Hacker News Reader

Comment by bradleyjg | Hacker News Reader

I don't have any love for NYC construction unions and their ridiculous work rules, but the fact that France -- France! where strikes shutting down the whole country is a national pastime -- does better means that can't be the number one answer. Similarly we can rule out corruption because even Chicago is a piker compared to Shanghai. I think both unions and corruption play a part but not of the first rank.

The first candidate I'd propose for the top slot is diffusion of responsibility stemming from a US obsession with localism. An MTA project might well be paid for by: the federal government, NY state, NYC, and bonds to be paid back by a dedicated tax in 10 downstate counties. There are contracting rules imposed by each of those sources. And it is administered by the MTA board, a monstrosity with appointments from eight different sources. The buck stops nowhere.

The second candidate is, paradoxically, the progressive era procurement rules designed to rein in corruption. They leave too little discretion. In a two sided competitive game, if one side is hampered in by slow changing inflexible rules the other side is going to dominate. That's exactly what happens. Government contractors play the system like a fiddle and there's little the government negotiators even when they are supremely competent and diligent can do about it.

View on HN

> The first candidate I'd propose for the top slot is diffusion of responsibility stemming from a US obsession with localism

If it were only localism, there would be localities that were fantastic at investing in their own infrastructure. Responsibilities, even funding local schools, have been nationalizing slowly over the years without an explicit change in the U.S. Constitution. So the national government has to 'hack' the system with grants and taxes to indirectly get national policies enacted. It's a corruption of the system to work around the Constitutional protection of federalism.

One solution would be to amend the Constitution to make things like education national policy once and for all. Another would be for the national government to leave things like building roads and funding schools to the local governments.

If a transportation system crosses more than one jurisdiction than it doesn't really matter if some localities are fantastic at investing in their own infrastructure. It is inherently a problem that can be solved optimally only at a different level. It might be possible to design a system such that metro areas are always a single jurisdiction, but given that such areas grow and shrink over time it wouldn't be easy.

I don't think that "corruption of the system" is a cause in and of itself of the high cost of infrastructure construction or repair. Whether or not the Constitution would need to be changed, that's an implementation detail. First we need to figure out if diffusion of responsibility is actually a major problem (as I suggest), then get widespread agreement that that is the case, only then should we start worrying about what legal barters may or may not be in the way of mitigation strategies.

>> One solution would be to amend the Constitution

Err. That's probably a tall order.

Maybe. It used to be amended every 10-20 years, depending on how you count. It has been 25-40 years (again, depending on how you count) since the last successful amendment attempt. I think that, to some degree, the amendment process has been abandoned because people think the government can hack around it with good results.

I think it's worth considering that this sort of corruption is a side effect of that approach versus the cleaner debate-and-amend process that the U.S. practiced historically.

I think the U.S. government will be ultimately doomed if the U.S. can't figure out a way to amend the Constitution again. All sorts of new wrinkles have popped up since the early 90s and we have abandoned the process for democratically deciding how to deal with them. It really erodes the legitimacy of the process.

No less a man than Scalia himself said that, if he could wave a magic wand and change the constitution, the change he would make would be to make it easier to amend.

I didn't agree with most of what he said, but I thought he was spot-on with that.

It has been abandoned because of a slew of SCOTUS rulings that have significantly broadened the powers implied in some clauses (most notably, the Commerce Clause, and the Due Process clause of the 14th Amendment), allowing the federal government to work around the system more often than not; but also because of increasing partisanship of the political system making the 3/4 majority of states necessary to amend anything extremely unlikely on any issues where there isn't a near-unanimous consent already.

I think nationalizing more issues also leads to more divisiveness. If you get fed up with your city government, you can move. What do you do if you get fed up with your national government?

On top of that, if relatively slim majorities (55%, say) can drastically change the rules (Citizens United, ACA), then there is a loss of legitimacy in the government. This also causes divisiveness since the country becomes a nation of power (who can scrape together a majority?) and not a nation of laws.

The feds could also do some easy, central infrastructure things that would reduce school and local govt funding needs in the first place. If the schools or localities want to buy their own, great. Three examples:

1. It seems every school uses a mishmash of Moodle, Schoology, and some Sharepoint abortion. If the feds would spend around the cost of 3 or 4 F-35 helmets ($1.6M) they could build and offer a decent school site. You know, homeworks, attendance. It could be the same software everywhere, even if you changed schools. Teachers wouldn't need retraining and kids would have their stuff follow them.

