption was false, and they should have known that. The distribution of returns in the financial markets is not bell-shaped. In fact, the distribution of returns has what are called fat-tails. That means that there are likely to be more periods of both exceptionally high and exceptionally low returns than a bell curve would predict. The best and most recent example was the crash of 1987. As Lowenstein pointed out (p.71) Eugene Fama, Myron Scholes thesis advisor at the University of Chicago, had done a study on the distribution of stock returns. Here is what Fama found: “If the population of price changes is strictly normal, on the average for any stock…an observation that is more than five standard deviations from the mean should be observed about once every 7,000 years. In fact such observations seem to occur about once every three or four years.” Lowenstein points out that by the time LTCM was founded it was well documented that virtually all financial assets behaved like the stocks Fama had studied. If you get the fat tail in the wrong direction AND you are highly leveraged and thus cannot wait for the markets to possibly recover, financial death is highly likely. LTCM not only didn’t anticipate the 100-year type flood, it levered up so highly it couldn’t survive one if it occurred even once. LTCM would have done well to heed the advice of John Maynard Keynes: Markets can remain irrational longer than you can remain solvent (p. 123).
5. They confused risk with uncertainty. LTCM treated the output of their financial models (which had databases that had very short lives) as if there was no uncertainty, just risk. In other words, their models anticipated all the potential trading environments. They could then determine the exact likelihood of possible outcomes. They didn’t anticipate that there might be environments outside of the model’s database. Thus they were unprepared for them. They simply made the mistake of confusing risk with uncertainty. For example, an insurance company might be willing to take on a certain amount of hurricane risk in Dade and Broward Counties in Florida. They would price this risk based on maybe 100 years of data on the likelihood of hurricanes occurring and the damage that they did. But only a foolish insurance company would place such a large bet that if more/worse hurricanes were to occur than had previously been experienced they would go bankrupt. That would be ignoring uncertainty—the future might not look like the past. LTCM modeled risk based on their databases, priced for that risk, and ignored the possibility that the world might look worse going forward than it had ever done. In other words, their models would work; until they came across an environment they had not model (uncertainty) and then they would no longer work. LTCM also ignored the most basic of all risk management principles: never make the bet that if you are wrong you don’t get to play again because you are dead.
6. They assumed markets would always be liquid. This assumption was key because if their positions started to go against them they might have to unwind them to meet margin calls. But what if the markets were to become illiquid and LTCM were forced to unwind huge positions in order to meet margin calls? The result might be financial ruin. The potential for this to occur should have been foreseen by LTCM as it had occurred less than 10 years earlier in the crash of 1987 when believers in what was known as portfolio insurance learned that markets are not always liquid and you cannot always get out of positions, especially at “market prices.”
7. Hubris can be a deadly sin. LTCM was a victim of its own success. They delivered such great returns (with seemingly great ease and consistency) that they began to believe they could do no wrong. This encouraged them to not only take much greater risks (bigger bets with more leverage) but they also began to take risks of the type that they have never engaged in previously. Instead of simply making bets on spreads narrowing or tightening they began to take outright positions on securities. They even had short positions on stocks such as General Electric, Dell and Microsoft (p.128). This is a much more risky proposition, especially when you add huge amounts of leverage. It was also not what the firm had told either its investors or its bankers (the providers of leverage) it would do. It also had nothing to do with modern financial theory. In fact, it went against everything modern financial theory believes in. LTCM’s short positions in these stocks were simply a bet that the market was inefficient and had somehow mispriced the securities. As one example, in they took a huge position of almost $500 million in the junk-bond issue of Starwood Hotels and Resorts (p.129). Clearly this was well beyond what the firm had positioned as its mandate. The traders simply began to believe that they were infallible.
I believe that the following analogy that sums up the rise and fall of LTCM. LTCM used the best technicians to build them the fastest racing car in the world. They then gave that car to a bunch of drunk teenage drivers. The fault was not in the design of the engineers. Instead, the fault was in how the elegant design was used. The company simply exposed itself to unnecessary risk. While you can survive making big bets, and you can survive being highly leveraged, but if you do both the only way to survive is to be right EVERY SINGLE DAY. One bad day, and you are dead. In the end, it was the efficiency of the market and the hubris of the firm’s partners that led to the death of genius.

I hope this is helpful
Larry

Re: QSPIX - thoughts on interesting fund

Posted: Wed Sep 02, 2015 4:09 pm
by lack_ey
garlandwhizzer wrote:[...]AQR discounts low vol (emphasis added), while RAFI is enthusiastic about it. The other point that low vol detractors make is that the only meaningful outperformance comes from the lottery effect which involves only the most volatile quintile or decile of stocks. Throw those extremely volatile stocks away and the premium largely disappears they argue.[...]

Are you sure?

Their quality/defensive is low volatility, low beta, though extended here perhaps bizarrely (or not) to bonds as well as stocks and equity indexes. Because the style premia fund (QSPIX) is long-short, it shorts high beta stocks, so if the so-called anomaly exists mostly as the top quintile/decile being the worst, then that will be captured here.
Defensive, or low beta/low risk, is a strategy with a long pedigree that has experienced a resurgence in recent years. The initial motivation for
defensive strategies dates back to Fischer Black, who in 1972 saw that the security market line in U.S.equities was too flat[...]


