hafius500 wrote:Akiva wrote:...no one says that small-cap and value tilts are "quant" strategies...........This isn't a "quant" strategy because it's entirely passive. It's conceptually the same sort of thing as these fundamentally derived indexes -- a different way of estimating what weights you should assign in an attempt to overcome various problems with capitalization weights. The only difference is that it uses somewhat more sophisticated mathematics that happen to have better theoretical support. (The method also has some theoretical justification from behavioral finance.)
So you call an investor, who attempts,...who estimates...who tries to avoid less sophisticated maths...who searches theorectical support that is better than average... a passive investor?? And the active investor would be the investor who uses less sophisticated maths or the monkey that randomly selects stocks??
There's a lot of really sophisticated math that goes into the theory of market cap weighting as well (e.g. stochastic discount factors and intertemporal general equilibrium models), not to mention a whole raft of (absurdly) unrealistic assumptions about utility functions, risk preferences, etc. that you need to derive the conditions under which a market cap index makes sense.
In some sense, the alternative that I'm referencing is actually simpler, since it relies only on (advanced) econometrics and basic portfolio mathematics. In fact, but for the above complicated theory telling you that market cap weighting was right, the logical thing to do would be to use a risk factor model to estimate covariances and then solve for the minimum variance portfolio. (i.e. you admit that you can't predict returns, and unlike above, you say that you don't even know which risks will be compensated and which won't, so all you can do is minimize your overall risk.) IOW, the only reason you think that market cap weighting is passive and volatility minimizing weights are not is because there's a popular theory telling you this is the case for various esoteric reasons that most on this board don't understand and don't care to. Under different sets of assumptions, each of the various alternatives being discussed in this thread will be the resulting "market" portfolio and would thus represent the "passive" investment.
DFA funds are quant funds.
Everyone I know who works at an actual quant fund would take exception to this. DFA offers funds based around implementing the best academic evidence on investing. Nothing they do is proprietary or secret, and their turn-over is way too low to be considered "active".
Neither of the links you bring are technically apropos to this discussion because they are talking about the phenomena that a fund with returns of much above 20% ends up holding ludicrous amounts of the market over a period of time (and thus brings the market average in line with its returns). I presume the point of all this is to challenge my claim that small cap and value tilts are passive, but I would suggest that you just look around the board. No one here is going to claim that the tilts that Bogleheads use make these investments "active".
Quant funds are always active funds by definition. Google ("DFA + quant") found, e.g.,:
What Will Be Obvious 20 Years From Now? (With a Sharpe of 0.76)?. August 22nd, 2012 by Mebane FaberWhat is really interesting to me is that often the great investment approaches seem so obvious in retrospect. Buffett has one of the best track records ever by buying cheap, safe, high quality stocks in a structure that allows for sub Tbill leverage. But the key is that he came to this realization before most. Ditto for the top endowments like Harvard and Yale...– it is quite obvious now, but they were pioneering these concepts decades ago. The pioneering trendfollowers and managed futures shops caught onto an idea long before computers made it simple, as did Bogle (indexing), and DFA (quant multifactor)
This post has nothing to do with DFA. Second you are confusing the use of a multifactor investment strategy with quantitative investing. Tilting towards value and small cap is not a quant strategy and the Fama French 3 factor model has nothing to do with quantitative investing.
Wikipedia gives a short explaination of what a quant fund does:
Quantitative investing represents an investing technique typically employed by the most sophisticated, technically advanced hedge funds. These quant shops employ fast computers to find predictable patterns within financial data. Some of the larger quant shops include but are not limited to Renaissance Technologies' Medallion Fund, Winton Capital Management, D. E. Shaw & Co., Barclay's Global Investments (now known as Blackrock), Numerics, GMO, First Quadrant, Robeco, etc.
Typically, quant investing is implemented by people who have spent time in the physics, math, computer science, or statistics disciplines. The condensed results of quantitative analyses, however, can be readily accessible to all far-from-quantitative investors, when presented in an intuitive framework.
The process consists of thorough examination of vast databases searching for repeating patterns—persistent occurrences of a phenomenon, correlations among liquid assets ("statistical arbitrage" or "pairs trading"), or price-movement patterns (trend following or mean reversion).
