EXPENSE RATIOS ACROSS INSURANCE DISTRIBUTION SYSTEMS: AN ANALYSIS BY LINE OF BUSINESS

• DOI: http://doi.org/10.1111/j.1540-6296.1999.tb00051.x
• Author: Laureen Regan
• Published date: 01 January 1999
• Date: 01 January 1999

EXPENSE RATIOS ACROSS INSURANCE

DISTRIBUTION SYSTEMS: AN ANALYSIS

BY LINE OF BUSINESS

by

Laureen

Regan

Abstract:

The author examines whether inde-

pendent agency insurers are more expensive

to operate

than

direct writers in personal auto

insurance, workers compensation,

or

general

liability.

INTRODUCTION

wodistinct classes of insurance

T

distribution systems are used

in

the U.S. markets. Direct writing

ar-

rangements include insurers which

operate through mass merchandising,

where no salesperson is employed,

companies which sell through employ-

ees, or companies which use exclusive

agents. Exclusive agents are autono-

mous contractors, but are constrained

to represent the products of only one

insurer. Under any of these arrange-

ments, the insurer owns the customer

list, and thus the residual profits which

arise from the insurance transaction.

Under independent agency, the

agent is

an

autonomous contractor, and

may represent the competing products

of several insurers. The principal dis-

tinction between distribution systems

is the agent’s ownership rights to the

customer list under independent

agency. Agency ownership of the list

means that the insurer can not replace

the agent or contact clients directly

without the agent’s permission. The

agent, however, has the unrestricted

legal right to terminate business and

move her portfolio.’

Prior research suggests that the in-

dependent agency system is character-

ized

by higher expense ratios (Joskow,

1973;

Cummins and Vanderhei,

1979;

Barrese and Nelson,

1992;

Flanigan et

al,

1993).

Indeed, the independent

agency system has continued to lose

market share over the past two decades,

most notably

in

personal lines. In

1978,

independent agency insurers

controlled

62.52

percent of the mar-

ket,

with

a

45.4

percent share of per-

sonal lines.2 By

1994,

market shares

declined to

52.14

percent overall, and

just

33.2

percent of personal lines.

Much

of

this loss of market share has

been attributed to higher costs under

independent agency

(See,

for example,

Barrese and Nelson,

1992).

It is often

argued that independent agency insur-

ers have higher costs as

a

result of pro-

viding some differential level of ser-

vice to customers. However, the sup-

port for this service hypothesis is

mixed (See, for example, Etgar,

1976;

Cummins and Weisbart,

1977;

Doerpinghaus,

1991;

Barrese, et al,

1995).

Moreover, regulators have ar-

Laureen Regan

is

in

the

Department

of

Risk

Management and

Insurance,

School

of

Business and Management

at Temple University. This

research

was

supported

in part

by

a

grant

from

Temple University.

44

Risk Management

and

Insurance Review

gued that insurance distribution ex-

penses are too high and that regula-

tory attention should be paid to reduc-

ing insurer expenses.

Interestingly, over the same time

period independent agency share of

commercial lines remained relatively

stable, with shares declining from

75.25 percent in 1978 to 71.8 percent

in 1994. This suggests that indepen-

dent agency offers some advantage in

commercial lines and markets. Addi-

tionally, it might be the case that inde-

pendent agency is not at a cost disad-

vantage to direct writing

in

these lines.

This is the hypothesis explored in this

paper-

Most previous studies of expense

ratio differentials have examined over-

all

firm level expenses, and have con-

trolled for line of business differences

by including specialization measures

as independent variables.

An

excep-

tion is Flanigan, et

al.

(1993) which does

examine expense ratio differentials for

alternative distribution systems across

lines of business. Controlling for pre-

mium volume, the authors find that

direct writing insurers have signifi-

cantly lower expense ratios in six of

the eight lines studied. However, the

analysis does not control for other fac-

tors which might influence expense

ratios. Moreover, the data used

in

the

study was drawn from

Best’s Aggre-

gates and Averages, By Line Under-

writing Experience Exhibit,

which

publishes information only for those

insurers with more than $10 million

in net premiums written for each line

of business. Thus,

it

is possible that

the results are biased by including only

the largest insurers in each line.

The current paper adds to the analy-

sis by examining expense ratios across

lines of business, controlling for own-

ership form, line of business special-

ization, and other variables that might

influence expense ratios. The litera-

ture which examines the relative ad-

vantages of alternative distribution

systems is reviewed below. The

hy-

pothesis that the expense ratio advan-

tage to direct writing is consistent

across lines of business is then tested.

Using a random sample of insurance

companies and groups, the empirical

findings indicate that expense ratio

differentials are line specific,

with

di-

rect writers enjoying an advantage in

only two out of the six lines analyzed.

RELATIVE COSTS AND

ADVANTAGES

OF

INDEPENDENT

AGENCY

Several explanations for the over-

all lower expense ratios of exclusive

agency insurers have been advanced.

First, renewal commissions are typi-

cally lower than new business commis-

sions under direct writing, because di-

rect writing agents do not have alter-

nate placement opportunities. D’ Arcy

and Doherty (1990) argue that direct

writer insurers have a cost advantage

because agents can not trade on the

value of the private risk information

that accrues over the term of the rela-

tionship with the policyholder. Be-

cause of their ability to capture the

gains from this information, indepen-

dent agents are more often paid an

additional commission which is con-

tingent on the quality of the business

placed (Regan and Tennyson, 1996).

45