I think the answer is pretty simple: The people who work in coops are more likely to select for conscientiousness and other traits that correlate with being more-productive-than average. Another way of saying that is that they’re true believers. The other reason is that capitalist ownership likely scales easier than worker coops. It’s faster to make decisions, workers don’t need to put as much thought, there’s probably less bureaucracy and procedure. So even if the coops are 5-10% more productive when they hit maturity, it may be that it doesn’t take as long to reach maturity for more capitalist firms because the lack of egalitarianism streamlines the process.
Anecdotally, the people who care about co-ops in America are more likely to be, like, booksellers or whole foods-esque grocery shoppers. Working class people in my experience only want to own a business if they don’t want to be told what to do (I think there is social science on small business owners saying as much) and otherwise are very much just at their workplace to collect a paycheck. Heck, I’m not even a working class person and I wouldn’t want to be a co-owner at my hypothetical workplace. Too much responsibility. And yeah, there are low wage workers who’d generally benefit from more input and protections in the workplace, but I think if you gave them that, many of them would still not want to be owners, and many of them would still be lower productivity.
> The people who work in coops are more likely to select for conscientiousness and other traits that correlate with being more-productive-than average.
That's what the selection bias section was about. This could maybe explain away our observational findings, but not our experimental findings.
> The other reason is that capitalist ownership likely scales easier than worker coops.
That's certainly possible, though I haven't seen any evidence/data/studies on it.
> Too much responsibility
You can still work with representative democracy e.g. you still have managers, but if you think your manager is doing a bad job, you can vote for another one. Or if even that is too much responsibility you can work with voting proxies, giving one or more colleagues you trust the ability to vote for you. This has the benefit that you can immediately get the power back if you ever change your mind.
That last point has way too much friction for the average worker, including me. Most people (in America) are “clock-in, clock out, and don’t ask me to come in if you’re not paying me.” The people who benefit from greater rights and protections or whatever are likely service workers who don’t really have time in the work day for a mid-day meeting. I sincerely believe that people who are all-in on coops have a poor theory of mind about these things.
I think the evidence and data in favor of scalability is the fact that like 99% of the American economy is not coops!
I would have to see the actual methodology about the selection bias selection. Because I don’t see how you isolate such variables that aren’t in the studies being surveyed. Because as far as I can tell this study was on the firm-level, and so it was my impression that the selection effect was on studies on the type of firms (as in what industry they were, implying what technology they used), not the type of workers within the firm. Unless all of these studies surveyed had that information, I just don’t think that’s feasible
Edit: Yeah, I looked at the summary at the beginning and they did NOT control for personality type (but I wouldn’t expect them to). Anyway, you’re right about the experimental finding not being invalidated, but the problem is that it lacks external validity if the experimental findings can’t be generalized, given the selection bias.
> are “clock-in, clock out, and don’t ask me to come in if you’re not paying me.”
Which you can do perfectly well if you've given someone your proxy. If *even that* is too much you can also just abstain from voting, which is not the same as not being able to vote since you can still change your mind later.
> 99% of the American economy is not coops!
That's...what the entire post was about...? Specifically from "Why aren’t there more worker coops?" onward?
> Because I don’t see how you isolate such variables that aren’t in the studies being surveyed.
You don't with observational studies, which is why we invented experiments. So in e.g. 'Workplace democracy in the lab' they don't even look at existing firms but rather set up a lab experiment where they randomly add participants to different groups (so the psychological profiles are randomly distributed). These groups had to do paid labor (different problems to solve), but some of the groups could vote on their pay-scheme, while for others it was randomly selected by the computer.
Allowing a vote raised the amount of effort (problems attempted) by about 7%, and raised effective effort (problems solved correctly) by roughly 9%, compared to the no-vote groups. The effects remained statistically significant (p < 0.05) after controlling for math ability, gender, and the chosen pay scheme.
That study is behind a paywall, for all the studies you've cited, I've CTRL + F "personality type" and "conscientiousness" and those have not been controlled for. How do we know that the sample has a "randomly distributed" psychological profiles.
Setting up a lab experiment in a college setting may be a convenience sample, but it doesn't evade the criticism of selection effect. Because often times those convenience samples are not representative of the population (we know this notoriously for WEIRD societies, but it's also true of college students and the kind of people who sign up to do studies within WEIRD societies).
There are so many ways this sampling could be bad, that more keep sticking out as I think about them (they controlled for math ability, gender, and chosen pay scheme, but not conscientiousness!). So, again, the studies you're citing don't actually do the control for sampling you say they do to avoid the criticism.
> How do we know that the sample has a "randomly distributed" psychological profiles.
