Hacker News Reader

Comment by gen220 | Hacker News Reader

I think we're all pretty familiar with the arguments against non-competes. Does anybody have a compelling argument for them, as practiced in 2023 in the US?

While we're at it, does anybody have a compelling justification for allowing non-poach agreements, too?

From my perspective, they both seem like a non-durable benefit to $current_employer at the expense of a compounding opportunity cost to the greater economy (insofar as it hinders an efficient execution of the labor-employer matching process).

Curious if anybody has an interesting argument for why this is OK. The only thing that comes to mind is "trade secrets" arguments, but don't we have separate laws for that?

View on HN

Only thing that comes to mind is a scenario where a sales person goes to another company and takes all their sales leads with them. This happened at a previous employer, at which time they instituted noncompetes for all the salespeople and most of the engineering force.

They forgot to have me and several other very senior engineers sign the noncompete, which was extremely hilarious to me when I went to another company in roughly the same industry and my old boss asked me about it...

Any stories after that conversation?

Not really. My new job was in CA which doesn't honor noncompetes, so they really didn't have a leg to stand on. Mostly just funny.

It matters more where your old company is than your new one. If you work in a state that does allow non-competes and move to a new job in CA which does not, your old employer can come after you in their/your old home state courts where CA law does not matter one bit.

The argument is basically that trade secrets violations are incredibly hard to detect and pursue. You can’t see inside a person’s brain, and (short of industrial espionage) you can’t see inside a competitor either. So a person could come work for you, gain some knowledge of your trade secrets, and then go get a huge offer from a competitor based on that knowledge.

In contrast, it is extremely easy to detect a person’s employer.

I’m not saying this is a valid or correct argument, I’m saying this is one of the arguments.

Another argument (that I know less about) is that there is knowledge that is valuable and proprietary, but does not actually qualify as “trade secrets” under those laws.

"Trade Secrets" have always sounded like a form of protectionism to me, anyway. If you believe software patents shouldn't exist, then you should follow the same logic to lead to the conclusion that trade secrets shouldn't exist either. Both are expressions of business processes that shouldn't otherwise be restricted by blanket clauses or gag orders.

To me (a non-lawyer and non-legal scholar), an NDA seems more appropriately legal than a non-compete agreement, because it at least depends on enforcing a specific action rather than blanket gagging a person's ability to do an entire job where they might not necessarily disclose pertinent information.

It begs the question, where is the line between NDA and non-compete? If I have an NDA and then go to a competitor, my former employer still has a right to enforce that NDA, but how would they ever determine (prove) I violated it? I guess this difficulty is where the justification for non-compete comes in, but I think the onus should be on the employer. It's not fair to the employee if the employer can just blanket forbid them from working somewhere because they might disclose some information to their new employer in a violation of an NDA which the old employer is either too lazy or incapable to detect.

With a non-disparagement agreement, it's often pretty easy to see if someone publicly disparaged them. A public tweet about how your last job was a hellhole or your boss was an idiot coupled with a Linkedin profile showing where your job was, or just a series of posts on one platform where one indicates your employer and another is disparaging.

Edit: Oops, I misread NDA as Non-Disparagement instead of Non-Disclosure.

A reasonable counterargument here is that if the improvements from misusing trade secrets are that hard to detect, then they were perhaps not sufficient to be worth suing over. Especially given that the discovery phase of a lawsuit does indeed let you see inside a competitor.

This is a solid summary. Another, related argument I've heard is that it incentivizes R&D + upskilling. Much of the value is "we now have people who can do X", not X itself. Without noncompetes, it's a better strategy to try to poach your competitor's R&D employees, rather than invest in your own, which very quickly leads to less research + upskilling.

That's a somewhat reasonable concern. But if it really matters to an employer...pay for it. For example, offer a security that vests after 5 years or whatever. Saying 'it's already built into the compensation' doesn't cut it, because then it starts becoming the norm for everyone.

The issue is that it's very difficult to tell trade secrets from expertise in certain fields. In a field that is sufficiently niche, like non flagship chip production - a lot of architectural ideas are well known across the industry are but the implementation details that enable them are trade secrets. You can't "un see" these details and IP law is loose enough that you could easily reimplement those features or products for a competitor without violating IP laws because you've seen the idea before and can come up with infinite implementations.

IANAL but I had (kinda unconsciously) assumed that what can be protected by trade secrets was broader than what patents protect (patents merely implementations - which might be broad.) So I'd like to know more; though I wouldn't be shocked (now that I think about it) if you're quite right.

