Allocation in Volatile Markets: procyclical or countercyclical?

First off, here is the relevant section from the YouTube version (from 24'40" until 36'40"):

Section on countercyclical and procyclical adjustments to a RP portfolio: 24'40" to 36'40"

And in podcast form, here is the same section, from 21'53" until 33'53".

Brief summary:

This section concerns different ways to implement a RP portfolio, centering around how to size and allocate a RP portfolio as volatility changes. If volatility hits bonds especially, with their price dropping - do you buy more? or, to compensate for their increased volatility, do you lessen your commitment to bonds?

Adam Butler breaks responses down to two camps: the countercyclical and the procyclical, based on whether your portfolio adjustments are going against the movement of the markets by selling off what’s going well and buying what’s doing badly, and the procyclical approach where you adjust your portfolio in the direction the market is moving.

He identifies the countercyclical approach with Bridgewater which holds that portfolio construction is a function of fundamental economic drivers. Bridgewater sets a strategic portfolio with fixed weights for asset classes and adjusts the portfolio to meet those longer-term, more stable allocations. In March 2020, for example, as equities plunged and Treasuries rose, they'd be selling bonds to buy equities to get them back to their targets.

Contrast that with some of the other funds like AQR's or Invesco's where the constitution of the portfolio changes over time due to changes in volatility and estimated correlation of component assets. These funds would sell assets in volatile times in order to bring the total volatility of the portfolio down, with the idea of expanding when markets are calmer. To use the March 2020 example, portfolios following this approach would pursue a general sell-off and move to cash to stamp down the portfolio’s volatility.

Implications for DIY investors:

This section of the discussion caught my attention since my understanding has always been that RP was defined by the countercyclical, Bridgewater approach using stable, long-term allocations. Given my needs and horizon as an investor, this approach seems to make more sense. You buy low and sell high, and through this kind of continued rebalancing to targets, reap a reward for rebalancing.

To be fair, a weakness of this approach is that you are implicitly betting on a return to the mean. In March 2020, this could have easily been a situation where you were selling off bonds to buy equities at a 30% discount, but there was no guarantee at the time that the markets would snap back. Japanese investors from the late 1980s might still be waiting for the rebalancing alpha if they followed this approach.

Meanwhile, the more active, procyclical approach seems favored by professionals who have a shorter investing timeframe. This is the nature of the beast when clients want to see results and are paying for active management. But given professional oversight and expertise, there could be examples where being procyclical works out and helps to meet investor demands. Perhaps the risk parity fund is the stable bedrock of a client’s portfolio; they might want that to act like bedrock above all and look elsewhere for higher returns by chasing sources of alpha.

In terms of fidelity to Risk Parity principles, it's hard to say which is more correct, and I suppose both have their merits based on objectives. As Butler says, “there’s more than one way to skin the risk parity cat” (funny aside: I used this metaphor with my 8th grade students once, and they were shocked and wondered what I had against cats. I had to convince them that this was an actual idiom and not something my sick mind made up). It seems to come down to your threshold for changing allocations: do you want to be sensitive to changes and adjust frequently (procyclical), or are you less sensitive and favor more enduring guidelines (countercyclical). What say you?