You are on page 1of 15

Journal of Cultural Economics 27: 5771, 2003. 2003 Kluwer Academic Publishers. Printed in the Netherlands.

57

Research Note

What Becomes of Undergraduate Dance Majors?


SARAH S. MONTGOMERY and MICHAEL D. ROBINSON
Department of Economics, Mount Holyoke College, South Hadley, MA 01075, U.S.A., e-mail: smontgom@mtholyoke.edu, mirobins@mtholyoke.edu Abstract. This paper examines the earnings, work patterns, and work histories of individuals employed in the eld of dance using data from a survey of graduates of the Five College Dance Department. The respondents to our survey include not only those currently working in dance, but also some who never entered the profession and many others who entered but later left the eld. We nd that graduate education in dance, age, and prior dance experience are highly correlated with employment in the eld of dance. Our ndings are similar to those of other studies on artists, in showing that the returns to dance are low and that many of those employed in dance rely on non-dance jobs to supplement their dance earnings. Key words: artists employed in the eld of dance, earnings, education, labour economics

1. Introduction We report here the results of a survey of the graduates of the Five College Dance Department (FCDD).1 In 1998 about half the respondents were employed in dance, while 85 percent report having been employed in dance at some point after their graduation. The unique nature of our data arises because we survey the alumnae of a dance program not all of whom are currently employed in dance. This allows us to examine which factors are important in whether or not a dance graduate is employed in dance and which factors are related to dance and non-dance earnings. We nd that graduate education in dance, age, and dance experience are highly correlated with both employment in the eld and dance earnings. We also have data from all who were ever in dance about years of professional experience between their college graduation and our survey year of 1998, which allows us to do some analysis of the career paths of our respondents. Our information, on the other hand, is limited to college graduates, who typically go into modern dance. Our results should be applied with caution to others in the dance world. There is a substantial body of empirical literature on the working conditions of performing and creative artists, on the determinants of their earnings, and on the nature of the supply function in artistic labor markets. These include those of Filer (1986), Wassall and Alper (1984, 1992a), Throsby (1992, 1994b, 1996), and

58

SARAH S. MONTGOMERY AND MICHAEL D. ROBINSON

Montgomery and Robinson (1993, 2000). The ndings of these and other analyses are summarized in Wassall and Alper (1992b), Throsby (1994a), and Towse (1995, 1996). Data come from national censuses or from specialized surveys of working artists. Those working in dance frequently have been included in these databases, but have not been the focus of the research in these studies. In addition there have been several studies especially devoted to those in dance. Two early papers are by Santos (1976), who employed data on dancers and singers from the 1960 U.S. Census of Population, and Gray (1984), who surveyed dancers in Minneapolis and St. Paul. There are also monographs by Netzer and Parker (1993) and by Jackson et al. (1994). The rst includes training, performance, and 1989 earnings data for choreographers in four U.S. cities. The second is based on a 1993 survey of the training and careers of those in British dance and drama, with detailed information on the training and current work of members of the dance profession. The study covers weeks worked both in and outside of dance, but contains no earnings data. In addition to having information on the respondents work in the preceding twelve months, it also has data on their rst year in the profession. We developed a survey that was sent to all the individuals on the FCDD mailing list. To allow for comparisons of results, our questions, to a substantial degree, parallel those asked in several earlier surveys of artists work lives. In particular we are indebted to Wassall, Alper, and Davidsons 1983 study of New England artists. However, because we address only dance graduates, we were able to adapt our questions specically to the patterns of training and employment in their eld. The survey was conducted in 1999 and contains detailed information on the activities of the graduates in 1998 as well as substantial information about their careers.2 Five hundred and thirty-four surveys were distributed and 193, or 36 percent, were returned. This paper presents a summary of our ndings. Our results allow us to speculate on the career paths and choices these dance graduates have made. Many graduates pursue careers in dance shortly after graduation, often supporting their dance through non-dance work. Over time the number actively involved in dance declines. It appears that some stay in dance and seek formal graduate work in the eld, while many others obtain non-dance related degrees and leave dance. As is true for many arts, earnings in dance are very low, particularly compared to potential earnings in non-dance elds.

2. Results of the Study Table I includes the means of the demographic variables, breaking these down by whether or not the respondent was employed in dance and by year of graduation. Fifty-four percent of the FCDD graduates were currently employed in dance and 85 percent had some dance experience prior to 1998. The graduates had a mean age of 37 and very few children (less than one on average). Seventy-three per-

WHAT BECOMES OF UNDERGRADUATE DANCE MAJORS?

