Scaling up to the cloud: Cloud technology use and growth rates in small and large firms

arXiv:2409.17035 · econ.GN, q-fin.EC · Submitted 2024-09-25 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Economics. The day in economics: theory, mechanisms and the evidence beneath them.

Ruth: I'm Ruth.

Warren: And I'm Warren, here with Iris.

Ruth: We're looking at the paper: "Scaling up to the cloud".

Warren: Here's the gist of the paper. Using a unique combination of micro-level data sources on French firms, this paper investigates the relationship between cloud technologies and firm growth rates, specifically distinguishing between Infrastructure-as-a-Service (IaaS) and Software-as-a-Service (SaaS).

Ruth: The analysis utilizes four micro-level data sources: the 2016 and 2018 waves of the French ICT surveys, administrative balance sheet data from 2011–2018, matched employer-employee data for the same period to derive firm-level shares of ICT technical and R&D workers, and information on firm establishment stocks from the French business register.

Warren: The study focuses on long-run growth rates.

Ruth: The main findings demonstrate that cloud use is positively associated with firm growth, but this relationship exhibits a distinct size-contingent effect: the advantage is significantly stronger for smaller firms.

Paper summary: Ruth: To recap what we're looking at with this paper, "Scaling up to the cloud: Cloud technology use and growth rates in small and large firms," they are using micro-level data from French firms—specifically surveys, balance sheets, worker data, and business register stocks—to explore how cloud technologies impact firm growth rates.

Warren: The core thesis they put forward is that while digital technologies usually favor larger companies, cloud use itself has a positive association with firm growth across the board. But the key finding they highlight is that this association becomes significantly stronger when you look at smaller firms >

Iris: And to make that finding more specific, they break down cloud services into IaaS and SaaS, showing that this size-contingent advantage is driven exclusively by Software-as-a-Service. This means the mechanism isn't generic cloud access; it’s something specific to how SaaS operates >

Ruth: That leads them to a deeper interpretation: this advantage for smaller firms only emerges among those firms that already possess complementary assets, meaning their ICT capabilities are aligned with what the SaaS is offering >

Warren: They back this up by showing that in smaller firms, using SaaS cloud services is associated with an internal shift of the workforce toward ICT jobs and higher management roles, which they interpret as evidence of SaaS acting as an organizational innovation enabling scaling >

Iris: That’s a really strong claim because it moves beyond just measuring adoption; they are linking usage to actual internal structural change in how the firm operates >

Ruth: It seems like the main contribution here is showing that cloud use isn't a universal growth driver for all firms; it depends entirely on the firm's existing digital readiness and what specific cloud service it employs >

Warren: So, if you’re thinking about this paper, you should remember that it shifts the focus from just firm age to firm size as a critical determinant of whether cloud technology provides a real growth benefit >

Conclusion: Ruth: So wrapping up this discussion on "Scaling up to the cloud: Cloud technology use and growth rates in small and large firms," what this paper really suggests is that we can't treat digitalization as a blanket positive for every company.

Warren: They are emphasizing that the benefits aren't just about being in the cloud; it’s about *how* you are using it, specifically preferring SaaS over IaaS when you’re a smaller player looking to grow >

Iris: And they suggest that this advantage is conditional on having those complementary ICT capabilities already in place, which means firms need to be set up correctly before the cloud can really help them scale up >

Ruth: So for listeners who just want the plain idea, think of it this way: SaaS broadens access to digital tools, but only firms that are already ready with the right internal assets can actually use those tools to get a disproportionate return >

Warren: That’s the practical implication—it means small firms have an opportunity for scaling up if they have the right digital foundation, but they can’t just jump in and expect massive growth without that preparation >

Iris: And they also point out that the study has limitations, specifically that their instrumental variable strategy uses spatial variation in lightning strike density as an instrument for cloud use, which is a specific choice of data source >

