Almost everyone thinks we should broaden the tax base -- that is, reduce the number and size of tax deductions, exclusions and credits.The Bowles-Simpson fiscal-reform plan went after tax breaks so aggressively it was able to slash the deficit and still cut tax rates. Most conservative reformers follow that model. Even economists who aren't conservatives generally say the government should raise revenue by getting rid of tax breaks before it raises rates. That is the position Republican leaders in the House and Senate have taken, too.As pleasing as it is to find a consensus in this polarized era, this one is mistaken. At the top of the Congressional Budget Office's list of the biggest tax breaks is the exclusion of employer-provided health insurance from income and payroll taxes. You pay taxes on your wages, but not on your benefits, encouraging employers to provide more of your compensation in the latter form. The policy arose after employers used benefits to evade World War II-era wage controls. If we were designing a code from scratch, we would certainly tax these benefits.We have had this break for an awfully long time, though. It would be disruptive, as well as unpopular, to get rid of it. The exclusion should be modified so it doesn't reward people for choosing more-expensive policies, and so people who have no access to employer plans can buy insurance. But we won't get big savings from shrinking this break and shouldn't try.Several other large breaks benefit savings and investment, the most famous being the low tax rate on capital gains and dividend income. In a way, though, these aren't tax breaks at all: They are a partial corrective to the code's bias against saving and investment.Economists generally dislike the mortgage-interest deduction, another big one. With housing finally showing signs of a recovery, though, it seems like an inopportune time to fight against this deduction.Anyway, Realtors in every congressional district in the country, all of whom know a lot of people in their communities, would surely jam the phone lines on Capitol Hill if anyone tried it.The tax credit for children also can be justified as a partial corrective to other federal policies. By raising kids, most parents make a contribution to the future of Social Security and Medicare beyond what they pay in payroll taxes. Fully recognizing that double contribution would require expanding the credit, not shrinking it.The deduction for charitable contributions is another sacrosanct tax break -- and that's fine. How many people really think the tax code has wreaked damage on American society by encouraging more charitable giving than would otherwise take place?Rounding out the list of the top breaks is the deductibility of state and local taxes. It's hard to see the justification for making people in low-tax areas pay more to subsidize states that choose to have larger governments -- especially when many high-tax states have higher incomes than the low-tax ones and the biggest beneficiaries of the break are high earners.Even this break won't be curtailed easily. Politicians in high-tax jurisdictions tend to be strong supporters of this deduction. Their voters would get hit hard if this deduction were scaled back, and it might force changes to their budgets as those voters grew more sensitive to state and local taxes.The Reagan administration tried to eliminate this deduction in the 1986 tax reform but failed because of opposition from high-tax areas.The more closely you look at the tax code, the less promising the idea of broadening the base becomes. That may be why Mitt Romney's campaign proposed imposing a cap on the total value of all deductions anyone can take (and why so many congressmen are now talking about the idea). It is more or less an admission that paring back the number and size of specific tax breaks is more political trouble than it is worth in revenue.Ramesh Ponnuru is a Bloomberg View columnist and a senior editor at National Review.