POETICS

ELSEVIER Poetics 28 (2001) 241-254 ii i

www.elsevier:nl/locate/poetic

Artists as workers:
Theoretical and methodological challenges *

Pierre-Michel Menger* _.
Centre de Sociologie des Arts. EHESS CNRS, Paris, France

Abstract

Artistic labor markets expand along a path of highly unbalanced growth: competitive pres-
sure, flexible specialization of the work organization and pervasive work contingency cross
individual as well as entrepreneurial ideals such as self-achievement and innovativeness in
ways that challenge conventional views of the skilled working process as well as the conven-
tional survey and measurement methods.
High differentiation of artistic products and steady oversupply, which are common traits of
an imperfect monopolistic competition, are magnified by work arrangements that evolve
towards increasingly fragmented and brokered employment relationship. Work trajectories com-
bine features from professional as well as from entrepreneurial careers, under constraining con-
tingency. Yet, uncertainty of the creative process, that goes along with the 'functional flexibil-
ity' requirement in the arts, helps to explain artistic behavior: neither a rational actor nor a
deterministically driven agent, the artist may be depicted as a bayesian actor learning to balance
self-actualization against occupational risk. © 2001 Elsevier Science B.V. All rights reserved.

1.1ntroducfion

W h e n one tries to summarize s o m e o f the findings o f numerous studies on artis-

tic labor markets (for a review and for the detailed references, see Menger, 1999),
the following picture e m e r g e s :

Artists as an occupational group are on average younger than the general work force, are better educated,
tend to be more concentrated in a few metropolitan areas, show higher rates of self-employment, higher

* I would like to thank Joan Jeffri for helpful remarks and suggestions, and the participants to the
working conference 'Research on Artists', organized by Paul DiMaggio (Princeton University) and Joan
Jeffri (Columbia University) and held in Princeton in May 2000, for comments on an earlier version of
this paper.
* E-mail: [email protected]

0304-422X/01t$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S0304-422X(00)00027-9
242 P.-M. Menger / Poetics 28 (2001) 241-254

rates of unemploymentand of several forms of constrained underemployment(non-voluntarypart-time

work, intermittentwork, fewer hours of work), and are more often multiplejob holders, They earn less
than workers in their referenceoccupationalcategory,that of professional,technical and kindred work-
ers, whose members have comparable human capital characteristics (education, training and age), and
they experience larger income inequalityand variability.

This portrait seems crudely unattractive. Yet artistic labor markets are steadily
expanding, and growing numbers of new candidates to fame or to self achievement
enter the market. I will not explain the trend by exploring the causal factors of the
growing demand for art and culture. Rather, I will highlight some points of theoret-
ical and methodological interest. Mainly, these points stem from the basic observa-
tion that artistic labor markets expand along a path of highly unbalanced growth:
competitive pressure, flexible specialization of the work organization and pervasive
work contingency cross individual as well as entrepreneurial ideals such as
self-achievement and innovativeness in ways that challenge conventional views of
the skilled working process as well as the conventional survey and measurement
methods.

2. Labor supply and labor demand: Aggregated and disaggregated levels Of

analysis

Let me begin by pointing to discrepancies between the supply and the demand
approach of artistic labor markets.
From the labor supply standpoint, one should say: one individual equals one
career, however successful that career might be. But there is a gap between the voca-
tional commitment and the way it transforms into work and it results in a career.
Self-employment, freelancing and contingent work, which are the 'prevailing work
arrangements in the arts, have quite similar effects: they bring into the picture dis-
continuity, repeated alternation between work, compensated unemployment, non
compensated unemployment, searching and networking activities, cycling between
mutliple jobs inside the arts sphere or across several sectors related or unrelated to
the arts. Therefore, statistics on numbers of artists, such as those from Census sur-
veys or from professional guilds or unions sources, have to be considered cautiously
as indicators of the artistic labor supply and, accordingly, of the actual level of
unemployment.
From the labor demand standpoint, things seem simpler: the focus is on contracts,
hirings or with works sold on the market. Thus, in contrast to the aggregate view of
art as an occupational world, the labor market here is investigated at its most disag-
gregated level, that of the series of hirings, of work opportunities and of bargaining
relationships. These represent an individual artist's working life in a given period;
when considered longitudinally, they display a career trajectory. Disaggregated data
lead to statistics on the amount of labor demand expressed in terms of quantities and
prices of work.
If one brings supply and demand together, the main issue is: what is the impact of
the fact that labour demand is expressed mainly in terms of contingent work?
 P.-M. Menger / Poetics 28 (2001) 241-254 243

