Successful Australian Classical Musicians Abroad – A Statistical Trial

Introduction

Dick Letts’s list of high-level Australian classical music artists enjoying international success (Australian Classical Musicians Successful Abroad at a High Level) is a pioneering project in Australia. It is purely narrative but perusal of the list suggested that it might have potential statistical value.¹ There are of course shortcomings – these findings represent patterns, not hard data. The indicators, with main categories in brackets, are:

• Gender
• Field (concerts; opera; composing; other)
• Specialty (musical instrument; voice category; composing; other)
• Position (conductor; concertmaster; principal, soloist, major)
• Current country of residence (identified with reasonable confidence for 60% of individual artists)
• Observations on venues in each country visited/compositions performed (identified with reasonable confidence for 93%).

These indicators provide a structure for further insights, adding life to the statistical description.

The statistics are based on the list as it stood on 30 June 2015, when it numbered 316, of which 309 were included in the analysis as individuals (excluding three orchestras and four smaller ensembles).² Since the beginning of June 2015, additions to the list have been significantly reduced, making the end of June the appropriate time for an update. The addition of 51% new entries (from 205 to 309) makes the list more representative, but we will keep adding to Australian Classical Musicians Successful Abroad at a High Level and extract the statistics with future updates in mind.

The list was compiled not to aim at statistical purity but to demonstrate that many high-level Australian classical musicians enjoy international success. The length of the list exceeded our expectations. The criteria for describing a particular artist were based on the Music Trust’s accumulated knowledge and experience, but some details were omitted subjectively. Career characteristics are almost inevitably more complex than can be conveyed in short verbal descriptions. More opera singers, for example, may perform at concerts; such activities were not always noted in the artist summary. Some composers conduct their own works but were not recorded as being both composers and conductors. Many orchestral players play chamber music as well, which presents less of a problem because the term “concerts” used as shorthand for instrumental music does not distinguish between orchestral and chamber music, though the general impression from the descriptions is that orchestras dominate these careers.

Generally, statistics must present a simplified picture to be manageable. As noted above, a statistical treatment was not intended when the list was being compiled, and the criteria for the descriptions are not altogether consistent. So there are shortcomings in the categorisation of these successful Australian musicians.

The statistics, nonetheless, form patterns that feed our knowledge. The patterns make sense despite the shortcomings. The data derived from a range of verbal descriptions naturally must be considered “soft”, including the counts of performance locations in each country where some descriptions give a lot of detail and others are more laconic. As just discussed, we can only look for patterns in these particular findings, but to demonstrate the extent of internal consistency this paper presents two alternative ways of tabulating the country information from the lists (described in the section on countries represented).

Three observations are appropriate:

1. The provision of survey data by the Australian Bureau of Statistics has grown immensely over the past 50-60 years, from a time when the Commonwealth Statistician was hesitant using surveys. He reflected that sample surveys carry considerable uncertainties due to statistical sampling error, and some also because of their design.³ This is still the case but quality standards have been formalised,⁴ and on balance the need to broaden the range of statistics in a cost-effective manner within strict budget limits has won out. The statistics presented here fall into the “softer” end of survey data, but still extend our knowledge.
2. National income and expenditure accounts, first published in 1945 by the then Commonwealth Bureau of Census and Statistics, are generally considered to be the quintessential “hard” data, supported by the requirement in national accounting to achieve consistency between three sets of sources: income, expenditure and product. But these statistics do involve some detailed estimation. Furthermore, as discussed in several Knowledge Base papers, GDP (the almost universally used growth indicator) is not always the most suitable measure – it is geared to money rather than community and ignores the damage to cultural and ecological capital from economic activity. GDP growth and general wellbeing are far from synonymous.
3. The music sector is starved of data. If “soft” statistics like those derived here can provide reliable additional insights into its status, it is justified to extract them despite their limitations. The Music Trust is currently engaged in developing as comprehensive a statistical database as possible to conclude the scenario project reported most recently in A First Set of Music Sector Scenarios. Nothing else even begins to illustrate the success of Australian classical musicians abroad, which has economic and cultural benefits for the nation despite the expatriate status of many of these musicians. So should we toss out this evidence because it is imperfect? We think not – present the findings but point out the imperfections.

