Songwriters vs. the recording industry: the use and abuse of statistics in UK streaming debates

Abstract In Britain, the Select Committee of the government department for Digital, Culture, Media and Sport has investigated the economics of streaming and recommended that the share of revenues for record companies should decrease so that songwriters can earn more. This article addresses lobbying activity that has resulted from this recommendation. To support their causes, songwriter representatives and record company organisations have made incorrect or misleading use of data from the report Music Creators’ Earnings in the Digital Age. This article looks at the impact of these uses and provides corrections and alternatives to the statistics that have been employed. It also looks at the importance of the issues that have been raised. In conclusion, it addresses aspects of record company accounting that will need to be considered if an increase for songwriters is to be made at their expense.


Introduction
This article addresses lobbying activity that has resulted from a British campaign to reform music streaming revenues, focusing on a proposal that the money should be redivided so that songwriters gain more and the recording industry receives less.Whereas songwriter lobbyists have rallied in support of this proposal, record company lobbyists have rebuffed it.Both parties have drawn on statistical evidence.This article examines their uses of data, assessing the intentions of the lobbyists, outlining the impact of their arguments and correcting the inaccuracies in their analyses.
In October 2020, the Digital, Culture, Media and Sport (DCMS) Select Committee of the UK's House of Commons launched an inquiry into the economics of music streaming.After holding seven oral sessions and gathering nearly 300 pieces of written evidence it issued a report in July 2021, which found 'fundamental, structural problems within the recorded music industry' and concluded that 'Streaming needs a complete reset' (DCMS Select Committee 2021a, §41).The Select Committee identified two issues relating to the division of revenues.In the first instance, it felt that recording artists should receive a higher share of the money.This problem was situated in their contractual agreements with record companies, some of which were deemed unfair.The Select Committee therefore recommended a series of changes to copyright law so that royalty rates might be improved (DCMS Select Committee 2021a, § §77, 123). 1 This article concentrates on the second issue.The Select Committee also felt that songwriters were being short-changed.However, rather than locating this problem in their royalty agreements, it placed the blame on the 'disproportionate' division of revenues 'between the song and master rights' (DCMS Select Committee 2021a, §130).
This issue can be summarised as follows.It is because of their contracts with musicians that record companies gain hold of the copyright in their recordings (known as master rights or recording rights).Similarly, it is because of contracts with songwriters that publishers gain the copyright in their songs (known as song rights or publishing rights).This enables the record companies and publishers to licence the use of these recordings and songs to third parties, including digital service providers (DSPs).Agreements with DSPs such as Spotify have witnessed a larger share of revenues going to the master rights than the song rights.In 2021, the average allocation in the UK was 53% to the former and 15% to the latter, with the remaining 32% being retained by the DSPs (Competition and Markets Authority (CMA) 2022b, figure 5.3).The Select Committee did not feel this imbalance was fair.It reported that streaming had engendered a 'song economy', whereby 'the quality of songwriting is as important as performing talent in determining the extent of music consumption', yet it found that the economic situation of songwriters had worsened rather than improved (2021a, § §78-85).Therefore, the Committee urged the Government 'to consider how to ensure that the song is valued in parity with the recording' and stated that 'If necessary' it would 'bring forward legislative proposals' to support this change, although it did not state what these proposals would be (2021a, §88).
Other countries were debating streaming revenues at this time but the British arguments became atypically polarised.Master rightsholders and song rightsholders do not negotiate against each other; they instead gain their streaming shares through separate negotiations with DSPs.However, rather than concentrating on the ability of DSPs to pay a higher share to the song rights, the UK debates focused on the ability of the master rights to take a cut so that the song rights could earn more.Moreover, rather than being fought between record companies and publishers, it is songwriters with whom the master rightsholders became engaged.This bifurcation was fostered by an alleged conflict of interests.There are three major music companies: the Universal Music Group (UMG), Sony Music Entertainment (SME) and the Warner Music Group (WMG), each of whom has a record company division and a music publishing division.These three companies dominate the market for recordings and songs.In 2021, they gained 70% of the global copyright revenue for master rights and 60% of the global copyright revenue for song rights (Music & Copyright 2022).Songwriter representatives alleged that it was in these companies' interests to maintain the skew of streaming revenues towards master rights, as this enabled them to retain more of the spoils.This is because contractual agreements generally have lower royalty rates for recording artists than they do for songwriters: in 2021, average rates in the UK were 26.3% for the former and 84% for the latter (CMA 2022b, §5.98; see also Table 1).The Select Committee concurred, stating that 'As long as the major record labels also dominate the market for song rights through their publishing operations, it is hard to see whether the song will be valued fairly as a result' (DCMS Select Committee 2021a, §134).
