## Traditional Model (wpiea2020127-print-pdf)

## Source details

**Canonical URL:** [Traditional Model (wpiea2020127-print-pdf)](https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020127-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2020/english/wpiea2020127-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2020/english/wpiea2020127-print-pdf.pdf.json)

---

### Background and definitions
- Digitalization and innovative use of digital technologies are changing work, learning, communication, and commerce, with cloud computing of growing importance.
- Cloud computing services provide on-demand Internet access to ICT resources: computing power, data storage capacity, software services, and operating system functionality, running on servers, storage devices, and networking equipment housed in data centers operated by the cloud provider.
- Virtualization and containerization enable pooled access to hardware and operating system functionality.
- Service types:
  - Infrastructure as a Service (IaaS): access to computing, storage, and/or networking resources alone.
  - Platform as a Service (PaaS): operating system and development tools plus infrastructure for application development.
  - Software as a Service (SaaS): access to software applications run on the provider’s infrastructure.
- Definition preserved: “Data latency is the time it takes for data packets to be stored or retrieved.”

### Prices (pricing characteristics and drivers)
- Pricing models and units:
  - Subscription and usage fees; numerous service packages with varying price-determining characteristics and pricing units.
  - IaaS compute capacity pricing categories: on-demand, pre-reserved increments, off-peak (low usage) periods.
  - Typical unit: per hour (sometimes per second); some providers moved from per hour to per second charges.
  - IaaS data storage pricing: typically per gigabyte per month, with lower rates for each subsequent gigabyte above thresholds; discounts for longer-term storage commitments.
  - Data transfer: charged based on amount transferred in and out.
  - PaaS: charged on computing capacity and data storage usage similar to IaaS.
  - SaaS: typically subscription fee per user per month; discounts for more users and/or longer service periods; usage-per-session charges less frequent.
- Location and currency effects:
  - Prices vary based on infrastructure location; higher prices where demand is higher.
  - Payments typically assessed in provider’s home country currency by default; some providers allow buyer currency with charges determined by seller’s offered exchange rates at transaction time.

### Cloud computing — industry size, growth, and concentration
- Industry revenue and forecasts:
  - Global public cloud computing services revenue: increase 17.5 percent in 2019 to US$214.3 billion.
  - Market revenue projected to reach $331.2 billion in 2022, which is 54.5 percent higher than the 2019 estimate.
  - SaaS expected to retain share at around 43 percent.
  - IaaS forecast to grow an estimated 25.9 percent annually on average, increasing its portion from 16.7 percent in 2018 to 23.1 percent in 2020.
- Market concentration:
  - Top five cloud service providers accounted for nearly 77 percent of the global IaaS market in 2018 (up from less than 73 percent in 2017).
  - Large providers cited include Amazon (AWS), Alibaba, Microsoft (Azure), Google, IBM, and others.
  - Market consolidation is expected as organizations seek standardized, broadly supported platforms.

### Demand-side adoption (OECD and emerging markets)
- OECD adoption:
  - Share of businesses in OECD countries purchasing cloud computing services increased nearly 6 percentage points between 2016 and 2018.
  - Information and communication sector increased cloud use by almost 10 percentage points between 2016 and 2018.
  - Professional, Scientific and Technical sector increased cloud use by almost 8 percentage points between 2016 and 2018.
  - In 2018, share of businesses purchasing cloud computing services ranged from 65 percent in Finland to 10 percent in Latvia.
  - Over 60 percent of firms in the Information and Communication sector use cloud computing services; Accommodation and Food and Beverage services: 21 percent.
  - Larger firms are more likely to purchase cloud services than small firms (All businesses vs 250+ vs 50–249 vs 10–49).
- Emerging markets:
  - China public cloud market grew 65.2 percent in 2018 and is expected to expand almost threefold by 2022 (China Academy of Information and Communications Technology).
  - Public cloud services revenue in India expected to reach $2.4 billion in 2019, an increase of 24.3 percent from 2018.

### Illustrative accounting example — Traditional Model (selected transactions and amounts)
- Four firms considered: IT equipment manufacturer (servers), accounting firm, clothing wholesaler, computer processing firm (provides cloud services).
- Traditional Model — selected items:
  - Gross Fixed Capital Formation – IT Equipment:
    - IT Equipment Manufacturer: 50,000
    - Accounting Firm: 50,000
  - Output:
    - Servers: 100,000
    - Accounting Services: 25,000
    - Wholesaling Services: 40,000
  - Intermediate Consumption:
    - Inputs related to internal IT services: Accounting Firm 5,000; Wholesaler 10,000
    - Other inputs: IT Equipment Manufacturer 50,000; Accounting Firm 10,000; Wholesaler 15,000
  - Gross Value Added:
    - IT Equipment Manufacturer: 50,000
    - Accounting Firm: 10,000
    - Wholesaler: 15,000
  - Compensation of Employees:
    - IT employees: Accounting Firm 3,000; Wholesaler 5,000
    - Non-IT employees: IT Equipment Manufacturer 25,000; Accounting Firm 5,000; Wholesaler 7,000
  - Operating Surplus, Gross:
    - IT Equipment Manufacturer: 25,000
    - Accounting Firm: 2,000
    - Wholesaler: 3,000

