Capital incentive policies are traditionally geared towards boosting tangible investments. However, with the rise of cloud computing, firms can now also access IT as a digital service. This column examines how a UK capital incentive successfully increased tangible IT investment by firms but slowed cloud service adoption. The policy also slowed the diffusion of advanced technologies that rely on the cloud, such as AI and big data analytics, and reduced the demand for data analytics workers.
Policymakers have long tried to influence the direction of technological adoption, using a variety of policy tools to promote innovation, exports, or investment. One common tool – capital incentives – has a strong track record of boosting investment in tangible capital, such as new types of machinery or IT equipment, through tax allowances or grants. But with the rise of new digital technologies, like cloud computing, artificial intelligence (AI), and big data analytics, an important question is whether these traditional capital incentives have unintended consequences for the path of future technology adoption.
Historically, firms acquired new technologies primarily through capital investments. For example, firms access information and communication technologies through the purchase of computers or servers (Jones and Liu 2022). To encourage such investments, every OECD economy currently offers some form of capital incentives, many of which include or even explicitly target IT investments (Tax Foundation 2018).
However, the advent of cloud computing has changed the way firms access IT, allowing them to rent, rather than purchase, computing resources (OECD 2014). This is shifting IT from being a capital investment to a service expense driven by both supply-side cost reductions, as cloud providers benefit from economies of scale (Greenstein and Fang 2020), and demand-side factors, as firms increasingly embrace the cloud for its flexibility, scalability, and cost efficiency (DeStefano et al. 2023). Since the launch of Amazon Web Services in 2006, the substitution towards the cloud and away from traditional IT investment has been dramatic: by 2016 cloud services comprised 25% of IT budgets in Europe.
Despite the rapid growth of cloud services, cloud expenses are typically excluded from capital incentive programmes. Capital incentives subsidise firm investment in IT but not their expenditure on the cloud. The substitutability between IT investment and cloud services raises the question of whether capital incentives inadvertently distort firm technology adoption.
The potential distortion becomes even more important if the cloud is a gateway to next-generation technologies like AI and big data. The skills and knowledge developed from the use of earlier technological vintages can be important for the path of future technology (Atkeson and Kehoe 2007). The effect of capital incentives on next-generation technologies is not obvious though. Their effect depends on the extent to which firms can use their own IT capital, or rather, whether there are advantages to using cloud computing. On the one hand, if capital incentives increase investment in IT hardware to store and process data, then this may encourage the adoption of big data analytics and AI. On the other hand, if cloud computing is strongly complementary to the use of big data and AI, then capital incentive policies (by slowing cloud diffusion) may slow the adoption of these technologies.
Our recent research (DeStefano et al. 2024) examines the impact of capital incentives on the diffusion of cloud, AI, and big data technologies. We focus on the UK’s Annual Investment Allowance (AIA) policy that allows firms to deduct the cost of capital investment against profits up to a certain threshold. The policy was introduced in the UK in 2008 and then modified many times thereafter. Using firm-level panel data, we use changes in these AIA investment ceilings across time – and therefore changes in who was eligible for the programme – to measure the effects on firm investment in traditional IT and their adoption of cloud services and data technologies. Our analysis also considers the labour market effects, tracking the demand for data-analytics workers within firms (Schmidt et al. 2023).
Our findings suggest that capital incentives can distort firm technology choices, slowing the diffusion of next-generation data technologies. The AIA policy boosted tangible IT investments, as expected, by on average 41.3% for hardware and 27.8% for software (Figure 1). However, these capital incentives also slowed the adoption of cloud services by 17 percentage points (Figure 2).
Figure 1 The effects of the AIA capital incentive policy on firm investment in hardware and software
Notes: The table presents the estimated coefficients for the effects of the AIA on hardware and software investments estimated via Callaway and Sant’Anna (2021). All regressions include year and firm fixed effects and control for lagged employment, multi-establishment status, foreign ownership, and age.
Figure 2 The effects of the AIA capital incentive policy on the propensity to adopt cloud computing, big data, and AI
Notes: The table presents the estimated coefficients for the effects of the AIA on the propensity to adopt cloud computing, big data, and AI, respectively. All regressions include year and firm fixed effects and control for lagged employment, multi-establishment status, foreign ownership, and age.
This is a significant distortion, given that 56% of firms used the cloud by the end of our study period. Crucially, the capital incentive also slowed the diffusion of other data-intensive technologies, reducing AI and big data adoption by 3 and 18 percentage points, respectively. These effects were particularly pronounced for small and medium-sized enterprises, precisely the firms most likely to benefit from cloud computing’s flexibility and variable cost structure (DeStefano et al. 2023).
The capital incentive policy also dampened demand for data-analytics workers. Workers in occupations that perform data analytics experience a 1.1% fall in wages in firms that become eligible for the incentive, relative to these same occupations in ineligible firms. However, there is no discernible impact of the policy on other types of workers, such as those who do not perform data tasks or those who input data (rather than analyse it). Thus, we do not see that the AIA affects labour demand in general, but rather it slows demand for data-analytics occupations.
These results are of economic importance. By slowing the diffusion of cloud services, capital incentives have inadvertently hindered the adoption of complementary next-generation technologies like AI and big data analytics. Back-of-the-envelope estimates suggest that big data adoption in the UK could have been 14% higher, and AI adoption 30% higher, in the absence of the AIA’s distortive effects. There has also been much recent discussion of the role of AI as a general-purpose technology capable of having far-reaching effects across the economy (Bonfiglioli et al. 2024, Fillippucci et al. 2024). These findings underscore the need for policymakers to reconsider how capital incentive programmes are structured in the digital age, to avoid unintended consequences that may slow the very technological progress they seek to promote.
Source : VOXeu