Numerous studies have shown that an increase in the number of artists as reported in
the Census data may be far from corresponding to a similar increase in the level of
activity, since the former trend may have different and contradictory meanings. It
there is more work but an ever more rapidly growing number of individuals, a fiercer
competition takes place that implies higher inequalities in the access to employment,
more variability in the level and schedule of activity and, on the whole, work
rationing for those who share the labor pie and cycle more often from work to unem-
ployment or from arts work to arts-related or non-arts work.
Results from research on the French performing arts labor market may highlight
these discrepancies. The data reported in the Figs. 1 to 3 are from the Caisse des
Cong6s Spectacles, a fund that collects employment records from the artists and
other technical and administrative personnel employed as intermittent workers in the
performing arts in France - nearly 90% of the labor force in that sector. Records
come from each individual hiring, with mention of the working time and the amount
of fees. These records are registered in order to provide the personnel with monetary
compensation for the vacation time they are entitled to claim, according to the legal
rules governing contingent employment.

450

400

350

300
m

250
|

~ 150

100
~4~Totd ~ n b e r of worked a~yl dec/arid by artplts

50 ~ N u m b e r of artists v~rking
I - ;X-" -TOI~ amount of eltmi.l~

0 I I I I I ( I f 'l i'
1966 1987 7988 7909 1990 ! 9g 1 1992 ~993 1994 19~5 ~996

Fig. 1. Trends in the French performing arts labor market: 1986--1997.

Disaggregated data.

As shown in Fig. I, the French labor market for the performing arts has constantl3
expanded over the period 1986--1997. However, the trend of the supply of work (th~
number of artists working) and the demand side trend (the number of hirings, th~
number of worked days declared and the total amount of earnings) evolved at rate,
of increase that differ significantly. As a result, displayed in Fig. 2, the mediar
 244 P.-M. Menger / Poetics 28 (2001) 241-254

160

140

120

40

- " 0 " - I ~ l l ~ wn~tmt of ~ q a

~ p,w Irt~t I
i
20 .--li---M~ltln number of hlrlngl per attiC. ]

0 ,i ', I I I' i' I , I I I I I

1986 1987 1988 1989 1990 ~991 1992 1993 1994 1995 1996 1997

Fig. 2. Trends in the French performingarts labor market: 1986-1997.

Median values (1).

160

140

120

40

* ~edim kmith'Qfhlrm~ [
20 ,'...41---Meellmitmomem workeddl~

0 ' I I' I I I I' ' I I '" I .......... I ....... l

1986 198? 19a8 1989 1990 199? 11992 1993 19114 1995 1996 1997

Fig. 3. Trends in the French performing arts labor market: 1986-1997.