59

Table I. Characteristics of respondents Variable All Employed in dance Not employed in dance Recenta graduates Earlier graduates

Dance employment Employed in dance in 1998 In 1998 employed as: Dancer Teacher Choreographer Administrator In more than one category Employed in dance prior to 1998 Age Female White Black Hispanic Number of children Living with spouse or partner Spouse employed full-time Age began dance training Pursued formal graduate studies Field of degree Dance Body/movement Non-dance

54.4% 22.8% 34.2% 15.5% 9.8% 22.8% 85.0% 37.0 94.8% 90.4% 3.7% 3.3% 0.8 72.8% 85.1% 8.7 55.4%

100.0% 41.9% 61.9% 28.6% 18.1% 41.9% 96.2% 34.6 95.2% 89.4% 4.8% 3.0% 0.6 71.2% 80.8% 8.9 47.6%

0.0% 71.6% 38.4 94.3% 91.7% 2.4% 3.6% 1.1 74.7% 90.5% 8.5 62.5%

78.3% 40.6% 39.1% 20.2% 15.9% 34.8% 85.5% 28.2 95.7% 91.2% 4.4% 7.5% 0.2 55.1% 81.0% 8.6 31.8

41.1% 12.9% 31.5% 12.9% 6.4% 16.1% 84.7% 41.2 94.4% 90.1% 3.3% 0.9% 1.2 82.8% 89.2% 8.8 66.9

23.3% 10.4% 20.7%

32.4% 8.6% 6.7%

12.5% 12.5% 37.5%

15.9% 5.8% 10.1%

27.4% 12.9% 26.6%

a Recent graduates graduated in 1990 or later.

cent were living with a partner. Those employed in dance were less likely to have partners who were employed full-time, perhaps in part because their partners also were working part-time in dance or some other performing art. The respondents interest in dance began early. On average they began training before the age of nine and over 75 percent rst took lessons at age 12 or younger. The respondents are largely white (90 percent) and female (95 percent). The proportion who are white is slightly higher than the 84 percent for American choreographers (Netzer and Parker, 1993, p. 17) and the 87 percent for those in British dance (Jackson et al., 1994, p. 22). The gender make-up of the group probably reects both the fact that two of the ve colleges whose students are in the FCDD are exclusively for women and the overall dominance of women in dance. In the survey of artists done by the Research Center for Arts and Culture at Columbia University in 1989 (1990), 75

60

SARAH S. MONTGOMERY AND MICHAEL D. ROBINSON

percent of those in dance were women,3 as were 73 percent of the choreographers studied by Netzer and Parker (1993, p. 16), and 76 percent of those in the Jackson et al. study (1994, p. 22). The younger graduates were substantially more likely to be employed in dance and substantially less likely to be living with partners or to have children. Over half (55 percent) of the dance graduates had pursued formal graduate studies. There are differences, however, between those employed in dance and those only employed in non-dance jobs. The former were somewhat less likely to have done graduate work, and their studies were predominately in dance. Thirty-two percent of those employed in dance pursued graduate work in dance, compared to only 13 percent of those not employed in dance. On the other hand, 50 percent of those not employed in dance had body/movement or non-dance degrees, compared to only 15 percent of those employed in dance.4 In order to explore the timing of this graduate work, we broke down educational attainment by date of graduation. Not surprisingly, fewer of the graduates of the 1990s had done graduate study of any sort by 1998. However, 67 percent of those graduating before 1990 had pursued a graduate degree. This suggests that young dance graduates rst try to work in dance and later decide on graduate work. They then may choose further dance training, if they have hopes of continuing in dance, especially as teachers at the college level. If they decide to leave the eld, they may return to school to get a non-dance degree. In 1998, among the respondents who were employed in dance, 42 percent were dancing, 62 percent teaching, 29 percent doing choreography, and 18 percent, working as administrators. It is common for the artists to have been engaged in more than one of these dance activities. Among those employed in dance in 1998, 42 percent report holding jobs of different types. The respondents to the British survey also frequently play multiple roles in the profession. For instance, 32 percent are both dancers and teachers (Jackson et al., 1994, p. 22). Recent graduates were much more likely to be employed in dance and to be dancers. Seventy-eight percent of the respondents from the 1990s were in dance and 41 percent were dancers. In contrast, 41 percent of the graduates from before 1990 worked in dance in 1998 and only 13 percent were dancers. The graduates from the 1990s were not only more likely to be dancers, but also more likely to be teachers, choreographers and administrators, though the difference in the percentage between recent and earlier graduates in these latter occupations was much less than for dancers. These results suggest that many of those who stop dancing, seek only non-dance employment, while others stay in dance to pursue careers in teaching, administration, or choreography. Among our respondents employed in the eld, but currently not working as dancers, 62 percent report past experience as a dancer. The career paths of our respondents parallel those found by Jackson et al. (1994). In discussing British dance they report,

