By John Lourenze Poquiz
What does depreciation mean to you? The wear and tear of physical assets may come to mind, such as buildings getting run down by the elements and machinery losing its efficiency after years of use.
Depreciation for tangible assets is often visible and measurable, making it relatively straightforward to account for their declining value over time. However, the concept of depreciation becomes far less intuitive when applied to intangible assets, such as software, artistic originals (songs, films and TV series), or research and development. These assets do not physically deteriorate, yet we can argue that their economic value diminishes over time due to factors like obsolescence (when the asset becomes technically outdated or no longer economically useful, even though it may still function), changing consumer preferences, or the introduction of newer versions of the asset.
A new ESCoE discussion paper from John Lourenze Poquiz looks at measuring the depreciation of intangibles using Google Trends.
Understanding how to measure this decline in value is essential for capturing the true dynamics of the modern economy
Most economies are becoming more reliant in intangibles. A report by the IntanProd Project, a collaboration by the World Intellectual Property Office (WIPO) and Luiss Business School, shows that growth in intangible assets has outstripped growth in tangible assets. Intangibles now play a central role in driving innovation, productivity and competitiveness. In fact, for sectors such as the Information and Communication industry, the level of investments in intangibles already exceeds investment in tangible assets, as shown below.
Source of basic data: Office for National Statistics.
The 2025 System of National Accounts (SNA) will also recommend emphasising figures net of depreciation such as Net Domestic Product (NDP) over gross figures like Gross Domestic Product (GDP). Net figures provide a more accurate measure of economic welfare by accounting for the value loss of capital in production. However, this shift requires reliable estimates of depreciation rates across all asset classes, including intangibles. Without these estimates, the move toward net accounting risks being incomplete or inaccurate, especially in a world where production is becoming more intangible-intensive.
Compilers of official statistics currently use various approaches to estimate depreciation. These include conducting surveys to ask firms about the expected service life of their assets, consulting experts and focus groups for industry-specific insights, adopting practices from comparable countries, or using revenue streams to calculate net present value. However, these methods often rely on assumptions, are highly resource and data-intensive, and may not fully capture the true nature of intangible assets. Additionally, existing approaches tend to lack regular updates, making them less responsive to changes in technology, market conditions and consumer behavior. This highlights the need for innovative, data-driven methodologies that are both scalable and adaptable, providing more accurate and timely estimates for depreciation in a rapidly evolving economic landscape.
Using Google Trends data
Our study proposes a novel approach to estimating the depreciation of intangible assets using data from Google Trends. This freely available tool captures the intensity of internet searches for specific terms over time, offering a unique lens into the popularity of certain search terms.
The core idea is that search volumes for products linked to intangible assets – such as films, songs, TV series, and software – typically peak at the time of release and gradually decline over time. This decline in search volume is assumed to reflect the diminishing economic value of the intangible asset associated with the product. We interpret this rate of decline as depreciation.
As an illustration, let us look at the Google Trends results for the popular software package, “Microsoft Office 2010”, shown in the image below. The Google Trends Index for the search term spiked during the release of the software in 2010. However, searches began to decline shortly after its release, marking a gradual decrease in the software’s popularity and relevance. We interpret this rate of decline as an indicator of depreciation, capturing how the asset’s economic value diminishes over time.
Our analysis focused on Google search patterns for a sample of films and software titles. To measure this, we used a panel regression using data from a sample of software and film titles. This allowed us to see how each film’s/software title’s popularity changed over time and compare different films/software titles to each other. By analysing these trends separately for software and films, we could calculate an overall rate of decline in popularity for each asset category. This was then used as an estimate of how quickly their value decreases over their service life.
For software, we estimated depreciation rates of 13.4% to 19.4% per year, slightly lower than the 20% to 25% rates often assumed by statistical agencies. For films, our estimates aligned more closely with published estimates. Interestingly, newer releases tended to exhibit steeper depreciation, likely reflecting shorter lifecycles in today’s fast-evolving media landscape.
We also find that by applying this methodology to estimate the capital stocks of software originals for reproduction to an intangible-intensive sector such as Industry J (Information and Communication Industry) in the UK, estimates of depreciation would be lower, while gross capital stocks and net capital would both be higher (see image below). This suggests that the current assumptions used in official statistics may overstate the depreciation rates for intangible assets.
Source: Office for National Statistics and author’s computation.
Evaluating the methodology
This methodology offers several advantages. Unlike traditional approaches that rely on static assumptions or costly surveys, our method uses real-time, publicly available data, making the approach both flexible and scalable. While our study focused on films and software, it should be possible to extend this approach to other intangibles, such as TV series, music albums, and other entertainment originals. It may also be possible to extend this approach to intangibles assets categories that are currently not capitalised in the National Accounts, such as marketing and branding.
Of course, some challenges remain. For instance, not all changes in search volume reflect depreciation; market-wide factors or one-off events, like sequels, can temporarily boost interest in older assets. However, these events tend to be relatively uncommon and small in scale. Statistical agencies can apply time dummies to control for these factors, if they choose to apply our regression approach. Time dummies are a way to isolate and measure the impact of events or changes that occur at a particular point in time. Our results suggest that Google Trends is a good substitute for capturing the lifecycle of intangibles, particularly when compared to other indicators of economic value such as revenues.
In conclusion
As we transition to a “weightless economy”, where intangible assets play an increasingly central role, economists and compilers of official statistics must continually improve their tool kits to better capture the complexities of the modern economic landscape. Accurate measurement of intangible assets is crucial for producing reliable productivity statistics, which serve as a cornerstone for understanding economic growth. High-quality data on intangibles enables policymakers and researchers to make informed decisions, design effective development strategies, and address critical challenges in economic policy making.
ESCoE blogs are published to further debate. Any views expressed are solely those of the author(s) and so cannot be taken to represent those of the ESCoE, its partner institutions or the Office for National Statistics.