### Illustrative accounting example — Cloud Model (selected transactions and amounts)
- Cloud Model — selected items:
  - Gross Fixed Capital Formation – IT Equipment:
    - IT Equipment Manufacturer: 100,000
  - Output:
    - Servers: 100,000
    - Accounting Services: 25,000
    - Wholesaling Services: 40,000
    - Cloud Computing Services: 22,000
  - Intermediate Consumption:
    - Cloud Computing Purchased Services (IaaS, PaaS, SaaS): Accounting Firm 8,000; Wholesaler 14,000
    - Other inputs: IT Equipment Manufacturer 50,000; Accounting Firm 10,000; Wholesaler 15,000; IT Services Firm 15,000
  - Gross Value Added:
    - IT Equipment Manufacturer: 50,000
    - Accounting Firm: 7,000
    - Wholesaler: 11,000
    - IT Services Firm: 7,000
  - Compensation of Employees:
    - IT employees: IT Services Firm 6,000
    - Non-IT employees: IT Equipment Manufacturer 25,000; Accounting Firm 5,000; Wholesaler 7,000
  - Operating Surplus, Gross:
    - IT Equipment Manufacturer: 25,000
    - Accounting Firm: 2,000
    - Wholesaler: 4,000
    - IT Services Firm: 1,000
- Implication: accounting firm and wholesaler no longer record capital outlays for servers nor hire labor to manage servers; they record cloud services expenses purchased from the IT services firm, shifting where investment, value added, compensation, and operating surplus are recorded.

### Illustrative accounting example — Cloud Model with Rationalization (Table 3 key items)
- Selected transactions and accounts:
  - Gross Fixed Capital Formation – IT Equipment: 50,000
  - Output:
    - Servers: 50,000
    - Accounting Services: 25,000
    - Wholesaling Services: 40,000
    - Cloud Computing Services: 20,000
  - Intermediate Consumption:
    - Cloud Computing Purchased Services: 7,000 13,000
    - Other: 25,000 10,000 15,000 7,000
  - Gross Value Added: 25,000 8,000 12,000 13,000
  - Compensation of Employees (IT): 0 0 3,000
  - Compensation of Employees (Non-IT): 12,500 5,000 7,000
  - Operating Surplus, Gross: 12,500 3,000 5,000 10,000
- Aggregate outcome noted: economy-wide output will decrease since both the output of the manufacturing firm and cloud services firm declined (total output=135,000).
- Rationalization effect: reduction in output of servers, labor, and measured cloud computing services because cloud providers leverage average rather than maximum capacity.

### Cross-border provision, trade classification, and multi-territory examples
- Cloud provision can be located in a different economic territory than the purchaser; purchases may therefore be imports rather than domestic purchases.
- Table 4 — Cloud Model with Rationalization and Foreign Provision (selected items):
  - Exports of Servers: 50.000
  - Exports of Cloud Services: 20,000
  - Imports of Cloud Services: 7,000 13,000
  - Gross Fixed Capital Formation – IT Equipment / Input of IT Equipment: 50,000
  - Output:
    - Servers: 50,000
    - Accounting Services: 25,000
    - Wholesaling Services: 40,000
    - Cloud Computing Services: 20,000
  - Intermediate Consumption:
    - Cloud Computing Purchased Services: 7,000 13,000
    - Other: 25,000 10,000 15,000 7,000
  - Gross Value Added: 25,000 8,000 12,000 13,000
  - Compensation of Employees (IT): 0 0 3,000
  - Compensation of Employees (Non-IT): 12,500 5,000 7,000
  - Operating Surplus, Gross: 12,500 3,000 5,000 10,000
- Table 5 — Cloud Model with Rationalization and Foreign Provision with Multiple Data Centers (selected items):
  - Exports of Servers: 50,000
  - Exports of Cloud Services: 10,000 10,000
  - Gross Fixed Capital Formation – IT Equipment / Imports of IT Equipment: 25,000 25,000
  - Output:
    - Servers: 50,000
    - Accounting Services: 25,000
    - Wholesaling Services: 40,000
    - Cloud Computing Services: 10,000 10,000
  - Intermediate Consumption:
    - Cloud Computing Purchased Services: 7,000 13,000
    - Other: 25,000 10,000 15,000 5,000 2,000
  - Gross Value Added: 25,000 8,000 12,000 5,000 8,000
  - Compensation of Employees (IT): 0 0 1,000 2,000
  - Compensation of Employees (Non-IT): 12,500 5,000 7,000
  - Operating Surplus, Gross: 12,500 3,000 5,000 4,000 6,000
- BPM6 measurement guidance preserved: trade measured on an ownership basis — if a consumer in country A is billed by an establishment in country B that manages delivery, the import is attributed to country B even if actual computing occurs in country C; payments under contracts between B and C are imports by B from C.