The Committee recommended that the UK's Competition and Markets Authority (CMA) conduct a market study 'to consider how the majors' position in both recording and publishing has influenced the relative value of song and recording rights' (DCMA Select Committee 2021a, §134).This offer was taken up in October 2021, when the CMA announced that it would undertake the study, which in turn would determine whether it would make a full Market Investigation Reference (MIR) and thus 'consider whether there are features of a market that have an This article assesses lobbying activity that took place before, during and after the market study.There is a personal aspect to this work.Concurrent with the DCMS inquiry, I was part of a research team that investigated the earnings of British recording artists and songwriters to see how they were affected by the turn to streaming.This research was funded by the UK's Intellectual Property Office (IPO) and published in the report Music Creators' Earnings in the Digital Age (MCE) (Hesmondhalgh et al. 2021).Lobbyists in the streaming debates have used data from MCE in support of their causes.This article addresses three such instances, each of which involves inaccurate or misleading use of the statistics.First, to help prompt the investigation of revenue shares, songwriter lobbyists utilised MCE figures to suggest that the combined annual earnings of UK songwriters were less than those of Sir Lucian Grainge, the CEO of UMG.This was incorrect.Although the lobbyists under-estimated the amount that Grainge would earn, his pay was ultimately less than that of the songwriters.Second, seeking to preserve the status quo, recording industry trade bodies used statistics from the report to indicate that songwriter earnings had grown while those of record companies had declined.This analysis was accurate for the period 2008-2019, but in more recent years the growth in record company revenue has outpaced that of songwriters.Third, representatives of independent record companies utilised data from MCE to claim that the song rights' share had grown 'at the expense' of the master rights' share and suggest this trend should be reversed (IMPALA 2022, p. 7).This argument rested on a false assumption.Since 2010, the rise or fall in the song rights' share has been set against the DSP share of revenues, rather than the master rights' share.
These uses of statistics all concern an issue that is pivotal for the UK music business.If songwriters are to improve their remuneration from streaming, this will be most comprehensively realised through an increase in the song rights' share; if this share is to increase, this might now result in some diminution in the master rights' share.The equitability and extent of such a change is difficult to determine and has been further occluded by the misuses of data.Therefore, as a step towards clarifying the economic environment, this article provides corrections and alternatives to the statistical misrepresentations.In conclusion, it points towards issues that are yet to be addressed.If the master share is to diminish, this will require assessment of whether the record companies can balance their books.If the record companies can withstand a reduction, there is a need to determine whether this will be beneficial for UK music creators.

Songwriter revenue vs. CEO revenue
Songwriter lobbyists provided the first tactical use of MCE statistics, making a comparison between creator pay and executive pay to promote issues of fairness and sustainability.They also inaccurately represented the data, starting a trend that has remained prevalent in the revenue share debates.
On 21 September 2021, the largest of the three major music companies, UMG, floated 60% of its stock on the Amsterdam Euronext.Its opening valuation was $54.3 billion (Stassen 2021).According to reports, the company's CEO and chairman Sir Lucian Grainge stood to earn a 'transaction bonus' of at least $170m from this initial public offering (Sweney 2021).At the launch event for MCE, which took place the following month, there was a presentation of some of the main findings.This included the calculation that total songwriter revenues from streaming, physical sales and downloads in the UK amounted to £150m in 2019.Witnessing this data, a representative from a songwriters' organisation noted that it fell short of Grainge's earnings.
This comparison was then submitted to the British press.In early November, the Guardian published an article with the by-line 'After a bonus payment, Grainge will earn more this year than all UK songwriters did from streaming and sales in 2019' (Beaumont-Thomas 2021a).The article includes a quote from Crispin Hunt, who was then chair of the songwriters' trade body the Ivors Academy: 'This is evidence of a business which is completely out of control.For songwriters who are struggling to make a living, there's only one word for itobscene' (Beaumont-Thomas 2021a).The Guardian solicited the opinions of politicians.Conversative MP Esther McVey commented, 'It's shocking that record label owners are earning more out of artists' works than the artists themselves', and Labour MP Jo Stevens noted that 'artists get a pitiful amount while streaming sites and record companies cash in' (Beaumont-Thomas 2021a).A few days later, the Labour MP Ellie Reeves referenced the article in the House of Commons, complaining that the head of UMG would earn 'more than UK songwriters made from all UK music streaming, downloads and sales put together' (Reeves 2021).
