What drives car demand?

The Indian car market is exploding. Demand for passenger cars is expected to grow at more than 8 per cent a year till 2011-12. By that time, analysts estimate that there will more than 12 lakh new cars on Indian roads every year.

But what determines consumer choice? And how do car manufacturers determine what consumers want in a crowded market? An increasingly-popular statistical technique is the hedonic analysis, which identifies the contribution of different features to the final price of a car. A recent study by Crisil using this method came up with some surprises.

What do consumers pay for when they buy a car? Do they pay a premium for manufacturer pedigree, factor in technological specs or simply go by the "look"?

These are questions that usually fall in the domain of consumer surveys that form rough and ready guidelines for manufacturers to go by when they launch new models.

An alternative, however, is to work backwards from car prices, using the assumption that each car model is an amalgam of various attributes such as fuel efficiency, length and so on.

Thus, the objective is to split the final or showroom price of a car into prices of individual attributes or features. In the jargon of the statistical modeling technique that attempts this, these "attribute" prices are known as "hedonic" prices.

This approach is increasingly becoming popular in both the US and Europe where car models and variants are a dime a dozen. Research in this area is being funded heavily by some of the auto sector heavyweights like General Motors (GM).

Interestingly, one of the first "hedonic" models ever to be used related to the US car market in the 1930s when a GM statistician, Andrew Court, used it to measure quality change in GM cars and their impact on prices.

The advantage of this approach is that it enables a separation of the different elements that go into the consolidated price of an automobile. It thus enables analysts to home in on the impact of a single factor on price by filtering out the impact of other factors. This can be used strategically in a number of ways.

Suppose a particular feature, say the inclusion of power steering, is associated with a certain premium, say 5 per cent, over a version that does not include one. Hedonic analysis would suggest that a car priced at a higher premium would find few takers since it is higher than the price consumers would be willing to pay for this feature.

However, pegging the price at less than 5 per cent is likely to lead to a sharp spurt in sales since consumers would see a strong "value for money" proposition.

Until recently, the Indian car market did not afford such analysis since the number of models was somewhat limited. In technical terms, the size of the sample was a trifle too small for any kind of meaningful statistical analysis.

However, with the flurry of model launches over the past two years, hedonic price analysis has become possible in the Indian market as well.

Credit rating agency Crisil worked out a hedonic price model, based on a sample of 100 car models and their variants in the mini, compact and mid-size categories. In its list of attributes or features, Crisil took a whole range of technical and other specifications and tried to figure out how these impacted on prices for the aggregate sample.

What counted in this hedonic analysis
Price

• Ex-showroom price in Mumbai (in Rs)

Technical specifications

• Brake horse power (unit of power)
• Engine displacement of a car (in cubic centimetres)
• Revolutions per minute (or speed) of a car
• Width of the car (in mm)
• Kerb weight (overall weight) of the car (in mm)
• Length of the car (in mm)
• Dummy for the body type of the car. It takes a value of one when the car is a "two-box", zero otherwise
• Tank capacity (in litres)
• Ground clearance (in mm)
• Dummy variable for the type of fuel used by the car. It takes a value of one when the car uses diesel and zero when the car uses petrol/unleaded petrol

Standard equipment "" comfort and safety

• Power windows dummy
• Power steering dummy
• Air-conditioner dummy

Type of manufacturer

• Dummy for the type of manufacturer. It takes a value of one when the manufacturer is foreign, zero when it is domestic
• Japanese manufacturer dummy
• European manufacturer dummy
• American manufacturer dummy
• Korean manufacturer dummy

Technical factors included things such as weight and brake horse-power while the non-technical, "soft" factors looked at factors such as the size, the pedigree of the manufacturer and width.

The information on these factors was drawn from company brochures. Prices are ex-showroom (excluding insurance, registration charges and so on). The implicit assumption is that there is a separate market for each attribute where prices are determined following the laws of demand and supply.

There are obvious problems in quantifying some of these "soft" factors. For instance, how does one quantify the simple fact of "foreignness" "" that a car earns a premium because a foreign manufacturer produces it?

To model this, Crisil used "binary" or "dummy" variables "" these take on a value of one if the model possesses a particular quality (say, produced by a foreign manufacturer) and takes the value of zero otherwise.

