Washington thrives on perversity: The more something is worth doing, the less likely it is to be done. Obstacles to change usually lie in popular prejudice and vested interests -- not of the few, but of the many.

Your textbook example is a gasoline tax. A stiff increase in the tax in 1974 would have better prepared the country for the inevitability of insecure oil supplies. But public opinion wouldn't have it; its prospects were zilch.

The same logic is at work today. We have a tax system that impels consumers to borrow too much, save too little and throw too much into housing. But no one is about to abolish the system, because a cherished tax provision lies at its center: the tax deduction allowed Politicians talk incessantly these days about "supply-side" economics: the sensible proposition that inflation results from distortions between supply and demand, caused in part by twisted tax policies. By that test, the interest rate deduction would fall in a moment. It distorts in spades: First, it has fanned housing speculation and inflation: and second, by denying business needed capital, it has hampered efficiency.

But politicians do not take their own rhetoric seriously, especially when it means confronting a well-entrenched and popular privilege. The interest-rate deduction, associated as it is with home ownership, enjoys such sanctuary.

That's a pity. It's conceivable that the deduction, by encourageing overconsumption of housing for interest payments on mortgages and consumer loans.

When government invites people to borrow, they do. Combined with high inflation, the tax system now overwhelmingly favors housing investment against other investments -- say, corporate stocks or bonds. Consequently we get too much housing, though not necessarily enough housing. Upper-middle-class taxpayers (the main beneficiaries of the deductions) simply buy larger homes. Meanwhile, they aren't investing in business, which helps explain why American industry has so much trouble raising funds.

We face here Washington's invisible, institutionalized hypocrisy-by the upper-middle class, now may be making home ownership more difficult for the young. More important, overhaul of the tax system now may depend upon the elimination of existing distortions and the use of resulting revenue gains for general rate reductions. Spending pressures, primarily for defense and Social Security, will make any broad tax reduction difficult.

Elimination of the interest-rate deduction, therefore, need not (and should not) mean higher taxes for the upper-middle class. In 1978, two-thirds of the interest claimed as deductions came from taxpayers with incomes between 18,000 and $50,000. This group, which also paid more than half the total personal tax, represents immense economic power. Its wealth needs to be rechanneled.

That the current system favors housing is not seriously disputed. Suppose you invest in a 10 percent corporate bond. Most of the interest simply compensates for inflation, which erodes your original investment. Nevertheless, the government taxes all the interest (not simply the amount above inflation) as income. The same problem applies to most increases of stock. Original investment is being taxed away.

Now suppose you take out a mortgage for a home. If your marginal tax rate is 30 percent, the interest-rate deductin reduces the effective rate of a 13 percent mortgage to 9 percent. Not bad when inflation is between 9 percent and 11 percent.

Moreover, much of the gain in a house's value isn't taxed. Any proceeds from a sale escape the capital gains tax if reinvested in housing within 18 months. For those over 55, the first $100,000 of housing gain escapes the tax entirely.

Given these inducements, everyone wants to own a home. The number of owners today may be about 10 percent higher than it would be with low inflation, according to Purdue University economist Patric H. Hendershott. And the cash savings flowing from the tax deduction enable people to buy slightly bigger homes -- estimated at as much as between 5 percent and 15 percent bigger.

Coinciding with the housing demands of the "baby-boom" generation, all this has augmented inflation. Demand for land and building materials has increased, helping boost the price of new homes from $39,300 in 1975 to $62,900 last year. The young, who didn't get in on the boom early, are increasingly priced out of the market. That is just what government policy isn't supposed to do.

But the evils of the system go deeper. We may be perpetrating a giant investment hoax. People have skimped on savings in part because they see their appreciated homes as savings. For many, they have been. But savings materialze only when houses are sold for less costly shelter.

And there's the rub. Markets go up, and they also go down. Today's baby-boom buyers have created hugh upward price pressures, but the increase in first-time buyers gradually will subside and someday may be replaced by huge selling pressure. Aging families may want to trade in large houses for something smaller and use the surplus cash. Will they be able to? Will appreciated values survive?

Just as artificially low gasoline prices gave misleading signals about what cars to buy, the tax system may be giving misleading investment signals.

Nothing prevents Congress from undoing this mess. The interest-rate deduction results in huge tax losses, between $19 billion and $20 billion in fiscal 1981. Surely it would be unfair to abolish it overnight on existing loans, but eliminating if for new loans would provide a natural phase-out period. Home tenure averages seven to eight years; consumer-loan maturities are shorter.

Don't hold your breath. The first effect would be a cut in housing price inflation or even a price decline; new buyers simply couldn't afford to pay as much. The size of new homes (which increased one-seventh in the 1970s) would diminish slightly to reduce their costs. None of this is good politics. Although everyone favors lower inflation, we have created a hugh constituency of existing homeowners who want and expect rising home prices. It's perverse. o