Because the participants were randomly distributed. So if I set up an experiment and I randomly distribute the participants over different groups, then we should expect different attributes to also be randomly distributed over the different groups. Not only things like conscientiousness but also things like blood types etc. If someone says, "you didn't check for bloodtypes: it's technically possible that the random process put all the same bloodtypes into the same groups and that explains the result!" we can say "yes, that's technically possible, but it's very improbable, so we should expect the bloodtypes to be randomly distributed". Same with conscientiousness.
But just because you can assume it, doesn't mean it's been controlled for. You're basically admitting it hasn't been controlled for!
Let's say there's a blood disease, a virus perhaps. Some blood types will naturally recover, while some won't with current medication, while some will recover with current medication. If we test a new drug and there's a general recovery on average, it's very important that it's controlled for the blood type. Because it's saliently important for the group that doesn't improve. Because it can be mathematically true that everyone gets better on average faster, but if the "naturally recover" blood type recovers faster, as does the "medical recover" people, while the unresponsive one does nothing, that's important information to control for. You can't just "assume it"
To transfer over to the employment statistics, we know that some big five personality traits are correlated with income, earnings and productivity. We also know that being deficient in some traits is correlated with lower earnings, productivity, etc. If it's my contention that the reason why these studies show that increase is because it's a sampling bias in favor of higher-productivity personality traits, it's not good enough to say "it's randomly selected." Because "randomly selected" does not mean the same as controlled for! I can randomly select people on my street and ask them what political party they support and it would be disproportionately Democrat because I live in an urban area, even if theoretically there's an even distribution of republicans and democrats.
It's actually pretty basic research methodology to not say the thing has been controlled for when it is, in fact, not controlled for!
Actually, the experiment I linked *is* a randomized controlled experiment. It was only over a three week period, so it's possible that it fades after that, but at least for the first three weeks we have a statistically significant result.
Interesting! However, I'm skeptical about the marginal productivity of worker co-ops over incentive pay. It can be tricky to distinguish between the two because worker co-ops almost always intrinsically implement a form of incentive pay, that is, the worker's pay varies with the firm's performance, whereas in many traditional capitalist firms, salary is fixed. Economic theory *can* predict that certain compensation schemes have free-rider problems, but other compensation schemes---especially pay for performance---do not. In fact, one of the most popular models today (principal-agent) predicts that pay for performance is optimal, group bonuses are good, and fixed salaries are worst.
This has created another mystery: why the heck do so many companies pay fixed salaries? It's probably because of risk aversion. People like having a consistent salary rather than getting paid $0 today and $250,000 in two years. (And that seems like a reason that might apply to worker co-ops as well.)
Anyway, the empirical evidence on worker co-ops productivity is not particularly convincing to me for this reason. For example, the experiment you cited cannot distinguish this because it compares a group that votes on its compensation contract to a group that does not, but the only allowable compensation contracts are group bonuses or tournament pay rather than individual pay-for-performance. Of course we should expect the group that can vote to perform better: the group vote is *itself* mimicking a selection effect, where people vote for the incentive contract they know will motivate themselves. If you as a manager in a capitalist firm knew your employees well enough to know which compensation contract would work best, you could re-create this effect in a capitalist firm (note: I'm not saying there is selection bias; I'm saying that the experiment's average treatment effect is equivalent to a sort of selection effect in a natural environment).
Also, we (economists) don't really have a good understanding of organizational structure in general. Early organizational economics models predicted that managers are useless, and newer models don't do much better. We still have no idea why managers are everywhere! But that too is an empirical observation that we shouldn't ignore. Does it mean that hierarchical organizational structures are better? No. But it does suggest that cooperative structures have some major frictions that prevent their implementation, at least.
> the marginal productivity of worker co-ops over incentive pay. [...] This has created another mystery: why the heck do so many companies pay fixed salaries?
Incentive pay is indeed a possible explanation, but I think that explanation still advantages coops. As I said on my post on employee stock (https://bobjacobs.substack.com/p/isnt-it-harmful-if-worker-co-ops ) giving employees stock without giving them voting power isn't enough. It may be that the reason why coops (by and large) can maintain incentive pay, while capitalist firms (by and large) cannot, is *because* of the voting power. If the owners of a capitalist firm want to pull shenanigans like diluting the shares, they can do so, whereas in a coop everyone has to vote on it first. In other words, employees might not trust the incentive pay offered by capitalist firms, while they do trust the incentive pay offered by coops.
> And that seems like a reason that might apply to worker co-ops as well
Possible, but also keep in mind that coops tend to fire people less. If a capitalist firm hits a downturn a lot of people are going to be fired, whereas this a lot lower for coops (and people who are fired are often quickly rehired later). I think this does a lot to alleviate risk-aversion. (another would be better social-safety-nets/UBI, which a socialist economy would have, but since that's outside the firms I'm kinda cheating with that answer).
> Does it mean that hierarchical organizational structures are better? No. But it does suggest that cooperative structures have some major frictions that prevent their implementation, at least.
