Management, these days, is full of corporate metrics (otherwise called buzzwords) that managers are urged to follow, to maximize or to minimize: EVA, CFROI, shareholder value, costs of some sort etc... to name a few. Sometimes, the metric is a bit more involved but follows the same vein or puts some constraints on them: balanced scorecard, sustainable development etc...
In any case, the theory around these catch-all metrics boils down to the idea that following them is good for the corporation as a whole, that the corporate decisions should always be benchmarked against them. For instance, what is good for shareholders is also good for the firm as a whole.
Surprisingly enough, the discussion around the true relevance of these benchmarks is rather tenuous. Yes, market imperfections are discussed and principal-agent models, game-theoretic-models have been crafted to accommodate them.
However, it seems rather puzzling that a principle known under the name of "Goodhart's law" is hardly mentioned in business books. The law is named after Professor Charles A. E. Goodhart, Norman Sosnow Professor of Banking and Finance at the London School of Economics, who crafted with central banking in mind (see Download Goodharts_Law.pdf ). Here is what the University of Cambridge has to say about it:
"The original form of Goodhart's law arose in economics. According to the 99th edition of Pears Cyclopaedia (1990--1, pp. G 27, G31), the law states that
- `As soon as the government attempts to regulate any particular set of financial assets, these become unreliable as indicators of economic trends.'
This, of course, is because `financial institutions can... easily devise new types of financial assets.'
Professor Charles Goodhart FBA was Chief Adviser to the Bank of England. The Bank used to have a web page about him at www.bankofengland.co.uk/cvs/goodhart.htm, giving his own statement of the law, as published in his book Monetary Theory and Practice, page 96:
- `Any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes.'
Professor Marilyn Strathern FBA, following Hoskin (1996), has re-stated Goodhart's Law more succinctly and more generally (see also Download STATISTICS_AND_TARGET_SETTING.pdf ):
- `When a measure becomes a target, it ceases to be a good measure.'
Goodhart's law is a sociological analogue of Heisenberg's uncertainty principle in quantum mechanics. Measuring a system usually disturbs it. The more precise the measurement, and the shorter its timescale, the greater the energy of the disturbance and the greater the unpredictability of the outcome."
The last re-phrasing of the law is clear: Too much emphasis on a measure as a target kills the measure that may have been a good one in the first place.
Moreover, in economies like our today economies which are highly non-linear (knowledge-based economies, complex systems, winner-take-all etc...), it is not even clear that any such measure does exist. More often than not, obliquity is at work and single metrics are very poor at capturing the richness and complexity of the economies and societies in which we live. They may simply defeat their own purpose!
The recent business history is full of cases that illustrate the perils of putting too much emphasis on narrowly defined corporate metrics: Stock-options, Enron, Adelphia etc...
The more distant history is also quite rich in such stories, chief among them that of former USSR and its obsession for production targets (yielding the results that we all know about).
So, why is that most business books authors and professors forget about Goodhart's law as if they were petrified telling students that, indeed, the world outside is awfully complex?
Thanks,
This is excellent,
I'm teaching a class tomorrow at HEC-ISA on how to manage intagible asset, a class for a metrics and measures, and i shall you the goodhart law as a health warning during my intro...
Marc.
Posted by: Marc Goldberg | November 30, 2004 at 12:05 AM
Well, what can I say except that I fully agree with what you have written Eric.
Regarding my personal experience, when I tried to tie performance and knowledge management, I met "intangible assets" (u are right Marc, this law is a good place to start a course on that topic)...BSC (Kaplan & Norton), Skandia's Navigator (Edvinsson & Malone)...and so what!!! This metrics are so complicated that you spend much more time and energy to transform knowledge into something
Two major findings I have made:
1. Companies who have tried to measure and report its intangibles have encountered bad financial results at the same time! Check the Skandia website and tell me if the intangibles report is still made today...of course it is not! (it was made in 1994)
2. You can manage what you cannot measure. I think there are only engineers who can say that! I tell you a short story: interviewing a manager in the company I am doing my PhD I ask him:
- what is the value brought by KM?
- the value...you want a figure??? I give a figure...12!!!...this is my figure, my value of KM...
This guy was right. It is impossible to measure such a complex process in a knowledge based economy.
So far, I believe in success stories to measure that kind of value. When people talk, they measure the benefits by creating a sense of wht is going on.
What do you think?
Posted by: Alex | November 30, 2004 at 07:16 PM