WHAT BECOMES OF UNDERGRADUATE DANCE MAJORS?

61

Table II. Years of experience prior to 1998 Dancer Full sample In dance in 1998 Not in dance in 1998 By type of degree None Dance Body/movement Non-dance 4.6 6.4 2.4 Choreographer 3.6 5.9 0.9 Teacher 5.5 7.8 2.8 Administrator 1.8 2.8 0.6

3.8 8.6 3.8 1.9

2.8 8.2 0.6 1.8

3.8 10.5 4.8 3.7

1.7 3.7 0.3 0.7

. . . our survey respondents were most likely to start their careers as dancers working in dance/ballet. However, many dancers started their careers in other areas and being exible in terms of employment options would seem to be one hallmark of working in the dance labour market. Few people continue working only as dancers for more than ten years. . . in the longer term the survey suggests that after ten years, teaching, choreography and other dance related occupations are the areas where the majority will be working if they remain in the profession (p. 78).

Our data allow for a modest longitudinal analysis of dance careers. Table II examines the respondents experience in the eld of dance prior to 1998. On average the graduates had ve years dancing experience, six years teaching experience, four years experience as a choreographer, and two years experience as an administrator prior to 1998. Those employed in dance in 1998, however, had substantially more previous experience. The had worked, on average, six years as dancers and choreographers, eight years in dance teaching and three years as administrators. Because we believed that educational attainment might be related to dance experience, Table II also breaks down dance experience by type of degree. The respondents who pursued formal graduate studies in dance have substantially more experience than the others in all aspects of dance. Especially striking is that they had spent nine years as dancers while those with formal graduate studies in nondance elds had only two years of dancing experience. It appears that formal training in dance is clearly related to ones attachment to a career in dance. This, combined with the postponement of advanced study, reinforces the impression that there is a division between dance graduates in the sample. There are those who return to school to get more dance training and continue in the eld and others who seek graduate training in non-dance elds and change careers.

62

SARAH S. MONTGOMERY AND MICHAEL D. ROBINSON

Table IIIa. Dance/non-dance employment by decade of graduation (%) Decade of graduation 19901998 19801989 19701979 Not employed in dance 21.7 59.2 60.0 Employed in dance No non-dance job 26.1 17.1 17.1 Employed in dance Have a non-dance job 52.2 23.7 22.9

Table IIIb. Dance/non-dance employment by type of graduate work (%) Graduate work Not employed in dance 36.8 24.4 55.0 82.5 Employed in dance No non-dance job 19.5 46.7 0.0 7.5 Employed in dance Have a non-dance job 43.7 28.9 45.0 10.0

None Dance Body/movement Non-dance

Those employed in dance rely heavily on non-dance jobs. While 76 percent of these artists consider dance their principal profession, 61 percent also work at a non-dance job. Seventy-three percent of these say they require this work to support themselves. As further evidence of their need, 41 percent of those in dance with non-dance jobs get health insurance and 29 percent get retirement benets from their non-dance jobs. Netzer and Parker (1993, p. 55) report a similar percentage (55 percent) of choreographers who work outside of dance. Non-dance employment, however, is substantially lower among those in British dance surveyed by Jackson et al. (1994, p. 64). Only 39 percent reported working outside of dance during the preceding 12 months. We already have observed that a much larger percentage of recent graduates are employed in dance. This is conrmed by the results in Table III. What is particularly striking here, however, is the relationship between the simultaneous holding of dance and non-dance jobs and the decade of graduation. Fifty-two percent of recent graduates are employed in both dance and non-dance jobs. This percentage falls substantially to 24 percent for graduates from the eighties and to 23 percent for those graduating earlier. The differences among the decades are less pronounced, however, when we look at the percentage of graduates holding only dance jobs. Twenty-six percent of the graduates from the 1990s hold only dance jobs, while 17 percent of the graduates from earlier decades work exclusively in dance. This is consistent with the hypothesis that recent graduates are very committed to dance and try to make it in the dance world by supporting themselves with other jobs. After some years of doing this, however, many of those needing non-dance jobs

WHAT BECOMES OF UNDERGRADUATE DANCE MAJORS?

