Good article, but Matt is a bit off the mark on the reasoning for non-competes. I think he's conflating them with Non-disclosure agreements (NDAs). I work in engineering / tech, and non-competes just aren't that common in those generally considered innovative industries. Where they are EXTREMELY common is in my wife's industry: veterenary medicine (and medicine in general) where you don't have trade secrets so much as a client base. Their main purpose, in addition to eliminating employee leverage like Matt says, is to eliminate any risk of those professionals moving and taking clients with them. In this way, it's actually in favor of freedom to contract (between the provider and their clients) to ban these awful clauses. I actually was in a social conversation with a lawyer about them, and apparently non-competes are against the ethics of the legal profession to allow clients to choose their jurists or something. Same conclusion though: they are awful and should absolutely be banned.
A friend of mine in the corporate public relations space is contending with recruiting offers from rival firms in part on the basis of him bringing not trade secrets or training, but his established client base (largely biotech firms) and teams with him. He is subject to a non-compete (although, intriugingly, the interested firm has offered to *fully fund the necessary court battle* to beat the clause, largely on the basis that judges are typically very loth to actually enforce such agreements
The fact that there is a larger gap between the productivity growth of "frontier" firms in IT (Heavily concentrated in California) and a smaller gap for other businesses doesn't seem to support the idea that banning non-competes would help with the diffusion of talent and expertise from innovative firms to the rest of the industry.
In investment banking, firms have have what is called "garden leave." If you quit to go to a competitor, your old firm pays you for 90 days to sit in your garden (the term came from Britain; Americans might have called it "back yard leave"). During garden leave, time-sensitive information becomes stale and worthless, but other knowledge does not. After garden leave, you join your new firm without having had an interruption in pay/benefits.
It is a compromise and it seems to be working pretty well. It is widely accepted and one hears few complaints.
Interesting. I knew my friends who work for the big banks jump around a lot which I always thought odd given how much the firms hate each other and compete tooth and nail.
I mean, I think some noncompetes should not be allowed but because your behavior is controlled in a specific way (you have to "work") while you're not getting paid. And we do have laws against that....
Compensated noncompetes are different.
And protection of trade secrets is sometimes important and that should be possible for corporations to do when it's real and necessary.
I don't think the framing of "difficult" versus "expensive" is useful because a) in economy and law, as numbers of people scale we really only think in money and b) sure that Walton could generate a lot of innovation with $17B. You hire a bunch of smart people with the first billion and then let them do the rest. Yes, if you're going to micromanage anything that will mess your returns up but that's true even at Intel in the 70's / Google / Facebook / Apple.
Innovation is something we want to promote because it creates productivity growth. It's kind of by definition the only thing that does that! We want people to be free to innovate as much as they can. That's really the reason why we want (unpaid) non-competes to be disallowed.
> The main idea of the non-compete agreement is that employers want to stop people from walking off the job and taking trade secrets to rival companies. If companies weren’t able to secure those protections, they’d need to pay lower salaries, and we’d all be worse off.
I've worked in tech for three decades, and as far as I recall, I have never heard anyone make an argument about lower salaries. I don't really even understand the argument, but regardless, everyone is pretty well aware that the reason companies put in non-competes is that they don't like the idea of you walking across the street to work for a competitor.
Companies also do legitimately worry about trade secrets, although there are already laws against intellectual property theft, as evidenced by various law suits against high-level employees who go to work for rival firms. I think the broader concern, though, is less about theft of extremely valuable IP and more about the transfer of valuable skills and knowledge, as in the NYT example.
The major tech companies in silicon valley had a cartel agreement to avoid getting into bidding wars with each other over engineers. They may not talk about it but management absolutely cares about keeping labor costs down.
I know that, but that doesn't seem to be what Matt is saying at all. The paraphrased argument is that if companies didn't have non-competes in place, they would be forced to pay *lower* salaries, which would make us all worse off.
But you are making me think that this is just a typo and it should say *higher* salaries.
That is at least directionally correct, but it still isn't the "main idea." The main idea is that employers want to stop people from walking off the job and taking trade secrets to rival companies because employers want to stop people from walking off the job and taking trade secrets to rival companies. I've never heard anyone attempt to argue for a broader societal benefit. Maybe lobbyists are saying this? It's not an argument I see getting a lot of traction in other contexts.
I don't understand the NYT example here. Is MY saying its a good thing that the Times gobbles up every remotely successful media company, and faces essentially no real competition in "serious" national news?
