Tag: law

The “Read the Bill” Debate and Government Growth

There’s an interesting back-and-forth over at the Volokh Conspiracy about whether legislators should have to read the actual legislative text of bills they vote on. Most people’s intuitive reaction is: “Duh, of course!” But if you’ve ever actually spent time poring over legislative text, you know that reading the bill itself seldom leaves you with a very good sense of what it does. Legislation is typically a tangle of modifications along the lines of “Strike paragraph 2, replace the period with a semicolon, insert the word ‘reasonable’ in the following sentence…”—which is why legislators have staffers who prepare plain-English summaries of the effects of legislation. Now certainly it would be possible to render bills somewhat more readable to ordinary people. Saving paper is not a huge concern in the digital era, so there’s no good reason legislation couldn’t simply contain the full text of the statutory provisions it amended, perhaps including a side-by-side comparison highlighting the changes. Even this, however, wouldn’t necessarily be all that illuminating. I’ve got a reference book on my desk that contains the 80-or-so pages of the Foreign Intelligence Surveillance Act, and then a few hundred pages explaining what it actually means. It’s not enough to know what the verbatim text says; you need to understand how it interacts with other statutes, how key terms are defined in the law, how courts have interpreted the law’s provisions, and so on.

Legislation could be written in a somewhat more transparent way, but in light of all these complex interactions, it can’t actually be that much more transparent, for the same reason computer programs are a lot longer and more impenetrable than a plain-English description of what the program does. Achieving a result in a complex rule-based system requires a level of precision and sensitivity to how terms are used within the system that’s at odds with colloquial description. Of course, for precisely the same reason that summaries will give an ordinary person a better understanding of a law than scrutiny of the verbatim text, they also give a very incomplete understanding. An ordinary language description will tell you what a computer program is supposed to do. If you want to know whether it’s going to crash or open up a security vulnerability under certain conditions, perhaps when it interacts with other software running simultaneously, you need to have a look at the source code. Again, if you’ve spent any time digging through legislation, you know that the staff summary of a bill often glosses over many interesting little details and ambiguities you can ferret out while reading the text.

Most legislators, of course—even those with legal training—cannot possibly have the kind of expertise needed to undertake meaningful scrutiny of the details of legislative text outside a tiny number of issue areas. So does it make sense to insist that every member of Congress literally “read the bill”? Probably not. The actual text will contain important details not captured in a summary, but only an expert will really understand what those are on the basis of the text anyway. Crucially, this is not a function of needless obscurantism on the part of Congress: it is a necessary feature of legislation in a legal system as complex as ours. Which means that there’s a pretty basic tension between the value of democratic transparency and a large, complex government. Past a certain point, it’s more or less impossible for any individual legislator—let alone ordinary citizens—to really understand the vast majority of bills Congress takes up in any detailed way.

Prosecutors Should Not Be Allowed to Fabricate Evidence

In 1977, county attorney David Richter and assistant county attorney Joseph Hrvol worked side by side with police to investigate and “solve” the notorious murder of a former police officer in Pottawattamie County, Iowa. The prosecutors fabricated evidence and used it to charge and convict Curtis McGhee and Terry Harrington, sending them to prison for 25 years.

After the convictions were overturned for prosecutorial misconduct, McGhee and Harrington sued the county and prosecutors. The defendants in that civil suit invoked the absolute immunity generally afforded prosecutors to try to escape liability. After the Eighth Circuit ruled against them, the Supreme Court agreed to review the case.

On Friday, Cato joined the National Association of Criminal Defense Lawyers and the ACLU on a brief supporting the men unjustly imprisoned. We argue that prosecutors should be responsible for their role in manufacturing a false “case,” just as police officers would be under the same circumstances. As the Court has held, prosecutors enjoy absolute immunity only during the prosecutorial phase of a case, not its investigatory phase. Were prosecutors to receive absolute immunity here, citizens would have no protection from or recourse against prosecutors who frame the innocent by fabricating evidence and then using that evidence to convict them.

To read Cato’s brief in the case of Pottawattamie County v. McGhee, see here.

A Chance to Fix the PATRIOT Act?

As Tim Lynch noted earlier this week, Barack Obama’s justice department has come out in favor of renewing three controversial PATRIOT Act provisions—on face another in a train of disappointments for anyone who’d hoped some of those broad executive branch surveillance powers might depart with the Bush administration.

But there is a potential silver lining: In the letter to Sen. Patrick Leahy (D-VT) making the case for renewal, the Justice Department also declares its openness to “modifications” of those provisions designed to provide checks and balances, provided they don’t undermine investigations. While the popular press has always framed the fight as being “supporters” and “opponents” of the PATRIOT Act, the problem with many of the law’s provisions is not that the powers they grant are inherently awful, but that they lack necessary constraints and oversight mechanisms.