63

appear to drop out of dance employment. Wassall and Alper (1992b) note that older artists spend more time at their art work, but that it is not clear whether this is because the less successful drop out of the arts labor market or, alternatively, because as they become older artists have more and better employment opportunities. Our data support the rst of these explanations. We saw that education is highly related to years of experience in dance. The comparison in Table III of those holding dance and non-dance jobs by type of graduate training shows similar results. Seventy-six percent of the respondents who pursued formal graduate work in dance are employed in dance in 1998, with 47 percent of them having only a dance job. In contrast, only 18 percent of those with non-dance graduate work are in dance. Sixty-three percent of those with no graduate work are employed in dance, but they typically also have a non-dance job. None of those with graduate study in body/movement work exclusively in dance, although 45 percent continue some employment in the eld. Of great interest are the earnings of the artists and in particular their dance earnings. Previous research has found that returns to performing and visual artists are typically quite low and below what they could earn in the non-art world. Table IV reports the 1998 annual earnings for the graduates. Our survey asked the graduates to list their earnings by type of dance activity: work as a dancer, choreographer, dance teacher, dance administrator, other dance work, and grants. We dene total dance earnings as the sum from all these activities. Those employed in dance had mean annual total dance earnings of $16,150. Those with no non-dance job earned $27,725, while those with a non-dance job had dance earnings of only $9,097 and non-dance earnings of $17,232. Since as we already have seen a large number of artists combine dancing with choreography, teaching, and other dance world activities, it is frequently the case that their total dance earnings come from several dance activities. For example, the average earnings of those who are dancing from dancing alone are $3,514, while their total earnings from all dance activities are $13,920. Teaching comprises the largest component of dance earnings making up 45 percent of all dance earnings. Total individual income for those only employed in dance was $32,744 while for those who also held non-dance jobs it was $26,638. Those not working in dance had a substantially larger mean individual income of $44,321. All the graduates reported quite high total household income. For those employed only in dance it was $101,898, for those employed in dance with non-dance jobs, $52,174, and for those not working at all in dance, $99,130. Although they are substantially different populations, it is interesting how similar our respondents earnings are to those of the choreographers studied by Netzer and Parker (1993). In 1998 dollars, the choreographers mean dance related earnings were $17,431, their non-dance earnings $8,159, and their total individual incomes $28,867. Only total household income differs substantially between the two groups with the choreographers averaging $40,610 as compared to $71,689 for our respondents who are in dance.5 This difference may be related to the higher proportion

64
Table IV. Average income and earnings

SARAH S. MONTGOMERY AND MICHAEL D. ROBINSON

For those employed in dance Total No Have a non-dance job non-dance job Total dance related earnings Non-labor income Non-dance earnings Total individual income Total household income $16,150 1,872 10,245 29,021 71,689 $27,725 4,130 32,744 101,898 $9,097 461 17,232 26,638 52,174

Not employed in dance

$ 1,597 40,816 44,321 99,130

Note: Because of some non-responses the components of total individual income do not add up precisely to total individual income.

of our respondents working in dance (71 percent) than of the choreographers (58 percent) who live with a spouse or partner (p. 41). Since we have information on the number of hours and weeks worked both at dance and non-dance employment we can estimate the average hourly earnings for each activity. Conrming what other studies have found for all artists, the hourly return to dance activities at $6.77 is substantially lower than the $15.58 hourly return for non-dance work. Those who held both dance and non-dance jobs earn $6.20 an hour at the former and $13.75 an hour at the latter. This suggests that these artists are highly committed to dance and that much of their reward is nonmonetary. Throsby (1994a) pointed out that artists suffer an earnings disadvantage both because of low hourly art earnings and less than full-time artwork. At least part of the explanation for low dance earnings may be the sporadic nature of dance employment. In our sample the annual weeks worked ranged from 25 as choreographers to 32 as dancers, dance teachers, and administrators. In spite of the fact that many respondents hold multiple jobs this employment does not appear to be full-time year-round work and thus conforms to our expectation that much of the respondents dance employment may be short-term. Jackson et al.s ndings suggest a similar work pattern for their respondents, who reported working in dance an average of 30 weeks per year (p. 64). Graduates who hold both dance and non-dance jobs typically work longer weeks than those who work only in dance or exclusively in non-dance jobs. Those employed only in dance, however, work substantially more than the 40-hour week reported by those with only non-dance jobs. Table V shows the hours per week spent at various dance and non-dance activities for each of the three groups of respondents. These graduates like the artists surveyed elsewhere work unusually long hours, especially if they are employed both in dance and non-dance work.

