Oversight of cost overruns does not always help

In the United States, 42% of public infrastructure projects report delays or cost overruns. To mitigate this problem, regulators scrutinize project operations. We study the effect of oversight on delays and overruns with 262,857 projects spanning 71 federal agencies and 54,739 contractors. We identify our results using a federal bylaw: if the project’s budget is above a cutoff, procurement officers actively oversee the contractor’s operations; otherwise, most operational checks are waived. We find that oversight increases delays by 6.1%–13.8% and overruns by 1.4%–1.6%. We also show that oversight is most obstructive when the contractor has no experience in public projects, is paid with a fixed-fee contract with performance-based incentives, or performs a labor-intensive task. Oversight is least obstructive—or even beneficial—when the contractor is experienced, paid with a time-and-materials contract, or conducts a machine-intensive task.

That is from a new paper by Calvo, Cui, and Serpa, via the excellent Kevin Lewis.


shocker: increasing bureaucracy and paperwork increases expense, even when it's supposed to somehow decrease it. And contractors who are used to the bureaucracy are most accustomed to dealing with it...

Republicans are quickly becoming a parody of Obama’s mantra. The No We Can’t! Party.

Posts like this are just red meat for the No We Can’t base. The cause of Good Government is now strictly partisan. Sad!

Yes, I've tried to tell people that increasing bureaucracy and paperwork is a sign of Good Government.

Oversight of delayed projects does not mean more bureaucracy.

Occam’s razor would indicate the causality is either backwards or exogenous. Chemo does not cause cancer.

The Mythical Man Month would argue that, yes, the causality is in the right direction. Government procurement has a number of issues, and in places where firm fixed price is appropriate, the incentives mean that detailed oversight is not as useful as in T&M, where the differing lack of contractor incentives for cost control makes oversight comparatively more useful.

"We also show that oversight is most obstructive when the contractor has no experience in public projects,..."

Why work with a contractor with no experience in public projects in first place?

President Trump knows the answer to that question.

It isn't hard to figure out, to be honest - at least with a bit of familiarity with how things work in a country where this court case occurred - https://en.wikipedia.org/wiki/Citizens_United_v._FEC

Money is a political tool protected by the 1st Amendment - and money is more than able to make itself heard.

This web site tends to be interested in certain types of corruption, without ever noting other forms. Which is strange in its way, as this particular sort of behavior would seem to be a wonderful validation of the insights of public choice economists.

Instead, we have a post concerning oversight being counterproductive involving those companies without any experience in public projects, and not why such companies get their contracts in the first place. (Apart from the fact that everybody is a beginner the first time, obviously.)

A nice reminder that prior is a troll. Irrelevant Trump reference, off-topic Wiki dump about political advertising, followed by vague ad hominem insinuation about this blog, because some other economist wrote a paper on one topic and not another.

'Irrelevant Trump reference'

Really? Recently, it seems as if various Trump properties have been receiving the largesse of the federal government, though previously, they had not been involved in providing accomodations to military personnel - or foreign leaders. You do remember President Trump's invitation to his fellow G7 members to stay at his property, right? Of course, the Trump Organization is just demonstrating why some people, when given the chance, tap public coffers to make money for themselves. I seem to recall that was considered blatantly obvious when one used the name Clinton.

'Wiki dump about political advertising'

Well, depends on your perspective on how effective a corporation's various forms of political support are worth. Though likely, only a troll (or public choice economist) would suggest that a company spending significant money to help a candidate win an election would expect to just 'coincidentally' receive support from that lawmaker in receiving a government contract. Here is another irrelevant example involving President Trump - 'President Trump has personally and repeatedly urged the head of the U.S. Army Corps of Engineers to award a border wall contract to a North Dakota construction firm whose top executive is a GOP donor and frequent guest on Fox News, according to four administration officials.

In phone calls, White House meetings and conversations aboard Air Force One during the past several months, Trump has aggressively pushed Dickinson, N.D.-based Fisher Industries to Department of Homeland Security leaders and Lt. Gen. Todd Semonite, the commanding general of the Army Corps, according to the administration officials, who spoke on the condition of anonymity to discuss sensitive internal discussions. The push for a specific company has alarmed military commanders and DHS officials.