Median values (2).
 P.-M. Menger / Poetics 28 (2001) 241-254 245

amount of working time and earnings per artist (unadjusted for inflation) decreased
over the period, although the number of hirings increased: individual intermittent
work was increasingly fragmented in shorter hirings, and competition turned out to
become fiercer among the growing numbers of artists sharing the less rapidly grow-
ing 'work pie'. The last figure illustrates the way people are partly compensated for
the increasing risk that goes along with the shortening of individual hirings. On the
whole, however, the decrease of median earnings over the period indicates that
employers do not insure the artists they choose to hire under such a working scheme
against the consequences of the unbalanced growth of that labor market: employers
and consumers may benefit from the increasing variety of talents supplying their
work, at the expense of the increasing level of unemployment or underemployment,
or at least at the expense of increasing variability in the individual working situa-
tions, both across the workforce and during the career of each artist.
The overall result is that of an increasingly wider gap between the trends depict-
ing the evolution of each side of the market.
For several years, theoretical as well as empirical work on artistic labor supply has
been remarkably pioneering (Throsby, 1994). But it had to be supplemented more
and more with no less pioneering studies on the demand side. For the supply side
studies may be hampered by several important limitations as to the way artists are
identified and information on socio-demographic and economic characteristics is col-
lected, if only available through unbiased declarations. Of course, such studies exist
more and more, and most of us have done some, with varying levels of satisfaction.
Ideally, one should follow true cohorts of artists through longitudinal surveys to
get reliable and valuable informations on the main issues of core interest which I will
review i n n moment. Since cross-sectional data provide us with a set of snapshots
drawn from individual courses of life that very often are everything but regular and
predictable, we may find some devices to produce pseudo-longitudinal information
(retrospective appraisal of the course of a career with its main steps and tuming
points). To be sure, these are rather imperfect devices when one tries to get the
slightest idea of the variability and complexity of a working life in the arts. That's
why such sociographic investigations have lacked substance, although they would
claim to play a revealing role in setting the vocational commitment to the arts in a
simple deterministic frame of social class causation, against idealistic celebration of
the irresistible inner drive to the arts.

3. Data and surveys on artistic contingent work

The labor demand study has specific characteristics in the arts. Concerning
employment and Wage structure, information focuses less on individual firm policies
than on industry wide work agreements and institutions. One unexpected result of
the early development of contingent work in the arts is that We have become expert
at identifying employment and earnings records over time and at using completely
disaggregated data to identify a number of characteristics of each individual job
worker throughout his or her career.
246 P.-M. Menger / Poetics 28 (2001) 241-254

As these sources were better known and became available, new research opportu-
nities have appeared. The research done by Christopherson and Storper (1989) on
data provided by the Motion Picture Health and Welfare Fund uses records on work-
ers with multiple employers. More recently the research done by Bielby and Bielby
(1999) on data from the Writers Guild of America does the same. In France we have
completed research from the same kind of sources (Debeauvais et al., 1997; Menger,
1997), and in Canada, we have collaborated with Benoit Laplante using more sophis-
ticated statistical methods - mainly random effects logistic regression for panel data,
rather than duration analysis or structural equation models (Laplante and Htbert,
2000).
The statistical information taken from such sources may fit the purpose of build-
ing and testing models of job allocation, of earnings prospects, of segmentation of
the workforce along several independent (sociodemographic characteristics) and
dependent variables. These dependent variables include the work and income track
records, portfolios of contractual recurrent and nonrecurrent employment ties and the
corresponding network structure of employment relationships. Access to such data is
presumably the main way we may expect to model and understand the building of a
career through a discontinuous line of engagements.
Issues like the workforce segmentation can be studied in fine detail and carefully
modelled as we rely not only on sociodemographic variables depicting initial endow-
ments in human capital resources, but also on the mechanics of work allocation and
of reputation building, that is on working and bargaining relationships with employ-
ers, with colleagues and with clients whose market characteristics can be known and
put in the model. The human capital framework turns out to fit unequally well for
catching this second dimension: to mention only one item, the age-earnings profile
may vary considerably according to the art form and to the sector considered. Thus
work experience may have a positive value only over a short period of time, as in the
case of screenwriters studied by Bielby and Bielby (1999), or may display a non-typ-
ical profile of evolution, when young artists are much in demand, then experience a
more or less long eclipse, before a possible (and often carefully managed) reap-
praisal of their work and reputation in the later part of their creative career. But
things are even more complicated since work experience accumulation may have a
positive impact on the odds of managing the portfolio of multiple jobs, for instance
on the process of switching from one main occupation to another related one, using
human and social capital accumulated.