The larger coops indeed tend to gravitate towards a "more hierarchical" system (e.g. mondragon). However, these are often in the form of ~'representative democracy', which is still vastly more democratic than traditional firms. If it turns out some form of decision-making-centralization is necessary, I would be fine with that, as long as it has the form of a ~'representative democracy' (or maybe a demarchy?) and not of a ~'autocracy'.
I agree with you on these points and do think that cooperatives are a not-tried-enough model of the firm, though I'm just more hesitant that their benefits will turn out to be all that large in practice.
As a side point, I wouldn't cast worker co-ops as necessarily socialist. Functionally, they're simply another way to structure firm decision-making, and that decision-making needn't churn out wealth equality. Frankly, selling it this way is a lot more palatable to prospective entrepreneurs and employees.
It is necessarily socialist because it puts the means of production in the hands of the workers. I agree that it won't help *that* much with wealth inequality, though I think it will help *somewhat* because it decreases inequality within firms (which is especially a big deal with monopsonies)
But yeah, they're just an additional tool to create an equitable economy, not a silver bullet or anything. And I definitely wouldn't present it as socialist to entrepreneurs (though I would to employees depending on their country of origin).
Not an economist so forgive my ignorance, but I worry your definition of socialism is overbroad. I'm a salaried worker, and I own stock in corporations thus owning some share of the means of production. But that doesn't make my 401(k) a socialist part of our economic system, does it?
Interesting post, although your comments about how investors get a worse deal has me concerned that cooperatives, if they became the standard business model. would discourage investment, especially given the possibility of a lock in effect which you discussed near the end of your post. In addition,, I am a little worried given how they generally operate that cooperatives create a financial incentive against hiring new workers. Also, while it’s intriguing that cooperatives have been spreading over the last decade that just makes me wonder why if they are so great, they only started gaining momentum recently.
Then again elite law firms appear to be doing fine, even though they are basically worker cooperatives run by the partners.
If the present trend of cooperatives increasing share in the economy continues, these empirical questions should be easy to resolve given most possible concerns would be easier to resolve with a larger sample of cooperatives especially as it would mitigate factors like the selection effect of which workers end up in a cooperative.
>your comments about how how investors get a worse deal has me concerned that cooperatives, if they became the standard business model. would discourage investment
Investors can still make money (a lot of money even) investing, just less than they do now. Which, given the enormous power and privileges they have in society (legal or otherwise) I wouldn't mind that much.
> create a financial incentive against hiring new workers
There are really large coops (e.g. mondragon) that hire a lot of workers, but you're right that game-theory-wise it only makes sense to hire workers if you think the worker can at least bring in enough revenue to pay their own salary. I think in most cases this will be the case, but if coops do err on the smaller side, I don't think I would mind it that much. This would be an organic way to reduce the amount of monopolies/monopsonies, which are perhaps the biggest danger to a (political) economy.
> why if they are so great, they only started gaining momentum recently.
My post suggests that it was delayed that long because of propaganda + ROI + firm design late start + legislative hurdles. My guess is the first and the last one are the most important in the timing.
> If the present trend of cooperatives increasing share in the economy continues, these empirical questions should be easy to resolve given most possible concerns would be easier to resolve with a larger sample of cooperatives especially as it would mitigate factors like the selection effect of which workers end up in a cooperative.
We could also look at countries like France and Italy, which already have a large amount of worker coops.
I think this is generally quite a good article, and am sympathetic to most of it. However, I think the strongest argument, that of investment, is reasonably effectively rebutted by what David Friedman says here (https://youtu.be/SAcR394NoAw?feature=shared&t=4682).
Essentially, if we expect capital to be the major limitation then economic sectors that require less capital should have far more workers coops, but this is not the case. Curious what you make of his argument.
Is it true that we don't see that? If we look at the countries that do have a lot of worker coops, e.g. France and Italy, it seems that they are indeed mostly started in sectors that require less capital (at first glance, I haven't studied this in depth or anything). It may be that once the other factors I mentioned start to lessen, that ROI becomes the major limiting factor, which creates the split in the economy that seems (at first glance) to be happening in countries like Uruguay and Italy.
One simple answer to the capital issue is for the co-op workers to put up their own capital for the initial capitalization of the business. If the co-op is more productive and can outcompete the capital firms, then the workers would reap an outsized reward on their earnings because they don't even need to pay the capital owners.
The obvious downside to this is the risk to the owners. It's tough to even get a loan in lieu of capital investment because I believe most lenders would want guarantees from the members.
It's simply a benefit of working for a capitalist firm in that I am not expected to contribute capital nor am I at risk for any accumulated wealth I have. They exploit me by scraping the excess value of my labor, but I exploit them by not having to put my own capital at risk.
A second and better solution would be for these worker co-ops to solicit investment from other like-minded people who would act as some type of investor or lender. I think there would be people out there willing to invest in projects that have a lower return when it is an entity structure that they want to support.
