three, in many instances and indeed most instances, sales tax, the total tax base for the education in that area.

That is the type of situation I visualized when you said some are underpaid. How does the Administration's proposal, and evidently yours does not differ in this respect from the Administration, how does that completely compensate that school district for the cost of education in that district?

Dr. HOVEY. Assuming we have perfect data, we calculate the added cost for each of those students which we will assume is what the district is spending per pupil from local sources for other students in the district. You multiply what that district is spending per student by the number of new students and that becomes the change in cost. You then look at the change in revenues to the school district. If the change in revenues of the school district by stimulated economic activity is zero, then the appropriate payment would be that cost change.

The absorption provision works in a pure A case, by, in effect, presuming that there will be some impact on revenues associated with having that particular activity that gives rise to those students. It would say in substance that some one and a half percent of the additional cost of those students would not be paid because of absorption. The payment formulas devised by both the Administration and Battelle are not identical for calculating the incremental cost of educating that child in that district a hard number to calculate. I hope that answers the question.

Mr. MEEDS. Yes; I don't know that I quite understand your answer. However, I am sure it is me.

Let's back off. You said will not reach by one and a half percent the cost of educating that child.

Dr. HOVEY. Right. What happens to the district can be best understood from the standpoint of the superintendent. You are a superintendent: you have 48 new children and you write an application for payment for those children, but you have to absorb 3 percent of your nonfederal students. The nonfederal students in that case are 52. So you have to absorb 3 percent of 52, which is about 1.5. The federal government will only pay you on 48 minus 1.5, or 46.5 students.

From your point of view the federal government is not picking up some percentage of their apparent burden. Let's say that the figure is $800 for local costs of 1.5 students. If you have stimulated revenues in excess of $800, then you are all right. You are ahead. If you have no stimulated revenues, you are out that $800.

Mr. MEEDS. So you don't reach full compensation in the illustration which I made because I don't see any, well, perhaps there is some. I don't know how you calculate it. Any increase in economic activity which does pay some tax which is utilized by the local school district. You see, the illustration I used was, one, a military base, two, military people living on that base where they didn't pay income tax, sales tax, or property tax. What other factor are you putting in there or I should say do I understand it to be the other factor you are putting in there is increased economic activity brought about by the location of that military installation?

Dr. HoVEY. The absorption concept, in the context that you raise this question, in substance either presumes that or has another argument which I previously discussed. But looking at that situation, you can go out around bases and have all of the people living on the base and see certain kinds of economic stimulation that tends to add to the tax base. Somebody opens a motel. The bars expand their business, and the convenience grocer does a little more business, that kind of thing, which does have an indirect revenue stimulating effect which is extremely difficult to measure.

Mr. MEEDS. As I understand it, under the Administration proposal, this is equivalent to 40 percent of the cost of educating that child.

Dr. HOVEY. I wouldn't say it quite that way. The district, to take an example, may spend $600 a child, of which it raises $300 locally. If another child comes in, state aid will pick up $300 and the district's additional cost is $300. So you can't focus on the total changing cost of educating the child, but only the change in cost associated with local money.

Mr. MEEDS. You are shifting now the obligation of the federal government to the state. You say the state will pick up $300. Is it proper that the state be required to pick up $300 when it is not receiving anything?

What I am talking about here is the federal obligation to the local school district.

Dr. Hovey. All of the bills before you, including the current legislation, force the states to pick up this added cost. There is a limited exception to this in the Battelle recommendations, but in general the state is forced to pick up its state aid costs for any federal pupil caused by the federal government.

We recommended against your providing full pupil cost, which would require you to make both local payment and state payment, for a variety of reasons. Not the least of these is that the whole impact aid program arises from the balkanization of finance sources in a large number of little school districts. You can have a base over here which may stimulate economic activity over there and if they were all in one school district, you would be all set. But they are not. When you get to the state basis, you have a much different situation. We have about five pages in our report on the notion of paying to the states which has tables which may be of interest to you in that they show some of the conceptual difficulties in paying the states. (The material referred to follows:)

THE ECONOMIC RATIONALE OF PAYMENTS TO STATES

The primary difficulty with the concept of making all impact aid payments to states is that there would be little if any economic rationale for whatever payments were made. From the analysis in the preceding chapter it will be recalled that the appropriate payment to offset the net burden of federal activities would seek to calculate the number of students added to enrollment, multiply that number by the costs of educating each student, and pay the difference between that product and the gain in school revenues resulting from the federal impact. The same difficulties in determining the situation "without" federal activities exist in the case of states as in the case of school districts. However, the inherent overcompensation of an impact aid formula based upon student counts becomes more apparent in the case of states.

