Cliff Asness, Andrea Frazzini, Ronen State of Israel too Toby J. Moskowitz receive got a lovely SSRN paper "`Fact, Fiction too Momentum Investing."
The whole momentum literature is then huge, it's difficult to know where to amount up, too this is a non bad place, especially if you're didactics an MBA class. Since they're manifestly a chip conflicted (momentum + value is i of AQR's centre strategies), their emphasis on the elementary facts, which yous (or your students) tin post away replicate from Ken French's databse, is great.
Some "Myths,"
Myth #1: Momentum returns are likewise “small too sporadic”. Like whatever factor, it's non an aribtrage opportunity.
Myth #2: Momentum cannot live captured yesteryear long-only investors equally “momentum tin post away only
live exploited on the brusk side.
(evidence that the long side industrial plant equally well)
Myth #3: Momentum is much stronger with pocket-sized cap stocks than large caps.
Myth #5: Momentum does non function for a taxable investor.
too then on.
This isn't a 100 per centum endorsement. I've seen other cuts of the information that propose other results, especially Fama too French's finding that momentum didn't be 1926-1963, the large left tail of the fiscal crisis, the query whether it's lastly beingness arbitraged away, too Fama too French's prove that it's weaker inwards large stocks, though non absent. I postulate to reconcile all the evidence, too it's soundless a huge projection that I've set off for around other day.
Also, momentum is either large or pocket-sized depending on how yous expect at it. Suppose returns receive got a 0.1 autocorrelation, too thus a 0.01 R2 inwards R_t+1 = a + b R_t + error_t+1. The locomote yesteryear 1/10 of stocks inwards the previous twelvemonth went upwards virtually 50%. The bottom 10% of stocks inwards the previous twelvemonth went downwards virtually 50%. So, a 100% previous twelvemonth render spread, too a 0.01 R2 implies a 10% render spread side yesteryear side twelvemonth -- virtually what nosotros meet inwards the 1-10 momentum portfolios. You mightiness tell that a 0.01 R2 is a tiny chip of forecastability. You mightiness tell that a 10% render spread is a huge phenomenon. They are the same.
But if yous desire i non bad newspaper to read too assign your students virtually momentum, this is it.
The whole momentum literature is then huge, it's difficult to know where to amount up, too this is a non bad place, especially if you're didactics an MBA class. Since they're manifestly a chip conflicted (momentum + value is i of AQR's centre strategies), their emphasis on the elementary facts, which yous (or your students) tin post away replicate from Ken French's databse, is great.
Some "Myths,"
Myth #1: Momentum returns are likewise “small too sporadic”. Like whatever factor, it's non an aribtrage opportunity.
The render premium is evident inwards 212 years (yes, this is non a typo, 2 hundred too twelve years of information from 1801 to 2012) of U.S. of A. equity data
So, to amount up, who yous calling pocket-sized too sporadic?Influenza A virus subtype H5N1 writing mode to a greater extent than bold than my own.
Myth #2: Momentum cannot live captured yesteryear long-only investors equally “momentum tin post away only
live exploited on the brusk side.
(evidence that the long side industrial plant equally well)
Myth #3: Momentum is much stronger with pocket-sized cap stocks than large caps.
Putting it starkly: in-sample, out-of-sample, calculated inwards Greenwich Connecticut, Chicago, Boston, Palo Alto, Santa Monica, Austin, or inwards the library with a candlestick, wherever or withal yous desire to look, along whatever dimension, those who brand the claim that momentum fails for large caps, spell beingness supporters of value investing, are non merely mistaken, they receive got it backwards.Myth #4: Momentum does non survive, or is seriously express by, trading costs.
Frazzini, Israel,and Moskowitz (2013), “FIM”, which uses trades from a large institutional investor [hmm, I wonder who that could be] over a long menses of time. . Their decision is that per dollar trading costs for momentum are quite low...
... Trading patiently (by breaking orders upwards into pocket-sized sizes too setting boundary social club prices that provide, non demand, liquidity) too allowing around tracking error to a theoretical mode portfolio tin post away significantly trim back trading costs...
Where did this myth come upwards from? ... First, the studies that discovery much larger trading costs produce then because they guess costs for the average investor...which plough out to live virtually 10 times larger than the costs of a large institutional manager,...Second,...these other studies attempt portfolios that produce non consider transactions costs inwards their design, which tin post away significantly trim back turnover too hence trading costs further.This actually is the DFA sales pitch. Momentum, a elementary strategy may work. But if you endeavor to produce it, you'll teach swamped yesteryear trading costs.
Myth #5: Momentum does non function for a taxable investor.
too then on.
This isn't a 100 per centum endorsement. I've seen other cuts of the information that propose other results, especially Fama too French's finding that momentum didn't be 1926-1963, the large left tail of the fiscal crisis, the query whether it's lastly beingness arbitraged away, too Fama too French's prove that it's weaker inwards large stocks, though non absent. I postulate to reconcile all the evidence, too it's soundless a huge projection that I've set off for around other day.
Also, momentum is either large or pocket-sized depending on how yous expect at it. Suppose returns receive got a 0.1 autocorrelation, too thus a 0.01 R2 inwards R_t+1 = a + b R_t + error_t+1. The locomote yesteryear 1/10 of stocks inwards the previous twelvemonth went upwards virtually 50%. The bottom 10% of stocks inwards the previous twelvemonth went downwards virtually 50%. So, a 100% previous twelvemonth render spread, too a 0.01 R2 implies a 10% render spread side yesteryear side twelvemonth -- virtually what nosotros meet inwards the 1-10 momentum portfolios. You mightiness tell that a 0.01 R2 is a tiny chip of forecastability. You mightiness tell that a 10% render spread is a huge phenomenon. They are the same.
But if yous desire i non bad newspaper to read too assign your students virtually momentum, this is it.
