Last week Robert S. Mueller III, the special counsel, indicted 13 Russians for intervening in the 2016 United States election. Two of the charges — buying political advertisements and mandatory disclosure — bear on free speech.

Much of the indictment documents activities during the election that would be both normal and protected by the Constitution if undertaken by American citizens. The defendants bought political advertisements, staged political rallies, and even “posted derogatory information about a number of candidates,” Hillary Clinton in particular. Lacking all scruples, they are said to have “solicited and compensated real U.S. persons to promote or disparage candidates” which means paying an actress to impersonate Hillary Clinton in jail. The defendants tried to create “political intensity through supporting radical groups, users dissatisfied with [the] social and economic situation and oppositional social movements.” Overall the Russians hoped “to sow discord in the U.S. political system.”

As it happens, all this activity may be illegal because the Russian government supported these activities. The Federal Election Commission concisely explains regulation 110.20: , “The Federal Election Campaign Act (FECA) prohibits any foreign national from contributing, donating or spending funds in connection with any federal, state, or local election in the United States, either directly or indirectly.” The Commission notes that this ban “was first enacted in 1966 as part of the amendments to the Foreign Agents Registration Act (FARA), an “internal security” statute. The goal of the FARA was to minimize foreign intervention in U.S. elections by establishing a series of limitations on foreign nationals.” FARA also required agents of foreign principals to register with the federal government presumably, as the indictment says, so “the people of the United States are informed of the source of information and the identity of persons attempting to influence U.S. public opinion, policy, and law.” (It should also be noted that the defendants are charged with several counts of fraud and identity theft).

These two parts of the law establish different rules for different audiences. Voters in an election are prohibited from hearing speech funded by a foreign power. Arguably, they are prevented from hearing any speech by an employee of a foreign government; such speech would involve indirect spending on an election. Other listeners, unnamed in the law, need not be prevented from hearing speech of foreigners “attempting to influence U.S. public opinion, policy, and law.” The public, apart from electoral appeals, and public officials, including above all members of Congress, may hear foreign speech assuming disclosure of its source. Voters, however, should be, and are protected from such speech.

The law seems informed by the following assumptions. Public officials need to hear foreign views, especially about international affairs. They have the ability to sort out the false from the true if they know who is behind the arguments. In contrast, voters might be moved by foreign speech to the detriment of the nation. Hence, voters must be prevented from hearing such speech even if its source is disclosed. To put it mildly, this distinction between officialdom and voters is contrary to the foundations of freedom of speech. If voters lack this basic capacity of citizenship, why protect speech through the First Amendment? The distinction seems paternalistic.

There’s little evidence that the Russian efforts had much effect on the voters in 2016. The New York Times indicates that the Russians may have persuaded a few people to show up at a small anti-Muslim rally in Texas. Speculation about others effects does abound, as the Times article shows. However, as Brendan Nyham indicates, political science research shows how hard it is to change votes even with significant spending. The Russian effort was a miniscule portion of overall spending in 2016.

The law is the law, and the indictment made a good enough case against the defendants that at least 12 grand jury members endorsed the charges. But the indictment comes at a difficult time for free speech. Facebook and other tech giants have been put on the defensive in DC. For some, only foreign malevolence could propel a man like Donald Trump to the White House. Fears about national security can foster public actions that otherwise would be rejected. This might be one of those moments when something must be done and the “something” that is selected does real harm to freedom of speech. That outcome could easily be worse than whatever threat the Russians pose to American democracy. 

Ballotpedia asked five redistricting buffs to comment on how the Supreme Court might rule on the two gerrymandering cases it is considering, Gill v. Whitford from Wisconsin and Benisek v. Lamone from Maryland. My response: 

One clue is timing. Rather than fast-track consideration of Benisek v. Lamone, the Court instead set an oral argument date of March 28. That leisurely pace ensures that any decision will come late enough in the term to wreak havoc on this year’s [election] cycle if it announces the imposition at once of a new constitutional standard. Few if any Justices wish to wreak such havoc avoidably. From which I deduce that the Court either 1.) does not expect to announce a new standard, or 2.) expects to do so only in a staged or prospective way that allows states time for compliance or kicks in with the next Census cycle.

