20% Equity vs 100% SPY

shipwreckedandalone

I enjoy reading the MFO articles. Anybody have thoughts regarding the article mentioning .....a 20% equity portfolio translates into receiving 60% of the returns of an all equity portfolio with about 25% of the volatility? Just a Thinking Fast and Slow type observation....volatility (SD) is a annualized metric that would remain in this case relatively consistent over time around 25% but the 60% yearly returns vs SPY compounded over a 30 year retirement horizon would not be 60% after 30 years. A significant difference in return. IOW's the value one receives from low SD does not compound over time. Although there is value inherent.

yogibearbull

A 15-85 or 20-80 stock-bond portfolio is also called parity portfolio that means that the SDs of stock and bond portions are nearly equal.

There are several funds/OEFs that offer this mix - FASIX, VASIX.

Hedge-fund managers may put high leverage on such parity portfolios.

Some may use these as next level of risk up from core bond funds; another way is core-plus. Idea is to add some stocks or HY to core/inv-grade bonds.

There are some multisector and HY bond funds with up to 20% equity, RPSIX (MS), FAGIX (HY).

So, there is something to this 15-20% equity solution.

shipwreckedandalone

fyi.. ...
since inception VASIX vs SPY runs 56% of CAGR and 29% of SD. 10/1994. Lo as 16% Equity.
since inception FASIX vs SPY runs 55% of CAGR and 31% of SD. 2/1993. Hi as 25% Equity.
since inception VBMFX/SPY vs SPY runs 60% of CAGR and 28% of SD. 2/1993.

The 60/20 may also be a good trade off depending on market valuations, business cycle and future perceptions of return (or loss). 100% equity portfolio in retirement normally not used.

carew388

BACPX is another option.

msf

I believe the unnamed article referenced is one from Prof. Snowball:
Investing in 2022: The Indolent Portfolio (Jan 2022), citing
The case for a stock-light portfolio, version 4.0 (April 2021), or
Mutual Fund Commentary Nov 1, 2014

They all reference the same T. Rowe Price analysis of data from 1949 to 2013. For more current data and something a bit more interactive, Portfolio Visualizer has an efficient frontier tool.
https://www.portfoliovisualizer.com/efficient-frontier

Worth noting is that the TRP analysis probably uses either Treasuries or total US (IG) bond market for the fixed income portion. Those have virtually zero correlation with the US stock market.

This matters because as one uses "junkier" bonds (moving from IG funds to core-plus funds, multi-sector, and ultimately to HY funds) the correlation with stocks increases substantially. This in turn reduces the diversification effect of the bonds and thus does less to temper the equity volatility. Interestingly, according to PV, using HY rather than IG bonds results in a lower return efficient frontier. That is, for a given target volatility, HY + equities returns less than IG + equities.

This PV output shows the efficient frontiers and optimal allocations (i.e. highest Sharpe ratios) using data from 1987 through 2021.

The higher efficient frontier (EF) curve, i.e. the one with the higher return for a given level of volatility, is the IG + equity portfolio. Mixing HY and equity doesn't do as well. An 80/20 IG/equity mix, according to the curve, has a standard deviation of around 4.4%, or around 30% of the 15.3% volatility of a pure equity portfolio. It also has an average return of 7%, which is about 60% of the 12.3% return from the equity portfolio. Figures that are consistent with the TRP results.

You can play around with this, e.g. using a portfolio of HY, Intermediate Treasuries, and equities. The result of that experiment suggests that once one goes above 40% in equities, one is better off leaving out the junk bonds.

davidrmoran

FASIX 1/3/5/10/15y = 3.76 7.66 5.53 4.78 4.46

VASIX 1/3/5/10/15y = 1.56 7.49 5.61 4.92 4.60

(if I am reading the M* data right)

so for me no

msf

Is the object is to get the maximum return for the minimum volatility (which IMHO is not the same as risk, but a subject for another day) with a traditional fixed allocation? That seems to be what's implied by the observation that a 20/80 allocation has the potential for 60% of the returns with 25% of the volatility.

