Factors and Risks

I am not sure who my audience is anymore, so I will try to keep this post as generic as possible. If you recognise where the example comes from, kudos to you 🙂

There is an investor with an 80% stocks, 20% bonds portfolio; the stock portion is heavily tilted towards the value factor, with an added sprinkle of small caps on top. The portfolio is divided into 50% USD-denominated assets and 50% non-USD.

Considering that value is deemed to be a defensive factor, a friend of the investor questioned him if it would not be the same to just move to a 70% stocks, 30% bonds split and call it a day. To further “test” this idea, they go and check how value performed during the dot.com, GFC and Covid crises.

This is 2023, we can do better than that, no?

What I have in my mind is that there are quite a few parts in this portfolio and how they interact might not be as evident. Luckily for us, we have one of my favourite tools: PortfolioCharts! Compared to a naive backtest a la’ PortfolioVisualizer, PC offers another angle:

• Inflation-adjusted CAGR (ok ok, you can do that as well with PV but usually other analysts just show you average annualised nominal returns)
• a risk measure in the Ulcer Index
• a suite of easy-to-digest charts
• data starting from the 1970s

I will compare a 70/30 portfolio, defined like this:

with this 80/20, value+small cap tilted:

I will keep this order going forward, i.e. the first chart in each chapter refers to the 70/30 and the second one to the 80/20.

I kept the “Home Country” as US because I do not like how PC manages each portfolio’s International Translation; I tried to keep the asset allocation as passive as possible, i.e. a World-type exposure as a blend of US and ex-US developed markets.

Annual Returns:

Despite the value tilt, the 80/20 is still riskier than the 70/30 (the loss frequency is marginally higher and the left tail got fatter). The additional risk is more than compensated by an increase in returns though.

Equalizer:

In my opinion, these charts are very effective in showing that the two portfolios behaved in quite a different fashion in the past. The 80/20 is more volatile, it had higher highs and a deeper low during the GFC, but it also spent “more time” with positive returns.

They show how hard it is to unequivocably define the better portfolio between the two. If you assume the 70/30 as closer to the “benchmark”, the 80/20 investor must be able to endure a sensible tracking error; in other words, they have to be a true believer in the value factor in order to stick with it.

The only possible conclusion here is that the two portfolios are not equivalent at all, despite the initial intuitive idea!

Drawdowns:

Moving away from annual-only losses, these drawdown charts are great in showing the benefits of the value tilt: drawdowns are not shallower but the recovery is way faster. Statistically speaking, the 80/20 investor spends less time stressing about their investments than the 70/30.

Withdrawal Rates:

The drawdown profile is a key ingredient for the Safe Withdrawal Rate calculation. Shorter drawdowns mean the investor is less likely to sell investments while they are “cheap”. The riskier 80/20 allows the investor to ‘safely’ withdraw higher amounts than the 70/30 (please note that SWR calcs are still very fragile. Small changes in methodology, ingredients, source data and so on can lead to materially different conclusions. This is for illustrative purposes only, do not try this at home as do not retire simply based on what you see on PC).

---

It is hard to define a ‘safe(r)’ strategy because there is not even a good definition of risk out there 😉 The value factor (is supposed to) outperform standard market-cap weighted indexes, therefore risk has to manifest in one form or another.

I hope that this quick analysis demonstrated, to the new readers of this blog since I beat this drum for a while now, that the relationship between single asset classes is not (always) linear and easy to grasp.

Verdad just did a blog post about Asset Class CAPM:

Based on this picture, no one in their right mind would invest in Oil or Gold: both commodities offer the same returns as HY Bonds with much higher volatility.

But if we turn to asset class correlations:

no equity investor would add HY to their portfolio 😉 Or, in other words, they would not put HY and Govies in the same ‘bond’ bucket. Now think about, how many people you know despise gold while defining Govies and Corps as similar assets?

The Growth Factor

Since the above example involved the value factor, I want to take the opportunity to clarify something that I have heard for ages. To be more precise, since the Nasdaq climbed back to the top of the leaderboard of stock indexes.

There is no growth factor.

At least, according to the infamous Fama-French model.

Value + Growth = Total Market. In other words, if you want to have exposure to value and growth, just buy the market, because that’s what you are going to have anyway. If value outperforms the market, as the academic literature suggests, then growth has to underperform.

I never looked for a “growth ETF” because…well, you just read why, it is like looking for a strategy to win at the roulette when you know that you cannot win (in the long term). In fact, there is just one growth index out there: the Russell 1000 Growth. And curiously, no retail investors that profess to have exposure to growth in their portfolio bought a product linked to it. They all have other stuff like QQQ, ARKK, IPOS, a tech sector ETF, you name it…

Those vehicles might mean “growth” to you but they are something else. I am not saying they are better or worse, just something else. In 2022, Russell had Facebook, Google AND Netflix in its value index (at least part of them, since Russell can split the same stock between Growth and Value, depending on their criteria)!

I wasn’t that particularly shocked about it, the first two names are incredible profit machines and in 2022 they declined 64% and 38%; but for the average retail investor, they were and are growth names.

The Russell 1000 Growth has indeed outperformed the S&P500 in the recent past:

Before doing some research for this piece, I thought the outperformance was simply coming from the index market-cap weighted methodology: if you further increase the weight of the Magnificent 7 (Apple, Google, Microsoft, Nvidia, Tesla, Meta and Amazon) by removing the value chaff, you have to outperform for sure…no?

Actually, “growth” (I had to take the Nasdaq here because there was no index for the Russell) outperformed almost the same CAGR amount even on an equal-weight basis:

Does it mean that, actually, the value factor does not exist?!?

I guess only time will tell…but in any case, pick your side and do not combine value with growth in your portfolio, otherwise you are simply paying unnecessary fees for nothing.

[for me, this was an interesting research because I thought the Fama-French model was EW-ing stocks but, in fact, they are MCW-ed (then the portfolios are EW-ed)]

What I am reading now:

Follow me on Twitter @nprotasoni