The NYT doesn't really face any serious competition for "being the NYT", at least for now.
But could any thinking person believe they don't face competition for attention and subscription dollars? Just speaking for myself, I subscribe to (off the top of my head), the NYT, WaPo, LAT, the SF Chronicle, the San Jose Mercury, the Sacramento Bee, the Guardian, the Atlantic, Slate, Vox, TNR, The American Prospect, and several Substack newsletters.
I don't subscribe to Bloomberg or the Wall Street Journal, because while I think they have some good content (and in particular I find it kind of annoying to miss out on Cathy O'Neil and Noah Smith in their opinion pages), they have set their price point too high, compared to the other things I could devote my reading minutes to. Competition!
That Jimmy John's non-compete was even worse than I thought. For example, it not only prevented you from taking a higher wage at a Jersey Mike's, it forbid you from taking a job at *another Jimmy John's.* (See section 3: https://big.assets.huffingtonpost.com/FACExhibitA.pdf). Protecting "trade secrets"? Yeah, right.
The American Booksellers Association made a post saying that they had put a book on their lists that was anti-trans and that this was, among many other things, "violence."
Matt, along with many other people, snarked about the idea of including a book on a book list being "violence."
It's a joke about the Streisand Effect. The outcry against Irreversible Damage will probably sell more copies of the book than any positive marketing scheme the author and publisher could have come up with. Matt is jokingly trying to get the same publicity for his book.
Most non-competes I've seen are at the consulting firms and in transactions. Consulting firms have them to prevent star performers from leaving with their clients. In transactions, the buyer doesn't want the old owner to set up a competing business across the street. Matt's article sort of seems to ignore the big issues with disallowing non-competes.
I know a person who did the latter- sell her firm to a publicly traded parent company, wait out the non-compete for a few years, and then recreate the firm & recruit all of her star employees out of the one she sold to work for her again. She's basically running the same company she was 5 years ago, she's just now $25 million richer and took a long vacation in the middle. Good stuff
Some of the proposals I have seen have exceptions for business sale cases that block someone from effectively recreating the business they just sold to you. I think that’s a pretty clear exceptions that would cover some of the most legitimate uses of non-competes without being too broad—if you aren’t an owner, you can’t be bound by it.
If it were up to me the principle should be something like “ongoing substantial consideration.” So if you were selling your plumbing supply business for $5 million and they wanted you not to compete they’d have to pay you 0.5% per month. Obviously in that case the upfront price would be lower.
These aren't "non-competes" -- they're sales contracts where you exchange something of value for money and then include conditions where you promise not to kill the goose that lays the golden egg. If someone wants to buy your company you can negotiate whatever terms you want, or say no.
A non-compete in the sense of Matt's article is a nonnegotiable condition of employment that *outlasts* the period of your employment with whoever you're working for. You've stopped working for them (or been fired / laid off / let go) and you're not paid AND not allowed to use your own industry & knowledge to earn a living. That's bad and substantially different.
For consulting firms that’s what garden leave is for. For business owners who sell that’s what a 18 month “consulting” contract for $xx,xxx a month is for.
This describes the non-competes of my youth. Back in the day the owners of one set of burger parlors didn’t worry about people leaving their outfits for a like establishment down the street.
Thoughtful post, Matt. One thing I wanted to point out re California: even though there is a policy generally making noncompetes unenforceable, there is an exception where it is necessary to protect trade secrets. The burden is on the employer, and it’s not super easy to meet. Your general points stand, but the reason why those tech companies can succeed could be because those employees who actually could harm the business by stealing trade secrets and giving them to a competitor can be restrained by a noncompete. There is of course a separate trade secret misappropriation law, but once the cat is out of the bag, the damage is done and sometimes irreparable. So there is probably a good case for a narrow noncompete protection, so long as the employer has the burden and can’t just baldly assert that they have trade secrets and use noncompetes indiscriminately (as happens in many other states).
I am confused by this. In California trade secret protection is exclusively by statute, not by contract. Yes, some employer contracts contain trade secret protections, but they are superseded by the statute. And that’s more efficient, because the statute applies to everyone, whether or not they acknowledge or agree that they have trade secrets. There is no additional protection afforded by contract - the “cat out of the bag” issue you raise is equally addressed by both. Source: I am a trade secret litigator in California.
The additional protection is that if an employee with knowledge of trade secrets is restrained from employment by a competitor, they are less likely to misappropriate trade secrets.
