No Pelosi-Trump Infrastructure Deal, Please
After the midterm election 2.5 weeks ago, there began calls for an infrastructure deal. The details, as always, were always vague, but the idea is that congressional Democrats and President Trump will agree on a bill to spend about a trillion dollars on infrastructure. What infrastructure is at stake is not specified, except that some New York-based commentators (and Senator Schumer) are calling for federal funding of the Gateway project; whether to pay for the program with deficit spending, tax hikes, or cutting other spending is not specified either. The good news is that such a deal isn’t likely to happen, for roughly the same reasons such a deal would be a bad idea in the first place. However, just in case some people reading this blog might like the idea of such a grand bargain, I’d like to spell the reasons why such a deal would be a waste of money.
What is the purpose of an infrastructure deal, anyway?
Given around a year of something approaching full-time work, I could identify a trillion dollars’ worth of useful public transportation investment in the United States. Given that I’d also look for ways to cut construction costs (which I’m almost certain Congress has not seriously tried), and given that there are other infrastructure priorities than transit, it should not be hard to come up with a long-term 13-figure program.
However, I’m fairly certain there hasn’t been any serious attempt to list infrastructure projects that should be covered under this plan. The main clue is that if there were any, the people trying to sell the public on such a deal would mention them as concrete benefits. This has happened with Gateway: people around the New York area are desperate for federal funding to cover the project’s extreme cost, and do not shy from mentioning it as a beneficiary of a grand bargain. But with anything else, there’s nothing.
For example, nobody in California has said anything about federal funds for the state’s flagging high-speed rail project, even though it would be a natural candidate for a bipartisan deal between Trump and congressional Democrats (the state’s Republican delegation opposed the project, but much of it was wiped out in the midterm). Elsewhere, there are both road and transit projects in red state cities that are hungry for funding, some of which were on the Trump administration’s list of projects to fund last year, in one of the interminable Infrastructure Week pushes that went nowhere. Nothing comparable has surfaced this month.
The lack of detail about the plan suggests it’s not really serious policy. It’s a trial balloon – one that’s failing because of the political situation. But in the event anything comes out of it, it will be a half-thought plan, created for the purpose of spending money and doing something that gives the appearance of bipartisan consensus.
The US economy is not in a recession
The point of a Keynesian stimulus is to prop up the economy during recessions. The American economy right now has 3.7% unemployment, which is more or less full employment, and 2.5% inflation, which is a hair above target. Additional spending would be great for me – it would strengthen the dollar, personally helping me as someone who earns dollars and spends euros. But for the putative target of the bill – the American people – the only effect would be fiscal constraints. The country needs to think about reducing the deficit, not about increasing it in a show of bipartisan unity.
Worse, the stimulus effect of new government spending comes from the net change in annual spending, whereas the deficit effect comes from overall annual spending. A big infrastructure bill would only act as economic stimulus in the earliest phases, when the spending rate would ramp up. Subsequently, it would have no effect on growth or on employment. David Dayen made this point regarding the 2009 stimulus: it had a big effect on American economic growth in 2009, but as the spending rate reached its maximum in 2010, the net effect of federal spending on growth turned negative in the third quarter of 2010, even before the Republican victory in the midterm, long before most stimulus funds were actually spent.
This does not mean that infrastructure funding is out of the question. A serious bill that is crafted to be deficit-neutral in the short as well as long term could do good; it is also close to impossible. Some Democratic pundits have trolled the conversation by proposing pairing it with repealing Trump’s tax cuts, but the probability of a grand bargain that raises taxes to pay for extra spending is approximately zero. Cutting other spending is extremely unlikely as well – unlike state and local governments, domestic federal spending doesn’t have enough waste to fund a trillion-dollar infrastructure bill, and what waste does exist is locked up in Medicare, which is politically untouchable.
The state of the American economy is such that it’s a great idea to design an infrastructure bill, to be deployed at the next recession. There could be a list of priority projects for public transportation (or other forms of infrastructure) chosen for a combination of cost-effectiveness and nationwide spread. While designing this plan, the federal government would make the process open, to let local and state governments know what is happening and offer them the opportunity to submit their choice projects for consideration. The federal government should also insist that they not defer maintenance now hoping to score state of good repair money later – for example, I would propose to credibly commit to only funding expansion but not maintenance, and to defund projects run by agencies that defer maintenance (such as Boardman-era Amtrak). The plan would be funded, with deficit spending, at the next recession, which analysts expect to start in the next few years.