While I believe Justice Kennedy remains open to the development of some constitutional standard in this area, I also think he is looking for a fix that is objective and mechanical enough that 1.) it yields the same results from state to state and from Republican-appointed as from Democratic-appointed lower court judges, and 2.) the clarity of what it calls for and how to comply cuts off a need for continual massive litigation and judicial supervision (compare here the success of the one-person, one-vote revolution). Kennedy’s having joined with the conservatives to stay the Gill ruling leaves me thinking that as of then he wasn’t convinced that the “efficiency gap” standard fit that demanding bill.

I’m also pleased to note that my essay on gerrymandering from a libertarian/​classical liberal perspective, which appeared originally in Cato Unbound, is featured in the latest Cato Policy Report. You can read it here. 

George Hodgin’s mission seemed simple: manufacture uncontaminated, chemically consistent cannabis for use in scientific research on marijuana’s medical effects, all while complying with federal regulations surrounding the production of a drug still classified by the Drug Enforcement Administration (DEA) as highly dangerous. Despite new rules the DEA promulgated eighteen months ago, with the stated goal of allowing expanded cultivation of marijuana for scientific research, George Hodgin is still in administrative limbo. 

Hodgin, a former Navy SEAL, approached us recently for advice after encountering numerous regulatory roadblocks. We have no special knowledge or ability in that direction; but perhaps publicizing his endeavors will nudge public opinion (and regulators) in the right direction. 

Expanding research access to high-quality marijuana is important. The Marijuana Policy Project estimates that roughly 2.5 million patients use medical marijuana – just in states with legal medical marijuana programs. This number is likely an underestimate, as it does not account for individuals obtaining marijuana for medical use through non-medical channels. Marijuana’s illegality at the federal level prevents the collection of much needed data that could help drive future research.

Veterans represent a particularly important category of medical marijuana users. Although Veterans Health Administration physicians are prohibited by federal law from recommending medical marijuana, new guidelines issued in December 2017 revise existing standards to encourage doctors and patients to discuss the use of medical marijuana without fear of recrimination. 

A January 2017 report from the National Academies of Sciences, Engineering, and Medicine acknowledges the difficulty researchers face in acquiring appropriate cannabis products and recommends actions be taken to ameliorate the situation; yet, obstacles remain. 

The major difficulty facing those who want to produce or obtain cannabis products for research purposes is that the DEA currently classifies marijuana as a Schedule I drug, which means the DEA views it as having “no currently accepted medical use.” But thirty states and the District of Columbia have laws permitting the medical use of marijuana. Additionally, the research available suggests potentially widespread medical applications for marijuana. 

Legal restrictions on access have severely hampered medical research into marijuana’s possible effects. One solution would be to reschedule marijuana, yet the DEA has repeatedly denied petitions to do so. In a July 2016 denial, the DEA asserted that marijuana has no currently accepted medical use in the United States because, among other reasons, “the scientific evidence is not widely available.”

This will be true, by definition, so long as the DEA makes it virtually impossible to conduct scientific research. Catch 22.

In August 2016, the DEA changed its rules to allow registered entities to supply researchers with the quality and quantity of marijuana needed for scientific study. Despite the rule change, however, the DEA has yet to approve a single application. In testimony given to the Senate Judiciary Committee last October, Attorney General Jeff Sessions confirmed that 26 applications are outstanding from marijuana suppliers; but the DEA has not approved a single one. 

The DEA’s resistance is part of the long-running, tough on drugs approach it has taken since the 1970s. As recently as 2013, courts upheld the DEA’s monopoly on the production of (federally) legal marijuana. The current administration shows no intention of changing this stance. 

In a recent poll, 91 percent of Americans supported the use of medical marijuana. For years, the DEA’s opposition to legalization or rescheduling has hinged on the argument that no scientific evidence suggests medical benefits. Despite hopes that the August 2016 rule change would usher in a new era of research, the past year and a half has proven that inaction remains the status quo. Until something changes, countless sufferers of pain, PTSD, and other ailments will remain trapped in a legal gray zone. Allowing properly vetted companies to manufacture and distribute research grade marijuana legally is the least the DEA can do.

Research assistant Erin Partin contributed to this blogpost.

From 1983 to 1999, the CBO issued two-year forecasts that added up to a 2.7% growth rate, which would now be widely dismissed as a “rosy” forecast. Yet actual growth averaged 3.7% from 1983 to 1999 – a full percentage point higher – despite a recession in 1991. Today, the CBO forecasts that even 2.7% economic growth is impossible, and claims only 1.9% is within reach. 

