Case Study #1: Mortgage Interest Deduction for Owner-Occupied Housing
These results are part of an eleven-part series, The Economics of the Blank Slate, created to discuss the economic effects of repealing various individual tax expenditures. In these reports, Tax Foundation economists use our macroeconomic model to answer two questions lawmakers are considering:
- What effect does eliminating these expenditures have on GDP, jobs, and federal revenue?
- What would be the effect on GDP, jobs, and federal revenue if the static savings were used to finance tax cuts on a revenue neutral basis?
Eliminating the deduction of mortgage interest for owner-occupied housing would:
- Increase tax revenues by $101 billion on a static basis;
- Reduce GDP by $254 billion;
- Increase revenues by $39 billion on a dynamic basis;
- Reduce employment by the equivalent of approximately 659,000 full-time workers; and
- Reduce hourly wages by 1.1 percent.
Trading the static revenue gains solely for individual rate cuts would:
- Allow for an across-the-board rate cut of 8.7 percent;
- Lower GDP by $107 billion per year;
- Reduce federal revenues by $26 billion on a dynamic basis;
- Increase employment by the equivalent of approximately 187,000 full-time workers; and
- Reduce hourly wages by 0.9 percent.
However, using the static revenue gains to pay for 100 percent expensing for all non-corporate businesses and a 6.8 percent across-the-board rate cut would:
- Increase GDP by $50 billion on a dynamic basis;
- Increase federal revenues by $8 billion;
- Increase employment by the equivalent of approximately 255,000 full-time workers; and
- Increase hourly wages by 0.1 percent.
The federal individual income tax permits taxpayers who itemize to claim a deduction for the mortgage interest they pay on their primary residence. The government limits the amount of mortgage debt that qualifies for the interest deduction. A taxpayer can deduct the interest on no more than $1,000,000 of debt incurred to acquire or substantially improve a primary or second home. A taxpayer can deduct the interest on no more than $100,000 of home equity debt. (The limits are halved if married filing separately.)
The deduction is controversial. Two defenses are that there would otherwise be double taxation, in that the mortgage lenders are taxed on the interest they receive, making the current practice largely tax neutral; and that the deduction promotes a social objective of encouraging home ownership. The chief criticisms are that it is mostly claimed by upper-income taxpayers, and that housing is given less punitive tax treatment than many other forms of saving or investment. However, the purpose here is not to discuss those issues but simply to examine the growth effects if the deduction were repealed.
In a conventional static revenue estimate that holds the size of the economy fixed, the Tax Foundation's simulation model estimates that abolishing the federal income tax deduction for mortgage interest would have raised federal revenue by $101 billion in 2012. The Treasury estimate is $82 billion, and the Joint Committee on Taxation estimate is $69 billion. (The reason for this difference is that our model draws on 2008 IRS data, so it is likely that mortgage interest rates and deductions are probably higher in our base data set than they are at present.)
When the unrealistic static assumption is relaxed, the Tax Foundation model predicts that ending the deduction would cause some economic harm. Losing the deduction would push some people into higher tax brackets, and the people affected would respond to the higher marginal tax rates by working and investing less. In addition, the higher cost of home ownership would somewhat reduce the value of the owner-occupied housing stock, either through lower home prices or the building of smaller housing units over time. These incentive and price effects would make GDP lower by about $254 billion at 2012 income levels. Because of the negative economic feedback, the estimate of the dynamic revenue gain would be roughly one-third the amount of the static estimate, about $39 billion. (See Chart 1.)
If the revenue gain were used to finance an across-the-board federal income tax rate cut of 8.7 percent of the marginal tax rates, the decline in GDP would be only 40 percent as large as the impact of a straight elimination of the deduction, about $107 billion. Federal tax collections would fall by a net $26 billion. (See Chart 2.) From a growth perspective, this is not an attractive trade.
Capital (even owner-occupied housing) is quite sensitive to taxes, more so than the supply of labor. A tax change that is entirely focused on raising the cost of property may do more economic harm than may be offset by a dollar-for-dollar tax rate cut falling on wages, interest, and non-corporate business income. The trade would have to be justified on redistribution grounds, or on the grounds that other forms of capital formation are somehow more valuable to society.
In contrast, we also ran a dynamic simulation in which the tax increase on owner-occupied housing investment would be countered by a tax decrease in other investment. We used part of the static revenue gain, $23 billion, to increase the cost recovery of all non-corporate investments in equipment, software, non- residential structures, and residential (rental) structures. These were given 100 percent expensing (that is, they were written off immediately in the year they were bought, instead of being depreciated over time). This did not use up all of the presumed static tax increase. If the remaining static revenue gain of $78 billion were then used to fund a smaller 6.8 percent reduction in income tax rates, the combined offset would generate a net rise in GDP of $51 billion, although the federal revenue gain would be trimmed to $8 billion. (See Chart 3.) This approach, adding expensing, would yield more GDP than a tax rate reduction alone.
Finally, we determined the impact of these scenarios on employment and wages. We found that simply eliminating the mortgage interest deduction would reduce employment by the equivalent of about 659,000 full-time workers and cut hourly wages by 1.1 percent. With the rate cut offset, employment would increase by the equivalent of about 187,000 full-time workers but hourly wages would be cut by 0.9 percent. With expensing of investment and a smaller rate cut, employment would be increased by the equivalent of about 255,000 full-time workers and hourly wages would rise by about 0.1 percent.
 For instance, if a couple have a $1.2 million home acquisition mortgage at 4 percent interest, they may deduct the interest on $1 million of that, or $40,000. The cap applies to the amount of the debt, not the amount of the interest. If the interest rate on the couple's mortgage were 6 percent, they could deduct $60,000.
 Normal treatment of interest on an investment is for borrowers to deduct interest paid and lenders to pay tax on interest received. Except for the net profit of financial intermediaries, financing flows related to purchases of goods and services do not add to GDP and income. GDP is the total of the goods and services. As any economics textbook makes clear, adding the payment arrangements to the total would be double-counting.
 The remedy here would be to stop double taxing the other forms of saving.
 In evaluating the revenue and marginal tax rate effects, we used the SOI sample of tax returns for 2008 that are the basis for our model, and grossed the results up to 2012 values. In evaluating the impact on the housing stock, we took the lower JCT estimate of the tax increase for a more conservative estimate, assuming that the JCT may have a more recent estimate of the current housing stock following the bursting of the housing bubble, and may have made some adjustment for the recent decline in mortgage interest rates.
 We assume proportional cuts in all of the ordinary income tax bracket rates but no cuts in the lower tax rates on capital gains and qualified dividends.