A THEORETICAL EXPLORATION of the Adoption and Design of Benefit Plans

8/15/2019 A THEORETICAL EXPLORATION of the Adoption and Design of Benefit Plans

1/21

' Academy ol Monagement Review

1998,

Vol. 23, No. 2, 305-324.

A THEORETICAL EXPLORATION OF THE

ADOPTION AND DESIGN OF FLEXIBLE

BENEFIT PLANS: A CASE OF HUMAN

RESOURCE INNOVATION

MELISSA W BARRINGER

University of Massachusetts at Amherst

GEORGE T MILKOVICH

Cornell University

Hong Kong University of Science Technology

In this article we explore theoretical explanations of managers' decisions about flex-

ible benefit pl ans . First, we exa min e the adoption an d desig n of flexible benefit pla ns

throu gh four theoretic le nse s: (1) institu tiona l, (2) resou rce d epe nde nce , (3) age ncy,

and (4) transac tion cost. We then integra te the relevant insights ga ined from these

theories into a more complete model and derive propositions for future research.

Finally, we generalize the insights gained from exploring a specific innovation to

broader questions surrounding decisions about other human resource innovations.

In the early 1970s only a handful of co mp an ies

offered flexible benefit plans. Today, an esti-

ma ted one-third of organizations w ith more than

1,000 employees offer flexible benefits (Hewitt

As soc iates, 1995). Early pla ns w ere relative ly

sim ple, offering few cho ices an d lim ited flexibil-

i ty. Current approaches range from a simple

offering of pretax salary reductions, to pay for

benefi ts, to highly complex plans with many

options.

Beyond surveys of firm practices, scholars have

devoted little attention to these developments.

Very little research has been done to gain an un-

derstanding of why some firms adopt and others

do not or why some plans offer more choices than

others. We know virtually nothing about why the

variety of these plans has increased over time.

What objectives ar e achieve d? How do employe es'

attitudes, competitors' practices, rising benefit

costs,

and shifting regulatory conditions affect

management's decisions?^

We thank Anurag Sharma, Tony Butterfield, Pamela Tol-

bert, Madan Pilitula, and three of

AMR s

anonymous review-

ers for their invaluable comments on earlier drafts of this

article.

' When referring to decision s to adopt a nd d esign flexible

plans, we mean those organizational decisions made by

top-level management. For simplicity, we use the term

man agem ent 's decis ions in the art icle .

Our purpose is to explore the insights offered

by organization theories and research. Specifi-

cally, we first examine management 's adoption

and design decisions through four theoretic

len ses : (1) institutiona l, (2) resou rce de pe nd en ce ,

(3) age ncy , an d (4} tran sac tion cost. We then

integrate the insights gained from these theo-

ries and related research into a more complete

model, with testable proposit ions for future

empirical work. Finally, we generalize these

insights to other human resource (HR) innova-

t ions.

The principal contributions of our article are

threefold. First, by integ ratin g th e insig hts of mul-

tiple theoretical pers pec tives, we offer a m ore par-

s imonious model of management ' s decis ions

about flexible benefits and a set of testable prop-

ositions that can guide much-needed research.

None of the four theories considered here includes

th e full set of variables that may affect manage-

ment's decisions about employee benefits. Each

theory offers unique insights, although there is

also some ov erlap. Second, our model and analy-

sis offer a platform for understanding why or-

ganizations adopt other HR innovations. Third, we

bring together a diverse body of literature on flex-

ible benefit plans around theoretic perspectives.

Currently, the literature is largely segm ented into

benefi ts administrat ion, survey evidence, and

academic research.

5


2/21

6

cademy of Management Review

April

To acc om plish our three objectives, w e first

describe flexible benefit plans and review the

related l i terature and rese arch. Next, we p resent

a briei overview of the four theories and high-

light their relevance to flexible benefits deci-

sions. Following this overview, we ofier our in-

t eg ra t ed model based on spec i f i c i n s igh t s

gained from these theories, and we derive test-

able proposit ions. We conclude by discussing

the poten tial to generaliz e our id eas to other HR

innovations.

FLEXIBLE BENEFIT PL NS

Flexible Benefit Plans Defined

Definitions of benefits vary considerably. The

narrowest definit ion includes only those bene-

fits not underwritten or paid for by the govern-

ment, such as health insurance or private pen-

s io ns (Beam & M cFa dde n, 1996). Bro ader

definitions includ e virtually an y iorm of com-

pensation other than direct wages paid to em-

ploy ees {Rosenbloom, 1996: 3) and en co m pa ss

benefits that range from commonly observed re-

t irement and insurance plans to unique offer-

ings, such as spiritual counseling or fruit from

the orchard (Employee Beneiit Research Insti-

tute,

1995).

Definitions of flexible benefit plans are less

ambiguous, and the list of forms that are in-

cluded in such plans even more narrow. In gen-

eral , any plan that al lows employees to make

choices about the benefits they receive is con-

sidered to be a flexible benefit—or cafeteria—

pla n (Beam & M cFadd en, 1996; Rosenbloom,

1996). For the purposes oi this article, we use the

dei initio n set out in Sec tion 125 of the In tern al

Revenue Code. Specifically, we deiine flexible

beneii t plans as those that offer employees a

choice between quali i ied (nontaxable) benefi ts

an d c as h (Beam & Mc Fadden, 1996). A pla n that

does not offer a cash option—for example, a

plan that simply ofiers a choice oi medical

plans—is not considered a flexible benefit plan

under this deiinition, whereas a plan that offers

the option of payi ng for a (qualified, non taxa ble)

beneii t with pretax wages (cash) is (Borleis,

1996).

Although the options that can be included in

ilexible benefi t plans are somewhat l imited, the

plans incorporate a variety of different designs,

ranging in the amount oi flexibility (choice) and

the costs to the employer. We include four gen-

eral types of design s that are con sistent with the

IRS definition ad opt ed for us e in this artic le:

(1) salary reduction only, (2) modular options,

(3) core plu s o ptions, an d (4) mix an d ma tch. W e

summarize the four plan designs in Table 1

(Beam & M cFa dde n, 1996; Em ploye e B eneiit Re-

sea rch Institute, 1991). As we sho w in the ta ble ,

costs to the employer increase as the amount of

flexibility inc rea se s (Beam & McFa dden, 1996).

Hence, costs and flexibility are lowest under the

salary-reduction-only design and highest under

the mix-and-match design.

The Appeal of Flexible Benefit Plans

Flexible benefi t plans represent a substantial

departure irom the tradit ional , one-size-i i ts-al l

ph i losophy, where managers made decis ions

about the type and level oi benefi ts employees

received and al l employees were covered by the

same benefi ts. First introduced in the early

1970s, this in nova tive p ractic e did not start to

gain popularity until the mid 1980s (Hewitt As-

soc iate s, 1995). The num ber oi ad op tion s in-

cre as ed by 127 perc ent bet w ee n 1983 an d 1984

and has cont inued to increase s teadi ly , a l -

thou gh the rate h a s slowe d sinc e 1989. Cur-

rently, about 70 percent oi firms that oiier ilexi-

ble benefit plans are in the service industry

(Hewitt A ss oc iat es , 1995). The sim ple r, low-

flexibility type of design (salary reduction only)

seems to be the most widely used, al though the

incidence of the highly flexible plans seems to

be increasing (Employee Benefit Research Insti-

tute,

1995; He witt A ss oc ia tes , 1995).

Managers most frequently ci te two objectives

in their decisions to adopt flexible benefit plans:

(1) to enh an ce their emp loyees ' bene ii ts sat isfac-

tion or (2) to help con tain rising ben eiits costs

(Employee Benefit Research Institute, 1995). The

premise underlying the first objective is that as

the diversity of the workforce has increased, so

too has the d ivers i ty o i employees ' benef i t s

needs. Allowing employees to select a package

that meets their individual needs has increas-

ing appeal . In addit ion, f lexible beneii t plans

hold the promise of containing costs by moving

from a deiined beneiit (in which a certain pack-

age of benefits is promised, regardless of cost) to

a defined contribution (in which a certain level

oi benefi ts expenditures is promised, regardless

oi what the funds can purchase; Employee Ben-


3/21


4/21

8 Academy of anagement fleview

April

to the presence oi flexible benefits (Cable

Judge, 1994).

