Ned Davis Research (“NDR”) has some fantastic tools. One of the tools that we’ve found very interesting is their interactive Factor Performance Tool. The tool allows you to compare the performance of a wide range of equity factors in stock markets across the world. We decided to use the tool to compare the performance of several popular value factors in Australia.
Spoiler: How you define value really matters! Book/Price, the standard academic measure of value, may not be the best way to measure value in Australian equities. Read on for the details.
Interpreting the Charts
NDR’s analysis uses stocks in the MSCI Australia Index which represents 85% of Australia’s free-float-adjusted market capitalization. Each factor is shown as a yield, i.e. variable/price. For example, a price-to-earnings ratio of 20× is equivalent to an earnings yield (earnings/price) of 5%. All you need to remember is that higher yields equal cheaper valuations.
Factor portfolios are equally-weighted (not market cap weighted) and rebalanced monthly. This is an important point which we’ll come back to later on in this post. We’ve gone back to the inception of the data for each factor. Returns are price only (i.e. they ignore dividends).
Each chart works as follows. The solid black line shows the cumulative performance of the MSCI Australia over the period. Shaded areas are bear markets, or declines of 20% or more. The broken green line shows the cumulative performance of the cheapest decile (or 10%) of stocks as measured by the factor. The dotted red line shows the cumulative performance of the most expensive decile of stocks.
The blue line shows the cumulative long/short performance of the factor. In other words, it’s the cumulative return that an investor would receive if they went “long” (i.e. bought) the cheapest decile of stocks and sold “short” the most expensive decile of stocks. We’re most interested in the cumulative long/short performance because it isolates the performance of the factor. In theory, selling the most expensive decile hedges out market risk, leaving only the factor return.
The first chart shows the performance of the Book/Price (the inverse of the price-to-book ratio) factor.
For readers that don’t like charts, the table at the bottom tells the story using numbers. Had you invested in a long/short portfolio back in 1989, you would have lost 96.75% of your money! That’s -8.99% per annum for 29 years. Meanwhile the MSCI Australia index earned 4.93% over the period (ignoring dividends).
Its interesting to note that most of the under-performance is due to the cheapest stocks (top quintile Book/Price) under-performing. For most of the period, expensive stocks (bottom decile Book/Price) tracked the market. They’ve out-performed significantly since 2014.
Cheap stocks did enjoy a period of very strong performance during 2016 and 2017 (i.e. the green line spiking upwards). Performance has since plateaued.
The next chart shows the performance of the earnings/price (or the inverse of the price-to-earnings ratio) since November 1989.
Earnings/Price also under-performs the stock market, earning only 2.28% per year. At least it’s positive. Importantly expensive stocks (red line) under-perform. This is what we expect; but it’s not the case for Book/Price.
Once again, we see the performance of the cheapest stocks (green line) has improved since 2016. Expensive stocks (red line) have done even better (note the sharper rise over the same period). That’s why the long/short factor performance (blue line) has been trending downwards since 2016.
It’s a similar story with Sales/Price. Performance is positive but, still lags the market. Once again, expensive stocks (red line) under-perform, which is expected (unlike Book/Price).
Here’s a measure of value that trounces the market! 13.59% per year versus 5.35% for the MSCI Australian Index since March 1991. EBITDA/EV out-performs on both the long (green line) and short (red line) sides. Once again, expensive stocks under-perform just as we expect (unlike Book/Price).
Much of EBITDA/EV’s out-performance is attributable to the losses suffered by expensive stocks during the 2000=2004 period. Long/short factor performance (blue line) tracks the performance of the market quite closely (black line) until it takes off in 2000. This seems to be primarily due to the massive under-performance of expensive stocks (red line), which took a dive with the bursting of the tech bubble.
Free Cash Flow/Price
Our last value factor is the Free Cash Flow/Price. Once again, it beats the market by a very healthy margin (17.95% per year versus 4.46% since September 2001). Both the long and short sides perform as expected.
It’s important to remember that the portfolios shown above are not real portfolios! There are several reasons why this is true, including:
- Top and bottom decile portfolios may be highly concentrated in a few stocks
- Factor portfolios are equally-weighted, not market capitalization-weighted
- Factor portfolios are likely to be concentrated in only a few sectors
- Transaction costs are ignored
- Stock borrowing costs are ignored
- Dividends are excluded
All of these issues will affect results in real-world portfolios. Still that doesn’t mean that we should dismiss long/short factor research entirely. While they may not be useful as an investment strategy (without significant modifications), long/short factor portfolios are very useful for performance attribution. They help us to better understand the drivers of performance. This is an essential part of forming reasonable expectations for active investment strategies.
It’s also worth pointing out that it’s very difficult to make the claim that one factor (in this case Free Cash Flow/Price) is better than another. Most academic research shows that the performance of factors varies considerably over time.
For example, the best performing value factor in the 1st edition of James O’Shaughnessy’s What Works on Wall St published in 1996 showed that Sales/Price was the best performing value factor in the USA. By the time the 4th edition was published in 2011, EBITDA/EV was the best performing value factor.
How you choose to define “value” clearly matters!
All of the value factors, with the exception of Book/Price, perform as expected. That is, cheap stocks out-performed expensive stocks. But book/Price had the opposite result. Expensive stocks returned 6.90% per year, while cheap stocks lost 1.01% per year. This is significant because Book/Price is widely used as the standard measure of value in academic research.
The use of Book/Price as a measure of value has come under a lot of criticism recently. For example, Negative Equity, Veiled Value, and the Erosion of Price-to-Book by Travis Fairchild at O’Shaughnessy Asset Management and Time to Change Your Investment Model by academics Feng Gu and Baruch Lev.
A better way to measure value is to use a composite value factor, i.e. several value factors averaged together. This averages out the extremes in each individual factor, while still measuring the exposure to value. It also protects against the risk that a given measure of value becomes less relevant over time, which is what appears to be happening with Book/Price.
Grioli and Company Pty Ltd, Australian Financial Services Licence: 506003
This web page is for educational purposes only. It does not constitute financial advice or take into account the particular investment objectives, financial situations or needs of individual readers. Readers should consider whether any opinions or recommendations in this document are suitable for their particular circumstances and, if appropriate, seek professional advice, including tax advice.
This web page has been prepared with care. However, Grioli & Co. makes no warrant of any kind in regard to the contents and Grioli & Co. shall not be liable for incidental or consequential damages, financial or otherwise, arising out of the use of this document.
This web page is not an offer to sell or the solicitation of an offer to buy any security in any jurisdiction where such an offer or solicitation would be illegal.
The price and value of investments referred to in this web page and the income from them may fluctuate. Past performance is not a guide to future performance, future returns are not guaranteed, and a loss of original capital may occur.
The contents of this web page are subject to the provisions of the Copyright Act, 1968, and any unauthorized reproduction of it is subject to prosecution under the provisions of the Act.