The Measurement is Not the Thing

Measure what is measurable, and make measurable what is not so.

-Galileo Galilei

Any science contends with a difficult problem: there are things we want to understand, but we cannot easily measure those things.  Any tool for measurement will inherently have technical limitations (that is, limited by the technology of the time) and be subject to arbitrary choices by those doing the measuring.  Sometimes we may not even be able to directly measure the thing, instead we have to rely on proxies, or on a lack of information, to track what we’re interested in.  Consequently, the measurement one uses is not the same as the thing itself. It is merely a representation of the thing.

Unfortunately, those untrained in scientific thinking will often mistake the measurement for the thing.  This is very common in economics.  Take, for example, the Consumer Price Index (CPI).  CPI is a measurement of inflation using a basket of consumer goods.  The basic idea is that if prices are moving independently of one another, there should be limited movement in the Index.  However, if the Index is consistently changing in a certain direction, there is likely inflation (the Index is rising) or deflation (the Index is falling).  However, the Index can rise for reasons other than inflation.  If even one price changes and all else is held equal, the Index would change simply because of how the Index is calculated.  It is a weighted average of all prices in the basket.  One change therefore moves the whole Index.

To make my point more concrete, let’s say that the price of gasoline were to rise considerably and no other price were to change.  The CPI would naturally rise given gasoline is a component of the Index.  No economist would say there is inflation; inflation is a general rise in prices and this is just a single price rising.  But those who think that CPI is inflation would misunderstand and consequently misdiagnose the situation.  They confuse the measurement for the thing itself.

We see the same thing for Gross Domestic Product (GDP).  GDP is a measure of economic growth, but it is not economic growth in and of itself (nor is it a theory of economic growth).  GDP, like the CPI above, is an accounting identity that attempts to act as a proxy for economic growth.  But, like with CPI, those who confuse the measurement for the thing erroneously conclude that an increase in GDP necessarily means an increase in economic growth.  GDP is defined as New Consumption + New Investment + New Government Spending + New Net Exports.  If any of those variables change, GDP will necessarily rise.  That is true.  But it does not follow that the rise in GDP necessarily means economic growth is occurring.  America in World War 2 and the USSR showed that conclusively.  US GDP rose significantly in World War 2 because of the huge increase in government spending.  But, by many measures, people were worse off than during the Great Depression: consumer goods were hard to find because so many materials were needed for the war effort, people had to grow their own food—it was not an economy that supports a good life.  In the USSR, GDP was rapidly approaching the US.  Indeed, some were even predicting the USSR would overtake the US.  But once the USSR collapsed and we saw behind the veil, the standard of living for Soviet citizens had barely changed since the fall of the Tsars.  GDP was propped up by government spending, and thus became an unreliable indicator of economic growth.

The confusion between the measurement and thing I have discussed here is a perpetual problem for any sort of central planning or industrial policy.  The central planners must establish some goal, which in turn requires some measurement.  But then the project becomes all about hitting the measurement rather than promoting the goal.  Ultimately, this leads to the plan to fail in its goal even if it hits the measurements.