Explaining Variation in Decisions About the

Adoption and Design of Flexible Benefit Plans

The high costs of plan design and implemen-

tat ion, ever-changing regulations, paternalist ic

concerns about employees ' capaci ty to make

sound decisions, and actuarial concerns about

disproportionate part icipation rates among op-

t ions (adverse selection) have been oiiered as

reasons why more organizations have not im-

plemented flexible benefits (Beam McFadden,

1996; Bloom Tr ah an , 1986). Tha t som e org an i-

zations are deterred by these obstacles and oth-

ers are not suggests that decisions about the

plans are also affected by contextual factors.

To better understand these iactors, we turn

next to explore the insights offered by organiza-

tional theories. We have selected iour: (1) insti-

tutional, (2) resource dependence, (3) agency,

an d (4) tran sac tion cost . Our selec tions ar e

ba sed on two related cri teria. First, we co nsider

benefi t plans to be at tr ibutes of organizations,

akin to other organization proce sses a nd struc-

tures, such as iniormation system s, organization

design, or hierarchical arrangements. The or-

ganization is the basic unit oi analysis. Conse-

quently, we do not include such theories as ex-

pectancy and procedural just ice, because they

tend to iocus on the individual as the unit of

ana lysis . Agency theory also focuses on individ-

ua l s ,

but the unit of analysis is the contract

be tween p r inc ipa l s and agen t s (E i senhard t ,

1989).

Second, we have selected theories that al-

ready have been useiul in offering explanations

of other pay and beneii ts practices. Eisenhardt

(1988), for instance, has iound that both the

agency and inst i tut ional models part ial ly de-

scribe variat ion in the pay policies for s ales per-

sons at retai l stores. Similarly, Pfeffer and

Davis-Blake (1987) show that the resource de-

pende nce perspect ive helps explain var iat ion in

the wages paid to univers i ty adminis t rators .

G ood stein (1994) an d In gra m an d S imo ns (1995)

have appl ied the resource dependence and in-

s t i tu t ional mode ls to exp lain org aniz at ions '

work-family practices, and Williamson (1981)

suggests that management ' s decis ions about

pensions may be related to transaction costs.

CURRENT ORGANIZATIONAL THEORY

A REVIEW AND COMPARISON

In this section we describe and analyze the

four relevant theories. We sum ma rize the re sults

in Table 2, which contrasts the basic premises,

key assumptions, implications, and l imitat ions

o i t h e i n s t i t u t i o n a l , r e s o u r c e d e p e n d e n c e ,

agency, and transaction cost models. Based on

this analysis, we then derive a model that inte-

grates the ins ights gained .

Institutional Theory

As we note in Table 2, those with the institu-

t ional perspective posit that organizations wil l

adopt an innovation, even if technically inefii-

cient, in order to gain legitimacy and, hence, the

resources necessary to ensure thei r surv ival

(DiMaggio Powell, 1983; Meyer Rowan, 1977;

Zucker, 1987). Or gan izatio ns ma y ad opt prac -

t ices voluntari ly, in response to pressures to

conform to accepted standards oi practice, or

involuntarily, in response to coercion by power-

iul institutional iorces that control critical re-

so urc es (DiM aggio Po we ll, 1983; Scott, 1987;

Tolbert Zucker, 1996). Pro pon en ts of this theo ry

dist inguish between early, or preinst i tut ional ,

adopters, who introduce an innovation based on

its capacity to improve organizational perfor-

mance, and later adopters, who, when the prac-

tice has become semi- or fully institutionalized,

are more l ikely to behave according to prevail-

ing practice (DiMaggio Powell , 1983; Tolbert

Zucker, 1996).

The capacity of the institutional model to help

explain variat ion in organizational structures is

supported by several studies, including many in

which the authors examine innovative pay and

beneii t practices (see, ior example, Powell

DiM agg io, 1991; Scott, 1987; To lbert Zuck er,

1996, ior reviews oi this literature). The theory

has been used to explain organizations' iamily-

iriendly polic ies (Goodstein, 1994; Ingra m

Sim ons, 1995; O ste rm an , 1995), child ca re pro-

grams (Kossek, Dass,

DeM arr, 1994), exe cu tive

com pen satio n (W estphal Zajac, 1994), an d the

compensation oi retai l sales personnel (Eisen-

hard t, 1988). Ne verthe less, ins titutiona l theory

does not iul ly explain variat ions in these prac-

tices and structures. Critics note that it iails to

take into account the importance oi the organi-


5/21

1998 Barringer and

Milkovich

309

0)

a

o

01

I

0)

G

3

O

M

E

Ul

d

13

S c

1 d G 1)

«

di

3 d

G

3

.

O

E

E d

0 "-

-=: a

•5

a

X

2

O

U) (T

0) O U

|

g

O

0)

d

K

d >

o

i .-• G

g ^

>•

u y

d - B l

G

O Q) •

1 3

G

d

0.y

S

S ^ O

s § : 2

0 dTJ

G

O 1-

o

di.S

S —

d

d

d i3

g

:^ Ul

^

^ G

o

5 o

4=

a u

dS

S

u

S

d tri 0

^

ad-a _

^ O

II

i-f-

•

0

0

>

d

5^^

0

"a

0

0

o

y

>,_tn

S2

0 n O.

E

o

5

B-

a o

TI

G

D

Q)

tn

>

d

ui

tn

Z

g -I

g'u

5JU1

o

<

II

o

0

01

0

G

O

- d

tn

G

o

d

N

'G

d

0

U

'3

x :

u

0

o

&°^0

G

> « O

C


6/21

310

cademy of Management Review

April

zations' strategic goals and self-interests (Di-

M agg io, 1988; O live r, 1991).

The ins t i tu t ional model seems to provide

fuller explanations of organizational structures

when integrated with other theories. Research-

ers have integrated the insights offered by the

institutional model with those offered by re-

source de pe nd en ce {Goodstein, 1994; Ingram &

Sim ons , 1995; Oliver, 1991), ag en cy {Eisenha rdt,

1988),

an d tra nsa ctio n cost (Pouder, 1996; Roberts

& Gre enw ood, 1997) mo dels. We are aw ar e of no

research that has a t tempted a comprehensive

integration of the unique insights offered by all

four theories.

The resource dependenc e, agency , and t rans-

action cost perspectives al l suggest that expec-

tations about efficiency drive decisions about

such structures as flexible benefit plans. Fur-

thermore, as Table 2 indicates, the models sha re

some key var iab les a nd as sump t ions . Yet, each

focuses on different types of efficiency gains,

thereby offering unique insights into the factors

affecting management 's decisions about f lexi-

ble benefit plans. For the purposes of this arti-

cle, we group the three theories into one cate-

gory focusing on expected efficiency gains.

Expected Efficiency Gains

Resource dependence theory In Table 2 we

note that the resource dependence model as-

sumes that management ' s decis ions are heavi ly

influenced by internal and external agents, who

control critical reso urc es (Pfeffer & Sala ncik ,

1978). Internally, certain positions or employee

groups controll ing cri t ical resources can, theo-

retically, impose preferred structures on their

org an iza tion s (Pfeffer & Sala ncik , 1978). Balkin

and Bannister (1993) argue that employees hold-

ing critical jobs in an organization can influence

decisions about forms of pay (salary, incentives,

bene fits, and the like). High pay and attra ctiv e

benefits, in turn, can attract and retain those

individuals having the greatest abil i ty to obtain

cri tic al re so ur ce s {Pfeffer & Da vis-B lake , 1987).

External agents may control funds (e.g., gov-

ernmen t con t rac t ing agenc ies ) o r per sonne l

(e.g., unions) and can exert pressure on an or-

ganiza tion to adopt certain structures by l inking

com plia nc e with reso urc e alloc ation (Pfeffer,

1981;

Pfeffer & Sala ncik , 1978). Th es e ex ter na l

pressure s bear a remarkable resem blance to the

insti tut ional mod el 's coercive press ures, an d

dis t inguish ing between the two models ' expla-

nations of organizations' structures, therefore,

ca n b e difficult (Tolbert & Zuck er, 1996; Zuck er,

1987).

As Oliver (1991) and others point out, how-

ever, the resource dependence model views or-

gan iza t ions as r espond ing more s t r a t eg ica l -

ly— or au tonomous ly— to ex te rna l p res su re s

than does the in s t i t u t iona l model . Hence ,

whereas both models consider external pres-

sures ,

the resource dependence model also con-

siders characterist ics of the organizations that

affect the extent to which decision makers ac-

t ively manage the pressures , versus pass ively

com ply w ith them {Gre ening & Gra y, 1994; Oli-

ver, 1991). Goodstein {1994), for example, found

that resis tanc e to pre ssu res to adopt family-

friendly policies varied with the percent of

women in the organization 's workforce.