The federal government is unusually corrupt
If the above plan of coming up with a measured infrastructure plan, with incentives to encourage good behavior among state and local governments, sounds like science fiction, it’s because the federal government today doesn’t have the capability of carrying out such a program. Part of it is generic public-sector weakness within the United States, making it hard to make long-term plans; the civil service is weak, and politicians make capricious decisions, so nothing like the TGV, Grand Paris Express, High Speed 2, and Crossrail – all bipartisan projects within their respective countries – can happen.
But there’s a bigger problem now: Trump. Trump himself is corrupt in ways that go far beyond the affairs of scandal-ridden past presidents like Clinton and George W. Bush, and this affects how people think of infrastructure. The US has a public transportation cost premium of nearly a full order of magnitude over comparable countries. Such a premium must have multiple causes, but one cause is corruption: we’ve already seen how political interference by Schumer helped double the cost of Amtrak’s rolling stock procurement. Trump’s scandals easily surpass Schumer’s.
This goes beyond partisanship. Atrios has been a partisan Democrat since his blog’s early days, and yet he’s called for SUPERTRAINS (always in caps) since mid-2008, when the idea of stimulus became part of the American public conversation. At the time Obama was ahead in the polls, but he was not guaranteed to win, and years of Bush had gotten the Democratic base used to opposing anything a Republican president did; and yet, center-left writers like Atrios and Matt Yglesias (at the time transitioning from the Republican bloggers’ favorite Democrat to a conventional partisan liberal Democrat) were fine endorsing an infrastructure program in an uncertain partisan climate.
In theory, the extent of Trump’s corruption is small compared with the magnitude of the program. It’s billions of dollars at worst versus a trillion. In practice, the presence of the current president at the helm of any program screams at contractors, “make an effort to stay at Trump hotels and Mar-a-Lago, not to make a cheap and technically sound bid.” The extra cost coming from contractors slouching in the bidding and construction phases can easily soak up hundreds of billions of dollars out of the trillion: in Brian Rosenthal’s article about high New York costs, contractors quoted a premium of about 25% just from MTA red tape, and Trump’s personal corruption is probably on the same order of magnitude.
Ultimately, it’s fine to wait
In late 2008 and early 2009, the American economy was spiraling into the deepest recession since 1946; in that climate, rushing the stimulus was desirable. The situation today is not like that at all. There’s time to develop an infrastructure plan based on one’s combination of political preference and belief about the future (e.g. will Trump be reelected?, and who will control Congress after 2020?). There’s no point in passing a plan that exists purely to spend money and to show that Congress can enact big policies.
Since there’s no rush, and no need to deficit-spend right now, there’s grounds for demanding better of the government. Any infrastructure plan should be based on clear needs: that is, a national blueprint (such as reducing greenhouse gas emissions, or spreading infrastructure funding to poor states, or a similar political goal), a list of items designed to maximize cost-effectiveness within the blueprint’s parameters, and a federal civil service that can implement the construction of these items with maximum efficiency.
The incompetent and the corrupt should have no role to play in this program, and this begins with the current president. If it’s not possible to remove deadwood from the federal government, it’s fine to indefinitely postpone any big federal infrastructure plan. Nothing there would be indispensable; if Congress wants to deficit-spend money to create jobs, it can choose policies that are less sensitive to public-sector competence, such as tax cuts, unemployment benefits (not a big factor today but by definition a big one in a recession), and aid to states. With infrastructure that most of the developed world laughs at the US still manages to be one of the richest countries in the world; filling in the gap in public transportation is desirable, but the country won’t collapse if the gap persists.
Pedestrian Observations 
OK, but I don’t quite understand:
The US doesn’t really need any new mega-road projects but what it does need is to maintain those roads (and bridges etc) that are increasingly in poor state and which cost the freight industry and others countless billions in lost time & productivity, and doubtless cause more congestion than otherwise. Of course I have always said that Trump’s plan, inasmuch as it really provided anything like the err trumpeted trillion, would not fund what was most needed (maintenance) because, first, the big industry players couldn’t be bothered with maintenance contracts and second, politicians, especially Trump, want big flashy iconic ribbon-cutting infrastructure. I still reckon someone is missing a trick by not exploiting Trump’s public (if perhaps low-key and even dissipated) enthusiasm for HSR. However I see that CAHSR is in trouble again with a report showing $600m over-costs due to poor supervision of contractors etc, and that this puts at risk the Fed contribution of $3.5bn which is conditional on meeting a 2022 deadline. Having California and HSR linked like this is another factor in why Trump is probably not going to put another dollar into HSR. I mean Pelosi would support it!
