there will be continual pressures on the Congress and/or the enforcement agency to again expand the privileged sectors. Drawing a lesson from the histories of a number of regulatory agencies, a large majority of economists would not be optimistic about the possibility of maintaining distinctions as between privileged and less privileged sectors. Another frequent objection to schemes of credit allocation is that they are apt to interfere with the efficiency of the economy overall and the financial sector in particular. How much weight individual economists put on this point depends for the most part on how much they believe that markets work like textbook competitive markets. With competitive markets funds go to borrowers who are willing to pay the highest price, borrowers who by definition can use the funds most productively. If financial markets are reasonably competitive and credit allocation schemes are effective, they must reallocate funds away from more productive users to less productive users and hence the efficiency of the economy as a whole suffers. Similarly, financial institutions develop expertise in the types of lending that they can do most efficiently. If credit allocation schemes are effective they will force some or all institutions to engage in types of lending they would not otherwise engage in, types of lending they are less efficient at. Obviously how much weight one places on these arguments depends on one's prior conception of how competitive and efficient existing markets are. I want to make two points concerning this question. One, economists probably believe that markets are more competitive and efficient than do non-economists. Two, financial markets are severely limited and constrained by government regulation, and thus may not closely approximate the textbook notion of competitive markets. Some forms of regulation and control exist to give the government control over the money supply and to insure the stability of, and hence confidence in, the financial structure. However, in the view of many economists other regulations have the effect of restricting competition and exacerbating the problems credit allocation are designed to help. If this view is correct, then it is not at all obvious that adding new controls-credit allocation-to undo the effects of old regulations is the preferred course for public policy. One could simply eliminate the original outdated controls. 4. VARIABLE RATE MORTGAGES A number of observers have suggested that variable rate mortgages (VRM's) should be considered as a possible solution to the problems of thrift institutions.19 There are perhaps as many VRM schemes as there are proponents of the idea. To attempt a systematic discussion of all the proposals is simply infeasible. The common underlying feature is that the interest rate paid by a borrower would not be constant 19 Variable rate mortgages are only one of several possible variations in traditional mortgages that one might want to consider. In particular, there has been discussion of price level adjusted mortgages and graduated payment mortgages. Time and circumstances restrict the discussion here to variable rate mortgages. For a more complete discussion of other alternatives see Modigliani and Lessard, New Mortgage Designs . . .and Kearl's study of housing and mortgage innovation. over the life of the loan, rather it would vary in response to changes in some reference rate agreed to when the loan was originated.20 The basic idea, common to all the proposals, is to make the earnings from a mortgage portfolio more responsive to changes in short terin interest rates. If these earnings were more responsive to movements in short term rates then thrift institutions could afford to pay deposit rates that were competitive with short term market rates. In such a world it would be expected that institutions could do without deposit rate controls and that when short term rates rose they would not suffer massive deposit outflows. Thus mortgage lending would be stabilized over a cycle in interest rates. Several things should be pointed out. To the extent that long term rates, like mortgage rates, are correct predictors of the sequence of future short term rates, then VRM's tied to short rates would work only to change the time sequence of interest and principal payments. After a mortgage was paid back, the VRM or a traditional mortgage would, in retrospect, look the same to both the borrower and the lender. However, the VRM does introduce strong elements of uncertainty. Viewed before the fact. one simply does not know whether the unknown sequence of rates on a VRM will in the end be eg ivalent to the fixed rate on a traditional mortgage. Further, it may be of email consolation to some househoids that especially high mortgage paymer ta will, sometime in the future, be offset by especially low payments. The VPM does make a real difference, beyond snifta in the payment stream, when the sequence of short rates does not, in fact, for out to be ezivalent to the initial long term rate. In these cases there is currently an octions asymmetry depending on whether future short rates to cut to be lower or higher than implied in initial long term rates. If short term rates turn out to be lower than expected, then a VRM bermatere pazmenta will also be that much lower. With the TED DIRA riment one would expect that, when expertations were listed to reflect the realizations of lower than initially expected short term rates, long term rates, including the mortgags Ne vidi jele. Asuming the deeline to be of suficient mag..ide mortgage