Hedonic Pricing, an Economic House of Cards

Hedonic pricing can roughly be described as a tool used by government economic statisticians to deal with the problem of tracking the prices of goods over time when the actual goods themselves evolve. In particular, it is believed that such national economic statistics as consumer price indices, gross domestic product, and productivity can be improved upon by adjusting the actual goods prices paid in response to changes in the quality of the goods.

There are a number of things wrong with this approach, but for simplicity, I'll try to concentrate on the idea that increases in the quality of goods can compensate for the loss of purchasing power of the dollar due to monetary supply inflation. This is the most important use of hedonic pricing as it is used to attempt to adjust future mandated government transfer payments and union contract wages, for example.

Let's assume that the only economic difference between 2005 and 2006 is that a $10,000 Ford Escort for model year 2006 adds a deluxe trim level as standard and a satellite radio receiver as well, without increasing the $10,000 price.

It is the claim of the hedonic price adjusters that this change is equivalent to a reduction in the price since more consumer satisfaction is delivered for the same price. There is no argument that consumer satisfaction has indeed likely been increased for most consumers. It is the equivalence of a price reduction that is problematical.

If, in addition, the government increases the supply of dollars enough to increase the $10,000 price of an unimproved 2005 model Escort to $11,000, and all other prices to the same degree, this would be a price inflation rate of 10%. The hedonic price adjusters would use their computer models and arbitrary judgements to reduce the reported rate of price inflation below 10% in response to the quality improvements in this single good, the 2006 Escort. The degree of the price inflation adjustment would depend on somehow quantifying the amount of the quality improvement and the actual quantity of Escorts sold. Clearly, if no Escorts were sold, neither their prices or quality should impact the rate of price inflation.

A further reduction in the reported rate of price inflation would occur if GM made similar changes to one of its car or truck models. If a second improved good increases the reduction in the rate of price inflation, then there would seem to be no logical reason that an even further reduction would result if every good in the economy were improved between 2005 and 2006 without increasing their prices before taking monetary supply inflation into account.

In particular, if every good were improved to a degree that would be quantified as a 10% improvement, then the reported rate of price inflation would be 0%, even though the actual dollar prices paid would be 10% higher. While this has a reasonable sound to it, as more consumer satisfaction is delivered per 2005 dollar, and an equivalent amount is delivered per 2006 dollar, the final purpose of the reported rate of price inflation is to track changes in the purchasing power of the dollar.

If, in 2005, the government gives me a check for $10,000, I can buy a 2005 Escort.

If, in 2006, the government gives me a check for $10,000, I can't buy any Escort with it since only the improved 2006 model is available at $11,000. Only if the government check is adjusted for the actual 10% rate of price inflation, and not the hedonically adjusted 0% rate, can I buy an Escort.

Conclusion : A uniform increase in goods quality does not result in changes in prices, nor in changes in the purchasing power of money in terms of goods units. All goods may well be more satisfactory to their purchasers, but this doesn't imply a change in prices.

For a single good, the buy/ don't buy decision rests on the consumer's preference at the margin for the good itself or its purchase price in dollars. A good with increased quality may well turn a don't buy decision into a buy decision.

But if all goods have increased quality, the sacrificed purchase price will be of more significance than before since it has lost the purchasing power over higher quality future goods. Thus, there will be less tendency, if any, to turn a don't buy decision into a buy decision.

Prices are not values. Nor are they measures of value.

The dollar price of a good is the result of supply and demand. Supply and demand of and for the good itself, supply and demand of and for money to hold, and supply and demand of and for all other goods that can be purchased with money.

All economic values are subjective, but the effect on prices is indirect. Going back to the $10,000 2005 Ford escort, I might have been willing to buy it for up to $20,000. I might have been willing to buy the improved uninflated $10,000 2006 Ford escort for up to $22,000. But in all cases my unit demand would be for just one car, and this is my only incremental influence on market prices. Loosely speaking, you can't eat consumer surpluses.

A part of hedonic adjustments is based on the costs of the improvements. This, of course, is backwards. Any sustainable improvements must earn their costs by sufficiently increasing the consumer's subjective valuation of the improved end products.

Background References :

Non-technical WSJ article

Technical Hedonic Pricing Paper (pdf)

Share this

Hedonic methods, inflation,

Hedonic methods, inflation, and college textbooks
The blog Catallarchy has a post about hedonic methods and inflation. And there is a recommended free link to a WSJ article. The U.S. Bureau of Labor Statistics uses hedonic methods to calculate the Consumer Price Index. In other

Abiola, ...In the real world

Abiola,

...In the real world in which we live, the appearance of a 2006 Escort with improved widgets leads to a decline in the sale price of 2005 Escorts still in the car-lots, as well as a decline in the resale value of any of those which have fallen into consumers’ hands, however pristine their condition. ...

This is true but not a valid consideration for the question of Hedonic pricing. The first cars of a given model year are a different good, both for consumers and dealers, than end of the model year clearance cars. It would be incredibly stupid to record a decrease in price while comparing these different goods. However, there is some evidence that this kind of studipity does occur. In the pdf technical reference, they seem at one point to justify their price reductions by comparing one generation of pentium processors when it is the latest generation with its lower price later when a new generation has been introduced.

Also, what you describe is likely best thought as an increase in supply. The reduction in prices that results from an increase in the supply of goods is a real price decrease and does not require a Hedonic adjustment.

They most certainly are measures of value: they’re a measure of the value which all consumers and providers place on goods and services at any given point in time. That said value fluctuates, often erratically, doesn’t falsify this in the slightest.

Economic values are the result of subjective preferences of individual consumers. As such, they cannot be measured. Prices are what result from markets attempting to balance supply and demand.

An analogy -

The boundaries of two adjoining estates on Long Island would be defined by real surveyed co-ordinates and very real fences or walls.

OTOH, the boundary between the ocean and the shore is quite insubstantial and merely represents the contour along which the mass of the ocean happens to exhaust itself against the gradient of the land. This is the better description of price.

Regards, Don

"A uniform increase in goods

"A uniform increase in goods quality does not result in changes in prices, nor in changes in the purchasing power of money in terms of goods units."

The first claim in here is correct, but the second is manifestly false, as it relies on the almost universally untrue assumption that quality improvements always lead to the disappearance of unimproved versions from the market.

In the real world in which we live, the appearance of a 2006 Escort with improved widgets leads to a decline in the sale price of 2005 Escorts still in the car-lots, as well as a decline in the resale value of any of those which have fallen into consumers' hands, however pristine their condition.

"Prices are not values. Nor are they measures of value."

They most certainly are measures of value: they're a measure of the value which all consumers and providers place on goods and services at any given point in time. That said value fluctuates, often erratically, doesn't falsify this in the slightest.

There is indeed a subjective component to hedonic pricing, but this is the case with any attempt to come up with a composite index to measure the inflation experienced by a generic "consumer," as no two people's consumption baskets is the same. Besides, it's better to be somewhat subjective and mostly right than to be "objective" and totally wrong - using the distance of the earth from the moon as a measure of inflation would be subject to miniscule measurement error and as "objective" as anyone could wish, but I doubt most of us would take this option over the Fed's CPI.