Ruth: That’s right. They admit their method relies on linking infrastructure maintenance costs to slower broadband diffusion to get that instrument, which is a specific causal path they are testing >

European Commission · UNU-MERIT · Maastricht University · Università di Brescia

econ.GN, q-fin.EC

Submitted: 2024-09-25

Updated: 2026-08-24

Journal ref: Caldarola, B., & Fontanelli, L. (2027). Scaling up to the cloud: Cloud technology use and growth rates in small and large firms. Research Policy, 56(1), 105644

DOI: 10.1016/j.respol.2026.105644

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 85/100

The gist: Using a unique combination of micro-level data sources on French firms, this paper investigates the relationship between cloud technologies and firm growth rates…

Key concepts

Infrastructure-as-a-Service (IaaS)
IaaS provides fundamental computing resources like servers and storage over the internet. The study found that IaaS use does not show a significant size-dependent effect on firm growth, unlike SaaS.
Software-as-a-Service (SaaS)
SaaS delivers complete software applications via the cloud, meaning firms rent ready-made tools. This type of cloud use is strongly linked to positive firm growth and exhibits a distinct size effect where smaller firms benefit more.
Size-Contingent Effect
This means the impact of a variable (like cloud use) changes depending on another factor, specifically firm size. The paper discovered that the advantage of using SaaS is significantly stronger for smaller firms compared to larger ones.
ICT Complementarity
This refers to how well a firm's existing digital capabilities match the needs of new technologies. The research suggests that the growth benefits from SaaS are only realized when small firms already possess these complementary digital assets.

Terminology

Summary

Using a unique combination of micro-level data sources on French firms, this paper investigates the relationship between cloud technologies and firm growth rates, specifically distinguishing between Infrastructure-as-a-Service (IaaS) and Software-as-a-Service (SaaS). p. 3

Where the paper says it

In this paper, we examine the heterogeneous impact of the purchase of cloud services on firm size growth rates using a unique combination of four micro-level data sources on French firms: the 2016 and 2018 waves of the French ICT surveys, administrative data from firms’ balance sheets, matched employer-employee data, and the French business register. We focus on long-run growth rates, consistent with the view that the effects of digital technologies may take time to materialise owing to the large and complex organisational changes they entail, which are characterised by lengthy implementation lags (Brynjolfsson and Hitt, 2003; Brynjolfsson et al.

The analysis utilizes four micro-level data sources: the 2016 and 2018 waves of the French ICT surveys, administrative balance sheet data from 2011–2018, matched employer-employee data for the same period to derive firm-level shares of ICT technical and R&D workers, and information on firm establishment stocks from the French business register. p. 3
Where the paper says it

In this paper, we examine the heterogeneous impact of the purchase of cloud services on firm size growth rates using a unique combination of four micro-level data sources on French firms: the 2016 and 2018 waves of the French ICT surveys, administrative data from firms’ balance sheets, matched employer-employee data, and the French business register. We focus on long-run growth rates, consistent with the view that the effects of digital technologies may take time to materialise owing to the large and complex organisational changes they entail, which are characterised by lengthy implementation lags (Brynjolfsson and Hitt, 2003; Brynjolfsson et al.

The study focuses on long-run growth rates. p. 3
Where the paper says it

In this paper, we examine the heterogeneous impact of the purchase of cloud services on firm size growth rates using a unique combination of four micro-level data sources on French firms: the 2016 and 2018 waves of the French ICT surveys, administrative data from firms’ balance sheets, matched employer-employee data, and the French business register. We focus on long-run growth rates, consistent with the view that the effects of digital technologies may take time to materialise owing to the large and complex organisational changes they entail, which are characterised by lengthy implementation lags (Brynjolfsson and Hitt, 2003; Brynjolfsson et al.