If the US doesn’t need new roads (and it doesn’t) then it shouldn’t fund roads at all. It does need new transit infrastructure, and it should make sure to fund it, and not state of good repair backlogs. SOGR was a good target in the 1980s and 90s in New York, but subsequently it’s turned into a scam in which agencies defer maintenance under pressure to not raise taxes and then when money becomes available they cry poverty and point out to the backlog. Amtrak fired David Gunn and replaced him with the more pliable Joe Boardman when Gunn insisted on fully funding maintenance rather than undermaintaining infrastructure in order to look good for future privatization (P.S. DB’s doing the same, which occasionally explodes in its face as with Eschede or the Berlin S-Bahn meltdown). Then in 2009 when stimulus money became available it suddenly cried poverty and talked about the SOGR backlog and didn’t release any HSR plan until it was too late to get any funding for it.
And Trump’s enthusiasm for HSR is entirely performative. Generally it’s a good idea to ignore 100% of what Trump says and focus only on what he does. He says MAGA; what he actually does is steal money in ways that make Fillon look like a saint.
You are describing a utterly corrupt system in which all actors at all levels are not genuinely interested in fixing things, and are insincere. A truly grim picture but it explains why things are barely hanging together there.
Of course I understand about Trump, but his HSR comment–which was a throwaway line to reinforce the MAGA supposition that they weren’t great anymore–was unscripted and genuine. I don’t suppose he has thought of it since.
Meanwhile Pelosi (and Schumer) is going to go unchallenged just at the point they seriously need some new blood and new direction (ie. in the critical two years leading to 2020). But the whole top swathe of Dems are heading up to 80 years old … and don’t want to move aside.
Schumer is just bad at his job as both senator and minority leader and should retire.
The MAGA sentiment that the US is deficient because of infrastructure is genuine (though Trump himself isn’t), but there’s a serious problem with fixing it: the same people who express it are far too proud to learn from France or Japan or Germany. They prefer fixing what’s wrong with America with what’s right with America to fixing it with what works elsewhere, and this leads them to chase mirages like Hyperloop. Of course Musk was not born in the US, but he’s lived there since age 21 and his business culture is entirely American.
[The above paragraph is a condensed version of an entire post I should write sometime.]
One of the things Trump said in the Republican primaries that I found astounding was his complaint that, under George W. Bush we had launched wars that had cost about $4trln. How much better off, he asked, would America have been if we had spent that money here at home, on infrastructure. This was an absolutely stunning thing for him to say, and an indication that a sea change might be approaching.
What he actually did was to raise “defense” spending by about $50bln, and go along with the ridiculous tax cut passed last year. One understands WHY he dropped another $50 billion down the MIC rathole; presumably he figured that the military was the only part of the permanent state that wasn’t opposed to him. Of course, given the difference in drone strikes over the first two years of his administration in contrast to Obama’s, he has at least killed fewer foreigners who might have a grudge against us in the future. And he has so far avoided starting any new wars, unlike W or Obama (Libya.)
With the loss of the Republican House, maybe some change can be effected. I’d love to see the Democrats, with power of the purse, get the $150bln Amtrak wants for HSR from the $1 trillion we spend on the military (including not just DoD, but also interest on defense debt and spending for pensions and healthcare). They won’t, of course, because the goal isn’t to spend on lacking needs here at home but to maintain that gravy train doled out down in DC.
The entire Democratic Party needs a complete turnover at its top. It is amazing and depressing to see how much of the US has devolved into an gerontocracy. But it’s always been there, from J Edgar Hoover to Robert Moses. Can it possibly hold much longer? Pelosi may well have been a “great operator” over her 16 years in that job, however when does she accept she has overstayed? And perhaps clever operators are not what the US needs any more.
Musk is typical of a certain “immigrant done good” that becomes a hyper-American (no joke intended … kinda, but his adulation of this kind of capitalism is well past its sell-by date).
OTOH, re Hyperloop, I have always said that if it evolves into a version of standard maglev HSR–as it is in fact doing–then if it requires to be “American born” that’s ok. However I don’t think there is enough commitment, by him or the other hitch-hikers for it to be sustained. There is potential for a solid industry but not for any hyper-profits … so it will be left to the Asians and Europeans.
