The most powerful auto designers don't work for car companies.

They are employed by the privately funded Insurance Institute for Highway Safety and the tax-supported National Highway Traffic Safety Administration.

Both organizations affect nearly everything that goes into, onto -- or is taken off of--new cars and trucks.

Look at the disappearance of spare tires from the rear hatches of sport-utility vehicles. They are going away mostly because auto insurers don't want them there.

Findings by the insurance institute show that bracket-mounted spare tires cause considerable damage to the rear hatch doors and windows of SUVs in low-speed collisions. Insurers don't want to pay those property damage costs. So, the spares are going underneath new SUVs; or, they are being installed in the rear cabins of those vehicles, as is the case with the 2003 Hummer H2.

Ever heard of LATCH? It stands for "lower anchors and tethers for children"? It's a new, federally required anchoring system for child-safety seats. NHTSA believes that the LATCH system does the best job of holding those seats in place during crashes. Seats that hold infants and toddlers in place are seats that help to keep them alive. So, LATCH is in.

Both NHTSA and the insurance institute rely on crash-test results and real-world experience to help influence automotive design. But, when they feel it's necessary, they also use brute force.

For example, if a particular automotive design proves especially costly, either in terms of property damage or medical claims, insurance companies can raise the cost of insuring the affected vehicles; or they can refuse to insure them altogether.

High insurance costs on a vehicle can reduce its attractiveness to buyers. Car companies know this, which is why they are willing to design cars and trucks or change their designs to please insurers.

NHTSA can force a company to recall faulty vehicles. Technically, that means NHTSA asks the Justice Department to bring the target company to court on what amounts to a charge of public endangerment. But NHTSA very seldom has had to do that, because the publicity surrounding a forced recall can kill the affected product, or possibly cripple the company.

In the early 1980s, for example, General Motors Corp. decided to take on the government in a recall battle over its Chevrolet Citation and its other compact X-cars, which were cited by the government for allegedly defective rear brakes that could cause fatal crashes.

GM won the legal fight, but lost the car -- and a good bit of its reputation for quality and customer care in the process. Most automakers don't want to go through that sort of thing, which is why the overwhelming majority of safety recalls are voluntary.

Ultimately, it all comes back to tests and interpretations of real-world events, what auto safety researchers call "field experience" -- and the ability of both the insurance institute and NHTSA to generate publicity for their work.

The insurance institute, for example, has mastered the public relations game. It has a deal with Dateline NBC, a national TV news and entertainment magazine. Dateline gets first rights to air the institute's crash-test footage, which always makes for "good TV," as they say in the business. The insurance institute gets its message out, usually unchallenged, or with meek disclaimers from car companies that get poor test results.

The automakers' reticence stems from the two-faced manner in which they use crash-test findings. If the results are good, the companies include them in their advertising campaigns, thus giving those results high-profile value. But if the results are bad, the companies either publicly ignore or minimize them.

None of this is to suggest that testing done by IIHS or by NHTSA is invalid. Quite the contrary is true. Testing done by the insurance institute and NHTSA often are complementary and frequently lead to major improvements in vehicle design.

You'll note that said "improvements in vehicle design" as opposed to "vehicle safety." That is because IIHS and NHTSA have two very different interests. IIHS primarily wants to save its clients' money. That is why its clients, the auto insurers, are so willing to pay the institute's considerable bills.

NHTSA's basic goal is to save lives -- which, in the cynical world of dollars and cents, can undermine the insurance industry's profitability.

Ask any insurer who is willing to speak frankly: Is it more expensive to put a crash victim in the grave, or in a long-term care hospital bed, followed by extensive rehabilitation therapy? I've asked insurance agents that question many times; and "the grave" always was chosen as the least expensive way to pay for victims of serious auto crashes.

With that in mind, look at the latest IIHS crash tests, these involving a bumper's ability to withstand a "low-speed" crash of five miles per hour with minimum property damage. Repeat, "minimum property damage." Bumpers don't protect people. Bumpers protect property, the property insurers will have to pay for after subtracting your deductible.

The insurance institute tested two midsize and three small cars this time around. The midsize models included the 2002 Audi A4 near-luxury sedan and the 2003 Honda Accord. The small cars included the 2003 Toyota Corolla, the "premium small car" 2002 Mini Cooper, and a 2002 Suzuki Aerio.

The institute does a series of four bumper worthiness tests -- front into flat barrier, rear into flat barrier, front into angle barrier, and rear into vertical pole. A repair cost estimate is given for damage incurred in each test; and average cost estimates are given for damage incurred in all four tests.

The media often report these costs as costs to the consumer, ignoring why consumers pay insurance premiums in the first place. Again, the reality is that the bulk of those costs go to the insurer after you pay a deductible. Yes, your insurance rates are likely to go up as a result. But that assumes that you caused the accident, that no one bumped into you or forced you to bump into a pole or a wall.

Real-world accidents generally don't happen the way they happen in controlled, laboratory tests. But I'll say more on that in a closing note with a dramatic real-world example.

Suffice it to say that the Audi A4 was the winner in the latest insurance institute bumper tests with an average damage cost of $216 for all four crashes. IIHS gave the A4 a "good" rating. The Honda Accord, with an average damage estimate of $339, got an "above average" rating. The Mini Cooper (one of which I own, by way of disclosure), with an average damage estimate of $800, received a "marginal" rating. The Toyota Corolla, $296, received a "good" rating. The Suzuki Aerio, with an average damage estimate of $1,131, got slammed with a "poor" rating.

So, does that mean you should go out and buy an Audi A4 because of its "good" bumper performance? Hardly. The A4's base sticker prices range from $24,950 to $34,340, compared with $15,800 to $27,900 for the Honda Accord; $16,245 to $19,245 for the Mini Cooper; $13,570 to $15,480 for the Toyota Corolla; and $14,019 to $15,519 for the Suzuki Aerio.

The total cost of the sticker price comes out of your pocket. If you are willing to pay from $6,440 to $9,150 more for an A4 over an Accord because if you happen to have an accident it may save your insurer money on property damage, so be it. But keep in mind that the substantially more expensive A4 usually comes with a higher insurance premium in the first place.

Again, as for what the bumper tests don't mean: They don't mean a thing about vehicle occupant safety. For example, look at that "marginal" bumper rating on the tiny Mini Cooper and how that car actually performed in a horrendous real-world auto crash in California last November.

This comes to us from Marcus_San Diego, as posted November 28, 2002 on the owner-controlled Mini Cooper Message Board.

Marcus said his wife was at the wheel at the time of the rollover crash at 8:50 p.m. in late November on the 52 West Freeway in San Diego. He said the car was traveling at the 65 mph speed limit when his wife swerved to avoid a collision, overcorrected her steering and flipped the Mini, which skidded on its roof a considerable distance before coming to rest. The couple was belted. The car was totaled. But "we both, miraculously, only sustained minor -- and I mean minor -- cuts and bruises," Marcus said.

Frankly, I don't mind sustaining higher property damage costs in a low-speed fender-bender . . . as long as my Mini can protect me the way it protected Marcus and wife in a major smashup on the road. Here's hoping that the insurance institute will keep that in mind in doing whatever it's going to do to improve Mini bumper strength.