In any case, even if the employee refuses to work on a given task, that's information. Having to say: "Yeah, I think you want somebody else to design that particular part" (implying "because they'll be freer to innovate and free to choose the optimal solution") is a big fat clue that could attract a lot of resources to the design of that part.

As far as I understand it, the only time trade secrets are protected is when they are forced to be revealed in court. If for example Zildjan cymbals is in a lawsuit, and discovery requires that they divulge their trade secret for making cymbals, then the court will issue an NDA for everyone that sees it, and seal the records. However, if a former Zildjan employee writes a book with the details, then there is no protection. Which is why they only limit the secret to family members. Fun fact, Sabian uses the same secret technique, because it is the same family.

No, that's one aspect of protection. Penalties apply otherwise see the Waymo/Uber case.

Apparently the US statutes of relevance are:

The Economic Espionage Act of 1996

The Defend Trade Secrets Act of 2016

Re publication I think you're referring to a possible US constitutional issue: https://www.dmlp.org/legal-guide/publishing-trade-secrets

Difficult to defend "as practiced". For example trying to enforce a non-compete for sandwiches is ridiculous.

There are obvious scenarios in competitive industries where losing key people to competitors creates risks for everyone involved (not just the losing employer). The problem is when employers expect the individual employee to absorb the costs to mitigate that risk. Or where the non-competes are applied overbroadly. Rather than ban non-competes, it would be better to require compensation for the term of the non-compete. This prevents frivolous applications of non-competes, which is the actual problem.

So, here are a couple of arguments for:

1. I have a relative who is an MD. He was recruited cross-country at great expense. (Average cost to recruit an MD can be about $250K). So, if his comp was $200K/year and it cost $250K to recruit, a neighboring practice could monitor for new incoming docs, and make an offer of $220K/year in salary to the newly hired doc. If that happened, it would be in the best interest of the doc to switch jobs, but the original practice would be out $250K in recruitment costs.

2. In the case of an acqui-hire, the team is often the special sauce. You embed a bit of non-compete in the form of stock options that vest on a particular schedule, but it may be tricky to structure the deal in an attractive way without a non-compete and non-poach agreement.

3. Trade secrets are often hard to cover in NDA's. Your trade secrets may become embedded in the employee's mind in a manner that they cannot extricate. So, if your employee receives training that includes your trade secrets, those trade secrets will be implicitly used at the next job.

So, I think the argument basically boils down to there being a vast upfront cost to the employer for getting a new employee. If the employee switches to another company, the value of that upfront cost transfers to the new company with no compensation to the old company. It seems a new, more pernicious workaround to non-competes is where employers are charging their employees for training if they leave early. That seems even more hostile than a non-compete.

(As a side note, I think non-competes can be quite damaging. In the case of the MD relative, he was fired, essentially without cause, and his non-compete forced him to be unemployed for a year before he was finally able to convince the former employer to waive the non-compete. So, there should be very hard parameters around non-competes. One thing I think should be mandatory is a written buyout amount for any non-compete that has some basis in reality. For example, if my MD relative was recruited at a cost of $250K with a 2-year non-compete, then he could buy himself out at $250K, minus about $20K for each month of service he completed. Obviously, I haven't fleshed this idea out all the way.)

I think those are all reasons a company might want it. But asking the government to force people to not work requires reasons why it's good for society, not just the company.

In the first case, you're talking about a company that wants to pay below-market salaries. Why should that be the employee's problem?

In the second, there's a case for carrots to make the acquired team stay, like the stock options you mention. But from a societal perspective, why should the company be able to use the courts as a stick if the carrots turn out to be insufficient?

In the third, I again get why companies want to treat employees like property. But I don't see any societal argument for that other than "rich company wants things".

> But asking the government to force people to not work requires reasons why it's good for society, not just the company.

Nobody is forced not to work. That's pure hyperbole.

Sorry, I thought it was obvious from context that I was talking about forcing people not to work at particular jobs. Which is what a noncompete is. You say you won't do X, and if you do X they ask the government to force you to stop doing X. If you refuse to comply long enough, men with guns will come and drag you away (contempt of court, and if you keep going, resisting arrest). So yeah, force. "Come and see the violence inherent in the system" and all that.

1. If the doc is worth $220K, why couldn't the practice that hired him match the offer to save themselves the $250K they spent hiring him? This feels like exactly the kind of wage suppression that the FTC is arguing against.

