This article reports results from a survey designed to capture information on the drivers of economic performance in the information age. We find that firms’ spending on a broad set of intangible assets was less widespread in 2010 than in 2008, and overall spending on these assets, while still considerable at over £33 billion, was down by some 15 per cent. Within this broad picture, employer funded training is the most widespread knowledge-building activity, and R&D is the least widespread. Larger firms, and those in the production sector, are more likely to engage in intangible investment than smaller firms and those in the service sector. Spending on training increased in cash terms between 2008 and 2010, and spending on design was little changed, but spending on R&D and on Business Process Improvement both fell by around 40 per cent over this period. Reported benefit lives from intangible assets are typically 3-4 years. For the first time we are able to report a geographical breakdown of intangible activity. Our results show such activity to be concentrated in England, with lower than average activity reported in Scotland and Northern Ireland.
ONS is grateful to NESTA for financial support and to all survey respondents for information provided. The authors are grateful to Gaganan Awano of ONS for analytical support.
This article reports on a second survey of intangible assets, covering business expenditure on six categories of intangible assets and the life lengths of these assets (that is, the period over which the business expects to benefit from the asset). The first such survey was conducted two years earlier (Awano et al (2010)). In October 2011, questionnaires were sent to 25401 UK private sector firms with 10 or more employees in the production and service sectors of the economy, drawn from the UK business register. For the first time the sample frame included firms in Northern Ireland.
This exercise was undertaken as part of the National Endowment for Science Technology and the Arts (Nesta) Innovation Index programme and the survey was conducted by the Office for National Statistics (ONS). Responses were obtained from 1180 firms, giving a response rate of 46 per cent which is comparatively high for a voluntary survey (and a little higher than the response rate in 2009). Survey responses have been weighted to be representative of the UK business population, including firms with fewer than 10 employees.
The impact of new technology on developed economies has caused structural changes in the way that many businesses operate. In modern economies there is less focus on conventional forms of tangible capital (machinery and vehicles for example) and more emphasis on “knowledge capital”, which can be embodied in highly trained staff, complex information management systems and sophisticated management techniques.
These developments are beginning to impact the way that national statistical offices measure economic activity. Computer software has been treated as a capital asset in the UK national accounts since 2006 (having previously been treated as intermediate consumption, “used up” in the process of producing other goods and services), and R&D will be similarly incorporated in the national accounts from 2014.
The influential work by Corrado, Hulten and Sichel (2005) makes the case for expanding the asset base further, to include “innovative property” over and above that embodied in conventional measures of R&D (including design) and “economic competencies”, a broad heading including investments in human capital (training), product branding and organisational capital.
The purpose of this survey is to complement other sources of information on intangibles (including R&D surveys, the Community Innovation Survey 2009, and surveys of business investment on currently capitalised assets). In particular, this survey focuses on the distinction between purchased assets and asset creation inside the firm (sometimes referred to as own-account investment).
Another feature of this survey is that it asks explicitly about the life lengths of investments in each category of intangible asset. Information on benefit lives is of interest in the assessment of the case for capitalisation, and also in the computation of flows of capital services from intangibles, which are needed for growth accounting purposes2.
As in the previous survey, firms were asked for their expenditure on six categories of intangible assets: Training, Software, Reputation and Branding, Research and Development (R&D), Design, and Organisation or Business Process improvement (BPI).
This article is not the place for a detailed comparison of survey results with other sources, and we confine ourselves to the following observations. Firstly, other available sources are patchy and not always strictly comparable. Second, R&D spending by businesses is one area where comparability is achievable, but in this particular case we would not expect a small random sample approach as used here to deliver comparable results with the large and tightly targeted approach used in the Business Enterprise R&D (BERD) survey (which reports about three times as much R&D spending after adjusting for double counting).
Third, in the case of purchased software, our results are reasonably close to those from official surveys. Fourth, where comparable data are not available, our estimates are generally much lower than estimates derived indirectly, for example from employment of occupations such as software engineers and designers, or based on information from trade bodies.
Though ours is a small survey, the results are broadly comparable with those from the previous survey and it seems improbable that both sets of results are wrong by an order of magnitude. More work is needed in these areas. Lastly, our reported benefit lives tend to be somewhat shorter than the prevailing assumptions in the literature.
The main differences between this survey and the previous survey are as follows:
The sample size of this survey is approximately 25 per cent larger than the previous survey.
The increase in sample size has allowed the sample frame to be stratified by country (England, Wales, Scotland and Northern Ireland), as well as by industry and size class.
The definition of training has been widened to include staff costs of trainers delivering on-the-job training.
Some other definitions have been clarified in the light of feedback from the first survey.