And yes, there are many cheaper ways to get partial exposure to many of the components here. I personally don't feel that comfortable with a lot of factor tilts in taxable accounts and have limited tax-advantaged space to work with now; I mostly like the fund for broad coverage of different strategies and high density per dollar invested of these alternate potential sources of return. The very high expense is a bitter pill to swallow for all that.

Personally I'm not really looking for lowering portfolio volatility, though that would be nice. I just want other shots at making money if stocks are flat or down.

Anyway, low/min vol funds and certainly backfilled indexes have existed for quite a few more years than just Vanguard's fund.

Re: QSPIX - thoughts on interesting fund

Posted: Wed Sep 02, 2015 4:37 pm
by Johno
larryswedroe wrote:RE LTCM
LTCM blew up because they actually ignored the evidence and simply started making directional bets, like going long Brazilian debt and Russian debt, thinking they were uncorrelated perhaps. But this had nothing to do with the premise the fund was founded on, arbitraging anomalies. And then they applied 100x leverage which means even 1% moves against you and you can die. QSPIX is nothing like this. There is systematic exposures to factors that are well documented, not personal bets made by some managers and very minor leverage.
there is simply no logic to comparing these two IMO. NONE. They are very different animals
BTW-note that if LTCM actually could have waited out their positions (if they had used no or limited leverage) they would have been fine.But that's not all that relevant (except to show that it was leverage that killed them, not their positions), because what killed the fund was ACTIVE positions, not betting on factors or anomalies unwinding.
Larry

I generally agree. As mentioned above, LTCM's principal loss positions have usually been reported as being short equity option vol and long swap spreads (meaning you lose money when spreads rise) and leverage more like 'only' 30:1, explicit anyway. Prabhu (2001) estimated those two categories as 64% of LTCM's losses, but emerging market debt was for example another 10%, and it all tended to go in the same direction. Though people will often give LTCM as example of 'correlations being unpredictable' and there's some truth to that (correlations are indeed unpredictable), anyone could tell you swaps spreads and equity option vol would blow out in a serious flight to quality, along with emerging market debt yields, etc. The headline of 'model risk' while not entirely irrelevant is somewhat misleading. 'Very big, risky and known to be generally similar positions blow up in a crisis' is more accurate for LTCM.

As you say, this has basically no applicability to QPSIX other than via the background agency risk in any fund, that the management firm could be doing something completely different than what they say they will, or their employees could, or could just make costly mistakes doing other than what they intended to do. But that sort of risk is again just not very relevant to LTCM. The positions which blew them up were not really the signature type of the fund originally, 'yield curve arbitrage' and 'convergence' except by stretching those terms, but LTCM wasn't breaking any rules changing to huge short volatility positions (short equity option vol, long spreads swap and emerging...everyone knew those would all go south in a general pick up in market volatility): it was a hedge fund. Totally different situation.

Re: QSPIX - thoughts on interesting fund

Posted: Wed Sep 02, 2015 7:29 pm
by afan
larryswedroe wrote:They then gave that car to a bunch of drunk teenage drivers.

Wow! Talk about hindisght hubris. Bad as the outcome was, these were not "drunken teenage drivers", these were some of the leading financial economists in the world.

I find it hard to accept these retrospective declarations that the problems were "obvious". If they were, then explain how people of this quality completely missed them?

The problem arose because they thought they knew the covariances of the bets they made. They thought that their bets were highly diversified and uncorrelated.

They were wrong.

They based their estimates on limited data sets. They thought these data were enough to predict the behavior of the bets when extreme events occurred.

They were wrong.

The use of leverage mattered only because they were wrong. Had they been right, the leverage would not have been a problem.

The solution for people who want to do this sort of investing is simply to find managers who are a lot smarter than these Nobel laureates. How hard can that be?

Or accept that market prices are the best estimate of expected risk adjusted return, buy the market, and save the turnover, transaction costs, tracking error and expense ratios associated with the LTCM/QSPIX approach.

Re: QSPIX - thoughts on interesting fund

Posted: Wed Sep 02, 2015 7:38 pm
by grok87
garlandwhizzer wrote:There is no doubt that QSPIX has done well relative to US and INTL equity during its short lifetime in these volatile markets. Lets compare its results to another fund that invests globally with the aim of long term capital gains and lowering overall portfolio volatility, Vanguard's Global Minimum Volatility Fund, VMVFX. Like QSPIX, VMVFX has too short a lifespan to fully evaluate its potential in all market cycles. It uses only one factor, low volatility, an unpopular choice among factor enthusiasts relative to say V, MOM, and S.

i also like VMVFX. But i think its story/strategy is actually a little bit different than you may think. Here is my take of the different ways to play low volatility:

1) Low Volatility: invest in low volatility stocks. ie stocks whose trailing returns have shown low standalone volatility, ie low price swings. Generally big safe companies. the approach tends to cluster in certain industries like utilities etc. so you actually end up with a fair amount of sector risk

2) Low beta: invest in stocks with low beta compared to the broad market index- e.g. S&P 500. Stocks that tend to go their own way. These stocks could actually be very volatile- just not correlated to the market. Think biotech stocks whose fate hinges on a single experimental drug getting FDA approval or something like that...

