Hacker News Reader

Comment by darkwizard42 | Hacker News Reader

Why aren't contracts to build the projects tightly written to avoid cost ballooning?

When a service is to be provided at a certain cost, it should be provided at certain cost. Some contingencies should be built in (if something gets discovered that was unexpected) but how on earth does every municipal project seem to have so much invested in budget oversight and contractors overbilling or needing overtime to make things happen.

If someone is painting my house, I get a quote for the job and the date it will be completed. I don't find myself paying them more unless its for something outside of what we specified together and they quoted.

Building on government contracts should be solid money but it seems like every time the taxpayer gets fleeced and the end result is barely functional/impactful.

View on HN

It is easy to say, and tough to do.

Big projects have almost unlimited liability. The entities involved can get taken to court for pollution concerns, environmental concerns, get caught up in local politics. And once something is in the courts, the whole timeframe is completely up in the air.

One of the places I lived had Walmart wanting to build a Walmart Supercenter, and convert their current Walmart into a Sams Club. The non Walmart grocery store (which was garbage in terms of quality and service) and a Sams Club competitor tied up Walmart in court over the status of some endangered garden snake.

It was obvious nonsense, but it would have taken so long to untangle legally that Walmart said screw it and moved on.

Firm fixed contracts seem like the way to go, until you try it. What happens is you get a bunch of contractors in a bidding war trying to undercut each other until the actual bid is below cost. That's not sustainable so what happens then? The winner will cut every possible corner to try to make a profit margin and read the contract in the narrowest way possible so they can hit the government with a fat change order. Quality of the product takes a nose dive and final cost still isn't controlled.

On the other hand, time and materials contracts have exactly the problems you would expect.

I honestly don't know how these things should be done so that taxpayers get the most value for their dollars. But I do know it's more complex than just "fixed price contracts."

Maybe have the military operate as a contractor as well and any 'profits' they'd make go towards the general federal budget? At the least it's then a bid by an open party with incentives that should be aligned to public service to compare all the other bids against.

Where's the evidence that contractors will bid below-cost? Even if some do, they'll just go out of business, and the contractors who know how to do basic math will live another day.

Meanwhile you end up with a bridge half-built because the cheapest contractor went bust.

(I don't know if contractors bid below cost.)

Which seems like an improvement, given the current failure mode is projects get cancelled after a billion dollars are spent on them anyways. At least it was partially finished.

No?

How often in software do we see a half finished product or product not designed to scale and we have to start over from the beginning to get it where it needs to go?

You think this doesnt happen in construction projects too? Half building a bridge that you then need to tear down is awful and should not happen. You are just assuming its possible to finish the rest of the bridge later and that critial parts will be available later and that it being in a half finished state for a period of time doesnt cause major issues to the whole thing.

For firm fixed price contracts you can write in requirements for passing independent inspections. And incentives for quality measures. For example I've seen road paving contracts that paid a bonus for achieving a certain level of smoothness. Of course writing a good contract requires deep technical and legal expertise which some local governments simply lack.

Another model is having a government run project, with all the employees direct hires of the relevant government department. Though that has its own downsides.

One model I haven't heard of, is the government as general contractor, hiring and managing the subs directly. I could imagine a hybrid mode like this could work.

Those things have legal remedies, the US legal system has just decided that the investor class should be absolved of any risk or liability. If you intentionally underbid to win a contract, you've committed fraud and are supposed to go to prison to disincentivize the next guy. If you are a contractor and are incapable of giving a good faith estimate of your costs, you're supposed to go out of business. If you cut a bunch of corners to stay under budget and do not render satisfactory performance, you're liable for breach of contract (and are supposed to go out of business).

The first two options have been removed from the toolbox of American government almost entirely.

>If you intentionally underbid to win a contract, you've committed fraud and are supposed to go to prison to disincentivize the next guy.

How would you measure and enforce this?

It's not as hard as it might seem. If you are actually held to the contract and you have underbid, you either have a plan to continue the fraud (using substandard materials or illegal labor to save costs) or you will be forced to take a loss or go out of business. If your investors, employees (or their union), or client(s) are unhappy with that, and there is any evidence of intentionally underbidding to be found (and there generally will be, these are decisions made by teams. There are emails. There are notes) those people will have incentive to find it.

An investigation, and the legal system. Corporations exist because they make a deal with the state. In exchange for limited liability, they're required to respect the corporate form which entails things like meeting minutes, preservation of certain business records and communications, conducting audits, imposing oversight, etc. If they don't abide by these rules, plaintiffs can sue the company and bypass the limitation of liability to recover from individuals in their personal capacity.

This system of incentives and disincentives does work, and the US had a better handle on it in the past.