Pelosi is old, but her electoral strategy was so successful that Democrats are on track to win an even larger number of seats than they did in the aftermath of Watergate. She has already tampered down an internal rebellion against her and has plans for ambitious new legislation on voting rights and gun control. I agree that any infrastructure deal with Trump will be a bad move, but for her pure political effectiveness, I don’t see how Democrats could be better off with someone else. And remember she is the one who passed Obamacare when almost everybody else had given up.
Ha! What’s the old maxim: success begats a hundred parental claimants while failure is an orphan.
Seriously, no matter how clever she is, I don’t believe the wins were much, if anything, to do with Pelosi. I mean you really think she encouraged Orsini-Cortez in overthrowing that oh-so-typical machine man?
Your logic is empty and would never lead to change. And to argue about “ambitious new legislation” when it will all hit the McConnell wall in the senate is not worthy of discussion. All that really matters now is that they work towards winning all three arms of government (because alas it is the only way to try for real change) in 2020. And this is the exact same leadership that so cleverly brought you Trump.
It’s time.
Michael. I assume you mean Ocasio-Cortez and no that is not Pelosi’s job. Internal primaries are not her responsibility as long as they don’t affect general election prospects and the district is solidly blue anyway.
Most states have closed primaries. Or don’t have primaries.
Republicans brought us Trump.
And he’s exposing the Republicans for what they are.
Josh, way to miss the point. Which is that the change, all positive, that is sweeping the Dem Party of which Ocasio-Cortez is one sign, had nothing to do with Pelosi. That is part of my point. If Beto O’Rourke had pulled off a defeat of Cruz, you and others would have also claimed that as “a great victory for Pelosi” but it had less-than-zero influence (in Texas, quite probably the opposite). Ocasio-Cortez is significant precisely because it came as a shock to the complacent party managers and machine-men.
It is time for Pelosi to stand aside. Perhaps the best strategy is for her to say she will stay for 6 months (max 12 months, absolutely not 2 years) while her replacement (someone not in the current gerontocracy power structure) to learn the ropes as her deputy during this period.
Incidentally this feature of American political structure shows up its grave weakness: no real (ie. accepted as valid by voters) leadership for 80% of each presidential term. De facto, Pelosi and Schumer (or whichever time-server rises to that level; Harry Reid FFS!) assume that position but only as considered by the party machine men, not by the only people who count. I don’t know of any other political system that has this weird feature.
What the US needs is a federal gas tax increase to 40-50c and keep it indexed to inflation. That will bring in money for transportation year after year, and the increased funds next year can go to items that otherwise would have just missed the cut, or be saved up to give time for more ambitious thought-out programs to be developed. An infrastructure bill does not have to be lump sum.
What the US needs is a federal gas tax increase to $4/gallon and use it to plug holes in the general fund (and if the deficit ends up too low then cut other taxes, as British Columbia did with its revenue-neutral carbon tax).
But anyway, the big infrastructure plan has to be long-term anyway, on the order of 10-15 years.
There is no way to justify a $4 a gallon gas tax. At the current estimated social cost of carbon of $40, that leads to a 36c a gallon externality, and state gas taxes are already around 31c. Smog is not an issue in most of the country. Sure, you could generate tax revenue for other programs, but that is discriminatory and makes no sense. You could also generate tax revenue by having a tax specifically on computers or televisions or restaurants. Economically, that makes no sense as it causes shifts in consumption that waste value, so we instead have a flat sales tax.
Beyond carbon, an additional large social gain due to high gas taxes is decreasing traffic congestion.
Though maybe this is better implemented on the state level (congestion is more of an issue in New Jersey than Wyoming).
Fuel taxes aren’t really great at that, though – part of it is the Jersey vs. Wyoming problem, but it’s even more granular (Jersey City vs. Kearny, let alone Hudson County vs. Warren County). That’s why there are congestion pricing zones in some cities – they focus on trying to keep vehicles of any fuel economy level outside the most congested and transit-replete areas.
Congestion should be dealt with very locally and targeted, with different rates at different times, and highway by highway, and in some cities on key streets/areas. And since congestion is an externality that only harms other cars, the money should be spent on roads.