WHAT BECOMES OF UNDERGRADUATE DANCE MAJORS?

65

Table V. Hours per week of dance and non-dance activity in 1998 Activity Dance job only 13.4 3.8 11.5 5.1 14.0 3.8 51.6 1.1 52.7 0.1 Dance job and non-dance job 13.6 4.7 6.6 4.6 4.6 2.6 36.7 24.1 60.8 1.2 Non-dance job only 0.8 1.3 0.1 0.2 0.1 0.2 2.7 37.8 40.5 0.3

Dancing and rehearsing Self-imposed practice and taking classes Teaching and coaching dance Working as a choreographer Working as a dance administrator Doing other dance work Total dance related Doing non-dance work Total hours working Looking for work

Graduates with only dance jobs spend more time at their dance work than those who combine this with other jobs, but primarily because they spend more time teaching and working as administrators. One of the unique features of our data set is that we have data both on dance graduates who are employed in dance and those who are not. We have observed that a number of factors seem to play a role in determining whether one is employed in dance, including decade of graduation, and post-graduate education. To expand this analysis further we estimate a probit model of employment in dance in 1998 for all the graduates. Since some of them never worked in dance, we estimate a second model limited to those with some dance experience prior to 1998 to see the factors related to continuing in the eld. Table VI reports the results of these estimations. The model controls for dance experience,6 graduate training (dummies for graduate work in dance, body/movement, and non-dance elds), age, race (dummy for white), gender (dummy for male), regional location (dummies for living in New York and for living in New England), whether one has a partner who shares living expenses, partners income, and number of children under six. We expected that years of prior dance experience and graduate training in dance would increase the probability of being employed in dance in 1998, while age would have a negative effect. In addition, we expected having young children to lower, and sharing expenses and partners income to raise, the probability of currently working in dance. We nd that those with graduate training in dance have a higher probability of being employed in dance, while training in a non-dance eld is associated with a lower probability. Past dance experience is also an important determinant of employment in dance. The older graduates, however, are less likely to be employed in

66

SARAH S. MONTGOMERY AND MICHAEL D. ROBINSON

Table VI. Probit model of working in the eld of dance in the survey year 1998 Dependent variable = 1 if employed in dance in 1998, 0 otherwise Variables Full sample 3.58a (13.73) 0.06a (20.63) 0.75b 0.21 1.20a 0.09a 0.02 0.29 0.02 1.06a 0.03 0.56b 0.54 172 67.22 (3.27) (0.34) (10.47) (33.54) (0.01) (2.05) (0.24) (5.58) (0.02) (3.07) (1.97) Those with some dance experience 5.08b (14.38) 0.08a (14.78) 1.33a 0.34 1.03a 0.14a 0.10 0.21 0.02 0.40 1.08 0.34 0.52 130 42.80 (5.32) (0.52) (4.96) (23.80) (0.04) (0.69) (0.15) (0.45) (1.38) (0.75) (1.10)

Intercept Dance experience Graduate work Dance Body/movement Non-dance Age Share living expenses Children under 6 Spouses income White Male New England New York N Log-L

Notes: a Signicant at the 95 percent level; b Signicant at the 90 percent level. Chi-square statistics in parentheses.

dance. Those living in New England are more likely to work in dance than FCDD graduates in the rest of the country. Generally the results for the full sample and for those with prior dance experience are similar. Above we examined earnings and other characteristics of the artists in some detail. In order to extend this analysis we estimated both a dance earnings and nondance earnings equation. The samples consist of respondents with positive dance or non-dance earnings, respectively. For both models we use the log of annual earnings. We control for age,7 race, gender, dance experience, type of graduate work, region, type of dance activity, and weeks worked in dance (non-dance). Since all the respondents in the present study are college graduates, we can only test for the effects of the three types of post-graduate training on dance and nondance earnings. Because our respondents chose whether to be employed in dance or to hold non-dance jobs, it is possible that this selection affects our results. We therefore re-estimated the dance and non-dance earnings models using Heckmans selectivity correction. The correction for the dance earnings regression is based on the probit model shown in Table VI and the correction for the non-dance earnings

WHAT BECOMES OF UNDERGRADUATE DANCE MAJORS?

