Now, if I may, I would like to make reference to table 4, because table 4 is a more detailed breakdown of the differences that occurred between previous and adjusted.

If I may, Senators, I would like to mention at this time that of necessity. I'm talking about summary data. I've got two brief cases here that are full of all the monthly data, so if you should desire to pick some number out of the hat and let's see what happened on the account, I've got it, if you want to look at it. But I'm interested in looking at summary data.

So in table 4 what I've shown here are the results of the 865 accounts, and what the difference in monthly finance charges was between previous and adjusted.

As indicated earlier, the average was 15 cents a month, the median was 4 cents; that is for half the accounts, it amounted to 4 cents a month less, but this shows you exactly how much difference it was. This is the whole thing.

If you will look, for example, close to 40 percent of the accounts, there would have been no difference at all.

In other words, on 342 accounts previous and adjusted would have cost identical.

Now, that "identical" is defined in this case as being less than a penny. In other words, there might be somewhere was a .001, or something like that. But, anyway, 40 percent, no difference.

Another 21 percent, there was a difference, of course, being less expensive, but by no more than 10 cents a month. What you have got here is 60 percent of the cases there would have been no difference, or no more than $1.20 a year difference.

Obviously, the differences go up, and for some customers, it amounted to $1 a month for four of them out of 865, it amounted to more than $12 a year. The biggest difference that I found was $1.86 a month.

Now, in 93 percent of the instances, the difference amounted to no more than 50 cents a month or $6 a year. So the conclusion that I reach on this is that although the hypothetical examples often show tremendous differences, I don't think the hypothetical examples are anywhere near realistic of actually what happens. I think this data illustrates that.

Now, on table 5, is the same type of table except it compares previous balance to one of the ADB methods, that is, the one that includes debits.

Now, in this particular case again, 43 percent of the accounts roughly-let me move this chart over. The reporter might want to look at it. Previous versus ADB including-43 percent of the instances, again there was no difference. Now, this is one of those methods where if a store uses the ADB method including debits, some people gain under one and some people gain in the other. But in 43 percent of the instances, there was no difference. In 87 percent of the cases, these two methods resulted in either zero difference or no more than 10 cents one way or the other.

Now, actually, the previous balance method cost more, a slightly larger percentage of the time. If you look down in this chart there were 268 positive values out of the 865, which means previous cost more. There were 228 negative values, which means previous cost less.

So the tendency was for previous to cost more, and this is a greater percentage of the time, but the mean difference again between these methods is zero. On the average there was no difference.

Now, the other thing that is interesting about this, though, is that when the average daily balance method, with debits, when it did cost more, it cost more than where the previous cost more. If that makes sense. It probably doesn't. What I mean by that is if you concentrate on the positive values, which means where previous cost more, it average costing 6 cents a month more when it did. If you concentrate on the negative values, it cost more by 9 cents on the average. When ADB does cost more, it cost more than the previous does.

But on the average, there was no difference, and 87 percent of the differences, the average was zero, or plus or minus 10 cents a month or less.

Table 6 compares previous with another type of average daily balance, and that is one where you do not include the debits. I refer to it as ADB excluding debits. Again in 42 percent of the instances, no difference, another 48 percent a difference of between a penny and 10 cents a month. So in something like, what, 78 or 79 percent of the instances, there was no difference, or a savings of 10 cents a month or less.

It has been claimed by some, Senators, that this method actually saves nobody any money. Well, it actually does save 60 percent of the customers money.

Now, it has also been pointed out by Senator Proxmire himself, I think, that the differences are not striking, and I would agree with that. But the point I would make is that the differences are not striking no matter what billing methods you are talking about, unless you are comparing maybe ending or true actuarial with the others. But the differences illustrated here are not striking. In most cases there is 10 cents a month or less, but so is the same analysis true regardless of any combination you might want to pick.

But the point is that they paid no more under this method, and 60 percent of them paid less under this method, by an average of 6 cents

a month.

