Forums  > General  > "Risk parity"  
     
Page 1 of 4Goto to page: [1], 2, 3, 4 Next
Display using:  

Cheng


Total Posts: 2619
Joined: Feb 2005
 
Posted: 2013-01-22 11:30
I stumbled across this concept in an article this morning. From what I understand:

- Bridgewater and other HFs pitch a "new" way to increase meager bond returns: leverage
- Either HFs offer prepacked products / funds or bond players like pension funds do it themselves, either by levering themselves or using derivatives like futures
- Everyone is relaxed since bonds went up for the last 30 years (more or less at least), volatility is still small compared to equity portfolios and of course you can get out or have enough cash at hand if markets turn against you
- Not to forget: it is cool because you get equity like returns using bonds without the risks associated with equities

Is my understanding correct so far ? Can anyone provide further insight, eg how common this concept is, who uses it and the like ? This reminds me of 2006, kinda...

"Sun's going down / Moon's rising high / We'll pay you Bodom beach terror"

AndyM


Total Posts: 2242
Joined: Mar 2004
 
Posted: 2013-01-22 13:34
CPD'oh!

I used to be disgusted; now I try to be amused...

Cheng


Total Posts: 2619
Joined: Feb 2005
 
Posted: 2013-01-22 14:25
Worse... not AAA rated.

"Sun's going down / Moon's rising high / We'll pay you Bodom beach terror"

Corey


Total Posts: 237
Joined: Feb 2008
 
Posted: 2013-01-22 15:30
[Redacted]

"Then there was the man who drowned crossing a stream with an average depth of six inches." W. I. E. Gates

mmport80


Total Posts: 25
Joined: Jul 2010
 
Posted: 2013-01-22 16:31
Equalising marginal risk (however it is defined or whatever are the current marketable bogeymen).

Genius!!

--- http://johnorford.blogspot.com

quantie


Total Posts: 879
Joined: Jun 2004
 
Posted: 2013-01-23 02:09
Well there is a ton of assets going in to these funds. The problem I see is we are at the top of the rally in fixed income(perhaps) so when rates sell-off there is going to be a bloodbath and all this levering of FI (aggs etc) in risk parity will be interesting to watch.

Cheng


Total Posts: 2619
Joined: Feb 2005
 
Posted: 2013-01-23 11:33
Thanks Corey for the insight. This confirms my initial concern that people move further and further out the risk curve hunting for yield. We definitly live in interesting times.

"Sun's going down / Moon's rising high / We'll pay you Bodom beach terror"

Corey


Total Posts: 237
Joined: Feb 2008
 
Posted: 2013-01-23 15:41
[Redacted]

"Then there was the man who drowned crossing a stream with an average depth of six inches." W. I. E. Gates

NIP247


Total Posts: 474
Joined: Feb 2005
 
Posted: 2013-01-23 23:01
Here's my 2c: risk parity (or balanced portfolios) is just a general term for the portfolio construction process whereby you focus on minimizing volatility without regards to expected returns.

The intuition was born out of pension funds realising that the 60/40 "long term" allocation of capital implied a 90% exposure to equity risk. So they needed to move away from thinking of investments in terms of capital exposure and start thinking in terms of risk.

What goes into a risk-parity portfolio seems open for debate though. Bridgewater say they take their investment decisions solely based on systematic fundamental models. From my understanding, their portfolio construction is then such that the sum of their risk/exposure should be neutral to macro/business cycle regime. That's one definition of "balanced", which is integral to their trading decisions.

The simplest approach to implementing a "risk-balanced" portfolio strategy is by equal weighting the expected volatility of the assets (and assuming zero correl). Then you introduce correl, and finally there are ideas about looking at subset baskets and their individual effect on the overall risk of the portfolio. These are all just portfolio construction techniques though (as far as I understand), which seem to be the way the good old trend-followers have always viewed risk management (looking at dollar volatility of your futures and/or having equal dollar value of entry to stoploss per trading signal)

My understanding of AQR's risk parity fund (Again, it would be interesting to hear Aaron here), is that it is a "balanced risk" fund of "classical" risk premia. So having identified a few risk premia that are assumed to exist and will continue to exist, the fund constructs a balanced portfolio of sort equalising the risk (however that is defined) and potentially taking into account assymetric behaviour/downside correlation of some of these risk premia.