If you don’t spend the money on alternatives to roads….. people won’t have alternatives… They’ll still drive.
adirondacker12800 There is always the marginal case of discretionary long road trips or buying online versus driving to the store or carpooling. If they still drive after the gas tax that means driving is valuable enough. If there is enough demand for alternative means of transportation then build them, but that doesn’t justify huge subsidies to underutilized public transit.
Seriously? It affects, not only the entire economy, but its effects are built into the very fabric of the nation: urban design etc. Not to mention an existential reality for the planet.
Get back to us once cars start paying for their own needs. The last time I saw something it was a buck a gallon just to keep things paved. Then throw in policing. Half of a police budget can be for traffic control. And the courts. And the people who survive the accident and need expensive medical care for the rest of their lives. It’s somewhere between 3 and 4 bucks a gallon.
Total gas tax is over 90c in California and would be over $1 if federal gas tax is raised. Police also have to patrol public transit, and they do get revenue from traffic control. And medical care should be covered by increased car insurance liability minimums, up to economic value of life at $7M, thus making risky drivers pay, not everyone.
I will also note, that your original article on Amtrak’s inflated costs merely noted a sole bidder and that Schumer announced the winner. Do you have any concrete info on him being the source of the inflated costs?
There were two bidders, but Schumer prematurely announced that one of them – the one with the Upstate New York factory – had won. So in practice it was a one-bid contract, and the reason there was just one bidder was political sabotage.
So many politicians are so incompetent.
Um, that’s EXTREMELY competent politiciianing there, Matt. Just because he’s advancing something that advances his agenda and not yours, it doesn’t mean he’s incompetent.
That the system is set up to reward people who individually profit at the expense of the whole (see Harry Reid and the millions he made insider trading on stocks of companies about to be awarded government contracts, for just one example) doesn’t make the people who know how to exploit it incompetent. When there’s a 4.5 trillion dollar honeypot doled out in DC, you’re going to attract a lot of flies.
No, he is bad at advancing his agenda, too. His agenda does not include letting the Senate Republicans confirm tons of judges, and yet his deal with McConnell let that happen and then failed to save most vulnerable Democratic incumbents anyway. He’s genuinely pro-choice, much more so than Harry Reid ever was, and yet he’s slouching in keeping the courts committed to status quo as well as in defending vulnerable Democrats who are needed to vote down right-wing judges.
He faced a situation where a bunch of democratic senators were up for re-election, and not so many Republicans. The situation in 2020 is the opposite for the Republicans: two Republicans running for every Democrat. So his time to make hay is 2020, but to seize the Senate he has to not lose too many positions this year. Keeping the Senate at 53-47 means he is easily within striking distance for 2020. He cannot stop Republicans from appointing judges, only slow them down a bit. He could have used a procedural move to do just that and did not do so. Perhaps that was also a political call, in the sense that he did not want to stir voters on the right at all.
Since the Republicans have removed the last restrictions on Judicial appointments, a Democratic President in 2020 in combination with a flipped Senate can appoint Progressive judges with no ability of the right to stop them. I think he has set himself up well for that possibility, rather than wasting troops and energy on the battle this fall that he cannot win.
That having been said, I detest the man. Maybe some young Democrats can stage a coup in the Senate, just like the House.
He faced a dilemma and somehow managed to get the worst of both worlds: all the vulnerable Democrats got booted and even Tester and Manchin barely made it, and McConnell got all of his judges.
No matter what they did, McConnell was gonna get his judges. They could have handled it better but when idiots are digging themselves into a hole you don’t wrestle the shovel out of their hands. Ya let them dig.
Schumer famously got 1600 on his SATs. He really thinks he’s brighter not just than everyone in the room, but any combination of people.
Again, he prematurely announced the winner. Yes, he was tooting his own horn as the factory was in his own state, but that doesn’t prove that he manipulated the contest, perhaps he just had inside sources. It is not obvious that the Senate minority leader has such power over an executive branch agency.
The Gateway project may be bloated and have undesirable aspects….but don’t we need SOMETHING in the way of new tunnels, ASAP?
It’s useful, but North Jersey won’t collapse if the tunnels aren’t built. It’s telling that the apocalyptic rhetoric about tunnel closures began not right after Sandy but later as it turned out there was no political will to sink $20-odd billion into Gateway when sold as a capacity expansion.
To beat a dead horse, ARC was scheduled to open this year. It probably wouldn’t but it was scheduled for this year. New Jersey and everybody else in the Northeast and therefore the rest of the country has a problem if there is a problem.
