«Jonathan Haskel Imperial College Business School, CEPR, Ceriba, IZA and UK-IRC Keywords: copyright, IP, innovation, knowledge, investment, ...»
(1) Upstream Input costs: A) the-numbers.com B) OFCOM C) ASHE D) ASHE E) ASHE (2) Downstream Rental Payments, allocated over lifetime of individual assets, adjusted for additional capital revenues. Data unavailable for this report (3) Downstream Rental Payments: D) Based on aggregates: PPL/PRS/MCPS/VPL/BPI/ Live performance. Indicative only, accurate estimation requires a full longitudinal analysis as in (2) (4) Proportion of downstream revenues. Estimates for Books and Music are based on ONS method with new estimates for components (5) Approximation derived from aggregate industry output. Presented for information. Not suitable to use in estimating GFCF. A) ABI, alternative available from UKFC presented in Annex B) ABI D) ABI (6) Official ONS estimate: A) input costs B) input costs adj. for market power in private sector broadcasting C)proportion of revenues D) proportion of revenues E) no official estimate Note to table: All data are nominal and refer to 2007. In cases where we have more than one dataset for that method we choose what we consider to be the superior source. Green highlighted cells in rows (1) and (2) indicate the preferred method for final estimates. Patterned green cells in rows (1) and (3) indicate the second-best approach where data for the preferred method were unavailable. Orange highlighted cells in row (6) contain official data as recorded in the National Accounts.
To see how our new estimates compare with those from the US, we update Table 2 and recalculate investment as a proportion of total GFCF in Artistic Originals and as a proportion of GDP. The results are presented in the Table 13.
Table 13: GFCF by asset, comparison with US and official UK estimates, 2002
Note to table: Data for 2002 for comparison with US data, constructed on the same basis as Table 2.
New estimates for artistic originals are 0.35% of GDP in 2002. Looking at the data by asset category shows our results to be more in line with those from the US. It is worth noting that where data on production costs are not available, estimates in Soloveichik (2010) are based on revenues including for television, music and books. For instance, in the case of television, estimates were based on revenues for cable subscription and advertising. Therefore implicit in the US estimates is the mark-up for monopoly power in the innovation sector. In contrast all of our new estimates for the UK are based on either direct or indirect measures of upstream production costs, with no allowance for upstream market power. With this being the case, we would expect the relative UK shares to be lower than those for the US.
Bearing this in mind, two areas that still stand out are Film and Music. Although estimates for Film are still low relative to the US, we consider this to be a reflection of the role of Hollywood in worldwide Film production.
For Music Originals, our final estimate is just 0.02% of UK GDP in 2002. We think that this is likely due to our use of occupational data on wages from ASHE. Although sole traders, freelance workers and the self-employed are covered by ASHE, we do not believe they are covered as adequately as say employees in a more conventional industry. This is one of the reasons why we had intended to use data from the collecting societies as our preferred source for this asset. Unfortunately due to factors beyond our control we have been unable to make use of this data.
Of the remaining assets, data for Television is also considerably higher in the US. Two related factors come to mind here. First, the US estimates are based on revenue and therefore include additional rents earned by the upstream sector in the US. Second, in recent years the US television industry appears to have begun to acquire a similar status to the US film industry, and is generating revenues from big budget productions that it exports worldwide.
This should be borne in mind when comparing US data with UK data that are based solely on costs of production. Finally, data for Books and Miscellaneous Art appear relatively in line with the US data.
14. Comparison with US estimates: Data and methodology Throughout this report we have made extensive reference to the recent work of Soloveichik of the BEA. The following section compares our data and methodology to that used in Soloveichik (2010), and documents discussions between us, including practical and conceptual issues in the data and some checks on the robustness of the data used. Issues regarding depreciation are discussed in a later section, where we also make use of data kindly supplied to us by Soloveichik.
14.1. Film Our preferred estimates are based on the production budgets for individual films. In estimates for the US, a similar approach was taken that also made use of data from IMDB.com, which as well as production budgets gives information on the production studio and country of ownership.
To provide a check on our data, Soloveichik has sent us data on joint UK/US production back to 1915 from her own dataset. Figure 35 compares current price UK investment in joint UK/US productions with our estimates, from 1991. Total UK expenditure from our own dataset, with no adjustment for ownership shares, is represented by the solid bold black line.
After applying our assumptions on ownership shares we generate a central estimate and upper and lower bounds. The central estimate and upper bound are represented by the black unbroken line and the black dashed line respectively.
As can be seen, the series for US/UK co-productions is fairly consistently at around 40-60% of total expenditure on UK films. It also tracks our higher bound of UK investment fairly closely until the early 2000s, at which point it moves to a higher point than our assumed upper bound. This suggests that our estimation of film investment may be slightly too
conservative. We believe this may be due to the following factors:
- our dataset is more likely to include real data for big budget movies. Lower budget movies are more likely to have been imputed using data in box office revenues, sometimes based on data for the US. If low budget British movies tend to do fairly poorly at the US Box Office, as seems likely, our imputed budgets may be too low.
Additionally, it may be that low budget movies are not comprehensively covered in our dataset
- related to the first point, box office revenues from the UKFC and BFI are not as upto-date as those from the-numbers.com. That is they are box office revenues as of
2003. This may have introduced a downward bias to some of our data for budgets
- our assumptions for co-productions are fairly conservative due to the lack of real information on funding and ownership shares. Since greater information on this is available from IMDB.com, the UK share of joint UK/US productions may be more accurately calculated Since it appears that significant improvements could be made to our dataset using data from IMDB.com, we plan to investigate this issue further in future work.
