I have previously written about the good, and not so good, parts of Alexander Edeling’s, Shuba Srinivasan’s, and Dominique M. Hanssens’ 2020 review paper on the marketing finance interface, see here. Here I will comment on a specific assertion they make that strikes me as misleading. It is important to highlight this as I fear their words might suggest to some scholars that Total Q is a reasonable metric to use as a measure of firm value in marketing. I believe it is not, and have seen no meaningful arguments to the contrary. The fundamnetal problem is that, contrary to the assertion in their paper, there seems no reason to think Total Q will resolve crticisms of Tobin’s Q as a performance metric. Put bluntly adjusting for unrecorded intangibles is really hard and Total Q doesn’t solve this problem.
Given this, I argue that the idea that Total Q makes up for the flaws of Tobin’s Q as a performance metric should not be taken seriously until people fully engage with the challenge of adjusting for unrecorded intangibles.
Total Q And Tobin’s Q
Elsewhere I write about Total Q and Tobin’s Q, see here and here. A core problem with Tobin’s Q is that the approximations based on financial accounting data exclude many unrecorded assets.
The core idea of Tobin’s Q is to compare the value of a firm to the replacement value of its assets. Still, it doesn’t take a genius to realize that you have a dodgy measure if you omit a bunch of assets. The biggest challenge for marketers is that the assets you omit are not random. You generally miss intangibles, and these are often marketing related. This makes Tobin’s Q a truly appalling measure for anyone who believes that marketing creates assets. (And how can you be a marketing professor if you don’t agree with that?)
Natalie Mizik criticized Tobin’s Q back in 2009 but collectively marketers seem to have missed this work. Indeed, it looks like they started to use Tobin’s Q even more after 2009. Moeen Butt and I had another go in 2018. Finally, marketing scholars seemed to be moving away from Tobin’s Q. I had even become a little hopeful.
Yet, now something else is being offered: Total Q, as new and improved Q. Total Q is a little better but I think about this as similar to “would you rather a) be shot or b) tortured first and then shot.?” I guess a) is better but I’d take almost anything else if you gave me any other option.
Current Usage Of Total Q
The argument for Total Q as a replacement for Tobin’s Q is that it adds in the unrecorded (mostly intangible) assets. This, if done properly, would remove a major source of bias. It would be great.
The currently used measures of Tobin’s Q clearly need improvement and so people have started to adopt Total Q as an improved measure. Edeling and colleagues tell us this, saying that Total Q has been applied in marketing and is popular in finance.
The metric, applied recently in the marketing literature by Du and Osmonbekov (2019) and popular in the finance literature (301 citations on Google Scholar as of May 1, 2020), is closer to the true Tobin’s q measure than AATQ and thus could overcome many of the problems Bendle and Butt (2018) discuss.
Edeling, Srinivasan, Hanssens (2020)
This is not the compelling social proof that one might hope. Firstly, Q when used in finance is typically employed as an independent variable/control. The authors state Total Q might solve a problem and suggest the evidence shows that it is in wide use. But it is in wide use doing something else. The value of a metric depends upon its use, see here. The fact that your toaster is a popular toaster doesn’t mean that it is a good idea to use it to heat up your water.
Setting A Terrible Example
Du and Osmonbekov (2020 (this is the same as the 2019 paper) do use Total Q as a dependent variable in a paper in marketing. Unfortunately this paper, as I’ll discuss next post, is very disappointing. The Du paper betrays a fundamental misunderstanding of the nature of the problem with Tobin’s Q that Total Q is supposed to solve. On several occasions it misstates Total Q or does not give proper detail on what the authors did. No one should follow the Du paper. To be honest, I don’t think the authors of the review piece should be giving oxygen to this paper. Putting it up as an exemplar of the use of Total Q, when the Du paper so obviously misstates Total Q, is not good.
Easy Computable? I Thought Adjusting For Unrecorded Intangibles Is Hard
The fact that the only usage in marketing pointed to should never be copied does not mean that Total Q could never have value in a better paper. Total Q exists, which is better than some theoretical measures. Total Q can be computed which seems to be the main argument for it. (This mirrors the initial argument for Tobin’s Q which I’d paraphrase as ‘it isn’t great but it exists’). Edeling, Srinivasan, and Hanssens indeed tell us that Total Q is easily computed.
One of these metrics could actually be a variant of the traditional AATQ, a measure called “Total q” developed by Peters and Taylor (2017) in the finance field. This easily computable measure accounts for intangible capital in the denominator as the sum of so-called knowledge capital (based on R&D expenditures) and organization capital (based on selling, general, and administrative expenditures).
Edeling, Srinivasan, Hanssens (2020)
It is fair to argue if you follow Peters and Taylor (2017) and treat 30% of selling, general, and administrative expenditures (S,G&A) as an investment and then amortize that you can compute a number relatively easily.
But Is This Any Good?
Some obvious questions to ask are:
- Why 30%?
- Surely this can’t be the same for all firms and across all industries?
- Is the argument that all unrecorded assets should be amortized in the same way? Why?
- Why do you think the amount invested equals the amount of value created? We often try to assess the value of investments so is it a problem that your performance measure assumes that investments all merely return what was invested?
Clearly the choices you make determine the number you get. I don’t want to bore anyone but I’ll repeat: adjusting for unrecorded intangibles is really, really hard.
If one thinks of the question narrowly it is easy to compute Total Q. If one thinks a bit more deeply these authors are telling us that it is easy to adjust for unrecorded intangibles. And that these adjustments are of sufficient robustness you can use the adjusted figures as key inputs into your performance metric. Really?
What The Authors Are Suggesting
To make this explicit the advice in the review paper is that it is easy to correct Tobin’s Q.
All you need to do is to adjust for unrecorded assets with a number which is unknown, and likely unknowable. The numbers you choose will then determine if your research finds a result or not. The good news is that all tests going forward should have positive results given the researcher can pick numbers that work; everyone is a winner.
Here the authors hope that we judge success against;
- a metric that **almost** no manager has ever heard of, never mind used,
- that is based upon a metric known to be biased, and
- which includes mysterious ad hoc adjustments that can’t be independently verfied.
Perhaps it is just me being a coward but this seems not that simple to me. Adjusting for unrecorded intangibles is really hard, pretending it is easy doesn’t help anyone.
For more on Tobin’s Q see here.
For more on Total Q see here.
Read: Alexander Edeling, Shuba Srinivasan, Dominique M. Hanssens (2020) The marketing–finance interface: A new integrative review of metrics, methods, and findings and an agenda for future research, International Journal of Research in Marketing, available online 19 September 2020