Empirical research generally has supported

resource dependence-based explanat ions of or-

ganizational structures. Pfeffer and Davis-Blake

(1987), for instance, found that pay-level deci-

sions for employees occupying the same posi-

t ion in different organizations could be ex-

plained partially by differences in the role the

posit ion played in the acquisi t ion of monetary

resources. As noted above, in a number of stud-

ies ,

scholars have successfully integrated the

resource dependence and ins t i tu t ional models

to explain, part ial ly, a variety of organizational

practices, including benefi ts.

Agency theory. The focus of the agency model

is on the design of optimal employment con-

tracts. The model 's premise is that the interests

of principals may conflict partially with those of

the agen t s and tha t compensa t ion packages

should be designed to motivate employees to

act in the best interests of the principal (Jensen

& Meckling, 1976; Ross, 1973). Outcome-based

contracts provide powerful incentives for agents

to be as productive as possible, but when out-

comes are uncer tain , agents wi ll demand h igher

pay to offset attendant risk. Behavior-based con-

tracts shift the risks to the, presumably, risk-

neutral principal , but when information about

agent effort is difficult to obtain, incentives for

agents to work in the best interests of the prin-

cipal are weak. When outcome uncertainty is

high

and

monitoring employee productivity is

difficult or costly, optimal contracts are more

difficult to des ign . A po ssib le so lution to this

dilemma is to pay so-called efficiency wages—

tha t is , w ag es tha t a re above the marke t -


7/21

1998

airinger and Milkovich

3U

cl ea ri ng r at e {Na lban tian , 1987; Stiglitz, 1987).

Paying these h igher wage s creates an incentive

for employees to work at least hard enough to

kee p their jobs. At the s am e time , their pa y is not

dependent upon the outcomes of their efforts,

which can be affected by factors other than em-

ployee effort.

We suggest that the agency and efficiency

wage concept, which has been applied most

frequently to issues of direct pay, can be ex-

tended to total compensation (pay plus benefits).

Base pay and benefits that are not tied to per-

formance generally are considered to have a

weak or indirect impact on motivation (i .e. ,

through promotions). This prem ise su gge sts that

base pay and benefi ts are tied to continued em-

ployment, which, if high enough, will induce

employees to work at least hard enough to keep

their jobs (Sh apiro & Stiglitz, 1984). A sim ilar

argument is made by Shepard, Clifton, and

Kruse {1996), w ho app ly th e efficiency w ag e th e-

ory to ex pla na tion s of dec ision s to adop t flexible

work hours. The authors argue that employees

who value flextime will prefer to work in or-

ganizations that provide it and will work hard

enough to keep such employment because the

cost of being fired {and losing flextime) is high

{Shepard et al., 1996).

In a large body of research, scholars explore

the implications of agency theory for employee

com pen satio n (see Eisen hard t, 1989, for a re-

view).

In general , these scholars support the

agency model. Once again, many of the authors

use both agency and companion theories to ex-

amine research hypotheses—an approach sup-

ported by Eisenhardt , who argues that agency

theory pre sen ts a pa rtia l view of the world

that, alth oug h valid, also igno res a good bit of

the com plexity of org an iza tion s (1989: 71). Sp e-

cifically, agency theorists tend to overempha-

size efficiency explanations of organizational

structures and neglect the inst i tut ional context

in which they are estab lishe d (Kalleberg Reve,

1993;

Nila kan t & Rao, 1994).

Transaction cost . Those having this perspec-

t ive assume that organizations establish struc-

tures minimizing the costs of their transactions

(the exchange of goods or services) with other

part ies when the potential for opportunist ic be-

hav ior is high (W illiamson, 1981). In the ca se of

HR transactio ns, the potential for opportunist ic

behav ior is related to two important huma n

ass et chara cterist ics. First , whe re cri t ical skil ls

specific to the organization a re acqu ired primar-

ily on the job, turnover can be costly, and or-

ganizat ions adopt in ternal governance s t ruc-

tures to stabil ize employment. Second, where

individual productivity is difficult to monitor,

organizations adopt governance structures to

provide incentives for employees to act in the

or ga niz ati on s' int ere sts (W illiamson, 1981: 564).

To the extent that employees value flexible ben-

efits, adopting the plans increases the value of

continued employment, hence reducing volun-

tary turnover and shirking.

Transa ction cost exp lana tions of organization-

al structures have, for the most part, been sup-

ported by research (see Klein Shelanski, 1994,

for a review). Much of the work has focused on

vert ical integration; however, authors hav e use d

the model to explain a variety of governance

is su es {Klein & Sh ela nsk i, 1994). W illiam son

{1981) ar gu es th at tra nsa ction costs can also ex-

plain the adoption and design of specific HR

practices (such as flexible benefits), yet empiri-

cal evidence in this area is limited. In one study

researchers found that the extent of transaction

costs,

such as firm-specific training, can help

explain the adoption of contingent work ar-

ra ng em en ts (Davis-Blake & Uzzi, 1993). In re-

spon se to cr i tic ism that the model n eglec ts

contextual variables, such as power and inst i tu-

t ional pressures, researchers recently have com-

bined transaction cost explanations with other

perspectives, most notably inst i tut ional theory

{Pouder, 1996; Ro ber ts & G ree nw oo d, 1997).

Summary Sources of Con vergen ce and

Divergence

As Table 2 i l lustrates, there are both conver-

gence and divergence in the insights offered by

the individual theories. The potential for inte-

grating the four theories without creating redun-

dancy and incongruity, therefore, may seem lim-

ited. Yet, a recurring theme in the empirical

l i terature is that r icher explanations of organi-

zational structures can be derived by integrat-

ing the insights of mult iple theoretical perspec-

t ives. Oliver (1991) ar gu es that con verg enc e

demonstrates the potential for integrating com-

p l e m e n t a r y e x p l a n a t i o n s o f o r g a n i z a t i o n a l

structures into a single model. Further, diver-

gence highlights the underlying assum ption s

about organizational behavior that are not ac-

knowledged by a s ingle theory and demon-


8/21

312

Academy

oi

anagement

eview

April

strate s the value of integra ting mult iple theories

(Oliver, 1991: 146).

Although all four perspectives assume that

management ' s decis ions about s t ructures are

constrained by environmental condit ions, they

differ in assumptions about the nature of those

const rain ts . External pressures are assumed to

be the primary source of influence on decisions

under the inst i tut ional model, whereas internal

conditions related to the nature of the work con-

st i tute the primary constraints under the agency

and transaction cost models. The resource de-

pendence model assumes that in ternal

and

ex-

ternal agents impose const rain ts on decis ions

about structures. Further, both the agency and

transaction cost models assume that constraints

arise as a result of goal conflict, information

asymmetries, and the tendency of people to act

in their own self-interest.

Differences in assumptions about organiza-

t ions' responses to environmental constraints,

as we note in Table 2, are p erh ap s the m ost

striking. Organizational interest and agency are

assumed to play an important role in determin-

ing these responses under the resource depen-

dence, agency , and t ransact ion cost models .

These rat ional mod els al l ass um e that organi-

zat ions act ively manage environmental con-

s t rain ts , adopt ing s t ructures that ensure the

flow of resources, or minimize agency or trans-

action costs. Hence, the primary source of influ-

ence on management ' s decis ions under these

persp ective s is what we refer to as expected

efficiency ga ins. The institution al model, how-

ever, assumes that organizations do not exercise

active choice; rather, they more passively con-

form to their environments. Organizations theo-

ret ical ly adopt s t ructures that enhance thei r

legit imacy in the external environment, regard-

less of the impact on the technical efficiency of

in ternal operat ions . Thus, ins t i tu t ional pres-

sures are the primary source of influence on

management ' s decis ions about such s t ructures

as flexible benefits.

We argue that the integration of the four the-

oretical perspectives is not only feasible but

necessary to understand decisions about f lexi-

ble benefit plans. All of the theories seem to

offer insights, yet none offers a full explanation.

On the one hand, inst i tut ional theory has been

criticized for ignoring the role of organizational

inter est an d ag en cy (DiM aggio, 1988; Oliver,

1991). The agen cy a nd t ran sac tion cost mo dels .

on the other hand, have been criticized for over-

emphasizing efficiency and ignoring social con-

text va ria bl es {Eisenhardt, 1989; Ka llebe rg &

Reve , 1993; Nila kan t & Rao, 1994). Fur ther, al-

though the resource dependence, agency , and

transaction cost models al l suggest that ex-

pected efficiency gains drive decisions, each fo-

cuses on different types of gains. In response to

these cr i t ic isms, researchers have in tegrated

different combinations of the theories. We are

aware of no research that has at tempted to in-

tegrate the unique and tenable insights of al l

four models. Accordingly, we take the logical

next step and propose a model for explaining

managemen t ' s dec i s ions abou t the adop t ion

and design of flexible benefit plans, integrating

the insights of the inst i tut ional , resource depen-

dence, agency, and transaction cost theories.