3) global minimum volatility. like the vanguard fund and also ishares has a suite of funds.ACWV for instance. these funds do something different than 1) or 2) that is a little less easy to explain. basically they try and build portfolios out of stocks where the portfolio has the LOWEST possible volatility. I think the starting point may well be low beta stocks. But the funds try to match the broad market index in terms of sector (and country?) weights to eliminate unwanted biases and tilts that are present in 1) and 2) (avoiding the utility bias of 1) for example). and out of the low beta stocks they pick the ones that are least correlated with each other to drive down the overall portfolio volatility. and the stocks themselves can't be too volatile either i bet. So again a totally different approach.

Re: QSPIX - thoughts on interesting fund

Posted: Wed Sep 02, 2015 7:53 pm
by grok87
I haven't read this whole thread, so no doubt the leverage of the fund has already been discussed. But just to reiterate for those looking at this thread new: QSPIX is about 7x levered (about 3.5 times per side). In other words for every 1 dollar you give them them will deploy 3.5 dollars into "long bets" and 3.5 dollars into "short bets"

see the fact sheet
https://funds.aqr.com/our-funds/alterna ... ative-fund

to my mind that is a relatively high degree of leverage that i am not personally comfortable with...

Re: QSPIX - thoughts on interesting fund

Posted: Wed Sep 02, 2015 8:16 pm
by lack_ey
grok87 wrote:I haven't read this whole thread, so no doubt the leverage of the fund has already been discussed. But just to reiterate for those looking at this thread new: QSPIX is about 7x levered (about 3.5 times per side). In other words for every 1 dollar you give them them will deploy 3.5 dollars into "long bets" and 3.5 dollars into "short bets"

see the fact sheet
https://funds.aqr.com/our-funds/alterna ... ative-fund

to my mind that is a relatively high degree of leverage that i am not personally comfortable with...


I agree to some extent, but do note that a lot of the leverage is being used for the bonds/interest rates segment, assets that don't move that much. With these taken out and other positions scaled up to bring it back to the ~10% yearly standard deviation target, the net leverage would be considerably less. It's nothing close to being for example 3.5x long and short stocks, or commodities, or something like that. Still pretty risky, but maybe not as much as it looks on the surface.

Yes, the positions could go really bad, but take a look at the holdings and you see something like ~40% notional exposure to 90-day Eurodollar futures and other positions in short-term rates with large notional denominations (making up a nontrivial chunk of that ~700% net). That almost might as well not even exist.

And again, if somehow you like the strategy but are freaked out by the amount of leverage, there's the half-as-levered low volatility version. I don't think it's a great idea, but if you're imagining doomsday sequence-of-return nightmares with the underlying holdings combined with just too much leverage, it would be less affected?

Re: QSPIX - thoughts on interesting fund

Posted: Wed Sep 02, 2015 8:33 pm
by grok87
lack_ey wrote:
grok87 wrote:I haven't read this whole thread, so no doubt the leverage of the fund has already been discussed. But just to reiterate for those looking at this thread new: QSPIX is about 7x levered (about 3.5 times per side). In other words for every 1 dollar you give them them will deploy 3.5 dollars into "long bets" and 3.5 dollars into "short bets"

see the fact sheet
https://funds.aqr.com/our-funds/alterna ... ative-fund

to my mind that is a relatively high degree of leverage that i am not personally comfortable with...


I agree to some extent, but do note that a lot of the leverage is being used for the bonds/interest rates segment, assets that don't move that much. With these taken out and other positions scaled up to bring it back to the ~10% yearly standard deviation target, the net leverage would be considerably less. It's nothing close to being for example 3.5x long and short stocks, or commodities, or something like that. Still pretty risky, but maybe not as much as it looks on the surface.

Yes, the positions could go really bad, but take a look at the holdings and you see something like ~40% notional exposure to 90-day Eurodollar futures and other positions in short-term rates with large notional denominations (making up a nontrivial chunk of that ~700% net). That almost might as well not even exist.

And again, if somehow you like the strategy but are freaked out by the amount of leverage, there's the half-as-levered low volatility version. I don't think it's a great idea, but if you're imagining doomsday sequence-of-return nightmares with the underlying holdings combined with just too much leverage, it would be less affected?

thanks that's interesting
https://funds.aqr.com/~/media/files/fac ... spalvi.pdf
qslix
only 3.5 x levered (i.e. 1.75 long and 1.75 short)

interestingly qslix (the low vol version) is up 0.59% YTD (as of 9/2/15) while qspix is up 0.4%. Not to make too much out of small numbers but i think that makes my point about leverage. It doesn't always work the way one thinks it should and it is not easy to understand. If it were simple and linear QSPIX should be up twice as much as QSLIX. ie QSPIX should be at 1.2% instead of 0.4%.

Re: QSPIX - thoughts on interesting fund

Posted: Thu Sep 03, 2015 7:55 am
by larryswedroe
afan
You are so off base it's absurd, many knew and were pointing these things out at the time, because you were unaware is irrelevant. Wall Street firms were all well aware of what was happening and actually exploiting it, by pushing prices against LTCM because they were aware they would be forced to unwind positions due to the margin calls.

LTCM also made simple directional bets, not trying to exploit some perceived anomaly or gain exposure to a well known factor. Just simple bets that they knew better, that prices were somehow wrong. (QSPIX makes no assumption about pricing being wrong)

And leverage isn't the problem, it's the EXCESSIVE use of it, they went from using relatively small amounts to about 100 counting use of derivatives--
That was the hubris.


Larry

Re: QSPIX - thoughts on interesting fund

Posted: Thu Sep 03, 2015 9:47 am
by Johno
afan wrote:
larryswedroe wrote:They then gave that car to a bunch of drunk teenage drivers.

