8 ms·

Firm Inefficiency

230 points by a3voices 12y ago

fiatmoney 12y ago
On the inside, firms are usually set up as something like communist autocracies (formally everyone subordinate to the CEO, no internal prices, budgets set by fiat, etc). This has the advantage of reducing transaction costs (which in classical theory of the firm, is why they exist in the first place) but results in some of the same pathologies.
- TheOtherHobbes 12y ago
  Including a Cult of Personality, in too many cases. Even if it's not always wreathed around individuals, the role has been mythologized to an unhealthy, even a ridiculous extent which has allowed many CEOs to ignore consequences and accountability. But my suspicion is that in the general case, all systems tend towards aristocracy and oligarchy, no matter how they dress on the outside, and that every system evolves its own Theory of Social Privilege to legitimise and promote certain kinds of relationships and transactions and prevent others.
  - rhizome 12y ago
    Indeed, we see entire industries rewarding failure.
- nhaehnle 12y ago
  > communist autocracies That's a pretty crazy misuse of the term "communist". Just say "autocracies", the inside setup of typical firms really has nothing to do whatsoever with communism.
  - fiatmoney 12y ago
    Resource allocations are set by fiat and there are no price signals or "private property" per se (the accounting department doesn't "own" their computers, they're allocated them, and if that department is outsourced they'll be allocated to someone else). Departments don't technically control their own investments; any revenues they bring in are the company's as a whole and redistributed as they see fit. Seems apt to me; what would you call it?
holograham 12y ago
A very interesting list to say the least. The major component of the mega-corp money saving justification is the consolidation of overhead functions. e.g. finance, payroll, HR, media At tiny companies, these tasks are either outsourced or done by the c-level employees. As a company scales you need departments to handle these things. A company of 50 people may need 2 HR folks but the tasks for HR can scale (especially with technology) such that if that same company scaled to 1,000 employees it wouldnt need to hire 20 HR folks (maybe just 10 total needed) thus saving 10 heads in HR of salary + benefits + office + IT assets + support. Not insignificant sums.
- porlw 12y ago
  I think that in most large companies the efficiencies produced by economies of scale are mostly captured by the management hierarchy (in terms of executive compensation and the support apparatus provided by consultants) And why not? - the main shareholders are institutional investors who can't manage investments in hundreds of small companies, so it's not like they're going to take their money elsewhere.
aasarava 12y ago
It's really import to look for ways to make your organization more efficient (and these are good starting points) -- but I think there's a common misunderstanding between this sort of efficiency and the efficiency that economists talk about in relation to firms. Correct me if I'm wrong, but economists don't mean firms are "smooth flowing". They are saying that firms are more efficient relative to other firms and one-on-one marketplaces. E.g., it's inefficient for individuals to go directly to a marketplace to buy parts for a printer and then to find someone to assemble those parts. It's much more efficient for a firm to step in and handle the sourcing of the parts and to create a single place where people can buy them. That doesn't mean the firm has to be completely without internal friction.
staticshock 12y ago
The market pressures faced by firms (which, I'm assuming, are responsible for anything you might view as "firm efficiency") are similar to evolutionary pressures faced by any living creature. Evolution doesn't create "perfect" or "efficient" creatures, but predominantly creatures that are just "good enough." If you can live long enough to reproduce and pass on your genes, you're in the club. An animal with no natural predators, for instance, may get bigger and less adept at fighting off potential threats (after some generations,) which is kind of similar to the stagnation of a monopoly.
rayiner 12y ago
> My working hypothesis to explain these inefficiencies is that the people and supporting coalitions closest to them tend to gain from them, and that selection pressures on political coalitions are often much stronger than selection pressures on firms. Economists have a term for this, the principal-agent problem: http://en.wikipedia.org/wiki/Principal%E2%80%93agent_problem http://en.wikipedia.org/wiki/Principal%E2%80%93agent_problem. It is a key issue in understanding why corporations deviate from the efficient ideal. Most of the examples listed boil down to principal-agent problems. Consider, for example, poison pills. A company might adopt a poison pill because its executives and board feel that they might get replaced in a takeover. On the flip side, they may realize that the acquisition will be disastrous, and the other company's executives may be pushing for it because it feathers their own nests. I don't think any economist thinks that firms are actually efficient. They use it as an idealized model for analysis, just as an engineer might use an idealized linear model of a more complex physical system.
