McLEAN, Va.— After a thorough review of the Cash for Clunkers program, the chief economist of the National Automobile Dealers Association (NADA), Paul Taylor, determined that the cost of each incremental vehicle sold was around $4,587. An incremental sale is a sale that would not otherwise have occurred without the Clunkers program.

His findings bring into serious question the methodology behind the $24,000 estimate promoted last week by the car-buying Website Edmunds.com.

“It’s really not that hard to determine a credible cost estimate for the Clunkers program,” says Taylor. “You subtract projected sales versus actual sales for July and August when the Clunkers program was operating, and divide the program’s $3 billion by that number.”

Taylor says that, based on sales volume for previous months, a realistic projection of auto sales for July and August would be around 1.6 million. Actual sales for those two months totaled 2,253,963. The difference is 653,963. That’s the number of incremental sales generated by the Clunkers program. Divide the program’s $3 billion by that number and you get $4,587, the average cost per incremental Clunker sale.

Taylor says the methodology used by Edmunds.com in its analysis of the Cash for Clunkers program is “fundamentally flawed.”

“The analysis by Edmunds.com is wrong on the two main points that it tries to make. First, because of is flawed methodology, the study can’t form the basis for measuring the program’s impact or costs. Secondly, and more importantly, the analysis clouds understanding and misleads rather than clarifies the true state of auto sales and the economy,” Taylor added.

Edmunds.com says it came up with its estimates by examining the sales trend for luxury vehicles and others not in Cash for Clunkers, and applying the historic relationship of those vehicles to total SAAR (Seasonally Adjusted Annual Rate). Taylor says this method virtually assured that cost estimates would be “overstated and inflated.”

“Historically, over the past 20 years, auto sales have been lower in July and August than in June, in the absence of strong incentives. Edmunds ignores this,” says Taylor.

“The Dow Jones and broader market indexes made strong recoveries over the summer, assisting luxury light vehicle sales. But there is a fundamental difference between what drives luxury car sales and non-luxury sales,” says Taylor. “An improving stock market, for example, may boost luxury car sales but it has little effect on non-luxury sales. Job growth, income growth and housing affect non-luxury vehicle sales. And in each of those categories the numbers are not good. Unemployment is up. Income is down. And housing prices continued to fall through July. This has not been the kind of economic environment that encourages a purchase by the average car buyer.”

Taylor says that Edmunds.com also painted a too-rosy picture for auto sales for the rest of the year. He points out that the Conference Board’s Consumer Confidence Index actually dropped to 47.7 in October from 53.4 in September. “This does not support a dramatic increase in non-luxury car sales for the final months of 2009, as suggested by the estimates in the Edmunds.com analysis,” says Taylor.

“Edmunds.com has obviously underestimated just how much the Clunkers program stimulated car sales,” says Taylor. “The Clunkers program lit up the market. Auto showrooms went from almost empty to overflowing. It’s hard to imagine how anyone who takes an objective look at the Cash for Clunkers program can reach any conclusion other than it gave a dramatic boost to retail sales and manufacturing output.”