AQR Paper On Alternative Risk Premia

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
Topic Author
Random Walker
Posts: 4229
Joined: Fri Feb 23, 2007 8:21 pm

AQR Paper On Alternative Risk Premia

Post by Random Walker » Sun Mar 18, 2018 3:36 pm

https://www.aqr.com/-/media/AQR/Documen ... ia_vff.pdf

Fairly short summary from AQR on benefit of diversifying into alternative risk premia. Each of the premia show benefit across individual asset classes. More powerful to diversify across multiple styles and multiple asset classes. The premia are uncorrelated with each other and the behavior of each style uncorrelated between asset classes.

Dave

Dominic
Posts: 211
Joined: Sat Jul 02, 2016 11:36 am

Re: AQR Paper On Alternative Risk Premia

Post by Dominic » Sun Mar 18, 2018 4:00 pm

I'm still a little unclear about what carry and value mean.

In fixed income, value is long government bonds with high real yields and short government bonds with low real yields. If I understand correctly, this is based on the real yield curve for a single issuer of bonds. That is, buy the highest yielding Treasuries unless the real yield curve inverts.

In currency, carry is long high-interest currencies and short low-interest currencies. I assume that the interest rate is defined by T-bills and their international equivalents. So hold currencies whose ultra-short-term sovereign bonds are yielding the most.

There is also carry in fixed income, and I assume this the same as value, but with nominal yields instead of fixed yields. That is, ignore the implied inflation rate found by the spread between TIPS and nominal Treasuries.

Is any of that right?

On top of that, "defensive" is new to me. Is this a new name for the quality/profitability factor? And is volatility the same as AQR's "betting against beta?"

User avatar
nisiprius
Advisory Board
Posts: 39751
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: AQR Paper On Alternative Risk Premia

Post by nisiprius » Sun Mar 18, 2018 4:25 pm

Random Walker wrote:
Sun Mar 18, 2018 3:36 pm
...Fairly short summary from AQR on benefit of diversifying into alternative risk premia. Each of the premia show benefit across individual asset classes. More powerful to diversify across multiple styles and multiple asset classes. The premia are uncorrelated with each other and the behavior of each style uncorrelated between asset classes...
I think you meant to say:

"Each of the premia show has shown benefit across individual asset classes. It has been more powerful to diversify across multiple styles and multiple asset classes. The premia are have been uncorrelated with each other and the behavior of each style has been uncorrelated between asset classes."

There's no reason to expect any of that to change, nor is there any reason to assume it is sure to persist.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

lack_ey
Posts: 6701
Joined: Wed Nov 19, 2014 11:55 pm

Re: AQR Paper On Alternative Risk Premia

Post by lack_ey » Sun Mar 18, 2018 4:42 pm

Dominic wrote:
Sun Mar 18, 2018 4:00 pm
I'm still a little unclear about what carry and value mean.

In fixed income, value is long government bonds with high real yields and short government bonds with low real yields. If I understand correctly, this is based on the real yield curve for a single issuer of bonds. That is, buy the highest yielding Treasuries unless the real yield curve inverts.

In currency, carry is long high-interest currencies and short low-interest currencies. I assume that the interest rate is defined by T-bills and their international equivalents. So hold currencies whose ultra-short-term sovereign bonds are yielding the most.

There is also carry in fixed income, and I assume this the same as value, but with nominal yields instead of fixed yields. That is, ignore the implied inflation rate found by the spread between TIPS and nominal Treasuries.

Is any of that right?

On top of that, "defensive" is new to me. Is this a new name for the quality/profitability factor? And is volatility the same as AQR's "betting against beta?"
AQR likes to define some of these things and extend some of these concepts in ways that are not that standard, so let's see if I can get this right...

They define carry in fixed income as the spread between long-term and short-term rates. So that's a riding-the-yield curve kind of concept: choose steeper yield curves. Value to them in fixed income is based on real (after-inflation) yield. So select countries where real rates are higher. So again, it's choosing between countries.

Defensive in equities would be low beta or low risk (maybe low vol, maybe low idiosyncratic risk—these things are different but related). It's related to their betting-against-beta. There's some relationship with quality/profitability, but no, not a name for that. They have funds that target quality in equities, and they call it quality there.

The definitions for value, momentum, carry, and defensive, and a bit more context, is in their "Investing with Style" paper:
https://www.aqr.com/-/media/AQR/Documen ... -Style.pdf

Sometimes it's better to read the journal publications like that, which are more vetted and may have more details, rather than their white papers.

Random Walker wrote:
Sun Mar 18, 2018 3:36 pm
The premia are uncorrelated with each other and the behavior of each style uncorrelated between asset classes.
Um, I think they do show correlations that don't look like zero in Exhibit 3: specifically, negative between value and momentum in the way they've defined, and positive between momentum (cross-sectional) and trend (time-series momentum). Largely or mostly uncorrelated, I suppose, usually?

Also, wasn't the point of some of their other papers that styles are positively correlated between asset classes and this is part of the evidence that they exist and are legitimate? Not so high that there wasn't a diversification benefit in using multiple asset classes to target a given style, but certainly nontrivial and many pairs very statistically significant.

User avatar
Robert T
Posts: 2608
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: AQR Paper On Alternative Risk Premia

Post by Robert T » Sun Mar 18, 2018 6:47 pm

.
Interesting.

Two observations:

(1) Exhibit 5 shows a 60:40 stock:bond market portfolio with 20% "alternative risk premia" added. This implies 48% stocks: 32% bonds: 20% alternative risk premia which is equivalent to a 48% stock:52% bond portfolio with a factor tilt (assuming the alternative risk premia series is market neutral). There are alternative ways to get a similar stock:bond mix, with a factor tilt. For example:

1990 - 2016: Annual return / Standard Deviation
  • 7.2% / 8.1 = 48:32:20 MSCI World:Barclays Global Aggregate: Alternative Risk Premia (from the exhibit 5 in the paper)
    7.9% / 8.6 = 48:52 DFA Balanced Equity:DFA Balanced Fixed Income (data from the DFA matrix book)
    7.7% / 8.1 = 45:55 DFA Balanced Equity:DFA Balanced Fixed Income
    7.2% / 7.3 = 40:60 DFA Balanced Equity:DFA Balanced Fixed Income
(2) Exhibit 4 seems to show a structural break from about 2006. Not sure the reason, but low interest/t-bill rates on the underlying collateral to cover short positions in the style premia fund is not helping.

Robert
.

Topic Author
Random Walker
Posts: 4229
Joined: Fri Feb 23, 2007 8:21 pm

Re: AQR Paper On Alternative Risk Premia

Post by Random Walker » Sun Mar 18, 2018 8:25 pm

Lack-ey,
Yes some very low correlations. I think of anything -0.3 to +0.3 as uncorrelated.

Nisiprius,
Yes, we only have past data to look at. There are intuitive reasons to expect returns and behavior to persist, but you are right; future may well look different.

Dave

Post Reply