The Administration thinks the economy can grow a percentage point faster. The 2019 Budget estimates the economy will grow by 2.9% a year for ten years. The Committee for a Responsible Budget (CFRB) argues that this “strains credulity, especially if interest rates and inflation also remain under control, as the budget predicts they will.” [This appears to suggest higher inflation would be good for growth.]

“Given population aging and other economic fundamentals,” says the CFRB, “the United States is likely to ultimately achieve growth of 2 percent per year or perhaps less – not 3 percent. The Federal Reserve projects long-term sustained growth of 1.8 percent per year [1.7–2.2%], and the Blue Chip average for sustained growth is only slightly higher at 2.1 percent. Prior to the tax deal, CBO projected a long-run growth rate of 1.9 percent.”

Is the OMB unrealistic to estimate the economy can grow by 2.9% a year or is the CBO unrealistic to assume it can’t grow faster than 1.9%?

The real contest here is between current OMB projections after the tax deal, and CBO projections “prior to the tax deal.” Fed and Blue Chip forecasts are unofficial and unpersuasive.

The Federal Reserve does not make official long-term projections. The quarterly FOMC Summary of Economic Projections (SEP) defines “longer-run projections” as rates of growth to which “a policymaker expects the economy to converge over time – maybe in five or six years.” Last December 13, the 19 survey participants thought that after five or six years real GDP would settle down to 1.7% to 2.2%, but that is not a ten-year average since growth in the previous five or six years might be rapid.

Blue Chip Indicators collect monthly forecasts from over 50 business economists (I used to be one of them). It constructs a “consensus” by averaging a possibly wide array of different estimates. All such frequently-revised forecasts are most reliable as lagging indicators – becoming pessimistic after economic news turns bad and optimistic after things pick up. The models and techniques used to make monthly or quarterly forecasts have no predictive power at all beyond two quarters, if that.

What about the CBO? Their projection of 1.9% growth is a full percentage point below that of the Trump administration. Could the CBO possibly be that far off? Sure. They’ve done it before.

From 1983 to 2000, the CBO’s two-year forecasts of real GDP growth were exactly one percentage point too low, on average. A two-year forecast for 1983–84 was made in 1983 and combined estimated growth rates for both years, so the fact that 2‑year growth rates kept being underestimated repeatedly in all but two years (1990 and 1991) from 1983 to 2000 was a triumph of theory over experience.

The graph, from “The CBO’s Economic Forecasting Record,” shows the CBO systematically underestimated growth of real GDP after the Reagan tax rate reductions were phased in during 1983–84 and 1988–90 (TRA86), and again after the capital gains tax was slashed from 28% to 20% in 1997. Conversely, the CBO overestimated GDP growth after Bush 41 raised tax rates in 1990, after Obama raised tax rates in 2013, and during the high-tax bracket creep years of 1976–82. The CBO appears to suffer from a pro-tax estimating bias – assuming higher tax rates do no harm, and lower tax rates do no good.

The CBO argues that its “five-year forecasts of output and inflation are more accurate than its two-year forecasts of those variables, in part because long-term forecasts rest more on underlying trends in the economy than on short-term cyclical movements, which are very difficult to predict.” Unfortunately, CBO five-year forecasts also underestimated real GDP growth in all but one 5‑year period between 1981–1985 and 1999–2003. The exception was 1987–91, when the CBO estimate proved slightly optimistic (0.28%) thanks to recession on the heels of the ill-fated Bush 41 “tax increase.” A single mild recession (aggravated by higher taxes) can’t explain why the CBO consistently underestimated the 4.4% rate of real GDP growth for seven years from 1983 to 1989, or the same 4.4% pace for four years from 1996 to 2000.

A recent paper, “How CBO Produces Its 10-Year Economic Forecast” explains that “CBO projects potential TFP on the basis of historical trends in TFP growth. However, projecting trends in TFP is particularly challenging because it is, by definition, a measure of unexplained growth in output.” On the contrary, extrapolating “historical trends” is just lazy, not “challenging.”

The so-called “economic fundamentals” the CFRB mentioned essentially consist of projecting recent trends into the future. That means assuming output per hour (productivity) keeps growing at the anemic 1.3% pace of 2006 to 2015, and that hours worked grow at half that rate because labor force participation is assumed (“projected”) to remain extremely depressed and most part-timers are assumed to reject longer hours.