AN INTEGRATED APPROACH TO EXPLAINING

THE ADOPTION AND DESIGN OF FLEXIBLE

BENEFITS

As dep icted in Figure 1, we hypo thesize tha t

management 's decisions about f lexible benefi ts

are influenced prim arily by two factors: (1) insti-

tutiona l pre ssu re s and (2) exp ected efficiency

gains. These two factors provide the framework

for our discussion of the model. First, we exam-

ine the theoretical premises underlying each of

the factors. We identify antecedent variables

that are l ikely to shape the strength and direc-

tion of each factor's effects on management's

decisions about the adoption and design of the

plans , and we develop tes tab le proposi t ions .

Second, we examine variables that affect the

relat ive strength of the two primary factors,

again developing a set of related proposit ions.

Institutional Pressures

The inst i tut ional model suggests that internal

decisions about structures are heavily influ-

enced by pressures in the external environment

(Go odstein, 1994; Ingr am

Sim on s, 1995). Below ,

we outl ine the ways in which inst i tut ional pres-

sures to conform to such practices as flexible

benefi ts are generated theoretically.

Antecedents of

institutional pressures: Com-

petitors

practices. As a practice becomes more

diffused, organizations increasingly view i t as

standard—or best—practice, theoretically tying

an organization 's legit imacy and abil i ty to ob-


9/21

1998

arringei a nd Milkovicb

313

FIGURE 1

An Integrated Model of Organizations De cisions About Flexible Benefit Plans

Labor

market

conditions

Nature

o

the work

Competitors'

practices

Coercive forces

•Legislation

•Unions

Benefits

objectives

Uncertainty about

goals and technology

Organizational

Interconnectedness

Expected efficiency gains

Organization size

•• Stage 0 1 ^ ^

\institutionalizatioiv

Decisions about flexible benefit plans

•Adoption

•Design

tain reso urces to com pliance (Meyer & Rowan,

1977; To lbert & Zuck er, 1996). He nce , a s the por-

tion of competitors in an organization's field or

industry adopting a f lexible benefi t plan—or a

par t icular plan des ign—increases , the s t rength

of the pressures to adopt the plan a l so in-

c rea ses . The p res sure s to adop t shou ld be

greater in the heal th care industry, for example,

where over two-thirds of organizat ions offer

f lexible benefi ts , than in manufacturing, where

the incidence is lower. Similarly, the pressures

to include a f lexible spending account in the

plan should be highest in the services industry,

where the incidence of the accounts is highest .

Thus,

we propose the fol lowing:

Proposition ia; The ra(e of new flexi-

ble-benefit-pian adoptions will be

higher in industries in which the

portion of organizations offering the

plans is high than in industries in

which the portion is low.

Proposition Ib: Organizations imple-

menting a new flexible ben efit plan

are likely to match the plan design

e.g., core plus optional coverage or

mix and match) most comm only used

in their field or industry.

Antecedents of institutional pressures: Coer-

cive forces. Inst i tut ional pressures that are more

coercive in nature may be generated by external

agents who control cr i t ical resources and have

the power to impose preferred structures on an

org an iza tion {DiMaggio & Pow ell, 1983; Scott,

1987;

Tolbert & Zucker, 1996). Ou r m od el sug -

ges ts that unions may exer t such pressures .

Powerful unions that control the supply of labor

should theoret ical ly be able to influence man-


10/21

4

Acadeniy oi Managem ent Review

April

agement 's decisions about benefi ts practices.

Unions tend to view flexible benefit plans with

disfavor, primarily because of the concomitant

reductions in employer benefits that often occur

and because the resultant lack of uniformity in

benefi ts is counter to the unions' egali tarian

philosophy, making contract negotiat ions more

difficult {Employee Benefit Research Institute,

1991; Mulcahy, 1990; Wiatrowski, 1994). The re-

source dependence model suggests , however ,

that the strength of these pressures wil l depend

on the importance of the resources and the ex-

tent to which they

can he

obtained from other

sources (Meznar Nigh, 1995; Pfeffer, 1981; Pfef

fer & Salan cik, 1978). Hence , coercive pre ss ure s

are likely to be strongest in highly unionized

and labor-intensive sectors, where human re-

sources represent cri t ical inputs to production

and the power of unions to withhold or supply

labor is high. Close to half of the workers in the

labor-intensive public sector, for example, are

union members, suggesting high union control

over critical res ourc es (Freem an, 1996). There-

fore, we predict the following:

Proposition

2:

The rate of new flexible-

benefit-plan adoptions w ill be lower

in highly unionized labor-intensive

organizations than in organizations

where the degree o f unionization is

low.

Government agencies also may exert coercive

pressure on organizations to adopt preferred

str uc tur es {DiMaggio & Pow ell, 1983). Restric-

t ions on employee compensation, for example,

can be imposed on organizations that have fed-

eral contracts. We are aware of no such contin-

gencies related to flexible benefits. The govern-

ment, however, has passed iegisiafion over the

pas t 20 yea rs that ha s variously encoura ged or

discouraged adoption of the plans. Prior to 1978

the federal tax code discouraged conventional

flexible benefit plans—that is, plans in which

employees could choose between taxable and

no nta xa ble benefits. The doctrine of construc-

t ive receipt esse ntial ly incre ased the costs of

such plans by requiring that employees' taxable

income be calcu lated as if they had received the

maximum amount of taxable benefi ts available,

rega rdle ss of their actual selections (Beam &

McFadden, 1996). The Revenue Act of 1978 ended

the application of this doctrine to flexible bene-

fit plan s, thus restoring favorable tax treatme nt.

Other laws, a l though not speci f ical ly ad-

dressed to flexible benefi ts, may have had the

effect of pushing organizations to build more

flexibility into their benefit plans. The Health

M ainte nan ce Act of 1973, for instan ce, requ ired

employers to give employees the option of join-

ing a Health Maintenance Organization (HMO),

if locally a va ila ble (Beam & Mc Fadde n, 1996).

More recent ly , leg is la t ion requir ing unpaid

leave for family medical needs and portable

heal th insurance has increased the pressure on

employers to become more flexible. Hence, the

strength of legal pressures to introduce flexibil-

i ty into benefi ts programs seems to be increas-

ing.^

Decisions about the

design

of flexible ben efits

also may be affected by coercive governmental

pres sure s. Concerned a bout the potential loss of

tax revenue s result ing from pretax contributions

to flexible spending accounts, the federal gov-

ernm ent estab lish ed the use it or lose it rule in

1984. The rule had the effect of limiting em-

ployee use of these accounts by requiring that

any money left in the funds at the end of the

enr ollm ent p erio d be forfeited. In 1989 the gov-

ernment enacted legislat ion that further dis-

couraged employers f rom offer ing f lex ib le

spending accounts , requi r ing them to make

available at the beginning of the enrollment

period the full amount of the yearly contribution

election, even if employees had not yet contrib-

uted the full amount to the account.

Changes in leg is la t ive pressures are not nec-

essar i ly related to cha nge s in m anag eme nt ' s

decisions a bout the adoption a nd d esign of f lex-

ible benefit plans. As Tolbert and Zucker {1983)

point out, strong resistanc e ca n lead to noncom-

pliance with unpopular legislat ion (e.g. , man-

dated integration of public schools). Further, the

r e s o u r c e d e p e n d e n c e m o d e l s u g g e s t s t h a t ,

rather than passively conforming to such pres-

sures, organizations wil l at tempt to actively

manage them wherever feasib le and desi rab le

(Oliver, 1991). Ind eed , Sec tion 89 of the Inte rna l

Revenue Code, which created complicated non-

discrimination rules for al l employer benefi t

plans , wa s repe aled just 3 yea rs after i ts pas-

sage, following objections that the rules were

^ Many of the ideas we di&cuse about the effects of gov-

ernmental legislation are based on comments from an anon-

ymous reviewer.