Semonite was summoned to the White House again Thursday, after the president’s aides told Pentagon officials — including Gen. Mark Milley, the Army’s chief of staff — that the president wanted to discuss the border barrier. According to an administration official with knowledge of the Oval Office meeting, Trump immediately brought up Fisher, a company that sued the U.S. government last month after the Army Corps did not accept its bid to install barriers along the southern border, a contract potentially worth billions of dollars.' https://www.washingtonpost.com/immigration/he-always-brings-them-up-trump-tries-to-steer-border-wall-deal-to-north-dakota-firm/2019/05/23/92d3858c-7b30-11e9-8bb7-0fc796cf2ec0_story.html

Almost as if Fisher could be seen as an illustration of the connection between political support and getting a public works contract. Or an illustration of how Trump knows the answer to the question as to why a first time federal contractor might be chosen.

Hope that was straightforward enough to meet your demanding standards for keeping things as plain as possible, by making the default assumption that commenters here are simply ignorant of current affairs - not to mention public choice economics. (Or as some like to call it, the Will Rogers School of the Obvious.)

"You do remember President Trump's invitation to his fellow G7 members to stay at his property, right?" Because the Secret Service already had the placed secured obviously.

This was not the Florida White House at Mar-A-Lago, this was another resort, the Trump National Doral Golf Club .

But don't worry, one can be sure that the Secret Service will be racking up all the expenses necessary to ensure security - rooms, meals, etc. As will the security services of at least 6 other nations.

Admittedly, if the president invites foreign leaders to spend money at commercial property he owns, that might not techincally be against the emoluments clause (depending on one's point of view, of course). Unlike military personnel staying at commercial property where the proceeds of their stays directly benefits their commander-in-chief.

Always fascinating to see how the originalists are able to effortlessly ignore one of the clearest examples of the Founders being constitutionally explicit in forbidding the sort of practices which were routine in the monarchies they were familiar with.

Note to prior: read your stupid wikipedia article before citing it. A donor to a political candidate or party is ipso facto not engaged in the independent expenditures covered by Citizens United, but then clueless people like you treat it like a bogeyman.

'... read your stupid wikipedia article before citing it. A donor to a political candidate or party is ipso facto not engaged in the independent expenditures covered by Citizens United'

And maybe you should read the first comment, which said this - 'depends on your perspective on how effective a corporation's various forms of political support are worth.'

Various forms of support include donations, which as you effectively pointed out to anyone who was actually unaware of the point, were unaffected by that case. Running a coordinated ad campaign ('electioneering') with a political campaign or advocating the election of a candidate, however, is just fine.

Seems to be adequately covered by 'corporation's various forms of political support,' assuming one actually reads the comment.

And yet you chose to respond to Tom T.'s comment by mentioning a GOP donor, which is not an independent expenditure (which, by definition, excludes donations or any coordination with a candidate).

Sometimes there are rules or laws about having to choose the lowest bidder in an effort to minimize cost, so an inexperienced contractor can come in, low bid the entire project, and the government's hands are tied and they have to choose that contractor.

other times, there are requirements that some work has to go to minority-owned, or women-owned businesses, and those contractors are also inexperienced, but by law, they have to be awarded a contract.

There are also rules for that. In my office we work for a government organization for country well-known for bureaucracy.

If a low bidder cannot finish the work in the offered terms, the bidder loses the contract to the 2nd place bidder and must pay back the price difference between his low offer and the 2nd place offer.

This rule lets all people play the game while minimizing the risk for the contract emitting organization.

How does this work if the bidder is halfway through the project? It's not like they can just pick up all their gear and leave. It would take months to years to get through all the legal challenges to kick the contractor off the project, then more money and time to get a new contractor on board at an even higher cost, because the new contractor would have to demo what the previous contractor did, and re-do it. A new contractor is not going to take responsibility for work that is not their own, and the government certainly isn't going to take responsibility for that unfinished work either.

Many times the constructor is required to hold a "performance bond" or a "surety bond" that, in the event of constructor non-performance, comes in and pays to finish the job. These bonds are ~1% of the total price.

As far as legal challenges go, the FAR (Federal Acquisition Regulations) allow pretty rapid Termination for Default, so the contractor would be outta there quickly, and would be sorting it out in the courts later on.

The performance bond is really to guard against the contractor taking all the money and skipping town. The contractor could really be making a good faith effort to complete the work, but is just not experienced enough to complete the project on time. In this case, it would be difficult to legally kick the contractor off the job.

Yes, unfortunately the US (rooted in understandable reasons of attempting to avoid corruption) has some pretty strict and poor procurement laws that force governments to accept the low bid, even when the low bid is made by a group that has little experience, or perhaps even a long history of many change orders and ending up costing much more than their lowball bid (e.g. Tutor Perini.)