Consider the much-contested “roving wiretap” provision allowing warrants under the Foreign Intelligence Surveillance Act to cover all the communications devices a target might use without specifying the facilities to be monitored in advance—at least in cases where there are specific facts supporting the belief that a target is likely to take measures to thwart traditional surveillance. The objection to this provision is not that intelligence officers should never be allowed to obtain roving warrants, which also exist in the law governing ordinary law enforcement wiretaps. The issue is that FISA is fairly loosey-goosey about the specification of “targets”—they can be described rather than identified. That flexibility may make some sense in the foreign intel context, but when you combine it with similar flexibility in the specification of the facility to be monitored, you get something that looks a heck of a lot like a general warrant. It’s one thing to say “we have evidence this particular phone line and e-mail account are being used by terrorists, though we don’t know who they are” or “we have evidence this person is a terrorist, but he keeps changing phones.” It’s another—and should not be possible—to mock traditional particularity requirements by obtaining a warrant to tap someone on some line, to be determined. FISA warrants should “rove” over persons or facilities, but never both.

The DOJ letter describes the so-called “Lone Wolf” amendment to FISA as simply allowing surveillance of targets who are agents of foreign powers without having identified which foreign power (i.e. which particular terrorist group) they’re working for. They say they’ve never invoked this ability, but want to keep it in reserve. If that description were accurate, I’d say let them. But as currently written, the “lone wolf” language potentially covers people who are really conventional domestic threats with only the most tenuous international ties—the DOJ letter alludes to people who “self-radicalize” by reading online propaganda, but are not actually agents of a foreign group at all.

Finally, there’s the “business records” provision, which actually covers the seizure of any “tangible thing.”  The problems with this one probably deserve their own post, and ideally you’d just go through the ordinary warrant procedure for this. But at the very, very least there should be some more specific nexus to a particular foreign target than “relevance” to a ongoing investigation before an order issues. The gag orders that automatically accompany these document requests also require more robust judicial scrutiny.

Some of these fixes—and quite a few other salutary reforms besides—appear to be part of the JUSTICE Act which I see that Sen. Russ Feingold (D-WI) introduced earlier this afternoon.  I’ll take a closer look at the provisions of that bill in a post tomorrow.

Washington Legal Foundation Opposes GBS Deal

Via James Grimmelmann, the Washington Legal Foundation, a group known for its defense of property rights, filed an objection to the Google book deal earlier this month focusing on concerns related to those I raised in my posts earlier this week.

WLF points out that the Supreme Court has mandated that plaintiffs seeking to certify a class must make a diligent effort to notify all affected class members. According to the high court’s Shutts decision, this effort must include—at a minimum—sending a letter to every identifiable member of the class. In this case, this would mean sending a letter to every address in the US Copyright Office’s database of authors. WLF questions whether this was done; the foundation reports that it never received notification related to any of the books for which it holds the copyrights.

Now, it might be objected that this process would be prohibitively expensive. But if the class is so large that it’s impractical to notify all of its members, then the class is certainly too large to expect a judge to verify that the interests of all class members is being served by the settlement. If the class is too large to notify, then it’s too large to certify.

WLF also points out that the sprawling and heterogeneous class of plaintiffs makes it unlikely that the plaintiffs’ lawyers can fairly represent all parties who would be bound by the settlement:

For example, only those class members whose works have already been copied by Google are entitled to cash payments under the Settlement Agreement. Thus, the financial interests of those whose works have been copied diverge from the interests of class members whose works have not been copied. The former have an interest in maximizing cash payments, while the latter would prefer to see a smaller portion of the settlement pot allocated to those cash payments and a larger portion allocated to compensation for copying to be performed by Google in the future.

Another distinction among class members involves orphan works – that is, works whose owners are difficult (if not impossible) to ascertain. The owners of orphan works (who may not even know that they hold any ownership interests) have an obvious interest in ensuring that a large portion of settlement funds is dedicated to identifying the ownership of orphan works. On the other hand, the owners of works whose ownership is readily identifiable have an interest in holding down such search costs. If less money is devoted to such search efforts, a correspondingly greater percentage of settlement funds will be available to provide compensation
to them.

The point here isn’t that the amount allocated to orphan works searches is too high (or too low), or that more (or less) should be allocated to pay for previously-scanned books. The point is that a settlement involving millions of plaintiffs will inevitably enrich some plaintiffs at the expense of others. This isn’t a problem that can be solved by changing the details of the settlement. It’s a problem that can only be solved by limiting the scope of the settlement to parties whose interests are actually represented by the plaintiffs’ attorneys.