Wow! Talk about hindisght hubris. Bad as the outcome was, these were not "drunken teenage drivers", these were some of the leading financial economists in the world.
I find it hard to accept these retrospective declarations that the problems were "obvious". If they were, then explain how people of this quality completely missed them?

...They thought that their bets were highly diversified and uncorrelated.

They were wrong.

This is simply wrong, and now a worse form since it's been explained to you fairly exactly which positions lost the bulk of money for LTCM, but you're just ignoring it and repeating that LTCM's had positions it thought or said were uncorrelated. That's simply untrue. You can easily look it up, though I think posts above might help in sifting through various jargon to the key points: 64% of losses from huge short in equity option vol and long in swap spreads. Nobody thought or said those are uncorrelated, and most of the other smaller LTCM position went in the same direction (long EM debt, long slightly less liquid forms of govt debt against short the most liquid forms, etc.). All those things will go the same way when general market hits the fan, and everyone knew as well as knows that.

So why would they have taken such large mainly one way bets and why wouldn't everyone run away from the fund if they did? Because they trusted their reading of the specific market situation to make a huge ad hoc bet. And those positions, especially at *30* time leverage, make a lot of money if the market calms down (recall this was a period of already fairly heightened vols, spreads etc following from the Asian fin crisis previous year) or even just stays the same. Volatilities and correlations of the positions (ie including volatilities of volatilies of options) did matter in LTCM's calculation of its 'Value At Risk' max loss, but that's not to be mistaken for the clearly wrong statement that 'they thought these bets were uncorrelated' which is nonsense, sorry to be blunt.

And as to 'why didn't anybody see this problem?', as Larry pointed out that's also just a false assumption. LTCM lost at one point $250mil *in one day* on just one position, short in 5yr over-the-counter equity option volatility. That was quite illiquid, and derivatives dealers were making money specifically squeezing LTCM on that position because they knew it was huge and hard to get out of. Like many such cases, a notable part of the meltdown was other speculative entities exploiting the situation, and there's nothing inherently wrong with that, all big boys and girls including the investors in LTCM, unless or until it reaches the point of putting public money at risk to stabilize the situation.

But this story, LTCM, is wholly removed from the actual risks of a ~5:1 not 30:1 leveraged mutual fund attempting to 'harvest' particular 'premia', according to a specific investment plan in writing, with apparent low correlation to the ERP (by both history and simple common sense), using liquid exchange traded futures and diversified individual stocks. We could have a useful discussion of the actual risks of a fund like QSPIX, but your now insistently misinformed harping on LTCM is just not it.

Re: QSPIX - thoughts on interesting fund

Posted: Thu Sep 03, 2015 11:32 am
by larryswedroe
Johno
And add that there is no human intervention with human's making judgments about what is a good position or a bad one--it's systematic and replicable approach--wasn't the case in LTCM which wasn't systematic at all and was employing human judgments, pure active management.
Larry

Re: QSPIX - thoughts on interesting fund

Posted: Thu Sep 10, 2015 10:32 pm
by Tier1Capital
I don't believe I saw this in the thread anywhere. Apologies if I missed it somewhere.

QSPIX is available in my Schwab 401k PCRA. $100,000 minimum initial trade amount.

Re: QSPIX - thoughts on interesting fund

Posted: Fri Sep 11, 2015 6:43 am
by matjen
Tier1Capital wrote:I don't believe I saw this in the thread anywhere. Apologies if I missed it somewhere.

QSPIX is available in my Schwab 401k PCRA. $100,000 minimum initial trade amount.


Hmmm. It is in my wife's Schwab PCRA as well but there is no minimum amount or if there is it is much, much less than that.

Re: QSPIX - thoughts on interesting fund

Posted: Fri Sep 11, 2015 9:12 am
by grap0013
Tier1Capital wrote:I don't believe I saw this in the thread anywhere. Apologies if I missed it somewhere.

QSPIX is available in my Schwab 401k PCRA. $100,000 minimum initial trade amount.


These minimums may be incorrect. My Fidelity account said 100K minimum right up until the point when I clicked to purchase it. I bought in at <100K.

Might also be able to buy QSPNX in place as well. Build up a bunch as portfolio grows and then convert over to QSPIX.

Re: QSPIX - thoughts on interesting fund

Posted: Fri Sep 11, 2015 11:16 am
by nisiprius
Johno wrote:...This is simply wrong, and now a worse form since it's been explained to you fairly exactly which positions lost the bulk of money for LTCM, but you're just ignoring it and repeating that LTCM's had positions it thought or said were uncorrelated. That's simply untrue.... So why would they have taken such large mainly one way bets and why wouldn't everyone run away from the fund if they did? Because they trusted their reading of the specific market situation to make a huge ad hoc bet...
But this in turn brings up a meta-problem. You say that the problem with LTCM is that they quit doing what they said they were doing, They told everyone they were going to safely pull nickels out of the air, and leverage up these safe bets to make dollars out of them. You say this was perfectly sound and if they'd just kept doing it, LTCM would still be making money today. But, instead, they took their investors' money to a casino and kept doubling their bets until they lost. The question then arises: did LTCM make the change in strategy clear to their investors? Did their investors say, "OK, we don't actually care about the strategy, we just believe in the cult of personality, if John W. Meriwether and two Nobel laureates say it's not risky, that's good enough for me?"

