Hi Arnie,

We are currently doing an in-depth analysis of the hemscott data and how it stacks up to the data provided by OPS (S&P). While we have not concluded the study (i'd imagine it will be an ongoing thing, even after our initial analysis is complete), we have come across similar issues in regard to the reporting of data for financial institutions. I asked Ellis Traub, an expert on data matters, about your question and he has given the following in reply:

"In a word, Hemscott has it right; OPS/SDS does not.

The "correct" data for bank and many other financial institutions which
lend their money is the sum of the Net Interest Income and Non-Interest
Income less the value of the Loan Loss Provision. The first item is the
crucial one because, whether the interest rates go up or down, the net
interest income remains pretty much the same since the money the bank
borrows is subject to the same interest increase or decrease as is the
money then lend. This keeps things on a reasonably even keel over time.

There is a also a Tax Equivalent Adjustment which means simply that any
interest accruing from tax-exempt securities will be adjusted to reflect a
value were that income to be taxed. However, that's not available in any
of the data except what Compustat (S&P) provides in their line item, "niit."

For that reason, you can see that those who use only the gross interest
income (without subtracting the interest expense to arrive at the net)
will have considerable distortion as the interest rates go up and down.

Value Line never had it right. You had to dig up the answer and do the
math.

OPS/SDS started by using the gross interest income; but now, they don't
even do that. Their explanation of what this item includes for banks reads
as follows:

niit + (sales - tii) -(cogs-tie)

niit: Net Interest Income, Tax Equivalent - This item represents net
interest income with the interest income on non-taxable securities
adjusted to a taxable equivalent amount.

sales: Sales/Turnover - For banks, this item includes total current operating
revenue and net pretax profit or loss on securities sold or
redeemed.

tii: Total Interest Income - This item represents the revenue received
from all earning assets.

cogs: Cost of Goods Sold - This item includes all of the things that are
customarily considered COGS. But Compustat then says that this
information is "not available for banks." However, you will find this
populated as a line item when they display their expanded data.

tie: Total Interest Expense - This item represents the interest expense
on deposits, long-term debt, and all other borrowings.

It's easy to see the logic behind this definition. If you start with the Net
Interest Income (as Hemscott does), then you should be able to calculate
the "Non-Interest Income" by subtracting the Interest from the Sales (since
it would have to include it) and from that figure subtract the expenses,
which they cobble together by deducting the cost of producing that income
and adding back the Interest Expense which "Net Interest Income" should
have already taken into account.

This is how they describe the calculation of revenue for banks. However,
when you look closely at the numbers, you will find that the values found in
the OPS/SDS data represents only the value for "sales" as defined above.
None of the other items are considered, even though it's advertised that
way!

If you look at WIBC, for example, there is NO value in the expanded data
for the "niit." Yet this should be the basis for the bank data.

So it appears that they just don't have it and have given up trying to put
it together as they have advertised they do."

Posted w/ permission. We hope this helps, if questions please post them here, thanks!

Dave Forgianni
Product Manager
ICLUBcentral Inc.

Thanks for this information, Dave. What is WIBC?

david parker