This activity formed part of the campaigning that surrounded a legislative Bill that had been introduced by Kevin Brennan, one of the members of the DCMS Select Committee (Brennan 2021).On the same day that UMG made its initial public offering, the British Government's response to the Committee's streaming recommendations was made public.The Government noted the evidence that had been gathered for the inquiry and trailed the forthcoming publication of MCE, which it described as 'the most comprehensive study of music creators' earnings ever completed in the UK' (DCMS Select Committee 2021b, p. 1).It nevertheless decided that supplementary research was required before it could 'decide what action it should take' (DCMS Select Committee 2021b, p. 1).Rather than waiting for this research, Brennan had sought to incorporate some of the Committee's recommendations within UK copyright law. 2 Although ostensibly of more use to recording artists than songwriters, his Bill was employed to set several complaints on record.Therefore, when it was read in Parliament in December 2021, the debate moved beyond the specifics of the legislation to discuss the broader division of streaming revenues.Some Members of Parliament had been briefed by representatives of the music creators' community, whereas others were 'assiduously' targeted by recording industry lobbyists ('Copyright (Rights and Remuneration of Musicians, Etc.) Bill' 2021, column 1180).Alongside Brennan, the Labour MPs Geraint Davies, Julie Elliott and Seema Malhotra all referenced Grainge's pay ('Copyright (Rights and Remuneration of Musicians, Etc.) Bill' 2021Bill' : columns 1165Bill' , 1202Bill' , 1211Bill' , 1216)).The main support for the recording industry came from the Conservative MPs Andy Carter and Sir John Whittingdale, the latter of whom defended Grainge's bonus, arguing that he is a British music industry executive who has built Universal Music into the most successful company in the world.As a Conservativesomebody who can celebrate that successand a British citizen, I am delighted that he is going to do so well, but the whole company will do well, too.('Copyright (Rights and Remuneration of Musicians, Etc.) Bill' 2021Bill' : column 1173) ) The music journalist Tim Ingham also defended Grainge's remuneration.In a podcast for Music Business Worldwide, he stressed that the bulk of Grainge's pay would be derived from the 'one-time financial event' of UMG's stock flotation, and compared this with similar amounts earned by the songwriters Bob Dylan, Bruce Springsteen and Paul Simon, who had recently sold off the rights in their compositions to investment companies in their own 'one-time financial events' (Nawaz Kahn 2022).Ingham also noted that under Grainge's leadership, UMG's value had experienced a five-fold increase, worth 'somewhere in the region of $32 billion'; therefore, if he had a $200m bonus, it would 'still only be worth around 0.6 percentand probably lessof the thumping increase in UMG's worth that he's overseen' (Nawaz Kahn 2022).
Grainge's pay packet was higher than predicted.The annual accounts for UMG show that it amounted to E274,284,383 in 2021E274,284,383 in (UMG 2022, p. 179), p. 179).Utilising average exchange rates for that year, this equates to £235,857,141 or $324,478,425 (Exchange Rates UK 2023a, b).Consequently, Grainge earned over 1% of the valuation increase that had occurred during his tenure.A substantial amount of this pay was derived from unrepeatable stock exchange events: Grainge received E17,530,000 owing to a Tencent-led consortium purchasing 10% of UMG; E20,909,789 because Pershing Entities acquired their own 10% share; and E194,982,887 from the initial public offering of 60% of UMG stock (UMG 2022, p. 179).These financial events can be differentiated from the one-time purchases of song rights.Grainge has been in receipt of an ongoing remunerative salary.In contrast, the songwriters that Ingham referenced sold their copyright interests outright and will therefore receive no more royalties or otherwise have them considerably reduced.Furthermore, Grainge's ongoing remuneration has been substantial.He received £35m as his annual salary in 2021 and £40m in 2022 (UMG 2022(UMG , p. 179, 2023, p. 221), p. 221).
As well as under-estimating Grainge's earnings, the lobbyists underrepresented the revenue that songwriters received.The £150m total was taken from figure 4.11 in MCE, which depicts the 'estimated UK combined revenue for physical sales, downloading and on-demand streaming 2000-2019', and provides 'music creator and rightsholder divisions of revenue' (see Figure 1). 3Having been supplied with this total by the lobbyists, the Guardian claimed that it showed the amount that 'UK composers and lyricists earned in 2019 from streaming, downloads and sales' (Beaumont-Thomas 2021a).This was later amended to indicate the correct source of the statistic: composers and lyricists earned £150m from 'streaming, downloads and sales in the UK' (Beaumont-Thomas 2021b).
The difference between the two calculations is significant.To reach the actual figure for the amount UK songwriters earned in 2019, we first need to work out what proportion of UK revenues went to domestic songwriters.Data is not available for song rights, but master rights' sources indicate that about 43% of UK recorded music revenues were generated by British artists (Green 2020, pp. 36-7).If we apply this percentage to UK songwriting revenues, we can approximate that local writers were in receipt of £65m of the 2019 total for streaming, downloads and sales.We then need to assess their global earnings.It has been suggested that British music creators generate fourth-fifths of their copyright revenue abroad (Hesmondhalgh et al. 2021, p. 202).Consequently, the total earnings for UK songwriters from streaming, downloads and sales can be estimated at £325m in 2019, more than double the figure the Guardian reported.This does not represent their total copyright earnings, however.Songwriters also gain revenue from synchronisation, broadcast and public performance licensing, as documented in MCE (Hesmondhalgh et al. 2021, figure 4.12). 4If we add these sources to those previously accumulated and incorporate revenues from live music that were omitted from MCE, 5 we are left with the following: UK revenue for songwriters was £396m in 2019; British songwriters received approximately £170m of this amount; globally, their copyright earnings were around £850m.
It is also possible to provide figures for 2021, thus giving a more temporally appropriate comparison with Grainge's salary. 6Composers and lyricists earned £220m from streaming, downloads and sales in the UK, and £453m once public performance, broadcast and synchronisation licensing are included.British writers received approximately £181m of the latter total.Globally, their copyright revenues can be placed in the region of £905m, which is around three times the amount that Grainge took home.

