In my last commentary, I noted that the Treasury and IRS recently issued proposed regulations relating to the classification of cloud transactions and transactions involving digital content.1 That commentary addressed the element of that guidance which proposes to amend the guidance on the source of income from the sale of digitally delivered software and content under the existing §1.861-18 regulations dealing with computer programs.2 This commentary will address the principal purpose of these proposed regulations, which is to add a new section to the §861 regulations, §1.861-19, to deal with the classification of cloud transactions.

The proposed regulations start with the proposition that the purpose of the regulation is to classify a cloud transaction as either a provision of services or a lease of property.3 The proposed regulations define a ‘‘cloud transaction’’ as follows:

A cloud transaction is a transaction through which a person obtains on-demand network access to computer hardware, digital content (as defined in §1.861-18(a)(3)), or other similar resources, other than on-demand network access that is de minimis taking into account the overall arrangement and the surrounding facts and circumstances. A cloud transaction does not include network access to download digital content for storage and use on a person’s computer or other electronic device.4

This definition apparently is based on a definition of cloud computing formulated by the National Institute of Standards and Technology which is frequently cited in the industry.5 This definition would seem to cover those arrangements commonly described as Software as a Service (‘‘SaaS’’), Platform as a Service (‘‘PaaS’’), and Infrastructure as a Service (‘‘IaaS’’). The effect of the second sentence of the definition is to ensure that those transactions where the main point of the transaction for the user is to obtain access to downloaded content are classified under the software regulations of §1.861-18.

The first classification rule is that transactions will be classified ‘‘solely’’ as either a lease of computer hardware or content or a provision of services, taking into account the factors described below. This is an important point, and is a prudent foundation of this guidance. In many cases, a cloud transaction may include elements that, if analyzed separately, could demonstrate different characters. For example, a transaction which has as its predominant character a SaaS transaction, still might involve some modest amount of computer code downloaded to the user’s hardware that performs some part of the software functionality. There would be no particular benefit to the proper administration of the law to require taxpayers to bifurcate those elements of a single transaction, and apportion the consideration for the cloud transaction into separate parts for the two elements.

The core of the proposed regulations is the list of factors that are to be used to classify the transaction as a service or a lease. The proposed regulations list nine factors. Those factors are not necessarily exclusive, as the proposed regulations state that ‘‘all relevant factors’’ including the ones listed, are to be taken into account. The proposed regulations further provide that ‘‘the relevance of any factor varies depending on the factual situation, and one or more of the [listed] factors ... may not be relevant in a given instance.’’6

Cloud transactions, of course, have been around for some time, so Treasury is playing a bit of catch-up in this area. In the meantime, taxpayers have had to look to analogous areas of law to classify their transactions. In general, the most analogous framework is the list of factors contained in §7701(e), which is to be used to determine whether a contract ‘‘which purports to be a service contract’’ should instead be treated as a lease of property.7 Section 7701(e) codified prior case law which had expressed various factors to distinguish service from lease contracts, but Congress did not include all of the factors relied upon in the prior case law in §7701(e). In the proposed regulations, Treasury has dipped into the prior case law to supplement the §7701(e) factors. The proposed regulations list the following nonexclusive factors demonstrating that a cloud transaction is classified as the provision of services rather than a lease of property:

(i) The customer is not in physical possession of the property;

(ii) The customer does not control the property, beyond the customer’s network access and use of the property;

(iii) The provider has the right to determine the specific property used in the cloud transaction and replace such property with comparable property;

(iv) The property is a component of an integrated operation in which the provider has other responsibilities, including ensuring the property is maintained and updated;

(v) The customer does not have a significant economic or possessory interest in the property;

(vi) The provider bears any risk of substantially diminished receipts or substantially increased expenditures if there is nonperformance under the contract;

(vii) The provider uses the property concurrently to provide significant services to entities unrelated to the customer;

(viii) The provider’s fee is primarily based on a measure of work performed or the level of the customer’s use rather than the mere passage of time; and

(ix) The total contract price substantially exceeds the rental value of the property for the contract period.

With some rewording to tailor the factor to the cloud context (and to reverse the unusual construction of §7701(e) which was drafted to overcome the purported character of a transaction documented as a service), the Reg. §1.861-19 list incorporates all six of the existing §7701(e) factors, plus it adds numbers (iii), (iv), and (viii) from the pre-§7701(e) case law.

With so many factors in the mix, the first question is whether there are any superfactors that will normally be critical to the distinction. A second question is whether perhaps Treasury thought that the three new factors were particularly germane to cloud transactions, and thus might attract extra attention in the weighting analysis.