Again using Portfolio Visualizer, one can see that the optimal allocation (using 20/20 hindsight) varies depending on the calendar years used. "Optimal" here meaning highest reward/"risk", i.e. highest Sharpe ratio. Using VTSMX and VBMFX as proxies for the US stock and bond markets, the optimal allocations (per PV) are:

Jan 2019 - Dec 2021 (3 years): 20.4/79.6, Sharpe ratio 1.64 (same as 20/80)
Jan 2017 - Dec 2021 (5 years): 21.8/78.2, Sharpe ratio 1.28 (vs. 1.27 for 20/80)
Jan 2012 - Dec 2021 (10 years): 26.5/73.5, Sharpe ratio 1.35 (vs. 1.33 for 20/80)
Jan 2007 - Dec 2021 (15 years): 13/87, Sharpe ratio 1.15 (vs. 1.11 for 20/80)

Even though the optimal allocation varies widely (from 13% stocks to 26% stocks) depending on years and time frame, a fixed 20/80 allocation appears to come very close to optimal in all periods. So it's likely not productive to try to vary one's allocation based on future predictions.

The Sharpe ratios (per PV) over 3/5/10/15 year periods for VASIX, FASIX, and BACPX are lower than the fixed 20/80 Sharpe ratios above:

Period	VASIX	FASIX	BACPX
3 years	 1.43	1.28	1.46
5 years	 1.15	1.01	1.10
10 years 1.22	1.10	1.27
15 years 0.86	0.76	0.66

https://www.portfoliovisualizer.com/backtest-portfolio

If one periodically rebalances one's portfolio, then there's no additional work in buying two index funds and maintaining a 20/80 mix. But if one wants a set-and-forget portfolio, then a single fixed allocation fund (or a target date fund if one wants the mix to gradually change) or a cheap robo advisor that uses basic broad market index funds could be the best and simplest choices.

shipwreckedandalone

Thanks for the contribution to this thread. I tried to retrieve/search the T Rowe Price Report Fall 2004 article referenced without luck. Behind a wall. Wish I could find it. The article would give us more clarity and specific allocation holdings. However, I did notice the column headings from the 4/2021 article 1949-2013 study is using 20% equities 50% bonds and 30% cash producing the 6.8% annual number referenced. The Nov 2014 article used column headings of 30% Short (assume to be in error?) also producing 6.8% annual return. (not the 20/80 allocations we have been discussing).
As far as what is held within the bond holdings of the 80% (or 50%)...... references are made to Fidelity Asset Manager series and RPSIX for comparisons but as you said we do not know what TRP used in the study.
Interesting subject matter.
References are also made to M* Conservative Retirement Saver Portfolio using the Lifetime Allocation Indexes. Christine Benz wrote an article providing suggested funds with allocation %'s which I plan to analyze for metrics.

davidrmoran

AOK for all!

>> if one wants a set-and-forget portfolio

(unless too aggressive at 30-70)

msf

See my concluding sentence in the thread Seeking Suggestions for Vanguard Asset Allocation Funds: "Or even push it to a 30/70 fund (VTINX)."

This vanilla Vanguard fund of funds returned more than AOK over one day (YTD), one year (5.03% vs 4.37%), three years (9.46% vs. 9.20%), five years ( 6.78% to 6.66%), and 10 years (5.81% to 5.41%).
http://performance.morningstar.com/fund/performance-return.action?t=VTINX (compare with AOK)

Its volatility (std dev) was lower over three years (5.87% vs. 6.34%), five years (5.02% vs. 5.25%), and ten years (4.42% vs. 4.62%).
http://performance.morningstar.com/fund/ratings-risk.action?t=VTINX

The main area in which VTINX (or more generally, retirement income funds) falls short is tax efficiency. These funds are designed to generate income, and are thus not great in taxable accounts.
http://performance.morningstar.com/fund/tax-analysis.action?t=VTINX

davidrmoran

yeah, AOK is burdened for having more foreign and more cash somewhat (probably the case historically too)

interestingly, lower Sharpe, slightly, and higher ulcer, slightly (I may not be understanding something here)

msf

Fair point about the foreign allocation. A drag over the past decade, could be a plus going forward.