Record company revenue decline vs. songwriter revenue growth
Recording industry lobbyists were also quick to utilise MCE statistics.For the reading of Kevin Brennan's Bill in Parliament, Andy Carter and Sir John Whittingdale were supplied with data from figure 4.13 of the report, which shows inflation-adjusted UK revenues for physical sales, downloads, streaming, synchronisation licensing, public performances and broadcasts (see Figure 2).The politicians utilised this information to argue that the 'the share of money from streaming that goes to artists has gone up' and therefore legislation was unnecessary ('Copyright (Rights and Remuneration of Musicians, Etc.) Bill' 2021: column 1175).
These interventions helped to prevent the Bill from progressing.George Freeman, Minister for Copyright and Intellectual Property, spoke on behalf of the Conservative Government, acknowledging 'there is a problem' with streaming and accepting the 'fundamental case made by the Select Committee' (column 1221).He nevertheless stated that, rather than legislating, the Government's instinct 'was to try to solve the problem through an industry-led package of measures that artists and musicians support' (column 1230).Although MPs on both sides of the house spoke in favour of the Bill, Freeman singled out the statement of Whittingdale as being 'important and well noted' and that of Carter as demonstrating 'his experience in the industry' (column 1218).In addition, he cited submissions from recording industry trade bodies, which said the Bill would make the UK 'a less attractive place to invest and between composers and publishers in accordance with average royalty rates.The master rights revenue for broadcast and public performance licensing was derived from the annual reports of PPL.After deducting a 14% rate for this collecting society, the remainder was split 50:32.5:17.5 between record companies, featured artists and non-featured performers in accordance with PPL rules.The song rights revenue for broadcast and public performance licensing was derived from the annual reporting of PRS for Music and MCPS.After deducting 15% for their administration rates, the remaining revenue was split 75:25 between composers and publishers in accordance with collecting society rules and average royalty rates (Hesmondhalgh et al. 2021, p. 135). 5Live music was worth £54m in 2019 and £8m in 2021 (PRS for Music 2023). 6The figures in this paragraph utilise BPI data for master rights revenues (Crutchley 2022, p. 15).The total song rights revenue is calculated via the master rights' figures, using revenue shares outlined by the CMA (2022b, figure 5.3).Aside from synchronisation licensing, for which a 65% royalty is employed, the songwriters' share is calculated using the 84% royalty rate from the CMA report (2022b, §5.98).
The share going to British songwriters is calculated at 40%, which is the figure the BPI yearbook suggests was distributed to domestic artists in 2021 (Crutchley 2022, pp. 42-3).
record' and that the solutions it proposed would not 'lead to the outcomes its supporters hope for' (column 1221).
A fortnight later, the British Phonographic Industry's (BPI's) CEO Geoff Taylor used the same MCE figures to question 'the basis for far-reaching proposals for legislative reform' (Paine 2021b).As well as targeting the suggested copyright revisions, Taylor addressed the subject of revenue splits, arguing that any increase in the song rights' share would be unjustified because 'realities aren't in dispute' that 'since 2008 real-terms songwriter earnings have risen 11%, with label revenues down by 19%' (Paine 2021b).This data was then adopted by the International Federation of the Phonographic Industry (IFPI) in its Global Music Report 2022.The IFPI argued against 'unnecessary regulatory interference', citing MCE evidence that 'between 2008 and 2019 artists and songwriters saw revenues grow at a higher rate than those of labels' (IFPI 2022, p. 48).
Although the BPI and IFPI referenced the MCE statistics correctly, their arguments should be treated with caution.They were implying that the growth and fall in revenue were the result of changes in streaming revenue shares.This could only be said with certainty if the figures concerned streaming only and creator royalty rates remained unchanged.This was not the case.The MCE data address the combined copyright revenues from recorded music.As such, we need to consider various trends that determine the progress of song rights and master rights in relation to this overall market.We also need to assess whether the pattern identified for 2008-2019 remains true for subsequent years.The economist Will Page has indicated that this might not be the case.In an analysis of the global value of music copyright, he compared the years 2014 and 2021, and found that record companies were now outpacing songwriters: the song rights share of total music copyright revenues fell from 45% to 34%, while the master rights share grew from 55% to 66% (Page 2022).
The market for recorded music is split between two main types of revenue.There are the 'consumer spend' revenue sources of physical sales, downloading and streaming, which have traditionally witnessed at least three-quarters of the copyright revenue going to the master rights (Page 2022).In addition, there are the 'business licensing' revenue sources of synchronisation, public performance and broadcast licensing, where the copyright revenue is either split evenly between the two sets of rights or the song rights gain a higher share (Page 2022).The fortunes of song rights and master rights can vary if one of them achieves revenue share gains or falls within either the consumer spend sources or the business licensing sources and this is not matched by the other.They can also vary if the market is rebalanced between consumer spend and business licensing.