2. Fund a decent central, unbiased curriculum with open source text books. Kansas can suck an egg. Localities really _don't_ need their own curricula. The 3 R's don't need to vary with district.

3. Same as 1 but for local government operations software. Do they really all need their own?

Number 2 is never ever going to happen. Do you remember all of the bugaboo about Common Core? That wasn't even the Feds getting involved in setting curriculum directly.

The right of a parent to indoctrinate their children trumps the right of the child to have an actual education.

I don't know the answer, but it's a gross feedback loop.

>an actual eduction

I'm not one of those people, but here it should be "indoctrinated by the state", because I think that's what our schools do. I think we play fast and loose with the term "school". For being an institution supposedly concerned with imparting knowledge on kids, the current system does a pretty shitty job of it. Most kids don't even know the algebra they "learn" in school, let alone stuff like how to pay rent, taxes, how to drive, how to budget, how to vote, how to work effectively in team situations on common tasks, how to survive in the wilderness (yes, that skill is taught in some countries). Our "schools" are awful. Maybe it's as simple as a big money injection. I dunno. But as is, it's awful.

I think you're being a little hyperbolic. And the question is whether the parents' exercising their "right" to promote their religion of choice and cut out the teaching of evolution and sex ed might not leave the child with some deficiencies in their education. It generally depends on how qualified the alternatives are, but the answer is probably "yes it does".

I'm not saying anything about those awful people. I'm saying that our education system is a giant mess. I am the polar opposite of those parents, but I know a few of them on Facebook and find that I agree with their sentiment about schools more often than not. And that is a problem, when the system is in a condition such that I find myself agreeing with those people about schools.

I wouldn't bet on a money injection fixing anything - I've read a number of reports that show almost no correlation between performance and per-student-dollars.

Yeah, you're right, douche.

On paper that's interesting. I'm skeptical that the national government can do that well, however. The rollout of the ACA websites makes it clear that the government isn't good at producing that sort of product.

Furthermore, the premise of the article is that the government is very inefficient. I'm not sure $1.6M is a fair estimate for the cost of this sort of product. There are millions of teachers in the U.S. not to speak of parents, administrators, and kids.

I'm open to providing incentives to make this sort of progress more feasible. I'm partial to better support of open source software that meets these ends (like tax deductions for time or cash contributions to OSS projects the government uses).

>The rollout of the ACA websites makes it clear that the government isn't good at producing that sort of product.

You mean that the private contractor hired to do the HealthCare.gov website, CGI Federal, did a lousy job?

Most of the things the government "produces" are not produced by the government. They're sent out for bidding by private companies - roads, guns, websites, whatever - and the government is in charge of oversight.

And the fact that they've been proven incapable at oversight and management should figure into the planning.

But what makes you so sure the government wouldn't do better taking it in-house? And even if we throw out that possibility, we're talking about implementing off-the-shelf software here, no producing custom software.

Great point! And since - as described - it's all open source, there doesn't even need to be Federal oversight or management.

If the only thing standing between now and doing it is the will and tech support, that's an easy problem to solve.

> You mean that the private contractor hired to do the HealthCare.gov website, CGI Federal, did a lousy job?

That may be true as well, but any customer that's changing requirements at the last minute and demands idiotic deployments (nationwide, hard deadline, untested, lots of PR, etc.) is ultimately responsible for what happens.

Maybe every single vendor should have sensed that outcome, but it's more realistic to think that someone will take that contract, even if it's doomed.

ACA was very complex and the time from the law actually passing to the site going live was pretty short. Until the law was completely signed there was basically no spec for GCI to work with other than some vague generalities about what it'd have to do. I'd put a fair amount of the blame for how shoddy it was to that compressed timeline as much as contractor waste and inefficiency.

> ACA was very complex and the time from the law actually passing to the site going live was pretty short.

And the federally-run state exchanges were a fallback that was intended to not need to be used, anyway; the whole point was that every state was going to have an exchange and that not setting one up wouldn't affect that, with the idea being that in that scenario even states that didn't support the idea would prefer to have control of the state exchange rather than handing it to the feds. The timeline from determining that there'd actually need to be federally-run exchanges to the site needing to be live was even shorter from the time from the law actually passing to the site going live.