4. Methodological challenges

The peculiarities of the artistic labor markets I mentioned, therefore, have all a
strong impact on the use of customary Census indicators on the occupation as a
whole. And of course, they do challenge the normal design of quantitative research
on occupations.
For instance, extensive job and sectoral mobility as well as multiple j o b holding
considerably affect the use of conventional work and unemployment indicators.
 P.-M. Menger / Poetics 28 (2001) 241-254 247

Unemployment rates may be mismeasured for several related reasons: individuals

with artistic occupations may switch temporarily to work mainly in non-artistic
occupations when unable to make a living in their primary vocational field, without
stopping to produce art works. They would therefore not be classified as unem-
ployed in their artistic occupation if they are primarily engaged in non artistic work
during the Census week.
Thus there might be little sense to state that artists supply a unique labor market
and hold a unique occupational identity. Not only do artists often supply several
related and intricate labor markets, via their multiple job holding practice (Throsby,
1996), but they may also cumulate several occupational roles within their artworld -
e.g. creators, performers, teachers, managers, entrepreneurs.
Methodologically, the most successful research design I have tried out was that of
combining work on micro-data from professional sources and results from a specifi-
cally designed survey, i.e. a survey design based on sampling stratification variables
including the amount of work, earnings and sectoral diversification of artistic work,
as they are documented in the personnel records collected on a sector or industry-
wide basis (mainly by guilds, unions or pension funds).
Thus one can take advantage of the survey methods which remain invaluable as
soon as one has to explore basic dimensions like social origin, early familiarization
to culture and the arts, general and specific learning (these are the basic human cap-
ital variables), but also to survey some highly specific issues like multiple job hold-
ing, occupational role versatility, work schedule, cycling behaviour between differ-
ent jobs and between work, unemployment and nonworking time, and more
generally on the patterns of use of time, portfolio of resources, monetary and non-
monetary ratings of the work situation, the meanings of work, the nexus of ties that
links an artist to his counterparts and the to different actors of his occupational world
(agents, dealers, critics, support personnel), the family context, the way of life, and
so on.
To take just one or two core issues, I would never have suspected how complex
the training process of an actor can be had I not investigated it carefully in my sur-
vey research on actors. And the same is true of the sectoral differences in work orga-
nization and in hiring processes (brokered or not), and of the proportion of income
drawn from the vocational occupation.
Of course, had I not had the information on earnings from a non survey source
(the records from the main French performing arts pension fund), I would not have
been able to control the reliability of some information about amount of work and
income and to adjust for sampling biases.
It is important at this point to consider some key issues as to the overall structure
of artistic labor markets and the idiosyncrasies of artistic work.

5. Labor markets under imperfect monopolistic competition

The artistic labor markets seem to correspond to the spot market model of text-
book economics.
248 P.-M. Menger / Poetics 28 (2001) 241-254

But it is in fact a rather paradoxical competitive market: on one hand, employ-

ment is more and more contingent, as we used to see for secondary labor markets,
but on the other hand individuals are highly skilled and nonsubstituability is a core
value, as in the so-called primary labor market. At the same time, segmentation of
the workforce may be so strong that in fact the seemingly old-fashioned distinction
between primary and secondary markets remains somewhat relevant.
Things might be better conceived in terms of a monopolistic competition with
some of its essential features: high differentiation of products (goods and labor ser-
vices, that is works and performances) and a large variety of consumer preferences
(Lancaster, 1979). The application of a monopolistic competition model to the mar-
ket of art works, which are durable and highly differentiated goods, is well known,
since Moulin's pioneering work on art markets (Moulin, 1987). But one might spec-
ulate about extending the model to the market of inputs in itself, that is to the labor
market. Two core features may sustain that suggestion, Firstly, an artist's reputation
provides him with a temporary monopolistic position on the market, as long as his or
her skills and talents are in demand. Secondly, the artist may be viewed as a
microfirm when attention is given to the supply of the numerous bits of work and
efforts out of which the artist makes a living: the way he acts to supply work in sev-
eral related markets and to perform various occupational roles within his artworld
points to the differentiation of the working profiles that underly the competition
among artists not only for success in the vocational labor market but also for secur-
ing and maintaining a position in the art world taken as a whole.
Possibly a strong objection should be raised: may we apply the same analytic
tools to the different strata of a population of artists since the level of reputation, the
amount of work and the occupation success vary considerably among them? In other
words, shall we consider as monopolistic suppliers of their own work artists who
face a rather inelastic demand curve, due to their (temporary or strongly established)
fame, as well as artists who are induced to act as monopolistic suppliers (i.e, to trade
their human capital and work as freelancers) although their market value and market
power are very poor? In the latter case, artists, either young aspiring ones, or mod-
erately or poorly successful experienced ones, face the risk attached to a self-
employed or quasi self-employed position without benefitting from the rent attached
to the seemingly monopolistic supply of their work.
This has to do with the competitive pressure that characterizes art worlds. As we
know, excess capacity of production is a common trait of imperfect monopolistic
competition, Now, the oversupply argument appears to be a constant feature in the
more or less elaborate views of artistic labor markets.
The organizational characteristics of the arts sphere, as they are magnified by the
trend towards a more disintegrated, more brokered system of production, act as an
inflationary labor supply factor. An increasing number of aspiring young artists are
attracted; barriers to entry are low for organizations as well as for individuals, but a
substantial share of the risks are transferred from the firms to the artists who face a
fiercer competition and more uncertain career prospects.
Here one should think not only in terms of artists' oversupply at an aggregate
level, as a result of growing inter-firm competition in more contestable markets, but
 P.-M. Menger / Poetics 28 (2001) 241-254 249

