3753337827_269929eb38_o

“Scary-bathroom” by Flickr user Miss Nixie

Title: The “Starving Artist” – Myth or Reality? Earnings of Artists in the United States

Author(s): Randall K. Filer

Publisher: Journal of Political Economy

Year: 1986

URL: http://www.jstor.org/stable/1831960?seq=1#page_scan_tab_contents

Topics: artists’ incomes and labor markets, demographic characteristics of artists

Methods: Descriptive analysis of US Census data from 1980 regarding artists’ earnings and demographic characteristics, as well as a descriptive comparison of artists’ labor market characteristics compared to other professions.

What it says: Filer seeks to explore the conventional wisdom that artists earn substantially less than they could in other professions, that only a very few artists will earn the majority of available earnings and thus the income distribution will have high variance, and that artists tend to be very young and leave the artist work force very early to pursue more stable and reliable professions. Prior research exploring artists’ labor markets have suggested that risk-seeking individuals, drawn to the potential high reward of an artistic profession, are particularly drawn to arts professions. Other explanations suggest that artists need to gain experience in the labor market to get an accurate sense of their earnings potential, which will draw a less experienced and more transient population. Further, some have contended that artists are motivated by the non-monetary benefits of artistic participation that may make them more willing to take the earnings penalty of pursuing an arts career than another, less risky field.

Filer examines these assumptions using earnings and occupation data from the 1980 US Census. This dataset classifies professions based on what respondents spent the most time doing during the prior week. Therefore, someone who spent most of their time as a receptionist in an office but also went to dance auditions would be classified as an office worker, not a dancer. The author acknowledges that this is an imperfect classification of artists and that a more accurate data set might be somewhere between self-identified artists and people who successfully make their living as artists, but that a dataset using this classification does not exist. Regarding earnings, Filer finds that while artists’ average earnings are six percent lower than non-artists, which he contends is a smaller difference than most might think given the conventional wisdom. Additionally, there is heterogeneity across artistic disciplines, with actors being the highest earners. Artists also reported working less hours than the general labor market population, were more educated, and were less likely to be black (.055 compared to .105 in the general labor market) and more likely to be male (a proportion of .604 compared to .563 in the general labor market). He finds that artists are younger than the labor market on average, but also that there is less mobility among professions among artists than other professions.

In terms of determinants of artists’ earnings, artists experience a far lower return on earnings per year of education ($511.59 vs. $730.99 in 1980 dollars), which suggests that artists develop human capital more on the job and less in school. Filer notes that it’s interesting that artists choose to spend longer in school than the general population since their returns are lower, and speculates that artists may choose to stay in school to escape from the competitive world of the labor market while continuing to pursue their art. Additionally, black artists face a smaller penalty in the arts than in the general workforce, and noncitizen artists also tend to earn more than noncitizens in other fields.

What I think about it: Filer acknowledges that the US Census’ methodology to identify artists is imperfect, as it only includes those who primarily make their livings as artists. He uses a quantitative method to address the endogeneities in the choice to become an artist, but cannot address the problem that successful artists are an incomplete representation of artists on the whole, as it excludes those who work other jobs to make a living. In light of the fact that this sample only includes those who successfully make their living as artists, I am not surprised that their earnings are not substantially lower than the rest of the workforce. If successful artists did earn substantially less than the general workforce while spending the majority of their time making art, I would think that a larger number of them would change their occupations. Similarly, I would be more surprised at his findings on how artists tend to stay in their fields for longer spells than the rest of the workforce if their earnings were far lower than other occupations. If an artist is one of the lucky few to make their living doing art and also makes art because they love to make art, shouldn’t we expect them to stay in the labor market for long periods of time? Finally, since this data is from the 1980 Census, we should be hesitant to draw conclusions about today’s artists’ labor market based on these findings.

What it all means: Artists in the 1980 Census, when successful at earning a living, didn’t make that much less than the rest of the population, were younger than the rest of the workforce, and tended to stick to their professions for a long time. Other than noting that these working artists are more educated than the general workforce, there is little information in these findings about the socioeconomic status of these artists, including other sources of income from parents or spouses. I’d argue that our research questions are not exclusive to those who have found success at artists, but include those who may be struggling to get by as artists because of barriers to developing human capital as artists due to socioeconomic status. People struggling to get by as artists, and who therefore take work in another field, seem like a key group to understand to answer these questions.