With respect to the three additional factors, numbers (iii) and (iv) — the right to determine the specific property used in the transaction, and the property being a component of an integrated operation — properly reflect characteristics of cloud service transactions. The latter in particular, by its inclusion of both hardware and software functionality, would seem to be a core descriptor of many cloud service business models. At first blush, one might think the first factor doesn’t really add anything to the factor looking to which party ‘‘controls’’ the property. It is worth noting, however, that the ‘‘control’’ factor has been challenging to apply in prior cases due to the ambiguity of the word ‘‘control’’ when applied to property, so the factor (iii) supplementary guidance on the sort of control that is relevant in the cloud transaction space is useful.8

It is less clear that the third new factor, (viii), crisply distinguishes service from lease transactions. It seems correct that a compensation model based on work performed or level of use is more likely to indicate a service transaction, as the most common business models in those cases would relate to payments 6 Prop. Reg. §1.861-19(c)(1). for processing power or transaction processing. On the other hand, payments measured by the passage of time might not necessarily indicate a lease transaction. SaaS subscriptions normally are concluded on a time basis (e.g., annually), but SaaS transactions normally are clearly on the services side of the line. This is one case where the other factors would apply to cause the overall character to be that of a service.

The proposed regulations include 11 helpful examples, eight of which conclude that the example given falls within the definition of a cloud transaction, while three are deliveries of software that are classified under Reg. §1.861-18. All eight examples of cloud transaction conclude that ‘‘service’’ is the proper character. There is no apparent superfactor applied in these examples; in all cases, Treasury methodically applied most of the factors to reach its conclusion.

One forlorn factor, the factor of whether the total contract price exceeds the rental value of the property for the contract period, was applied in only two of the examples. It would seem that in most cases, this factor would be a clear pointer towards a service characterization. Even if ‘‘the property’’ mentioned in the factor would be only the hardware, the hardware in a cloud service transaction would be configured with an operating system and applications necessary to provide the data hosting or other service, and the service provider normally would be responsible for providing connectivity, a secure environment for the equipment, and the like. Once the service includes other software elements as necessary to constitute a SaaS or PaaS offering, the service fee normally would exceed a proportionate part of a hardware rental fee by a greater margin.

On the other hand, one of the major value propositions of a cloud computing model is that providers can offer computing power more cheaply through their centralization and management of shared hardware assets, as compared to the old model where each user needed to purchase (or lease) its own equipment. That ability to reduce prices for users could interact strangely with this factor, as it doesn’t make much sense to cause this factor to point toward a lease characterization if the efficiencies of the cloud service provider can reduce the total contract cost to less than the rental value of the hardware.

Finally, does Treasury’s conclusion in all of the eight cloud transaction examples that the proper characterization is ‘‘service’’ suggest that this analysis should not be expressed as a balancing with two possible results? Treasury has specifically requested that taxpayers provide ‘‘realistic’’ examples of cloud transactions that would be treated as leases under Prop. Reg. §1.861-19. That will be challenging, as the nature of cloud computing is the aggregation of assets, both hardware and software, in a form in which the provider can provide the hardware and software functionality more efficiently than individual users could if they had to build their own data centers and host software applications themselves. By and large, this economic essence of cloud computing appears in various of the Prop. Reg. §1.861-19 factors.

Users who choose to buy or lease assets are following a different investment, functional and risk strategy than those who meet their infrastructure and software needs through cloud transactions. In light of the fact that Treasury so far has not been able to come up with a realistic cloud transaction that is properly characterized as a lease, perhaps the better framework of these proposed regulations is to conclude as a general matter that cloud transactions as defined in the regulations should be classified as a service, absent unusual factual circumstances. The regulations would continue to rely on the §7701(e) factors, as enhanced.

That would reduce controversies, by removing the implication that ‘‘cloud transactions’’ as defined in the regulations should be arrayed on a continuum between services at one end and leases at the other. The definition itself, referring to ‘‘on-demand network access’’ to hardware and software resources, neatly limits the transactions falling within that definition to service business models.


First published in the Bloomberg Industry Group's Tax Management International Journal, Vol. 48 No. 11, November 8 2019.



1 REG-130700-14, Classification of Cloud Transactions and Transactions Involving Digital Content (Aug. 9, 2019).

2 All section references are to the Internal Revenue Code (the ‘‘Code’’), as amended, or the regulations thereunder existing or proposed, unless otherwise indicated.
3 Prop. Reg. §1.861-19(a).
4 Prop. Reg. §1.861-19(b).
5 Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction. NIST, Special Publication 800-145, at 2 (Sept. 2011)
6 Prop. Reg. §1.861-19(c)(1).
7 Sprague, Characterizing Cloud Transactions — Applying Section 7701(e) to Remote Access Transactions, Part I, 42 Tax Mgmt. Int’l J. 559 (Sept. 13, 2013); Sprague, Characterizing Cloud Transactions — Applying Section 7701(e) to Remote Access Transactions, Part II, 42 Tax Mgmt. Int’l J. 701 (Nov. 8, 2013).
8 Sprague, Characterizing Cloud Transactions — Applying Section 7701(e) to Remote Access Transactions, Part II, 42 Tax Mgmt. Int’l J. 701 (Nov. 8, 2013).
Explore More Insight