The MCE data shows that between 2018 and 2019 the former trend predominated (see Figure 3).Among the consumer spend revenue sources, the market shifted from physical sales to streaming.As this happened, the share of streaming revenue allocated to song rights almost doubled, whereas the share for master rights ebbed and flowed by a few percentage points (Hesmondhalgh et al. 2021, p. 115).This was set against other market trends that did not shift a great deal.The balance between consumer spend and business licensing was 68:32 in 2008 and 65:35 in 2019.Meanwhile, once inflation was factored in, total copyright revenues experienced a moderate decline.The increased share that song rights gained for streaming more than compensated for this market fall, leading to an 8% increase in revenue. 7In comparison, master rights revenue fell by 14%. 8 We can combine revenue share and royalty rate data from the CMA market study with revenue sources used in MCE to provide an analysis for the years 2017-2021 (CMA 2022b, §2.69, §5.98, fig. 2.5, fig 5.3).It shows the second trend prevailing (see Figure 4).The balance between consumer spend and business licensing shifted from 63:37 to 72:28.This was largely due to streaming, even though the song rights and master rights both encountered falls in their revenue shares in this period: the former by two percentage points; the latter by three percentage points.Having been worth 34% of UK recorded music revenues in 2017, streaming rose to 57% by 2021.Streaming was also the major factor in an overall growth in revenue.With inflation factored in, it rose from £1.6bn in 2017 to £1.9bn in 2021.Against this background, the revenue for song rights increased by 1% but was considerably outpaced by revenue for master rights, which increased by 23%. 9  The UK market therefore mirrored the move towards master rights that Page documented in his global analysis.Posing the question 'What caused that?', Page writes, 'The answer is the recovery in consumer spend on music, which traditionally favours labels over publishing' (Page 2022). 10  Master rights' revenue shares vs. song rights' revenue shares In November 2022, the CMA published its final report relating to its market study.Among its findings, it determined that 'a publisher would be at a competitive disadvantage in signing songwriters if it did not keep up with the increase in the publishing share of revenues negotiated by other publishers' (2022b, §5.162).In addition, the CMA had asked DSPs about their negotiation processes with major companies and found no evidence 'of terms in the majors' licences for recording rights being linked to terms for publishing rights (or vice versa) to any significant degree' (2022b, §5.115).The CMA also quoted data from section 4.2.2 of MCE, demonstrating that between 2008 and 2019 the average streaming revenue share allocated to the song rights 'increased significantly more than that of recording rights', and concluded 7 During this period there were changes in the division of song rights revenues.The growth of streaming led to collecting societies getting a lower share of revenues, while publishers and songwriters gained more.As a result, collecting societies suffered a revenue decrease of 7%, whereas songwriters gained an increase of 11% and publishers an increase of 10%. 8During this period there were changes in the division of master rights revenues between record companies and recording artists, with the latter gaining higher average royalties for streaming than for physical sales.As a result, record companies suffered a revenue decrease of 19%, whereas recording artists had a rise of 5%. 9With a growth of 1.6%, the revenue for songwriters narrowly outpaced the song rights total.A 22% rise in record company revenues was similar to the master rights total. 10Page adds, 'This has been accentuated by the pandemics [sic] adverse impact on business licensing, which traditionally favours publishers over labels'.However, while Covid-19 did affect public performance and synchronisation licensing revenues in 2021, the general trend is still towards consumer spend revenue.This can be assessed by comparing figures for 2017 with figures for 2022, when lockdowns had ceased.In 2022, the market was balanced 69:31 between consumer spend and business licensing, with streaming accounting for 56% of total revenue.Between 2017 and 2022, revenue for song rights increased by 3%, and revenue for master rights by 18%.
that this increase was 'more consistent with there being competition in publishing over this period to drive up this share rather than there being tacit collusion' (2022b, § §2.120, 5.162).
Although it did not give the streaming market a 'clean bill of health', the CMA decided 'it is unclear or unlikely that the alleged poor outcomes that concern many stakeholders are primarily driven by how firms compete with each other'; therefore, there was no need to conduct a full MIR (2022b, §7.4, §7.8).This marked a turning point in debates about the economics of streaming, as the main avenue for examining the divisions of revenue had been cut off.The outcome was embraced by record company organisations.The BPI commended the CMA's analysis, welcoming its decision that 'co-ownership of recording and publishing are not suppressing the value of the song' (BPI 2022, 3).The most tactical use of the CMA's findings was nevertheless made by IMPALA, a pan-European trade group for independent record labels, which argued for a reassessment of the revenue share debates.
IMPALA reached its position in stages.Preceding its final report, the CMA had issued a market study update.This document references the MCE figures analysed in the preceding section of this article, suggesting that they show how 'the share allocated to publishing rights has increased significantly more than that of recording rights (with songwriters seeing a corresponding increase)' (2022a, §3.89).IMPALA seized upon this information, claiming that, because 'The IPO's analysis shows that [. ..] the share allocated to publishing rights has increased significantly at the expense of label's recording rights', there could be 'no valid reason for increasing the value that songwriter and publishers are able to extract from the digital cake' (IMPALA 2022, p. 7).When the CMA reiterated its assessment of the MCE figures in its final report, this prompted IMPALA to raise a 'contentious question' (CMA 2022b, §2.120;Cooke 2023b).In April 2023, the organisation issued a white paper, which accompanied its own proposals 'to make the most of streaming' (IMPALA 2023a).IMPALA was now asking 'whether the label share is being undervalued compared to other parts of the sector, which are experiencing increases' (IMPALA 2023b, p. 7).