Or did they fail to parse the disclosure documents carefully enough to detect that LTCM had changed what they were doing?

Re: QSPIX - thoughts on interesting fund

Posted: Fri Sep 11, 2015 2:29 pm
by Johno
nisiprius wrote:
Johno wrote:...This is simply wrong, and now a worse form since it's been explained to you fairly exactly which positions lost the bulk of money for LTCM, but you're just ignoring it and repeating that LTCM's had positions it thought or said were uncorrelated. That's simply untrue.... So why would they have taken such large mainly one way bets and why wouldn't everyone run away from the fund if they did? Because they trusted their reading of the specific market situation to make a huge ad hoc bet...
But this in turn brings up a meta-problem. You say that the problem with LTCM is that they quit doing what they said they were doing, They told everyone they were going to safely pull nickels out of the air, and leverage up these safe bets to make dollars out of them. You say this was perfectly sound and if they'd just kept doing it, LTCM would still be making money today. But, instead, they took their investors' money to a casino and kept doubling their bets until they lost. The question then arises: did LTCM make the change in strategy clear to their investors? Did their investors say, "OK, we don't actually care about the strategy, we just believe in the cult of personality, if John W. Meriwether and two Nobel laureates say it's not risky, that's good enough for me?"

Or did they fail to parse the disclosure documents carefully enough to detect that LTCM had changed what they were doing?

"this has basically no applicability to QPSIX other than via the background agency risk in any fund, that the management firm could be doing something completely different than what they say they will, or their employees could, or could just make costly mistakes doing other than what they intended to do. But that sort of risk is again just not very relevant to LTCM. The positions which blew them up were not really the signature type of the fund originally, 'yield curve arbitrage' and 'convergence' except by stretching those terms, but LTCM wasn't breaking any rules changing to huge short volatility positions (short equity option vol, long spreads swap and emerging...everyone knew those would all go south in a general pick up in market volatility): it was a hedge fund. Totally different situation."

Your response seems to relate more the passage I just quoted from my previous post than what you quoted. But as you see, it does recognize that there is some agency risk in any fund, index funds too, or even broker as agent to do specifically what one chooses to do oneself. I however reject any special connection between LTCM and QSPIX in this regard as compared to say QSPIX and agency risk in an index fund. Clearly one can't directly compare the nature and degree of agency risk in QSPIX to that of an index fund, but I don't see why LTCM and QSPIX are any more comparable.

Hedge funds and mutual funds have very different disclosure requirements, in HF case as much a matter of custom and commercial considerations as anything else, especially at that time (what does the particular investor demand to know and how much does the fund care about satisfying them). There was no detailed disclosure document for LTCM even originally describing a very specific investment theory and plan, a targeted volatility etc. And it's open to debate whether the principals of LTCM ever misled anybody, according to the accepted ground rules that applied to them at the time. They returned significant capital to investors in '97 saying the original opportunities were not as great. Naturally some investors would later view that as misleading (some sources claim 'investors assumed the fund would reduce its risk' commensurate with the return of principal, rather than ramp it up as happened) but nobody won any suits against LTCM in that regard. It's further off on an already useless tangent, to QSPIX, to re-litigate what Meriweather told investors and their mindset about 'star power' v detailed information in their decisions to become or stay investors in LTCM. Completely different situation.

And it's quite a stretch IMO to go from my comment "The [LTCM] positions which blew them up were not really the signature type of the fund originally, 'yield curve arbitrage' and 'convergence' except by stretching those terms" to your characterization "You say this was perfectly sound and if they'd just kept doing it, LTCM would still be making money today." I assume the latter is extrapolated from the former, a huge leap, but I'm at a loss where else it could come from in anything I said about LTCM. But I never said LTCM was ever 'perfectly sound' and LTCM itself generally communicated to investors that its previous strategy had diminished potential as of 1997. This seems on the whole to effectively put words in my mouth implicitly comparing LTCM and QSPIX, albeit 'good' LTCM pre 1998, but I believe the QSPIX/LTCM comparison is useless all around, based on the facts.

Re: QSPIX - thoughts on interesting fund

Posted: Fri Sep 11, 2015 5:33 pm
by larryswedroe
[OT comment deleted by admin alex]
Comparing the two is IMO absurd, they literally have virtually nothing of any significance in common. LTCM took positions based solely on human judgments, people's opinions, not based on academic research and loaded on massive leverage. AQR only invests in systematic manner, not based on human judgments, and only in factors supported by deep academic research, and uses minimal amount of leverage. In addition LTCM by necessity was totally opaque in what it did while AQR is almost totally transparent including doing attribution analysis on quarterly basis to show the sources of returns and out/underperformance. And they have a good record of doing exactly what they say they do. Like I said, nothing in common.
Best wishes
Larry

Re: QSPIX - thoughts on interesting fund

Posted: Sat Sep 12, 2015 12:00 am
by Biffer
larryswedroe wrote:And add that there is no human intervention with human's making judgments about what is a good position or a bad one--it's systematic and replicable approach

larryswedroe wrote:AQR only invests in systematic manner, not based on human judgments, and only in factors supported by deep academic research, and uses minimal amount of leverage.

Larry, You seem to be saying that QSPIX does not rely on human judgments. But doesn't every actively managed fund rely on some human judgments?