2. Workers are not serfs to be bought and sold. If the acquiring company wants the talent, then they should structure the deal in a way that makes the talent want to stay, not use legal handcuffs to force them to.

3. I can see this as an argument for noncompetes being legal in some very limited cases, but most jobs don't need this. Others have suggested requiring garden leave in lieu of a noncompete, and in the few roles where this applies I suspect that would work out fine.

You’re focusing on the specific numbers in the example, but missing the point. Imagine the salary difference is larger.

Basically, you have one company that has already paid $250K recruiting the doc and another company that paid close to $0. So if it comes to a bidding war over salary, the former company will always be at a financial disadvantage. And budgets always have limits.

No, I'm not focusing on numbers. In fact, if I focus on the numbers they seem suspect: $250k is the cost to recruit, not the relocation costs[0]. A good chunk of that (upwards of $220K) is going to occur even with a local recruitment, which makes the disadvantage of the initial recruiter much smaller.

Further, as others have mentioned, there are other ways to contractually recoup relocation costs without a non-compete. A "you must pay back your relocation costs if you leave within a year" clause is far more justifiable than a "you can't work as a physician within 30 miles for at least 1 year if you ever leave us ever" clause.

(The same type of payback clause could apply if you did something drastic like pay off their entire student loans in one lump sum, though I am assuming that in most cases practices don't do that.)

All that said, my point had nothing to do with the numbers and everything to do with the principles: non-compete clauses are an extremely blunt instrument and are inappropriate in most cases. Firms should be required to come up with more limited contracts that accomplish their stated goals and nothing more, rather than throwing in something that is so damaging to the worker because it's easier for their lawyers.

[0] https://healthrecruitlink.com/blog/what-will-it-cost-to-recr...

In your first case, I'm pretty sure I've seen contracts that require the repayment of e.g. relocation costs if the employee leaves within x time period.

I was recruited with paid relocation at my current job, and I would have had to repay relocation costs if I left within a year.

Often a signing bonus also needs to be repaid (pro-rated).

> In the case of the MD relative, he was fired, essentially without cause, and his non-compete forced him to be unemployed for a year before he was finally able to convince the former employer to waive the non-compete.

It seems like an easy and obvious solution that the non-compete is automatically null and void if the employee is let go for any reason.

Unfortunately, that also has easy workarounds, like making the employee's life hell until they quit. So you'd also need to solve for constructive dismissal: https://en.wikipedia.org/wiki/Constructive_dismissal

1. I think the end-game here is salary-parity, no? If Practice A learns that Practice B is taking their employees for a marginal salary difference, that forces Practice A to improve their compensation package.

But I don't think I fully understand this example. Why does it cost $250k to hire an MD? Why does Practice B not need to pay this cost, is it because they can skip the vetting process since A has already performed it?

2. If it's an acui-hire, the only way it can be successful (in real terms, not just fake "retention" terms) is if the acquired team is consenting to the acquisition and partnership. If the acquirer can't create a deal (involving stock options, work lifestyle guarantees, whatever) that's successful in the eyes of their prospective future employees, non-competes and non-poaches simply delay the inevitable. This is to the detriment of "we the people", because we want good people to be productively working on important things.

In my not incredibly informed opinion, NCs for acquihires smell lazy and inefficient, because they optimize for the wrong metrics (employee retention vs value creation + satisfaction). If there's uncertainty about the long-term success of the merger, it can be factored into the acquisition price.

To me, this implies that outlawing NCs would lead to fewer acquihires, on the margin. This seems like it would probably be a good thing.

3. I think there's some "basement" of trade secrets that we just need to accept are going to spread around. As a company, you have to understand that this bottom 20% of ideas are going to osmosis their way out with every departing employee, and there's nothing you can do about it other than work to retain employees and innovate new ideas.

Similarly to the acquihire issue, NCs simply delay the inevitable here, and don't seem to provide much benefit "we the people" (or protection to "we the entrepreneurs").

1. There's two costs being confused here. The impact of a doctor burning out or otherwise retiring and no longer using their qualifications, requiring the training of a junior doctor with required specialisations at a potential cost of USD$250k-$1M[1] and many years of waiting. And the much less significant impact of a doctor using their qualifications to find an equivalent job with a different employer. In the later example, it's a fairly standard professional recruitment process and cost.

