Risk Parity

[Sea Island]

I am trying to learn more about this strategy and what it means. Does anyone have a resource they could recommend for learning more about this? Thank you.

[rpadebet]

Hopefully you already saw this from bridgewater (this strategy has its intellectual genesis from here)

https://www.bwater.com/home/research--press/risk-parity--portfolio-construction-white-papers.aspx?agreed=64913

 

In my opinion it extends the concept of risk=volatility=standard deviation and basically generates portfolio allocations where the low volatility stuff is highly weighted compared to the high volatility stuff. If you think in terms of bonds and equity, as opposed to the mutual fund standard 60%/40% benchmark, this strategy gives about 90%/10% allocation.

 

Almost all backtest's I have seen employing this strategy have a good historical performance, which I think is mainly because Bonds have been in a secular bull market now for 30 years almost. It is anybody's guess how it will pan out in the future.

 

 

[thepupil]

On a very broad level, it means having the same amount of "risk" (defined as volatility) in each asset class in order for those asset classes to "act" as proper diversifiers. My understanding comes from Bridgewater's writings as well.

 

According to them, a portfolio of bonds, can be expected to return its initial yield to maturity. If you buy AGG your return over 5-6 and ten years will be the YTM because even if prices change, coupon reinvestment should (before defaults) bring your yield to your initial YTM.

 

Because  bond yields are generally low and bond "risk" is generally low (even 1994, which saw the fed raise rates like 300 bps in a short time, bonds made close to 0%if i remember correctly), Bridgewater would tell you that you don't get enough return from bonds to act as a good diversifier unless you lever them, i.e. you match or pare your expected volatility with your equity book. They argue levered bonds with more duration are a better diversifier of returns. The 2008 experience would confirm this as if you went in levered long of duration (of the government sort, not the credit sort), your bond book would've acted as a great divirsifier.

 

Like rpadebet, I think they are fighting the last war and the strategy does not acknowledge that bonds and stocks can move down in tandem like they did in the early 70's.

 

Buy long duration mREITS if you think what I've told you has any merit.

[Sea Island]

Thank you both.

[TwoCitiesCapital]

On a very broad level, it means having the same amount of "risk" (defined as volatility) in each asset class in order for those asset classes to "act" as proper diversifiers. My understanding comes from Bridgewater's writings as well.

 

According to them, a portfolio of bonds, can be expected to return its initial yield to maturity. If you buy AGG your return over 5-6 and ten years will be the YTM because even if prices change, coupon reinvestment should (before defaults) bring your yield to your initial YTM.

 

Because  bond yields are generally low and bond "risk" is generally low (even 1994, which saw the fed raise rates like 300 bps in a short time, bonds made close to 0%if i remember correctly), Bridgewater would tell you that you don't get enough return from bonds to act as a good diversifier unless you lever them, i.e. you match or pare your expected volatility with your equity book. They argue levered bonds with more duration are a better diversifier of returns. The 2008 experience would confirm this as if you went in levered long of duration (of the government sort, not the credit sort), your bond book would've acted as a great divirsifier.

 

Like rpadebet, I think they are fighting the last war and the strategy does not acknowledge that bonds and stocks can move down in tandem like they did in the early 70's.

 

Buy long duration mREITS if you think what I've told you has any merit.

 

I believe you both are being very quick to write off a firm that who has been largely successful in risk parity AND alpha generation AND has better understanding of global macro scenarios then the Federal Reserve.

 

As mentioned above, risk parity is based on volatility, leveraging asset classes to the same volatility, and a very large amount of diversification. Its not as simple as stocks and bonds. Throw in precious metals, real estate, futures, domestic and foreign equities, currencies, high yield, corporate bonds, sovereign bonds, etc.

 

Secondly,  I'm sure we all agree that risk parity has generally benefited from the bull market in bonds.  In general, a reversal will result in higher volatility in bonds which will result in less leveraged exposure to them as an asset class meaning that they will be far less of a counterweight then they were a boon if the strategy is implemented correctly. Seriously, Bwater is the worlds largest hedge fund for a reason and it's for a whole lot more than "the bull market in bonds".

[thepupil]

I don't write them off; and did not mean to do so. they are the best macro guys out there. i love Dalio's writings and the writing of the broader organization.  And they have compounded capital at a crazy rate , and huge amounts of it.

 

I just don't totally buy into the risk parity thing and am not going to manage my capital in that manner.

 

Edit: but i did not mean to write off Bridgewater's capability in doing so.

[rpadebet]

Zach,

I wasn't writing off Bridgewater  :). I have a lot of respect for their analysis, performance and intellectual contributions. I think Ray Dalio is probably one of the few who gets the macro picture right AND is able to profit from it.

 

The question was about risk parity strategy and its practical implementation. On a certain theoretical level I agree that one should have their asset allocations  such that they have equal risk contributions from all underlying's. That part makes sense - If I had knowledge before hand what the risk of each underlying was ( at least in relation to one another) I would be able to then weight them appropriately according to this logic.

 

The problem I have though is with its practical implementation. Risk is equated with volatility (to me it makes intuitive sense to think of risk as something which can cause a permanent loss and not something that wiggles a lot) and then volatility is equated to standard deviation of historical returns ( I have an issue with that, because I don't think +10% should be same as -10% )

 

So I am not writing them off. I just have a nuanced disagreement with their and others implementation of it.

[Palantir]

Correct me if I'm wrong, but my understanding of Risk Parity is:

 

Stock: 8% Return, 15% Risk

Bond: 1% Return, 1.5% Risk

 

Lever up bond by 10x, we get:

 

Bond = 10% Return 15% Risk

 

This should replace part of the stock allocation. In other words, turning your bonds into stocks, but with lower risk.

[thepupil]

Generally yes that's the basic reasoning.

 

But as zach pointed out, it's an oversimplification of the implementation and Ray Dalio isn't worth billions for just levering up every asset class.

[CorpRaider]

Two words:  transcendental meditation.