All right; one more table of this type, and that is table 7, which compares previous with the true actuarial average daily balance. In this particular instance, in a majority of the cases, the previous balance method was less expensive.

That is, in 88 percent of the cases, or more than that, in about 90 percent of the cases, previous balance cost less than true actuarial average daily balance. Of course, the median difference in this case was 23 cents a month in comparison with the previous balance. There were some accounts, 93, where a true actuarial daily balance method would have cost less money, and when it did, it would have cost less by an average of 7 cents a month.

Now, where previous was more expensive, though, out of the 93, 42 of them, it was not more than 5 cents a month more costly, which would be 60 cents a year.

So the true actuarial method is not the most expensive method for everybody, but it is by far for the vast majority of customers.

Senator PROXMIRE. Could you bring your remarks to a conclusion in the next minute or two.

Mr. MCALISTER. Let me comment about the impact of the store. My thesis is that to the customer, the impact is minimal, amounting to no more than $2 a year. I have read the $200 million savings figure in several places, and I doubt seriously that it is a realistic figure, because if we assume that the customer saves $2 a year by switching from previous to adjusted, there would have to be a lot of accounts for that to be so.

But at any rate, the impact on the store is substantial, and so I have tried to explain in table 8, which is on page 26, I've tried to show what the yield to the stores are under the six billing methods.

Under the previous method, which is the method actually used, the yield was 15.9 percent, under the adjusted 14 percent and so on. Now, the index shows you the relationship between the two. If a store moves from previous to adjusted, he will lose 12 percent of the revenue, which is a pretty substantial drop. Ending balance would have produced 14 percent more. The average daily including debits, approximately the same revenue, actually a little less than a half percent more, an average daily excluding debts, which is the method that Sears is now using in, I think, 45 States, results in 5 percent less revenue, and the true actuarial, of course, would be resulting in more revenue, 21 percent more, roughly.

So the impact of the store is substantial, although the impact of the customer is not that substantial.

Let me make just a couple of concluding remarks, and then I'll quit. It has been claimed also that one of the advantages of standard billing methods would be that you could compare the cost more easily. I don't believe this to be a fact. Even if the billing methods were the same, and even if the rates were the same, you could still not predict the cost to the customer because you would have to have the same purchase and payment patterns, and the same repayment schedule.

Now, one last thing. On the impact of the 50-cent minimum, table 9 shows a simulation of what the actual finance charges would have been had these accounts not incurred a 50-cent minimum. As you recall earlier, under the previous balance method, with the minimum, they paid $1.24. Without the minimum, simulating what it would have been without it, they would have paid $1.23. So they would have saved a penny a month, or 12 cents a year, which may buy a pack of gum. I'm not sure these days. But 12 cents a year is about all they would have saved.

Under some billing methods they would have saved more. So the impact of the billing method is a bogus issue. I don't think it has that much impact.

Now, my personal feeling based on having worked on this for 18 months, is that sections 167 and 168 are really not that much of a concern to the customer, and the impact on the store is substantial, but the average customer-me, for example, I might save a couple of dollars a year, maybe-but even in that instance, if the loss of the revenue to the store forces them to have to raise their prices to offset it, then I wouldn't experience that.

[Mr. McAlister submitted the following for the record :]

AVERAGE TRUE ACTUARIAL ANNUAL PERCENTAGE RATE, 6 BILLING METHODS, 865 ACCOUNTS

1 These figures for this billing method are not very meaningful since imposition of a 50 cent minimum charge once or twice during a year on accounts with very small average monthly balances ($10 or less) produces extremely high average rates of charge (over 200 percent in some instances) even though the dollar cost for the year may not be more than $0.50 to $1 or so.

Source: Study of 865 Sears SRC accounts in Texas by Ray McAlister, NTSU, Denton, Tex.

NUMBER OF ACCOUNTS EXCEEDING TRUE ACTUARIAL AVERAGE DAILY BALANCE BY BILLING METHOD-865

Note: It is possible that other accounts could have exceeded an annual rate of 18 percent under any of the billing methods, including adjusted-but if this were true, it would be a result of the 50 cent minimum charge-not the billing method used.