2. Stock options are a risky gamble for employees and employers alike. Neither party can rely upon stock options too heavily as a retention tool because no one knows what the stock options would be worth 3 years in the future. If it's really essential to keep employees on for numerous years then guaranteed salary increases would be a better way to ensure employees are adequately compensated for the detriment to their career of staying in the same organisation doing the same work for a long period of time. And of course, proactively ensuring employee salaries are _always_ in the top decile of industry/specialisation salaries is needed too because it is not uncommon for some labour rates to move +10% in a single month. Many employers with a high NIPE/PPE[2] could easily pay higher salaries if they desired to keep employees for longer periods of time.

3. Aren't patents are meant to protect such R&D investments? Employers benefit from hiring from each other creating a mixture of technical knowledge and culture that is gained from employees having worked in different roles and projects elsewhere. Thus I struggle to comprehend why employee movement would be viewed as a net negative overall that justifies non-compete agreements.

[1] https://www.ama-assn.org/practice-management/physician-healt...

[2] https://tipalti.com/profit-per-employee/

> So, if your employee receives training that includes your trade secrets, those trade secrets will be implicitly used at the next job.

That's not a compelling argument for NCCs. It's an argument stifling competition, which should not be what NCCs are used for.

Every company takes a risk hiring someone that may leave with field expertise specifics to the company. Them using that expertise elsewhere makes the market healthier and more robust because it increases competition.

It's conceivable that NCCs are more anti-competitive than protective of the company's trade secrets, at least that's how it appears they're being effectively used nowadays.

In theory, it lowers the risk of hiring in certain situations. So, if the alternative is not hiring, that's beneficial.

For a lot of companies, the "sauce" is a personal relationship that isn't trade secret. If you hire a salesman, and give them access to Mr. Big, you create an incentive for the employee to trade the relationship for your competitor's money.

Penalizing the change in employer is not the only way to deal with it, of course. Without a non-compete, you might have to buy/keep the salesman's loyalty. Which may or may not be an economic deadweight.

That makes sense as to why an employer would want a non-compete.

Is there some way that this translates into why "we the people" would want to allow companies to enforce non-competes?

From my perspective, if a company's sauce is personal relationships, the non-compete is not particularly effective in engendering loyalty and extracting the greatest value from that relationship.

A disaffected salesperson is not going to be enthusiastic about leveraging their relationships for an employer they're unhappy with -- whether they're under a non-compete, or not. This means they'll extract the minimal value to keep their salary, but not the maximal value out of the potential they have access to.

In fact, a non-compete prevents that salesperson from bringing that relationship to a different company, where that relationship might produce greater dividends for "we the people" (as measured by the new company's ability to attract salespeople and engender their loyalty).

1. Shifting profits from the entrepreneur (with skin in the game of providing consumer wants) to the middleman can be deadweight. You'd want some evidence that rewarding disloyal middlemen results in we the consumer surplus, not just a flourishing middle-man economy.

There can be price-discovery benefits to middlemen, like market makers at the NYSE. But the financial markets usually involve fungible goods. Consumers seem to benefit from finding ways around the middlemen, though.

2. We the people have egalitarian ideals about jobs and opportunities. The old-fashioned way to 'lock in ' people to loyalty is to hire family or from the same social strata. Historically, those benefits have been distributed unfairly.

So the non-compete provides some mechanism for the "haves" to spread around opportunities to the "have-not-so-many-reasons-to-be-loyal," understanding that the law will enforce the agreement to be not-disloyal.

Good points! Let me try to dissect them, let me know what you think.

1. I agree with the themes of not paying disloyal middlemen, but I'm not sure if it applies to the example.

> You'd want some evidence that rewarding disloyal middlemen results in we the consumer surplus,

If entrepreneur Y is willing to pay 2x the salary of entrepreneur X for the same employee, presumably its because Y thinks they can use the salesperson's relationships (at least) twice as productively.

In that case, Y has a product to sell that is twice as valuable as X's. As a consumer, don't we want companies we engage with to use the most valuable products?

I'm not suggesting that we pay unaligned people more (obviously, this is bad). Rather, I'm saying we shouldn't erect barriers to prevent disgruntled employees from leaving, because we're paying the opportunity cost of them not being a more productive employee at a competitor.

2. I'm not sure if this holds up in 2023, where talent networks are global and talent discovery is so cheap.

I think this argument makes sense if the skills of people who are signing non-competes were not so rare and in-demand. If I exclusively depended on family ties to run a company in 2023, I'd expect my competition to run me over.