> I'd put a fair amount of the blame for how shoddy it was to that compressed timeline as much as contractor waste and inefficiency.

That's more or less my take, but I consider that part of the government's terrible project management.

> There are millions of teachers in the U.S. not to speak of parents, administrators, and kids.

How's the number relevant? It's not like they'll charge per user, and each school only cares about O(10,000) users at most.

Huh? I worked in school district IT for quite some time and a huge number of services are billed per-student.

> How's the number relevant?

I'm just saying $1.6M is underfunded for something that complex. That's on the right order of magnitude to build out a proof-of-concept to pitch for an actual production rollout.

Keep in mind that you'll be dealing with a lot of relational data with some interesting security requirements, some of them legal. It's not just photos of cats.

I believe the US recently paid about as much for the functionality of a coin flip on an iPad and the software was only used for a few months.

OP is talking about much more complicated software than simulating a coin flip.

But the reason the Fed don't build cheap schools is the same reason the Feds do buy million dollar helmets.

Regulatory capture extends to every level of government so switching the level of government you deal with won't change the problem.

> Fund a decent central, unbiased curriculum with open source text books.

The key with this would be to fund the effort but not mandate it. The carrot for using the federal standard books is that they would be without license fees, which means localities would have to justify spending their own money to use something else. But if the feds do such a crap job that localities find it worth it to do that, they should still be able to. Or do that with some books but not all of them etc.

The free software model (something like the GPL) would be very interesting here, because you could have local schools making contributions to the national curriculum that other schools could adopt or not based on merit rather than local budgets. But not have some textbook publisher doing what Apple does to BSD, i.e. making sufficient improvements that theirs is the dominant version and then prohibiting further improvements by anyone else.

> Another would be for the national government to leave things like building roads and funding schools to the local governments.

But then you'll have people from poorer neighborhoods/towns drive on worst roads and sending their kids to crappier schools compared to wealthier parts of the country. The roads' problem let's say is not the end of the world per se, but the fact that your kids will have less chances in life only because your neighborhood is not as wealthy as the one 5 or 10 miles down the road seems moraly wrong and a misallocation of resources.

School funding is already essentially through local government. It's where the vast bulk of property taxes go to. And, yes, it leads to high-priced suburbs with good schools, with schools that are often less good elsewhere. (It also means that people with children in large urban areas often feel they either have to move to the suburbs or put their kids in private schools.)

Bad urban schools often get more funding than better suburban schools[1], and how's that working out?

[1] https://randomcriticalanalysis.wordpress.com/2016/05/09/my-r...

Fair enough. Funding is often one factor but there are obviously many reasons why suburban schools in good neighborhoods are "better" (have superior outcomes) than those in many city neighborhoods. And, yes, there's not much evidence that more money is a particularly effective fix in many poorly performing school districts.

Mostly I was just responding to a comment that seemed to not be aware that school funding was already mostly local.

It's in the national interest to have well maintained roads the military would be able to traverse. That's really why we have the interstate highway system in the first place.

It's also in the national interest to have a well educated population with a generally common knowledge foundation. Education is a cultural and nationalistic cornerstone.

If you use "in the national interest" as your threshold, you can apply it to almost anything:

It's in the national interest to have a healthy population. For everyone to work out regularly. For everyone to read every day. For everyone to travel the world and learn about other cultures. For everyone to get a full night's sleep.

That's why that is not the US standard. We explicitly have restraints and constraints on government. If you want to remove those restraints and constraints, that's fine. There's a process for that.

But if you ignore the restraints and constraints instead, then government power is based on whims and preferences of whoever is in power. And we know how that works out.

The government takes steps to encourage many of the things you've listed.

Highways are much less of a military necessity than when they were built. The nature of warfare is no longer based as much on troop movement.

Realistically, I don't see how an "every municipality for itself" (or even at the state level) system wouldn't make things much slower and more inefficient no matter what. It seems like a recipe for either too much duplication of function or else an unclear chain of command.

> If it were only localism, there would be localities that were fantastic at investing in their own infrastructure.

Seattle. Great water, renewable energy, lots of transit investment.

