|
Article Excerpt
We derive a theoretical model of how jumbo and conforming mortgage rates are determined and how the jumbo-conforming spread might arise. We show that mortgage rates reflect the cost of funding mortgages and that this cost of funding can drive a wedge between jumbo and conforming rates. Further, we show how the jumbo-conforming spread widens when mortgage demand is high or core deposits are not sufficient to fund mortgage demand, and tightens as the mortgage market becomes more liquid and realizes economies of scale. Using Mortgage Interest Rate Survey data for April 1997 through May 2003, we estimate that the government-sponsored enterprise funding advantage accounts for about 7 basis points of the 15-18 basis point jumbo-conforming spread.
**********
Congress created two government-sponsored enterprises (GSEs), the Federal National Mortgage Association (Fannie Mae) and the Federal Home Loan Mortgage Corporation (Freddie Mac), with the goal of providing banks, thrifts and other mortgage originators with a liquid secondary market that would provide an alternative to funding mortgages with deposits. A secondary mortgage market allows mortgage originators to respond more quickly to fluctuating mortgage demand and to lower mortgage rates for some homeowners when mortgage demand is high.
When Congress created the GSE charters, they provided the GSEs with a variety of special benefits. (1) Initially, many viewed these benefits as a way to enhance the GSEs' efforts in establishing a secondary mortgage market. However, with the secondary mortgage market well established and with many other well-functioning purely private secondary markets, the justification for the GSEs' benefits has shifted to the GSEs' success in lowering mortgage rates and in encouraging affordable housing.
The GSEs benefit from the government-sponsored status because purchasers of their debt assume that the government will not allow the GSEs to fail, even though the government has made no explicit promise to bail out the GSEs should problems arise. This ambiguous government relationship creates an implicit government subsidy to the GSEs that is worth billions of dollars (CBO 2001). In this article, we estimate the proportion of this implicit subsidy that GSEs transmit to homeowners via lower mortgage rates.
Like many previous studies, this article focuses on the difference in mortgage rates observed on mortgages that exceed the size limit imposed on GSE mortgage purchases (jumbo mortgages) and mortgages below this size limit. Many previous studies assume that the spread between jumbo and conforming mortgage rates is a measure of the effect of GSEs. This assumption ignores segmentation of the jumbo securitization market as well as the effects of bank funding capacity, banks' investment alternatives and fluctuating mortgage demand on mortgage rates. When one considers a hypothetical world without GSEs, researchers need to consider both the possibility that conforming rates will rise and that jumbo rates will fall. Currently, the jumbo mortgage securitization market is artificially segmented from the conforming market because of the conforming loan limit, and it therefore cannot realize the economies of scale or scope of the conforming mortgage-backed securities (MBS) market. (2) Thus, the jumbo-conforming spread is an upper bound on the extent to which mortgage rates might rise if GSEs lost their special status. Our research suggests that the typical jumbo-conforming spread is between 15 and 18 basis points and that the GSE funding advantage likely accounts for about 7 basis points of this difference.
The Effect of the GSE Implicit Subsidy on Mortgage Rates
Given the many intermediaries between the source of the subsidy (investors who view the GSEs as backed by the government) and the target of the subsidy (homeowners), the GSEs' presence does not necessarily change mortgage rates very much. As argued by Goodman and Passmore (1992) and Hermalin and Jaffee (1996), much of the subsidy may not be transmitted to homeowners. This is because the conforming mortgage market has many of the conditions required for imperfect competition among competitors: two competitors (Fannie Mae and Freddie Mac) with roughly equal market share, homogeneity of product, high entry and exit barriers and each with almost infinite production capacity. (3) In addition, Fannie and Freddie--like all insurers of credit risk--face an adverse selection problem, which requires that they include a "lemons premium" in the purchase price they offer for mortgages. Theoretically, the GSEs' efforts to avoid adverse selection could then completely absorb the subsidy (Passmore and Sparks 1996). This outcome has become more likely with the advent of automated underwriting because mortgage originators can determine, with little cost, whether GSEs will purchase a mortgage (Passmore and Sparks 2000). Finally, because the mortgage originators who are depository institutions always decide first which mortgages to keep and which to sell (a "first mover advantage"), the GSEs--even if they desire to pass on a subsidy to homeowners--might not be able to use the mortgage banking system to actually transmit the subsidy because of the banks' relative advantage in bargaining over pricing and underwriting standards (Heuson, Passmore and Sparks 2001).
Congress created the GSEs with the goal of providing banks with a source of mortgage funding other than deposits. The GSEs enable banks to sell mortgages into a secondary market during times when high mortgage demand and limits on bank capacity for holding additional assets constrain mortgage funding. Putting aside the more complicated arguments about the relative bargaining power of GSEs outlined above, a secondary mortgage market--whether GSEs are involved or not--enables mortgage originators to respond more quickly to fluctuating mortgage demand and lowers mortgage rates for some homeowners when mortgage demand is high. (4) GSEs may or may not provide this "extra capacity" to the banking system at a lower cost than private securitizers.