borrowers to refinance. It and secrets of mort, rates burns out to be higher than initially 2ft make a real diference of sucstantial impact. dese ames that thrift institutions get into trouble med to help them out. The VRM helps them out payments from borrowers. While these larger payServer vond heip lenders, they obviously are not so var mi, 10t mrprisingly, have been resistert for this rent nor gage instrument. a borrower's payments mmarize, when future short term rates are expected, a VRM mandates lower payments Lutz Ima vanud be expected to lower payments with the riment. If future short term rates are higher rns of the reference nie proposed the adinstment of apference are hanges-toes the size of the payment change arhimitation on the mulative magnitude of poesthie hanges, and other arenarts and onions that would de avullmentioneri that the reference are deed it * 1 dingle Teraze of several mies than originally expected, a VRM mandates higher payments while the current mortgage instrument leaves payments unchanged. Borrowers obviously lose under a VRM with unexpected increases in short term rates as compared with the alternative of a traditional mortgage with a fixed interest rate and a fixed stream of payments. However, under some circumstances, borrowers with traditional mortgages are big gainers. One could view VRM's as requiring borrowers to share some of their current big gains. Specifically, if short term interest rates rise unexpectedly because of unexpected inflation, then borrowers would be expected to have capital gains on their property and higher normal incomes as a result of the inflation. With traditional mortgages the capital gain accrues entirely to the borrower. The VRM would require the borrower to share part of the capital gain with the lender. The sharing analogy can perhaps be carried too far. The VRM requires payment to the lender when interest rates rise, that is when the capital gain takes place. The borrower would probably not realize that capital gain for some time. The borrower is thus forced to finance the early payment of part of the as yet uncertain capital gain to the lender. There may also be significant distribution effects. Interest rates respond to inflation in the economy as a whole, i.e., the average of price increases over all sorts of goods in all sorts of places. It is possible that houses in general, or particular houses in particular places, would not experience capital gains that matched economy wide measures of inflation. There are other circumstances where, when making higher payments under a VRM, borrowers would not simultaneously be receiving other benefits that might finance these payments. If the unexpected increase in short-term rates is in response to a change in real interest rates, not changes in the rate of inflation, then borrowers would not be expected to have capital gains with which to finance higher mortgage payments. What impacts might VRM's have on thrift institutions and housing? As with the general question of the impact of financial reorganization, this question can not be answered with certainty. The wide scale introduction of VRM's could change borrower and lender behavior in ways that would be exceedingly difficult to capture with existing simulation models. Further, the variety of VRM proposals means that even if we had the perfect simulation model, there might not be one answer to the question; the answer could well differ from proposal to proposal. With these caveats in mind, mention might be made of some simulation work done by Professor James Kearl. Kearl simulates the complete adoption of VRM's over the period 1962 through 1973 and finds that the introduction of VRM's, accompanied with the elimination of deposit rate ceilings, does three things: (1) the housing stock is smaller, (2) profits of savings and loans, as measured by their reserves, are larger, and (3) the size of savings and loans, as measured by their total deposits, is smaller.21 Kearl attributes these results to the implicit change in the role of savings and loan associations. With earnings tied to short term rates, 21 Kearl's simulations involve the extreme assumption that all mortgages are converted to VRM's. In such situation there is a less pressing case for broader investment powers for thrift institutions. If VRM's formed only a small part of the mortgage portfolio of thrift institutions, there would still be a strong case for wider investment powers. savings and loan associations are no longer intermediating between long, and short term interest rates and hence no longer earning the profits that come from taking that risk. These lower profits on average mean lower deposit rates on average, a smaller industry, less mortgage lending, a higher mortgage rate and hence a smaller housing stock. În particular most of these effects appear to come in the last part of Kearl's simulation. In particular, the variable rate on mortgages and, with a lag, the deposit rate paid by savings and loans drops from mid 1969 through 1972. This lower deposit rate means fewer deposits, less mortgage lending, etc. The role of reserves at savings and loan associations in the model used by Kearl is difficult to interpret. Kearl's particular results about the growth in reserves needs to be interpreted with care. Within the model transfers to reserves are, for the most part, measured as the difference between mortgage interest income and deposit interest expenses. There is no feedback from large reserve accumulations to higher deposit rates, a feedback one would expect in the real world. To