### Price measurement and productivity evidence
- New service price indexes are required for cloud computing because services were previously deflated using goods prices when firms purchased equipment.
- Bundling, contracts, discounts, and incentives complicate price collection and compilation.
- Empirical findings:
  - Bank of Japan publishes a Services Producer Price Index for Entrusted computing services (cloud applications), 2015=100, showing steady decline since 2010 (category closely aligns with SaaS).
  - Byrne, Corrado, and Sichel (2018): constructed price indexes for selected AWS IaaS products in the U.S. and recorded quality-adjusted price declines of approximately 50 percent between 2009 and 2016.
  - Coyle and Nguyen (2019): found even larger price decreases for similar AWS products in the U.K. over the same period.
  - U.S. Producer Price Index for Hosting, ASP, and other IT infrastructure provisioning services was little changed over the same period, despite the product-specific price declines reported by researchers.

### Classification, business registers, and data source needs
- Classification gaps:
  - Phrases “cloud computing,” “IaaS,” “PaaS,” and “SaaS” do not appear in CPC Version 2.1 or ISIC Version 4, producing inconsistent classification across statistical agencies.
  - Present classification inconsistencies:
    - General agreement: IaaS classified in CPC subclass 83159, Other hosting and IT infrastructure provisioning services.
    - PaaS treatment varies; Eurostat Task Force suggests CPC class 8314; U.S. concords PaaS with CPC subclass 83159 in practice.
    - SaaS: Eurostat Task Force suggests classification with software publishing (CPC subclass 84392, Online software); U.S. distinguishes based on provider copyright — if provider holds copyright, classify with software publishing; otherwise classify as CPC subclass 83152, Application service provisioning (ASP).
  - Recommendation: consider renaming ISIC 6311 to Cloud Computing Services and include IaaS, PaaS, and SaaS (except where provider holds copyright) within the activity.
  - EBOPS classification (telecommunication, computer and information services) should be updated to include clear classes for cloud products.
- Business registers: NSOs must add emerging cloud computing enterprises and determine nature of affiliates’ operations (sales office vs. local delivery vs. R&D/IP production) for correct coding.

### New data collection needs (Table 6 summary)
- Purchases of IT processing services:
  - Vehicles: Business Surveys and Administrative Tax Forms
  - Required update: Most countries only collect IT consulting; few collect purchased service expenditures related to cloud computing (software subscriptions, processing time, storage space).
- Purchases of IT processing equipment:
  - Vehicles: Business Surveys, Administrative Tax Forms and Customs Forms
  - Required update: No change needed for collection of IT equipment purchases; possible greater consolidation among fewer firms.
- Purchases of Software:
  - Vehicles: Business Surveys, Administrative Tax Forms and Customs Forms, Own-account estimation methods
  - Required update: Need decision on distinction between purchased software (capitalized) and software subscriptions (service) because recording subscriptions as non-capital services would significantly affect macroeconomic accounts.
- Prices of cloud services:
  - Vehicles: Producer Price Reports, Consumer Price Reports
  - Required update: Develop new price indexes; hedonic methods likely required to capture pure price change amid rapid technological change.
- Trade in cloud services:
  - Vehicles: International Trade in Services Surveys
  - Required update: Expand surveys to include detailed questions on imports and exports of cloud services to clarify residence of providers within complex global supply chains.
- Example survey coverage issues: Statistics Canada and U.S. business surveys group cloud-related expenses into broad categories without specific guidance to identify cloud computing expenses.

### Sample sizes and survey coverage implications
- Wide-scale adoption of cloud computing expands the range of potential importers to include smaller firms in less-traditional importing industries.
- Without expansion of importer samples, import statistics risk under-coverage and drift.
- Surveys and statistical programs to examine for design and sample adjustments:
  - Imports of Commercial Services Surveys
  - Exports of Commercial Services Surveys
  - Surveys related to computer design services
  - Price report surveys
  - Annual and quarterly business statistics surveys for cloud activities
  - Business expense surveys