If you add 1.3% projected growth of productivity to 0.6% projected growth of the labor force, you end up with a 1.9% limit on potential economic growth (slower than 2010–2017 when the economy grew at a 2.2%). But productivity and labor force participation depend on incentives, not past trends. And incentives to invest and work just changed dramatically.

We know that the CBO is perfectly capable of underestimating economic growth by a full percentage point over a 10-year period since it already managed to do that over a 17-year period. If history is any guide, the CBO’s 1.9% long-term forecast is once again much too low, as it has been whenever the highest tax rates on income and/​or capital gains were reduced.

The White House released President Trump’s infrastructure plan today, which calls for spending $200 billion federal dollars as seed money to stimulate a total of $1.5 trillion on “gleaming new infrastructure.” Almost lost in the dozens of pages of documents issued by the administration is that the reason why the federal government supposedly needs a new infrastructure program is that our existing infrastructure is crumbling, and the reason it is crumbling is that politicians would rather spend money on gleaming new projects than on maintaining the old ones.

The White House proposes several new funding programs. The administration could have dedicated one or more of these programs to maintenance and repair of worn-out infrastructure. Instead, all $200 billion can be spent on new projects, and knowing politicians, most of it will be. To make matters worse, funds for most of the programs would be distributed in the form of competitive grants, but experience has proven that competitive grants are highly politicized. 

“In the past, the Federal Government politically allocated funds for projects, leading to waste, mismanagement, and misplaced priorities,” agrees White House economic advisor Gary Cohn. The administration’s solution, Cohn continues, is to “stimulate State, local, and private investment.” In other words, instead of most decisions being made by Washington politicians, they will be made by local politicians. But if local politicians were any better at maintaining infrastructure, then we wouldn’t have tens of thousands of local bridges classed as “structurally deficient” and the New York, Washington, Boston, and other subway systems wouldn’t be falling apart.

The White House says that the federal funds it proposes to allocate to infrastructure may be spent on either new construction or maintenance, which is an advantage over some existing federal programs that can only be spent on new construction. But just because they can be spent on maintenance, doesn’t mean they will be.

The New York subway system is falling apart because the city doesn’t have enough money to maintain it. Yet it has enough money to spend $10 billion on a tunnel between Penn Station and Grand Central Terminal for Long Island Railroad trains, which the New York Times has called “the most expensive subway in the world.” It also has enough money to build the eight-mile Second Avenue subway, which at $2.1 billion a mile must be the second-most expensive subway in the world.

The Washington Metro system is falling apart because the region is short $10 billion to maintain it. Yet Virginia was able to find enough money to build the Silver Line and Maryland to build the Purple Line; the costs of these two projects would have been enough to fix most of the existing Metro system.

The Boston rail transit network is also falling apart because the region can’t find the $470 million a year needed to maintain it. Yet is was able to find $2.3 billion to build a 4.3‑mile light-rail extension. In all of these cases, local politicians decided to spend money on new projects even though 50 to 80 percent of the money in each case could have been spent on maintenance.

The truth is that our infrastructure isn’t in as bad shape as some claim. State highways are in good condition; local roads less so. Rail freight systems are in good condition; rail transit systems less so. In general, infrastructure paid for out of user fees are in good condition; infrastructure paid for out of tax dollars less so. 

This means there is a simple way to take the politics out of infrastructure and make sure that managers maintain it: fund it out of user fees, not tax dollars. Managers of user-fee-funded infrastructure tend to do the best job of maintenance because they know users will stop paying to use their facility if it becomes unreliable.

To its credit, the White House program does make some token movements in the direction of more user fees. For example, it would allow states to charge tolls for all interstate highway, something that is now, for the most part, prescribed by Congressional edict. But even this depends on state politicians asking their constituents to pay tolls for something they have been getting for “free,” which is unlikely to happen.

In the end, however, the major impact of the Trump plan is to drop $200 billion new dollars onto state and local governments on top of existing federal spending programs. This will merely provide more insulation for state and local politicians from having to ask users to pay for the infrastructure they build. The result will be that most, if not all, of that $200 billion will go to build new infrastructure that we probably don’t need and can’t afford to maintain rather than maintaining what we have.

As I was preparing for a Demand Progress-sponsored panel on Congressional oversight of intelligence matters on the afternoon of February 9, Demand Progress Policy Director Daniel Schuman and I agreed that if President Trump was going to refuse to “declassify” the House Intelligence Committee Democrats rebuttal to the “Nunes Memo,” he would wait until the late Friday news cycle to do it. We didn’t have to wait long for that prediction to come true. 