In the case of the 3(b) students, a strong argument can be made that most of the civilian employees of the federal government-particularly, the large category of civilian employees working at military bases-would be residents of

the state whether they worked for the federal government or not, in which case their federal employment does not cause additional students to be living in the state. In the case of other 3(b) students and many of the 3(a) students, the asymmetry of the student-counting procedures becomes a particularly significant. If a New York citizen with two school-age children accepts federal employment as a military doctor and is moved to an air base in Texas, the student-count procedure gives Texas credit for having been burdened by the costs of educating his family. However, that same move reduces the educational costs of New York State much more than it is likely to have reduced the tax base in New York. Thus, any program of assistance to states based upon a count of children of federal employees will significantly overestimate the number of students for whom the federal government should assume responsibility. In considering the impact of the federal government on school revenues at a state level, it must be recognized that the situation may well be quite different from the situation considered at the level of a school district. Because of the division of states into many school districts, situations do arise in which the costs of educating federally connected students fall upon one school district, while the economic stimulus of the federal activity takes place in another school district. Under these circumstances, the federal government pays impact aid to the burdened districts, but cannot collect a special tax from the benefited districts. Thus, the program of payments of impact aid are financed from general tax revenues. Viewed at the level of the state, however, it would be incorrect simply to base the state entitlements upon the summation of all the net burdens of federal installations on certain school districts without subtracting the net benefits of federal installations to other school districts in the same state.

Put in more practical terms, it is quite reasonable to find that the federal government has placed a special burden on the school systems of China Lake, California, Norfolk, Virginia, or Bellevue, Nebraska. It is doubtful however that the federal installations in those states have left the states of California, Virginia, and Nebraska worse off economically than they would have been without those installations. While one cannot accurately calculate what the situation would have been in those states if there had never been federal installations in them, it is as reasonable to assume that tax base per pupil in those states has been increased as it would be to assume that tax base per pupil has been decreased by the federal impact.

The State of California, for example, can claim that it should be paid federal impact funds because the federal government has caused an adverse economic impact resulting from federal activities.. On the other hand, the State of Vermont can contend that it should be paid federal aid funds because the federal government takes funds from Vermont residents through the federal tax system and spends those revenues to pay salaries of federal employees in California. This later contention would suggest that, at the state level, an impact aid program might be used to compensate those states that do not have federal installations rather than the states that do.

Impact aid deals with a problem that exists because states are fragmented into small districts for educational purposes and those districts rely primarily upon local tax sources for their educational programs. If the federal government is willing to assume either that states are not thus fragmented or that states rather than the federal government should handle the effects of that fragmentation, then the appropriate impact aid policy is, for most states, to have no federal program at all and let the states raise revenues in districts that benefit from federal installations for the use of those districts that are burdened by those installations.

Another important economic aspect of the federal burden on states stems from the difference in taxing sources between state government and individual school districts. Although there are exceptions, local school finance in the United States is generally based upon taxes upon property. The property tax does not reach the places of employment of federal employees because the federal government is not liable for local property taxes. The property tax does not reach the residences of military personnel and others living on federal property for the same reason. Although there are exceptions, state tax sources generally concentrate upon income and sales taxes on individuals and such special tax sources as excises on liquor and cigarettes, mineral extraction taxes, and taxes upon public utilities. These tax sources are not affected by the tax exemption of federal property. For example, in the case of a civilian federal

employee the liabilities for sales and income taxes are comparable to those of a non-federal employee.

DIRECT PAYMENTS TO STATES VIEWED AS AN EDUCATION PROGRAM

As an alternative to considering payments to states in terms of the economic impact of the federal government on state school revenues and costs, one might consider such payments in terms of the degree to which they would correspond to the distribution that might be made to states under other potential federal education programs or revenue sharing. The data for this analysis are provided by Table 3.1.

TABLE 3.1. SELECTED FINANCIAL DATA OF REPRESENTATIVE STATES

1 Both average daily attendance (ADA) and average daily membership (ADM) are measures of the number of pupils in a district or State.

Source: See text.

