June 14, 2009

How we measure inflation

Traditional inflation measures have been looked at from the perspective how much a basket of goods cost today compared to last year. However this approach fails to measure the output or the use of this basket of goods. The question I am asking is: do we garnish the same utility of that basket of goods or do we yield higher output with the better technology?

As technology progresses, consumption of anything is yielding higher utility of use as things are becoming more efficient, last longer, use less energy,...etc. Aside from using more of things, the cost per unit of consumption should have gone down over the years with increase in efficiency, innovation and productivity of things. Lets look at energy as an example. Please note that the point I making here is only valid over multi-year periods, decades maybe as technology does not progress overnight.

We have always seen the graph of energy cost over the years going up and up,

The oil put into our cars now allows us to travel much longer distance than say 20 years ago. When you went from point A to B in the 1970s, you consumed more energy than travelling the same distance now. The energy consumption per mile traveled has gone down. See the second graph which illustrate that each mile traveled is costing us less and less in term of energy consumption as measured by BTU per mile traveled. Drivers most certainly realized improved output by increasing the number of miles per unit of energy used. So the question is the cost per mile traveled have gone down?

A barrel of oil produces 5,800,000 btus, a measure of energy. Then if I divide the historical cost of a barrel of oil by that factor I would get an approximate cost of the passenger mile. The result show that there is still inflation in figures but at much lesser extent, see graph 3. Oil prices from 1960 has gone up 21 folds while cost per passenger mile has gone up some 14 folds.

The same concept can be applied to the basket of goods that we measure inflation with. Electricity can be analysed in the same manner.

The cost of production and technology would have declined resulting in higher output per unit consumed. Off course on the aggregate we are using more of everything compared to 20 years ago. We are driving more, using more lights, using more electricity...etc So our total use of energy and electricity has gone up. But the cost per per unit of energy used has not gone up as high as we think.

This is me thinking out loud, I may have missed some critical issues in my analysis as I am no expert on the economics of inflation. Well, may be I am, I can't be more wrong than anyone else.


Neil said...

Interesting analysis. This factor is probably already more or less accounted for through the CPI formula as quality improvement. In greatly simplified terms, when the government puts together CPI numbers, they look at items like cars and say "this car costs 20% more than last year's model, but because it burns less gas, has a longer expected lifespan, and has more features, it's really only 5% more expensive." Turning around and reducing the effect of increased energy costs would be double dipping on that same quality improvement.

CPI isn't really an exact science, and there's probably more people who think CPI understates "real" inflation than think the opposite. Which camp is right, I couldn't tell you.

Sami said...

with subjective measures no wonder there are so many shadow CPI statistics out there.