The main findings demonstrate that cloud use is positively associated with firm growth, but this relationship exhibits a distinct size-contingent effect: the advantage is significantly stronger for smaller firms. p. 0

Where the paper says it

Whereas digital technologies have generally been found to disproportionately benefit larger firms, we document that cloud use is positively associated with firm growth, and this association is significantly stronger for smaller firms. Distinguishing cloud services between Infrastructure-as-a-Service (IaaS) and Software-as-a-Service (SaaS), we find that the size-contingent association is driven entirely by SaaS.

Crucially, this size-contingent association is driven entirely by SaaS cloud services, whereas IaaS does not show a comparable size-dependent effect. p. 0
Where the paper says it

Distinguishing cloud services between Infrastructure-as-a-Service (IaaS) and Software-as-a-Service (SaaS), we find that the size-contingent association is driven entirely by SaaS. Crucially, this small firm advantage is conditional on complementary assets as it emerges only among firms endowed with ICT capabilities.

This finding suggests that SaaS broadens access to digital technologies, but its effective use is conditional on firms’ ICT readiness, implying that the disproportionate returns accrue only to firms endowed with complementary ICT assets.

Furthermore, the paper establishes an interpretation rooted in ICT complementarity: the SaaS-driven growth advantage of smaller firms emerges only among those already possessing complementary digital capabilities. p. 4

Where the paper says it

We find that the SaaS-driven growth advantage of smaller firms is concentrated among those endowed with ICT capabilities, measured by the use of other digital technologies. This supports an ICT complementarity interpretation: SaaS broadens access to digital inputs, but the disproportionate returns accruing to smaller firms are conditional on the presence of complementary ICT assets, which neither IaaS nor SaaS cloud technologies can substitute for.

This is supported by evidence showing that SaaS use in smaller firms is associated with an internal reallocation of the workforce toward ICT occupations and upper-layer managerial and intermediate roles, consistent with SaaS operating as an organizational innovation enabling digitalization and scaling. p. 0
Where the paper says it

SaaS cloud use is further associated with internal workforce reorganisation toward ICT, managerial, and intermediate occupations and away from manual and clerical ones, particularly among smaller firms, consistent with SaaS cloud operating as an organisational innovation enabling digitalisation and scaling. Finally, we present descriptive sectoral evidence of a negative correlation between cloud and market concentration.

The study employs a sophisticated econometric framework, including instrumental variable (IV) strategies using spatial variation in lightning strike density at the municipality level as an instrument for cloud use, leveraging evidence linking increased infrastructure maintenance costs to slower broadband diffusion. p. 3

Where the paper says it

We further address endogeneity concerns through instrumental variable (IV) strategies, including control function approaches and two-stage least squares (TSLS) models designed to account for treatment heterogeneity (Wooldridge, 2015). Our instrument exploits spatial variation in lightning strike density at the municipality level, which is associated with investments in IT infrastructure (see also Andersen et al.

Endogenous treatment models and Two-Stage Least Squares (TSLS) models are also employed to address endogeneity concerns. p. 3
Where the paper says it

This result is robust across a range of specifications, including alternative samples and variable definitions, non-linear specifications with employment, sales and age classes, varying growth rate measurement windows, and long-difference estimation. We further address endogeneity concerns through instrumental variable (IV) strategies, including control function approaches and two-stage least squares (TSLS) models designed to account for treatment heterogeneity (Wooldridge, 2015).

The interaction terms within the model reveal specific size-contingent effects: Cloud SaaS × Size Class = 5 (1000+) yields a negative coefficient (-0. p. 4
Where the paper says it

Aggregating cloud intensity and market concentration at the two-digit sectoral level, we document a negative correlation between the two. While this association is descriptive in nature, it is consistent with the interpretation that the size-contingent returns to cloud use documented for smaller, ICT-capable firms may, in aggregate, be associated with less dispersed market structures across French industries.

1292), while Cloud SaaS × Size Class = 1 (50-249) yields a smaller effect (-0. p. 5
Where the paper says it

Not all ICTs and cloud technologies affect firms in the same way. In our analysis only SaaS generates a growth premium that is disproportionately large for smaller firms, and this size-contingent effect emerges only among firms with complementary ICT capabilities.