It seems that the reason coops do better with other coops around is because they then start cooperative leagues, which aid in the creation of new coops. Cooperative leagues facilitate the acquisition of equity through economies of scale which gives cooperatives more capital to fall back on, and also creates opportunities to open more franchises.
Another avenue is *consumer* coops. I can't remember where I read this, but I remember reading that consumer coops and worker coops work well together. This makes intuitive sense, and would especially be an interesting avenue for lending from credit unions (consumer coop banks). I'm a fan of credit unions in general (since it's easier to make them divest from firms that produce negative externalities), but this would be another reason to be in favor their creation.
There are also more ways I can think of to tackle this problem; like consumer-worker coop hybrids, selling non-voting shares (see my previous post on coops), sponsorships, pre-selling... However, this is getting more into the territory of finance and away from political economy, meaning my expertise starts to decrease and you should start taking my suggestions with a bigger grain of salt.
Thanks for the reply. Is there a profit incentive in creating new coops by an existing, successful coop, i.e., does the aiding coop receive some type of equity interest in the new coop in exchange for the capital support?
Do you think the coop structure lends itself to being large? By that I mean would you envision a Google or Visa being able to function in a coop entity structure? From what I can tell, the coop structure works well for focused enterprises, but not so much beyond that.
I can see there being two main ways of having a profit incentive for creating new coops.
The first is for creating economies of scale; if you e.g. have a bunch of different stores (or similar) in different locations then things like bulk orders, fixed costs, marketing, etc, all become cheaper (per unit of output), and you'll likely be able to borrow more/at cheaper rates from banks.
Another is vertical integration, which will be especially attractive for coops in sectors with large/volatile supply chains.
As for the size. Mondragon (a coop) has over 70,000 workers, while Visa has 31,600 employees, so yes, they can scale that large. However, the structure does change slightly. Smaller coops tend to function akin to a direct democracy, whereas larger coops tend to function more like a representative democracy. I know some anarchist have problems with representative democracy, but personally I'm totally fine with that.
It's my understanding that Mondragon is more a federation of coops that are in some ways integrated but much more disparate in the coop outputs and aren't necessarily coordinating business direction throughout the business divisions (please correct me if I'm wrong). I would agree that in a conglomerate structure, maybe coops make more sense (for example a Berkshire Hathaway could probably map onto a coop federation/league fairly well).
However, if I were to try to map the structure into the various business divisions of Visa (card, R&D, processor networks, bank relationships, vendor relationships, etc.) that wouldn't necessarily stand on their own as a separate enterprise, it's harder for me to envision it working or being competitive with capitalist firms that can scale without the worker vote/management that is a big selling point of coops.
Yes, mondragon is more decentralized than Visa, so if you really need a more centralized structure it would probably be more akin to a singular council/person that everyone votes on that makes the decisions. Those seem to work pretty well for running the much more complex machinery of countries (while simultaneously accommodating many more voters), so I don't see why they couldn't work for large companies too.
Probably if you need *really* large companies you have elections at multiple levels, like we have elections for mayors, governors and presidents etc. So e.g. in your local store the people of the store vote for the manager, then e.g. they select amongst themselves the regional manager, etc...
I think this is fine. There's less direct input from workers than in a ~direct democracy, but I'm more so interested in raising the floor/preventing the worse excesses of capitalism, rather than wanting a consensus on every decision. Coops are a realistic way to improve upon the existing system, they're not the end goal or anything.
Partnerships are the norm in a bunch of industries: law, consulting, advertising, accounting, medical practices, plenty more. They have a lot in common with worker coops, so I wonder if you're more inclined to emphasize the differences or the similarities. They're owned by workers and decisions are made collectively, but not all workers are owners.
They may be a stepping stone towards more coops, but I don't think (most of them) count. While a higher percentage of the workers have ownership, it's most often still only a subset. I also think that, by itself, giving workers stocks is not enough: https://substack.com/@bobjacobs/p-160177501
Partnerships also vary a lot, so it's hard to say anything definitive about them as a group, but *usually* they don't operate by a one-person-one-vote principle. Personally, I think it's fine to give someone in a coop maybe half a vote if they work part-time, but in partnerships it's not uncommon for some people to have orders of magnitude more voting power than others, which is not very democratic. So I guess they're better than conventional firms, but by and large, not as good as coops.
> The return on investment for society may be larger (especially if we take into account all the costs traditional firms externalize to society at large that cooperative firms internalize) but that doesn’t mean they have a higher return on investment for investors.
Trade is what increases social trust. There is a lot evidence on this idea. If you want to increase social trust you need to cut the government aggressively. I suppose that's not incompatible with market socialism
I think the answer is pretty simple: The people who work in coops are more likely to select for conscientiousness and other traits that correlate with being more-productive-than average. Another way of saying that is that they’re true believers. The other reason is that capitalist ownership likely scales easier than worker coops. It’s faster to make decisions, workers don’t need to put as much thought, there’s probably less bureaucracy and procedure. So even if the coops are 5-10% more productive when they hit maturity, it may be that it doesn’t take as long to reach maturity for more capitalist firms because the lack of egalitarianism streamlines the process.