### Conceptual challenges for national accountants
- Capitalization of software and subscription services:
  - SNA guidance: software purchased and used in production for more than one year should be capitalized; multi-year lump-sum licenses recorded as assets; regular payment streams (monthly, quarterly) treated as service payments.
  - Ambiguous cloud scenarios requiring guidance:
    - Scenario a: firm sets up account, employees use specific software on as-needed basis for two years; billed according to usage.
    - Scenario b: firm subscribes month-to-month for three years.
    - Scenario c: firm subscribes for baseline of 20 users on a three-year contract with monthly billing for additional users.
  - Need guidance to identify when cloud services constitute an asset versus intermediate consumption and how to value assets under mixed or dynamic pricing models.
- Geographic attribution of cloud production:
  - Multi-jurisdictional storage and frequent relocation of data across data centers complicate assignment of production to an economic territory.
  - Pragmatic recommendation: attribute production to the location where activity predominately (greater than 50 percent) takes place.

### Aggregate implications for national account users (explicit list)
- Wide-scale shift to cloud computing can be expected to produce:
  - a softening in output and investment in IT equipment;
  - investment in IT equipment to become increasingly industrially and geographically concentrated;
  - increases in the cross-border flow of commercial services;
  - increases in the relative size of IT services industries;
  - increased concentration of IT related labor in those industries providing cloud services;
  - potential gains in productivity due to the rationalization associated with cloud computing;
  - appearances of new products and processes made possible by cloud computing; and
  - change in the geographic concentration of imports of ICT equipment.

### Major findings on computer services and cloud computing trade
- Quoted figures and country highlights:
  - "14.3 percent and 7.5 percent respectively."
  - Ireland is reported as the biggest exporter for computing services among OECD countries in 2016.
    - Ireland’s exports accounted for 29 percent of the total computing services exports of OECD countries in 2016.
    - Computing services represent 43 percent of Ireland’s total exports of services for the year.
    - Comparators: Germany 10 percent; the United States 3 percent.
    - Distinctiveness partly explained by multinational enterprises crediting a large proportion of global computing sales to business units resident in Ireland.
- Goods trade and server imports:
  - Servers are imported under HS codes: 847141; 847149; 847150 (computer servers without keyboard and monitors; largest and most relevant).
  - Trends:
    - The United States is the dominant importer of servers and its portion has continuously increased since 2014.
    - The import value for the United States increased more than threefold between 2009 and 2018.
    - Ireland reported significant growth in imports of servers (location of a number of data centers).

### Macroeconomic implications and measurement needs
- Areas affected by a wide-scale move to cloud computing:
  - Investment patterns
  - Trade flows
  - Economy-wide output
  - Value added
- Specific data improvements required:
  - A modified EBOPS classification for cloud computing services.
  - More information on residency and volume measures for GDP estimates.
  - Information regarding import and export prices of cloud computer services.

### Conclusion and policy implications
- The paper is a first attempt to outline challenges associated with accounting for wide-scale sales and purchases of cloud computing services.
- Key conceptual and measurement points:
  - Delivery of cloud computing services is concentrated and global.
  - The potential number of buyers is very large; given scale and global nature, implications on macroeconomic indicators and measurement could be large—impacting output, trade, investment and real value added.
  - Most NSOs and IOs are currently not well-placed to measure this transformative change; measurement requires a new set of tools and instruments.
  - While much of the challenge is measurement-related at the border, some conceptual lines may need to be crossed or moved so users can interpret and appreciate the changes occurring.

*Source: wpiea2020127-print-pdf — https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020127-print-pdf.pdf*

### 1. Traditional Model ...................................................................................................

### 1. Traditional Model

### Background
- Digitalization and innovative use of digital technologies are changing work, learning, communication, and commerce, with cloud computing of growing importance.
- Cloud computing changes ICT investment patterns from diversified across many industries to more concentrated ICT investment, and alters cross-border flows of commercial services because the provider need not be in the same economic territory as the purchaser.

### Cloud computing services (definitions and types)
- Cloud computing services provide on-demand Internet access to ICT resources: computing power, data storage capacity, software services, and operating system functionality, running on servers, storage devices, and networking equipment housed in data centers operated by the cloud provider.
- Virtualization and containerization enable pooled access to hardware and operating system functionality.
- Notable early provider: Amazon began offering pooled computing resources to external customers in 2006.
- Service types:
  - Infrastructure as a Service (IaaS): access to computing, storage, and/or networking resources alone.
  - Platform as a Service (PaaS): operating system and development tools plus infrastructure for application development.
  - Software as a Service (SaaS): access to software applications run on the provider’s infrastructure.