Mortgage rate studies based on data from the late 1980s and early 1990s generally conclude that mortgage rates for conforming mortgages were about 20-40 basis points less than mortgage rates for jumbo mortgages (Cotterman and Pearce 1996, Hendershott and Shilling 1989). Passmore, Sparks and Ingpen (2002) show that better screening of the data combined with the use of more up-to-date information lowers the estimate of the typical spread between jumbo and conforming mortgage rates to about 20 basis points. McKenzie (2002) provides an extensive survey of this literature and estimates the spread to be 22 basis points over a long horizon (1986-2000) and 19 basis points during a more recent period (1996-2000). Torregrosa (2001) found similar results for this latter period (1995-2000), with estimates ranging from 18 to 25 basis points depending upon the estimation technique and screening of the data. (5)
Ambrose, LaCour-Little and Sanders (2004) conduct a unique study that, unlike many other studies of the effect of GSEs on mortgage rates, does not rely on the Federal Housing Finance Board's Mortgage Interest Rate Survey (MIRS). Using data from an unidentified large national lender, they have much better measures of borrower credit quality and of a mortgage's conforming loan status than studies based on the MIRS data. After taking into account borrower characteristics, house price volatility and endogeneity and sample selection issues, they find that jumbo mortgage rates are about 27 basis points higher than conforming mortgages. Further, they decompose this spread into the conforming-nonconforming (6) spread (9 basis points), the jumbo-nonconforming spread (15 basis points) and house price volatility (3 basis points). They conclude that the GSEs account for between 9 and 24 basis points of the jumbo-conforming spread, depending on whether one perceives the jumbo-conforming spread or the conforming-nonconforming spread as being a result of GSE activities.
GSEs, the Banking Industry and the Jumbo-Conforming Spread
The Demand for Mortgages
Using a simple model of mortgage demand, we illustrate that the jumbo-conforming spread changes with fluctuations in mortgage demand, the attractiveness of banks' alternative nonmortgage investments and the banking industry's capacity to collect core deposits. We assume that households maximize the utility they receive from their allocation of wealth. We use a quadratic utility function so that household utility increases as housing wealth increases until wealth reaches the bliss point, W*. Any increase in housing wealth beyond W* decreases household utility. Given a certain value of housing wealth, w, household utility is then
U(w) = -[1/2](w - W*)[.sup.2]. (1)
At the time a mortgage payment is due, the household must decide whether to pay or to default on the mortgage. In our one-period model, we assume that the household makes a down payment, p, which is a constant percentage of the total mortgage amount, M. Then the household must repay the entire mortgage balance with interest, (1 + [r.sub.M])M, or default on the mortgage. The house is then sold at its realized liquidation value.
The liquidation value of the house (as a percentage of the purchase price), l, has a distribution that is known ex ante to both the borrower and the bank extending the mortgage. The household decides whether to repay the mortgage or default after observing the realized liquidation value. If the amount of the mortgage plus interest, (1 + [r.sub.M])M, less the default cost, [c.sub.def] M, is greater than the liquidation value of the house, lM, then the household defaults. Otherwise, the household repays. The default cost encompasses the stigma that the household incurs from defaulting on such a large debt. This stigma mainly takes the form of a lower credit rating.
Given the distribution of house liquidation values, f(l), repayments and defaults have known probabilities. The household repays with probability q and therefore defaults with probability (1 - q), where
q = [[integral].sub.1+[r.sub.M]-[c.sub.def].sup.[infinity]]f(l)dl. (2)
Housing wealth depends on whether the household repays the mortgage. If the household does repay, housing wealth is the difference between the realized liquidation value and the total amount paid on the house:
[w.sup.R] = ([[l - p -[r.sub.t]p]/[1 - p]] - (1 + [r.sub.M]))M, (3)
where [r.sub.t] is the risk-free Treasury rate. Here, the first expression is the gross proceeds from selling the house less the down payment and the opportunity cost of the down payment. Housing wealth under default is the amount lost from the down payment plus the cost of default:
[w.sup.D] = ([[-p - [r.sub.t]p]/[1 - p]] - [c.sub.def])M. (4)
Expected household utility is therefore
E[U(w)] = qU([w.sup.R]) + (1 - q)U([w.sup.D]). (5)
The household maximizes expected utility with respect to mortgage size, resulting in the first-order condition
[dE[U(w)]]/[dM] = q[([[E[l] - p - [r.sub.t]p]/[1 - p]] - (1 + [r.sub.M]))M - W*] X ([[E[l] - p - [r.sub.t]p]/[1 - p]] - (1 + [r.sub.M])) + (1 - q) X [([[-p - [r.sub.t]p]/[1 - p]] - [c.sub.def])M - W*]([[-p - [r.sub.t]p]/[1 - p]] - [c.sub.def]) = 0. (6)
This implies that the household's optimal mortgage size is
M* = [[q([[E[l] - p - [r.sub.t]p]/[1 - p]] - (1 + [r.sub.M])) + (1 - q)([[-p - [r.sub.t]p]/[1 - p]] - [c.sub.def])]/[q([[E[l] - p - [r.sub.t]p]/[1 - p]] - (1 + [r.sub.M]))[.sup.2] + (1 - q)([[-p - [r.sub.t]p]/[1 - p]] - [c.sub.def])[.sup.2]]]W*. (7)
Under this formulation, mortgage demand depends on optimal housing wealth (W*), the risk-free interest rate ([r.sub.t]), the down payment ratio (p), the cost of default ([c.sub.def]), expected house prices (E[l]), the probability of default (q) and the mortgage rate ([r.sub.M]).
The Bank Cost Function and the Supply of Mortgages
Turning now to the supply side of the market, we differentiate between two types of financial institutions that issue mortgages: commercial banks and mortgage bankers. Commercial banks typically fund mortgages using deposits (and are therefore required to build the "bricks and mortar" associated with raising core deposits). Primary mortgage originators, however, sell their mortgages shortly after origination. Both entities can sell mortgages they originate in the secondary mortgage market, but each entity has very different cost structures, affecting the mortgage supply function.
To collect a given amount of deposits, banks must raise capital, K, to invest in bricks and mortar. Banks pay the market rate on equity, [r.sub.e], for this capital. Deposits...
|
