Three of his conclusions were:
- the small stock premium is not present in current market pricing;
- there was a small stock premium from 1926 to 1980;
- however the size premium was negative from 1981 to 2014.
Deep Thoughts from Roger Grabowski
Roger Grabowski of Duff and Phelps has been a leading advocate for the small stock premium. He has written a very thoughtful and balanced article in the Summer 2016 edition of Business Valuation Review: “The Size Effect – It is Still Relevant.”
He has surveyed much of the academic research in this area over the last 20 years. He has addressed head-on the following challenges to the small stock premium:
- the betas of small stocks are higher than those of larger stocks – therefore the raw observed premium needs to be adjusted for beta;
- the betas of small stocks can be understated due to thin trading and adjustment is needed for this factor;
- the tenth decile includes companies which are distressed, companies which are low revenue start-ups and some companies with relatively high total assets offset by high debt. These latter companies are small because they are risky, rather than being risky due to being small. It is therefore necessary to exclude troubled companies and to derive measures of size other than market capitalisation when considering the small stock premium.
What is meant by Size?
In order to ensure that the size effect has been filtered through different measures of size, Duff and Phelps use 8 different measures of size, including market capitalisation, sales, number of employees, EBITDA and total assets.
The article includes data on the size effect for the US markets based upon the various measures of size.
Roger Grabowski has looked at the data with considerable care: he has made adjustment for the above factors. He also recognises that it may not be size per se which is responsible for the so called size effect. There are clearly other possible culprits:
- inadequate analysts’ coverage;
- liquidity;
- wider than average bid-offer spreads.
The article includes a review of data for the period from 1990 to 2014 – this period was selected to counter the criticism that the size premium has disappeared in recent years. The high financial risk companies were excluded from the data. We are left uncertain as to whether similar results arise in the period from 1980 to 2014.
The Conclusions
The following conclusions are drawn by Roger Grabowski:
- Business risks, as measured by unlevered asset beta, generally increase as size decreases;
- Business risks, as measured by average operating margin, generally increase as size decreases;
- Business risks, as measured by the variability of operating margin over the prior five years, generally increase as size decreases;
- Business risks, as measured by the variability of return on book equity over the prior five years generally increase as size decreases.
The overall conclusion is that there is a still a small stock premium provided that adjustments are made for various correlated factors.
Further Voices
Cliff Asness and Others published an article in January 2015 with the arresting title: “Size Matters if you Control Your Junk”. In this context the term “junk” is used as a descriptor of those companies which are at the other end of the rainbow from high quality companies.
Their thesis, demonstrated with complex statistical analysis, is that there are two factors criss-crossing in the arena of small companies: there are the attributes of both size and quality. Provided that the population of small companies can be sifted and graded in order to retain only the quality smaller companies, there is still a small stock premium to be seen.
The conclusion which they draw: “Controlling for quality/junk means that a significant size premium emerges.” They also state: “Controlling for quality (QMJ), we find a large and significant size premium, which is stable across time, measures of size, seasons, industries, and international markets.”
This therefore remains one of the more polemical areas in current business valuation.
Andrew Strickland