In conclusion, it is acceptable to look for patterns derived from different sources, not just the conventional economic, demographic and industry statistics which are often insufficient, especially in social and cultural areas. New statistics, “soft” as they may appear, are justified if they add to our insights and their limitations are acknowledged.

Summary of Artist Characteristics

This presentation starts with an overview (Tables 1 to 3) followed by more detailed data, based on cross-classifications. The analysis is mainly derived from so-called pivot tables.

The main activities are concerts (instrumentalists and conductors), opera, and composing, together accounting for 99% of all individuals on the list (Table 1).⁵ Subsequent tables will indicate that these are primary activities; many artists combine activities, for example concerts and opera.

The list of living artists is shown in two parts in Australian Classical Musicians Successful Abroad at a High Level. As Dick Letts explains in his introduction, the smaller category of “early careers” covers young and emerging artists who have not had time to accumulate lengthy curricula vitae. It contains a relatively large opera-related number of artists, 21% of all males and 18% of all females associated with opera. The main activity of all the men included as the early careers group is opera, compared with nine of 14 women.

The proportion of males in the combined list was 63% overall: 61% for concerts, 57% for opera and 79% for composing (100% for the single observations of music information and instrumental ensembles).⁶ In relative terms, the same proportion of men and women were associated with concerts (49%). Forty-four percent of women were in opera, compared with 35% of men (38% in opera overall). Thirty of the 38 composers were males (15% of total males, compared with 7% of females). Chart 1 shows the actual number of persons on the list.

Other activities might have been added, including that at least 15 and probably several more of these prominent artists also hold academic positions abroad. These positions, however, were not always noted in the personal summaries, which tend to concentrate on what shows up as main activities” for most artists – composing, concerts and opera. Academics in all areas, of course, are essential in the development of knowledge and skills – they have another vital role apart from being, in this area, fine performers or composers.

Table 2 gives an overview along different dimensions, identifying “position” from the descriptions on the list. The categories explicitly included in the lists are shown in the left-hand column. The intention when the list was compiled was for instrumentalists to be Australians working as highly selected section principals overseas, or as instrumental soloists or ensemble leaders or members of an acknowledged top orchestra. Some were identified as soloists, which in a orchestral context is a largely European term related to “principals”. In the final analysis, not one instrumentalist on the list failed to qualify on one or more of these criteria. A major source of information on instrumental players was the Australian World Orchestra.⁷

Opera singers were counted as “Principals” only if explicitly described as such, but all others are by definition (and supported by their descriptions in the list) in major roles or soloists. The compiler (Dick Letts) strictly observed the primary criterion that all artists making the list must be “successful abroad at a high level”.

The information in the list, therefore, has been supplemented as far as possible to include “soloists”, which is used as an alternative term for principals in Europe, to include instrumentalists who play for major orchestras abroad, and to identify all listed singers who are not specified as operatic principals in the rubric of “major roles or solo”. The structure of opera means that relatively many singers are not formally called principals – despite their undoubted success and excellence.

In this overall view, males dominate “Management”, which includes persons such as CEOs of orchestras and artistic directors (75% are men). Scrutiny of the main spreadsheet suggests that seven persons (including the two women) are associated with instrumental music, one of these also with opera, and one with composing. Six opera directors on the list include one woman. There are also relatively many males among the conductors (88%; 29 of 33) and among composers as already revealed in Table 1 (30 of 38 = 79%).

Among the principals, soloists and other major performers (numbering 212 in Table 2), 55% were male. Tables 5 and 6 will show distributions for instrumentalists and opera singers, respectively.