On a related note to non-competes, what about "non-poaching" agreements? Both between companies and employer-employee.
Various companies in the same industry here in town have a gentleman's non-poaching agreement. They'll absolutely hire you if you apply to them, but they won't actively seek you out to recruit you. Now that's not an enforceable thing - that's just a kindly agreement.
For the employer->employee one
At one point for a contract job I signed a non-poaching agreement that I wouldn't, after I was done with the contract, immediately(had to wait a year) try to recruit someone I had just worked with. I'm sympathetic (and I signed it) because I was getting introduced to a bunch of people on a short-term basis and it would be connections I exploit to pull them away later. Note that other employees were free to apply to the same place I went, I just couldn't actively recruit them (unsure how enforceable that was in court - in any event I took it seriously ethically because I signed it)
I've never been offered or signed a non-compete. (I work in software, and not in California)
Google, Apple, Adobe and a few others had a genuinely massive fine levied against them in the late-00s for just that sort of "gentlemen's agreement". (It was large enough that my share of the settlement as an entirely undistinguished mid-level engineer at Google was in the low four figures.) I think the big players in SV have been a lot less prone to such shenanigans since.
This is quite an interesting column, particularly for the reference on the uneven diffusion of "best practices" across firms. It corresponds to the Autor et al stuff on "superstar firms" getting all the gains. Nick Bloom (Stanford) has written on how this is even a within-firm phenomenon. On non-competes and know-how diffusion & monopsony: interesting points here, but there is an extensive empirical literature using diff-in-diff/regression discontinuity methodology to compare outcomes across states that adopted cognate regimes. Matt should do a follow-up testing his ideas against those studies.
Good article, but Matt is a bit off the mark on the reasoning for non-competes. I think he's conflating them with Non-disclosure agreements (NDAs). I work in engineering / tech, and non-competes just aren't that common in those generally considered innovative industries. Where they are EXTREMELY common is in my wife's industry: veterenary medicine (and medicine in general) where you don't have trade secrets so much as a client base. Their main purpose, in addition to eliminating employee leverage like Matt says, is to eliminate any risk of those professionals moving and taking clients with them. In this way, it's actually in favor of freedom to contract (between the provider and their clients) to ban these awful clauses. I actually was in a social conversation with a lawyer about them, and apparently non-competes are against the ethics of the legal profession to allow clients to choose their jurists or something. Same conclusion though: they are awful and should absolutely be banned.
A friend of mine in the corporate public relations space is contending with recruiting offers from rival firms in part on the basis of him bringing not trade secrets or training, but his established client base (largely biotech firms) and teams with him. He is subject to a non-compete (although, intriugingly, the interested firm has offered to *fully fund the necessary court battle* to beat the clause, largely on the basis that judges are typically very loth to actually enforce such agreements
The fact that there is a larger gap between the productivity growth of "frontier" firms in IT (Heavily concentrated in California) and a smaller gap for other businesses doesn't seem to support the idea that banning non-competes would help with the diffusion of talent and expertise from innovative firms to the rest of the industry.
In investment banking, firms have have what is called "garden leave." If you quit to go to a competitor, your old firm pays you for 90 days to sit in your garden (the term came from Britain; Americans might have called it "back yard leave"). During garden leave, time-sensitive information becomes stale and worthless, but other knowledge does not. After garden leave, you join your new firm without having had an interruption in pay/benefits.
It is a compromise and it seems to be working pretty well. It is widely accepted and one hears few complaints.
Interesting. I knew my friends who work for the big banks jump around a lot which I always thought odd given how much the firms hate each other and compete tooth and nail.
I mean, I think some noncompetes should not be allowed but because your behavior is controlled in a specific way (you have to "work") while you're not getting paid. And we do have laws against that....
Compensated noncompetes are different.
And protection of trade secrets is sometimes important and that should be possible for corporations to do when it's real and necessary.
I don't think the framing of "difficult" versus "expensive" is useful because a) in economy and law, as numbers of people scale we really only think in money and b) sure that Walton could generate a lot of innovation with $17B. You hire a bunch of smart people with the first billion and then let them do the rest. Yes, if you're going to micromanage anything that will mess your returns up but that's true even at Intel in the 70's / Google / Facebook / Apple.
Innovation is something we want to promote because it creates productivity growth. It's kind of by definition the only thing that does that! We want people to be free to innovate as much as they can. That's really the reason why we want (unpaid) non-competes to be disallowed.