Re: QSPIX - thoughts on interesting fund

Posted: Sat Sep 12, 2015 9:17 am
by larryswedroe
Biffer, yes but QSPIX is NOT actively managed, no idea why you think it is. It gets systematic exposure to each factor determined by definitions of the universe of securities it buys or shorts and doesn't make judgments about them at all. Only judgment being made is when using the algo programs to trade and they see something weird going on in the market they MIGHT make some judgment. No different than DFA which does same thing, and Bridgeway.
Larry

Re: QSPIX - thoughts on interesting fund

Posted: Sat Sep 12, 2015 10:30 am
by nisiprius
larryswedroe wrote:...[in QSPIX there is] no human intervention with human's making judgments about what is a good position or a bad one...

...Biffer, yes but QSPIX is NOT actively managed, no idea why you think it is...
AQR Funds, Prospectus, click on "Fund Summary: Style Premia Alternative Fund," scroll to p. 84 (as shown on the page itself, page "89" in their navigation bar).
AQR wrote:The Fund is actively managed and the Fund’s exposures to Styles and Asset Groups will vary based on the Adviser’s ongoing evaluation of investment opportunities.
"Adviser" = humans, "evaluation" = making judgments.

Image

Re: QSPIX - thoughts on interesting fund

Posted: Sat Sep 12, 2015 10:51 am
by lack_ey
Nobody seems to agree what that really means, what the cutoff is. It might be clearer to just say that the securities selection and management is non-discretionary, being based on algorithmic screens. Maybe Larry has a bit more insight as to what goes on in terms of the trade execution and then macro-level details of fund style rebalancing and risk sizing based on "ongoing evaluation of investment opportunities" and so on.

For what it's worth, order placement and timing, managing of fund inflow and outflows, and sometimes securities selection are all part of passive index funds as well, particularly bond funds where the index must be sampled. Not really a good or bad thing, just how things are.

Some funds deviate from the broad market by holding a very concentrated portfolio and not trading it much. Others follow indexes that slice the market by certain qualities. Others don't rely on an index provider but create their own fund compositions in much the same way. Some of these may be called quant funds, with different levels and frequencies of trading being prevalent. There are traditional discretionary managers too, but many of those rely somewhat or heavily on their own rules or even quantitative models. I don't think people agree on which of the above are passive, but even if everyone followed the same definitions, the distinction may not even be that useful, as inevitably different funds under the same labeling may have very different properties and modes of operation.

Re: QSPIX - thoughts on interesting fund

Posted: Sat Sep 12, 2015 11:22 am
by nisiprius
lack_ey wrote:Nobody seems to agree what that really means...
It really means that "The Fund is actively managed and the Fund’s exposures to Styles and Asset Groups will vary based on the Adviser’s ongoing evaluation of investment opportunities." What could be clearer?
lack_ey wrote:For what it's worth, order placement and timing, managing of fund inflow and outflows, and sometimes securities selection are all part of passive index funds as well, particularly bond funds where the index must be sampled. Not really a good or bad thing, just how things are.
Maybe you feel Total Bond should be regarded as being actively managed, but Vanguard's lawyers do not think it is necessary to say so. They think there is no problem with saying that
Vanguard, in the Total Bond Fund prospectus wrote:The Fund employs an indexing investment approach designed to track the performance of the Barclays U.S. Aggregate Float Adjusted Index.... The Fund invests by sampling the Index, meaning that it holds a broadly diversified collection of securities that, in the aggregate, approximates the full Index in terms of key risk factors and other characteristics.
The word "active" appears in the prospectus only in three places: to say that index funds are inexpensive to run compared to active funds;" to say that Mortimer J. Buckley is part of the management of Vanguard's "active bond funds" as well as index funds; and in a glossary where "active management" is defined. The word "evaluation" does not appear at all. "Opportunity" occurs only where it is said that you will be given an opportunity to tell Vanguard not to promote Investor shares to Admiral.

QSPIX is actively managed and the Fund’s exposures to Styles and Asset Groups will vary based on the Adviser’s ongoing evaluation of investment opportunities. Total Bond is not actively managed and the fund's exposures to bond categories does not vary based on the adviser's ongoing evaluation of investment opportunities.

Re: QSPIX - thoughts on interesting fund

Posted: Sat Sep 12, 2015 11:41 am
by lack_ey
Personally if I had to choose a word I would call AQR's QSPIX (style premia) active for sure and Vanguard VBTLX (total bond) passive. I'm not really arguing Larry's point here. More broadly, I would say there's a range of investment strategies and behaviors, and they don't even fall neatly on an active/passive continuum (one axis), so I don't see much point in focusing too much on the one word as that chucks away a lot of the detail anyway. If you tell me a fund is active, that doesn't tell me how it goes about its business, so I still need to know more about it. Likewise, if you call a fund passive, I'm not sure what it's doing and if you mean a DFA-style fund, a small cap stock index fund, a fund-of-funds using index funds, some kind of low-AUM multi-tilted iShares contraption, or a broad market fund, all of which at least some people will call passive.

Didn't total bond get in trouble tracking its index in the early 2000s (more specifically, underperformed nontrivially) because they misevaluated the risk of some securities and tilted in a way that got them in trouble? The process now is different, to be sure.