This table selects a representative group of states and shows the characteristics of those states that might be relevent to viewing direct payments to states of impact aid funds as a federal education program. The first column shows impact aid entitlements under P.L. 874 for fiscal 1968 as reported in the Annual Report of the Commissioner of Education on the Administration of P.L. 81-874 and P.L. 81-815. Assuming that the same payments would be made to states as were made to the total of districts in the states, the second column shows the approximate amount per pupil that each state would receive.* Direct payments to the states would initiate a program that provided over 30 times as much assistance per pupil to New Mexico as to Vermont and about 15 times as much per pupil to California as to Vermont. It is reasonably clear that the relative educational problems of the two states could not justify this disparity. A program of direct payments to states would not accurately reflect the differences in tax effort (local and state revenue for public schools as a percent of personal income, 1966) as an examination of the third column of Table 3.1 will readily indicate. A similar comment applies in the case of measures of tax capacity such as the one shown in column 4 of Table 3.1. The conclusions to be drawn from this table is that impact aid rewards the states in a very haphazard way-totally unrelated to educational needs or (so argues the preceding pages) to statewide economic impact.

WHAT FORMULA COULD BE USED TO CALCULATE STATE PAYMENTS?

The final column of Table 3.1 shows the relative federal impact for each state, calculated by dividing the sum of fiscal 1968 federally connected pupils (counting the (b) pupils at 50 percent) by the total average daily attendance of each state. If the federal payments were made to states on the same basis as to districts, all states would be eligible because they would meet the present minimum requirement of 400 federally connected pupils. However, if a threshold for eligibility of 3 percent to total ADA were applied to states, some states would not be eligible for assistance at all. Because there is no good economic or educational rationale for distribution of impact aid funds to states rather than districts, there is no clear way to construct a formula for the distribution to states that would appear equitable.

Another problem arises in connection with determining what requirements would be placed on states to "flow through" a certain portion of assistance to heavily impacted districts. If no "flow through" is used at all, the resulting

program may do a very poor job of providing good education for federal pupils. If, on the other hand, a "flow through" provision could be designed that would ensure that the districts that "really needed" impact aid would be provided such aid through the states, then it would appear appropriate for the federal government to use that provision as its impact aid formula for districts and make no payments to the states.

Dr. HOVEY. You would be paying a great deal to California, for example, and nothing to Vermont. Vermont could make the case that indeed the burden is on Vermont because it doesn't have federal installations rather than on California which does, and therefore California should be taxed to pay Vermont rather than people in Vermont being taxed to pay the State of California.

Mr. MEEDS. I don't see that this concept follows your "but for" basic philosophy because the question, it seems to me, is, would the obligation arise but for, pursuing your basic premise, "but for," and once you have said that it would not, "but for," then it seems to me you have the total obligation rather than shifting part of it to the

state.

We at the federal level have this obligation.

Dr. HOVEY. That is correct, but let's keep pursuing our "but for." You have a new student in the State of Washington, and the State of Washington will pick up $300 in state aid cost. The first thing that happens in "but for" is that student came from some place. Let's say he came from the State of Kansas. "But for" the federal activity, you have imposed $300 in cost on the State of Washington, and you saved $300 in the State of Kansas.

Mechanically if you wanted to pursue "but for," you would tax Kansas $300 as the originator of the student if you take the student off Kansas' hands and pay it to Washington.

Mechanically you can't get that done, for the states as a wholesince the federal government doesn't cause any pupils nationwide, it just shifts them around - "but for" for the states as a whole is not a significant problem.

You also have in your "but for" the stimulated revenue.

That is nonetheless an extremely valid question in my opinion. We wrestled with this about six months before we were convinced to come out the way we did in that chapter on payments to the states. The first thought that must come to your mind, once you adopt a "but for" kind of approach, is why not pay the states.

Mr. MEEDS. This then is why you suggested that something ought to be worked out in lieu of the Administration's proposal of not allowing states or allowing states to take into consideration this federal contribution in their equalization formulas?

Dr. HOVEY. That is correct, and it also underlies our comment that if you have a minimum payment that exceeds the burden on the local district because the state is making a large effort, make part of your minimum payment to the state.

Mr. MEEDS. How do you regard Section 13 of the Administration bill? As I recall, you said that you had differences on that? Section 13 is the very heavily impacted district.

Dr. HoVEY. The language in Section 13 of the Administration's bill is to a significant degree the language recommended by the Battelle Report.