In a moment, I’ll get to the issue of whether Trump actually has the authority under the Constitution to do what he did, but I want to start with is this paragraph from the New York Times story referenced above:

But Donald F. McGahn II, the president’s lawyer, said in a letter to the committee on Friday night that the Democratic memo could not be released because it “contains numerous properly classified and especially sensitive passages.” He said the president would again consider making the memo public if the committee, which had approved its release on Monday, revised it to “mitigate the risks.”

In that same NYT story, House Intelligence Committee ranking member Adam Schiff provided further context:

In a statement on Friday night, Mr. Schiff said that Democrats had provided their memo to the F.B.I. and the Justice Department for vetting before it was approved for release by the committee. The Democratic memo was drawn from the same underlying documents as the Republican one.

“We will be reviewing the recommended redactions from D.O.J. and F.B.I., which these agencies shared with the White House,” Mr. Schiff said, “and look forward to conferring with the agencies to determine how we can properly inform the American people about the misleading attack on law enforcement by the G.O.P. and address any concerns over sources and methods.”

So if Schiff is to be believed, House Intelligence Committee Democrats ran their memo by Justice Department and FBI officials prior to the unanimous committee vote to release his memo, then sent the memo over to the White House for reaction. Trump and his team then demanded still more redactions. If the above account is correct, the same Justice Department or FBI officials who reviewed the original “Schiff Memo” apparently demanded still more redactions once it got to Trump’s desk.

It’s this sequence of events which brings me to the question of whether Trump has the authority under the Constitution to censor or rewrite Congressional work product, with or without Congressional assent, if it contains references to Executive branch information asserted as being classified, in part or in whole. The short answer is no. The longer answer is still no, but with some caveats.

Congress and Secrecy

The word “secrecy” appears only once in the Constitution, specifically in Article I, Section 5, which contains the following clause:

Each House shall keep a Journal of its Proceedings, and from time to time publish the same, excepting such Parts as may in their Judgment require Secrecy; and the Yeas and Nays of the Members of either House on any question shall, at the Desire of one fifth of those Present, be entered on the Journal.

From the beginning of the Continental Congress in 1774 through the adoption of the Constitution in 1789, the Congress was responsible for keeping such matters of state secret as they deemed necessary. It was not until 1818 that Congress passed a resolution directing President Monroe to publish previously secret material from the nation’s earliest history.

The ability of the Executive branch to keep military and foreign policy secrets was not something Congress would get involved in legislating until well into the 20th century. Examples include the Atomic Energy Act, the Freedom of Information Act, and the Classified Information Procedures Act. 

It’s worth noting that in none of these statutes did Congress renounce its authority to make materials covered by these laws public if it chose to do so—including material deemed classified by the Executive branch. The first real confrontation over this principle occurred in the aftermath of Congressional investigations of Executive branch domestic spying scandals that first surfaced in 1971.

The Pike and Church Committees

In 1975, the House and Senate each created Select Committees to investigate domestic spying and political repression operations carried out by the NSA, FBI, CIA, and military intelligence elements. The two committees became known by the names of their respective chairmen: Frank Church (D‑ID) in the Senate and Otis Pike (D‑NY) in the House. Both committees encountered deliberate efforts by Ford administration officials to block access to relevant agency or department records, resulting in months of often heated confrontations with CIA, NSA, FBI, and White House officials over committee demands for documents. 

As I noted in a recent piece in The Hill, the slightly differing approaches of the two committees led to very different outcomes:

In a now-infamous incident known as the “September compromise,” Pike agreed to allow the Ford administration to make the call about what executive branch documents could or could not be made public. When Pike moved to finalize his committee’s report and make it public in January 1976, Ford persuaded the House to block publication on the grounds that the entire Pike Committee report was a classified document. In contrast, the Senate refused to submit the Church Committee report for Ford’s review and published its findings in April 1976. Constitutionally, Church and his colleagues made the right call, Pike’s House colleagues the wrong one.

Indeed, in its preamble to its final report, the Church Committee made its position abundantly clear (from the committee’s final report, Vol. 1, p. 13):

It should be noted that the Ford administration made no attempt to challenge the Church Committee’s publication of the report in federal court—a de facto admission that the Congress did indeed have the authority to release—and thus simultaneously declassify—information previously deemed secret by Ford and his predecessors. 

