- Some argue this money would stimulate the economy like an equal-sized tax cut, but that view is simplistic. A tax cut boosts personal disposable income. Unless the mortgages being refinanced are held by the government, the Federal Reserve, or foreigners, refinancing merely redistributes pre-tax personal income between individual borrowers and lenders, with no effect on personal income.
- Therefore, any direct impact on consumer spending would depend importantly on the differential response of borrowers and lenders to the changes in their respective incomes. Empirical estimates suggest this differential is modest.
- In addition, recent empirical work suggests the take-up rate on such a program could be small, on the order of just 10%, or roughly 3 million mortgages.
- All told, we estimate that, at most, such a plan might boost GDP growth by 0.1 to 0.2 percentage point.
- A benefit of the plan might be to help stabilize house prices which, through wealth effects, could support consumption and, by breaking expectations of further price declines, help encourage an earlier turnaround in housing construction.
- Such a plan requires support of the Federal Housing Finance Agency (FHFA) or Congressional approval. Neither is assured.
In one version of this plan, proposed a year ago by Glenn Hubbard and Chris Mayer, the GSEs would require mortgage servicers to send streamlined applications for refinancing to eligible homeowners.[4] Servicers’ fees would be wrapped into the new mortgages to reduce costs to taxpayers while allowing borrowers, many of whom are financially distressed, to amortize these fees. The agencies would issue new mortgage-backed securities (MBSs) to retire mortgages in existing MBSs. New mortgages would receive priority over current second liens.
The main argument advanced for such a program is that the reduction in mortgage interest payments will free up cash flow that households will spend. The plan would also help homeowners keep their houses and, by reducing defaults and foreclosures, alleviate downward pressure on house prices while reducing defaults and hence losses incurred by the GSEs and investors in mortgage-backed securities. Investors who have not “priced in” such early refinancing by relatively risky borrowers would suffer losses.[5] Putting aside issues of political practicality, the idea holds some appeal. However, we doubt that it would give a quick and major boost to overall consumer spending.
Spending by Borrowers
Consider this first from the borrowers’ perspective. Some estimates suggest that if all 37 million or so GSE-backed mortgages are refinanced, nominal pre-tax household mortgage interest payments would be reduced by as much as $84 billion per year. Most homeowners claim mortgage interest as an itemized deduction, and we estimate that these deductions are claimed at an effective marginal tax rate of around 20%. Hence, the nominal after-tax saving comes to (1-.20)*84 = $67 billion, or $58 billion in inflation-adjusted terms. If homeowners spent all of this during the ensuing year, it would raise the annual growth of real personal consumption expenditures (PCE) by 0.6 percentage point, and the growth of real GDP by 0.4 percentage point, both before any multiplier effects.
This is substantial—indeed, on a par with our estimates of the stimulus associated with the current payroll tax holiday—but the calculation assumes the take-up rate on the program is 100%, that the mortgages are refinanced immediately at today’s rock-bottom interest rates, that homeowners spend all the after-tax cash-flow relatively quickly, and that investors who suffer losses don’t reduce their spending.
The Take-Up Rate
At least one empirical estimate suggests the take-up rate would be far lower than 100%. Researchers at the Congressional Budget Office and the Massachusetts Institute of Technology recently used a model of the refinancing decision to assess how many mortgages sponsored by Fannie Mae and Freddie Mac would be refinanced if, for one year (2012) and assuming CBO’s economic projections, current restrictions on loan-to-value ratios and credit scores were lifted but no principle forgiven.[6] The model considers the expected costs and benefits to the homeowner of refinancing. The authors projected that 2.9 million mortgages—or just 10% of the total—would be refinanced in the first year, with an interest saving to borrowers of only $7 billion.[7]
Spending by Borrowers
At the aggregate level, a typical estimate of the marginal propensity to consume (MPC) out of disposable income is 0.7. However, if disposable income is decomposed into transfer payments, labor payments, and asset income (which includes “rental income of persons,” the net income derived from home-owning), the MPC on asset income usually is estimated at between 0 and 0.4, consistent with the Life-Cycle Model (LCM) of household behavior; our own estimate is around 0.2. Such estimates imply that not all, or even most, of the mortgage interest saved during a refinancing would be consumed. This also is consistent with survey data suggesting that during previous refinancing booms much of the interest saving was used to pay down mortgages faster—that is, it was saved rather than consumed.
The counterargument is that such low estimates of the aggregate MPC on rental income conceal important differences between consumers. At least some homeowners likely to refinance under the program might have become cash-constrained during the housing collapse and now are so financially strapped as to spend all the interest saving. We believe the truth lies somewhere in a range between the maximum value of 1 and our aggregate estimate of 0.2, but the resolution of this thorny empirical issue, which requires micro or panel data, is beyond the scope of this Focus.[8] Still, even for values at the top end of this range, the effect on aggregate spending would be modest for the estimated take-up rates discussed above (see below).
Spending by Lenders
Some argue that a refinancing program would act as a personal tax cut, but this view is simplistic. In the National Accounts, a personal tax cut boosts personal disposable income. A mortgage refinancing might not, because one person’s mortgage interest payment is another person’s interest income. Hence, the refinancing plan could simply shift the composition of pre-tax personal income away from “personal interest income” towards “rental income of persons.” In that case, the net stimulus generated by the refinancing would reflect the difference between how much borrowers raise spending as their rental income rises and how much lenders reduce their spending as their interest income falls. Once could reasonably argue that the MPC of mortgage lenders is lower than the MPC of mortgage borrowers. However, in a typical consumption function the two MPCs usually are constrained to be equal, in part because, at least with aggregate data, it is difficult to reject the hypothesis that they are, in fact, the same. Hence, in such equations, the composition of personal income between interest and rental income does not affect total PCE.