11/21

1998

Barringer

and Milkovich

3 5

too difficult and costly to implement (Employee

Benefit R ese arc h Ins titute , 1991). Evid enc e su g-

ges t s , however , that organizat ions general ly

have been responsive to changes in legislat ion

pertaining to flexible benefits. In 1984—the year

the federal government issued regulations clar-

ifying the terms and restrictions on flexible ben-

efit pl an s set out by section 125 of the In tern al

Revenue Code—new adoptions of the plans in-

cre ase d by 127 perce nt. Based on theory a nd

evidence then, we predict:

Proposifion 3a: The rate o f increase in

new adoptions of flexible plans will

be highest following the introduction

of favorable legislation.

Proposition 3b: The variety of plan de

signs will be responsive to favorable

legislation.

Expected Efficiency Gains

All four theoretical perspectives provide in-

sights regarding the effects of efficiency con-

cerns on decisions about f lexible benefi t plans.

Under the institutional model, efficiency is only

expected to play an important role during the

early, preinst i tut ionalization stag e, wh en the in-

cidence of the plans is low and adoption deci-

s ions dep end on the de gre e to which the

ch an ge improves internal proce ss {Tolbert &

Zucker, 1983: 26). The res our ce de pe nd en ce

model implies that organizations adopt total

compensation systems that wil l ensure the flow

of critica l res ou rc es (Balkin & Ba nnis ter, 1993;

Pfeffer

8i

Da vis-Blake , 1987). The a ge nc y mo del

suggests that organizat ions adopt compensa-

t ion structures that wil l reduce shirking where

productivity is difficult and/or costly to monitor

(Eisenha rdt, 1989). Under the tr ans act ion cost

model, organizations theoretically use compen-

sation systems both to reduce shirking where

productivity is difficult to monitor and to reduce

costly turnover among employees who have re-

ceived firm-specifc training (Williamson, 1981).

As Figure

shows, we posit that expectat ions

about the efficiency gains associated with flex-

ible benefit plans are directly influenced by the

organization's benefits objectives. The institu-

t ional perspective assumes that during the pre-

inst i tut ionalization years of f lexible-benefi t-

p lan adopt ion , when external pressures are

weak, management makes decis ions about the

plans based on i ts expected impact on technical

operations. Hence, organizations wil l be l ikely

to adopt flexible plans during this stage only if

the promised outcomes (enhanced benefi ts sat-

isfaction and cost containment) are (1) important

to the effectiveness of the organization's bene-

fits pla n an d (2) likely to be ac hie ve d. A doption

decisions, therefore, should be determined, at

least in part, by the organization's benefits ob-

jectives. Similarly, the particular type of design

implemented should be logically related to the

achievement of plan objectives. The assumption

under the resource dependence, agency , and

transaction cost perspectives that organization-

al interest and agency play an important role in

determining structures also implies that deci-

sions about flexible plans will be rationally re-

lated to benefits objectives. Predicting decisions

about the plans, therefore, requires information

about organizations' benefi ts objectives.

Although the inst i tut ional model implies that

decisions about f lexible plans during the early

stage will be influenced by benefits objectives,

the model's focus is primarily on explaining en-

vironmental constraints on decisions during the

later stages. We can more readily derive in-

sights into the particular factors that constrain

the setting of benefits objectives from the re-

source dependence, agency , and t ransact ion

cost models. Together, these three perspectives

suggest two sets of factors affecting organiza-

t ions ' ben efits objec tives: {1) the n at ur e of the

work and (2) labor m arket cond itions.

Theoretically, the nature of the work affects

the value of employees and the ease of monitor-

ing their productivity. Under the resource depen-

dence model, attributes of the work affect the

extent to which organizations are dependent on

employees for the acquisition and/or efficient

us e of critic al res ou rce s {Bartol & M artin, 1988).

When resource dependency is high, organiza-

tions will emphasize satisfaction objectives as a

means of at tracting and retaining cri t ical em-

pl oy ee s {Balkin & Ba nn iste r, 1993; Pfeffer &

Davis-Blake, 1987). Bartol and Martin (1988) iden-

tify four work attributes that affect employee

control over c ritical re so urc es: {1) the d eg re e of

uncer tain ty surrounding the achievement of

t asks ,

{2) the im porta nce of task a chie vem ent to

the or ganiz ation (task centrality), {3) the ea se

with which the work can be observed and mea-

sured , and (4) the level of skill spe cializ ation

required to achiev e tasks . Such condit ions m ight


12/21

316

cademy of Managem ent Review

April

be found in the research and development divi-

sion of a high-tech company, where creative ef-

forts related to produ ct deve lopm ent c an be dif-

ficult to monitor, specialized skills required to

perform the work can be difficult to replace, and

the development of innovative products is cen-

tral to the com pa ny 's su cce ss (Bartol & Martin,

1988;

Eisenhardt, 1989).

We propose a fifth factor: the extent to which

the technology is labor intensive versus capital

intensiv e. Wh ere HR inputs rep resent a rela-

tively small part of the production process, op-

era t ions can cont inue even when these inputs

are withdrawn. Accordingly, capital- intensive

organizations (e.g., oil or chemical), although

dependent , are re la t ively less dependent on

workers wi th discre t ionary cont rol over re-

sources than are more labor-intensive organiza-

tions (e.g., service).^

The agency model implies that benefi ts sat is-

fact ion wil l take precedence over cost-contain-

ment object ives when organizat ions face both

high monitoring costs and high outcome uncer-

ta inty . Under these condi t ions , behavior -based

contracts provide l imited incentives for self

interested agents to act on behalf of the princi-

pal ,

and outcome-based contracts inefficient ly

al locate r i sks to r i sk-averse agents . Al terna-

t ively, und er a behavior-based/eff iciency w ag e

approach, compensat ion is t ied not to outcomes

but to continued employm ent, the desirab i l i ty of

which is increased by offering employees pay

and benefits that they value (Nalbantian, 1987).

The organizat ion reduces monitoring costs, be-

ca us e emp loyee s do not wa nt to r isk losing their

jobs (with the valued benefits) and, therefore,

have an incentive not to shirk, even when not

closely m onitored (Calvo, 1987; Sh ep ard et al.,

1996).

Like the resource dep en den ce model , then,

the agency model impl ies that where employee

effort is difficult or costly to monitor, organiza-

tions are likely to em ph asiz e offering bene fits

sys tems that enhance employee sa t i s fact ion.

A key var iable not sugg ested by the resou rce

dependence or other models i s outcome uncer-

ta i nt y^ th at i s , the extent to which outcomes are

af fected by factors beyond employee ef for t

and/or are difficult to measure. To illustrate.

We use industry examples such as these ior the purpose

of illustration only. Empirical tests of our model would re-

quire direct measurement of variables, such as task uncer-

tainty, rather than the use oi proxies, such as industry.

consider consult ing f irms, where professionals

with special ized knowledge may be diff icul t to

monitor and outcomes are affected by factors

beyond the consul tant ' s cont rol . When manage-

ment in such organizat ions decides to imple-

ment flexible benefits, the aim, more likely, is to

enhance employee sa t i s fact ion.

The ease of monitoring productivi ty emerges

as a key work at t r ibute explaining organiza-

t ions ' benefi ts object ives under the t ransact ion

cost model as well. When monitoring is difficult,

the potential for opportunist ic behavior is high.

The assumption is that employees are not apt to

work hard if they believe that lower effort is

unlikely to resul t in any negative consequences.

Incentives for employees to act in the interests

of the organizat ion are, in many cases, provided

by tying the employee's pay to organizat ion or

individual per formance. Kal leberg and Reve,

howev er, arg ue that in an economic sen se, in-

centive s are th e price vector of the contract , an d

include pay, bonuses, and fr inge benefi ts

(1993:

1113).

As wi th the resource dep end enc e and

age ncy models, this implies that the price of

employment—hence, the cos t of los ing i t—

affects em ploy ees ' prope nsi ty to act in the inter-

est of the organizat ion.

A second at tr ibute of work under the t r ans ac-

t ion cost model—one not specif ied by other

models— is the extent to which ski l ls required to

achieve the work are acquired on the job (hu-

man asset specif ici ty) . When employees acquire

critical skills specific to the organization pri-

mari ly on the job, an organizat ion's cost ly in-

vestments in t raining can be lost i f workers quit

before expected productivi ty gains are ful ly re-

al ized. Examples of such organizat ions include

consulting firms, law firms, and financial bro-

kerage f irms, where revenue creat ion is l inked

to bui lding re la t ionsh ips am ong c l ients . In these

ci rcumstances internal governance s t ructures

will be efficient if they stabilize employment.