67

regression is based on a separate model (using the same variables) for holding a non-dance job. Several economists have previously found that human capital theory was not strongly supported in examinations of the determinants of artists earnings. Filer (1990), who found a weak link between education and total earnings of most types of artists, found no effect for dancers. Dividing artists earnings into art, art-related, and non-art earnings, Wassall and Alper (1984) found a positive relation between years of schooling and both arts-related and non-art earnings, but a negative relation with core artistic earnings. In our study using data from the Research Center for Arts and Culture education positively affected non-art income but was not a signicant determinant of art income (Montgomery and Robinson, 2000). However, each of these studies only had data on total education irrespective of eld. Hoping to learn more about the relationship between education and earnings for artists, we sought data on education that contained information on the eld of graduate study. This allows us to control for dance, body/movement, and non-dance graduate work in our regressions on total dance and non-dance earnings. Table VII reports the results of the total dance and non-dance earnings regressions. In neither of the regressions that correct for selection is the inverse Mills ratio signicant, suggesting that selection is not a problem. In addition the results from the two sets of regressions are substantially the same. We nd education to be an important determinant of both dance and non-dance earnings. Formal graduate training in dance signicantly raises dance earnings while lowering non-dance earnings, while formal non-dance graduate training increases non-dance earnings, but does not affect dance earnings. Graduate work in body/movement does not affect either dance or non-dance earnings. Our results are, therefore, consistent with human capital theory with its emphasis on the role of education in determining earnings. In particular we nd that graduate work in dance and graduate work in non-dance elds are forms of specic human capital that do not have positive effects on earnings outside the eld. In our results dance experience positively affects dance earnings while age, in this physically demanding profession, has a negative effect. On the other hand, nondance earnings increase with age. We nd that choreographers have lower dance earnings than those in the other elds. We nd no effects of region, race, or gender on earnings.

3. Conclusions This paper has reported the results of a survey done of the graduates of the Five College Dance Department. One unique feature of our study is that by surveying these graduates, our data contain information not only on individuals employed in dance, but also on those formerly working there and on some who were never employed in the eld. This allows for a more extensive analysis of dance career patterns. Using our data we have examined earnings, work lives, and careers.

68

SARAH S. MONTGOMERY AND MICHAEL D. ROBINSON

Table VII. Dance and non-dance earnings regressions. Dependent variable log of annual earnings Variables Dance earnings 9.16a (8.55) 0.01a (2.16) 0.73a 0.08 0.47 0.04a 0.22 0.05 0.03a 0.07 0.20 0.17 0.22 0.19 0.46b 29.9 95 (2.34) (0.17) (0.83) (2.12) (0.53) (0.09) (3.31) (0.23) (0.64) (0.70) (0.82) (0.61) (1.77) Dance earnings 8.31a (6.60) 0.01 (0.37) 0.61b 0.01 0.92 0.02 0.45 0.17 0.03a 0.26 0.32 0.19 0.17 0.24 0.51b 0.62 32.2 88 (1.74) (0.05) (1.26) (0.54) (0.96) (0.25) (3.32) (0.81) (0.89) (0.73) (0.55) (0.75) (1.75) (0.94) Non-dance earnings 7.87a (9.50) 0.01 (0.68) 0.59b 0.37 0.55a 0.03a 0.18 0.08 0.03a 0.22 0.25 0.30 0.57a 0.25 0.10 (1.79) (1.09) (2.02) (2.06) (0.48) (0.21) (4.72) (0.95) (0.93) (1.11) (2.23) (0.65) (0.33) Non-dance earnings 7.86a (9.23) 0.01 (0.53) 0.70b 0.37 0.55b 0.03a 0.20 0.11 0.03a 0.26 0.29 0.25 0.64a 0.30 0.17 0.11 47.0 103 (1.69) (0.96) (1.75) (2.05) (0.52) (0.24) (4.49) (1.06) (1.01) (0.88) (2.37) (0.69) (0.48) (0.16)

Intercept Years dance exp. Graduate work Dance Body/movement Non-dance Age White Male Weeks emp. in dance Weeks emp. in non-dance New England New York Dancer Teacher Administrator Choreographer Inverse Mills ratio R-squared N

47.1 107

Notes: a Signicant at the 95 percent level; b Signicant at the 90 percent level. T-statistics in parenthesis. The sample for the non-dance (dance) earnings regressions consist of workers employed in non-dance jobs (dance jobs) with positive non-dance (dance) earnings. Because many workers hold more than one type of dance job, there is no omitted category for the dance job dummy variables.

