Risk Parity

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Risk Parity

Post by willthrill81 »

prioritarian wrote: Wed Dec 08, 2021 9:38 am Treasuries are often positively correlated with stocks so wealthfront is talking their book. However, the premise of risk parity is not that stocks and bonds are always negatively correlated but rather that during periods of market panic longer duration treasury bonds show extreme negative correlation with stocks. There is a large economic literature on this flight to liquidity and some examples can be found by searching my comments.
While I don't have a Ph.D. in finance, nor do I play one on TV, nor did I sleep at a Holiday Inn Express last night, it seems to me that 'betting the farm' (and I'm not insinuating that you have suggested doing so) on a negative correlation between long-term Treasuries and stocks 'during periods of market panic' would not be prudent. We don't have many such periods in the post-Volcker era to investigate, and even if there is a negative correlation during such a period, that does not capture the magnitude of any potential counter-balancing effect. For instance, when stocks dropped nearly 30% in a single week in October of 1987, Vanguard's LTT fund (VUSTX) slightly dropped in value over the same period. In the weeks immediately following the stock crash, LTT went up by about 11% but still managed to finish that year with a -7% real return. The only times in the last 50 years that I'm aware of where LTT had a strong counter-balance effect when stocks fell significantly were the 2000-2002 and 2008-2009 stock downturns, but that was in no small part due to a drop (or perceived future drop) in interest rates. Can that same effect occur when interest rates are so low? The convexity of bond returns indicates that it is possible, but will the Fed drop interest rates from their already record lows if the economy nosedives from here, or will they use other tools to accomplish their goals, including the repression of Treasury yields by buying trillions of dollars of them?

That said, I have a small stake in a version of HFEA, which is built on the premise of the negative correlation you're talking about. And apart from the negative correlation, I think that there is far too much attraction toward TBM around here and that many investors, especially those with 30+ year investment horizons, would be better served with LTT. But I would have a hard time counting on a negative correlation between stocks and LTT during 'periods of market panic' with a big portion of my portfolio. That's just me though.
The Sensible Steward
EfficientInvestor
Posts: 580
Joined: Thu Nov 01, 2018 7:02 pm
Location: Alabama

Re: Risk Parity

Post by EfficientInvestor »

willthrill81 wrote: Wed Dec 08, 2021 10:16 am While I don't have a Ph.D. in finance, nor do I play one on TV, nor did I sleep at a Holiday Inn Express last night, it seems to me that 'betting the farm' (and I'm not insinuating that you have suggested doing so) on a negative correlation between long-term Treasuries and stocks 'during periods of market panic' would not be prudent.
In my implementation of risk parity, I agree that we should not assume negative correlation between stocks and long term treasuries during market distress (or any other time for that matter). I also think we don't have any reason to assume positive correlation either. I think anything outside the range of -0.1 to +0.1 is probably akin to market timing. I take the "I don't know nothing" approach and just assume a 0 correlation.

FWIW....If I assume a 0 correlation, assume that future volatility will be similar to past volatility, and assume a 6% return for US Stocks and a 1.9% return for LTT, PortfolioVisualizer spits out a forecasted tangency portfolio of 62% stock and 38% LTT.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Risk Parity

Post by willthrill81 »

EfficientInvestor wrote: Wed Dec 08, 2021 10:54 am
willthrill81 wrote: Wed Dec 08, 2021 10:16 am While I don't have a Ph.D. in finance, nor do I play one on TV, nor did I sleep at a Holiday Inn Express last night, it seems to me that 'betting the farm' (and I'm not insinuating that you have suggested doing so) on a negative correlation between long-term Treasuries and stocks 'during periods of market panic' would not be prudent.
In my implementation of risk parity, I agree that we should not assume negative correlation between stocks and long term treasuries during market distress (or any other time for that matter). I also think we don't have any reason to assume positive correlation either. I think anything outside the range of -0.1 to +0.1 is probably akin to market timing. I take the "I don't know nothing" approach and just assume a 0 correlation.