I can tell anyone who says the sky is falling non competitive agreements are void in CA and life goes on
> The main idea of the non-compete agreement is that employers want to stop people from walking off the job and taking trade secrets to rival companies. If companies weren’t able to secure those protections, they’d need to pay lower salaries, and we’d all be worse off.
I've worked in tech for three decades, and as far as I recall, I have never heard anyone make an argument about lower salaries. I don't really even understand the argument, but regardless, everyone is pretty well aware that the reason companies put in non-competes is that they don't like the idea of you walking across the street to work for a competitor.
Companies also do legitimately worry about trade secrets, although there are already laws against intellectual property theft, as evidenced by various law suits against high-level employees who go to work for rival firms. I think the broader concern, though, is less about theft of extremely valuable IP and more about the transfer of valuable skills and knowledge, as in the NYT example.
The major tech companies in silicon valley had a cartel agreement to avoid getting into bidding wars with each other over engineers. They may not talk about it but management absolutely cares about keeping labor costs down.
https://en.wikipedia.org/wiki/High-Tech_Employee_Antitrust_Litigation
I know that, but that doesn't seem to be what Matt is saying at all. The paraphrased argument is that if companies didn't have non-competes in place, they would be forced to pay *lower* salaries, which would make us all worse off.
But you are making me think that this is just a typo and it should say *higher* salaries.
That is at least directionally correct, but it still isn't the "main idea." The main idea is that employers want to stop people from walking off the job and taking trade secrets to rival companies because employers want to stop people from walking off the job and taking trade secrets to rival companies. I've never heard anyone attempt to argue for a broader societal benefit. Maybe lobbyists are saying this? It's not an argument I see getting a lot of traction in other contexts.
this isn't even vaguely responsive.
I don't understand the NYT example here. Is MY saying its a good thing that the Times gobbles up every remotely successful media company, and faces essentially no real competition in "serious" national news?
The NYT doesn't really face any serious competition for "being the NYT", at least for now.
But could any thinking person believe they don't face competition for attention and subscription dollars? Just speaking for myself, I subscribe to (off the top of my head), the NYT, WaPo, LAT, the SF Chronicle, the San Jose Mercury, the Sacramento Bee, the Guardian, the Atlantic, Slate, Vox, TNR, The American Prospect, and several Substack newsletters.
I don't subscribe to Bloomberg or the Wall Street Journal, because while I think they have some good content (and in particular I find it kind of annoying to miss out on Cathy O'Neil and Noah Smith in their opinion pages), they have set their price point too high, compared to the other things I could devote my reading minutes to. Competition!
That Jimmy John's non-compete was even worse than I thought. For example, it not only prevented you from taking a higher wage at a Jersey Mike's, it forbid you from taking a job at *another Jimmy John's.* (See section 3: https://big.assets.huffingtonpost.com/FACExhibitA.pdf). Protecting "trade secrets"? Yeah, right.
Oh, and when the Illinois attorney general finally brought an end to the\is odious practice, Jimmy John's was required to pay a fine of . . . $100,000. (https://www.polsinelliatwork.com/blog/2016/12/14/that-was-fast-jimmy-johns-nixes-non-competes).
I'm not a Bernie supporter, but reality keeps asking me why the hell not.
What's going on with One Billion Americans being called violent? Is that a real thing? I'm confused
The American Booksellers Association made a post saying that they had put a book on their lists that was anti-trans and that this was, among many other things, "violence."
Matt, along with many other people, snarked about the idea of including a book on a book list being "violence."
I assume this is some kind of fallout from that.
It's a joke about the Streisand Effect. The outcry against Irreversible Damage will probably sell more copies of the book than any positive marketing scheme the author and publisher could have come up with. Matt is jokingly trying to get the same publicity for his book.
Oops, I misunderstood. My bad!
Read the current line up of articles at Vox and you can sure tell who the NYT didn't poach and why.
Most non-competes I've seen are at the consulting firms and in transactions. Consulting firms have them to prevent star performers from leaving with their clients. In transactions, the buyer doesn't want the old owner to set up a competing business across the street. Matt's article sort of seems to ignore the big issues with disallowing non-competes.
I know a person who did the latter- sell her firm to a publicly traded parent company, wait out the non-compete for a few years, and then recreate the firm & recruit all of her star employees out of the one she sold to work for her again. She's basically running the same company she was 5 years ago, she's just now $25 million richer and took a long vacation in the middle. Good stuff
I mean, that's fine. She complied with the noncompete, competed successfully for key employees and then beat her old business at its old game.