Re: QSPIX - thoughts on interesting fund

Posted: Sat Sep 12, 2015 12:08 pm
by larryswedroe
What everyone needs to understand is that prospectuses all are written to address the POTENTIAL for unusual situations that MIGHT POSSIBLY occur and can literally not reflect anything the fund intends to do. You see this in DFA prospectuses also.
This fund is NOT actively managed in any true sense of the word, as the allocations are all done in a systematic matter, not with human beings making judgments. Now you can accept that or not but it won't change the facts. People need to understand how prospectuses are written, to avoid any potential issue that comes up that would require having to go to investors to change some minor language in order to do something that is perfectly rational. AQR's fund is no more actively managed than say Vanguard small caps or DFA funds (both do some patient trading for example,just to different degrees). In this case AQR does manage the level of leverage but it's formulaic, not human judgment. It targets 10% volatility, and as vol changes the level of leverage changes--vol predicts future vol, so as vol rises leverage is reduced and vice versa. So if you want to call that active management, fine. But it's systematic, replicable, not based on human's making judgment in way LTCM did.
Larry

Re: QSPIX - thoughts on interesting fund

Posted: Sat Sep 12, 2015 12:34 pm
by Biffer
larryswedroe wrote:This fund is NOT actively managed in any true sense of the word, as the allocations are all done in a systematic matter, not with human beings making judgments.

Larry, The QSPIX prospectus states: "The Fund is actively managed..."

Do you believe that the QSPIX prospectus is in error? Misleadingly inaccurate?

Re: QSPIX - thoughts on interesting fund

Posted: Sat Sep 12, 2015 12:43 pm
by nisiprius
lack_ey wrote:Personally if I had to choose a word I would call AQR's QSPIX (style premia) active for sure and Vanguard VBTLX (total bond) passive. I'm not really arguing Larry's point here. More broadly, I would say there's a range of investment strategies and behaviors, and they don't even fall neatly on an active/passive continuum (one axis), so I don't see much point in focusing too much on the one word as that chucks away a lot of the detail anyway. If you tell me a fund is active, that doesn't tell me how it goes about its business, so I still need to know more about it. Likewise, if you call a fund passive, I'm not sure what it's doing and if you mean a DFA-style fund, a small cap stock index fund, a fund-of-funds using index funds, some kind of low-AUM multi-tilted iShares contraption, or a broad market fund, all of which at least some people will call passive.

Didn't total bond get in trouble tracking its index in the early 2000s (more specifically, underperformed nontrivially) because they misevaluated the risk of some securities and tilted in a way that got them in trouble? The process now is different, to be sure.
I don't know how much trouble they got into. They underperformed by enough to be annoying. Then around 2007 Fidelity did the same thing. It's kind of pretty on a chart, the two lines diverge and rejoin.

Image

The fund you are probably thinking of is the Schwab Total Bond Market Fund, which does not include the word "index" in its name, but tracked the Barclay's index as perfectly as any index fund... until 2007. At the time it happened, they were holding a portfolio that was very different from the composition of the Barclay's index and there are or were class action lawsuits about it. Hmmm...
Image

To me, the Schwab Total Bond Market Fund is a perfect example of the pitfalls of saying "Technically according to the prospectus it's not an index fund, but that's only a technicality, and we know what the fund managers intend to do and that's good enough."

March, 2015, "Charles Schwab must face U.S. lawsuit over bond index fund."
A U.S. appeals court on Monday revived a class-action lawsuit accusing Charles Schwab Corp of stuffing risky mortgage-backed securities into a bond index mutual fund, causing the fund to significantly lag its benchmark. The plaintiffs said that by investing more than 25 percent of assets in non-agency mortgage securities and collateralized mortgage obligations, Schwab portfolio managers ignored the fund's fundamental investment objectives of tracking the Lehman Brothers U.S. Aggregate Bond Index and avoiding big industry bets.

They said this caused the fund to lag its benchmark from Sept. 1, 2007 to Feb. 27, 2009, losing 4.80 percent while the index posted a positive total return of 7.85 percent.

Re: QSPIX - thoughts on interesting fund

Posted: Sat Sep 12, 2015 12:49 pm
by packer16
Larry,

Are you saying they have a static % exposure to each factor in each asset class over time and adjust the leverage to match expected volatility. Or do they have ranges of factor exposure by asset class over time that they adjust based upon their expected returns of those factors versus other factors (this is what the prospectus implies) and use leverage to then adjust the expected volatility. The former is more passive than the later.

Packer

Re: QSPIX - thoughts on interesting fund

Posted: Sat Sep 12, 2015 12:53 pm
by larryswedroe
biffer
Two issues here, as I stated above most investors don't understand that prospectuses are NOT written to what the fund intends to do and how it will be run, but to account for any possibility that MIGHT occur so the managers can take appropriate action without being in violation of the prospectus and thus liable to lawsuits. In other words it's written by lawyers to give the fund managers the freedom to act in rational manner to address issues that might theorectically come up.


And as I noted the leverage is "actively" managed but not in way most think of the term. They don't change leverage based on someone's views but in response to changes in volatility, to manage the fund so the vol stays at 10% (so lowering leverage when vol rises, reducing risks, and vice versa).
But the fact is the fund maintains the same allocations to each of the factors and the same percentages in each of the asset classes in each of the factors, not changing them based on someone's judgment.

So if you want to say Vanguard small index fund is active when it might do a block trade and not match the index exactly, I won't argue, and DFA does more of that, and it uses algo programs to trade to keep costs down. And this is of the same nature, very minor issues that should lead to totally random tracking error in return for lower trading costs.