FWIW....If I assume a 0 correlation, assume that future volatility will be similar to past volatility, and assume a 6% return for US Stocks and a 1.9% return for LTT, PortfolioVisualizer spits out a forecasted tangency portfolio of 62% stock and 38% LTT.
While it might not satisfy a given definition of risk parity, I believe that leveraging something like the Permanent Portfolio, Golden Butterfly, etc. might be at least as worthwhile as more 'traditional' RP portfolios since they have more return drivers (e.g., size factor, value factor, gold), though I haven't sufficiently investigated such a possibility to reach any conclusion on its viability.
The Sensible Steward
exodusNH
Posts: 10344
Joined: Wed Jan 06, 2021 7:21 pm

Re: Risk Parity

Post by exodusNH »

prioritarian wrote: Wed Dec 08, 2021 9:38 am
nisiprius wrote: Mon Apr 26, 2021 11:23 am
slowandsteadywins wrote: Mon Apr 26, 2021 10:20 am(Wealthfront replied) "It's unusual for both asset classes to move in the same direction." ...
Only if you have an unusual definition of "unusual."

Here, I looked only at 1/2000 through 2/2021, a restricted period of time that I chose because I knew that stocks and long-term Treasuries did show negative correlation over that time period (unlike the 35 preceding years).

There is negative correlation, yes; in fact ρ = -0.301. But that does not mean they mostly moved in opposite directions.

The monthly returns of the Vanguard Total Stock and Long-term Treasury Funds moved in the same direction 117 times out of 253 or 46% of the time. I don't consider something that happens 46% of the time to be "unusual."

Image

Treasuries are often positively correlated with stocks so wealthfront is talking their book. However, the premise of risk parity is not that stocks and bonds are always negatively correlated but rather that during periods of market panic longer duration treasury bonds show extreme negative correlation with stocks. There is a large economic literature on this flight to liquidity and some examples can be found by searching my comments.
LTTs being negatively correlated has been true "recently" but there's no guarantee that will hold. In our lifetimes, it was always the case.
EfficientInvestor
Posts: 580
Joined: Thu Nov 01, 2018 7:02 pm
Location: Alabama

Re: Risk Parity

Post by EfficientInvestor »

willthrill81 wrote: Wed Dec 08, 2021 11:05 am
EfficientInvestor wrote: Wed Dec 08, 2021 10:54 am
willthrill81 wrote: Wed Dec 08, 2021 10:16 am While I don't have a Ph.D. in finance, nor do I play one on TV, nor did I sleep at a Holiday Inn Express last night, it seems to me that 'betting the farm' (and I'm not insinuating that you have suggested doing so) on a negative correlation between long-term Treasuries and stocks 'during periods of market panic' would not be prudent.
In my implementation of risk parity, I agree that we should not assume negative correlation between stocks and long term treasuries during market distress (or any other time for that matter). I also think we don't have any reason to assume positive correlation either. I think anything outside the range of -0.1 to +0.1 is probably akin to market timing. I take the "I don't know nothing" approach and just assume a 0 correlation.

FWIW....If I assume a 0 correlation, assume that future volatility will be similar to past volatility, and assume a 6% return for US Stocks and a 1.9% return for LTT, PortfolioVisualizer spits out a forecasted tangency portfolio of 62% stock and 38% LTT.
While it might not satisfy a given definition of risk parity, I believe that leveraging something like the Permanent Portfolio, Golden Butterfly, etc. might be at least as worthwhile as more 'traditional' RP portfolios since they have more return drivers (e.g., size factor, value factor, gold), though I haven't sufficiently investigated such a possibility to reach any conclusion on its viability.
I shy away from trying to satisfy a strict definition of risk parity, especially in such a low interest rate environment. I like to think I'm trying to implement "risk-adjusted parity" in my approach. I'm not trying to time the market or make strong assumptions about what will happen next, but I do want to take into account reasonable assumptions about forward looking returns, volatility, and correlation to determine asset allocation.
prioritarian
Posts: 485
Joined: Tue Jul 16, 2019 6:00 pm

Re: Risk Parity

Post by prioritarian »

exodusNH wrote: Wed Dec 08, 2021 11:09 am
prioritarian wrote: Wed Dec 08, 2021 9:38 am
nisiprius wrote: Mon Apr 26, 2021 11:23 am
slowandsteadywins wrote: Mon Apr 26, 2021 10:20 am(Wealthfront replied) "It's unusual for both asset classes to move in the same direction." ...
Only if you have an unusual definition of "unusual."