The Caveats

The Pike Committee experience proved that the Executive branch could, under the right circumstances, still thwart an official Congressional release of information President Ford and his agency heads considered classified. However, he needed help to do it—in this case, Pike’s House colleagues, who failed to authorize the release of the Pike Committee report.

The subsequent legislation passed by the House and Senate creating standing Select Intelligence Committee’s largely adopted the system that had been used by the House in the Pike Committee episode. Molly Reynolds of Brookings recently wrote a good overview of this process at Lawfare. 

What the “War of the Memos” reveals is that the original Church Committee approach to this kind of Executive-Legislative confrontation was the right one. The process now being used—and more than likely abused—by the House Intelligence Committee GOP majority and the Trump White House makes it far less likely that the public will learn the truth about whether or not the FBI and Justice Department misled the FISA Court in the Carter Page episode. The larger and more long-term ramifications of this episode are that a new and dangerous precedent is being set by not just the House Intelligence Committee, but the Congress as a whole.

Letting any President dictate what Congress can or cannot publish is a clear assault on the Separation of Powers. It’s also possible for Congress to voluntarily breach the Separation of Powers and compromise its oversight of intelligence matters. That was the error committed by Senator Diane Feinstein (D‑CA) when she voluntarily submitted the Senate Intelligence Committee’s full report on the Bush 43-era CIA torture program to President Obama for a “declassification” review. Instead, Obama—a fellow Democrat—sat on the full committee report. To date, Feinstein has failed to get her Senate colleagues to hold a vote to make the full report public.

Trump’s quashing (at least for now) of the “Schiff Memo” only underscores why the House and Senate must modify their respective chamber rules to make it clear, as the Church Committee did, that while the House and Senate Intelligence Committees will consider Executive branch concerns and arguments against making something public (and thus declassifying it), the final call will remain with the committees. 

The Trump administration has rolled out its 2019 federal budget, which includes a plan to boost investment in the nation’s infrastructure.

I see five components to President Trump’s infrastructure policies so far.

First, Trump approved business tax cuts in December, which will boost private capital investment in pipelines, broadband, factories, and much else. The private sector owns more than three-quarters of U.S. infrastructure, so the tax cut will create wide-ranging benefits.

Second, the administration is taking steps to reduce costly regulations and speed permitting on infrastructure projects. The new budget proposes to “shorten the process for approving projects to 2 years or less.” The number of federal rules creating roadblocks to construction has increased over time. Kudos to Trump for removing some of the barriers.

Third, the budget proposes spending an additional $200 billion over 10 years of federal dollars on infrastructure, as summarized here. There is a new “Incentives” program, a “Transformative Projects” program, and a “Rural Infrastructure” program.

These new programs are unaffordable, especially given that last week’s budget bill exploded annual deficits to more than $1 trillion. Creating a new $50 billion program for rural areas is particularly ridiculous given that the government already has a range of wasteful rural subsidies.

In general, new federal subsidies for infrastructure are not needed. Any state wanting to improve its infrastructure can do so with its own funding. States can raise funds through taxes, debt, user charges, public-private partnerships (PPPs), and privatization. Federal infrastructure subsidies are often counterproductive, as I discuss here.

Giving Trump’s new subsides fancy titles such as “Transformative Projects” program” will not make them more efficient than current bureaucratic federal efforts. All federal spending for infrastructure comes from taxpayers who live in the 50 states, and there is no “transformative” magic that happens when cash is sent through Washington. 

Fourth, the Trump plan would boost incentives for the states to pursue PPP deals, including expanding various loan programs and broadening eligibility for private activity bonds. Here the end goal is a good one, but a better way to encourage PPPs and privatization would be to end federal subsidies altogether and repeal the tax exemption for municipal bond interest. That exemption dissuades the states from privatizing facilities such as airports.

Fifth, the Trump plan would make reforms to federal control over lands and structures. It would create a “capital revolving fund” to allow agencies to buy real property that they currently lease. That may save money, but the administration should instead focus on reducing the government’s size and its need for office space in the first place. (The government currently owns or leases 275,000 buildings). The Trump plan would also allow agencies to generate more revenues from public lands, and allow “for the sale or lease of federally owned assets.” That sounds positive.

In sum, Trump’s infrastructure plan includes both big government and small government policies. The plan states, “President Trump’s proposal will return decision-making authority to State and local governments, which know the needs of their communities.” But the best way to do that would be to end federal subsidies and all the related string-pulling from Washington.