Hedonics Hijinks

In our last post, we showed how the exclusion of vast swaths of real time prices causes price measures limited to final goods and services to lag actual inflation.  We also highlighted one of the consequences being that incipient increases in inflation cause false positives in measures of real GDP.  Among other things this stokes the illusion that inflation is caused by the economy running “too hot”.

Here we’ll peel back another layer of the onion skin and question whether the prices of things in general are the valid gauges of the value of the currency they’re presumed to be in the first place.

One of the places the flaw in this presumption shows up most vividly is with the introduction of “hedonic” adjustments to prices in the calculation of inflation proxies like the Consumer Price Index.  If for example the price of this year’s iPhone is 10% higher than last year’s, but due to more storage capacity or a faster processor is deemed by government statisticians to be 10% better, the effect of the price increase on the CPI calculation is zero.  By extension this means that if the prices of all goods were to rise by 10% and they are also deemed 10% better, the CPI increase is zero.  Yet despite this “zero inflation” environment if you were to try and live on the same number of dollars this year as last, you’d have to do without something you could afford before.  The potential for this patently absurd result seems not to trouble the econometric establishment.

We could apply the same reasoning to a number of goods and services.  Say your Internet service provider increases its monthly rate by 15%.  But it also raises your access speed by 20%, or your data allowance by 20%, with the implication that the inflation component of the price change was -5%.  Never mind that you can’t buy last year’s service at any price – your only option is to pay 15% more – by virtue of hedonic massage this technological advance resulted in 5% deflation.

What if access speed and data allowance both went up by 20%?  Does that mean the contribution of internet service to inflation was -25%?

Anyone else get the impression that this practice allows the government to report just about any inflation rate it wants to?

What about food?  Technological advance such as gene splicing has affected agricultural commodities such as corn and soybeans.  In this case the technological advance isn’t aimed at improving the product’s value to consumers, but at making it less costly to produce.  Inserting a gene to say, make a product more resistant to a herbicide contributes zero or less to its nutritional value.  Many consumers in fact object, and will pay more for the non-GMO versions of such products.  Does the government give the cheaper version a hedonic debit in its inflation reckoning?  Or does it merely compare the newer, cheaper product with the old as if there were no change in quality?

The underlying fallacy here is that technological advance has anything to do with inflation in the first place.  Technological advance is a real phenomenon.  Inflation is a monetary phenomenon.  If you’re trying to measure inflation – and especially if you’re using your measure to inform monetary policy – it makes no more sense to include real effects like technology into your calculus than it would to include New York wind speeds in money supply.

A little reductio ad absurdum helps illustrate this even more clearly.  It should already be clear that if you wanted an unbiased and accurate measure of dollar depreciation (inflation) in terms of goods and services, you can really only legitimately compare prices of identical items.  Comparing the price of this year’s iPhone or Intenet service to last year’s immediately introduces error potential.  Extending the conventional reasoning over longer time frames, how would you compare the value of buggy whips in 2018 to those of 100 years ago?  How would you compare the value of a 2018 iPhone to a 1918 iPhone?

But wait, Bill, I can almost hear you object.  Didn’t you earlier complain that these inflation proxies are flawed because they’re too narrow?  Wouldn’t just limiting them to unchanged goods and services make them even more so?

I’m afraid you have me there.  Yes, it would further narrow an already too narrow price base. But next I counter that that problem is a consequence of assuming a priori that goods and services are a valid foundation on which to build this edifice in the first place.  Who even questions this?  But one of the recurring themes of Financology will be how it is the questions that go unasked that are often the most important.

It seems that everything we can think of is subject to changing value over time.  Goods, services, real estate, commodities, stocks, bonds … you name it, it changes.  Even gold, used as money for thousands of years, and which tends to hold its value over long periods of time, is notoriously volatile and subject to change with cycles of supply and demand.

I think we’re not finding an answer just because we’re looking in the wrong places.  We’re looking at things, as if they possess some objective value apart from ourselves.  It’s not things that have value, but us.  Human time.  Our own time is the one thing that we have a fixed supply of every hour, every year, and which has the same value to us today as it did to our ancestors centuries ago.  I submit that an hour of a man’s time has the same value to him in 2018 as it did in the age of Shakespeare.  That value is subjective, and not a property of a thing unto itself.  The economic value of any material thing is ultimately measured by the time we will put into obtaining it.

So if we want to measure changes in the value of anything, including a currency, the standard against which we should measure it is not buggy whips or iPhones, but human time.  Though this may seem at first a venture into the abstract, it is possible to arrive at a quantitative measure of the value of the dollar, and thus of changes in the value of the dollar, and I have done so.  We’ll explore this index of inflation in future posts.