Anecdotally, the people who care about co-ops in America are more likely to be, like, booksellers or whole foods-esque grocery shoppers. Working class people in my experience only want to own a business if they don’t want to be told what to do (I think there is social science on small business owners saying as much) and otherwise are very much just at their workplace to collect a paycheck. Heck, I’m not even a working class person and I wouldn’t want to be a co-owner at my hypothetical workplace. Too much responsibility. And yeah, there are low wage workers who’d generally benefit from more input and protections in the workplace, but I think if you gave them that, many of them would still not want to be owners, and many of them would still be lower productivity.
> The people who work in coops are more likely to select for conscientiousness and other traits that correlate with being more-productive-than average.
That's what the selection bias section was about. This could maybe explain away our observational findings, but not our experimental findings.
> The other reason is that capitalist ownership likely scales easier than worker coops.
That's certainly possible, though I haven't seen any evidence/data/studies on it.
> Too much responsibility
You can still work with representative democracy e.g. you still have managers, but if you think your manager is doing a bad job, you can vote for another one. Or if even that is too much responsibility you can work with voting proxies, giving one or more colleagues you trust the ability to vote for you. This has the benefit that you can immediately get the power back if you ever change your mind.
That last point has way too much friction for the average worker, including me. Most people (in America) are “clock-in, clock out, and don’t ask me to come in if you’re not paying me.” The people who benefit from greater rights and protections or whatever are likely service workers who don’t really have time in the work day for a mid-day meeting. I sincerely believe that people who are all-in on coops have a poor theory of mind about these things.
I think the evidence and data in favor of scalability is the fact that like 99% of the American economy is not coops!
I would have to see the actual methodology about the selection bias selection. Because I don’t see how you isolate such variables that aren’t in the studies being surveyed. Because as far as I can tell this study was on the firm-level, and so it was my impression that the selection effect was on studies on the type of firms (as in what industry they were, implying what technology they used), not the type of workers within the firm. Unless all of these studies surveyed had that information, I just don’t think that’s feasible
Edit: Yeah, I looked at the summary at the beginning and they did NOT control for personality type (but I wouldn’t expect them to). Anyway, you’re right about the experimental finding not being invalidated, but the problem is that it lacks external validity if the experimental findings can’t be generalized, given the selection bias.
> are “clock-in, clock out, and don’t ask me to come in if you’re not paying me.”
Which you can do perfectly well if you've given someone your proxy. If *even that* is too much you can also just abstain from voting, which is not the same as not being able to vote since you can still change your mind later.
> 99% of the American economy is not coops!
That's...what the entire post was about...? Specifically from "Why aren’t there more worker coops?" onward?
> Because I don’t see how you isolate such variables that aren’t in the studies being surveyed.
You don't with observational studies, which is why we invented experiments. So in e.g. 'Workplace democracy in the lab' they don't even look at existing firms but rather set up a lab experiment where they randomly add participants to different groups (so the psychological profiles are randomly distributed). These groups had to do paid labor (different problems to solve), but some of the groups could vote on their pay-scheme, while for others it was randomly selected by the computer.
Allowing a vote raised the amount of effort (problems attempted) by about 7%, and raised effective effort (problems solved correctly) by roughly 9%, compared to the no-vote groups. The effects remained statistically significant (p < 0.05) after controlling for math ability, gender, and the chosen pay scheme.
That study is behind a paywall, for all the studies you've cited, I've CTRL + F "personality type" and "conscientiousness" and those have not been controlled for. How do we know that the sample has a "randomly distributed" psychological profiles.
Setting up a lab experiment in a college setting may be a convenience sample, but it doesn't evade the criticism of selection effect. Because often times those convenience samples are not representative of the population (we know this notoriously for WEIRD societies, but it's also true of college students and the kind of people who sign up to do studies within WEIRD societies).
There are so many ways this sampling could be bad, that more keep sticking out as I think about them (they controlled for math ability, gender, and chosen pay scheme, but not conscientiousness!). So, again, the studies you're citing don't actually do the control for sampling you say they do to avoid the criticism.
> How do we know that the sample has a "randomly distributed" psychological profiles.
Because the participants were randomly distributed. So if I set up an experiment and I randomly distribute the participants over different groups, then we should expect different attributes to also be randomly distributed over the different groups. Not only things like conscientiousness but also things like blood types etc. If someone says, "you didn't check for bloodtypes: it's technically possible that the random process put all the same bloodtypes into the same groups and that explains the result!" we can say "yes, that's technically possible, but it's very improbable, so we should expect the bloodtypes to be randomly distributed". Same with conscientiousness.