The contract should be ironclad on acceptable quality, consider including a term for longevity of the product in its acceptable state (bridge can't collapse in 10 years or something)

The US has one of the most well-funded legal industries and we are worried that we can't write a complicated and thorough legal contract to do work? Doesn't really make sense.

If the winner cuts the corners and doesn't deliver, they get sued in US's extremely thorough courts. I also think the bar for applying for these contracts should be higher. Firms with liquidity, track record, etc. get the best jobs.

> Why aren't contracts to build the projects tightly written to avoid cost ballooning?

Because no legitimate construction firm will take on a project where they are saddled with 100% liability for overruns.

Not without very generously padding their estimated costs, that is.

Which will make them lose the initial bid.

A fly-by-night operator may want to take that project, and when things go sour, they'll leave you with a half-built bridge, and an empty shell of a corporation that you won't squeeze a dollar out of.

One alternative would be to have a public works department which does the work itself. It would hire people, plan, write specs, buy or lease equipment, buy materials, do the work.

That, however, would be socialism, so it can't be done, even though it's done on a smaller scale in every US city for issues starting at road repair and going on up.

I think there's a real argument for the efficiency of outsourcing many aspects of public works projects. But I think you make a good point, that doing things in-house is sometimes the right move. Obviously this applies to more than turnpikes :)

One of the tricks Robert Moses had when he was basically building modern NYC was he had on staff a top-notch crew capable of the end-to-end implementation of whatever project proposal he came up with. Designers, engineers, project managers, accountants, even down to the guys digging the holes.

Not sure what happened and why cities stopped having all these capabilities on staff (seems obvious that e.g. NYC will eternally need a huge amount of people like this).

This would be a good alternative, but it has a cold-start problem.

We don't build enough because building is expensive, and building is expensive because we don't build enough.

When we don't build enough, a crown corporation dedicated to building is not going to have enough consistent work to justify its existence.

What if maintaining capability justified the existence

Then it will be axed the moment a 'small government' type gets elected.

This is what I'm also wondering. What is the point of making a contract in the first place if the price quoted is not binding? If the provider needs more material or manpower or whatever than expected, then it should be their fucking problem and not the clients. The possibility of unexpected stuff happening should, of course, also be taken into account in the initial price quote since the provider should be much better at estimating the probability of something happening than the client.

The problem when dealing with government or public works is that the provider simply isn't better at estimating that probability, since they aren't empowered to bulldoze ahead.

The contractor can accurately price in (and insure against) technical risks like the risk of accidentally striking a cable while digging, and they can mitigate this as well (looking at the quality of cable mapping data, using detectors etc.)

What the contractor can't factor in is all the (local) government politics and bureaucracy. Where things fall on the financial year calendar, whether they will have to spend the money before or after April, whether they will have to pause works because a resident legally challenges the works, whether the client has even asked for the right thing, or if it will emerge while planning the work that they have made major errors etc.

When dealing with a well-specified, fixed-scope piece of work, you are right - the provider should eat that. When dealing with an uninformed non-expert customer (i.e. most public sector or government contracts), nobody in their right mind would take on that risk, except some of the big government outsourcing contractors, who would all quote insane prices based on past "actual costs" after factoring in all the nonsense from previous work.

And there will be bias. The contractor will be optimistic about their own abilities and excited about the project. The client will become emotionally attached to the end goal. And everyone will underestimate the overhead of communication. Human factors apply and cannot be engineered away.

Moving away from cost-plus contract funding is something that is needed but is an enormous fight. Every government contractor loves cost-plus. Bid low on cost to get the contract then just keep amending the contract with added cost.

I agree that all government contracts should be fixed bid. It is wholly the contractor's responsibility to determine if they can be profitable on the contract or not. We should not be on the hook for added cost.

It's fine for most infrastructure projects to be firm fixed price (with bonus incentives for early completion or higher quality). But there are other projects for which cost plus is the only viable option. Sometimes there's so much technical risk and uncertainty that no sane contractor would even submit a fixed price bid.

As the article goes into detail, a big factor of costs ballooning is that the municipalities explicitly ask for much more stuff than initially planned - as you say "unless its for something outside of what we specified together and they quoted", which is the exact thing happening in the examples provided by the article.

You ask for A, get quoted $X; but if then after the public input (which, as the article states, can't be avoided or its suggestions refused, because votes and politics) you decide that you actually want A+B+C+D+E+F+G (it's not an exaggeration, it's common for the "add-ons" to require multiple times more work than the initially requested thing) then you're not getting that for $X no matter how you write the contract.

If someone is painting your house and halfway through the process you discuss with your family that it needs to be a different color (redoing the already painted parts) and the kids room also needs to be painted, and oh, they should fix the porch before painting it, that's going to result in cost ballooning far above the quote.

Reply on news.ycombinator.com