This is designed because of worries about corruption (which certainly did exist in the days of "honest graft") involved when you allow people to select the winning bid based on having a good long term relationship and experience actually getting the job done.

If you limit it to experienced contractors, there may be only a few that can bid the project, driving up the costs.

Eg, SpaceX would never have gotten its first NASA contract, and rocket launch costs would still be more than twice current pricing.

Whether it's building an aircraft, an electric car, or a highway, whether private or public, we aren't very good at it. In contrast, we excel at banking. Today's NYT has a flattering piece about Stripe, the internet payments startup founded by the Collison brothers, which just received a new round of funding that values the private company at $35 billion. Stripe, unlike many Silicon Valley startups, has a steady stream of income from the 2 to 3 percent it collects on every credit card swipe. Stripe is now expanding its banking business with loans to customers and a Stripe credit card. We are very good at banking, not building. To paraphrase Peter Thiel, we wanted infrastructure, instead we got money changers. https://www.nytimes.com/2019/09/19/business/stripe-valuation.html

My observation about public infrastructure projects in my low country community is that those projects more often than not aren't undertaken for their own sake but to facilitate private projects. Road and water and sewer projects are undertaken not for the sake of roads or water and sewer but to facilitate new construction in this growing community. For that reason, the infrastructure projects are undertaken in haste and contracted with the lowest bidder who promises the quickest completion date. Not surprisingly, the projects are poorly designed, poorly supervised, and poorly constructed and held together with haywire. But they serve their intended purpose, which is to facilitate all the new construction by private developers and builders. One won't be surprised to learn that the new construction by private developers and builders is shoddy too: shoddy public infrastructure to facilitate shoddy private construction. All of it funded by, you guessed it, bankers.

"To paraphrase Peter Thiel, we wanted infrastructure, instead we got money changers. "

Why is Thiel complaining? He made his first fortune from Paypal, a money changing business. If he wants infrastructure he should pay for it, but he hates taxes and government. Always complaining that Thiel.

I was paraphrasing. Thiel actually said: "We wanted flying cars, instead we got 140 characters". Thiel was one of the original investors in Stripe. I listened to an interview of one of the Collison brothers describe his interview by Thiel, in which Collison described to Thiel the deficiencies of Paypal's business model. Thiel has a thick skin (at least when it comes to business): he made the investment in Stripe.

What about the effect of the threat of oversight? If oversight is sufficiently annoying, I'd expect it to lower cost overruns overall.

Nope. I work for a federal agency sometimes. Annoying definitely costs more. a. I charge more because of it, and b. There are a large number of contractors who won't even bid because of 'annoying'.

Actually though, costs run over because the feds have so many redundant requirements, audits, paperwork, and approvals that everything drags on forever, wasting a lot of time. Time that I still get paid for.

the economic incentives for people spending government/taxpyer money are sharply different than for people in the purely private sector.

Rule-of-Thumb is that government projects usually cost twice what the private sector would charge.
No politician cares. Few economists care.

"In the United States, 42% of public infrastructure projects report delays or cost overruns" and the other ones lie.

The implication here is that oversight of public expenditures is a waste of time -- and even makes things worse! -- so obviously it shouldn't be done.

It's the kind of dreck you'd expect from the mediocre Kevin Lewis, which is always dutifully reposted by Tyler.

I don't think it's a lie so much as they assume it will cost more than it should and take longer than it should, but even with that assumption built in, 42% of the time, it's not enough.

Which way is the causation?

Project experiences delays and overruns, therefore more oversight.

Project experiences oversight, therefor more delays and overruns.

You also do not know what the "but for" is. Had there been no oversight once the project experienced an overrun or delay....

I think the causation may be flipped. The larger projects get oversight. The larger more complex projects are also harder to keep on schedule and budget.

The large more complex projects are going to involve the more experienced contractors, all else being equal.

But the study shows that: "Oversight is least obstructive—or even beneficial—when the contractor is experienced,"

Larger projects will affect more private property owners.

Conservatives seem to want to be more like China where there is no private land, thus no basis for private property claims against large infrastructure projects.

Until it's their land which might only see the infrastructure project, eg, Trump objecting to offshore wind farms off his golf courses.

The cost overruns on the Big Dig were primarily due to protecting private property. Eg, 200-300 year old buildings on rubble foundations. A century of settling made the foundations stable as long as the soil was not disturbed. The solution was first freezing the soil, then creating a slurry wall, then low impact excavation to build a structural wall. A few years delay and a lot of extra costs in equipment and labor.