> If it were only localism, there would be localities that were fantastic at investing in their own infrastructure.

There's at least one: Bellevue, Washington, USA (across the lake from Seattle). I lived near there growing up and watched in awe how they repeatedly refactored the city in record time, over and over and over again, improving it each time. It was almost like watching agile development happen physically in front of you.

I was talking to somebody about how education ends up being so expensive in the US, and I kind of feel like there are similar reasons.

Subcontracting out work + massive federalism creates $200 textbooks. Now I know national textbooks are a bit much, but I would be interested in seeing a state completely co-opt the textbook creation process and produce some textbooks and learning material at cost.

As for the budget distribution, this pretty much happens in other places too. My French school's expansion was half funded by the ministry of industry, some by the region, some by the national government... I think that's a side effect of how budgeting works.

> Subcontracting out work + massive federalism

I think you meant massive anti-federalism.

If we had a federal procurement system for math and english textbooks, after being put out for a nationwide bid, the per-unit cost would be about 10% over the cost of paper and printing. (the 10% premium is the bid adjustment for a women or minority owned business)

Now $8 textbooks wouldn't save a lot of money compared to the national education budget, but it would be a good start. And do we really need regional variations in math textbooks?

The only reason I know of for different textbooks for different states is that some states seem to want to push a religious agenda in their education system. I struggle to find any net negatives here except for religious fundamentalists and textbook printers.

> The only reason I know of for different textbooks for different states is that some states seem to want to push a religious agenda

Until the federal government starts pushing a religious agenda. Then everyone here would want textbooks from, say, California.

This can't be the whole story. Consider math books. Publishers have coordinated math curriculums, perhaps a dozen different ones used across the country. I feel that the people picking these books are persuaded by the different book publishers to buy expensive, colorful math books that fundamentally are inferior to simpler, less costly math curriculum/books, like the math program and books used in Singapore (which are in English).

CommonCore is the Solution to these 20 different MATH text books across USA. 40 States are onborad with CommonCore https://en.wikipedia.org/wiki/Common_Core_State_Standards_In...

This gives the WHOLE Sotry .. How Bill Gates pulled off the swift Common Core revolution

https://www.washingtonpost.com/politics/how-bill-gates-pulle...

The only reason I know of for everyone using the same textbooks nationally is that some people seem to want to push an agenda in their education system, and see that as tool for doing so.

Using the same textbooks nationally has an obvious benefit of entire nation learning the same things. Standardization of the curriculum is not a bad thing.

Or let's do it like a bunch of other countries: children have the reaponsability of taking good care of the books and pass them to the next generation at the end of their class. It's a great way of teaching respect for the public good.

A yet better way would be to have these books as open-source and anyone is able to print them at their homes/public livrary/etc.

A classic story in local elections is that books are 20 years old and the system needs a millage to buy new books.

If you assume that a student uses ~10 books a year and that they last 10 years, with $200 books you get grade schools spending about $200 per student per year on books. If it's 20 books and 5 years, it's $800 per student per year.

So while there is probably an obvious opportunity for grade schools to save money on cheaper books, it's also pretty obvious that it isn't the main component of educational spending at that level (the high price scenario puts it at less than 10%).

I think things have gotten worse, but I probably only spent $1500 on books in college (and ~$40,000 on in state tuition). So that's a noticeable fraction, but it wasn't in any way something to worry much over compared to the total costs.

It's not a question of total costs so much as how cheap this would be to fix. The only real issue is text book manufactures like tax software company's have a huge impetus to maintain the current system.

Which is who the government wastes so much money outsourcing in the first place. It's all defused pain for personal gain.

I think the total costs inform how outraged to be. They are probably less than 5%, so if you permanently reduce them by 1/2, you've managed to offset ~1 year of cost inflation. That sort of scenario is not the first thing you address when you look for systematic ways to improve the effectiveness of spending (you of course try do them along the way, but they aren't worth burning much time or political capital on).

Inflation adjusted costs are not going up 2.5% a year. Further, I don't think there are many easy ways to save 2+% of total tuition spending.

> Or let's do it like a bunch of other countries: children have the reaponsability of taking good care of the books and pass them to the next generation at the end of their class. It's a great way of teaching respect for the public good.

This already happens in most primary schools because buying new copies every year is prohibitively expensive.

