Chadnudj wrote: ↑Fri Apr 14, 2023 8:29 am
I think nisiprius's point with B was more that backtested data (like the Trinity Study, the long term 6-7% real return of the S&P 500, etc.) can be trusted as a forecast for the future, and thus "things will stay the same" (or, perhaps more precisely, "the future over the long term will behave much like the past did").
I'm not sure that is exactly what was intended with (b), but if so then my comments very much apply. Meaning I think the proposition phrased that way is very unreliable, and should be down close to (d) in terms of unreliability.
And actually, I think phrased that way, it is just really unlikely to be true. There are very good explanations as to why real returns on both equities and bonds would be drifting down over time (subject of course to shorter-term variation), which in fact would just be a continuation of a trend that has been happening for centuries.
So to me, the idea of projecting them out as staying level instead is particularly unlikely to be a good projection.
but B posits that the best practices of the past -- namely Boglehead ones of dollar cost averaging into and holding low cost index fund portfolios allocated to an appropriate equity-fixed income proportion and rebalancing -- are likely to continue working into the future.
That is mixing together a lot of issues.
I would agree the general concept of saving for retirement early and often by investing in diversified risky assets remains a good idea. The most efficient way to do that, though, is one of the things that has evolved over time. That is true in part because new types of assets that are consistent with that approach have been created, costs have evolved (usually lower, which is good), and so on.
I also agree that using very-low-risk assets in combination with a portfolio of risky assets to manage certain important forms of risks, particularly sequence of real returns risk, makes sense for retirement savers. I note that the available fixed income assets are also something that has evolved a lot over time, and today different people often have different options available to them. So this is definitely an area where efficient implementation rationally depends on the options available to you at the time.
I do not agree the idea of adopting a single unchanging equity-fixed income proportion and rebalancing to that target ever had a very good foundation. This was a sort of modification of modern portfolio theory except it unnecessarily confused risky fixed income assets with very low risk fixed income assets, and in reasonable models those two categories of fixed income have very different purposes. It also is typically argued for using a very unrealistic model of investor psychology.
However, it is true that there are a lot of people here who argue that a static approach involving the sorts of assets available circa 1990 is really good and all you need and so forth. As evidence, they typically cite a combination of their own anecdotal experience, backtesting of some sort, and maybe something Bogle once said.
This is precisely the sort of argument I was referencing above. The anecdotes, including the backtesting, these people are offering typically implicitly depend on the good results for nominal USD bonds that happened during the era of decreasing interest rates, aka the "golden age" of bonds. Rationally, people should understand that cannot possibly be sustained. Rationally, people should also understand that decline in interest rates means risk-adjusted real rates on bonds going forward are likely to be much lower. Rationally, people should understand that very likely means risk-adjusted real rates on equities are also likely to be lower.
OK, so using nisiprius's framework, I am very much arguing that some form of (c) should be put well ahead of (b). I don't think we can "know" these things, but I do think there are very good reason to believe that the real returns on risky assets like stocks and risky bonds will very likely be lower over the next, say, 30 years than they were over the last 30 years. And I think people who one way or another insist it is instead likely they will be around the same do not have a reliable argument for that sort of assumption.
And I also think people who ignore the evolution of available assets and say a combination of whatever was already available circa 1990 do not have a very good argument for doing that.