The impact of size
While it is common knowledge that price varies with the size of the car, the hedonic model gives an interesting thumb-rule for relating prices to length: a 1 per cent increase in length increases the price by 2.6 per cent (the length elasticity is 2.6), with all other features remaining constant.

How does this work? Take, for example, a compact car manufacturer selling a car whose length is 3.8 metres. He wants to launch a mid-size variant that would keep all other features constant and just increase the length to 4.1 metres. This roughly works out to an increase of length by 8 per cent. That translates into a price increase of 20 per cent, going by the elasticity of 2.6.

Thus, if the compact car was priced at, say, Rs 3.7 lakh, the increase in size alone should set the mid-size version at Rs 4.45 lakh. To charge a higher price than this, the manufacturer, of course, needs to add other features.

What doesn't matter
Of the large list of attributes that we considered, a number of factors did not appear, independently, to make any significant difference to price. This means that consumers do not appear to pay a premium for any of these features.

Take ground clearance of a car, for instance. "Tall boy" models such as the Santro may have made strong forays into the Indian market but the Crisil hedonic analysis suggests that high ground clearance per se (which the tall boy models offer) does not fetch a premium. Similarly, wider cars may seem a lot more comfortable but there is no independent premium associated with it.

The fuel efficiency puzzle
Some car companies declare an estimated fuel efficiency parameter that specifies the number of km per litre that the car is expected to give. Since this is not available for all cars, the study analysed the relationship between price and fuel efficiency for a truncated sample for which this data was available.

No significant relationship was found between fuel efficiency and prices. What this seems to suggest is that consumers don't find the official efficiency parameter credible and go by more informal assessments of fuel efficiency. Informal feedback from existing consumers is clearly an important source of information.

Consumers do not use the weight of the car as a proxy for fuel efficiency, either (one could expect an inverse relation between the two since a lighter car could be expected to be more efficient). As a matter of fact, for the lower price range of cars within the sample, the weight of the car has a mildly significant positive coefficient that implies that consumers are actually willing to pay a small premium for heavier cars.

The "foreignness" factor
How much are customers willing to pay for the fact that a car is made by a foreign manufacturer? Our estimates show that the average premium works out to about 5 per cent in the relevant categories.

Thus, without making any changes in a model, a foreign manufacturer can expect to charge 5 per cent extra due to the sheer fact of "foreignness". Among foreign cars, Japanese makers get the highest premium, followed by Korean cars.

Is this an irrational fetish that Indian consumers have for all things that are foreign? Perhaps not. Foreign cars are often associated with better quality in the long term, better post-sales service and so on, and the premium simply captures some of these factors.

The fuel factor
The choice of fuel clearly makes a difference to the price of car. Since diesel prices are significantly cheaper than petrol, this cost advantage manifests in a markup for diesel cars over the petrol variant.

The study shows that the markup on average was about 8 per cent for smaller cars (mini and compact) but went up significantly to 11 percent for the mid-size and the higher segments.

The corollary in terms of competitive strategy is obvious "" a manufacturer launching a diesel version of a mid-size car could hope to gain market share quickly by pricing at a markup of less than 11 per cent over the petrol version.

This seems to be the tack that Tata Motors is following in pricing the diesel version of the Indigo, which is priced at just 5 per cent over the corresponding petrol version.

The power window puzzle
Surprisingly, the inclusion of "power windows" is associated with a premium of close to 10 per cent. This is largely counter-intuitive since it is unlikely that consumers would be willing to pay that much extra just for power windows.

The key to this puzzle lies, perhaps, in the fact that power windows are typically associated with the topmost variant of a model. Usually power windows come as part of a package that includes other accessories such as defoggers. Thus, what is measured as the premium for power windows really picks up the "snob" premium associated with the "luxury" variant of a model.

The strategic implications
The sheer ability to isolate and quantify the price-impact of different attributes makes the hedonic approach useful from a strategic perspective.

Manufacturers can gauge what really matters to consumers when they decide to purchase a car and how much they can hope to extract from the consumer by offering different features.

A word of warning: this is just an exploratory model and there could be some wrinkles. However, as more data comes in, the accuracy and predictive power is bound to improve.

(Abheek Barua is a senior economist and Sonal Varma is an economist at the Crisil Centre for Economic Research)