rtainly gives all ratios of insurance expense lower than
they would appear if I had known, and used, the exact actual premium
reserve on death claims, and all probably bear nearly the same ratio to
each other as they would in that case.

As the object of this statement is to explain the new method, and not to
defend my particular estimates in applying it, I forbear to state on what
rules I have made them. Expense which is not ascribed to insurance must be
ascribed to investment, and as in comparing any two companies, their two
ratios of one kind or the other must be equated, to decide the question of
economy between them, it may well be left to any company to say what the
fair division of its own expenses is.

Moreover, there can be but little motive to make a false division; for to
successfully compete for business, a company having large investments has
as much need to show a high net rate of interest earned as a low rate of
insurance expense. Again, it is not my purpose to pass judgment on the
economy or extravagance of any ratio of expense shown in the table. It is
not a fact exhibited for the first time by my figures, that the ratios of
some companies are more than double those of others. The same fact would
be displayed in about as high a degree by ratios based on premium income,
or any other incorrect basis. Custom, the balance of opinions, and
competition may well be left to decide what ratios of expense are high,
and what are average, or low. And their decision is to be gathered only
from _statistics_.

What I do claim is that the mode of determining ratios herein explained is
the only intelligible and scientific one, and the only one proper to
employ in _statistical tabulations_ and _investigations_.

As such, it calls attention to the fact that the amount of insurance
claims met, and of interest receipts, _are limits_ which the corresponding
expenses cannot exceed, certainly for a series of years together, without
making the _expense_ more than the _advantage_ of the business. To keep
this fact in view, _as a preventive of extravagance_, is not the least
valuable service the new mode may render. It may be seen that there are
eight cases in the table, in which the ratio of insurance expense points
to expenses exceeding the insurance claims met in the same time, yet the
reader need not hasten to conclude that the same companies will
permanently show similar ratios, or have no good reasons to give for the
one