Benefits that accrue with seniority, for example,

will disc ou rag e q uittin g (W illiamso n, 1981). In-

deed, many f lexible plans include a provision

that the al locat ion of credi ts for pu rch asin g b en-

efi ts wil l increase with seniori ty (Employee

Benefit Research Institute, 1991). Compensation

packages that include pay or benefi ts that ex-

ceed what employees could obta in e l sewhere

also wil l discourage quit t ing. Hence, when cri t -

ical skills are firm specific and/or monitoring

em ployee perform ance is difficult , o rganiza t ions


13/21

1998

Barringer and Milkovich

31 7

m ay try to minimize their transaction costs by

offering total compensation packages that

en-

h a n c e the v a l u e of con t inued employm en t ,

thereby discouraging employees from quit t ing

or shirking.

In

this case benefits objectives will

emphasize employee satisfaction.

abor market conditions

theoretically affect

the extent

to

which

an

organization

is

depen-

dent on external agents for crit ical human re-

sources . As we noted e arlier, unio ns that control

the supply

of

labor

to an

organization

can di-

rectly influence decisions about flexible benefit

p lans .

At the same t ime, nonunionized organi-

zations

in

indust r ies where membersh ip

is

high

might seek to avoid un ion involvement by em-

phasizing employee satisfaction objectives and

reducing workers' felt need

for

collective repre-

sentation. Hence,

the

percent

of

union mem bers

in an organization 's labor market may help ex-

plain benefits objectives. Conditions

in the la-

bor market also determine the ease wi th which

critical skills can be replaced. Replaceabil i ty,

as ind icated

by

such labor market c ondit ions

as

unem ploymen t rates , theoretically affects organ -

izat ional dependence on employees ' resources

and, hence,

the

importance

of

satisfying bene-

fits pr ef er en ce s {Bartol M artin , 1988).

To summarize, an organization 's benefi ts ob-

jectives can be explained , at least in part , by a

set

of

variables describing internal work

at-

t r ibutes and external labor market condit ions.

Where the work is such that the opportunit ies for

sel f - in teres ted

or

opportunis t ic behavior

are

high and/or employee control over critical re-

sources is high, organizations wil l tend to em-

phasize benefits satisfaction over cost control.

Satisfaction also is likely to be the primary ob-

ject ive where union involvement is high and/or

there

is a

shortage

of

critical skills

in the

labor

market . Organizat ions facing relat ively less

pressure from these condit ions are less likely to

be concerned w ith providing

a

valuable benefi ts

p l a n and are more l ikely to focus on cost-

containment objectives. ' '

Obviously, organizations typically claim multiple bene-

fits objectives, which include both employee satisfaction

and cost containment (Beam

McFadden, 1996). Neverthe-

less, survey evidence suggests that organizations prioritize

their objectives

and

easily specify

a

single, primary objec-

tive behind their decisions to adopt such practices as flexi-

ble p lans .

In a 1992

survey

52

percent

oi the

part icipat ing

iirms indicated their primary reason

ior

implementing ilex-

Of course, not all organizat ions whose pri-

mary benefits objective is employee satisfaction

are going

to

adop t flexible ben efits. Evidenc e

suggests that employees ' react ions to the plans

vary , perhaps because some see the enrollment

procedures

as

confusing

and

time consuming

(Kossek, 1989). Nor can organizat ions alw ays ex-

pect to ach ieve c ost-containment objectives by

adopting flexible benefi t plans.

The

costs

of im-

plement ing the plans can be high, possibly out-

weighing potential cost savings. We posit that

the effect

of

benefits objectives

on

decis ions

about flexible plan adoption will be moderated

by current fnonfJexijbieJ

benefit plan costs,

flex-

ible plan implementation

and

adminis t rat ion

costs (pJan cos*s), and employee

preferences.

Organizations whose primary objective

is

cost

control

are

unlikely

to

ado pt flexible bene fits

if

the adminis t rat ive costs of the plans are high

relat ive

to the

expected returns. Small organiza-

t ions,

and/or organizations without established

information systems, are less likely to have the

r esou rces

to

cost-effectively im plem ent

and

manage flexible benefi t plans than are larger

organizations, part icularly in the early years of

adoption, when software packages simplifying

program adminis t rat ion are not yet avai l ab le

(Lawler, 1981). How ever, org an iza tion s w ith gen-

erous benefits, such

as

first-dollar medical

cov-

erage (i.e.,

no

deduc t ib le

or

coinsurance), might

expect to contain the rising costs of benefits by

moving from

a

defined benefit

to a

defined

con-

t r ibution arrangement, as often occurs with the

introduction of flexible benefit plans.

Proposifion 4a: Amon g organizations

whose primary benefits objective is

cost control,

the

propensity

to

adopt

flexible benefit plans will be posi-

tively related to the munificence of

current nonflexible) benefits.

Proposition 4b: Amon g organizations

whose primary benefits objective is

cost control the propensity to adopt

flexible benefit plans will be posi-

tively related

to

organizational

re

ible benefits was health beneiits cost containment, and 28

percent indicated their primary objective was to satisfy di-

verse em ployee n eed s (Employee Benefit Research Institute,

1995). Our

analy sis sugg ests thai

the

primary objective will

be more helpful than secondary objectives in explaining

management's decisions about flexible benefit plans.


14/21

318

Academy of Management Review

April

sources

as indicated by size and cur-

rent information systems).

An organization may expect f lexible plans to

enhance benefi ts sat isfaction when i ts work-

force is highly diverse (in terms of gender, eth-

nici ty, and age). Whether an organization actu-

ally enhances sat isfaction by adopting flexible

benefits is unclear. In a study of employees'

rea ction s to the introduction of a flexible benefit

plan. Barber et al. found no significant relation-

ships between satisfaction with the plan and

diver gen ce from the trad itiona l profile —that

is ,

long-tenure, married male s with children

an d non wo rkin g s po us es (1992: 59, 62). Never-

theless, survey evidence clearly suggests that

employers believe adopting flexible benefi ts

will enhance the benefits satisfaction of a di-

verse workforce (Employee Benefit Research In-

stitu te, 1995; Hew itt Ass oci ate s, 1995). O rga niz a-

tions also may expect flexible benefit plans to

increase the benefi ts sat isfaction of employees

who are relatively well educated if, as Lawler

(1981,

1990) argues, such employees prefer to be

involved in making decisions that affect them. It

may also be that employees in the h igher-

income tax brackets are at tracted to flexible

p lans because of opt ions (such as f lex ib le

spending accounts) that provide a tax shelter for

cash earnings. Hence, we predict:

Proposifion 5:

Amo ng organizations

whose p r imary

bene fits objective is

employee satisfaction, the propen-

sity to adopt a flexible benefit plan

will be positively related to the di-

versity, education, and income level

of the workforce.

The assumption that expectat ions about effi-

ciency gains drive decisions about f lexible ben-

efi ts implies that the plans wil l be designed in

accordance with benefi ts objectives. Thus, we

expect that programs implemented primarily to

meet the diverse benefi ts needs of a heteroge-

neous workforce will offer more flexibility to em-

ployees in selecting their benefits. As we indi-

cate in Table 1, plan des igns tha t provide the

most choices—or flexibility—are the core-plus-

opt ions and mix-and-match p lans . Organiza-

t ions emphasizing employee sat i sfact ion are

more l ikely to implement these plan designs

rather than the modular-options plan, which

limits choices to prepackaged plans, or the sal-

ary-reduction-only plan, which simply offers the

opportunity to pay for predetermined benefits

with pretax dollars. Therefore, we predict the

following:

Proposifion 6:

Organ izations whose

primary benefits objective, at the time

of flexible-benefit-plan adoption, is

employee satisfaction are more likely

to offer the core-plus-options or the

mix-and-match plan than are organi-

zations that empha size cost control.

Organizat ions emphasizing cost -containment

objectives, however, are unlikely to offer a lot of

choices . Ma nagem ent can m inimize adm inis t ra-

tive costs by offering the salary-reduction-only

or modular-options plans, which offer less flex-

ibility and, therefore, are less costly to imple-

ment, commun icate, and m an ag e (Beam & Mc-

Fa dde n, 1996). The se organ izati ons m ight a lso

be less likely to offer flexible spending ac-

counts, because al lowing employees to pay in-

surance premiums and uncovered medical ex-

penses with pretax dollars essential ly reduces

the cost to workers of medical care and, hence,

their incentive to use it carefully. Thus:

Proposition

7: Organizations whose

primary benefits objective, at the time

of flexible-ben efit-plan ado ption, is

cost control are more likely to offer the

modu lar-options or the salary-reduc-

tion-only plan than are organizations

that empha size employee satisfaction.