also of an intra-organizational process designed to deal with an uncertain and turbu-

lent market environment.
Organizational flexibility in the arts indeed plays a major causal role in structural
oversupply. Employers in project-based organizations seek todraw from a large pool
of artists and personnel in order to build efficient and well-matched teams, since they
may gain from the variety of talents and skills at hand, and to reduce overheads.
Similarly, for record companies or book publishers, overproduction of new items,
along with allocation of numerous personnel to boundary-spanning roles and coop-
tation of mass-media gatekeepers, is a rational organizational response to an envi-
ronment of low capital investments and demand uncertainty, especially in the most
speculative and entrepreneurial segments of the market: because of a strategy of dif-
ferential promotion of the numerous items released, the corporate sponsoring is only
focused on a small proportion of them.
More generally, however, the notion of oversupply has to be questioned, since it
refers to a disequilibrium in only one of the labor markets the artists supply, that of
their principal vocational work. As stated above, when multiple job holders cycle
between rationed and less- or un-constrained job markets, or when individual, coop-
erative and collective devices of compensation for and insurance against risk are at
hand, notions of underemployment or oversupply may be hard to apply, provided
that work under such a steady 'management of risk' scheme is ~more attractive than
occupational alternatives outside the arts sphere.

6. The artist: A professional? An entrepreneur? Or just a highly skilled con-

tingent worker?

Several studies, mainly in a socio-historical perspective, have equated the great

artist with a kind of innovative entrepreneur designing not only stylistic innovations
but also new organizational ways of working and distributing his or her work; see
for instance Alpers' study on Rembrandt (1988), or Elias' one on Mozart,(1993), or
Fitzgerald on Picasso (1995). Basically, current individual work trajectories combine
more and more traits from professional as well as from entrepreneurial careers: pro-
fessional careers rely on reputation certified by the community of peers and by the
inner circles of experts and middle men in the art worlds. But artistic careers exhibit
low loyalty to organizations, and artists very often have to operate like small busi-
nesses, by holding multiple jobs and by setting up more often than not companies or
fringe firms. For instance, the number of employers in the performing arts is sur-
prisingly high and so is the turnover.
One striking dimension of the segmentation of the labor force has to do with the
entrepreneurial skills a number of artists acquire on the job. For, although brokerage
has emerged as a major device for mediating labor allocation and matching
processes in a highly fragmented and segmented labor market (DiMaggio, 1977;
Bielby and Bielby, 1999), artists may be induced to exercise supervisory or manage-
rial skills and, in so doing, to blur the line between management and labor (Christo-
pherson, 1996). In that respect, the artist's earnings, like those of any self-employed
250 P.-M. Menger / Poetics 28 (2001) 241-254