The third indicator in this general section is for 130 instrumentalists on the list, shown by individual instruments in the side column and in the body of the table by which orchestral section it belongs to.⁸ Table 3 shows the number of observations for each of 24 instruments played by the 76 males and 54 females listed. They are ranked by frequency showing the violin way ahead, especially for women (21 of 35 = 60%). Male violinists numbered 14 – 18% of the total 76 male instrumentalists. Males were relatively highly represented among pianists (69%), about equal for strings (52%), while there were more female woodwinds (59%). All 10 brass players were men, and both percussionists (small sample again).

Chart 2 shows the five sections graphically, with males and females shown separately.

Details

Given that the descriptions in the list were not intended to form a statistical base, the reliability in this section once again depends on the rigour of the descriptions. Table 4 is indicative rather than precise, as is other information revealed by the statistical analysis. But it is legitimate to look for patterns which have never been revealed elsewhere.

The tables to this stage have shown total findings without any attempt to categorise the main activities of concerts, opera and composing (except for identifying orchestral as distinct from opera conductors). Table 4 goes further, although the distinctions derived from the narratives in the list are not clear-cut. Using the criterion of “first mentioned” in the lists (such as a description of a successful classical musician abroad starting with his or her appearances at concerts followed by appearances at operas, or vice versa, or composing followed by concerts), we find first that composing is apparently the only primary activity of 29 persons on the list, but four others also mention concerts and five other activities. In reality, the number of composers also participating in concert performances may be higher.

Concerts are mentioned without also mentioning opera by 142 persons — only nine combined concerts (mentioned first) with opera. The situation is different for opera, which was mentioned as primary by 118, but 55 of these also played concerts. Concerts mentioned either first or second therefore totalled 210, compared with 127 mentioning opera either first or second. The analysis is shown in the bottom of Table 4, including double-counted observations amounting to 71 of the total 309 cases (23%). The actual double count to the extent that the descriptions don’t reveal all details.

Based on the descriptions, 68% of these artists are involved with concerts either as a primary or secondary activity, 41% with opera, and 14% with composing. These percentages add to 123, including the 23% double counting.

Table 5 identifies the position of Australian classical concert performers working successfully abroad. They numbered 92 males and 57 females, a total of 149 of the 309 individuals on the list (48%). There were 22 conductors, excluding nine who apparently conduct primarily opera but some concerts (adding them to Table 5 would be double counting statistically, since they appear in Table 6). The instrumentalists made up 80% of male and 93% of female concert performers on the list. The proportion of women instrumentalists was 42%, 53 of 127.

Opera performers working abroad are treated similarly in Table 6. Of 106 artists, nine were conductors (8.5%, eight males, one female), 51 male (48%) and 46 female singers (43%). The ratio of female to total singers was 47%, 46 of 97 (higher than the ratio for instrumental players, which was 42%).

Countries Represented

The current residence of high-level classical musicians on the international scene was determined with reasonable certainty for 184 of the 309 individuals on the list (Table 7). The main countries outside Australia itself appear to be Germany and Britain followed by the US. But there are many uncertainties since Table 7 only covers 60% of the list. Britain may be the headquarters of more expatriate Australian classical musicians than Germany and the total number may even beat Australia itself given that the largest quantity of “not ascertained” concerns the homeland. Many musicians marked “AU?” may indeed use the UK as the headquarters for the continuing pursuit of their international careers.

What are the countries represented either by the places they visit or in the case of composers, where their works are performed?

Table 8 presents two measures for the very satisfying total of 287 cases for which the analysis could be done (93%): counts which underestimate the concentration on particular countries, and weighted data which give a value of one to the set of country observations for each artist. If locations in only one country is mentioned by an artist, that country scores one. If there are five locations mentioned, of which two are Austrian, one German, and two in China, the score is Germany .40, Austria .20, and China .40, total 1.00. If there are 20 observations (and there are more than that for several artists), each location gets a score of .05.