the extent that savings and loan associations were able to use their higher reserves to increase deposit rates, one would not expect as large a reduction in either the size of savings and loan associations or the size of the housing stock as implied by Kearl's simulations. In spite of this limitation of the simulations, Kearl's results raise some fundamental questions as to whether VRM's could be expected, by themselves, to solve the problems of thrift institutions. One can have serious doubts that the conversion of all mortgages to VRM's would solve the problems of thrift institutions. One could also oppose the conversion of all mortgages to VRM's on other grounds. However, neither of these two positions need imply that one should also be opposed to allowing the introduction of and experimentation with VRM's. As implied in the discussion above, the VRM shifts the risks involved in interest rate changes as between borrowers and lenders. In a world where both types of loans existed, there could well be borrowers and lenders who, at times, would find VRM's mutually beneficial. To prohibit the use of VRM's would, in these cases, relegate both borrower and lender to less preferred options. The basic principle here is that, with adequate safeguards and information, more choice is better than less choice. 5. TERM STRUCTURE INSURANCE Some observers have suggested that the federal government should offer interest rate insurance to mortgage lenders.22 Under such a program, the federal government would guarantee to lenders that if some reference interest rate, which might be a single interest rate or an average of several rates, rose above its value when the mortgage loan was made, then the federal government would pay to the lender the increment in the reference interest rate times the outstanding principal of the mortgage. Thus if interest rates rose, lenders with this rate insurance would see their income keep pace with the rise in interest rates instead of remaining unchanged. Note that mortgage payments by homeowners would remain unchanged. The purpose of this proposal is to solve the problem of the impacts of adverse move This proposal is discussed in more detail by James Pierce. ments in the term structure of interest rates on thrift institutions. In this regard the term structure insurance is very similar to the VRM, however there is a big difference in who is paying. What benefits would one expect in this world where the earnings of mortgage lenders were protected against adverse movements in the term structure? With the simultaneous abolition of deposit rate ceilings, one would expect that the forces of competition would pass these higher earnings on to depositors in the form of higher deposit rates. These higher deposit rates would enable thrift institutions to remain competitive with direct market securities. To the extent that these higher deposit rates prevented massive disintermediation then cycles in homebuilding would be moderated. It should be noted that the proposal makes little sense without the abolition of deposit rate ceilings. What happens when interest rates, including mortgage rates, decline? Exactly what would happen would depend upon the form of term structure insurance. One version of this scheme would have mortgage lenders swap the income from their mortgages for a flow of income determined by the reference interest rate. Under this version income to mortgage lenders would decline when interest rates in general declined. Again homeowners payments would be unchanged at a level determined by the original lending rate. In such a situation one would expect that institutions would compete for making mortgage loans by offering automatic rolling over of mortgage loans when interest rates decline. That is the institutions would treat homeowners as if they had paid off their existing high rate mortgage and had taken out a new mortgage that reflected current, lower interest rates. With this sort of response the government would cease to get the original higher mortgage payments and would in essence be offering insurance only against increases in interest rates. An alternative version of the interest rate insurance proposal would recognize this onesidedness and simply offer insurance against an increase in in terest rates. This term structure insurance solves the fundamental problem of the impact of adverse shifts in the term structure of interest rates. One could imagine offering such insurance only to institutions that invested a substantial portion of their assets in mortgages. That is the insurance could be offered only to thrift institutions. In this case the proposal would achieve a change in substance without a change in form. There would be no need to allow thrift institutions broader asset or liability powers because the term structure insurance would achieve the same result. Alternatively one could imagine allowing broader powers to thrifts and offering the insurance to any and all mortgage lenders. What would such a proposal do to mortgage rates and what would it cost the government? The impact on mortgage rates and the cost to government both appear to depend on how the term structure insurance would be implemented. In particular, would the insurance be free? If not, how would a price be determined? If the insurance were offered for free, I would expect new mortgage rates to look very much like short-term interest rates. If institutions did not need to worry about the risk of interest rates rising and if they could attract large amounts of funds at rates that were competitive to other short |