Can the states keep secrets from the federal government?

• Author: Mikos, Robert A.
• Position: II. Commandeering's Harms through III. Rebuffing State Autonomy: the Prevailing View Regarding Demands For States' Secrets, p. 121-153

2. COMMANDEERING'S HARMS

   In this Part, I identify and discuss the two primary harms caused by the commandeering of states' secrets--harms that have been overlooked or misjudged by courts and legal scholars. First, commandeering states' secrets has economic costs, which drain state resources. The direct costs of complying with a federal subpoena or reporting requirement are usually trivial, but the dynamic costs are not. By increasing the incentives of citizens to conceal activity from state regulators, the threat of commandeering makes it more difficult for states to gather information in the first instance. Second, commandeering states' secrets has political costs. One such cost is that commandeering forces state officials to advance federal policies that they or their constituents may deem objectionable. A second political cost is that commandeering blurs the lines of political accountability. State officials might be unfairly blamed for providing information to federal officials and advancing controversial federal policies. Importantly, these economic and political costs correspond to the costs animating the Supreme Court's anticommandeering decisions. They also form the normative basis for a viable constitutional federalism challenge to the commandeering of states' secrets under the anti-commandeering rule. (82)

   1. The Economic Costs

      To begin, commandeering states' secrets can make it far more difficult for states to gather data about regulated activity. Commandeering states' secrets increases citizens' incentives to conceal their activity from state monitors, because it links the work of those monitors to the enforcement of federal law and the imposition of federal sanctions. As a result, states must spend more to get the same information they would have obtained without commandeering. (83)

      As explained earlier, (84) individuals engaging in regulated activity have an incentive to conceal that activity from the government to evade the costs of regulation. Commandeering introduces new costs into a citizen's decision calculus. A citizen who is deciding whether to divulge information requested or demanded by a state official now must consider how that information will be used by both state and federal authorities. In particular, she must consider the increase in the expected federal sanction that stems from giving the federal government access to the information. If the increase is large enough, the citizen may withhold information from the state, (85) thereby forcing the state to obtain the information using other, more costly means.

      To see how even small changes (real or perceived) in expected federal sanctions can discourage citizens from cooperating with state officials, consider an example involving a proposed state marijuana tax. (86) In particular, suppose that a state legalizes the distribution of marijuana, but imposes a 10% tax on such sales--a level roughly equivalent to what it considers the harms attributable to marijuana. To help enforce the tax, the state requires all marijuana distributors to obtain a license and to record and report all sales to the state's revenue collection agency. To give this reporting requirement some teeth, the state threatens to impose a fine equivalent to 200% of unreported sales, and it employs gumshoe detectives to identify reporting violations. Suppose these detectives uncover 10% of all unreported sales. In this hypothetical, a risk-neutral marijuana distributor would have an incentive to fully report her taxable sales to the state. Doing so exposes her to a certain 10% tax on reported sales, but that is less than the expected fine of 20% (200% x 10%) on unreported sales.

      Notwithstanding state law, of course, the distribution of marijuana remains a federal crime. Suppose, for ease of illustration, that the federal government threatens to impose sanctions on marijuana distribution equivalent to 100% of the value of marijuana sold. Suppose as well that the federal government can detect only 5% of all marijuana sales using its own agents. This likely would not be enough to deter marijuana distribution; it effectively amounts to a 5% tax on the drug. But suppose the United States Attorney threatens to commandeer the state's tax records to bolster enforcement of the federal ban. With those records, the federal government could detect 100% of all sales reported to the state tax agency, in contrast to only 5% of all unreported sales. (87) In this case, the distributor would have a strong incentive not to report to the state agency or pay the state marijuana tax. If she reports the transactions, she now faces $10 in state tax for every $100 in sales plus $100 in federal sanctions--a total expected "sanction" of $110. (88) By contrast, if she does not report, she faces only $20 in expected state sanctions and another $5 in expected federal sanctions, for a total of only $25 in expected sanctions. The dramatic increase in the probability of being caught and sanctioned by the federal government makes reporting extremely costly to regulated parties and seriously undermines state information-gathering programs.

      Since the work product of state gumshoe detectives can also be commandeered by the federal government, this information-gathering strategy can also be undermined by federal commandeering. As discussed above, gumshoe detectives frequently depend on cooperation from private citizens-crime victims, witnesses, etc.--in conducting their investigations. Citizens may be less forthcoming, however, if the information they give to state agents is turned over to federal law enforcement. Immigrant communities, for example, seem reluctant to cooperate with local police departments because they fear local police will share information about their own (or a loved one's) immigration status with federal immigration authorities. (89) In short, whether states gather information via their own agents or citizens, they will have a harder time getting that information when the federal government threatens to take it and use it for its own purposes.

      The severity of the disruption to state information gathering depends on the marginal expected federal sanction attributable to commandeering. (90) The expected federal sanction due to commandeering is a function of two variables. First, it depends on the magnitude of the gross sanction that would be imposed by the federal government as a result of having the information. In some situations, of course, the federal sanction is small or even nonexistent, as is the disruption to state information gathering. In Nevada, for example, state regulators closely monitor brothels--legal in some counties--in order to protect the health of prostitutes and customers. (91) Among other things, the state requires prostitutes to undergo regular health tests to screen for various sexually transmitted diseases. (92) The state succeeds at gathering such sensitive information because (1) prostitutes gain a lucrative benefit by reporting themselves--namely, the state's permission to engage in the sex trade; and (2) the federal government does not punish prostitution. (93) In other situations, by contrast, the threatened federal sanction is large, and commandeering threatens to disrupt the flow of information to the state. The federal sanctions for marijuana distribution, for example, are quite large: distributors face long prison terms, large fines, and civil forfeiture actions under federal law. (94) Distributors might forego reporting to state regulators--and whatever regulatory benefits such regulators offered them--to avoid these federal sanctions. (95) Similarly, deportation is clearly harsh enough to dissuade some immigrant communities from cooperating with state law enforcement. (96)

      Second, the disruption also depends on how much the commandeering of the state's information would increase the probability of being detected by the federal government. For example, if a citizen is sure to be caught by federal authorities anyway, the marginal cost of providing incriminating information to the state will be zero, no matter how large the gross federal sanction. She would report her activity to the state despite the threat of commandeering, as long as state law gave her sufficient inducement to do so. More realistically, however, the probability of being detected by federal authorities absent commandeering will be less than one--indeed, in many cases, the probability approaches zero. (97) The point of commandeering states' secrets is, after all, to boost detection of federal offenses that would otherwise escape detection or be difficult to prove.

      In a related vein, the disruption also depends on citizens' expectations about whether the federal government would actually commandeer information. The federal government does not demand all of the information gathered by the states, even though it arguably could (given current law). For example, until recently, the federal government seemed reluctant to commandeer state medical marijuana registries, perhaps reflecting a reluctance to prosecute medical marijuana patients. (98) It stands to reason that if citizens believe the federal government will not commandeer (or use) the information, commandeering's theoretical availability will not necessarily burden state information gathering. (99)

      Nevertheless, the fact that commandeering is not inevitable mitigates, but does not eliminate, the costs imposed on the states. After all, the mere threat of federal commandeering--even if it does not materialize--may be enough to discourage some citizens from supplying states with information. Maine, for example, recently ended its compulsory medical marijuana registration system in part out of concerns that the system could be co-opted and exploited by federal agents. (100) Barring constitutional protection for states' secrets, there can be no assurance the federal government will refrain from commandeering altogether, (101)...