Although IMPALA's argument cannot be rejected out of hand, it was arrived at through misrepresentation of MCE information.In the first instance, the organisation may have been misled by CMA's phrasing, which refers to 'the share allocated' when the subject in question was the growth and fall in revenues.As discussed in the previous section, it can be inappropriate to conflate these two subjects, as the former is not always the main driver of the latter.IMPALA nevertheless took advantage of this slippage, which underpins its claims about the respective value of the rights.Secondly, IMPALA was wrong to claim that the share accorded to holders of song rights had increased 'at the expense' of the share accorded to holders of master rights.This is not what MCE shows and nor is it what the CMA claimed.The polarisation of British streaming debates has distracted from the fact that three parties are involved in the initial division of DSP revenues: the master rightsholders; the song rightsholders; and the DSPs themselves.This means that an adjustment to the share of one rightsholder need not result in an adjustment to the share of the other rightsholder; it can instead be set against the DSP share.According to the CMA this is what happened in practice.The song rights increase was not gained at the detriment of the master rights, it was 'accommodated by [a] fall in the share of music streaming revenues retained by music streaming services' (2022b, §5.101).
The situation was, in fact, more complex than this.Section 4.2.2 of MCE outlines revenue shares for streaming for 2008-2019 (see Table 2).Figure 5.3 of the CMA report utilises data sourced by the organisation to provide revenue share information for 2017-2021 (see Table 3).At times during these periods, the rights agreements were independent of each other; at others, they were linked.Moreover, it is important to note that the share agreements were not the only component of rightsholder pay.The percentage of DSP revenues they receive could be reduced owing to payment charges such as credit card fees.More commonly, it would have been increased owing to deal terms such as minima guarantees (which are implemented if a stipulated per-stream rate is not reached) and digital breakage (by which rightsholders are entitled to keep any shortfall if their royalties do not equal the advance payments they have been paid by DSPs) (Cooke 2020, pp. 66-7).MCE does not incorporate these elements of the DSP deals; its percentages relate to share agreements only and are drawn from trade literature.In contrast, the streaming shares in the CMA report are compiled from data supplied by Amazon, Apple and Spotify, and provide a fuller reckoning of the division of revenues.
Between 2008 and 2010, there was direct correlation between the song rights and the master rights.This is because the original methodology for allocating streaming revenue in the UK was to calculate the master rights payment as a share of DSP revenue after the song rights share was removed.For this period, information about the song rights is in the public domain.In 2007, the UK's copyright tribunal established a revenue share of 8%.This matched the agreement for downloading and  (Cooke 2015(Cooke , 2021a)).MCE employs the lower figure for each timespan.Also, as noted above, MCE does not factor in minima guarantees and digital breakage.These aspects of the deals may nevertheless have had considerable impact at this time, as the DSPs were in their infancy and not yet generating significant revenues.11Consequently, the low revenue shares outlined in MCE for the years 2008-2010 might be misleading and the allocations may have come closer to the shares for the latter years than is shown in the report.It is next worth looking at the years 2011-2016.At the start of this period, the methodology for the song rights changed.Rather than being determined by the copyright tribunal, the UK publishing sector decided that streaming rates should be negotiated in the free market in accordance with the 'option 3' licensing model agreed by the European Commission (Ivors Academy 2022, pp. 22-4).Subsequently, the song rights gained annual incremental increases in their revenue shares, albeit these can only be represented as averages in MCE.This is because the deals were now subject to non-disclosure agreements and could be negotiated by publishers on an individual basis. 12The methodology for the master rights also changed in 2011, as the share was henceforth set against net DSP revenue.MCE depicts a resultant increase, as some master rightsholders retained a 55% share, but this was set against the larger pool of revenue.As documented in MCE, this period therefore witnesses both sets of rightsholders advancing at the expense of DSPs.
For 2017-2019, MCE represents the song rights progressing similarly to the preceding period.The ambition that they reach 15% of net DSP revenue was realised (and in some instances exceeded) by most publishers by the final year (Jones 2021).The master rights experienced greater change.In 2017, the major record companies conducted negotiations with the DSPs and it was reported that their revenue shares decreased to 52% (Ingham 2018;Paine 2021a).This reduction did not result from direct pressure imposed by the publishing industry.It was instead instigated to help some of the DSPs achieve profitability.Barry McCarthy, former Chief Financial Officer of Spotify, noted that in the case of his company, 'the labels were acting in their own self-interest to shore up Spotify's economically-challenged margin structure' (Ingham 2018).However, although the master rights agreements were negotiated independently of song right deals, it is not clear to what extent the DSPs' profit margins were being challenged because of their increased expenditure on the latter rights.The interaction between the three interested parties is further complicated by the CMA's analysis of the years 2017-2021.On the one hand, it shows that the percentage of revenues gained by both sets of rightsholders was consistently above the revenue shares reported in the trade press.Rather than this being a reporting error, this is more likely the result of the continuing influence of minima guarantees and digital breakage.On the other hand, the CMA report depicts the master rights and song rights both experiencing a decrease in their revenue shares.The pattern therefore represents the reverse of the CMA's own statement about the division of revenues.These reductions took place to accommodate a rise in the share retained by DSPs.