1. > "presumably it's because Y thinks they can use the salesperson's relationships (at least) twice as productively."

I'm not sure that's true. Y thinks Y having the profits from S's relationships are worth the marginal cost of paying 2 x $X (the salary). Presumably, he could offer that to S's Boss -- but the angle is that S will usually sell out for less. S has less skin in the game.

As you note, it's hard to tell how much S is actually benefitting consumers, versus how much they're trying to extract rent once they've got Boss over a barrel. And it seems the best time to solve that is with a voluntary agreement between consenting adults before Salesman has Boss over a barrel.

If Boss runs the risk of being gutted by S, the usual response is to raise prices, or conclude that it's just too expensive to make this good. So probably consumer prices rise under this regime. Now, maybe if you're the party of middle management, that's a political good!

2. I don't know enough about that to BS beyond what I've BS'd already. But it strikes me as mostly a toss up -- sometimes it's good, sometimes it's bad. And, again, I think the libertarians are mostly right about respecting deals between consenting adults, even when there's a wealth or social gap. The parties will have a much better idea about when they're good and bad. I'm not sure why you'd outlaw them.

I guess the best argument I have against myself is that California has already done something similar for awhile, and it hasn't crippled their market for talent. So whatever effect I might be worried about, it's de minimus, and it seems to be a morale booster for middle-management types.

I think NC clauses are pretty bad in general. But, I can't shake a weird feeling that crops up from time to time. Anyone out there operating without one of these things over their heads that feels like maybe they should?

I started out at a tiny company that didn't think their app idea would make it as far as it did. It mostly has. Now I am looking at the terms of my employment and finding myself approximately as encumbered as a McDonalds employee on their first day.

Some of the things I know could be leveraged to turn our biggest competitor into an unstoppable explosion in very little time. I'd never actually do this (there are always higher-order effects w/ betrayal), but I do think I could use this concern as part of my salary re-negotiation this year.

Equity and other forms of compensation serve as a very powerful carrot to abate this temptation. Carrots don't always work though. You usually need some kind of stick at some level.

It depends entirely on your target goal.

If your target goal is something like "maximization of worker pay",for example (really, i'm not suggesting that's what you want, it's just a convenient example), then no, there will never be a compelling argument for non-competes, because they always stand in the way of that particular goal.

This is because costs by employers can always be expressed without non-competes somehow to within some reason, and not get in the way of that goal (but may harm others).

For example - there are fields where large amounts of on-the-job training is paid for by employers. Certainly they don't get nothing out of it in the end, but they have upfront costs too. It is more frontloaded cost wise, such that if the employee leaves immediately after training, the employer would lose money.

The literal training cost can simply be expressed as an actual penalty for leaving rather than a non-compete. But the loss of time and need to train someone else (IE the time during which you would have made more money on the person who just left) is harder to value as a penalty (since it depends on what your profit would have been, which is probably variable, and lots of other things). So that is often expressed as a non-compete. But that would still stand in the way of "worker pay maximization", so that particular difficulty is not compelling for that goal (IE you could just assign a value to it and deal)

Similarly, there are fields where there are not enough people, and great upfront cost may be spend in acquiring people (doctors in rural areas, etc). Normally they would non-compete them and give a significant hiring bonus. You can also express this as a penalty, but it still puts you out a doctor for the time it takes to recruit someone else, etc, which is bad for the community. But not for worker pay!

If you go for worker pay maximization, them taking a 250k bonus to go work in that area, then leaving a year later (or whatever), maximizes their pay. It also likely forces the community to pay more to the next person to have that not happen!

So again, if your target goal is maximization of worker pay, no, I think there is no compelling arguments for non-competes.

If you switch your goal to say something like "maximization of production of useful things", maybe you could get to more compelling argument, because constant job hopping by everyone, while maximizing worker pay, probably interferes with maximizing the production of useful things.

Or if you introduce some balance of goals, maybe.

I think my goal would be "maximization of [responsible] production of useful things".

Constant job-hopping requires an employer who wants to hire you. If you're a "constant-job-hopper", you'll find few people who want to hire you.

NCs are not only a meaningful barrier to fair compensation, they also seem to be a barrier to the allocation of labor to more productive businesses that, by virtue of their productivity, can afford to pay higher compensation.

The main time I see them make sense is with very skill-based, client-based practices like accounting. You're likely giving this hire a lot of clients, and you don't want him leaving in two years and taking half the clients with them.

Reply on news.ycombinator.com