### Prices (pricing characteristics and drivers)
- Pricing models: subscription and usage fees; numerous service packages with varying price-determining characteristics and pricing units.
- IaaS compute capacity pricing:
  - Categories: on-demand, pre-reserved increments, off-peak (low usage) periods.
  - Typical unit: per hour (sometimes per second); some providers moved from per hour to per second charges.
  - Price depends on operating system, CPUs, memory, intermediate file storage, and other factors.
- IaaS data storage pricing:
  - Typically per gigabyte per month, with lower rates for each subsequent gigabyte above thresholds; discounts for longer-term storage commitments.
- Data transfer: charged based on amount transferred in and out.
- PaaS: charged on computing capacity and data storage usage similar to IaaS.
- SaaS: typically subscription fee per user per month; discounts for more users and/or longer service periods; usage-per-session charges less frequent.
- Location effect: prices vary based on infrastructure location; higher prices where demand is higher; customers may choose locations for privacy laws or small latency improvements.
- Currency: payments typically assessed in provider’s home country currency by default; some providers allow buyer currency with charges determined by seller’s offered exchange rates at transaction time.
- Definition: “Data latency is the time it takes for data packets to be stored or retrieved.”

### Cloud computing—The industry (market size, growth, concentration)
- Industry revenue and forecasts:
  - Global public cloud computing services revenue: increase 17.5 percent in 2019 to US$214.3 billion.
  - Market revenue projected to reach $331.2 billion in 2022, which is 54.5 percent higher than the 2019 estimate.
  - SaaS expected to retain share at around 43 percent.
  - IaaS forecast to grow an estimated 25.9 percent annually on average, increasing its portion from 16.7 percent in 2018 to 23.1 percent in 2020.
- Market concentration:
  - Top five cloud service providers accounted for nearly 77 percent of the global IaaS market in 2018 (up from less than 73 percent in 2017).
  - Large providers cited include Amazon (AWS), Alibaba, Microsoft (Azure), Google, IBM, and others.
  - Market consolidation is expected as organizations seek standardized, broadly supported platforms.

### Demand-side adoption (OECD and emerging markets)
- OECD businesses purchasing cloud services:
  - On average, the share of businesses in OECD countries purchasing cloud computing services increased nearly 6 percentage points between 2016 and 2018.
  - Information and communication sector increased cloud use by almost 10 percentage points between 2016 and 2018.
  - Professional, Scientific and Technical sector increased cloud use by almost 8 percentage points between 2016 and 2018.
- Country- and sector-level variation:
  - In 2018, share of businesses purchasing cloud computing services ranged from 65 percent in Finland to 10 percent in Latvia.
  - Over 60 percent of firms in the Information and Communication sector use cloud computing services; Accommodation and Food and Beverage services: 21 percent.
  - Larger firms are more likely to purchase cloud services than small firms (figure shows All businesses vs 250+ vs 50–249 vs 10–49).
- Emerging markets:
  - China public cloud market grew 65.2 percent in 2018 and is expected to expand almost threefold by 2022 (China Academy of Information and Communications Technology).
  - Public cloud services revenue in India expected to reach $2.4 billion in 2019, an increase of 24.3 percent from 2018.

### Challenges for national accountants
- Cloud computing changes transactions relative to the traditional model where firms self-provision IT services; this affects how national accountants record transactions in production, generation of income, and capital formation accounts.
- Stylized example contrasting Traditional Model and Cloud Model with four firms:
  - Firms considered: IT equipment manufacturer (servers), accounting firm (requires server), clothing wholesaler (requires server), computer processing firm (provides cloud services).

- Traditional Model (Table 1) — selected transactions and amounts:
  - Gross Fixed Capital Formation – IT Equipment:
    - IT Equipment Manufacturer: 50,000
    - Accounting Firm: 50,000
  - Output:
    - Servers: 100,000
    - Accounting Services: 25,000
    - Wholesaling Services: 40,000
  - Intermediate Consumption:
    - Inputs related to internal IT services: Accounting Firm 5,000; Wholesaler 10,000
    - Other inputs: IT Equipment Manufacturer 50,000; Accounting Firm 10,000; Wholesaler 15,000
  - Gross Value Added:
    - IT Equipment Manufacturer: 50,000
    - Accounting Firm: 10,000
    - Wholesaler: 15,000
  - Compensation of Employees:
    - IT employees: Accounting Firm 3,000; Wholesaler 5,000
    - Non-IT employees: IT Equipment Manufacturer 25,000; Accounting Firm 5,000; Wholesaler 7,000
  - Operating Surplus, Gross:
    - IT Equipment Manufacturer: 25,000
    - Accounting Firm: 2,000
    - Wholesaler: 3,000