Our results appear to conrm that those employed in dance have low average earnings. In 1998, their mean dance earnings were $16,150. It is clear that these graduates sacrice substantial potential earnings when they choose dance. The total annual dance and non-dance earnings of those in the eld averaged $26,329 while the earnings of those not employed in dance were $40,816. Hourly dance earnings were $6.77 for all employed in dance, while hourly non-dance earnings for those with both dance and non-dance jobs averaged $13.75. To support themselves our respondents frequently work both dance and non-dance jobs. However, this is most pronounced among the more recent graduates. While many of those in our sample spend a considerable number of years employed in dance after graduating, it is the case that over time many leave the eld. Eighty-ve percent of our respondents had

WHAT BECOMES OF UNDERGRADUATE DANCE MAJORS?

69

worked in dance before 1998, but only 54 percent were employed there in 1998. Over time those employed in dance are less likely to work as dancers and more likely to work as teachers, administrators and choreographers. Forty-one percent of graduates from the 1990s were dancing in 1998, compared to only 13 percent of those who graduated before 1990. The data also suggest that if they stay in dance the graduates frequently decide, often several years after college, to pursue advanced study in the eld. We estimate a model of employment in dance and models of dance and nondance earnings. Not surprisingly we nd that graduate work in dance is highly correlated with employment in dance and that those with graduate work in other elds are less likely to be employed in dance. In addition we nd that graduate work in dance increases total dance earnings and decreases non-dance earnings, while non-dance graduate work only increases non-dance earnings. Acknowledgements A Mount Holyoke College Faculty Grant supported this research. We are indebted to the Five College Dance Department for their generous help. We also acknowledge with thanks comments received at the ACEI Biennial Conference, Minneapolis, MN, May 2000 and comments from the editors and two anonymous referees. Notes
1. The FCDD is a consortium of departments at ve closely allied Western Massachusetts colleges: Amherst, Hampshire, Mount Holyoke, Smith, and the University of Massachusetts at Amherst. By the early 1970s, dance faculty from all ve schools were meeting regularly, students crossregistered, and there was a joint annual concert. The program, which is set in the context of a liberal arts education, became increasingly integrated in the late 1970s. Since then curricular and personnel decisions have been made jointly, there has been a ve college major in dance, and students have regularly taken classes on the several campuses. The department is one of the largest in the United States. It is highly regarded in the dance world as well as by its graduates. Fortysix percent of our respondents agreed strongly that the program provided me with excellent technical training, while another 43 percent agreed moderately with this statement. Learning that the FCDD maintains a mailing list of its graduates, we saw the opportunity to survey them about their work lives. The department not only supplied us with the mailing list but also gave us other support. The chair wrote a letter to accompany the questionnaire, urging its completion. 2. A copy of the survey is available on request from the authors. We also conducted several followup telephone interviews. Summaries of these are available in a longer unpublished monograph from the authors. 3. For a more detailed description of this survey see Information on Artists Report Series (Research Center for Arts and Culture, 1990). This statistic was computed by the authors from the raw data le. 4. We asked the respondents if they had pursued graduate work after college and to describe the nature of that graduate work. We found three general categories of responses to this question and recoded them as: dance, body/movement, and non-dance. Graduate study in dance included an MFA or MA in dance, dance and creativity, choreography, dance education, dance history,

70

SARAH S. MONTGOMERY AND MICHAEL D. ROBINSON

and dance-theater. Many graduates pursued graduate studies that were body/movement-related, but were not dance degrees. This type of degree included health and dance, dance/movementtherapy, massage therapy, exercise science, physical therapy, and tness. Non-dance graduate work included any training clearly outside the eld of dance including, for example, law, business, anthropology, education, and medicine. 5. The earnings data for 1989 from Netzer and Parker (1993, pp. 41,57) were converted to 1998 dollars using the Consumer Price Index for All Urban Consumers. Source: U.S. Department of Labor, Bureau of Labor Statistics. 6. Dance experience is dened to be the sum of the total number of years the respondent reported as working in dance, choreography, teaching, and administration. Since these experiences may overlap our total experience variable could exceed the total number of years since graduation. 7. We do not include an age-squared term in either equation. We did not believe the normal age earnings prole would be appropriate for dance where we expected earnings to decline with age. We experimented with an age-squared term in the non-dance earnings equation and were not able to reject the null hypothesis that the returns were linear.