Conclusion
The three lobbying uses of MCE statistics can each be questioned, yet they have all had an impact.Songwriter lobbyists used MCE data to compare songwriting earnings with those of Sir Lucian Grainge, illustrating that coincidental with the low payments some music creators were gaining from streaming, the heads of major music companies were enjoying unprecedented returns.The Guardian story received widespread attention.It was repeated, errors included, in many other publications and received extensive airing in parliamentary debates. 13Arguments can be made against the lobbyists' analysis: the earnings of UK songwriters exceeded those of the CEO and a large proportion of Grainge's salary was the result of singular stock exchange events.The contrast between executive pay and songwriter pay was nevertheless startling.Moreover, even though the CMA chose not to launch an MIR and take a closer look at royalties and revenue shares, publicity about creators' earnings helped to ensure these subjects were being aired elsewhere.Prior to the publication of the CMA's final report, the DCMS Select Committee convened an economics of music streaming 'follow-up' to monitor the progress of its recommendations.It was informed that, while there had been progress on some points, there was little development in relation to 'the big issue' of creator remuneration (DCMS Select Committee 2023, §9).The Committee therefore made a new recommendation that the IPO establish a working group 'to consider the current evidence base and monitor developments in other countries' (DCMS Select Committee 2023, §12).
The BPI and IFPI used MCE data to argue that copyright revenues for songwriters had increased while those of record labels had declined.This analysis countered arguments about who had done well and who had suffered in the streaming environment, and was employed to suggest that any rebalancing of rightsholder shares would be unjustified.As we have seen, it was reliant on taking data from a particular span of years.If we look at a more recent era we find that revenue increases for master rights outpace those for song rights.Given this outcome, recording industry trade bodies would presumably no longer wish to base arguments about streaming shares on the rise and fall of total copyright revenues.However, when originally raised, these arguments influenced legislative debates and fed into the conclusions of the CMA report.
IMPALA used MCE data to make the claim that, at the same time as song rights advanced, 'market assessments conducted in the UK (the most exhaustive and independent work done yet) show[ed] that the revenue share of master rights [had] been steadily eroded' (2023b, p. 1).The organisation went one stage further than the BPI and IFPI, arguing that this pattern should be reversed.This brought condemnation from David Israelite, head of the National Music Publishers' Association in America, who branded IMPALA's position 'irresponsible and offensive' (Stassen 2023).It was also denounced by Merck Mercuriades, CEO of the song investment company Hipgnosis, who railed against IMPALA's 'thoughtless statements' (Mercuriades 2023).Neither party corrected IMPALA's factual error, however.The organisation claimed that MCE statistics show the shares of song rights progressing at the expense of the master rights.This is not the case.The last time there was direct correlation between the two revenue shares was in 2010, when streaming was only worth 2.5% of UK recorded music revenues.Since that year, all changes in rightsholder shares have been made against DSP shares.
Although IMPALA's analysis rests on a contentious basis, the topic the organisation raises remains worthy of consideration.This is because any future changes in revenue shares may involve some rebalancing of the song and master rights.During the DCMS inquiry, song rights' lobbyists suggested that the shares should be equivalent, leading the Select Committee to advocate for 'parity' (Cooke 2021a; DCMS Select Committee 2021a, § §88, 132).It would be impossible for this to be achieved without some of the increase for songs coming at the expense of recordings.It would also require Government involvement to enforce this change.If an increase for the song rights were to instead be negotiated on the open market, the rise would probably only be a few percentage points.Even so, the CMA believes the DSPs have reached their limit and even a small increase such as this would ultimately impact upon the master rights' share (2022b, §5.135).
What sort of change could the record companies afford?In assessing their capacity, the first thing to note is that any fall in the master rights' share would not only affect label revenues; it would also reduce the value of their recording artists' royalty deals.Here it should be borne in mind that current divisions of streaming revenue result in songwriters and recording artists receiving similar amounts: utilising the methodology of MCE and data from the CMA report, both constituencies received around 12% of total UK DSP revenues in 2021 (see Table 1).Therefore, if songwriters were to gain improvements because of new revenue agreements, this would redouble the efforts of recording artists to campaign for higher payments of their own.The British Government has suggested the best way of supporting this is by legislating for higher royalty rates.If these were achieved alongside a higher song rights' share, the record labels would be impacted from two directions.