This sample size is small compared with other ONS business surveys. Accordingly, care should be taken in drawing inferences from the survey results.
Growth accounting with and without intangibles is described in Goodridge et al (2012), which also summarises other sources of information on intangible investment.
|Firm Size||Usable response rate||Positive response rate|
Table 1 summarises the breakdown of the sample response rate by employment size bands. Higher response rates were obtained from smaller firms than for large firms, and the overall response rate of 46 per cent at the end of the response period in January 2012 was higher than the figure at the comparable point of the previous survey (42 per cent).
When analysing responses of those who reported positive spending in any of the 6 categories of intangible assets there is a clear trend for larger firms to be more likely to report such spending. However, the incidence of intangible spending is lower than in the previous survey for each size class, with the decline being particularly pronounced among small and medium sized firms. In this survey, only 40 per cent of firms reported any spending on intangible assets, compared with 58 per cent in the previous survey.
Table 2 shows the breakdown of intangible spending in more detail. Firms in the production sector are more likely to engage in intangible investment than those in the service sector, for the sample as a whole and in all countries apart from Northern Ireland. The incidence of intangible investment in England, Wales and Scotland is close to the UK average, while the incidence appears to be significantly lower in Northern Ireland.
Figure 1 shows the incidence of each category of intangible activity in this survey for 2010 and compares with the results from the previous survey. With the exception of Design, where the incidence remained the same, there has been a decrease in the incidence of intangible expenditure for all assets. The largest decline is in the incidence of firms investing in software, which decreased by 8 percentage points.
The ranking of incidence of spending across the asset classes is very similar between the two surveys. Employer Funded Training is the most common intangible activity, and R&D is the least common. In this survey, the incidence of BPI and of Design have swapped their positions in the rankings compared with the previous survey.
It follows that the incidence of non-R&D intangible activity is much more widespread than R&D alone. According to this survey, only 6 per cent of UK firms are engaged in R&D, while a further 34 per cent of firms are engaged in other forms of intangible asset accumulation. Moreover, R&D is concentrated in a few sectors (pharmaceuticals, aerospace, telecommunications) whereas other forms of intangible activity are much more diffused across the private sector, and much more common across the service sector. Thus much of the incidence of intangibles is not reflected in the R&D statistics.
Before reporting population estimates for investment in intangible assets it is informative to look at the average size of intangible spends by positive responders. Figure 2 shows the summary information from the current survey and comparisons with the previous survey. While the general pattern is similar between the two surveys, it is immediately apparent that the later survey shows a large drop in average spending on R&D.
In Figure 3, the broad industry split reveals that firms in the production sector typically invest larger amounts in R&D than service firms. However the opposite is the case for software, where average expenditure of positive responders in the service sector is four times larger than in the production sector. Compared with the previous survey (see Awano et al (2010), Figure 3) there is some evidence that intangible spending among service sector firms has been more resilient than in the production sector, consistent with the less cyclical nature of services compared with production.
Figure 4 brings together the data on incidence and average expenditure to examine the total population-weighted expenditure of each asset category, and shows the comparable data from 2008. Total expenditure on all six categories of intangible assets is estimated at £33bn, 15% lower than the comparable figure in 2008.
Training was the only category to show an increase in expenditure, rising by 3.5%. This may reflect changes to the questionnaire wording, designed to capture expenditure on on-the-job training. The largest percentage decreases in expenditure occurred in R&D and Business Process Improvement, both of which fell by around two-fifths compared with 2008.
As was the case in the previous survey, software is the largest category of intangible investment, at £10 billion. Reputation and Branding is the next largest (£8.3 billion) followed by training (£7.3 billion) and R&D (£5.7 billion). Expenditures on the categories of Design and Business Process Improvement are both estimated at around £1 billion.
Figure 5 focuses on the split between in-house and purchased accumulation of intangible assets. As expected, this shows that in-house (or own-account) activity is important for all six asset categories, and especially so for Training and R&D. Taking all asset categories together, the share of own-account in total expenditure is about 40 per cent.
Further analysis of the survey results (not reported here for reasons of disclosure) shows that own-account activity is widely distributed across industries and size-class of respondents. This is in contrast with the picture for tangible asset accumulation, where own-account activity is limited to a small number of sectors and the overall share of own-account asset accumulation is low.
It should be noted that the split between in-house and purchased asset accumulation is broadly similar to that of the previous survey except for Software, where there has been a very marked shift towards purchased software, and the share of in-house software has collapsed.
Figure 6 illustrates the split of expenditure between smaller and larger firms as measured by their number of employees. In the population as a whole, employment is split roughly 60:40 between the two categories of 0-499 and 500+ employees (although, in terms of numbers of firms, the larger category represent only around 0.1 per cent of all firms).