Obama: I Want Those Patriot Act Powers

Yesterday, President Obama’s lawyers informed members of Congress that the president does not want any provision of the  Patriot Act to expire.  Turns out that  Obama wants to have the sweeping powers.  This is just the latest example of the cacophony that pervades Washington.  When Bush was in the White House, the Dems postured against his runaway spending, his military quagmires, and his constitutional violations.  With Obama in the White House, Bush’s most misguided policies either continue or worsen.

Obama is in the news today for his “off-the-record” comment about Kanye West.  It would have been better had a reporter overheard Obama saying something like, “John Ashcroft was a terrific Attorney General, but  I’ll never admit that publicly.”

For related Cato work, go here and here.

Public Information and Public Choice

MalamudOne of the high points of last week’s Gov 2.0 Summit was transparency champion Carl Malamud’s speech on the history of public access to government information – ending with a clarion call for  government documents, data, and deliberation to be made more freely available online. The argument is a clear slam-dunk on simple grounds of fairness and democratic accountability. If we’re going to be bound by the decisions made by regulatory agencies and courts, surely at a bare minimum we’re all entitled to know what those decisions are and how they were arrived at. But as many of the participants at the conference stressed, it’s not enough for the data to be available – it’s important that it be free, and in a machine readable form. Here’s one example of why, involving the PACER system for court records:

The fees for bulk legal data are a significant barrier to free enterprise, but an insurmountable barrier for the public interest. Scholars, nonprofit groups, journalists, students, and just plain citizens wishing to analyze the functioning of our courts are shut out. Organizations such as the ACLU and EFF and scholars at law schools have long complained that research across all court filings in the federal judiciary is impossible, because an eight cent per page charge applied to tens of millions of pages makes it prohibitive to identify systematic discrimination, privacy violations, or other structural deficiencies in our courts.

If you’re thinking in terms of individual cases – even those involving hundreds or thousands of pages of documents – eight cents per page might not sound like a very serious barrier. If you’re trying to do a meta-analysis that looks for patterns and trends across the body of cases as a whole, not only is the formal fee going to be prohibitive in the aggregate, but even free access won’t be much help unless the documents are in a format that can be easily read and processed by computers, given the much higher cost of human CPU cycles. That goes double if you want to be able to look for relationships across multiple different types of documents and data sets.

All familiar enough to transparency boosters. Is there a reason proponents of limited government ought to be especially concerned with this, beyond a general fondness for openness? Here’s one reason.  Public choice theorists often point to the problem of diffuse costs and concentrated benefits as a source of bad policy. In brief, a program that inefficiently transfers a million dollars from millions of taxpayers to a few beneficiaries will create a million dollar incentive for the beneficiaries to lobby on its behalf, while no individual taxpayer has much motivation to expend effort on recovering his tiny share of the benefit of axing the program. And political actors have similarly strong incentives to create identifiable constituencies who benefit from such programs and kick back those benefits in the form of either donations or public support. What Malamud and others point out is that one thing those concentrated beneficiaries end up doing is expending resources remaining fairly well informed about what government is doing – what regulations and expenditures are being contemplated – in order to be able to act for or against them in a timely fashion.

Now, as the costs of organizing dispersed people get lower thanks to new technologies, we’re seeing increasing opportunities to form ad hoc coalitions supporting and opposing policy changes with more dispersed costs and benefits – which is good, and works to erode the asymmetry that generates a lot of bad policy. But incumbent constituencies have the advantage of already being organized and able to invest resources in identifying policy changes that implicate their interests. If ten complex regulations are under consideration, and one creates a large benefit to an incumbent constituent while imposing smaller costs on a much larger group of people, it’s a great advantage if the incumbent is aware of the range of options in advance, and can push for their favored option, while the dispersed losers only become cognizant of it when the papers report on the passage of a specific rule and slowly begin teasing out its implications.

Put somewhat more briefly: Technology that lowers organizing costs can radically upset a truly pernicious public choice dynamic, but only if the information necessary to catalyze the formation of a blocking coalition is out there in a form that allows it to be sifted and analyzed by crowdsourced methods first. Transparency matters less when organizing costs are high, because the fight is ultimately going to be decided by a punch up between large, concentrated interest groups for whom the cost of hiring experts to learn about and analyze the implications of potential policy changes is relatively trivial. As transaction costs fall, and there’s potential for spontaneous, self-identifying coalitions to form, those information costs loom much larger. The timely availability – and aggregability – of information about the process of policy formation and its likely consequences then suddenly becomes a key determinant of the power of incumbent constituencies to control policy and extract rents.