DFA is academic finance people (no hedge fund hires these guys), and they aren't looking for patterns in price data to arbitrage away, they are looking to implement the best academic understanding of passive investing.
Several posts on the Fama/French forum mention "passive" growth- or value-tilted portfolios. This quote (Gene Fama) demonstrates the errors of such propositions:
1) Fama's quote is provided without relevant context, but the argument seems to be that passive exposure to his "value" risk factor shouldn't entail trading and so the active managers have to cancel each other out, thus leaving their net returns due to value exposure to be equal to his passive benchmark less fees. Unfortunately he inserted the word "value" which makes his statement confusing -- it seems like you could say that if the active value managers win, it could be at the expense of other types of active managers, but Fama's point is that we have fully accounted for the returns to value investing with the "value" factor. So any active value manager that overweights some value stocks and underweights others is winning at the expense of other value managers, on net "value" investors can't beat the value index. The flaw in this argument is that his value index doesn't capture what active value managers typically do. Instead of using ratios, they use a discounted cash-flow model. So they aren't giving just a mix of exposure to the market factor and the value factor. There are other things going on, and consequently they could be winning at the expense of other (non-value) active managers.
2) More on topic, you are confusing someone who manages an active value fund with someone who tilts his portfolio towards value stocks as defined by exposure to Fama's own "value" risk factor. That post was about the ability of active value investors to beat the market by a substantial amount over a long period of time.
3) It is worth noting that from Fama's perspective you don't "beat the market" by tilting towards value, you simply earn an incremental return for taking incrementally more risk. But he would certainly admit that tilting in this way is entirely passive; in fact that's who he's talking about when he says "passive value managers holed value-weight portfolios of value stocks."
Active (or "passive") value investors trade against other active investors but they cannot trade against passive investors who hold the market-weighted portfolio because these investors do not trade. It would make no sense to call the seller of X a passive (value) investor and the buyer an active investor. Why not the other way around? The stock-picker or the technical investor would be passive investors as well, especially, if they "passively" stayed the course.
It does make sense to call a value tilt passive because just like holding "the market" in a market cap weighted portfolio, you are accepting known risk exposure based on market valuations.
IMO Fama's definition is tautological because the value benchmark is arbitrary and can be replaced arbitrarily:
While I take issue with the way many understand Fama's value factor, his benchmark is far from arbitrary.
"Passive stock-pickers who buy undervalued stocks hold value-weighted portfolios of undervalued stocks."
You obviously don't understand how the Fama-French three factor model works because there's no "stock-picking" involved.
As I wrote in another post, the only sensible defintion of active vs. passive investing that I have seen refers to the costs of information:
There are different methods of active management. Broadly speaking they can be grouped into two categories 1) those who try to predict returns (and thus earn higher returns for the same risk) and 2) those who try to predict how risks will change (and thus beat the market by getting equivalent returns for less risk).
Neither of these categories of techniques apply to what we are discussing here. Value-weighted indexes don't predict returns or risk, they simply weight by economic impact instead of market size. Equal-weighted indexes don't weight at all. Volatility minimizing indexes assume that risks from risk factors are constant and that exposure to those factors does not change (in contrast to method #2 above) and minimize volatility given the risk factor exposure of the stocks in the index.
Active investors make informed trades (value investors need to define and search value stocks), passive investors make informationless trades.
And none of the alternative indexing methods discussed in this thread require information. They are all the result of seeking "market" exposure under different assumptions about investor behavior.
From your second reply:
All of them agree with the maths of the two authors. The latter do not miss the point that some stocks are wrongly valued. Their proof assumes that securities are randomly mispriced ( overvalued and undervalued).
But this is question begging. The argument from the advocates of value weighting is that security mispricing is non-random. To reply by saying that "You are wrong if we assume that mispricing is random" does in fact miss the point entirely.
And there is substantial econometric evidence to support the claim of the fundamental indexing people that mispricing is non-random. Stocks with high residual volatility for example have dramatically lower returns.
In any event ... thanks for pointing this emerging asset class out to me.