worker, depend not only on her skill, talent and effort, but also on how well she
performs the managerial and entrepreneurial functions (Aronson, 1991). Histori-
cal studies mentioned above tell us that this is hardly new, but the way the line is
blurred evolves according to new technical, economic and managerial oppor-
tunities.
Now, conventional forms of occupational identity and occupational careers are
undermined by the current organizational designs of flexible specialization. Contin-
gent employment is expanding rapidly in highly skilled service sectors - law, human
resource management, accounting services, higher education, medicine(Smith,
1997; Cohany, 1998; Hippie, 1998). l In that respect, and ironically enough, the arts
appear to have been forerunners in experiencing the trend toward increasingly flexi-
ble labor markets; I should even say, toward hyperflexible markets: freelancing
means indeed that one may be hired for only two or three hours, without any costly
dismissal procedures (on Hollywood, see Storper, 1989; for a somewhat militating
view of the Hollywoodian flexibility model, see Kanter, 1995). In the arts sphere, the
whole range of work contingency has existed for a long time.
Two striking features of the contingent employment in the arts are of special
interest.
The first feature deals with the way people try to manage their occupational risks.
Which risks are manageable and insurable and which are not, which individually
(through different schemes of diversification of work and income sources), which
collectively (on a professionnal community basis) and which socially (through cul-
tural policy support)? Answers may vary greatly among the different art worlds,
depending on their whole institutional apparatus, but common features exist that
bring to light what special kind of workers artists actually represent: they have to
identify devices to secure as much as possible the supply of a good, their talent, that,
in theory, is infinitely differentiated, but whose evaluation and employment obey in
practice to much cruder schemes of hierarchization like the A-list vs. B-list distinc-
tion (Caves, 2000).
One should add that only some of the occupational risks in the arts are insurable,
since the kind of uncertainty at stake is questionable: is it exogenously determined,
as when a lack of jobs and an unsuccessful career should be attributed to insufficient
demand for the kind of ability with which the artist is endowed? Or is it endoge-
nously embedded, due to the artist's insufficient ability? But ability and talent them-
selves may be ambiguous: 'talent' should be considered not only as an exogenous
factor of market success but also as an endogenous factor shaped by competition
through innovation. The more competition raises the rate of innovation or, at least,
of differentiation between prototype-like works, in exploiting and stimulating con-

' Reporting the results of the second survey the U.S. Bureau of Labor Statistics has done on contingent
work, Hippie (1998: 24) notes that "as in 1995, contingent workers in the 1997 survey were found in
both low- and high-skilled occupations. For example, contingency rates were highest for those in pro-
fessional specialty (6,0 percent [against the 4.4 percent for the whole labor force]), administrative sup-
port (6.0 percen0 and farming occupation (5.9 percent), The fact that the probability of holding a con-
tingent job was relatively high among occupations with such varying skill levels tends to refute the
stereotype that contingent workers are primarily low-skilled".
 P.-M. Menger / Poetics 28 (2001) 241-254 251

sumer demand for novelty, the more the sorting mechanism will be based on shifting
specifications of marketable talent.
The second trait is the apparatus of collective agreements and social institutions
that substitute for the missing role of the normally unique employer, so that creative
industries, surprisingly enough, are heavily unionized. For instance, the three-tier
compensation scheme depicted by Paul and Kleingartner (1994) shows how collec-
tive bargaining contracts and regulatory mechanisms may indicate one possible
direction for unions. Instead of protecting jobs from increasing contingency, unions
may bargain for the payment schemes to adjust to the new opportunities offered by
an enlarged and more diversified distribution of motion pictures and to programs. In
France, unlike in most countries where the Unemployment Insurance is beyond the
reach of the freelancers, intermittent work is equated with a wage-earning position,
in the performing arts, and so allows workers who are eligible (i.e. those who meet
the criterion of a certain amount of work over a given period) to get compensated for
the unemployed periods of time (Menger and Gurgand, 1996). Indeed, the relation-
ships between professional work and organizational versus spot market labor con-
tract settings diversify more than ever, and so does the whole range of individual vs
collective bargaining models over one's employment terms.
One might speculate whether our current sociological models of professionaliza-
tion, even when trying to combine entrepreneurial and professional features of occu-
pational work, may account for contingency in highly skilled work settings and
occupational careers. Obviously, a new research agenda has appeared (Leicht and
Fennell, 1997).