Now to the debate on whether "leveraging up" bonds make sense or not is a mix between the portfolio construction technique and the investment strategy. If you believe in the existence of these risk premia, a risk parity investment in all of them is just a "beta" or "alternative beta". It doesn't assume that yields will continue to go down from this point; it is (again as far as I understand) just delivering the bond beta exposure at a certain target volatility. If the risk parity / rebalancing is dynamic, any increase in bond vol will automatically lead to a decrease in exposure. (We've discussed in another thread the assumption of higher vol and downside moves coinciding for this to work). If your strategy includes "forecasts" then I imagine you will do tactical adjustments to your exposure to the different premia.

I'd be very interested to hear everyone's thoughts on the subject (and rejections of any of my claims above). If you look up "risk parity" on wikipedia, there's a list of institutions purportedly "using" risk parity, none of them doing the same thing as far as I know. So at the most general level, it's just a catchy name describing allocation in terms of risk exposure rather than capital deployment, but now used to describe "alternative beta" investment...

On your straddle, done on the puts, working the calls...

quantie


Total Posts: 879
Joined: Jun 2004
 
Posted: 2013-01-24 00:17
There is a good paper in JoPM on the risk premiums to mix up and serve and there is the book here.

As NIP and Corey point out it would make sense (and hope)that these funds are doing more than just inverse scaling with risk (volatility or max drawdown or whatever) as that would have worked very well over the last decade or so and would look great in a backtest. This could well be the next portfolio insurance type idea. But given how much spreads have compressed in credit/HY and where tsys are at this point you'd want to use the relative attractiveness in the portfolio construction. (Presumably this is what Bridgewater, First Quadrant,PIMCO etc are upto)

briant57


Total Posts: 96
Joined: Feb 2009
 
Posted: 2013-01-24 13:59
Bridgewater's All Weather Strategy docs may be of interest.....
http://www.bwater.com/home/research--press/the-all-weather-strategy.aspx

Cheng


Total Posts: 2619
Joined: Feb 2005
 
Posted: 2013-01-24 14:39
A few random thoughts, mainly Re Corey:

There is nothing that says they haven't moved to very short duration with their FI exposure...

Which means in turn they have to increase the leverage since the volatility (or what other measure they use) will be correspondingly low Wink.

Backtesting is all nice and cosy here if you look back 10 or 20 years. Bonds have been performing since 1982 (some small drags in between but rates have come down for 30 years now), equities similar (treading water since 2000 because multiples come down to earth). I wonder how this would have performed in the 60s or 70s.

No matter how you define your risk measure (be it vol, max drawdown, whatever), you are short vol-of-vol (either directly or indirectly). As long as things behave well you generate nice returns. If some pooh hits the fan you have a fair chance of blowing up big time (and pls don't tell me about "we see it coming and unwind our positions in time"). Liquidity in the underlying assets is your friend here, if it vanishes you are up for levered trouble.

A colleague told me that LTCM employed similar strategies. Can anyone confirm this ? Or is this wrong ? Would be interesting to know in terms of hindsight.

"Sun's going down / Moon's rising high / We'll pay you Bodom beach terror"

nym


Total Posts: 15
Joined: Aug 2012
 
Posted: 2013-01-30 17:25
maybe a naive view... but it looks like just a small variation of the classical modern portfolio theory.

http://en.wikipedia.org/wiki/Modern_portfolio_theory

am i missing something? Confused

mib


Total Posts: 350
Joined: Aug 2004
 
Posted: 2013-02-01 19:54

It is a simple concept, easy to explain to clients/boards/trustees: "we cannnot be sure what the returns will be in the future, so taking equal risk in each risk/asset class"

The consequences are more interesting:

1. it is not sensitive to your estimates/guesses about future returns

2. it is very sensitive to your estimates of future volatilities, these estimates being often past volatilities

3. it is very sensitive to the definition of "asset class". Split the "equity" bucket into developed and emerging market equities - et voila, equity weight almost doubles!

4. It is natural if you assume equal sharpe ratios and equal correlations, but prescribed volatilities (there are versions that take correlations into account).

5. As soon as bonds are in the asset universe, it prescribes portfolios that are too low return for many institutional (not to mention retail) risk appetites, so one needs a fair amount of leverage to return back to the desired return. This creates an implicit curve slope position (long bonds financed short term). It would be interesting to ponder, want kind of instability is created by large players, like Danish ATP, allocating this way. Would we see a govt bond repo market breakdown on rising yields?