But just because you can assume it, doesn't mean it's been controlled for. You're basically admitting it hasn't been controlled for!
Let's say there's a blood disease, a virus perhaps. Some blood types will naturally recover, while some won't with current medication, while some will recover with current medication. If we test a new drug and there's a general recovery on average, it's very important that it's controlled for the blood type. Because it's saliently important for the group that doesn't improve. Because it can be mathematically true that everyone gets better on average faster, but if the "naturally recover" blood type recovers faster, as does the "medical recover" people, while the unresponsive one does nothing, that's important information to control for. You can't just "assume it"
To transfer over to the employment statistics, we know that some big five personality traits are correlated with income, earnings and productivity. We also know that being deficient in some traits is correlated with lower earnings, productivity, etc. If it's my contention that the reason why these studies show that increase is because it's a sampling bias in favor of higher-productivity personality traits, it's not good enough to say "it's randomly selected." Because "randomly selected" does not mean the same as controlled for! I can randomly select people on my street and ask them what political party they support and it would be disproportionately Democrat because I live in an urban area, even if theoretically there's an even distribution of republicans and democrats.
It's actually pretty basic research methodology to not say the thing has been controlled for when it is, in fact, not controlled for!
It would be better if we could have some kind of controlled experiment but that’s probably not practical.
Actually, the experiment I linked *is* a randomized controlled experiment. It was only over a three week period, so it's possible that it fades after that, but at least for the first three weeks we have a statistically significant result.
Ok thanks 😊
Interesting! However, I'm skeptical about the marginal productivity of worker co-ops over incentive pay. It can be tricky to distinguish between the two because worker co-ops almost always intrinsically implement a form of incentive pay, that is, the worker's pay varies with the firm's performance, whereas in many traditional capitalist firms, salary is fixed. Economic theory *can* predict that certain compensation schemes have free-rider problems, but other compensation schemes---especially pay for performance---do not. In fact, one of the most popular models today (principal-agent) predicts that pay for performance is optimal, group bonuses are good, and fixed salaries are worst.
This has created another mystery: why the heck do so many companies pay fixed salaries? It's probably because of risk aversion. People like having a consistent salary rather than getting paid $0 today and $250,000 in two years. (And that seems like a reason that might apply to worker co-ops as well.)
Anyway, the empirical evidence on worker co-ops productivity is not particularly convincing to me for this reason. For example, the experiment you cited cannot distinguish this because it compares a group that votes on its compensation contract to a group that does not, but the only allowable compensation contracts are group bonuses or tournament pay rather than individual pay-for-performance. Of course we should expect the group that can vote to perform better: the group vote is *itself* mimicking a selection effect, where people vote for the incentive contract they know will motivate themselves. If you as a manager in a capitalist firm knew your employees well enough to know which compensation contract would work best, you could re-create this effect in a capitalist firm (note: I'm not saying there is selection bias; I'm saying that the experiment's average treatment effect is equivalent to a sort of selection effect in a natural environment).
Also, we (economists) don't really have a good understanding of organizational structure in general. Early organizational economics models predicted that managers are useless, and newer models don't do much better. We still have no idea why managers are everywhere! But that too is an empirical observation that we shouldn't ignore. Does it mean that hierarchical organizational structures are better? No. But it does suggest that cooperative structures have some major frictions that prevent their implementation, at least.
Thanks for the comment!
> the marginal productivity of worker co-ops over incentive pay. [...] This has created another mystery: why the heck do so many companies pay fixed salaries?
Incentive pay is indeed a possible explanation, but I think that explanation still advantages coops. As I said on my post on employee stock (https://bobjacobs.substack.com/p/isnt-it-harmful-if-worker-co-ops ) giving employees stock without giving them voting power isn't enough. It may be that the reason why coops (by and large) can maintain incentive pay, while capitalist firms (by and large) cannot, is *because* of the voting power. If the owners of a capitalist firm want to pull shenanigans like diluting the shares, they can do so, whereas in a coop everyone has to vote on it first. In other words, employees might not trust the incentive pay offered by capitalist firms, while they do trust the incentive pay offered by coops.
> And that seems like a reason that might apply to worker co-ops as well
Possible, but also keep in mind that coops tend to fire people less. If a capitalist firm hits a downturn a lot of people are going to be fired, whereas this a lot lower for coops (and people who are fired are often quickly rehired later). I think this does a lot to alleviate risk-aversion. (another would be better social-safety-nets/UBI, which a socialist economy would have, but since that's outside the firms I'm kinda cheating with that answer).
> Does it mean that hierarchical organizational structures are better? No. But it does suggest that cooperative structures have some major frictions that prevent their implementation, at least.
