The Maryland digital advertising tax, applied to gross revenue derived from digital advertising services, has a rate escalating from 2.5 percent to 10 percent of the advertising platform’s assessable base based on its annual gross revenues from all sources (i.e., not just digital advertising, and not just in Maryland). The escalating rate scale works to exclude from the tax any entity with less than $1 million of gross revenues from digital advertising services in Maryland and $100 million in annual gross revenues. It is not a typical progressive tax, inasmuch as the rate applies to all taxed activity, not just the marginal amount. The enabling legislation grants significant authority to the Comptroller of the Treasury to develop regulations implementing the tax, and it is on these draft regulations that the Tax Foundation submits comment.

Definitional Ambiguity

The proposed regulations inadequately answer an extremely important question: what exactly is subject to tax? The definition of digital advertising services is sufficiently vague to risk being either overbroad or discriminatorily narrow. In the bill, the definition of digital advertising services “includes advertisement services on a digital interface, including banner advertising, search engine advertising, interstitial advertising, and other comparable advertising services,” where a digital interface is defined as “any type of software, including a website, part of a website, or application, that a user is able to access.” (Interstitial advertising is advertising served within or between videos, or which briefly covers the full screen of an app or website.)

In terms of digital advertising served in the state, there are relatively easy cases, like a banner ad served to a user with an IP address in central Maryland whose linked account shows a Maryland mailing address. In other cases, however, determining what constitutes advertising is considerably more complex. Sponsored and branded content, email marketing, and even Voice-over-IP (VoIP) sales calls could all theoretically constitute digital advertising, with obvious complexities in assigning value or allocating them to Maryland users. Moreover, while some effort was made in subsequent legislation to address rebroadcast or online streaming of over-the-airwaves television and radio, which might contain advertisements placed for linear broadcasts, this exclusion is narrow and may be insufficient to solve the issue, particularly for national broadcasts.

An advertisement or promoted content on a microblogging platform like Twitter or Instagram plainly meets the definition for purposes of the law, but what of sponsored content on a non-corporate account, such as when a company pays a prominent account to promote its product? Interstitial advertising on a video platform or audio streaming site is relatively straightforward, but it is hardly meaningfully distinguishable from a podcast with a corporate sponsor. Or consider product placement in digital content streamed to Maryland users. Maryland’s tax will either have significant loopholes which discriminate against more traditional forms of advertising, or it will be an administrative nightmare which reaches almost every form of online content. The regulations, as drafted, fail to clear up the considerable uncertainty surrounding the tax base.

Suspect Sourcing Rules

The next issue pertains to the actual sourcing of revenue derived from digital advertising services in the state. The regulation suggests applying a worldwide base to apportion revenues. While the apportionment formula simplifies calculations, it also almost certainly taxes out-of-state activity, and indeed does not even reasonably approximate in-state revenues.

Revenue would be calculated by using the number of devices that have accessed digital advertising services from a location in the state as the numerator, and the number of devices that have accessed the digital advertising services from any location as the denominator. This fraction is then applied to the digital advertising gross revenue.

This apportionment formula is a very rough approximation of digital advertising revenue that could be expected to source to a state. However, it fails to actually consider real economic activity in Maryland and is highly likely to tax out-of-state activity. Consider a taxpayer (Company A) which makes $1,000 in revenue from running expensive video advertisements in Canada (90 percent of the revenue) and cheap banner ads in the United States (10 percent of the revenue). If 500 devices saw the videos in Canada and 500 saw the banner ads in the U.S., that would be a denominator of 1,000 devices. If we assume that 20 of the 500 U.S.-based devices were in Maryland, the numerator would be 20. Following the apportionment formula, we get:

In this example, Maryland claims revenue equaling 2 percent of the global gross revenue without considering how much revenue actually derived from Maryland even though it was considerably less—around 0.4 percent. Under such a system, if another state (or country) imposes a digital advertising tax based on revenue rather than devices, Company A is likely to experience double taxation. Note that the tax is applied to major platforms—search engines, social media, advertising networks, and the like—in aggregate, so an expensive ad campaign served outside of Maryland on behalf of one client would be partially apportioned to Maryland if another client had ads (even much cheaper ads) served into Maryland.

There is also an apportionment formula in the original bill which is wholly circular. This formula states that the numerator should be the annual gross revenues of a person derived from digital advertising services in the state, and the denominator should be the annual gross revenues of a person derived from digital advertising services in the U.S.

To use the above example, we both start and end with $20:

Since gross revenue derived from digital advertising is calculated based on devices rather than actual revenue figures, taxpayers may find themselves surprised to be liable. A taxpayer could have many Maryland-based devices viewing content without earning revenue above the threshold from business in Maryland. However, using the device-based rather than revenue-based apportionment formula could give a different outcome, which may result in tax liability. Rather than making sense of the bill’s sourcing rules, the regulations create new ones seemingly at odds with the legislative text and which do not fairly apportion income, as required by the Dormant Commerce Clause under U.S. constitutional law.