Here, I looked only at 1/2000 through 2/2021, a restricted period of time that I chose because I knew that stocks and long-term Treasuries did show negative correlation over that time period (unlike the 35 preceding years).

There is negative correlation, yes; in fact ρ = -0.301. But that does not mean they mostly moved in opposite directions.

The monthly returns of the Vanguard Total Stock and Long-term Treasury Funds moved in the same direction 117 times out of 253 or 46% of the time. I don't consider something that happens 46% of the time to be "unusual."

Image

Treasuries are often positively correlated with stocks so wealthfront is talking their book. However, the premise of risk parity is not that stocks and bonds are always negatively correlated but rather that during periods of market panic longer duration treasury bonds show extreme negative correlation with stocks. There is a large economic literature on this flight to liquidity and some examples can be found by searching my comments.
LTTs being negatively correlated has been true "recently" but there's no guarantee that will hold. In our lifetimes, it was always the case.
I never suggested that LTTs are negatively correlated with equities overall. I merely referenced the well studied phenomenon termed "flight to safety" or "flight to liquidity". This phenomenon is not only robust in the USA but also robust in other markets. Moreover, it has occurred during periods when inflation/rates were high and during periods when inflation/rates were low.

I am literally betting my entire portfolio on this phenomenon as just about anyone who diversifies equities with longer duration treasuries does.

Here is one post where I point people to one of the many papers in the economics literature that study this phenomena (there is even a figure that shows how robust this phenomenon is):

viewtopic.php?p=6317427#p6317427
User avatar
martincmartin
Posts: 900
Joined: Wed Jul 02, 2014 3:04 pm
Location: Boston, MA USA

Re: Risk Parity

Post by martincmartin »

There was a presentation on Risk Parity portfolios today at the Boston Bogleheads meeting, which got me interested in the actual performance of investments that are open to retail investors. There are some graphs in this thread on performance since 2013, and I wanted to update them, but the links seem broken. Of course, long term bonds had a terrible 2022, but I'd be curious how this has affected the risk parity portfolios in practice.
User avatar
nisiprius
Advisory Board
Posts: 52211
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: Risk Parity

Post by nisiprius »

prioritarian wrote: Wed Dec 08, 2021 4:46 pm ...Treasuries are often positively correlated with stocks so wealthfront is talking their book. However, the premise of risk parity is not that stocks and bonds are always negatively correlated but rather that during periods of market panic longer duration treasury bonds show extreme negative correlation with stocks. ... Here is one post where I point people to one of the many papers in the economics literature that study this phenomena (there is even a figure that shows how robust this phenomenon is)...
In 2022, stocks fell -24% from January through September, unequivocally a "bear market." An article hosted on nasdaq.com used the word "2022’s Stock Market Crash" and claimed that "That makes 2022 the fourth-worst year in the stock market’s history." Doesn't sound right to me but that's what the article said.

And yet long duration Treasury bonds--the longest readily available in ETF form, Vanguard's EDV--fell -44%. PortfolioVisualizer is calculating a +0.45 correlation for VTI versus EDV... +0.05 if we change the basis to "daily," which is low but most definitely not "extreme negative correlation."

