What C.A.R.S. Can Teach Us About Government Programs

Strange that a government program costing $1 billion would be considered small, no?

A small government program can teach us a lot about government:


Referred to in the media and by politicians as “Cash for Clunkers,” the Federal Car Allowance Rebate System gave rebates of $3,500 or $4,500 for qualifying automobiles. It was allocated $1 billion and intended to last five months.  Those who designed the program expected roughly $50 million in rebates would be issued each week over that period.

One month after new car sales qualified and one week after rebate checks began to be issued to dealers, the program was nearly out of money.  A program meant to last five months lasted only one before needing additional cash.

Add to this the government’s listing of certain “qualified” models that were later re-evaluated and disqualified resulting in rejected rebates and returned sales, causing angst and despair among many buyers.  Some buyers had to return their cars while others were forced to pay back the rebate.

As the government scrambled to find funding for the overtaxed program, some dealers suspended offering the rebate.  Even those still offering the program feel dismay that the program takes longer: Sales that used to take 60 to 90 minutes are now taking three hours or longer, to ensure compliance with all the paperwork.  The program requires the dealers to follow-up with junk yards, getting certification that the Clunkers are actually junked.

The C.A.R.S. program is a huge success in that it is getting people to trade in their fuel-guzzling automobiles, but it’s turning out to be an administrative nightmare and costing far more than anyone expected.  This is an important lesson for Americans as we debate other government programs, such as the “Healthcare Reform” bills currently being debated in Congress.

Before we compare the two, however, we need to understand why the C.A.R.S. program ran out of money.  We will not have a government analysis for some time, but the likely cause is that human beings respond to incentives.  Millions of Americans examined the Cash for Clunkers program and determined that getting a $3,500 or $4,500 discount was an excellent reason to trade-in their old car.  What the politicians and bureaucrats were not counting on was that many more people would respond to this incentive than just those already thinking about buying a new car.  In fact, many families who had no intention of trading in their clunker in July chose to do so, something for which the government was totally unprepared.  The result?  One billion dollars spent in just a month instead of half a year.

The politicians and bureaucrats have learned a small lesson in economics well: Give people a strong and positive incentive and they will respond.  This is a basic concept in both psychology and micro-economics: Strong incentives influence behavior.

What the Beltway insiders did not understand was that strong incentives do not just affect those already inclined to a particular behavior.  Rather, incentives can actually result in excessive modification of behavior.  One need only learn about Pavlov’s dog and the Skinner Box to understand this.  While these are normally used as examples of learning and behavior modification, they are also excellent illustrations of the economics of incentive.  When animals (including people) are given an incentive to do something new, we change our behavior to do something we might not otherwise do or do it so often that a healthy behavior is suddenly very unhealthy.

Now, on to health care.

Right now, the bills proposing to overhaul our health care system have an estimated 10-year cost of $1 trillion.  That’s a huge number.  To put it in perspective, the average household in the United States has income of just over $50,000 per year.  One trillion dollars is the equivalent to the annual income of twenty million average American households.

It is expected that these bills will enable the majority of the 35 to 47 million uninsured individuals living in the United States get health insurance, one way or another.  This program is expected to cost about $100 billion per year.  There’s just one problem: In order to pay for this program, the government is going to have to tax businesses and individuals.  Increased taxes on businesses create an incentive to cut costs.  The largest single cost to businesses in the United States is labor.  So the incentive will be for businesses to cut labor costs to pay for the cost of new taxes, or to raise their prices to consumers, resulting in inflation.

Certainly some businesses will choose to lay-off employees while others will cut benefits, for some that will include health insurance.  Even penalizing businesses who drop insurance programs will not stop the tide, since the 8% tax on payrolls currently proposed is less than the cost of many benefits plans.  This results in an increase in the number of uninsured Americans, increasing the number of people who will then qualify for the “public option.”  This increase in enrollment will further increase the cost of the program.  The government will enter into a vicious cycle of increasing revenues (taxes), causing more people to become uninsured, adding more families to the rolls of the “public option.”

So like a large cargo ship that takes more time to get going and requires much more fuel to run than a fishing trawler, the Democrat’s Health Care plan will require much more money and will take longer to turn into the train wreck almost all government programs become.  If Cash for Clunkers is any indication, instead of reducing medical overhead and costing $1 trillion over 10 years, this program will cost three times as much and result in a new mountain of paperwork for medical practices and hospitals.

If we think of Cash for Clunkers as a wind tunnel test for government-payment incentive programs like the health care bill, we as engineers of good policy should reject this design for government.  It is inefficient, costly and while it appears to be effective at achieving its designed goals, we can also see that there are much better ways we could build such programs.