Mr. MCALISTER. I'll be glad to answer whatever questions you might have.

Senator PROXMIRE. Thank you, both of you gentlemen, very much. You've obviously done a whale of a lot of work, and good work, and helpful work, and it is most useful to the committee to have this kind of as you say, Mr. McAlister-empirical work done. You're not relying on theories or impressions or sentiments.

Mr. Dunkelberg, I'm not at all sure your conclusions coincide with your data.

For example, you say the results of your data, and I quote, suggest that the heaviest users of the revolving option have incomes between $10.000 and $20,000. However, your data does not seem to suggest this.

Table 1 shows that the heaviest users of the revolving option are in the $7,500 to $10,000 income bracket. Over 80 percent of the sales of this group were revolving, whereas only 64 percent of the sales to $10,000 to $20,000 income group were revolving.

How do you explain this apparent discrepancy between your statement and your data?

Mr. DUNKELBERG. First, let me recall that during my oral testimony, I said that that break wasn't quite as clean as the $10,000 to $20,000 that I wrote here, that indeed we probably should include some lower income families in there too. What I really need is more detailed income breaks to get the sort out a little cleaner.

But let's take a look at the data you point to. It is true that in table 1, row 3, we find that the groups that have the highest percent of revolving sales are indeed the 75 to 10 with 80 percent, the 10 to 15 with 64, and the 15 to 20 with 64. And that is very different from the under $7,500 and over $20,000. OK.

Now, there are some other measures that you ought to take a look at. Those are the shares of finance charge revenues compared to the distribution of consumers in column 1, row 1. These people (the $10,000 to $20,000 income) do tend to pay a larger share of finance charges, taken together, than the $7,500 to $10,000 group, compared to their sales shares. If we really cut it fine, the major group paying a disproportionate share is the $15,000 to $20,000.

So it is true that the $7,500 to $10,000 income consumers behave pretty much the same as the $10,000 to $20,000, except for the shares of finance charge revenue. I don't really see that that changes the conclusion too much except that we now have even the much lower income people behaving much more like the very high-income people, still receiving the subsidy from a "free ride," as the under 75 group is even much more like the $20,000-plus income consumers, in that they essentially have the same number of revolving months, and so forth.

If you average those numbers out, they essentially have similar relative shares of revenues, and they are disproportionately low in their shares of finance charge revenues, and also disproportionately low in terms of the percent of sales that were revolving.

So I really don't see that it changes the conclusion very much. The low income and the higher income people do indeed behave very much the same, and other things equal, they receive a subsidy as they certainly are the free riders.

Senator PROXMIRE. I'm not so sure that your data indicates that the low-income people are subsidized at the expense of the higher income people.

Let me go on and get your comment.

You argued the highest proportion of free riders is in both the very low-income groups and the very high-income groups in between paying a higher proportion in the share of finance charges. The implication is left that the subsidy element is just about washed out, that with lowincome groups and low-however, low-income groups in your sample, those were incomes below $7,500, only account for 4 percent of net sales. They tend to be free riders because their purchases are smaller. Middle income account for 58 percent of net sales, but paid 71 percent, 71 percent of all finance charges, by way of contrast.

Upper income groups, with incomes over 20,000, accounted for 33.5 percent, about a third of sales, but paid only 23 percent, less than one-fourth of all finance charges.

In terms of a rate, middle income consumers paid finance charges of about 11 percent, whereas the comparable charges of upper income consumers, is only about 8 percent.

Don't these figures taken from your own data confirm the notion that upper income consumers enjoy a special benefit from revolving credit plans?

Mr. DUNKELBERG. It may well suggest that they enjoy a slightly higher benefit, but nonetheless, I think the case is still clear, that people who are in this middle income category, in terms of the generation of revenues to cover these costs, are subsidizing the low- and high-income credit users.

The unfortunate thing here is that these are averages for the month, and that is one of the problems we hope to correct in the New York study.