I hope that is helpful.
Larry

Re: QSPIX - thoughts on interesting fund

Posted: Sat Sep 12, 2015 12:55 pm
by nedsaid
Nisiprius has the annoying habit of looking things up. :happy

I think Larry is correct, the fund is probably passive but the lawyers probably made them put "active" in the language. The managers at any time could put their hands on the levers and exercise judgment particularly in a crisis market. The intent is to be passive. My guess is that what they are doing is somewhere in the continuum between active and passive. Factor investing is somewhere in the middle though much more passive than active.

Re: QSPIX - thoughts on interesting fund

Posted: Sat Sep 12, 2015 2:59 pm
by larryswedroe
Just to be clear
If AQR was shifting between weights on factors and which asset classes it was using that would be active, but it doesn't do any of those type things
So I personally consider it passive, being systematic in implementation,
Larry

Re: QSPIX - thoughts on interesting fund

Posted: Sat Sep 12, 2015 3:06 pm
by HomerJ
larryswedroe wrote:Just to be clear
If AQR was shifting between weights on factors and which asset classes it was using that would be active, but it doesn't do any of those type things
So I personally consider it passive, being systematic in implementation,
Larry


Just curious. How do you know they are not? Do they publish how they are weighting factors every month?

Re: QSPIX - thoughts on interesting fund

Posted: Sat Sep 12, 2015 3:28 pm
by larryswedroe
They have a presentation to educate investors/advisors before they invest, they also provide a quarterly attribution analysis showing the results and the sources of the returns, at least to advisors.
Also note they do this for ALL THEIR FUNDS and they have a good track record of delivering just what they say they do, just like DFA, Bridgeway and Vanguard, and are highly transparent. Which is why people like me and Robert T are confident when we invest with them. In our case I assure you that a considerable amount of due diligence was done, including spending time in their trading rooms, watching them execute.
Larry

Re: QSPIX - thoughts on interesting fund

Posted: Sat Sep 12, 2015 3:43 pm
by lack_ey
larryswedroe wrote:They have a presentation to educate investors/advisors before they invest, they also provide a quarterly attribution analysis showing the results and the sources of the returns, at least to advisors.
Also note they do this for ALL THEIR FUNDS and they have a good track record of delivering just what they say they do, just like DFA, Bridgeway and Vanguard, and are highly transparent. Which is why people like me and Robert T are confident when we invest with them. In our case I assure you that a considerable amount of due diligence was done, including spending time in their trading rooms, watching them execute.
Larry

Okay, then I had the (mistaken?) impression that there could be some shifting of styles/weights.

The very basic fact sheet says
Style agreement - Allows for the risk levels of asset classes to vary depending on the degree of agreement across styles


To me that implies that if for example a beaten-down commodity has strong positive momentum and has the most backwardation, with an expensive commodity with negative momentum that happens to be in contango, then the algorithm (or a person) can decide to shift up the risk weighting of commodities a bit because the styles agree on what to invest in and would not be contradicting each other when implemented. To me, if you suppose that the styles will make money in the long term, that sounds like a pretty reasonable approach and furthermore would constitute a shifting in weightings not related to overall volatility targeting. There to me would seem to be less point in spending the same risk weighting in an asset class if for example what momentum says to go long and short is the opposite of what value says to go long and short, where there's style disagreement. Is this not the case at all, or generally not the case, or maybe just hasn't happened so far?





nisiprius wrote:
lack_ey wrote:Personally if I had to choose a word I would call AQR's QSPIX (style premia) active for sure and Vanguard VBTLX (total bond) passive. I'm not really arguing Larry's point here. More broadly, I would say there's a range of investment strategies and behaviors, and they don't even fall neatly on an active/passive continuum (one axis), so I don't see much point in focusing too much on the one word as that chucks away a lot of the detail anyway. If you tell me a fund is active, that doesn't tell me how it goes about its business, so I still need to know more about it. Likewise, if you call a fund passive, I'm not sure what it's doing and if you mean a DFA-style fund, a small cap stock index fund, a fund-of-funds using index funds, some kind of low-AUM multi-tilted iShares contraption, or a broad market fund, all of which at least some people will call passive.

Didn't total bond get in trouble tracking its index in the early 2000s (more specifically, underperformed nontrivially) because they misevaluated the risk of some securities and tilted in a way that got them in trouble? The process now is different, to be sure.
I don't know how much trouble they got into. They underperformed by enough to be annoying. Then around 2007 Fidelity did the same thing. It's kind of pretty on a chart, the two lines diverge and rejoin.
[img]

The fund you are probably thinking of is the Schwab Total Bond Market Fund, which does not include the word "index" in its name, but tracked the Barclay's index as perfectly as any index fund... until 2007. At the time it happened, they were holding a portfolio that was very different from the composition of the Barclay's index and there are or were class action lawsuits about it. Hmmm...
[img]

I meant the Vanguard fund, which lost out that ~1% or so. I believe I read here that it had something to do with taking more credit risk on short maturities in an attempt to track the index better by outperforming there and covering more of the gap created by the expense ratio, but getting burned. I forget exactly, sorry. Obviously not by a huge amount, but something we can see. I think they've learned their lesson from that episode, though. It was just intended as a minor example.

Re: QSPIX - thoughts on interesting fund

Posted: Sat Sep 12, 2015 4:14 pm
by backpacker
This confirms my view that no one knows what this fund is doing. We can't even figure whether or not the style weightings are actively managed! :oops: