Pumping Vol in Swiss Retirement Portfolios

The following post details my experiences in getting VIAC, a Swiss provider of third pillar retirement portfolios, to do weekly rebalancing. It’s a repost of what I’d already contributed to Reddit and LinkedIn, so it’s written for a wider audience and you find it a bit too detailed in parts.

The Background

Over the past few years I’ve been digesting a number of books and academic papers on investing, with a particular interest in ideas originally rooted in sports betting and later applied to financial markets - particularly Shannon’s Demon, the Kelly Criterion, and Parrondo’s paradox. Some of you might be reminded of Poundstone’s great book “Fortune’s Formula”, and you wouldn’t be far off.

The long and short of these is that it should be possible to 1) generate returns in otherwise flat markets, 2) there exists an asset allocation so as to maximize the expected log returns of wealth (in other words, maximize wealth growth over time) and that 3) returns can be extracted not just in flat / sideways markets, but also with negative expectation trades, e.g. in environments where both stocks and bonds are expected to sink in the near term

Why is this possible?

A good definition comes from the EDHEC Business School:

The potential source of additional performance because of the simple act of resetting portfolio weights back to the original weights is referred as the rebalancing premium. It is also sometimes known as the volatility pumping effect or diversification bonus because volatility and diversification turn out to be key components of the rebalancing premium

How can this be applied to Pillar 3a?

VIAC doesn’t exactly let you click a button to do weekly rebalancing. What can be done instead is setting up a weekly recurring standing order to invest into your portfolios (you have more than one, right?).

I’m currently filling up #3 of 4, maybe 5, and have been sending money in on a weekly basis ever since Q1/2022 into that third portfolio. As such, the first two portfolios did not see active contributions for most of the year.

Until I realized I could just get my bank to transfer 1,00 CHF into each of them - as any inflow is enough to trigger a rebalance free of charge.

And as the current market environment is somewhat volatile - both for stock, real estate and foreign exchange markets - this should be as good a time as any.

I’ve since created a reporting for the Sunday before the start of this, and the Monday before initially writing up these experiences.

What stands out:

  1. With the filled-up and otherwise inactive portfolios, the “rebalancing premium” is about 1.45% and 1.72%, respectively. I calculate this by multiplying the difference in fund counts with the current fund price and dividing by the portfolio sum.
  2. Typically, an inflow of 1 CHF triggers rebalancing transactions of between 66 and 150 CHF. Those are free of charge, as VIAC pools and nets all internal transactions between its customers - so the effective volume of funds traded is orders of magnitude lower than the gross volumes.
  3. With the third portfolio that’s currently being filled, there’s a slight loss (-0.05%) versus contributions. I chalk that up to 2022 being a “down market” (-20% at its worst), which I’m using to “average down” - one of the few instances where Dollar-Cost Averaging excels.

One caveat: fund counts shown in the manual reporting may or may not be as precise as their internal weights. So some of the slippage detailed above might just arise from differences in rounding.

With that, I’m hoping to inspires more of you to try out exotic strategies in the Third Pillar investment scheme.

3 Likes

First, what kind of assets allocation you are talking about? And which investment horizon?

All these premia will look insignificant once the bull run starts and you are not in max stocks. MSCI provide historical data for MSCI world index from 1970, feel free to backtest your allocation strategy.

I did some tests with finpension. The results are: sell and buy transaction prices are the same (the same fund, buying in one portfolio, selling in another portfolio) but it is off from NAV by the issue or redemption spread as indicated in the fund’s factsheet. For me it means that they book transactions in both directions at the price they effectively transacted with the fund issuer, and it was mostly buying.

Feel free to test your assumption.

1 Like

@Dr.PI
First, what kind of assets allocation you are talking about? And which investment horizon?

Both do not matter, the only assumption is that the allocation stays stable across time. Obviously, if you start early, you’ll benefit of the additional CAGR for longer.

@Dr.PI
All these premia will look insignificant once the bull run starts and you are not in max stocks. MSCI provide historical data for MSCI world index from 1970, feel free to backtest your allocation strategy.

You can apply them to a 99% stocks strategy. The only thing required for harvesting rebalancing premiums / pumping volatility is fluctuating assets.

@Dr.PI
I did some tests with finpension. The results are: sell and buy transaction prices are the same (the same fund, buying in one portfolio, selling in another portfolio) but it is off from NAV by the issue or redemption spread as indicated in the fund’s factsheet.

Score one for VIAC then. :wink:

To obtain the “What stands out” section, I looked at fund counts first and prices second. Assume you don’t touch a VIAC portfolio after the first two years, those counts would stay stable across time (unless seriously out of whack). Now add weekly additions of CHF 1,00 to trigger rebalancing, and do that for a few months. Then count those funds again. Compare what’s in your portfolio now at current prices, vs. what would have been, also at current prices.

Thank you very much. It’s really cool that we can implement Shannon’s demon without transaction costs!

You can apply them to a 99% stocks strategy. The only thing required for harvesting rebalancing premiums / pumping volatility is fluctuating assets.

Well, my intuition would say that the rebalancing premium is very small if all the assets in the portfolio have high correlation to each other (as would be the case if the portfolio is 99% stocks). Would you agree?

1 Like

@tartufo:
Well, my intuition would say that the rebalancing premium is very small if all the assets in the portfolio have high correlation to each other (as would be the case if the portfolio is 99% stocks). Would you agree?

It would indeed work better with negatively or uncorrelated assets. In two out of three portfolios I added the maximum amount of gold I could (7% I believe) to achieve that. I’ve switched to 1% cash for the time being, will give this a spin until April and then consider a higher cash allocation.

From the best paper I’ve read so far on the subject [1]:

We have established the surprising result that these strategies generate port-
folio growth rates in excess of the individual asset growth rates, provided
some volatility is present. As a consequence, even if the growth rates of the
individual securities all have mean zero, the value of a fixed-mix portfolio
tends to infinity with probability one

I also found your old post on Why we rebalance and other tales about volatility. Great stuff, I’ve been reading Hollerbach / Abdelmessih / Eifert for years and may or may not have got an inspiration or three from The Great Age of Rebalancing Begins… Birds of a feather, and all that.

So I’ll leave you with a related paper from the absolute goats MacLean, Ziemba & Li [2] that may be a little more immediately useful for financial independence & early retirement than pumping vol in 3a pillars you cannot reach for some decades:

A dynamic Bayesian fractional Kelly strategy, where the investor
rebalances the portfolio based on the performance to date, is shown to be optimal
assuming that the risky assets are jointly lognormally distributed. The strategy
minimizes the expected time to reach a wealth goal while maintaining a high
probability of reaching that goal before falling to a subsistence level of wealth.

[1] Dempster, Evstigneev, Schenk-Hoppé: Growing Wealth with Fixed-Mix Strategies (2009)

[2] Maclean, Ziemba, Li: Time to wealth goals in capital accumulation (2005)

2 Likes

Earlier in April, I was revisiting this with more up-to-date numbers.

The rebalancing premium for Q1/23 is around 1.9% in two portfolios, and between 3.1-3.5% since I started the experiment last September.

Not bad considering all it took was two monthly recurring 1.00 CHF transfers :sunglasses:

1 Like

Wow, Shannon’s demon and Parrondo’s paradox, now we’re talking :grin:! For some reason, I’ve missed your original post, but glad I’ve found it now.

I believe monthly rebalancing is not necessarily ideal, since you want to catch long-term trends, especially with regards to stocks, no?

Or were you just referring to new 3a contributions?

1 Like

The underlying assumption I operate with is that I get to profit from imperfect correlation between similar-enough assets which I assume are mean-reverting, otherwise there would be market imperfections to be arbitraged away.

I once read that even trend-following could be approximated as special case of Shannon’s Demon but with ample leverage (whereas with 3a there’s no leverage and always a cash position, minimum 1%)

As for rebalancing period… I’ve tried with weekly rebalancing in Q4 in all three portfolios and settled on monthly rebalancing (= 1CHF inflow times two) for the two portfolios I try to hold “stable” and weekly inflows for the one I’m currently trying to fill: (7056 - 24)/52 ~= 135 CHF weekly contribution. Not much visible difference in my calculations either way.

There’s some debate whether daily/weekly/monthly/quarterly is best and there’s ample data to justify either way.

For my purposes it doesn’t have to be optimal, it has to give me results to study. It’s an experiment, don’t try this at home, etc etc.

1 Like

Good luck, keep us posted, I’m currently 100% in stocks, but that might change soon, so I’m definitely interested.

1 Like

Last time I simulated rebalancing of cryptos (BTC, ETH) against cash, I got the best results for very big rebalancing boundaries. Of course, because selling at the peak and buying at the bottom. I think no-rebalancing won though. Also of course, because being significantly in cash even part of the time cuts into the astronomical return of crypto in that period.

Since crypto is very volatile I would have expected that more rebalancing would have a positive effect, but the opposite was the case.

Rebalancing in smaller steps is also mathematically worse than taking one big step. An even bigger step is only worse, if it doesn’t take place (because the price didn’t go as high/low). Theoretically a high-frequency low-amplitude rebalancing could beat low-frequency high-amplitude.

I didn’t research deeper, so I don’t know if there are real world price patterns that would favor high-frequency even before fees. I could imagine that they are arbitraged away like every free lunch.

1 Like

As mentioned above, rebalancing definitely works with mean-reverting situations. When something is trending upwards consistently you’ll be better off going all-in (or more) on expected growth (Tesla, FAANG of old, BTC or ETH), or set a static allocation between some trending stocks – and maintain that.

But then you wouldn’t find any of those in a typical pillar 3a portfolio :man_shrugging:t2:

1 Like

I know that rebalancing works in your favor in many cases (just not all of them). My doubt was that high-frequency can beat high-amplitude.

1 Like

BTW, any of you invested in trend-following / managed futures funds?

1 Like

Not me, doesn’t fit my risk profile at this point - I’ve seen clever implementations of permanent portfolio style portfolios using this however.

1 Like

New quarter, new numbers:

The three portfolios post between 0.31% and 0.84% outperformance due to rebalancing (probably confounded by dividend reinvestments which I won’t strip out)

The first half of 2023, thus, looks pretty great: measured excess gains come in between 2.40% and 2.81%.

And by the time Q4/23 comes around I should finally be able to post annual numbers.

So far it’s enough for me to conclude that contributing CHF 1.00 per month for each VIAC portfolio is a great idea, almost a (risk-)free lunch

Can you disclose the portfolios asset allocation and the total TWR figures per quarter?

Have done a little of that upthread -

The youngest of the three is VIAC Global 100, the other two with gold admixture are custom-built & a little more ‘idiosyncratic’. Spreadsheet says it doesn’t matter much :wink:

Using TWR numbers outside of broker/fund marketing materials is a bad idea imho - relevant @NonMeek Offshorecomic

My experiment with forcing VIAC to do monthly if not weekly rebalancing by adding 1 CHF to my portfolios has been running for 11 months now.

I’ve published the quarterly results by January (OP), April and July, am very happy with them and will continue this until I find some better optimizations.

So now I’m curious.

Has this thread convinced any of you to contribute CHF 1,00 to your ‘old’ VIAC portfolios?

  • Sounds nice but not for me
  • Am considering it
  • Have implemented it
  • Not with VIAC
0 voters

So, you think there was ~2% of re-balancing bonus. But we didn’t yet see how you calculate it - nor the portfolios TWR. I still think that 75% of these 2% are a calculation error on your side…

I would use time weighted returns to compare two solutions where one gets new inputs regularly (albeit small ones) and the other does not. How do you compare them otherwise (I’m open to learning but can’t think of a method that would fit)?

Other than that, I’d stay it’s still a pretty short test period with only one market condition (bull market) so I’d want to see more before drawing any conclusion.