“It’s useful, but North Jersey won’t collapse if the tunnels aren’t built.”
Unless something happens to one of the existing train or car tunnels.
It’s telling that this rhetoric of urgency was not deployed until funding for the tunnel based on an appeal to costs and benefits failed. The study Port Authority did on this question recommended that Gateway be built to facilitate long-term closure for repairs on the existing tunnel, but did not say the existing tunnels were at risk and did say that closing the existing tunnels one tube at a time during weekends (which already happens) is feasible but more expensive, without a dollar amount attached because then people would notice it’s an order of magnitude off of the new tunnel’s cost.
I wouldn’t rule out road maintenance as a viable stimulus project. Recessions happen very roughly about every 10 years, which is also how long asphalt roadways tend to last. Private-sector construction is strongly cyclical; governments shifting more repaving work to recessions would smooth construction employment.
A lot hangs on “roughly”… if the recession strikes after 13 years, then you have a rapidly deteriorating road in the last 3 years.
They don’t do it on a schedule. People drive over it with machines that record how bumpy it is. Checking how bumpy it is, is on a schedule but it doesn’t get repaved until it’s unacceptably bumpy. The schedule for checking and how bumpy can be finagled which is why we are where we are.
In other words, an ideal deal would actually be a reform not a spending bill?
I’m always convinced that too little is said about how weirdly the US devolves certain policy decisions, and infrastructure is a great example* (but I’m sure you’ll tell me if it isn’t really so, relatively-speaking; though, it certainly seems like the Feds control spending while states et al control actual decisions 🙃)
___________________
*The way HSR monies were doled out to state(s) with “shovel-ready” projects, while necessary by our policy framework, always struck me a short odd. It’d have been smarter to have had the FRA/Amtrak do what has been done with NEC Future but with national scope (and an actual semblance of priority projects/corridors, costs, so on) that states could then have implemented on their own and/or could inform future decisions in periods of federal stimulus.
As it stands, the Obama Administration made a decision that I’m not so certain was *wrong* given the lack of “infrastructure” at the federal level to do such a thing?
I don’t know… I’ve been increasingly thinking that reconstituting all these agencies (eg. FAA, FRA, USDOT, Amtrak, FTA, HUD, etc) under a new Dept. of Infrastructure/Development is a good direction we could go in — along with actual national-level, long range planning.
Your assertion that Trump is corrupt needs some evidence. How is he corrupt? What exactly has he done? Are you referring to his various business dealings when he was a private citizen? What is your calculus here? I’ll readily concede that Trump is largely a buffoon. He is narcissistic, thin-skinned, has the impulse control of a 10 y/o, is ignorant of most things, is not by any serious standard an educated man.
Do you mean that he colluded with Russia? Show me some evidence.
I don’t mean the collusion; I mean the expectation (e.g. during the transition) that foreign visitors to DC must stay at the Trump hotel as a show of respect, or the complete lack of separation between politics and business (whereas Carter put his business in a blind trust).
Ok. I was unaware of that.
They haven’t finished gathering the evidence. They are still getting information from multiple cooperating witnesses. Actual conspiracies, there seem to be many, can take a long time to disentangle.
What kinds of projects should be covered in the plan to be written now and then implemented when there’s a recession? Game-changing stuff like real HSR? Or stuff more like finishing up the CREATE basket of projects in Chicago, electrifying MBTA, extending light rail lines, etc? Quality or quantity? I don’t actually know which I’d go with and so I’m curious to hear your thoughts.
^Quality vs. quantity is probably the wrong phrasing, since some cheap projects (like MBTA electrification) are high-quality. Really I am asking whether lots of affordable projects or a few expensive but very significant projects are better.
I’m curious to hear my thoughts too! My answer is that this depends on several factors.
1. Total amount of money available. Big projects (like California HSR or some regional rail tunnels) are expensive and often useless if you only build half of the project (like San Jose-Bakersfield HSR, gah). So even if they have high ROIs you need to have enough money on hand to build the entire thing.
2. Relative construction costs. The subway : light rail construction cost ratio is on average 5-6 but varies widely across cities.
3. Project dependence. You can’t build NSRL in Boston without electrifying. This interacts with the other factors.
4. Local idiosyncratic issues. Maybe a city has a lot of good medium-impact projects but no good big projects, or maybe it’s the other way around. As I understand it Philadelphia is in the former category and San Francisco is in the latter.