7. Contingency and uncertainty

On the one hand, increasing flexibility goes along with increasing differentiation in
cultural production: as a result, more and more of the occupational risks are transferred
down onto artists. Employment in the arts is increasingly on a short-term contract or
freelance basis, which magnifies the shift towards numerical flexibility observed else-
where in the economy (Smith, 1997). Several factors play a major role in boosting con-
tingency in the cultural sphere - e.g. the dominance of the single-project organizational
scheme, the development of a brokerage system (DiMaggio, 1977) that enhances the
role of talent agencies in mediating the labor market for contingent employment
(Bielby and Bielby, 1999), the segmentation of the artistic labor force according to
the reputation level of each of its members within each class of occupational roles,
which may explain why even artists may be or become much more substitutable
workers, although they claim to be endowed with a unique set of skills, talents and
abilities that command their - volatile - market value. Thus the attractiveness of artis-
tic occupations may be exploited to secure a pool of underemployed personnel form-
ing a 'reserve army' and to allocate work on a numerical flexibility basis.
On the other hand, production in creative industries and in the arts relies on func-
tional flexibiliy requirements that truly originate in the uncertainty of the creative
process. In order to catch the operating mechanisms of the artistic labor markets, a
252 P.-M. Menger / Poetics 28 (2001) 241-254

model of behavior under uncertainty is needed: note that in a sense, this model
brings us back to the distinction made by a long list of thinkers, from Marx (see
Elster's interpretation, 1985) through Hannah Arendt, Albert Hirschman or Eliot
Freidson (to mention only a few), a distinction between labor and work. The point
has to be discussed at length whenever the issue of the rationales of occupational
choice in the arts is at stake. Here I want only to point to the main underlying
assumption on which this distinction relies, that of uncertain prospect. Work as
opposed to labor refers to nonroutine and noninstrumental dimensions of the creative
activity: the more the work escapes routine and predictability of its outcome, the less
one can be certain about the chances of successful achievement.
To be sure, artistic work also entails routine aspects, both in relative terms - the
various artistic occupations and the various individual achievements in each of them
can be ranked according to how routine or nonroutine the work is - and in absolute
terms - no artist could reconstruct afresh every time his own frame of activity and
no collective work could be achieved either without conventional frames. The fact
remains that the unpredictability of artistic creative work is the most demanding, the
most rewardir~g and the most acclaimed dimension.
Insofar as nonroutine activity refers to a wide range of changing and challenging
work situations, it therefore implies that abilities may be revealed and skills acquired
only progressively, in the course of action, through a process of learning-by-doing
which is highly informative and which cannot be perfectly anticipated ab initio. The
attractiveness of artistic jobs can therefore partly derive from their high learning
potential, at least as long as the work is nonroutine and unpredictable enough.
Thus dealing theoretically with the issue of uncertainty may help to escape the
dilemma that is raised when one asks whether the artist should be considered as a
rational actor (the condition of survival in an ever more competitive market), or as a
bounded-rational actor, or as myopic actor induced to take such occupational risks
only because he or she forms probabilistic miscalculations of his or her chances of
success, or even as a causally driven agent programmed by his initial socialization to
enter an artistic occupation.
Consequently, it appears to be much more efficient to consider and to model the
artist as behaving like an imperfect bayesian actor, i.e. as gathering information,
learning by doing and revising his or her skills, expectations and conception of her
self, as building networks in order to widen his range of work experiences, and to get
new psychic and emotional foods, in a word as self-actualizing without knowing
who exactly he or she is and what exactly he or she is able to do, or to express in his
or her work.
The artist, insofar as he acts as a monopolistic supplier, tries to expand the control
over his own work and over the market of the goods or services he provides. This
outward-oriented goal, driven by the competitive pressure in the market for the arts
and entertainment services, would, however, be meaningless, were it not matching
the inward-oriented goal of self-discovery and self-actualization, a goal that may be
pursued only as long as the variety of work experiences and challenges is optimal
and as the balance between invention, security at work and temporary routine
exploitation of innovation is secured.
 P,-M. Menger I Poetics 28 (2001) 241-254 253