The larger coops indeed tend to gravitate towards a "more hierarchical" system (e.g. mondragon). However, these are often in the form of ~'representative democracy', which is still vastly more democratic than traditional firms. If it turns out some form of decision-making-centralization is necessary, I would be fine with that, as long as it has the form of a ~'representative democracy' (or maybe a demarchy?) and not of a ~'autocracy'.
I agree with you on these points and do think that cooperatives are a not-tried-enough model of the firm, though I'm just more hesitant that their benefits will turn out to be all that large in practice.
As a side point, I wouldn't cast worker co-ops as necessarily socialist. Functionally, they're simply another way to structure firm decision-making, and that decision-making needn't churn out wealth equality. Frankly, selling it this way is a lot more palatable to prospective entrepreneurs and employees.
It is necessarily socialist because it puts the means of production in the hands of the workers. I agree that it won't help *that* much with wealth inequality, though I think it will help *somewhat* because it decreases inequality within firms (which is especially a big deal with monopsonies)
But yeah, they're just an additional tool to create an equitable economy, not a silver bullet or anything. And I definitely wouldn't present it as socialist to entrepreneurs (though I would to employees depending on their country of origin).
Not an economist so forgive my ignorance, but I worry your definition of socialism is overbroad. I'm a salaried worker, and I own stock in corporations thus owning some share of the means of production. But that doesn't make my 401(k) a socialist part of our economic system, does it?
No because, you don't have *democratic* control, see my (shorter) post on this: https://bobjacobs.substack.com/p/isnt-it-harmful-if-worker-co-ops
Interesting post, although your comments about how investors get a worse deal has me concerned that cooperatives, if they became the standard business model. would discourage investment, especially given the possibility of a lock in effect which you discussed near the end of your post. In addition,, I am a little worried given how they generally operate that cooperatives create a financial incentive against hiring new workers. Also, while it’s intriguing that cooperatives have been spreading over the last decade that just makes me wonder why if they are so great, they only started gaining momentum recently.
Then again elite law firms appear to be doing fine, even though they are basically worker cooperatives run by the partners.
If the present trend of cooperatives increasing share in the economy continues, these empirical questions should be easy to resolve given most possible concerns would be easier to resolve with a larger sample of cooperatives especially as it would mitigate factors like the selection effect of which workers end up in a cooperative.
>your comments about how how investors get a worse deal has me concerned that cooperatives, if they became the standard business model. would discourage investment
Investors can still make money (a lot of money even) investing, just less than they do now. Which, given the enormous power and privileges they have in society (legal or otherwise) I wouldn't mind that much.
> create a financial incentive against hiring new workers
There are really large coops (e.g. mondragon) that hire a lot of workers, but you're right that game-theory-wise it only makes sense to hire workers if you think the worker can at least bring in enough revenue to pay their own salary. I think in most cases this will be the case, but if coops do err on the smaller side, I don't think I would mind it that much. This would be an organic way to reduce the amount of monopolies/monopsonies, which are perhaps the biggest danger to a (political) economy.
> why if they are so great, they only started gaining momentum recently.
My post suggests that it was delayed that long because of propaganda + ROI + firm design late start + legislative hurdles. My guess is the first and the last one are the most important in the timing.
> If the present trend of cooperatives increasing share in the economy continues, these empirical questions should be easy to resolve given most possible concerns would be easier to resolve with a larger sample of cooperatives especially as it would mitigate factors like the selection effect of which workers end up in a cooperative.
We could also look at countries like France and Italy, which already have a large amount of worker coops.
I think this is generally quite a good article, and am sympathetic to most of it. However, I think the strongest argument, that of investment, is reasonably effectively rebutted by what David Friedman says here (https://youtu.be/SAcR394NoAw?feature=shared&t=4682).
Essentially, if we expect capital to be the major limitation then economic sectors that require less capital should have far more workers coops, but this is not the case. Curious what you make of his argument.
Thanks for the comment!
Is it true that we don't see that? If we look at the countries that do have a lot of worker coops, e.g. France and Italy, it seems that they are indeed mostly started in sectors that require less capital (at first glance, I haven't studied this in depth or anything). It may be that once the other factors I mentioned start to lessen, that ROI becomes the major limiting factor, which creates the split in the economy that seems (at first glance) to be happening in countries like Uruguay and Italy.
One simple answer to the capital issue is for the co-op workers to put up their own capital for the initial capitalization of the business. If the co-op is more productive and can outcompete the capital firms, then the workers would reap an outsized reward on their earnings because they don't even need to pay the capital owners.
The obvious downside to this is the risk to the owners. It's tough to even get a loan in lieu of capital investment because I believe most lenders would want guarantees from the members.
It's simply a benefit of working for a capitalist firm in that I am not expected to contribute capital nor am I at risk for any accumulated wealth I have. They exploit me by scraping the excess value of my labor, but I exploit them by not having to put my own capital at risk.
A second and better solution would be for these worker co-ops to solicit investment from other like-minded people who would act as some type of investor or lender. I think there would be people out there willing to invest in projects that have a lower return when it is an entity structure that they want to support.
