government is paying more than is necessary, students are still not assured of being able to obtain the loans they

need.

Basically, the system costs are now equal to the government's cost of borrowing all the capital loaned to students, plus 3.5 percent a year on that capital, plus losses from defaults.

Since the government's borrowing costs

per dollar will not be significantly affected no matter how the student loan programs are structured, the only parts of the system costs worth scrutinizing for possible savings are the 3.5 percent allowance and the default costs.

The real purpose of the 3.5 percent allowance is to reimburse GSL lenders for their origination and servicing costs and to allow a reasonable profit on their activity. But some portion of the allowance, although the amounts are still probably small in a relative sense, is going to finance state arbitrage gains. Even if that issue is ignored, there still remains the question of whether the government could obtain student loan origination and servicing for an annual cost of less than 3.5 percent of the capital in the system.

Our preliminary conclusion is that origination and servicing of the GSL program can be obtained for substantially less than 3.5 percent---that is to say, for less than the

$280 million a year now paid on that account.

Considering origination costs first, they now include

But

for every guaranteed loan the participation of both a lending institution and an educational institution. since the government is paying all the costs of putting up the capital, the role of the bank in such a loan is actually superfluous. The loan can just as well be originated by the educational institution alone.

Secondly, it is demonstrable that servicing costs are well below 3.5 percent of the capital being serviced. Sallie Mae has servicing contracts with four banking institutions and one state agency. The costs range from about $11 per year for borrowers who are still in school to a high of $27 per year for borrowers in repayment status.

If the government were the owner of the student notes,

it too could contract out for loan servicing, presumably on terms at least as favorable as those negotiated by Sallie Mae. In short, it would pay $11 to $27 a year instead of $70 a year on the average guaranteed student loan now being made.

The remaining element of the current system cost is the loss from defaults. The long-run costs involved are difficult to estimate with any precision. While claims paid on

defaulted GSLs are a known amount---about $200 million last year---it is impossible to state how much the government will eventually collect from the debtors whose defaults gave rise to the claims, just as it cannot yet be known how much the government will realize from its current effort to obtain payment from NDSL defaulters. But it does seem safe to predict that default losses will decline in proportion to the amount of effort the government---with its superior ability to locate defaulters---puts into the task. Whether there will be a net budget gain from the government's efforts remains to be seen. But so long as the collection responsibility is spread among thousands upon thousands of educational and commercial lending institutions, the default rate is likely to remain at a level most observers would find unsatisfactory.

All the foregoing suggests that it would be useful for the Congress to consider reshaping the system to acknowledge the reality of the government's present role in raising the necessary capital. One possibility would be to federalize Sallie Mae, to have it be the conduit by which the funds would flow to state lending agencies, who would be the retailers, so to speak, in dealing with the educational

institutions that would orginate all student loans. There

are many other ways in which the system could be restructured to the same end.

I will conclude this statement with a brief discussion

of the borrower subsidies in the present student loan systems. As I mentioned earlier, there does not appear to be a sound analytical basis for continuing to have one program with a 3 percent charge and another with 7 percent. Indeed, the differential introduces questions both of fairness and practicality, as well as of cost. The fairness issue arises because the effect of the differential is that students in like circumstances are treated in an unlike fashion. The practicality issue comes about because, with two different interest rates, it is virtually impossible to consolidate loans when---as is often the case---a borrower has both a GSL and an NDSL when he reaches repayment status. Without suggesting what the interest rate ought to be, whether 7 percent or something else, there ought to be just one basic rate.

My second comment grows out of last fall's legislation that extended the in-school interest subsidy to every GSL borrower, irrespective of family income. We cannot yet know

the costs and program implications of that change, but they may turn out to be severe, with costs rising when loans

are readily available, and lower income borrowers crowded out when loans are rationed. It is probable that a great many families with liquidity problems in financing their childrens' college costs, but who are fully able and willing to pay, say 7 percent, for the privilege of stretching those costs over time, will arrange to have their children take out the interest free loans. The Congress might well consider adopting an explicit program for such families, with the interest subsidized for those who fell below a stated income level, or who have already qualified for any other need-based aid, and 7 percent loans for those who choose not to supply family income information.

Finally, I want to re-emphasize that the short-run outlook for the GSL program is extremely clouded. When the impact of the special allowance ceiling occurs a month or so from now, you may be asked to turn aside from your deliberations about the long-run shape of student loan programs and do something quickly about the short-run shape.

Mr. Chairman, that concludes my prepared statement.