Table 8 shows how Europe dominates in the pattern of visits. On the weighted measures, Europe accounts for almost 79% of visits/mentions, with Britain, Germany and Austria leading with 28.8%, 23.6% and 8.7%, respectively. The German/Austrian classical tradition shines through, as well as the level of subsidy and openness to foreign artists in these countries. So does Britain because of its historical and language association with Australia, coupled with a strong musical life with major opera companies and orchestras. One might have expected other nations with a strong classical music tradition (at least as far as composition is concerned) to have shown up at a higher level, but France is a modest fourth with 3.6%, and Italy, Spain and Russia appear further down the list.

North America accounts for 15.3% (with the lion’s share, 93.5%, to the USA; 6.5% to Canada) and East and Southeast Asia for 6.4% (there were 41 observations for Japan, 31 for Hong Kong, and 24 for China. Malaysia, Korea, Taiwan and Singapore accounted for 38 of the remaining 51). Perhaps surprisingly, there were only two observations for the Indian subcontinent.⁹ New Zealand, like the rest of the world including Latin America from Mexico in the north to Argentina and Chile, had a very modest share.

Table 9 shows the weighted scores split into concerts, opera and composing – the significant statistical groups. Europe dominates all, especially concerts and opera (77% and 82% weighted), while composing appears to have a more dispersed impact (Europe 70%). Composing has a relatively strong impact in North America (23% compared with 12-13% for concerts and opera). For all countries, composing is most dispersed with 30% of observations outside Europe,¹⁰ compared with 23% for concerts and 18% for opera.

Table 10 shows the number of mentions on concerts, opera and composing, classified from “very few” (one to five mentions) to 21 and over. The footnotes to Table 10 contains further analysis. Most significantly, 44% of concert artists are attached to local orchestras, though this picture is confused by some concert performers who roam widely because they are not attached by a local orchestra. The three largest counts of locations apply to these performers.

Opera and composing have much higher proportions of multiple mentions according to Table 10 — especially composing which are represented widely.

The usual statistical qualifications apply that were discussed in the introductory section. The travels of some persons are shown in more detail than others. Nevertheless a pattern emerges which makes sense.

Chart 3 summarises the information in this section, expressed as the weighted “scores” with each artist on the list given a weight of one. This is the generally preferred measure for identifying the importance of individual countries, rather than the rough count of locations. But the similarities on a country basis are more prevalent than the differences, except that the highest scores for the two dominant European countries, Germany and Austria, benefit from artists focusing relatively more on each of them, and having less of a scatter over many countries. Consequently, the share of the weighted score is higher than the rough count for each of them according to Table 8. Together, the two countries account for 21.7% of total mentions, but 27.4% of the total score in Chart 3.

Conductors were recorded differently from the other indicators in the analysis. Table 11 contains the geographical information. On a weighted basis, Germany comes out on top with 26.3% (suggesting greater concentration on resident conductors than the average, given that the number of mentions for Germany is “only” 19.4%). Similarly, Austria totals 5.4% weighted compared with 5% on number of mentions. Britain leads on number of mentions (23.4%) but is short of the German weighted score at 21.3% – these conductors attend a greater range of countries than those that visit Germany, where more are resident conductors of an orchestra or opera company.

North America scored higher on the weighted distribution of conductors than on number of mentions (10.8% compared with 10.1%), indicating that Australian-born conductors are more likely to be in charge of particular orchestras. The converse was the case for Asia and other parts of the world.

North America accounted for about 11% of the conductor distribution and East and Southeast Asia for a higher share than was shown in Table 9 for concerts, opera and composing. Comparing the weighted conductor distributions with the distributions for concerts, opera and composing in Table 9 may yield some insights. North America, for example, accounts for 19.6% of all opera performances by successful Australian classical musicians abroad, 15.4% of concerts and 13.5% of composing, but only 10.8% of conductors’ performances. The UK accounted for 21.3% of conductors compared with 30.5% of concerts, 22% of opera and 28% of compositions. The limited sample size for conductors may explain some of it.