Presumably she served customers better somewhere along the line otherwise why would they switch?
Some of the proposals I have seen have exceptions for business sale cases that block someone from effectively recreating the business they just sold to you. I think that’s a pretty clear exceptions that would cover some of the most legitimate uses of non-competes without being too broad—if you aren’t an owner, you can’t be bound by it.
If it were up to me the principle should be something like “ongoing substantial consideration.” So if you were selling your plumbing supply business for $5 million and they wanted you not to compete they’d have to pay you 0.5% per month. Obviously in that case the upfront price would be lower.
These aren't "non-competes" -- they're sales contracts where you exchange something of value for money and then include conditions where you promise not to kill the goose that lays the golden egg. If someone wants to buy your company you can negotiate whatever terms you want, or say no.
A non-compete in the sense of Matt's article is a nonnegotiable condition of employment that *outlasts* the period of your employment with whoever you're working for. You've stopped working for them (or been fired / laid off / let go) and you're not paid AND not allowed to use your own industry & knowledge to earn a living. That's bad and substantially different.
And I’m saying that shouldn’t be an option. Any enforceable “non-compete” clause must require ongoing substantial consideration.
For consulting firms that’s what garden leave is for. For business owners who sell that’s what a 18 month “consulting” contract for $xx,xxx a month is for.
This describes the non-competes of my youth. Back in the day the owners of one set of burger parlors didn’t worry about people leaving their outfits for a like establishment down the street.
Thoughtful post, Matt. One thing I wanted to point out re California: even though there is a policy generally making noncompetes unenforceable, there is an exception where it is necessary to protect trade secrets. The burden is on the employer, and it’s not super easy to meet. Your general points stand, but the reason why those tech companies can succeed could be because those employees who actually could harm the business by stealing trade secrets and giving them to a competitor can be restrained by a noncompete. There is of course a separate trade secret misappropriation law, but once the cat is out of the bag, the damage is done and sometimes irreparable. So there is probably a good case for a narrow noncompete protection, so long as the employer has the burden and can’t just baldly assert that they have trade secrets and use noncompetes indiscriminately (as happens in many other states).
I am confused by this. In California trade secret protection is exclusively by statute, not by contract. Yes, some employer contracts contain trade secret protections, but they are superseded by the statute. And that’s more efficient, because the statute applies to everyone, whether or not they acknowledge or agree that they have trade secrets. There is no additional protection afforded by contract - the “cat out of the bag” issue you raise is equally addressed by both. Source: I am a trade secret litigator in California.
The additional protection is that if an employee with knowledge of trade secrets is restrained from employment by a competitor, they are less likely to misappropriate trade secrets.
On a related note to non-competes, what about "non-poaching" agreements? Both between companies and employer-employee.
Various companies in the same industry here in town have a gentleman's non-poaching agreement. They'll absolutely hire you if you apply to them, but they won't actively seek you out to recruit you. Now that's not an enforceable thing - that's just a kindly agreement.
For the employer->employee one
At one point for a contract job I signed a non-poaching agreement that I wouldn't, after I was done with the contract, immediately(had to wait a year) try to recruit someone I had just worked with. I'm sympathetic (and I signed it) because I was getting introduced to a bunch of people on a short-term basis and it would be connections I exploit to pull them away later. Note that other employees were free to apply to the same place I went, I just couldn't actively recruit them (unsure how enforceable that was in court - in any event I took it seriously ethically because I signed it)
I've never been offered or signed a non-compete. (I work in software, and not in California)
Google, Apple, Adobe and a few others had a genuinely massive fine levied against them in the late-00s for just that sort of "gentlemen's agreement". (It was large enough that my share of the settlement as an entirely undistinguished mid-level engineer at Google was in the low four figures.) I think the big players in SV have been a lot less prone to such shenanigans since.
This is quite an interesting column, particularly for the reference on the uneven diffusion of "best practices" across firms. It corresponds to the Autor et al stuff on "superstar firms" getting all the gains. Nick Bloom (Stanford) has written on how this is even a within-firm phenomenon. On non-competes and know-how diffusion & monopsony: interesting points here, but there is an extensive empirical literature using diff-in-diff/regression discontinuity methodology to compare outcomes across states that adopted cognate regimes. Matt should do a follow-up testing his ideas against those studies.
Fwiw the SV vs. Boston argument against noncompete enforcement is empirically contested
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3516397