- Cloud Model (Table 2) — selected transactions and amounts:
  - Gross Fixed Capital Formation – IT Equipment:
    - IT Equipment Manufacturer: 100,000
  - Output:
    - Servers: 100,000
    - Accounting Services: 25,000
    - Wholesaling Services: 40,000
    - Cloud Computing Services: 22,000
  - Intermediate Consumption:
    - Cloud Computing Purchased Services (IaaS, PaaS, SaaS): Accounting Firm 8,000; Wholesaler 14,000
    - Other inputs: IT Equipment Manufacturer 50,000; Accounting Firm 10,000; Wholesaler 15,000; IT Services Firm 15,000
  - Gross Value Added:
    - IT Equipment Manufacturer: 50,000
    - Accounting Firm: 7,000
    - Wholesaler: 11,000
    - IT Services Firm: 7,000
  - Compensation of Employees:
    - IT employees: IT Services Firm 6,000
    - Non-IT employees: IT Equipment Manufacturer 25,000; Accounting Firm 5,000; Wholesaler 7,000
  - Operating Surplus, Gross:
    - IT Equipment Manufacturer: 25,000
    - Accounting Firm: 2,000
    - Wholesaler: 4,000
    - IT Services Firm: 1,000

- Implication: Under the cloud services model, accounting firm and wholesaler no longer record capital outlays for servers nor hire labor to manage servers; instead they record cloud services expenses purchased from the IT services firm, shifting where investment, value added, compensation, and operating surplus are recorded in national accounts.

*Source: wpiea2020127-print-pdf — https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020127-print-pdf.pdf*

### introduction of the IT services firm and subsequent sale of cloud services. In the traditional

### wpiea2020127-print-pdf - introduction of the IT services firm and subsequent sale of cloud services. In the traditional

### Illustrative example: traditional model versus cloud services
- Traditional model: internal IT departments provided IT services and these services were not recorded as output by the firm because they were produced and consumed by the same firm.
- With the introduction of an IT services firm, IT services are produced outside the firm and sold on the market; national accountants must:
  - record the output of IT services in the IT services industry, and
  - show the purchase of these services by the consuming industries.
- The introduction of market-sold IT services adds a new product to both output and input matrices, producing a grossing up of economy-wide output and inputs by the value of the IT services.
- Economy-wide gross operating surplus increases due to efficiency gains under the illustrative assumptions.

### Table 3 — Cloud Model with Rationalization (key transactions and accounts)
- Gross Fixed Capital Formation – IT Equipment: 50,000
- Output:
  - Servers: 50,000
  - Accounting Services: 25,000
  - Wholesaling Services: 40,000
  - Cloud Computing Services: 20,000
- Intermediate Consumption:
  - Cloud Computing Purchased Services: 7,000 13,000
  - Other: 25,000 10,000 15,000 7,000
- Gross Value Added: 25,000 8,000 12,000 13,000
- Compensation of Employees (IT): 0 0 3,000
- Compensation of Employees (Non-IT): 12,500 5,000 7,000
- Operating Surplus, Gross: 12,500 3,000 5,000 10,000
- Aggregate outcome noted: economy-wide output will decrease since both the output of the manufacturing firm and cloud services firm declined (total output=135,000).
- Rationalization effect: reduction in output of servers, labor, and measured cloud computing services because cloud providers leverage average rather than maximum capacity.

### Cross-border provision, trade classification, and multi-territory examples
- Cloud provision can be located in a different economic territory than the purchaser; purchases may therefore be imports rather than domestic purchases.
- Table 4 — Cloud Model with Rationalization and Foreign Provision (selected items):
  - Exports of Servers: 50.000
  - Exports of Cloud Services: 20,000
  - Imports of Cloud Services: 7,000 13,000
  - Gross Fixed Capital Formation – IT Equipment / Input of IT Equipment: 50,000
  - Output:
    - Servers: 50,000
    - Accounting Services: 25,000
    - Wholesaling Services: 40,000
    - Cloud Computing Services: 20,000
  - Intermediate Consumption:
    - Cloud Computing Purchased Services: 7,000 13,000
    - Other: 25,000 10,000 15,000 7,000
  - Gross Value Added: 25,000 8,000 12,000 13,000
  - Compensation of Employees (IT): 0 0 3,000
  - Compensation of Employees (Non-IT): 12,500 5,000 7,000
  - Operating Surplus, Gross: 12,500 3,000 5,000 10,000
- Table 5 — Cloud Model with Rationalization and Foreign Provision with Multiple Data Centers (selected items):
  - Exports of Servers: 50,000
  - Exports of Cloud Services: 10,000 10,000
  - Gross Fixed Capital Formation – IT Equipment / Imports of IT Equipment: 25,000 25,000
  - Output:
    - Servers: 50,000
    - Accounting Services: 25,000
    - Wholesaling Services: 40,000
    - Cloud Computing Services: 10,000 10,000
  - Intermediate Consumption:
    - Cloud Computing Purchased Services: 7,000 13,000
    - Other: 25,000 10,000 15,000 5,000 2,000
  - Gross Value Added: 25,000 8,000 12,000 5,000 8,000
  - Compensation of Employees (IT): 0 0 1,000 2,000
  - Compensation of Employees (Non-IT): 12,500 5,000 7,000
  - Operating Surplus, Gross: 12,500 3,000 5,000 4,000 6,000
- Measurement guidance per BPM6: trade measured on an ownership basis — if a consumer in country A is billed by an establishment in country B that manages delivery, the import is attributed to country B even if actual computing occurs in country C; payments under contracts between B and C are imports by B from C.

