Thursday, July 5, 2012

Misuse of the Scientistic Analogy "Negative Feedback Loop" in Financial Services

It is a well-known fact that the financial services industry - and their regulators - often use bizarre language.  For instance, Alan Greenspan, during his tenure as Chairman of the Federal Reserve System, was famous for achieving a level of incomprehension that spawned a generation of "Fed Watchers" needed to interpret him.  However, the problem goes far beyond Mr. Greenspan, and one of the worst offenses committed by financial services is scientism.

Scientism is the application of the language and methods of the natural sciences (physics, chemistry, biology, etc.) to problem domains that are completely outside of the realm of natural science.  For instance, the term "financial engineering" has been used to describe a large set of practices which contributed significantly to the great financial crisis we are still living in.  Having worked in asset securitization myself, I can assure anyone that whatever these practices were, they were certainly no form of "engineering".

Which brings us to the term "Negative Feedback Loop".  This is a term that originated with real engineers, but which has since been appropriated by the financial regulators, and has then trickled down to the rest of the industry.  Here is an example of its use (admittedly in a quote) by the otherwise erudite Ambrose Evans-Prichard of the Daily Telegraph.

"Mr Roberts said the collapse in Spanish tax revenues is replicating the pattern in Greece. Fiscal revenues have fallen 4.8pc over the last year, and VAT returns have slumped 14.6pc. Debt service costs have risen by 18pc.
The country is caught in a classic deflationary vice: a rising debt burden on a shrinking economic base. “Once you get into such a negative feedback loop, you can move beyond the point of no return quickly,” he said."
[  - by Ambrose Evans-Pritchard, International business editor, 7:39PM BST 30 May 2012]

The term "Negative Feedback Loop" in its engineering sense means that the output of a process becomes an input to that process, and acts to stabilize the activity of the process.  For instance, the governor of a steam engine is a device at the top of a pipe that is rotated by the engine.  The governor includes two or more weighty metal balls hinged to close a hole through which steam would otherwise escape.  As the pipe rotates faster, due to engine activity, the centrifugal force lifts the balls off the hole and steam escapes, reducing the power delivered by the engine.  This puts an upper limit on the power the engine can provide to whatever it is driving, thus preventing "runaway" effects such as running a pump so fast that it breaks.

This example shows that a "Negative Feedback Loop" can be a very good thing in engineering - and actually it is typically so.  However, the great minds of the financial services industry have little acquaintance with  real science and engineering.  I presume they saw the word "negative" and equated it with "bad" and simply appropriated the whole thing.  So if a country owes $1 Trillion, and has a falling GDP (say due to government austerity), the debt-to-GDP ratio increases and the  debt is even more difficult to service.  This is more closely resembles a "Positive Feedback Loop", where the output of the system is amplified by the output becoming an input, often contributing to equipment damage in real engineering.  Take a look a Wikipedia if you want a fuller explanation.

So "Negative Feedback Loop" in financial services is either:

(a) Misuse of terminology to signify real positive feedback loops that do occur in financial services; or

(b) A  scientistic analogy that is conning us into thinking there is a real concept being signified, whereas in fact there is no intelligible concept.  And the analogy has gone wrong because of the desire for the term "negative" to mean "bad", rather than "dampening" as it does in engineering.

The example I gave above of the debt-to-GDP ratio has points in common with a true positive feedback loop, but that is what I would expect of any analogy.  It does not prove that we are dealing with an actual positive feedback loop (that happens to be mis-termed as a "Negative Feedback Loop").   A definition is not a few selected attributes that are in common with some other generalization.  In financial services strange things can happen because the domain is governed by human laws, not natural laws.  Debt can be repudiated in financial services, but matter and energy must be conserved in the natural world.

My conclusion is that the class of phenomena to which the term "Negative Feedback Loop" is applied in financial services breaks down into concepts that have never been given adequate definitions.  Therefore we are dealing with unintelligible concepts.   The fakery of using scientistic analogies (because scientific language is always so plausible) has been poorly applied in this case, and the terminology has pointed up the problem.


  1. Richard OrdowichJuly 8, 2012 at 8:21 AM

    The use of "fakery" is prevalent in business. Most of the terms are made up without any forethought to what they mean. Each time a database is designed or an application built, a new group of people come together and conjure up new terms. There is no established discipline even when naming data. I equate this lack of discipline to bad design and laziness. It takes time and effort to define a data element and that time is typically wasted discussing the esoteric of the data rather than its precision. The consequences of this bad design remain from decades; people misinterpreting or misunderstanding the data. But because those who created this abomination are off building yet other databases and systems, there are no consequences to them.

    People working with data should be obliged to read The Design of Everyday Things by Donald Norman and have a stove top installed with the controls misaligned to the burners they control (an example in the book). After suffering with this example of poor design, perhaps those designing data solutions would take more care when naming and defining the data!

  2. I have to deal with a "negative feed-back loop" in a financial text, which I am translating into German, and came across your very helpful blog entry. I would rather translate it with "Teufelskreis", which actually means "vicious circle". Shouldn't they simply use "vicious circle" in English as well instead of "n.f.l."?