0848).

Beyond firm size, the analysis also documents sector-level evidence, finding a negative correlation between cloud intensity and market concentration at the two-digit sectoral level. p. 4

Where the paper says it

Aggregating cloud intensity and market concentration at the two-digit sectoral level, we document a negative correlation between the two. While this association is descriptive in nature, it is consistent with the interpretation that the size-contingent returns to cloud use documented for smaller, ICT-capable firms may, in aggregate, be associated with less dispersed market structures across French industries.

The overall contribution of the paper is providing evidence for a conditional cloud exception to broader patterns of ICT-driven growth and concentration, shifting focus from firm age to firm size as a key determinant of cloud benefits. The conclusion emphasizes that SaaS reshapes internal organization, enabling small firms to digitalize and scale up, provided they possess the necessary complementary ICT capabilities. p. 0
Where the paper says it

Distinguishing cloud services between Infrastructure-as-a-Service (IaaS) and Software-as-a-Service (SaaS), we find that the size-contingent association is driven entirely by SaaS. Crucially, this small firm advantage is conditional on complementary assets as it emerges only among firms endowed with ICT capabilities.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improve cloud service adoption recommendations by distinguishing between Infrastructure-as-a-Service (IaaS) and Software-as-a-Service (SaaS). The improved system can recommend SaaS for growth benefits in smaller firms, while noting that IaaS use shows no significant base coefficient or interaction with size in growth analysis.

  2. Enhance firm growth forecasting by incorporating a conditional cloud exception based on firm capabilities. The system should suggest that the growth advantage from cloud use accrues to smaller firms when they possess complementary ICT capabilities, shifting the binding constraint on growth toward internal digital readiness.

  3. Refine workforce planning recommendations by linking SaaS adoption to specific occupational reallocation patterns. The improved system can model and suggest that SaaS use is associated with a reallocation of workers toward ICT occupations and upper-layer managerial and intermediate roles, and away from routine manual and clerical tasks.

  4. Optimize investment strategies by incorporating an identification strategy based on spatial exogenous variation. The system should use municipality-level lightning strike density as an instrument to predict cloud adoption, as it is associated with higher costs and reduced reliability of digital infrastructures, thereby capturing the tradeoff faced by internet providers.

  5. Develop risk assessment models for infrastructure investment using geographic controls. The improved system can incorporate municipality-level measures of elevation and terrain ruggedness to mitigate concerns that the instrumental variable is capturing geographical factors correlated with the incidence of lightning strikes.

  6. Inform industry concentration predictions by aggregating firm-level results sectorally. The system can use the equation Log Concentrations,t = α + β1Cloud Shares,t + βxControlss,t + 2-digit Ind.s + Yeart as a descriptive tool to show that industries where cloud use is more frequent tend to be less concentrated.

Abstract

Using a unique combination of micro-level data sources on French firms, this paper explores the relationship between cloud technologies and firm growth. Whereas digital technologies have generally been found to disproportionately benefit larger firms, we document that cloud use is positively associated with firm growth, and this association is significantly stronger for smaller firms. Distinguishing cloud services between Infrastructure-as-a-Service (IaaS) and Software-as-a-Service (SaaS), we find that the size-contingent association is driven entirely by SaaS. Crucially, this small firm advantage is conditional on complementary assets as it emerges only among firms endowed with ICT capabilities. Our results are robust across a range of specifications, including instrumental variable strategies. SaaS cloud use is further associated with internal workforce reorganisation toward ICT, managerial, and intermediate occupations and away from manual and clerical ones, particularly among smaller firms, consistent with SaaS cloud operating as an organisational innovation enabling digitalisation and scaling. Finally, we present descriptive sectoral evidence of a negative correlation between cloud and market concentration.

Sources

Related papers