An additional implication of the data is that from 2003, almost all UK production was cofunded by the US. When considering this, it should be borne in mind that UK films with higher budgets tend to be high profile Hollywood movies, whilst sole UK productions tend to have smaller budgets. The latter are also typically less likely to be of a genre that includes expensive features and/or special effects.
Figure 29: Comparison with BEA data on joint UK/US productions, CP £mns Source: Our estimates compared with Soloveichik data Note to figure: BEA data refer to year of release rather than year of production. Our estimates are for the year of production. All data are nominal and expressed in £mns. BEA data were converted to sterling using an ONS series for the annual weighted average of the dollar/sterling exchange rate.
More generally, when comparing with the data for the US, we see that total UK production is around 4% that of US production (around $400m compared to $10bn in 2009). Initially this may appear small, but the position of Hollywood as a world centre for film production should be borne in mind. Additionally, data sent to us by Soloveichik show that relative to the US, UK films only generate 3% of the DVD sales revenues generated by US films. For premium cable revenues the relative proportion is 4%. Similarly US television audiences on nonpremium channels for UK movies are 3.5% of those for the US. This provides some support for our estimate of the relative size of the film production sector in each country, although note that these revenue comparisons do not take account of how well UK films export to the US.
As explained in the description of our methodology for estimating investment in Film, we aggregate production costs (budgets) across what we believe to be the entire universe of UK films, but for many of those films we have to impute production budgets based on box office revenues and the imputation may be introducing some bias into our final estimates. First, it is more likely that data are missing for lower budget films, and in turn it is more likely that lower budget films are sole UK productions. We are more likely to have data for higher budget films and these are more likely to consist of US/UK co-productions. Therefore our imputation of budgets for ‘smaller’ films will be based on trends for bigger budget productions.
Second, we think there may be some impact from failures or ‘flops’. Consider investments made in mineral exploration, where a mining firm purchases the right to explore 100 sites.
Now it may turn out that 98 of those sites are unsuitable for extraction, and only 2 are actually mined. The revenues generated will greatly exceed the cost of the rights to explore those 2 sites. But the relevant costs are for all 100 sites, that is, the cost for failures, which have been built into expectations. So it is clear that the cost-based estimates should include costs for productions that did not reach the end of filming. Indeed, it should also include the cost of resources devoted to projects that did not even commence. At the moment this is not the case in our dataset. Additionally, as mentioned, we believe we may have a lack of data for lower budget movies. Sometimes such films are not even distributed to cinemas and go straight to DVD release. If such films are under-recorded in our dataset, as we believe may be the case, there will be some bias in our final estimates.
14.2. TV & Radio For Television & Radio, we have used the production spend figures released by OFCOM in their annual reports on the UK PSBs and Communications industry. Unfortunately this only extends back to 1998. To extend the series back further we use the growth rates of the existing ONS data.
The BEA approach for this asset category was to base estimation on purchases of advertising and cable subscriptions. Therefore a crucial difference is that the US data should implicitly include some mark-up for the market power of the innovator. Such an approach would be more difficult to apply in the UK since the BBC is a publicly funded organisation that is not permitted to sell advertising or charge subscriptions for its broadcasts.
14.3. Books For literary works US estimates are based on sales of books and also fiction magazines, 47 and industry data is used alongside microdata on sales (by edition and week) purchased from Nielsen, to back out estimates of the total profits earned from each original.
It had been our intention to use data on the royalties received by authors as a basis for estimating UK investment in this asset type. Since ownership of originals is split between the author and publisher, those royalties would have needed to be grossed up to reflect the total capital compensation earned including that by the publisher. However, due to legal obstacles, data from the relevant collecting societies has not been made available to us in time for the writing of this report. Therefore instead our estimates are based on the incomes of authors and writers reported in ASHE. A potential area of future work would be to use similar data to that used in the US and/or that held by the Collecting Societies, the Publishers Association and the British Library, if and when it becomes available.
14.4. Music For recording originals, US data is based on explicit and implicit royalties earned from music sales (including sheet music) and live performance. Historical data are estimated using data on the number of musicians taken from the US Census.
As with Books, it had been our intention to take a similar approach, based on the royalty revenues earned by musicians, grossed up to account for the ownership share of recording companies. However, again this data has not been made available to us due to legal and administrative issues. So again our estimates are based on the incomes of musicians reported in ASHE. In the future we would like to triangulate our estimates with those generated by the data on royalty revenues, held by the Collecting Societies. In the meantime we have produced estimates for the most recent years based on a cross-sectional sum of annual revenues and royalties. Although this relies on numerous strong assumptions that likely do This is more relevant to data for earlier years when books tended to be serialised in magazine publications.
not apply to the UK music industry in reality, it does indicate that our estimates based on ASHE remain very conservative.
14.5. Miscellaneous Artwork In both this report and the work by Soloveichik, this category is used a catch-all to cover other originals not already included among the primary asset types. Therefore the composition of the category is slightly different in each country. In the BEA estimates, it includes theatrical play scripts, commercial photography, greeting cards, fine art reproduction and radio soap operas. In this report we consider photography, art and choreography, which we estimate using occupational income data from ASHE. Of the categories used in the US not included in this report: we have been unable to source any suitable data on theatrical play scripts; the BBC is the only significant producer of radio stock programmes in the UK so those data are included in our estimates for TV & Radio; we consider greeting cards to fall under ‘Design’, a separate asset type in the intangibles framework.
15. UK Capital Stock of Artistic Originals