Another Year Down – Learning from Quant Models

Another Year Down – Learning from Quant Models

By Justin Carbonneau (@jjcarbonneau) —

We’ve been running quantitative model portfolios since 2003. While many of our models were launched after that, our initial set of guru models actually went live in July of 2003, so this year we’ll hit the 20-year mark in terms of running systematic investment strategies. After every year, we like to look back and ask ourselves some questions: What did we learn over the past year? Are there any improvements we think we can make (this mostly happens in our private money management business)? Were there any major surprises?

In reviewing the returns for our portfolios in 2022, which were difficult for the markets and investors, things mostly played out as you may have expected as we look back with hindsight, although there are a few surprises and important lessons I think we can draw from the results.

Guru Model Portfolios

On Validea we run 22 stock selection strategies extracted from books or academic papers and construct 10 and 20 stock model portfolios based on these investment approaches. We offer different rebalancing periods so that investors can choose how active they want to be with each portfolio. For the purposes of this article, I am going to use the monthly rebalanced 10-stock model portfolios as the default I’m referencing.

Overview

Of the 22 models, 13 outperformed the S&P 500’s return while 9 trailed the market. Take a guess on what style – value vs. growth – did better this year?

That’s right! Our value models were the big winners while the growth and momentum models saw the largest losses.

Let’s look at a few of the top and bottom performers.

Top Performers

The top performing strategy was one that outperformed by a lot and that is on model we run based on Joseph Piotroski’s paper on value stock selection. In his academic paper (“Value Investing: The Use of Historical Financial Statement Information to Separate Winners from Losers“), Piotroski outlined and tested a strategy that takes the cheapest 20% of stocks in the market (using price-to-book) and then separates companies by a series of fundamental and accounting tests to determine which stocks have improving financials. Piotroski called this fundamental scoring system the “F-Score”.

The model portfolio was up 17.2% in 2022 vs. a 19.4% loss for the S&P 500 (that’s a 36% performance difference). That relative outperformance has put the model in striking distance of the S&P over the very long-term. After a tough decade for simple systematic value strategies from 2011-2021 this model bounced back very well in 2022. Value models like this are streaky, which could bode well for 2023, but only time will tell.

Runner Up: The model we base on David Dreman’s contrarian investing method was our second best for the year. The model was down 0.2% vs. the 19% loss for the S&P 500. Exposure to energy stocks and mid-to-large cap deep value names were the driving factors in this model’s outperformance.

Worst Performers

The biggest decliner (down 32.3% for the year) was the strategy that is the complete opposite of the Piotroski model. It was the P/B Growth Investor investing model outlined by Partha Mohanram in his paper, “Separating Winners from Losers among Low Book-to-Market Stocks using Financial Statement Analysis”. This strategy starts by looking at the 20% of the most expensive stocks in the market and then uses a set of fundamental factors to separate those growth stocks that look like they can keep growing from the ones that don’t. Mohanram called this the “G-Score”. Throughout 2022 the most expensive stocks were the ones hit hardest as valuations started to normalize in a new world of higher interest rates.

It’s important to note, however, that the P/B Growth Investor is one of the top performers long-term and in an environment where growth stocks are favored by investors this model has shown its ability to deliver, so if growth reverts watch this model closely.  

Biggest Surprises

The first biggest surprise is actually the second worst performing model, and that is the approach outlined by Joel Greenblatt in his book, The Little Book that Beats the Market. This model’s seeks to combine the value style of investing espoused by Ben Graham with the high-quality style of investing followed by Warren Buffett. The set up in the market coming into 2022 would have you thinking this type of approach should produce in 2022.

Not so fast! The 10-stock portfolio fell by 30.3%, making it the biggest disappointment in the value model camp.

The other major surprise actually shows up in one of our consensus models. These are portfolios that blend different guru strategies together, we run two consensus portfolios on Validea. The Validea Hot List and Top Five Gurus. I will spare you all the detailed intricacies of how we construct these portfolios but the Top Five Gurus model was up over 12% vs. -19% for the S&P 500.

ETF Model Portfolios

In addition to the concentrated stock models, we also run ETF model portfolios that can be broken down into three categories:

  • Factor ETF Model Portfolios: Hold factor ETFs based on how they rank on value, momentum and macroeconomic factors.
  • Sector ETF Model Portfolios: Hold sector ETFs based on how they rank on the same factors.
  • Risk-Managed ETF Model Portfolios: Multi-asset ETFs with risk management inputs.

Overview

In total, there are 23 different ETF portfolios on the site and in 2022 all but four of them outperformed their respective benchmarks, which is very good outperformance rate (82%) in a tough year.

Top Performers

I have to give the top performer award to the entire group of Sector ETF Model Portfolios. All of the Sector models outperformed the S&P 500 and by a margin of anywhere from 11.8% to 21.9%.

It’s easy to see the drivers of the good results when you look under the hood. This is the current portfolio in Sector Rotation Momentum model. Energy, Industrials, Health Care, Utilities and Consumer Staples. Notice, no Technology, Consumer Discretionary or Communication Services – the hardest hit areas of the market last year.


Runner Up: Two of our risk-managed strategies – Generalized Protective Momentum (GPM) and Protective Asset Allocation (PAA) – produced very strong relative returns of -5.8% and -6.1%, respectively. Both utilize multi-asset strategies that combine relative momentum and crash protection techniques to help protect capital in major market downturns. PAA has a long-term return that is superior to the 60/40, which is impressive given the 60/40’s performance up until this year, while GPM is slightly behind since inception.

Worst Performer

The biggest laggard for the year was a version of the Permanent Portfolio (holds short term bonds, long term bonds, gold and stocks) that uses trend following. The portfolio took a hit early in the year, moved to the sidelines and then back into assets later in the year only to see losses on those positions.

Surprises Lead to Lessons

The biggest surprise is actually the first key lesson learned and it doesn’t have to do with an ETF model, but instead the benchmark used for our risk managed strategies, which is the 60/40 stock/bond portfolio.

Stocks and Bonds Both Down

It’s been a long-time since stocks and bonds both fell at the same time, but 2022 brought this new reality to investors. The 60/40 had its worst performance in years and the bond portion of the portfolio was down almost as much as stocks. For many balanced investors, bonds are there for both income and protection from market declines, but higher inflation and higher rates changed that dynamic this year and probably will result in investors seeking more levels of diversification in the future.

Too Much of a Good (or Bad) Thing

The Technology sector came into this year as a very large weighting in the overall S&P 500. Energy, on the other hand, had dropped from somewhere in the 6-7% range to a 2% sector weighting. When there is too much (or too little) of a good (or bad) thing, the chances are strong the trend will revert – we just don’t know when. But over the past year, energy is up over 58% and technology is down 30%.

One Year Doesn’t Make a Trend

No one should be reading this and betting on the Joseph Piotroski value model in the short-term, going all in on the Sector Rotation Value ETF portfolio or going 50% into energy. But we can take stock in what these data points may mean and think about ways we can find reasonable amounts of exposure to areas of the market we might need exposure too. If you have been loading up on growth stocks for the past 5-10 years, there is a reasonable chance value (statistically cheap) stocks may produce better returns over the next 5-7 years. Or, if you’re thinking that the 60/40 shouldn’t be the dominant approach, you could consider making an allocation to a different approach to balance out some of that stock and bond short term volatility.

Next year there will be a whole different set of results with both new lessons and different reminders on how to invest successfully using quantitative models over time.


Justin J. Carbonneau is VP at Validea & Partner at Validea Capital Management.
Social | Podcast | Interviews | Articles
about Justin all in one place