And do we really need regional variations in math textbooks?

Depends on if you trust the federal government to do a better job selecting a textbook than the state you currently live in.

Having seen how the Texas board of education works at adding creationist and revisionist junk to science and history text books, I absolutely trust the federal government to do a better job selecting a textbook than most states.

If a Rick Santorum or Mike Huckabee presidency would change your opinion on federal control of education, perhaps you should reevaluate the underlying principles.

Realistically it's really whether you expect them to do a better job than major states like California and Texas. Nobody is producing a special Rhode Island edition of a textbook.

The federal government does not need to select those books.

It can merely create a few options (from the best professionals it can find on the entire federation), and make them available for the federated entities.

> do we really need regional variations in math textbooks?

I think there is value in having our population exposed to a diverse range of textbooks. I can't prove it, but I suspect that a homogenous education system would be detrimental to our economy. One of our national strengths is innovation, which comes partly from having a number of people looking at the same problem in different ways.

For history and social sciences absolutely.

For high school level Math, not really. The downside I can see is that you stifle any innovation in teaching methodology.

The issue would be deciding what was the "best" math book. There are some really crappy ones, at least in my opinion. Someone else would hold those same books up and say they were the best.

Right! You're exactly right. You want a great example, look to Texas. They are one of the big reasons we live in this nightmarish dystopian present of $200 textbooks that suck to begin with.

When I mentioned federalism, it was more about the fact that local and state governments have a lot of control on the education level. Usually meaning a lot of different kinds of books.

If we weren't a federal system, the national government would be dictating much more about education than it currently is.

The ultimate cause is regulator capture. A series of large companies extract massive rent through their ability to direct state funds into their coffers.

If the US as it exists now tried a Federal textbook, those would cost $200 again at the very least and would be put together by all the major textbook companies combined (unless one of managed a clever end-run). We see this in military contracting for example.

The capture is probably strongest at the State and Federal level just because centralization is easiest but these hand-down regulations that force the hand of local authorities - not controlling them entirely but preventing any tactics that cut-out the mass of rent seeking monopolies that now count on state funds for their sustenance.

> Now I know national textbooks are a bit much

We have an enormous federal Dpt. of Ed. but we don't have no-cost, open source curriculae, textbooks, school operations software etc. That's a damn shame. That kind of resource could transform education in non-affluent school districts.

I agree, and I suspect the union problem is a symptom of the rigid procurement rules. Only a few contractors know how to navigate the system, and so the market is not really "contestable." Even though it's the private sector, unions retain bargaining power.

Also, with major urban projects there's usually political pressure from the local community to use local union labor. That might make it into contracts even if it's not a locally managed project. Again, the underlying problem is not unions, but excessive localism.

France has a long history of making good infrastructure (TGV anyone ?) so there's that.

Also, more vacation (5+ weeks) + better health care -> more productive workers, especially in physical labor work.

My wife manages large scale remediation projects and says that on average a government contract costs approximately twice what a non-government contract would due to time, not materials cost.

All contracts have to be proposed with estimates, awarded, negotiated, revised, and re-proposed. Everything has to be approved and the lead times for going into the field or requisitioning materials is counted in months, not weeks or days. All contract project managers have a corresponding Government assigned project manager who must approve everything.

She said procurement rules are onerous and have fun restrictions like "all materials must be sourced from the USA unless unavailable" that add to the time needed to find and vet sources.

Because of all this, the government typically awards contracts to companies it has worked with previously and not necessarily the lowest bidders. The contractors almost always subcontract out everything except management. It's also not uncommon when the government has to award a contract to new contractor, they will also contract with a familiar firm for PMs to manage the new contractor.

So maybe a little bit of your first candidate and less of the second?

Employees generally don't get paid during a strike.

Unions generally have a fund for this.

One might also expect the union to negotiate a signing bonus (as it were) for the members who missed out on pay during the strike. SPEEA did that during at least one Boeing strike.

They don't get paid by their employer, correct. To cover costs, the union backing them will typically issue strike pay.

But the project doesn't make any progress until they're back to work, and labor costs may be higher after negotiation. That's why parent commenter was surprised that France - a country with strong labor unions that strike often - still has lower per-mile infrastructure costs than the United States.

Reply on news.ycombinator.com