Chapter Three

Ratemaking

Our first and primary objective in Underwriting Analysis
was to calculate automobile insurance rates that would
reflect the company's own claims results. In preparation,
I studied every IBC example in our files and queried the chiefs
in other underwriting areas about their ratemaking processes.
Apparently automobile was the only area in which we would
apply actuarial practices and those practices seemed very
formal. The results, however, would be anything but cut and dried.


In the world of Property & Casualty the inhabitants
subscribed to one of two mindsets. There were those who saw
the rates as a process of developing histories, balancing
statistical considerations and forecasting the future.
Especially forecasting the future. They were acutely aware of
the fact that even the most recent history available to base our
rates on was for policies written more than a year before and
needed to be projected a long way through the mists. They
struggled to see beyond the date of approval and implementation,
beyond the twelve months those rates would be used in the
market place, beyond the twelve months after that when the
last covered accident of the last policy sold in this planned year
would occur, and even thirty months beyond that when those last
claims would be sufficiently along the process of being
paid and closed to be considered complete. Until then, they
would not know whether the premiums calculated in this
ratemaking exercise had been adequate to produce a profit
for the company. Their calendar had sixty months.


To improve their aim, they carefully studied several
years of claims history, looking for trends in frequency, in
severity, in the patterns of payment, in the adequacy, over
time, of reserving practices for the unpaid portion of claims.
They pondered over what sort of events might cause shifts in
those trends. Would the public tire of safety features in the
cars they drove, would law enforcement become lax in the areas
they felt were the leading causes of accidents, would the
economy force changes in driving patterns, would the
government be able to contain the inflation rates, would the
courts award even larger settlements? They put their fingers in
the air, or tried to remember situations that felt as dire, and
how much those past crises actually moved their earlier
results. Their soothsayers were gloomy and felt the company
needed to move the rates up rapidly.


The other group lived by the taxman's calendar, the
shareholders' laws. Their focus was on this twelve months,
here and now, the premiums written in them and the claims
reported in them. Every day was spent dealing with the
public, with the agency forces, with the taxman. They studied
the competition and queried the independent agents for their
assessment. Why were the premiums written this year by
agents in this territory down? They watched for signs of any
political movement that would impact their bookkeeping.
They wanted to scoop the competition, they wanted to have
good news to appease the agency powers. They feared a
different trend, one in which the agents stopped looking at our
rates when advising their clients. Market share was the key to
their success. Premiums and claims reserves were simply their
investment resources. Their opinion seldom matched the
soothsayers' and the divergence was becoming acutely
uncomfortable for the senior chiefs.


The success of the company depended on both sets of
requirements being met, but was it possible? If not, should they
compromise? Who was right? How would Home Office react?
Having a rating structure that reflected the company's
experience was hoped to be the basis for resolving the
divergence. Their desired structure would lower the premium in
profitable areas, hopefully by more than our competition,
making the twelve month inhabitants happy, while raising it
for business we found unprofitable producing the premiums
needed to cover those claims when the sixty months were
finally clear. The new rate structure would be closely watched
by the senior chiefs.


Being the actuarial trainee, I should have been among
the advocates of the sixty month mindset. But I was too new to
have a mindset nor any preconceived ideas about whether the
ratemaking process I saw defined by the IBC would deliver our
senior chiefs from their dilemma. My day-to-day proximity to
the underwriting chiefs, instead of an actuary, increased my
sensitivity to the twelve month mindset, somewhat, simply
because the underwriters were users of both calendars. As the
period approached for our first ratemaking exercises, I was
more concerned with performing in the limelight of this
controversy than in its resolution, however intriguing and
suspenseful.


In my role as winged beast, my mentor equipped me
with the latest in programmable calculators and I set out to fill
the books of rates with legions of numbers to match their moons
and portents...

... in most provinces the area factor
remained a problem, even under this scheme of
independent dimensions. Our data was still being spread into
unacceptably small samples for the area factors. The real
world reason was that political jurisdictions jealously insisted
on being handled separately from their neighbors. Even the
entire industry´s data for some of those areas was inadequate to
give samples that were unlikely to be unaffected by unusual
patterns. Now that may seem oddly phrased to some but it was
unfortunately appropriate to statistics since that branch of
mathematics strives to eliminate negatives, which then
requires its focus to be acutely negative. Anyway, the
distortions in these small samples was not just a matter of a few
flukes that could simply be discarded because disposing of one,
and its opposite counterpart for balance, could then materially
misstate the average by an equally unknowable amount in such
small samples. In fact, in some cases it could be the absence of
certain representative claim sizes that was the fluke.


The only viable way to deal with inadequate samples
was to get more data. To do that meant using more years of
claims history than just the most recent, or else using industry
data, which was to be our last, last resort.


Staying within these constraints of maximizing the
company´s input, led to more Byzantine complications because
claims histories, which had two dimensions of their own,
frequency and size, were not static. They developed.


To be precise...

© JHRaichyk