chris conte wrote: Fri Oct 17, 2025 12:31 pm
abc132 wrote: Fri Oct 17, 2025 11:53 am
I know that over 70% of the money (real) I have ever invested is now in bonds. This is not because of a 30/70 AA but rather because I chose to take stock risk. This was not a feat of luck, but rather something that has always happened over a 30+ year accumulation career and retirement sequence.Those making the biggest prediction are those relying on some non-historical and low probability sequence to justify their choices. I’ll take high probability bet with a significant pay-off over low probability bet that is has very low chance of a significant payoff. Yes it is possible to make a low probability bet and benefit, but betting on bonds over stocks for long periods of time is a risky bet. In the hypothetical future sequence where bonds do better, choosing stocks will still have been the better choice. The expected value of stocks over a lifetime of accumulation and decumulation is higher than that of bonds.
I won’t care when stock risk shows up because I took stock risk. I also value bonds (currently a 70/30 AA) and by not liking or disliking assets I can manage risk instead of managing my preferences.
We prefer a lower volatility asset so it would be irrational to suggest stocks will be priced with lower expected returns over long periods of time. Thus we have high confidence in an equity risk premium over decades, even in the instance where that bet does not pay off.
I agree completely.
Of course, the historical data should be viewed with a healthy dose of skepticism. But we also cannot ignore the reality, the most reliable data spanning over a century overwhelmingly shows that equities outperform bonds in the long run, and it is, quite frankly, the only reliable data we currently possess. I believe that we should just be pragmatic and base our decisions on the best available evidence, however imperfect. Time will tell if that’s the right decision.
My primary frustration is the tendency for people to dismiss this evidence without offering a single constructive counter-framework for portfolio construction. Simply generating panic about the long-run performance of equities serves no practical purpose without a framework to replace it. And it seems like none of the critics possess the conviction to openly advise an investor in the accumulation phase of life to be primarily invested in fixed income.
It’s not about ‘reliability’ as in whether the numbers are right but how (and what) one should project into the future. Very few ordinary people are interested in financial market history for its own sake (I may be a partial exception ), they use it to project. And we do need to project and ‘predict’ in some sense, as many posts as there are here trying to win arguments by saying ‘you’re predicting, I’m not’. Especially the subset which effectively says projecting past return as best estimate of future results isn’t ‘predicting’. There are two basic issues with past results on stocks v bonds.
1. 100 yrs US only is actually a pretty small sample of 30 yr investment horizons, ~3 throws of the dice (using overlapping periods is junk stats, it holds the ‘big picture’ constant effectively assuming what we care about is higher frequency noise when in reality change in the big picture over a long period is the most significant form of ‘risk’*). There’s no complete solution to the small sample problem. Looking further back or at other countries can change results considerably, but it’s also reasonable to point out how different the US was in distant past or other countries were.
2. This one is more addressable. We’re almost surely IMO better off correcting results for starting relative valuations of stocks and bonds. See the graph in my earlier post (pg.3), relative real yield of stock v 10 yr T. Current expected return for stock, by fundamental relationship of E and P is unusually narrow compared to 10 yr real yield now. The assumption expected ERP is constant at what it averaged in realized terms in past, or even that it’s just ‘unstable’ is incorrect. The market gives at least a broad signal what it is on expected basis via pricing. It’s pretty narrow now.
Of course it’s not a binary question of stocks or bonds. And there are actual analytical frameworks to determine that sliding % based on utility function, E[ERP], variance. The simplest, Merton Fraction, is discussed above. The point being not that there’s an easy equation, or more complicated black box, that just gives the real-world answer. It’s that people who ‘choose’ in a binary way, which on this forum is virtually always ”why any bonds?’ virtually never ‘why any stocks?’, are ignoring a, or the, major theme of modern financial theory: how to determine allocation based on risk v risk tolerance for the period we *don’t* know, the one that matters, the future. Where 100% (or more) stock could be the answer but it’s not somehow simply determined as 100% by stocks having higher historical realized, or E[r], than bonds.
*another easy to dismiss claim is that we should look at return on TIPS to a horizon other than their maturity. That’s introducing artificial variance in bond return. One of the basic risk advantages (in return for lower E[r]) of bonds being that they *do* have a maturity when stocks don’t. We can package bonds up in a way that has no maturity (funds), but there’s no way to package up stocks in way that gives them one. Ignoring the higher certainty of value (complete if really ‘credit riskless’, though they are not really IMO) of TIPS at maturity is biasing the choice in favor of risk.