Figure 6 shows that spending on Design and on Business Process Improvement is weighted towards smaller firms, while for spending on Reputation & Branding and on R&D there is a bias towards larger firms. This finding for R&D is quite different from the previous survey, where there was a strong bias towards smaller firms.
This, and the big downward revision to R&D expenditure, suggests that while the sample size is larger in this survey than in the previous survey, the random nature of the sample selection process has led to our missing proportionately more of those firms that are actively engaged in R&D.
The expenditure split by broad sector is shown in Figure 7. In interpreting this breakdown of the data it should be borne in mind that the service sector accounts for around 80 per cent of all firms by number, employment and value-added. Hence it is of interest that the production sector “punches above its weight” as far as overall investment in intangible assets is concerned, and especially so in the case of R&D. As in the previous survey we find that investment in Software and in Reputation & Branding is biased towards the service sector, even after allowing for the weight of services in the total population.
One new feature of this survey is that the sample frame was stratified by country (England, Wales, Scotland and Northern Ireland) as well as by industry and size-class. This allows estimates of expenditure on intangible assets to be broken down by country, as shown in Figure 8. It is not surprising that England dominates the overall picture, although it should be noted that the share of spending by English firms is larger than the English share of firms in the population (86 per cent) and of employment (88 per cent).
The share of intangible spending by Welsh firms is roughly commensurate with their weight in the UK population. But the share of Scottish and Northern Irish firms is materially lower than their population weights. Industry composition may be a factor, but unfortunately, the low number of responses from the smaller countries means that it is not possible to present the country breakdown by industry due to disclosure control. For the same reason we do not present the country breakdown by asset category.
|£ billion||95% confidence intervals|
|Training||5.1 - 9.5|
|Software||6.8 - 13.1|
|Reputation & Branding||4.0 - 12.7|
|R&D||4.1 - 7.4|
|Design||0.7 - 1.6|
|Business Process Improvement||0.6 - 1.1|
Table 3 reports confidence intervals for the total expenditure estimates for each category of intangible asset. By construction, confidence intervals are inversely proportional to the size of the sample and directly proportional with the mean and variance of sample responses. The comparatively wide confidence interval for Reputation & Branding reflects the lower incidence of this category compared with, say, Software.
As well as questions on expenditure on asset accumulation, respondents to the questionnaire were asked to report on the life lengths of those assets, that is how long a typical investment is expected to deliver business benefits. This information is useful in estimating deprecation rates for each asset category, as well as providing evidence as to whether intangibles should be treated as capital in the system of national accounts.
Survey results for weighted average benefit lives of each asset category are shown in Figure 9, together with the equivalent estimates from the previous survey. Benefit lives of all asset categories are comfortably greater than one year (the de facto cut-off period for the system of national accounts) and there is broad consistency between estimates from the two surveys for most assets. However, benefit lives for Design and BPI are around one year shorter than the previous survey.
The broad sector split (Figure 10) shows the production sector reporting longer benefit lives for all asset categories apart from Design and (marginally) R&D. One interesting feature is the large difference in reported benefit lives for BPI, which was also a feature of the previous survey. There is no material difference in reported benefit lives between smaller and larger firms.
|Years||95% confidence intervals|
|Training||2.4 - 3.0|
|Software||3.1 - 3.7|
|Reputation & Branding||2.1 - 3.1|
|R&D||2.4 - 6.0|
|Design||2.5 - 3.3|
|Business Process Improvement||2.8 - 3.6|
Confidence intervals around the benefit life estimates are summarised in Table 4. As with the confidence intervals for expenditure, these reflect the number and variability of responses received. This is a factor in the relatively wide interval for R&D, where as we have noted the reported incidence is relatively low.
Awano, G., Franklin, M., Haskel, J. and Kastrinaki, Z. (2010) 'Measuring investment in intangible assets in the UK: results from a new survey', Economic & Labour Market Review, vol 4(7), July.
Corrado, C., Hulten, C. and Sichel, D.(2005) 'Measuring Capital and Technology: An Expanded Framework', in Measuring Capital in the New Economy, edited by Corrado, Carol, Haltiwanger, John and Sichel, Daniel, National Bureau of Economic Research Studies in Income and Wealth, Volume 65, The University of Chicago Press.
Goodridge, P., Haskel, J. and Wallis, G. (2012) 'UK Innovation Index: Productivity and Growth in UK Industries', NESTA Working Paper No. 12/09, July.
Details of the policy governing the release of new data are available by visiting www.statisticsauthority.gov.uk/assessment/code-of-practice/index.html or from the Media Relations Office email: firstname.lastname@example.org