In determining whether record companies could continue to balance their books, policy-makers would need to consider their costs of production.In the UK, rightsholder revenue for streaming is split so that about 80% goes to master rights and 20% to song rights (see Table 1).Thus, it has more in common with physical sales (where copyright revenue is split 87.5% to master rights and 12.5% to song rights) than with radio broadcasting (where the division is about even) (Hesmondhalgh et al. 2021, pp. 61-2).The songwriter fraternity has argued that this is wrong (Ivors Academy 2020).In respect of streaming, the record companies do not have the costs of record manufacture and they do not have to physically distribute their products.They instead license their recordings to DSPs in a similar manner to the licensing of radio broadcasts.Therefore, they could take a reduced share of revenue and pass on more to the music creators.In response, record companies have argued that, while some of their expenses have been reduced, they are now spending more on artist advances and marketing (Hesmondhalgh et al. 2021, pp. 125-6).While there is truth in this, the elasticity of these costs can be contrasted with the fixed expenses of manufacture and distribution.It can also be argued that it is in respect of these costs, rather than in collusion between recording and publishing divisions, that competition has been stifled.The DCMS Select Committee claimed that 'the size of the major record labels means that they can apply their financial might in terms of the size of performer advances' (2021a, §115).Similarly, rather than being an imposition to the major companies, marketing expenditure provides a means by which they can cement their positions.It is because of their ability to outspend independent companies that they can attract top artists and retain dominance in the market.
When assessing record company finances, there would also be a need to look at executive pay.This is another area where there could be flexibility.Although Lucian Grainge's earnings may seem 'obscene', they do not appear disproportionate when compared with similarly positioned CEOs (Nawaz Kahn 2022).Nevertheless, when Grainge was reappointed in 2023 to lead UMG for another five years, his remuneration was amended so that, rather than receiving an 'all-cash compensation' package, he will have a 'compensation programme with a broad set of performancebased objectives' (Cooke 2023a).In the process, his main rewards will come through equity and stock options in UMG and his annual salary will fall to $5m.Music industry analyst Chris Cooke has suggested this may help Grainge to 'avoid further controversies' such as the uproar about his 2021 remuneration (Cooke 2023a).It will also mean that the salary overheads of his company are reduced.
Additionally, policy-makers would need to determine whether increases for song rights and artist royalties can be accommodated via record company profits.In this respect, the companies welcomed the CMA's finding that streaming had not led them to gain any 'substantial and sustained excess' (CMA 2022b, §4.38).Will Page nevertheless calculated that the operating profit margins of major labels increased from an already healthy 8.7% in 2015 to 11.8% in 2019 (2020, p. 10).Reflecting this, Sony Music Entertainment's CEO Rob Stringer stated, 'our margins are way better when compared to the last great era of profit 20 years ago; our margins are amazing now' (Ingham 2019).
The analysis should not end there.On the one hand, there would be a need to assess whether independent record labels are similarly positioned to the major music companies: do they have the same margins in respect of costs and profits?On the other hand, even if it were decided that record company profits are excessive and they should receive a smaller share of the streaming money, it would be necessary to determine whether this would benefit British songwriters and recording artists.Any new agreements about revenue shares or royalty payments would be for the domestic market only.Consequently, it has been argued that multinational music companies would turn their attention to countries where they retain better margins, thus resulting in less investment in the UK (BPI 2022, pp. 6-7).
These topics are on the agenda again.In May 2023, the British Government confirmed it would act on the DCMS Select Committee's recommendation and convene a music industry working group 'to discuss music-maker remuneration' (Cooke 2023c).John Whittingdale, who was appointed as Minister of State for DCMS the same month, announced: The Government wants to see a thriving music industry that delivers sustained growth in an increasingly competitive global music market alongside fair remuneration for existing and future creators.We believe that these aims are complementary, and that reasonable action can be taken by industry to address creators' concerns.(Cooke 2023c) It was reported that the working group would employ data from MCE to inform its discussions (Paine 2023).We await its use of the statistics.

Figure 1 .
Figure 1.Estimated UK combined revenue for physical sales, downloading and on-demand streaming 2000-2019: music creator and rightsholder divisions of revenue (£m)

Figure 2 .
Figure 2.Estimated UK revenue for physical sales, downloading, streaming, synchronisation licensing, public performance and broadcast: music creator shares and rightsholder divisions of revenue 2008-2019 (£m), inflation adjusted

Table 2 .
Shares of UK streaming revenues (%) 2008-2019 (Music Creators' Earnings in the Digital Age, MCE) ://doi.org/10.1017/S0261143023000508PublishedonlinebyCambridgeUniversity Pressparalleled the share of retail revenues the song rights were receiving for physical record sales(Hesmondhalgh et al. 2021, pp.115, 118 (note)).In 2009, a new Online Music Licence set the song rights rate for streaming at 10.5% for the next two years (PRS for Music 2009).The master rights deals were operating differently.Streaming agreements for these rights have always been subject to non-disclosure agreements and each record company or collective of record companies can have a differing revenue share.It has been estimated that in this period the various master rightsholders received between 55% and 60% of DSP revenues minus the song rights share, equating to 50.6-55.2% of net DSP revenues in 20082% of net DSP revenues in  , and 49.2- 53.7% in 20092% of net DSP revenues in  -2010 https