The Relative Strength of Institutional Pressures

and Expected Efficiency Gains

Although institutio nal an d efficiency per-

spectives offer com plem entary r at ion ales for de-

cisions about flexible benefit plans, these two

factors also may generate confl ict ing pressures

on managers. According to those holding the

inst i tu t ional v iew, organizat ions may of ten

adopt structures that do not improve efficiency

but do conform to preva iling prac tice (Tolbert &

Zucker, 1996). We sug ge st that the rela t ive

strength of the two factors depends on the fol-

lowing three condit ions.

1)

T he

stage of institutionalization

of flexible

practices affects the strength of inst i tut ional

pressures . During the preinst i tu t ional izat ion

stage, when few organizat ions have adopted


15/21

1998

Bairinger

and

Milkovich

319

these p lans , p ressures to adopt and design in

specific ways

are

weak. Hence, decisions

to do

so depend on economic and technical iss ues. As

more organizations adopt, the p ressu res to con-

iorm

to

preva il ing practices beco me stronger.

According

to

this logic,

the

econom ic eificiency

rationales become less important or, maybe, are

a s s u m e d

by

m a n a g e r s .

( If

these leading

com-

p a n i e s are doing it, so should I. )

(2) Organizations may vary in their desire t

resist the pressures . That is, if man agers expec t

flexible benefit plans to generate minimal , or

nega tive, efficiency gain s, they may d ecide they

should try to resist any p ressu res to adopt or

design in a cer tai n w ay (O liver, 1991). W her e the

p l a n s

are too

costly—and

the

organization

is

trying to contain costs—or are u n a p p e a l i n g to

employees—and

the

organization

is

trying

to

en han ce benefi ts objectives—the de sire to resist

is likely to be high.

{3) Or gan iza tions vary

in

their

ability

t

resist

ins t i tu t ional pressures . According to inst i tu-

t ional and resource dep end enc e theorists, differ-

ences

in

ability

to

resist

are

related

to

differ-

e n c e s in o rgan iza t iona l i n t e rco nnec ted ness ,

uncertainty about goals or technology, and or-

ga niz atio na l size (Gree ning & Gra y, 1994).

/nferconnecfedness, or professionalization oi

m a n a g e r i a l staff, is the density oi interorgani-

zat ional re lat ions among occu pants

of an

organ-

iza tio na l field {Oliver, 1991: 171). The idea is

that organizat ions managed pr imari ly by pro-

fessionals,

or

others be longing

to

interorganiza-

t ional groups, wil l be influenced h eav ily by

s tandard-set t ing academic ins t i tu t ions or pro-

fessional networks

and,

therefore, w ill tend

to

adopt the preva il ing HR practice s in their field.

In these ci rcum stances the collective s treng th of

the pressures makes them more difficult to resist

{Oliver, 1991). Interconnectedness is likely to be

observed in industries that have a lot oi busi-

ness ,

t rade,

or

professional organization s (Good-

ste in, 1994; Ing ra m & Sim ons , 1995), su ch as the

service (e.g., cons ulting or he alth care) and high-

technology industries. Professional organiza-

t ions, such as the American Com pensat ion As-

sociat ion or the Society oi Human Resource

Mana gement , tend to empha size leading-edge

pract ices . Benchmarking o ther organizat ions '

best practic es is another example of how in-

terconnectedness

is

manifested

in

practice.

Uncertainty about goals

or

technologies

the-

oret ical ly induces organizat ions

to

imi t a t e

the

prac t i ces of other organizat ions that are per-

ce ived as succe ssful (DiMaggio Powell, 1983;

Oliver, 1991; Zucker, 1987). Or ga niz ati on s in

high- technology indust r ies or hea l th ca re , for

example, o f ten operate in highly uncer tain en-

vironm ents . F inal ly ,

the

s c a l e

and

resour ces of

an o rgan iza t ion , and hence its abi l i ty to res is t

in s t i t u t iona l p ressu res , are re l a t ed to its size

(Green ing & Gray , 1994; Pfeiier & Sa lanc ik ,

1978).

Therefore, we offer the following:

Proposition

8: The

relationship

be-

tween efficiency-related factors ben-

efits objectives, labor market condi-

tions, nature

of the

work, current

bene fits c osts, flexible plan costs, and

employee preferences)

and

manage-

ment s decisions about flexible plans

will be stronger during the preinstitu-

tionalization than the semi- and full-

institutionalization stages.

Proposition 9: The relationship be-

tween institutional pressures and

mana gement s flexible benefit deci-

sions will be stronger when conform-

ity with the pressures (e.g., to adopt

flexible benefits) does

not

conflict

with cost-containment or benefits sat-

isfaction objectives.

Proposition 10: The relationship be-

tween institutional pressures

and

mana gement s flexible benefit deci-

sions will be stronger am ong small or-

ganizations or organizations

in

which

uncertainty or interconnectedness is

high.

IMPLIC TIONS ND CON CLUS IONS

Our purpose in this article has been to ana-

lyze and integrate four theoretical explana-

tions oi why organizations adopt flexible ben-

eiits and why such variety in plan designs

exists. We have generalized the theories into

two primary ioci—institutional pressures and

expected efficiency gains—that influence both

adoption and design decisions. As our inte-

grated model demonstrates, further analysis

of these decisions through the iour theoretic

lenses suggests sets of specific variables that

must be included when studying manage-


16/21

3

Academy of anagement Review

April

ment's flexible benefit decisions. Institutional

pressures are contingent upon competitors'

pressures and such coercive forces as legisla-

tion and unions. The expected eificiency gains

are a function of plan objectives, which are, in

turn, affected by labor market conditions, the

nature of the work, employee preferences, and

costs. The relative strength of institutional

versus efficiency factors is moderated by the

stage of institutionalization, as well as factors

related to the organization's resistance to in-

stitutional pressures.

Application to Other HR Innovations

Generalizing the insights gained from ex-

ploring a specific HR innovation— îlexible

benefits—to questions surrounding decisions

about other innovations remains to be done.

We believe that our analysis oi the iour theo-

ries and related research provides a useiul

framework ior such eiiorts. Further, we submit

that our integrated model oifers insights into

the decisions management makes about other

HR innovations. Our model may have limited

generalizability when it comes to some of the

moderating factors, such as workforce diver-

sity, which may be unique to flexible benefits;

however, even these variables oifer a platform

for future work.

Consider the case oi self-managed work

teams. Our model implies that explanations

for observed patterns oi team implementa-

tions must include an examination oi the in-

stitutional pressures related to teams. Consid-

ering first competitors' practices, we find that

adoptions oi this innovative practice have, in

recent years, increased substantially among

manuiacturing iirms (Banker, Field, Schroe-

der, Sinha, 1996; Cohen, Ledford, Spreitzer,

1996; Osterman, 1994). This trend suggests that

competitors' practices in the manuiacturing

sector will generate strong institutional pres-

sures favoring the adoption of the team ap-

proach. Hence, new team implementations are

likely to be higher in this sector than in other

sectors, where the approach is less prevalent.

At the same time, many unions oppose the

adoption of teams, which they view as under-

mining the role oi unions in representing work-

ers {Kochan, Katz, McKersie, 1994). The Na-

tional Labor Relations Act, which prohibits

organizations irom forming company unions,

has recently been invoked by the National Labor

Relations Board (NLRB) to bar certain types of

teams in nonunion settings {Flynn, 1996). Conse-

quently, in settings that are currently unionized

and/or in settings where unions are attempting

to organize workers, coercive legislative and

union forces generate institutional pressures on

organizations ot to adopt a team approach.

Variability in the strength of these pressures

may help explain adoption patterns over time.

Soon after the NLRB issued its first ruling

against the use of teams, many organizations

backed away irom employee involvement initi-

atives or modified existing teams to be in com-

pliance with the ruling (Flynn, 1996). However, a

bill recently introduced in the

U.S.

Congress pro-

poses loosening the restrictions on teams (the

Teamwork for Employees and Managers Act),

reflecting a current political climate that favors

the use oi innovative approaches to work design

(DeWine, 1997).

Decisions about structures such as teams

will also, according to our model, be inilu-

enced by the efficiency gains organizations

can expect to realize by adopting them. In-

deed, survey evidence indicates that organi-

zations implementing teams do so with the

hope of achieving gains in productivity, qual-

ity, costs, and employee satisfaction {Lawler,

Mohrman, Ledford, 1992). A key determinant

of expected efficiency gains in the case oi

flexible beneiif plans is benefits objectives.