### Price measurement and productivity evidence
- New service price indexes are required for cloud computing because services were previously deflated using goods prices when firms purchased equipment.
- Bundling, contracts, discounts, and incentives complicate price collection and compilation.
- Bank of Japan publishes a Services Producer Price Index for Entrusted computing services (cloud applications), 2015=100, showing steady decline since 2010 (category closely aligns with SaaS).
- Research findings:
  - Byrne, Corrado, and Sichel (2018) constructed price indexes for selected AWS IaaS products in the U.S. and recorded quality-adjusted price declines of approximately 50 percent between 2009 and 2016.
  - Coyle and Nguyen (2019) found even larger price decreases for similar AWS products in the U.K. over the same period.
  - U.S. Producer Price Index for Hosting, ASP, and other IT infrastructure provisioning services was little changed over the same period, despite the product-specific price declines reported by researchers.

### Classification, business registers, and data source needs
- Classification gaps: phrases “cloud computing,” “IaaS,” “PaaS,” and “SaaS” do not appear in CPC Version 2.1 or ISIC Version 4, producing inconsistent classification across statistical agencies.
- Present classification inconsistencies:
  - General agreement: IaaS classified in CPC subclass 83159, Other hosting and IT infrastructure provisioning services.
  - PaaS treatment varies; Eurostat Task Force suggests CPC class 8314; U.S. concords PaaS with CPC subclass 83159 in practice.
  - SaaS: Eurostat Task Force suggests classification with software publishing (CPC subclass 84392, Online software); U.S. distinguishes based on provider copyright — if provider holds copyright, classify with software publishing; otherwise classify as CPC subclass 83152, Application service provisioning (ASP).
- ISIC industry 6311 currently records data processing, hosting, and related services; recommendation: consider renaming ISIC 6311 to Cloud Computing Services and include IaaS, PaaS, and SaaS (except where provider holds copyright) within the activity.
- EBOPS classification (telecommunication, computer and information services) should be updated to include clear classes for cloud products.
- Business registers: NSOs must add emerging cloud computing enterprises and determine nature of affiliates’ operations (sales office vs. local delivery vs. R&D/IP production) for correct coding.

### New data collection needs (Table 6 summary)
- Data to be acquired and possible collection vehicles with required updates:
  - Purchases of IT processing services:
    - Vehicles: Business Surveys and Administrative Tax Forms
    - Required update: Most countries only collect IT consulting; few collect purchased service expenditures related to cloud computing (software subscriptions, processing time, storage space).
  - Purchases of IT processing equipment:
    - Vehicles: Business Surveys, Administrative Tax Forms and Customs Forms
    - Required update: No change needed for collection of IT equipment purchases; possible greater consolidation among fewer firms.
  - Purchases of Software:
    - Vehicles: Business Surveys, Administrative Tax Forms and Customs Forms, Own-account estimation methods
    - Required update: Need decision on distinction between purchased software (capitalized) and software subscriptions (service) because recording subscriptions as non-capital services would significantly affect macroeconomic accounts.
  - Prices of cloud services:
    - Vehicles: Producer Price Reports, Consumer Price Reports
    - Required update: Develop new price indexes; hedonic methods likely required to capture pure price change amid rapid technological change.
  - Trade in cloud services:
    - Vehicles: International Trade in Services Surveys
    - Required update: Expand surveys to include detailed questions on imports and exports of cloud services to clarify residence of providers within complex global supply chains.
- Example survey coverage issues: Statistics Canada and U.S. business surveys group cloud-related expenses into broad categories (Professional and business fees; Office and computer related expenses; Telephone, internet and other telecommunication expenses; Data processing and other purchased computer services) without specific guidance to identify cloud computing expenses.

