What it means
Most things we buy are bundles of features. A home has floor area, bedrooms, a garden and a location; a laptop has processing speed, memory and screen quality.
A hedonic regression collects prices for many such items and uses a regression (a technique that finds the best-fitting relationship between a number and the factors that influence it) to estimate what the market pays for each feature. The method solves a basic problem in measuring inflation.
If this year's phone costs the same as last year's but is twice as fast, has a better camera and holds more data, the buyer has received more for the same money. Simply comparing sticker prices would miss that improvement, so statisticians use hedonic models to adjust for quality and avoid overstating price rises.
In property, the approach is used to build house price indices and to value homes for lending and tax purposes. A model might estimate that each extra square metre adds a certain amount, that a garage adds another and that a poor school catchment area subtracts a third.
A valuer can then price a home that has never sold by plugging in its features. Businesses use hedonic ideas when pricing products with many options.
A car maker, a software firm or a hotel chain can estimate how much customers value an extra feature and decide whether it is worth the cost of providing it. The result guides decisions on bundles, premium tiers and discounts.
There are limits. The model is only as good as the features included, so leaving out something important, such as the condition of a building, can bias the answer.
Features also tend to be correlated, for example larger homes usually have more bedrooms, which makes it hard to separate their individual effects, and the estimated values can shift over time as tastes change.
In practice
Real-world examples.
Example
A national statistics office tracks laptop prices. Raw prices are flat over a year, but a hedonic model shows that average processing power and memory rose sharply, so the quality-adjusted price index records a fall in prices.
Example
A mortgage lender values a home that has not sold for 20 years. An automated valuation system uses a hedonic model built from recent local sales to estimate a value from floor area, age, plot size and number of bathrooms.
Example
A boutique hotel group analyses booking data to learn what guests will pay for sea views, balconies and breakfast. It finds that a sea view adds around $35 a night, which supports the cost of refurbishing six rooms.
Formula
Calculation
Price = base value + (value per feature x amount of feature) + (value per feature x amount of feature) + ...
Suppose a regression on thousands of home sales in one district finds that a home is worth a base of $60,000, plus $1,500 for each square metre of floor area, plus $12,000 if it has a garage, plus $8,000 for each bedroom.
Consider a home of 120 square metres with a garage and 3 bedrooms.
Floor area: 120 x 1,500 = $180,000.
Garage: $12,000.
Bedrooms: 3 x 8,000 = $24,000.
Estimated price = 60,000 + 180,000 + 12,000 + 24,000 = $276,000.
If the home is listed at $300,000, it is priced $24,000 above what its features alone would suggest, so a buyer would want to know what else justifies the premium, such as a prime location.Case study
Seen in the real world.
Alder & Finch Property Advisers is a fictional valuation firm that was asked by a council to estimate the tax value of 40,000 homes. Instead of valuing each one by hand, the firm built a hedonic model using the sales of 6,000 comparable homes and tested it against 500 homes held back from the model.
The average error on the test homes was about 6%, with larger gaps for unusual properties such as converted warehouses. In this illustrative project the firm sent the odd cases to human valuers and used the model for the rest, which cut the cost of the exercise by roughly two thirds while keeping the results within an acceptable range.
Watch out
Common mistakes.
- Treating the estimated value of a feature as a fixed price, when it is an average that changes by place and over time.
- Leaving out an important feature, such as condition or location, which pushes its value into the other features and distorts them.
- Using the model on items far outside the data it was built from, such as a luxury home in a district of modest flats.
Questions
People also ask.
Why is hedonic regression used in inflation measurement?
It separates a rise in price caused by a better product from a pure price increase, so inflation is not overstated or understated.
What does hedonic mean?
The word comes from the Greek for pleasure, and it reflects the idea that buyers pay for the satisfaction that each feature gives them.
Is it the same as a comparable sales approach?
No; comparable sales look at a few similar properties, while hedonic regression uses statistics across many sales to estimate the value of each feature.
From the founder's library
Take it further with the book.
Build your financial confidence beyond this definition. Shihan's full-length guide, Accounting Fundamentals, takes the same plain-English approach and turns it into a complete, practical playbook for non-finance managers, business owners and students - with chapter-end quiz answers and presentation slides included.
25% off with code MMHQ25, applied at checkout. Priced in USD - checkout may show the equivalent in your local currency.
View the book and save 25%Related