The Big Dig was reversing much of the Chinese style construction methods of the Eisenhower era which bulldozed through immigrant working class neighborhoods destroying their character, and driving out most. When the highway infrastructure was headed into the wealthy areas and business districts, the highway building was stopped, creating highways dumping lots of traffic onto city streets.

Well, maybe. More experienced at being contractors, but not necessarily more experienced at the thing being contracted.

For example, Northrop Grumman is an experienced contractor. They are leading the James Webb Space Telescope project. No-one has ever done anything like James Webb before, so an experienced contractor was chosen. Nonetheless, the whole thing is a disaster.

"Oversight is least obstructive ... or even beneficial - paid with a time-and-materials contract,"

This is a bit counter-intuitive but matches my experiences. You would think that with T&M contracts the costs would be higher because the contractor doesn't have a motive to cut costs.

However, in my experience, it's the exact opposite. When the customer wants a change on a fixed priced job, you have to go through a change order process. This involves a lot of paperwork and time and it may be denied anyway. Furthermore, the people doing the work hate those kind of meetings and paperwork and would much rather be doing what they are good at and what you hired them for. It's quite possible that the change is 1 manweek's worth of work, but the change order process itself adds an additional manweek of overhead time. And you don't even bother with small changes that might take a half a day unless you can roll several of them together.

However, if you have a T&M contract and it's a minor change, you just do the work. This is particularly effective with small changes, where trying to roll them back in later after a change order process makes the task much harder. Furthermore, since you aren't obsessed with trying to minimize time and maximize fixed priced profit, you'll often take a little bit longer to fix small issues that would have otherwise been ignored.

Yes, but contractors abuse T&M contracts because it is in their short term interests to be inefficient. The compromise between the two extremes is to agree to a scope and cost, but include within that fixed price contract a T&M formula for changes so each change doesn't involve an elaborate negotiation.

Agreed, most of our contracts are T&M Not to Exceed contracts.

If the project is in risk of running out of money before core scope is completed, you address that in the regular status updates. The customer either agrees to start the process for a change order, changes the scope or starts limiting changes.

Often changing the scope is just dropping the after project Support hours. Then getting another PO to cover that, so it's nearly identical to a Change Order.

However, since all the work is T&M, both the bureaucracy of the Change Order process and the lack of flexibility of the Fixed Priced bid are minimized.

From there article's lit review:" Without oversight, time-and-materials contractsare much worse than fixed-price contracts: delays and overruns are 7.3% and 8.8% higher.Kim et al. (2007) and Guajardo et al. (2012) explain this finding by arguing that time-and-materials contracts create moral hazard because contractors obtain revenue by working slowlyand spending lavishly. But our results suggest that oversight prevents this form of moral hazard.In the high-oversight regime, the efficiency gap between these two contract types is halved>"

"This is a bit counter-intuitive but matches my experiences. You would think that with T&M contracts the costs would be higher because the contractor doesn't have a motive to cut costs. "

It's a bit counter-intuitive, but the argument is that T&M needs oversight but fixed price with easily measurable outputs does not. It's not an argument that T&M is ultimately cheaper or more efficient. It's that if you're doing T&M you need oversight because of the lack of motive to cut costs; with fixed price there is a motive to cut costs, so oversight is less necessary (and can be harmful if the agency lacks expertise.) Get the incentives correct and detailed oversight is less useful.

It's not by itself an argument that fixed price is more efficient either; there may be some projects that have to be done T&M, usually because the government customer is insisting on a custom one-off solution, or because there's not a good metric to measure the outputs. But I do find that the government tends to ask for custom solutions where in the private sector you'd make do with an off the shelf solution that isn't 100% perfect but is much cheaper. (The government then comes up with GOTS, which are still really custom solutions.)

There is what I call the bureaucratic calculation; a regulatory body has no incentive to approve an improvement, they take all the risk of any downside but no share of upsides. So the only changes that are approved are ones absolutely necessary for the project to deliver, in other words on changes which increase the cost. In a commercial project we normally try to find savings as we go along.

You are doing cut and cover underground viaduct construction under a street of centuries old historic buildings. What costs saving do you imagine you would find?

Surely we could get it down to Western European levels, which are one-fifth to one-tenth the total cost for similar underground tunneling under even older streets with even older buildings.

It's a typically American disease to refuse to learn from the rest of the world, and it plagues progressives, liberals, and others when it comes to infrastructure costs.

Shorter Libertarianism: Let's just give up then.

A more moderate and pragmatic take would be that all human organizations need continuous process improvement, and when that fails, step up one level and improve your process of process improvement. IOW, needs more meta.

Comments for this post are closed