Criticism

The objections to this strategy are a different matter from the mistakes that happen when it is carried out; those are in the chapter Staying the Course under Typical Mistakes. This chapter is about the strategy itself: what is publicly disputed about the split, about the role of the bond leg and about the backtests.

Each entry names the objection, this page's counter-position and where it can be read.

The split comes out of a bond bull market

The objection that the common splits come out of a period of falling rates stands in the original thread in February and March 2019 already. The period spans roughly four decades in which long-dated bonds rose almost throughout. A weighting that looked good in that environment is therefore not a statement about an environment of rising rates. The objection is not against the idea of two blocks but against the number it is circulated with.

Where the numbers come from is documented: 40/60 comes out of the idea of equal risk contributions, because the swings of stocks were roughly one and a half times those of bonds (the original thread). The switch to 55/45 lowered the bond share once longer backtests reaching into the 1950s were available (backtest from 1955, reasoning behind the switch). That 55/45 does better in every sub-period than 40/60 has been recomputed later by ZahlGraf on data from 1943 (part 5). The adjustment therefore ran away from the bond bull market and not into it. That the number remains a child of its rate environment does not change that.

This page's counter-position: the split stands here as a preset with a reason rather than as a prescription.

The two blocks move against each other only in some regimes

The core of the strategy is the assumption that stocks and long-dated bonds catch each other. That assumption is not a constant: the German-language series computes the rolling co-movement per rate regime on data series from 1943 and finds, for the years 1965 to 1985, a picture in which both sides could lose together. For a whole rate regime, then, the relationship the strategy rests on is not supported but contradicted.

This page's counter-position: it carries the counter-movement as a property of particular decades and not as a fixed quantity; the derivation in the chapter The Rule sets out what it depends on.

Bands against dates is an open question

The objection that a band check has problems of its own is in the original thread already: it trades precisely when markets are turbulent, and it trades more often than a calendar does. The calendar in turn trades even when there is nothing to do. Which of the two does better has not been settled in public, and both versions circulate side by side.

This page's counter-position: both ways stand there as parameters, with their costs described in the same currency, and neither is carried as the right one.

2022 is not a slip

In 2022 stocks and long-dated bonds fell at the same time and sharply, which is exactly the case the strategy does not provide for. It was not new: the original thread already carries a backtest from 1955 in which the bond leg does badly in the decades of rising rates from the 1950s to the early 1980s. The decades in which the assumption did not hold are therefore in the strategy's own numbers, and every backtest looks different depending on whether it includes those decades.

This page's counter-position: the community backtests in the chapter Backtest report the result for every possible starting point and name their limits; the analysis by modern_football starts in 1986 and so leaves out the decades of rising rates. This backtest shows the recent years on real prices.

The account type carries part of the result

The original thread describes the strategy inside a Roth IRA, an account in which rebalancing triggers no ongoing tax (IRS Publication 590-B). Outside such an account that does not hold: every rebalancing sells, and the rhythm produces taxable events several times a year by construction. A published result that does not name the account type is therefore not transferable, and not at the margin but in the size it shows.

This page's counter-position: the chapter Implementation and Taxes names the account type as a quantity of its own, and the backtest computes without tax. The tax classification there is rough, states its date and is no substitute for tax advice.