I did notice the cash component which is odd. One of the benefits of ETFs (including CEFs) is that unlike OEFs they don't need to keep cash around for redemptions. Though a few old ETFs like SPY and QQQ are structured as unit investment trusts that have to keep cash divs around until they distribute them periodically.

yogibearbull

iShare allocation ETF series AOK, AOM, AOR, AOA follows S&P Target Risk series indexes and cash held is dictated by that.

carew388

Using AOR to put some cash to work.

msf

What ever happened to indexes that actually measured something as opposed to defining investable portfolios? Oh well.

Saying that the AO* series just invest in various combinations of IVV, IJH, IJR, IDEV, IEMG, IUSB, and IAGG doesn't answer the question about why there's so much cash in these ETFs. It just pushes the question down a level: why is there so much cash in IUSB (and why is IAGG short cash)?

De facto, these underlying funds have particular cash allocations. But are those allocations what they're supposed to be? (Part of the answer may be the difference between how M* and BlackRock count cash, but I haven't looked more closely into that.)

S&P Target Risk Index Series Methodology

IUSB portfolio allocation, per M* 6.44% cash
IUSB portfolio allocation, per BlackRock 0.69% cash
(That must be net; the top holding of IUSB is BISXX, 10.07% per factsheet.)

IAGG portfolio allocation, per M* -0.24% net cash
IAGG portfolio allocation, per BlackRock 0.88% cash

davidrmoran

>> foreign allocation. A drag over the past decade, could be a plus going forward

So Tillinghast has always thought too, over the decades, for some reason.

As I do my January rebalancings real and imagined, I'm trying to fathom what gaps (important gaps, not just on paper) VONG and DSTL 50-50 have for my (half or 2/3) equity holding.

carew388

Respectfully, why VONG as opposed to QQQ?

davidrmoran

5x as many holdings (yes, not equally)

carew388

+1

davidrmoran

I always have serious trouble deciding b/w VONE and VONG, as the outperforming latter ETF also has lower UI and less drawdown, presumably because of gogo tech stocks' constant bounceback and current-market invincibility of hope ...

Also trying to see where DSTL (a nontrivial diversifier for sure, plus an MFOP GO, ffs) adds real value.

Speaking of gogo, I recently asked my consultant kid if he owned any Apple, and he said 'A lot, only because it is so heavily in our funds'. Now, if I had taken a really healthy bite when he did a paper on it as an econ undergrad 19y ago (too expensive, look at the runup, said dad), or done the same 11y ago when he was in business grad school (get serious, look at the runup, way too expensive now, said dad), I would be able to give him and his sister serious downpayments on next housing. Damn.

carew388

+1 I missed the boat when George Gilder recommended JDS-Fitel(later JDSU) as the next Intel in a Forbes supplement I received in 1997-1998! Of course, I would have had to sell at the right time as well.

shipwreckedandalone

In the past I have found AOM useful for a tactical trade for smallish position size if one doesn't want to pay the $75 in some brokerages to get into VTINX or similar slightly superior fund. Or have the need to dollar cost average into a position over several days. Commission free trades. If one needs an AOM-like balanced vehicle. The iShares series is useful from that regard.

finder

Of course one can make money in some bonds, but is it a good idea to put 80% of the portfolio to treasuries when the interest rates may go up? Historical data since 1980 show falling rates, from about 15% in the past to about 1% now. The situation will be different when rates rise.

PRESSmUP

carew388 said:

+1 I missed the boat when George Gilder recommended JDS-Fitel(later JDSU) as the next Intel in a Forbes supplement I received in 1997-1998! Of course, I would have had to sell at the right time as well.

I got caught in the JDSU web. The late Joe Battipalia referred to JDSU as "Just Don't Sell Us". I lost some money, but learned a valuable lesson.

carew388

+1