References
Filer, Randall K. (1986) The Starving Artist - Myth or Reality? Earnings of Artists in the United States. Journal of Political Economy 94: 5675. Filer, Randall K. (1988) Labor Market Earnings of American Artists in 1980. National Endowment for the Arts, Washington. Filer, Randall, K. (1990) Arts and Academe: The Effect of Education on Earnings of Artists. Journal of Cultural Economics 14: 1538. Gray, Charles M. (1984) NonPecuniary Rewards in the Performing Arts Labor Market: A Case Study of Dancers and Choreographers, in William S. Hendon, Nancy K. Grant and Douglas V. Shaw, eds., The Economics of Cultural Industries. Association for Cultural Economics, Akron. Jackson, C., Honey, S., Hillage, J., and Stock, J. (1994) Careers and Training in Dance and Drama. A Report of Research for the Arts Council of England. Institute of Manpower Studies, Brighton. Kay, Ann O. and Butcher, Stephyn G.W. (1996) Employment and Earnings of Performing Artists, 1970 to 1990, in National Endowment for the Arts, Artists in the Workforce: Employment and Earnings, 1970 to 1990. Seven Locks Press, Santa Anna California. Montgomery, Sarah S. and Robinson, Michael D. (1993) Visual Artists in New York: Whats Special About Person and Place? Journal of Cultural Economics 17: 1739. Montgomery, Sarah S. and Robinson, Michael D. (2000) The Time Allocation and Earnings of Artists. Industrial Relations 39: 525534. National Endowment for the Arts (1999) Artist Employment in 1998. Washington, D.C. Netzer, Dick and Parker, Ellen (1993) Dancemakers. National Endowment for the Arts, Washington, D.C. Research Center for Arts and Culture (1990) Information on Artists Report Series. Columbia University, New York. Santos, F.P. (1976) Risk, Uncertainty and the Performing Artist, in Mark Blaug, ed., The Economics of the Arts. Martin Robertson, London. Throsby, C. David (1992) Artists as Workers, in Ruth Towse and Abdul Khakee, eds., Cultural Economics. Springer-Verlag, Heidelberg. Throsby, C. David (1994a) The Production and Consumption of the Arts: A View of Cultural Economics. Journal of Economic Literature 32: 129. Throsby, C. David (1994b) A Work-Preference Model of Artistic Behavior, in Alan Peacock and Ilde Rizzo, eds., Cultural Economics and Cultural Policies. Kluwer Academic Publishers, Dordrecht.

WHAT BECOMES OF UNDERGRADUATE DANCE MAJORS?

71

Throsby, C. David (1996) Disaggregated Earnings Functions for Artists, in V.A. Ginsburgh and P.M. Menger, eds., Economics of the Arts Selected Essays. Elsevier Science B.V., Amsterdam. Towse, Ruth (1992) Economic and Social Characteristics of Artists in Wales. Welsh Arts Council, Cardiff. Towse, Ruth (1995) The Economics of Artists Labour Markets. Arts Council of England, London. Towse, Ruth (1996) Market Value and Artists Earnings, in Arjo Klamer, ed., The Value of Culture: On the Relationship between Economics and Arts. Amsterdam University Press, Amsterdam. Wassall, Gregory H. and Alper, Neil O. (1984) The Determinants of Artists Earnings, in William S. Hendon, Douglas V. Shaw, and Nancy K. Grant, eds., The Economics of Cultural Industries. Association for Cultural Economics, Akron. Wassall, Gregory H. and Alper, Neil O. (1992a) The Earnings of American Artists: 19601980, Paper presented at the Seventh International Conference on Cultural Economics, Fort Worth, Texas. Wassall, Gregory H. and Alper, Neil O. (1992b) Towards a Unied Theory of the Determinants of the Earnings of Artists, in Ruth Towse and Abdul Khakee, eds., Cultural Economics. SpringerVerlag, Heidelberg. Wassall, Gregory H., Alper, Neil O., and Davidson, Rebecca (1983) Art Work: Artists in the New England Labor Market. New England Foundation for the Arts, Cambridge, MA.

You might also like