THE DETERRENT EFFECT OF PERCEIVED CERTAINTY AND SEVERITY OF PUNISHMENT REVISITED*

• Author: DANIEL NAGIN,STEVEN KLEPPER
• DOI: http://doi.org/10.1111/j.1745-9125.1989.tb01052.x
• Date: 01 November 1989
• Published date: 01 November 1989

THE DETERRENT EFFECT OF PERCEIVED

CERTAINTY

AND

SEVERITY

OF

PUNISHMENT REVISITED*

STEVEN KLEPPER

DANIEL

NAGIN

Carnegie

Mellon

University

This paper synthesizes and extends recent criticisms of cross-sectional

and panel studies ofperceptual deterrence and then uses those criticisms to

design a better cross-sectional study.

A

series of scenarios involving the

noncompliance

of

a hypothetical plumber were posed to a sample of

predominantly middle-aged administrators. The respondents were asked

to estimate the probability that the noncompliance would be detected and,

if

detected, the probability that

it

would result in criminal prosecution.

Respondents were also asked to project their own behavior

if

in the

plumber's position. The results contrast sharply with prior findings, espe-

cially of the panel studies. Not only do perceptions of detection risk play

an important role in deterring tax noncompliance, but fear of criminal

prosecution appears to be a powerful deterrent.

The deterrence literature has proceeded through three distinct methodolog-

ical waves distinguished by the type of data used to measure criminal involve-

ment, detection risk, and severity. The first wave

of

studies, pioneered by

Gibbs (1968) and Tittle (1969), relied on aggregate data. These studies

examined the relationship across geographical areas of the crime rate and

measures of the fraction of criminals apprehended and the severity of the

punishment

of

apprehended criminals.' The next wave of studies turned to

surveys

to

ask individuals about perceptions of detection risk and sanction

severity and also about past or intended future criminal involvement. Like

the aggregate studies, these studies generally suggested that certainty but not

severity

of

punishment deterred crime (Paternoster, 1987). The most recent

wave of studies use panel data from repeated individual surveys. In contrast

to the earlier two waves of studies, these studies generally find that neither

certainty nor severity of formal sanctions is an important deterrent to crime,

but that informal sanctions, socialization, and moral considerations are key

determinants of criminal activity. These findings call into question the effi-

cacy of conventional forms of punishment in deterring crime.

*

We thank Raymond Paternoster, Harold Grasmick, and three anonymous referees

for helpful comments. This material

is

based on work supported by the National Science

Foundation and the Internal Revenue Service, Grant

SES

8709573.

For reviews

of

this literature, see Cook (1980), Gibbs (1979, and Nagin (1978).

1.

CRIMINOLOGY

VOLUME

27

NUMBER

4

1989

721

722

KLEPPER

AND NAGIN

The last two waves of studies arose in response to methodological critiques

of the prior wave. The cross-sectional survey studies were spawned by criti-

cism that earlier studies relying on aggregate data failed to capture individual

perceptions of sanction certainty and severity (Waldo and Chiricos,

1972).

The panel studies in turn were prompted by the possibility that the link

between detection risk and prior behavior found in cross-sectional survey

studies reflected the effect of past criminal involvement on perceptions rather

than the reverse deterrent effect (Bishop,

1984;

Paternoster et al.,

1983;

Saltz-

man et al.,

1982).

Like its predecessors, the panel studies are now being sub-

jected to methodological criticism (Lundman,

1986;

Murray and Erickson,

1987;

Williams and Hawkins,

1986).

Indeed, proponents of cross-sectional

survey studies believe that if properly designed, cross-sectional studies can

provide at least as reliable a basis for making inferences about deterrence as

panel studies (Lundman,

1986;

Murray and Erickson,

1987).

The main purposes of this paper are to synthesize and extend the criticisms

of the panel studies and to use those criticisms to design a better cross-sec-

tional study. Using a scenario format to elicit perceptions about detection

risk and severity of punishment, we explore the determinants of intentions to

engage in various forms of tax noncompliance among a sample of predomi-

nantly middle-aged administrators. While the study was not initially con-

ceived as a vehicle for overcoming criticisms of both cross-sectional and panel

perceptual deterrence studies, we believe that the findings and the methodol-

ogy used are a first step toward addressing those criticisms.

Our approach yields results that contrast sharply with prior findings, espe-

cially of the panel studies. Not only do we find that perceptions of detection

risk play an important role in deterring tax noncompliance, but we also find

that for most of our respondents the fear of criminal prosecution appears to

be a powerful deterrent to tax noncompliance. At least within the domain of

tax noncompliance, this suggests that the severity of punishment plays an

important role in deterring law breaking.

CRITICISMS

OF

THE PERCEPTUAL DETERRENCE

LITERATURE

Recently, several critiques of the panel studies have appeared. Some of the

issues raised in these critiques apply as well to the prior wave of perceptual

cross-sectional studies. The issues that are most relevant to our inquiry are:

(1)

inadequate attention to the full range of the preventive effects of punish-

ment,

(2)

unstable perceptions of sanction risks, and

(3)

limits to the general-

ity of the findings.

PREVENTIVE EFFECTS

OF

PUNISHMENT

Among the most important

of

the substantive criticisms of the perceptual