References

Alpers, S., 1985. Rembrandt's enterprise. The studio and the market, Chicago, IL: The University of
Chicago Press.
Aronson, R.L., 1991. Self-employment. A labor market perspective. Ithaca, NY: ILR Press.
Bielby, W.T. and D.D. Bielby, 1999. Organizational mediation of project-base labor markets: Talent
agencies and the careers of screenwriters. American Sociological Review 64, 64-85.
Caves, R.E., 2000. Creative industries: Contracts between art and commerce. Cambridge, MA: Harvard
University Press.
Christopherson, S., 1996. Flexibility and adaptation in industrial relations: The exceptional case of the
US media entertainment industries. In: L. Gray, R. Seeber (eds.), Under the stars, 86-112. Ithaca,
NY: Cornell University Press.
Christopherson, S. and M. Storper, 1989. The effects of flexible specialization on industrial politics and
the labour market: The motion picture industry. Industrial and Labour Relations Review 42,
331-347.
Cohany, S., 1998. Workers in alternative employment arrangements: A second look. Monthly Labor
Review, November, 3-21,
Debeauvais, R., P.-M. Menger, J. Rannou, S. Vari and B. Laplante, 1997. I.~ spectacle vivant. Paris: La
Documentation Franqaise, Collection Contrats d'Etudes Prospectives.
DiMaggio, P., 1977. Market structure, the creative process and popular culture: Toward an organiza-
tional reinterpretation of mass culture theory. Journal of Popular Culture 11,433---45I.
Elias, N., 1993. Mozart: Portrait of a genius. Oxford: Polity Press.
Elster, J., 1985. Making sense of Marx. Cambridge: Cambridge University Press.
Fitzgerald, M.C., 1995. Making modernism: Picasso and the creation of the market for twentieth-century
art. New York: Farrar, Straus and Giroux.
Hippie, S., 1998. Contingent work: Results from the second survey. Monthly Labor Review, November,
22-35.
Kanter, R.M., 1995. World class: Thriving locally in the global economy. New York: Simon and Schus-
ter.
Lancaster, K., 1979. Variety, equity and efficiency. New York: Columbia University Press.
Laplante, B. and B.P. H6bert, 2000. Towards a model of careers in acting: A longitudinal random effects
logit model of getting work on a fragmented and changing labor market. Working paper. Montr6al:
INRS, Universit6 du Qu6bec.
Leicht, K.T. and M.L. Fennell, 1997. The changing organizational context of professional work. Annual
Review of Sociology 23, 215-231.
Meager, P.-M., 1997. La profession de com6dien. Paris: La Documentation Fran~aise.
Meager, P.-M., 1999. Artistic labor markets and careers. Annual Review of Sociology 25, 541-.574.
Meager, P.-M. and M. Gurgand, 1996. Work and compensated unemployment in the performing arts:
Exogenous and endogenous uncertainty in artistic labour markets. In: V. Ginsburgh and P.-M.
Menger (eds.), Economics of the arts, 347-381. Amsterdam: North Holland.
Moulin, R., 1987. The French art market: A sociological view. New Brunswick, NJ: Rutgers University
Press.
Paul, A. and A. Kleingartner, 1994. Flexible production and the transformation of industrial relations in
the motion picture and television industry. Industrial and Labor Relations Review 4, 662-678,
Smith, V., 1997. New forms of work organization. Annual Review of Sociology 23,315-339.
Storper, M., 1989. The transition to flexible specialisation in the film industry. Cambridge Journal of
Economics 13, 273-305.
Throsby, D., 1994. The production and consumption of the arts. Journal of Economic Literature 32,
1-29.
Throsby, D., 1996. Disaggregated earnings functions for artists. In: V. Ginsburgh and P.-M. Menger
(eds.), Economics of the arts, 331-346. Amsterdam: Noah Holland.
254 P,-M. Menger / Poetics 28 (2001) 241-254

Pierre-Michel Menger is director of the Centre de Sociologie des Arts in Paris. He is editor of the series
'Art, Histoire, Soci~t6' and has been the principal investigator on numerous studies of al~ists and author
of many books and articles, including a book on actors. He recently published 'Artistic labor markets
and careers' in the Annual Review of Sociology. Menger is also Dir~cteur d'6tudes at the Ecole des
Hautes en Sciences Sociales, Paris, and Professeur associ6 ~ l'Universit6 du Qu6bec, Montr6al.