This factor is clearly not relevant to decisions

about teams; however, a closely related fac-

tor—HR objectives—might prove a useful sub-

stitute. Whether an organization's primary ob-

jective in implementing teams is reducing

production costs or enhancing employee sat-

isfaction will be influenced by the nature of

the work and labor market conditions, as pos-

ited in our above discussion of the model. Or-

ganizations depending on employees with

specialized skills that are in short supply in

the labor market, for example, will tend to

emphasize employee satisfaction.

As our integrated model suggests, the impact

oi objectives on organizations' expectations

about eiiiciency gains will depend on moderat-

ing facfors. In the case oi flexible benefits, these

include costs and employee preferences. Em-

ployee preferences ior flexible benefits are the-

oretically related to workiorce diversity and ed-

ucation. Employee preierences will also affect


17/21

1998

aiiinger

nd

Milkovich

321

the capacity of teams to achieve objectives;

however, the literature suggests a different set

oi indicators. Hackman and Oldham (1980), for

example, suggest that a team design will not

lead to performance and satisfaction improve-

ments ii workers do not have a strong growth

need. Lawler (1986) argues that needs for

growth, achievement, competence, and social

interaction are important. Contextual factors

also will aifect the extent to which teams can be

expected to help achieve HR objectives. Specif-

ically, teams may be most successful in contexts

where production tasks are interdependent

and/or in greenfield sites (Campion, Medsker,

Higgs, 1993; Lawler, 1986).

As with flexible benefits, the relative influ-

ence oi institutional pressures versus ex-

pected eificiency gains on management's de-

cisions about teams will depend on the stage

of institutionalization as well as the organiza-

tion's ability and/or desire to resist institu-

tional pressures. Factors related to the eiii-

ciency gains organizations can expect to

achieve by adopting teams will exert more of

an influence during the preinstitutionalization

stage. During the semi- and iull-institutional-

ization stages, organizational uncertainty, in-

terconnectedness, and size will aifect the or-

ganization's ability to resist pressures to

adopt {or not adopt) teams. Finally, the desire

to resist these pressures in the later stages of

institutionalization will depend on whether

the adoption oi teams is consistent with the

organization's objectives. Organizations with

a long tradition of individual-based work have

found that the introduction oi teams can lead

to employee discontent, making the achieve-

ment oi employee satisiaction objectives difii-

cult (Lawler, 1986).

The above discussion suggests that we can

apply many of the factors posited to aiiect deci-

sions about flexible benefit plans to decisions

about other

HR

innovations. The applicability oi

institutional pressures, expected efficiency

gains, labor market conditions, employee prei-

orences, and the nature of the work appears to

gQ beyond the specific case of flexible benefits.

Other factors may vary somewhat across con-

texts. Benefits objectives are only an appropri-

ate factor when one is considering benefits in-

novations. Nonetheless, there are a number oi

such innovations for which this factor is appro-

priate—ior example, work-iamily benefits or

wellness programs. Further, we suggest that a

closely related iactor—HR objectives—can be

substituted easily into the model. Although we

suggest that employee preferences moderate

the effect of objectives on expected efficiency

gains, we also argue that the relevant indicators

oi preierences vary across innovations. Finally,

we suggest that it may be appropriate to con-

sider moderators other than costs and employee

preferences when explaining other types of in-

novative practices (e.g., employees' growth

needs in the case of teams).

Implicat ions or Theory and esearch

The integrated model we propose represents

the first attempt, that we can iind, at synthe-

sizing the implications of these four leading

theories for decisions about organization

structures. As we have noted, others have in-

tegrated two theories (e.g., institutional and

resource dependence, and institutional and

agency), but ours is the first to draw insights

irom all iour. Arguably, integrating four theo-

ries to better understand organizational be-

havior yields a more parsimonious model. Em-

pirical studies oi our model will provide tests

of its efficacy in explaining management de-

cisions about flexible benefits. We have noted

that very little is known about flexible benefit

decisions at the organizational level of analy-

sis.

Our intent is to support organization schol-

ars in the use oi multiple theories in their

study oi employee benefits and, by implica-

tion, other HR innovations.

We should note here that many oi the explan-

atory factors included in our model may be en-

dogenous. For example, the model implies that

workforce diversity is a determinant of the deci-

sion to adopt flexible plans. But it may also be

true that the presence oi flexible plans affects

the diversity of workers who seek employment

at an organization. Similarly, the resource de-

pendence, agency, and transaction cost models

all suggest that inadequate information about

work effort will cause organizations to increase

the value of their total compensation packages.

At the same time, the level of overall compen-

sation may influence employees' willingness to

share iniormation about their work eiiort. Empir-

ically sound tests of the model, therefore, may

require the estimation oi simultaneous equa-

tions.


18/21

Academy

of

Management Review

April

Implications

for ractice

Although our purposes in this article do not

include an applied iocus, we believe that man-

agers may also gain insight irom our model

and analysis. The model clearly delineates

the principal iactors that leading theories

specify influence whether and why types of

flexible plans are adopted. It iocuses manag-

ers'

attention on both institutional reasons

{based on benchmarking, following competi-

tors,

and regulatory pressures), as well as ex-

pected economic gains. Experience suggests

that although the rhetoric around adopting

many HR innovations is expressed in terms of

expected economic gains, the reality may be

that government regulations, as well as fads

and iashions, prevail {Dyer Reeves, 1995).

Our model presents, for the first time and in

one integrated guide, very specific factors

managers could consider when faced with de-

cisions about whether or not to adopt ilexible

plans or what type oi plan design to imple-

ment (Rosenbloom, 1996).

R F R N S

Baikin, D. B., Bannister, B. D .

1993.

Explaining pay iorms for

strategic employee groups in organizations: A resource

dependence perspective . Journal

o/

Occupational

and

Organizational Psychology 66:

139-151.

Banker, R. D., Field. J. M., Schroeder , R. G.,

Sinha, K. K. 1996.

Impact of work team s on manufacturing performance: A

longitudinal field study.

Academy of M anagement

Jour-

na l

39: 867-890.

Barber, A. E., Dunham, R. B., Formisano, R. A. 1992. The

impact of flexible benefits on employee satisfaction: A

field study. Personnel Psychology. 45: 55-75.

Barringer, M. W., 8i Milkovich, G. T. 1996. Employee health

insurance decisions in

a

flexible benefits en vironm ent.

Human Resource Management 35: 293-315.

Barringer, M. W.,

Mitchell, O. S. 1994. Work ers' preferen ces

among company-provided health insurance plans. In-

dustrial and Labor Relations Review 48: 141-152.

Bartol, K. M..

Martin, D. C. 1988. Influen ces on m anager ial

pay allocations: A dependen cy perspective. Personnei

Psychology

41: 361-378.

Beam, B. T., McFadden, J. J. 1996. Employee

benefits.

Chi-

cago: Dearborn Financial Publishing.

Bloom, D. E., Trahan, J. T. 1986. Flexible benefits and

employee choice; Highlights oi the liteTature. New York:

Pergamon Press.

Borleis, M. W. 1996. Cafet eria pla ns

in

operation.

In J. S.

Rosenbloom (Ed.),

The handbook of employee benefits:

505-540. Chicago : Irwin.

Cable , D. M., St Judge, T. A. 1994. Pay preierences and job

search decisions: A person-organization fit perspective.

Personnel Psychology 47: 317-348.

Calvo, G. A. 1987. The economics

of

supervision. In H.

R.

Nalbantian (Ed.), Incentives coopera tion and risk shar-

ing: 3-43. Totowa, NJ: Rowman and Liltlefield.

Cam pion , M. A., Medsker, G. J.,

Higgs, A. C.

1993.

Relations

between work group characteristics and effectiveness:

Implications for desig ning effective work groups. Per-

sonnel Psychology

46: 823-850.

Cohen, S. G., Ledford. G. E., Jr., Spreitzer. G. M. 1996. A

predictive model of self-managing work team effective-

ness . Human Relations 49: 643-676.

Davis-Blake, A., Uzzi, B.

1993.

Determinants of employment

externalization: A study of temporary workers and inde-

pendent contractors. Administrative Science Quarterly,

38: 195-223.

DeWine, M. 1997. Roll Call June 16: 10.

DiMaggio, P. 1988. Interest and agency in institutional the-

ory. In L. G. Zucker (Ed.), Institutional patterns and or-

ganizations: Culture and environment. Cam bridge, MA:

Baiiinger.

Documents

A THEORETICAL EXPLORATION of the Adoption and Design of Benefit Plans