Table 12 presents a small statistical analysis of travel patterns for conductors, composers, and artists associated with concerts and opera. Some of these groups are small but a pattern is emerging. The average number of locations recorded is highest for composers (11.3), conductors (9.9) and opera performers (9.3) – they travel more, or at least their compositions in the case of composers. The average is much lower for the concert group (7.3), despite the impact of musicians not tied to a particular orchestra.

The differences are larger when we consider the median – the middle observation of a distribution ranked in numerical order. Composers (10) and conductors (9.5) show the largest medians, ahead of opera (7.5) and especially of concert performers (5).

The standard deviation, in popular terms, is a measure of how “spread out” a distribution is. In the statistical model, 68% of observations are within one standard deviation of either side of the mean, assuming the observations are arranged arithmetically into the well-known “normal” bell-shaped distribution. The standard deviation is lowest for conductors (5.8) — these observations show the smallest variation provided the distribution has the “normal” bell shape. It was 7.9 for composers, 8.4 for concerts and 6.2 for opera. For concerts in particular, this is obviously not possible because it takes us into negative territory from an average of 7.3 The bell curve does not work here – the distribution is just bunched together at the lower end: 40% or more of these persons visit only one or two locations according to these descriptions. At least, the home location such as belonging to a resident orchestra or opera company dominate so much that other locations just aren’t mentioned very often.

There are ways of modifying these statistics, for instance by assuming that the distribution is not an arithmetic “bell” but perhaps a log-normal distribution that cannot become negative. We leave Table 12 without attempting such techniques.

Additions to the List

The list remains alive and growing. Ninety-eight were added in the first three weeks since the statistical analysis was commenced with 205 observations, an impressive 48% though the growth then reduced significantly. Whether the distribution will change significantly only time will tell. Up to now, the most significant additions, relative to the original number, have been to instrumentalists, and the number of opera directors doubled compared with the original list. Male and female singers were less prominent in the additional list. The additions will continue to favour particular categories for a while, but the list is bound to become more representative in the process.¹¹

The additional list also shows some changes in the country distributions. The weighted distributions suggest that the major countries in Table 8, Britain, Germany and Austria, have a lower share than in the original list, and other European countries including France, the Netherlands, Switzerland, Italy and Belgium a higher share. So do North America and Asia. These changes are accepted in the spirit that the greater number has rendered the sample more representative.

The intention is to keep the list of additions up to date and redo the entire analysis at a future stage when the list has changed sufficiently.

Final Words

The statistical interpretation has proven quite resilient considering its purely verbal base. One original weakness was the identification of “principals” from the list — this has been largely overcome by acknowledging the rigorous criteria used to identify the artists. We are dealing with what is confidently regarded as high-level successful musicians, based on definitions including “principal”, “soloist”, “major” and playing in top internationally recognised orchestras.

The dominance of a few European countries in the analysis (notably Britain, Germany and Austria) is striking. It may carry a suggestion that the overseas market could spread significantly in coming years, notably in East and Southeast Asia but also in North America. India, up to now virtually unrepresented but about to be visited by the Australian World Orchestra in October, remains a dark horse. The distributions will be further strengthened by extension of the analysis to the new entries on the list.

The statistics were based on a small number of variables which reoccurred in the descriptions on the list. This largely gives a picture of the current situation — half a dozen or less on the original list are retired, and none of the additional entries — altogether insufficient to distort a basically “now” situation.

Author

Hans Hoegh-Guldberg, first version (205 observations) 16 May 2015. Second extended version (totalling 309 observations) 1 July 2015. Insights and comments on the statistical descriptions by Richard Letts, who compiled the list, are greatly acknowledged.