### Sample sizes and survey coverage implications
- Wide-scale adoption of cloud computing expands the range of potential importers to include smaller firms in less-traditional importing industries.
- Without expansion of importer samples, import statistics risk under-coverage and drift.
- Surveys and statistical programs to examine for design and sample adjustments:
  - Imports of Commercial Services Surveys
  - Exports of Commercial Services Surveys
  - Surveys related to computer design services
  - Price report surveys
  - Annual and quarterly business statistics surveys for cloud activities
  - Business expense surveys

### Conceptual challenges for national accountants
- Capitalization of software and subscription services:
  - SNA guidance: software purchased and used in production for more than one year should be capitalized; multi-year lump-sum licenses recorded as assets; regular payment streams (monthly, quarterly) treated as service payments.
  - Cloud scenarios creating ambiguity (classification unclear whether asset or intermediate consumption):
    - Scenario a: firm sets up account, employees use specific software on as-needed basis for two years; billed according to usage.
    - Scenario b: firm subscribes month-to-month for three years.
    - Scenario c: firm subscribes for baseline of 20 users on a three-year contract with monthly billing for additional users.
  - Need for guidance to identify when cloud services constitute an asset versus intermediate consumption and how to value assets under mixed or dynamic pricing models.
- Geographic attribution of cloud production:
  - Multi-jurisdictional storage and frequent relocation of data across data centers complicate assignment of production to an economic territory.
  - Pragmatic recommendation suggested: attribute production to the location where activity predominately (greater than 50 percent) takes place.

### Aggregate implications for national account users (explicit list from source)
- All things equal, wide-scale shift to cloud computing can be expected to produce:
  - a softening in output and investment in IT equipment;
  - investment in IT equipment to become increasingly industrially and geographically concentrated;
  - increases in the cross-border flow of commercial services;
  - increases in the relative size of IT services industries;
  - increased concentration of IT related labor in those industries providing cloud services;
  - potential gains in productivity due to the rationalization associated with cloud computing;
  - appearances of new products and processes made possible by cloud computing; and
  - change in the geographic concentration of imports of ICT equipment.

*Source: wpiea2020127-print-pdf - introduction of the IT services firm and subsequent sale of cloud services. In the traditional (PDF).*

### 14.3 percent and 7.5 percent respectively.

### wpiea2020127-print-pdf - 14.3 percent and 7.5 percent respectively.

### Major findings on computer services and cloud computing trade
- "14.3 percent and 7.5 percent respectively."
- Ireland is reported as the biggest exporter for computing services among OECD countries in 2016.
  - Ireland’s exports accounted for 29 percent of the total computing services exports of OECD countries in 2016.
  - Computing services represent 43 percent of Ireland’s total exports of services for the year.
  - Comparators: Germany 10 percent; the United States 3 percent.
  - The distinctiveness of Ireland’s figure is partly explained by multinational enterprises crediting a large proportion of global computing sales to business units resident in Ireland.
- Current trade statistics provide a rough picture of cloud computing services trade; accurate capturing of the trade will grow in importance as the cloud computing industry expands.
- Measurement challenges arise from:
  - The scalability of cloud services.
  - Mobility between data centers.
  - The need for timely measurement of trade in services associated with cloud computing.

### Goods trade and server imports
- Servers are imported under three sub-categories of Harmonized System Codes (HS) 8471:
  - 847141
  - 847149
  - 847150 (computer servers without keyboard and monitors; largest and most relevant)
- Trends in server imports:
  - The United States is the dominant importer of servers and its portion has continuously increased since 2014.
  - The import value for the United States increased more than threefold between 2009 and 2018.
  - Ireland, a location of a number of data centers, reported significant growth in imports of servers.
- Figure notes preserved from source:
  - Figure 8: Relative Size of Computer Services Export in OECD Countries (2016, in U.S. Dollar).
  - Figure 9: Computer Services Export Ratio of Major Exporters (Percent).
  - Figure 10: Portion of Major Server (HS847150) Importers (Percent).
  - Figure 11: Share of Server Imports in Total Imports (Top five countries in 2017, percent).

### Macroeconomic implications and measurement needs
- Areas affected by a wide-scale move to cloud computing:
  - Investment patterns
  - Trade flows
  - Economy-wide output
  - Value added
- National accountants and balance of payments compilers need to provide granular, timely data to capture the transformation.
- Specific data improvements required:
  - A modified EBOPS classification for cloud computing services.
  - More information on residency and volume measures for GDP estimates.
  - Information regarding import and export prices of cloud computer services.

### Conclusion and policy implications
- The paper is a first attempt to outline challenges associated with accounting for wide-scale sales and purchases of cloud computing services.
- Key conceptual and measurement points:
  - Delivery of cloud computing services is concentrated and global.
  - The potential number of buyers is very large; given scale and global nature, implications on macroeconomic indicators and measurement could be large—impacting output, trade, investment and real value added.
  - Most NSOs and IOs are currently not well-placed to measure this transformative change; measurement requires a new set of tools and instruments.
  - While much of the challenge is measurement-related at the border, some conceptual lines may need to be crossed or moved so users can interpret and appreciate the changes occurring.

*Source: wpiea2020127-print-pdf - 14.3 percent and 7.5 percent respectively.*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020127-print-pdf.pdf_