> > Dahiya and Yermack found that the performance of the typical endowment fund [from 2009-2016] was so poor that it would have earned substantially higher returns if its trustees had followed a simplistic investment strategy of holding 100% Treasury bonds and taken no equity market risk whatsoever.
As a non-expert this intrigued me so I found an online calculator for that period (https://dqydj.com/treasury-return-calculator/) and was confused when it showed an annualized return of 1.72%.[1]. But then I read the actual paper where they make it clear their "simplistic strategy" is investing in a "CRSP 10-Year U.S. Treasury Bond Index". (See page 10 of PDF at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3291117)
Digging further, as explained on this bond index calculator, https://portfoliocharts.com/bond-index-calculator/, a bond index fund uses mechanical rules, similar to an equities index fund, for buying and selling bonds, making money on changes in the market price, which can be greater or lesser than the nominal coupon return. IOW, such an index isn't simply buying and holding onto treasuries.
If you download the spreadsheet from the above link (the calculator is a spreadsheet) you'll see that YoY swings are huge, ranging from -4.5% in 2009 to 17.5% in 2011 for 10-year treasury index.[2] Those very good years are why a treasury bond index can achieve such amazing returns over longer periods--greater than 2x the nominal + reinvestment return, without using derivatives for leverage.
Now, I initially assumed the YoY bond index returns would be negatively correlated with equity returns. That's the point, right? But that doesn't seem to be the case; rather, it looks like a mixed bag. Here's the YoY returns of the example treasury bond index as compared to the S&P 500 (from https://ycharts.com/indicators/sp_500_total_return_annual):
bond | equity
-----------------------
2009 | -4.5% | 26.46%
2010 | 9.6% | 15.06%
2011 | 17.5% | 2.11%
2012 | 4.8% | 16.00%
2013 | -6.9% | 32.39%
2014 | 10.3% | 13.69%
2015 | 1.8% | 1.38%
2016 | 1.6% | 11.96%
2017 | 3.2% | 21.83%
On their face the years 2010, 2014, 2015, and possibly 2012 don't look negatively correlated. Maybe they are for some technical reasons, but it makes me skeptical about the degree to which such a bond index fund hedges equity risk.
As a total layman in this domain but with some experience pouring through these sorts of academic research notes, what this tells me is that the Dahiya and Yermack paper is interesting but hides some significant assumptions and, presumably, some pretty deep theoretical disputes within the academic and investor communities.
[1] 1.72% is for January 2009 to January 2017. For June 2009 to June 2017 the return shown is 2.75%. That suggests we should be a little skeptical about the effects of the starting and stopping points of any analysis.
[2] AFAICT, actual CSRP models and data are proprietary. The calculator and historical returns are based on that author's own model. (EDIT: Table A3, page 62 of the Dahiya and Yermack paper give the benchmark returns. I'm just eyeballing them, but the 10-year treasury figures seem to match up well with the returns from the portfoliocharts.com spreadsheet.)