I never believed the "negative correlation" or the "flight to safety" story as anything you could count on. But for the record 2022 falsified my own belief that you could pretty much count on Total Bond not falling more than -10% (It fell -17%). It also falsified my belief that you could count on Total Bond not losing money if held for its 6.5-year duration, when in fact it averaged -0.34%/year 4/30/2016 through 10/30/2022.
martincmartin wrote: Sat Jun 03, 2023 3:17 pm ...Of course, long term bonds had a terrible 2022, but I'd be curious how this has affected the risk parity portfolios in practice...
Here is a representative sample of how some "risk parity" mutual funds and ETFs performed, compared to the Vanguard Balanced Index Fund (60/40 stocks/bonds).

Source

Image

Not all that terrible. If I held such a fund I would chalk it up to "stuff happens," it wasn't wealth-vaporizing. But still, broadly, risk parity funds lost about -19% when a plain vanilla balanced fund was only losing -14%. The "weather" was bad for balanced funds, but risk parity made things worse.
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.
Logan Roy
Posts: 1838
Joined: Sun May 29, 2022 10:15 am

Re: Risk Parity

Post by Logan Roy »

nisiprius wrote: Sat Apr 24, 2021 8:42 am
corp_sharecropper wrote: Fri Apr 23, 2021 9:13 pm...plenty of AQR/Asness papers/publications...
And plenty of AQR/Asness funds, and the last time I looked, 38 out of 48 AQR funds underperformed their AQR-chosen benchmark--79% of them--and the average underperformance was -1.46% per year.

Image
watchnerd wrote: Fri Apr 23, 2021 9:28 pm ...But there are rational enhancements to indexing (e.g. modest use of leverage), many well-tested by institutional portfolios...
And for every Yale there are a dozen Harvards that have failed to outperform a dumb 60/40 policy portfolio. [Added] Correction, make that "for every Brown there are seven Harvards..." Source

Image

The closest thing to an "institutional portfolio" that I could personally invest in would be the Vanguard Managed Allocation Fund:
“Vanguard Managed Allocation Fund provides investors with a sophisticated, actively-managed, endowment-like portfolio at a low cost,” said Matt Brancato, head of Vanguard’s Portfolio Review Department,
with a portfolio that hits every buzzword, but has not hit its own benchmark:

Image

and has a two-star Morningstar rating, and has underperformed Morningstar's benchmark (orange) and the unsophisticated LifeStrategy Moderate (green) fund, while having noticeable higher risk as seen most visibly in 2008-2009.

Source

Image
To be fair .. All of these examples are meaninglessly short-term – while risk parity and All Weather strategies are designed to drag in single regimes (like stimulus-fuelled growth) in order not to lose significantly in regimes like stagflation.

As so many investors who got into the market around 2021 found out: the longer we've been in a single regime, the more cautious one should be .. 60:40 had such a bad time through the 1970s, in real terms, a repeat of that could set the stock/bond investor back 10-20 years .. Endowments can't be in that situation, because they have to meet spending commitments .. The choice to hold real assets and hedge funds is never going to bolster returns in markets just driven by stimulus.
User avatar
martincmartin
Posts: 900
Joined: Wed Jul 02, 2014 3:04 pm
Location: Boston, MA USA

Re: Risk Parity

Post by martincmartin »

nisiprius wrote: Sat Jun 03, 2023 6:03 pm Here is a representative sample of how some "risk parity" mutual funds and ETFs performed, compared to the Vanguard Balanced Index Fund (60/40 stocks/bonds).
Thanks. Here's a re-do of your graph on page 1, but extended through 2023:

Source

Image


SRPFX, Salient Risk Parity wasn't available in Portfolio Visualizer, I don't know why.
User avatar
nisiprius
Advisory Board
Posts: 52211
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: Risk Parity

Post by nisiprius »

With regard to Logan Roy's posting upthread, my overview of AQR funds as a family is seriously out of date, because QSPIX and some other did well last year. Constructing that overview is highly labor-intensive and an update is on my list of things that would be nice to do someday... but only when I get a Round Tuit.
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.
User avatar
martincmartin
Posts: 900
Joined: Wed Jul 02, 2014 3:04 pm
Location: Boston, MA USA

Re: Risk Parity

Post by martincmartin »

Image

Here you go
Post Reply