Thanks Connor, I largely agree.
It seems that the reason coops do better with other coops around is because they then start cooperative leagues, which aid in the creation of new coops. Cooperative leagues facilitate the acquisition of equity through economies of scale which gives cooperatives more capital to fall back on, and also creates opportunities to open more franchises.
Another avenue is *consumer* coops. I can't remember where I read this, but I remember reading that consumer coops and worker coops work well together. This makes intuitive sense, and would especially be an interesting avenue for lending from credit unions (consumer coop banks). I'm a fan of credit unions in general (since it's easier to make them divest from firms that produce negative externalities), but this would be another reason to be in favor their creation.
There are also more ways I can think of to tackle this problem; like consumer-worker coop hybrids, selling non-voting shares (see my previous post on coops), sponsorships, pre-selling... However, this is getting more into the territory of finance and away from political economy, meaning my expertise starts to decrease and you should start taking my suggestions with a bigger grain of salt.
Thanks for the reply. Is there a profit incentive in creating new coops by an existing, successful coop, i.e., does the aiding coop receive some type of equity interest in the new coop in exchange for the capital support?
Do you think the coop structure lends itself to being large? By that I mean would you envision a Google or Visa being able to function in a coop entity structure? From what I can tell, the coop structure works well for focused enterprises, but not so much beyond that.
I can see there being two main ways of having a profit incentive for creating new coops.
The first is for creating economies of scale; if you e.g. have a bunch of different stores (or similar) in different locations then things like bulk orders, fixed costs, marketing, etc, all become cheaper (per unit of output), and you'll likely be able to borrow more/at cheaper rates from banks.
Another is vertical integration, which will be especially attractive for coops in sectors with large/volatile supply chains.
As for the size. Mondragon (a coop) has over 70,000 workers, while Visa has 31,600 employees, so yes, they can scale that large. However, the structure does change slightly. Smaller coops tend to function akin to a direct democracy, whereas larger coops tend to function more like a representative democracy. I know some anarchist have problems with representative democracy, but personally I'm totally fine with that.
It's my understanding that Mondragon is more a federation of coops that are in some ways integrated but much more disparate in the coop outputs and aren't necessarily coordinating business direction throughout the business divisions (please correct me if I'm wrong). I would agree that in a conglomerate structure, maybe coops make more sense (for example a Berkshire Hathaway could probably map onto a coop federation/league fairly well).
However, if I were to try to map the structure into the various business divisions of Visa (card, R&D, processor networks, bank relationships, vendor relationships, etc.) that wouldn't necessarily stand on their own as a separate enterprise, it's harder for me to envision it working or being competitive with capitalist firms that can scale without the worker vote/management that is a big selling point of coops.
Yes, mondragon is more decentralized than Visa, so if you really need a more centralized structure it would probably be more akin to a singular council/person that everyone votes on that makes the decisions. Those seem to work pretty well for running the much more complex machinery of countries (while simultaneously accommodating many more voters), so I don't see why they couldn't work for large companies too.
Probably if you need *really* large companies you have elections at multiple levels, like we have elections for mayors, governors and presidents etc. So e.g. in your local store the people of the store vote for the manager, then e.g. they select amongst themselves the regional manager, etc...
I think this is fine. There's less direct input from workers than in a ~direct democracy, but I'm more so interested in raising the floor/preventing the worse excesses of capitalism, rather than wanting a consensus on every decision. Coops are a realistic way to improve upon the existing system, they're not the end goal or anything.
Partnerships are the norm in a bunch of industries: law, consulting, advertising, accounting, medical practices, plenty more. They have a lot in common with worker coops, so I wonder if you're more inclined to emphasize the differences or the similarities. They're owned by workers and decisions are made collectively, but not all workers are owners.
They may be a stepping stone towards more coops, but I don't think (most of them) count. While a higher percentage of the workers have ownership, it's most often still only a subset. I also think that, by itself, giving workers stocks is not enough: https://substack.com/@bobjacobs/p-160177501
Partnerships also vary a lot, so it's hard to say anything definitive about them as a group, but *usually* they don't operate by a one-person-one-vote principle. Personally, I think it's fine to give someone in a coop maybe half a vote if they work part-time, but in partnerships it's not uncommon for some people to have orders of magnitude more voting power than others, which is not very democratic. So I guess they're better than conventional firms, but by and large, not as good as coops.
> The return on investment for society may be larger (especially if we take into account all the costs traditional firms externalize to society at large that cooperative firms internalize) but that doesn’t mean they have a higher return on investment for investors.
Trade is what increases social trust. There is a lot evidence on this idea. If you want to increase social trust you need to cut the government aggressively. I suppose that's not incompatible with market socialism
I would love to see this evidence that to increase social trust you need to cut the government aggressively.
Excellent points.