 

Seems like a marketing fad to me, although more elegant than some of the earlier fads, like 130/30


Head of Mortality Management, Capital Structure Demolition LLC

JamesBonds


Total Posts: 3
Joined: Dec 2008
 
Posted: 2013-02-02 01:14
A critique from a value-investing viewpoint: http://news.morningstar.com/pdfs/GMOHiddenRisks.pdf

Soon2Bgreat


Total Posts: 16
Joined: Jun 2007
 
Posted: 2013-02-02 19:10
A couple points on risk parity vs. the 60/40 mentioned in the article.
-In the 3yrs since the article, risk parity continues to outperform
-In 2009 when bonds were down >20% and equities up >20%, it wasn't nearly the bloodbath some indicate and wouldn't expect an environment going forward where bonds lose 20% pa

IMO, the main points have been covered and it is a tool like any other - I personally like the idea of equalizing volatility and think it has merit in a variety of market conditions.Smiley

Corey


Total Posts: 237
Joined: Feb 2008
 
Posted: 2013-02-04 21:07
[Redacted]

"Then there was the man who drowned crossing a stream with an average depth of six inches." W. I. E. Gates

NIP247


Total Posts: 474
Joined: Feb 2005
 
Posted: 2013-02-05 14:04
I know it's not the right thread, and I've already pestered Quantie, but if anyone has this article, it would be highly appreciated.

On your straddle, done on the puts, working the calls...

NIP247


Total Posts: 474
Joined: Feb 2005
 
Posted: 2013-02-05 14:44
@James Bond, agree 100% about the comments on whether there is a risk premia in most "asset classes", nevertheless (1) allocating risk rather than assets does not need to be static and (2) allocating risk to asset classes is just one "pillar". Next step is to look at "style premia". Obviously, there's a legitimate debate to whether such "style premia" (e.g. momentum, carry etc) should continue to perform/exist just because they've worked in a backtest and someone was able to post-fit a theory that sounds good knowing the results of those backtests...

On your straddle, done on the puts, working the calls...

mib


Total Posts: 350
Joined: Aug 2004
 
Posted: 2013-02-06 10:19

Soon2Bgreat

It is true that risk parity continued to outperform over the last three years - not really surprising with the rates coming down. But to me this is more likely to be a reason why it will underperform over the next few years - hard to see another 200bp rally at the long end of the yield curve.


Head of Mortality Management, Capital Structure Demolition LLC

eye51


Total Posts: 184
Joined: Oct 2004
 
Posted: 2013-02-12 09:08
Hi NIP247,

Just send you the paper.

regards,
Eye51

Peace means reloading your guns

NIP247


Total Posts: 474
Joined: Feb 2005
 
Posted: 2013-02-12 15:38
Thank you. Much appreciated!

On your straddle, done on the puts, working the calls...

apine


Total Posts: 1005
Joined: Jun 2004
 
Posted: 2013-05-22 17:26
any ideas about the amount of aum in these types of strategies? has this migrated to the non-HF world?

Too many people make decisions based on outcomes rather than process. -- Paul DePodesta

DocAdam7


Total Posts: 150
Joined: Nov 2007
 
Posted: 2013-05-22 21:24
I would say so. AQR has retail risk parity funds already and I don't think they're the only ones. Bridgewater is still probably the biggest player in these types of strategies.

That which counts cannot always be counted. That which can be counted does not always count.

kaehler


Total Posts: 24
Joined: Sep 2007
 
Posted: 2013-06-14 23:37
Maybe one way to think about risk parity is as sort of a "maximum entropy portfolio". That is, if your only belief is "over the long run, you get rewarded for supplying capital", you could view risk parity as the portfolio that maximizes utility (e.g. log wealth) with respect to the maximum entropy distribution of returns under that belief. Random thought and haven't yet figured out how to make it more formal, let alone useful, but will reply if I think of anything more interesting.

Also, not useful for anything, but I like to think of the apocryphal Rothschild formula -- invest one-third of your wealth in each of government bonds, real estate, and art -- as the first risk parity portfolio. Arguably, it was probably a decent approximation to risk parity across all of the asset classes in which a wealthy person in the early 19th century could invest.
Previous Thread :: Next Thread 
Page 1 of 4Goto to page: [1], 2, 3, 4 Next