Business is plagued these days with academic studies that come up with sensational conclusions, usually based on statistical regression. Many seek – and often get – wide media coverage if their claims are wild enough. To experienced business people, these reports usually draw a wry smile. If having greener policies or more diversity or better governance really were such sure drivers of financial success, life would indeed be easy.
However, I was surprised to see Grant Thornton on this bandwagon, not least as auditors, you would expect them to specialise in precise statistics and impartial analysis. Moreover, their report, ‘Corporate governance and company performance‘, makes some of the most spectacular claims I’ve ever seen of such business performance studies;
Apparently they have found: ‘A proven link between effective corporate governance and value creation’ [my emphasis].
Companies with strong governance are;
- 29% more efficient at generating profits;
- 43% more efficient at making or selling products/services;
- Generate 3.4x more cash flow;
- Generate double the return for shareholders;
- and so on.
Let’s leave aside the basic error here, that correlation doesn’t mean causation. It is just as likely that better performing companies can afford better governance as it is that better governance causes better performance. Correlation will never tell you which way round the causation flows. So no, there can be no proven link.
But what about the statistics themselves?
- Grant Thornton uses a proxy score for governance, heavily influenced by box ticking annual reports, and we have no way of knowing if this is a true reflection of good governance in practice.
- For some unexplained reason, it uses comparisons of top and bottom quartiles rather than the more usual method of using each individual company results.
- It claims something called ‘outperformance probability’ which is underfined, and sometimes quoted at an improbable 100%.
- Despite the apparently massive impact of good governance on financial performance, the actual correlation coefficients are very modest at around 0.2. In other words governance explains only 20% of the variance.
- Statistical significance (ie the probability that the apparent relationship is just random chance) seems to have been set at 70% . No serious academic study would set such a low bar, as it implies that about almost a third of the relationships paraded here are just random. 95% is the standard benchmark for statistical significance.
- A 70% probability of a 20% explanation is a very weak relationship at best, not consistent with the big claims made.
- The report says at one point that it found that companies that improved their governance subsequently improved their financial performance. This would be an important conclusion, but no evidence is provided to substantiate it.
I could go on, but let’s not prolong the pain. Good corporate governance, like diversity and green policies are very important, and probably all contribute to being a successful company in the long run. However, dodgy statistics and wish-fulfilment do not assist anyone in understanding how companies should work and just provide click bait for the media. Business deserves better analysis than this. I can forgive a struggling academic trying to get some publicity, but when audit firms join in with this nonsense, it does make you wonder about their day job.