Even so, a differential effect on PCE can arise if the mortgages or mortgage-backed securities are held by foreign investors or if the mortgages are held directly by the federal government, a government agency, or the Federal Reserve—entities not treated as “persons” in the National Accounts. We estimate that roughly 40 percent of the relevant investments fall into these categories. Hence, with every dollar of interest saved through the refinancing program, personal disposable income might rise by 40 cents, reflecting a one-dollar increase in rental income of persons only partially offset by a 60-cent decline in personal interest income. Even if borrowers and lenders share a common MPC, there would be some net stimulus. That stimulus would be greater if, indeed, borrowers responded more strongly than lenders to a shift of income between them.
A Range of Estimates
Simple calculations can establish a reasonable range of estimates for the stimulus to consumption likely to be generated by the refinancing program. At one extreme, suppose that: (a) the take-up rate is 30% with nominal interest saving to borrowers of $21 billion in the first year; (b) borrowers spend all the after-tax cash flow; (c) lenders don’t reduce their spending in response to the decline in their interest income. In that case, real PCE would rise by roughly $15 billion (in 2005 dollars), or 0.2 percent, which would add between 0.1 and 0.2 percentage point to GDP growth before multiplier effects. At the other extreme, assume that: (a) the take-up rate is 10% with nominal interest saving to borrowers of $7 billion in the first year; (b) borrowers spend only 60 percent (or $4.2 billion) of the extra cash flow; (c) personal investors, who see their investment income fall by $4.2 billion (60% of $7 billion), reduce their spending by 20% of that, or $0.8 billion. The net effect would be to boost real PCE by $2.9 billion, which essentially would be lost in the rounding, even with multiplier effects.
Other Benefits
Such calculations leave us skeptical that a refinancing program would quickly jolt consumer spending. However, it could help some homeowners avoid foreclosure, and thereby work gradually to alleviate downward pressure on house prices. Since housing wealth is an important determinant of aggregate consumption, the plan could help backstop consumer spending indirectly through this wealth channel. Furthermore, helping break the expectation of falling house prices could encourage an earlier turnaround in the housing market by lowering the “real” mortgage rate. Complicating that story, however, is that public discussion of the proposal could alert investors in mortgage-backed securities to their potential losses if the plan is enacted. Fear of such losses could encourage capital to leave the sector, pushing nominal mortgage rates higher long before there is any change in the expected rate of change in house prices. Higher mortgage rates now would, of course, work to delay any turnaround in housing activity.
Moot Point?
To implement the plan would require cooperation of the FHFA, the regulatory agency that oversees Fannie Mae and Freddie Mac. To ensure that cooperation, the Acting Director of the FHFA, Edward DeMarco, who was not appointed by President Obama, would have to be convinced that the plan would not lower taxpayers’ return on the assets of the two agencies.[9] Some observers expect such a large-scale refinancing would lead to a significant reduction in revenues on the assets held by Fannie and Freddie. The issue is whether the cost of the lost revenues might be offset through a lower default rate on the newly issued mortgages. Even so, the FHFA could resist the proposal to maintain the agency’s political independence. Without the support of the FHFA, the Administration would have to propose legislation to Congress in order to enact the program, passage of which seems only a remote possibility.
This is from a commentary that was published on October 18, 2011.
[1] “President Obama’s American Jobs Act” (September 8, 2011), p. 17. This set of “talking points” was made available to the press shortly before the President’s speech before Congress.
[2] These are Freddie Mac, Fannie Mae, Ginnie Mae, The Department of Veterans Affairs (loan guarantee program), and the Federal Housing Administration.
[3] Ronald D. Orol and Gregg Robb, “The Regulator Who Could Block Mortgage Refi Plan,” MarketWatch (August 31, 2011).
[4] Glenn Hubbard and Chris Mayer, “How Underwater Mortgages Can Float the Economy,” The New York Times (September 18, 2010).
[5] The failure of pre-payments to rise as would have been expected given the recent decline in yields has resulted in an increase in the value of these securities.
[6] Other proposals would also forgive some principle for underwater mortgagees, with the resulting losses shared by private investors and taxpayers. Given the philosophical and legal issues surrounding such plans, we view them as much less likely to be enacted than plans limited just to refinancing.
[7] Mitchell Remy, Deborah Lucas, and Damien Moore, “An Evaluation of Large-Scale Mortgage Refinancing Programs,” (The Congressional Budget Office, Working Paper 2011-4; September 2011).
[8] And then there is this: at the peak of the housing boom in 2005-06, homeowners’ equity in real estate reached roughly $13 trillion. The rapid pace of refinancing then was accompanied by net mortgage equity withdrawals (MEW) that exceeded $750 billion per year. Our readers know we are not huge fans of the argument that MEW provided a strong independent boost to PCE during that period. However, to the extent it did, it almost certainly won’t be the case this time around. With the sharp drop in house prices since 2006, homeowners’ equity has plunged to only $6 trillion, and measures of MEW have been negative since 2008. We certainly would not expect a “supercharged” (by MEW) effect on spending from the refinancing encouraged by this program.
[9] DeMarco joined the agency in 2006 during the Bush Administration and has served as Acting Director since 2009. The Obama Administration nominated Joseph Smith (banking commissioner of North Carolina), but Smith withdrew his name from consideration in the face of Republican opposition in the Senate.
Contact Macroeconomic Advisers
TweetFollow @macroadvisers
Posts