Unworkable Geolocation Requirements

The last issue discussed in these comments relates to locating devices. The regulation stipulates that, in ascertaining where an advertisement is served, taxpayers use:

• internet protocol (IP),
• geolocation data,
• device registration,
• cookies, or
• any other comparable information.

None of these data points is certain to actually place a device in or out of Maryland. For instance, users commonly utilize Virtual Private Networks (VPN), and there is ambiguity surrounding IP addresses near state lines that create scenarios where ads are improperly determined to have been served into Maryland. A Maryland-based corporation could, for instance, require employees to connect to company servers through VPN when working from home. In such a situation, the IP address would indicate that content was viewed from Maryland. Yet, the employee could, technically, be anywhere in the world.

The regulation recognizes that devices may be impossible to locate. In that event, the apportionment formula throws out all the non-located devices from the denominator. Going back to the example above: if 1,000 devices viewed advertisement, but 400 could not be located, they would be thrown out from the denominator.

This throwout rule increases the amount of gross revenue apportioned to the state—again without consideration of actual economic activity. Instead of $20, it is now $33.33. In other words, taxpayers are punished for being unable to locate the devices viewing their advertising services.

This is highly problematic, not only because using devices as a proxy for revenue is suspect, but also because the throwout rule could result in taxation of more than 100 percent of revenue. In this example, Maryland’s apportionment formula disregards 40 percent of the devices, and by doing so, it assumes a higher percentage of revenue is attributable to Maryland than is otherwise justified by the method. It assumes that any revenue attributable to the 400 devices remains untaxed by other jurisdictions.

Throwout rules—and their more common cousin, throwback rules—exist in the corporate income tax codes of 21 states and the District of Columbia. They are deeply flawed, but as they currently exist, they do require at least one-half of any transaction to be attributable to a given state, a precept that Maryland ignores.

In a traditional throwout rule situation, a company might sell something from State A into State B, but lack taxable nexus in State B, so State A “throws back” (or “throws out”—a question of whether the sale is excluded from the numerator or the denominator) the sale to itself for apportionment purposes. At least in these situations, the sale originates in the taxing state.

There is no such requirement here. Many have argued (Tax Foundation included) that a state-level digital advertising tax could increase costs for in-state businesses and consumers. This remains a crucial issue with the tax, but because the proposed apportionment is based on devices rather than revenue, it is likely that Maryland’s tax will increase costs of advertisement nationwide (maybe worldwide). Taxpayers will face higher taxes in Maryland when they earn more—regardless of where revenue is earned. As a result, taxpayers may need to adjust prices elsewhere to account for Maryland’s tax. The only other option available is to limit access by limiting access to services for Maryland-based devices.

The only other economic option, at least. Companies may have legal recourse, particularly given the design of these regulations, which impermissibly interfere with interstate commerce.

There may be no good way to apportion a digital advertising tax, especially when devices cannot be identified, but defaulting to user addresses, or even using population-based apportionment, would be superior to a throwout rule which, if similarly implemented in other states, could see all taxing states dramatically overestimating the share of activity associated with their state.

Broader Considerations

In addition to the regulatory issues raised above, Maryland is currently involved in legal proceedings regarding several aspects of the design of the tax. First, traditional advertising is not taxed in Maryland, only digital advertising, which is likely in violation of the federal Permanent Internet Tax Freedom Act, which protects online businesses from punitive or discriminatory taxation. Second, the application of a threshold related to global revenues may violate the Commerce Clause as it results in a higher tax rate for multistate or multinational companies than Maryland-based ones, and because the tax is arguably neither internally nor externally consistent. Third, targeting one type of speech with this tax could violate the First Amendment of the U.S. Constitution.

Internal and external consistency are crucial, as both are part of the test for whether a state’s tax unconstitutionally interferes with interstate commerce. External consistency is usually a given; it simply requires that the tax is reasonably related to the taxed activity. Internal consistency is at least moderately harder: the rule is that, if every state adopted the same tax system, this would not result in double taxation.

By focusing on the number of devices, Maryland may have strengthened its case for internal consistency at a cost of external consistency, though both remain active legal issues, particularly given the design of the throwout rule. The proposed regulations do not come even remotely close to apportioning based on the actual revenues derived from advertising into Maryland; they simply look at Maryland’s percentage of all ads, regardless of cost or value, served by a particular platform. And it is not even clear that the regulations’ approach to apportioning ad revenue to Maryland is consistent with the language of the statute. While the bill granted the Comptroller significant authority on apportionment, the proposed rule seems to advance a new approach altogether, separate from the one contemplated in the bill.

Digital advertising taxes are bad taxes even without these legal and administrative challenges. Even if the state had proposed a more reasonable method for determining the tax base and solved the legal ambiguities surrounding other parts of the design, a digital advertising tax remains nontransparent and non-neutral. Digital advertising profits are already subject to ordinary business taxes, and policymakers have not articulated a vision for why they should be subject to a second level of targeted taxation.

While it is not within the purview of the Comptroller’s office to reject a legislatively-approved tax no matter how poor—or constitutionally infirm—the underlying policy, the regulations fairly to fully resolve several important issues raised by the bill, while potentially creating additional legal infirmities in the regulation’s approach to apportionment.