- csdrane 12y ago
  In case you weren't aware, the author of that post, Robin Hanson, is indeed an economist. So I imagine he is well aware of the principal-agent problem.
dj-wonk 12y ago
To give this some context, Dr. Hanson lists a number of ways why firms go not behave as suggested by traditional economic theories. See also: https://en.wikipedia.org/wiki/Theory_of_the_firm https://en.wikipedia.org/wiki/Theory_of_the_firm
clairity 12y ago
one major inefficiency hinted at but not directly named is the coordination problem (a la mythical man month). the tech startup ecosystem addresses that by having each startup "do one thing well" and having a simple way to coordinate efficiently (monetary payments in exchange for products/services). there are a number of business school case studies that talk about addressing coordination inefficiencies in large firms -- toyota (TPS/kanban) and haier (OEC/SST - http://zaipul.wikispaces.com/file/view/The+Haier%27s+Tao+of+innovation+-+a+case+study+of+the+emerging+total+innovation+management+%28TIM%29.pdf http://zaipul.wikispaces.com/file/view/The+Haier%27s+Tao+of+...) come to mind.
- duckingtest 12y ago
  This doesn't solve the problem, it just makes it invisible. When you buy something complex from other company, time spend communicating is not seen as bureaucracy, but productive work - on your company's end, communication may well be as efficient as possible. The same can be true at other company. When you sum the time/money spent on both sides, it's often worse than 'inefficient bureaucracy' at large company (which does both things internally). It may seem more inefficient just because people are able to see the entire process. If small companies really were more efficient, they wouldn't be able to grow too much, as they would be outcompeted by already existing small companies. The reverse actually happens.
  - clairity 12y ago
    not sure i get your point. i'm talking about the services ecosystem around tech. when i need servers, i just sign up at heroku. when i need source control, i get github. when i need agile tools, jira. need an html framework? bootstrap. analytics? email marketing? sales automation? bookkeeping? transportation? lodging? etc. etc. it's super efficient--just go to a website and pay. i spend no time coordinating my efforts with those companies, and the price i pay tends to compare very favorably to building it ourselves. the coordination problem is wrung out of the equation because teams are small and must compete. haier replicated this within the company with their STS model: each group in the supply chain effectively "paid" the upstream group and that became a simple metric for efficiency of coordination.
Htsthbjig 12y ago
There is not such a thing as a perfect company. There are companies that are more efficient than others. Not that big companies for example are "more efficient" because they are big enough. For example, Apple or Google pay very low taxes, because they could move, leaving lots of highly qualified people and technology behind. No state wants that, so they have negotiation power with gobertment, and they could pay lower taxes.
mynegation 12y ago
Many things (and not just in a corporate world) seem to be inefficient until you understand the undercurrents, hidden factors and incentive structure. Take for example hiring consultants. Why would you want to hire someone to get the watch off of your hand and tell you the time, as the joke goes? One factor is that consultants are used as the way to spread responsibility for important decisions. Consulting firms know that they may take (part of the) fall for someone and everyone plays this game. Another thing is that department or division may need to spend the budget, lest next year they are allocated less money.
- eru 12y ago
  The behaviour is understandable. But it is still inefficient in the economically sense of the term.
josephpmay 12y ago
I have an interesting hypothesis: firm inefficiency leads to greater income inequality. Think about it. Generally, smaller, newer firms are more efficient than older, larger firms. Greater firm inefficiency means that more older firms will be replaced by newer firms (whether by market position or outright bankruptcy). Average consumers can only invest in older, established firms, and, indeed, many 401k's are tied-up in stocks of 50+ year old companies. Millionaires, however, by being certified investors, can invest in many startups, a small portion of which will eventually become large, established firms in the future. This causes the rich to gain money, and others to loose it. Let me give a (contrived) example. Let's say you wanted to invest in a car company in 2005. If you were an average investor, you may have put your money in (old) GM, whose stock is now worthless. If you were a billionaire, you could have split your money between Tesla, ZAP, and Project Better Place. ZAP and PBP are now defunct, but you would have made a boatload of money off of Tesla.
- dnr 12y ago
  How does that square with the observation that venture capital returns aren't actually that good? http://pando.com/2013/07/30/ouch-ten-year-venture-returns-still-lag-the-broader-markets/ http://pando.com/2013/07/30/ouch-ten-year-venture-returns-st... http://avc.com/2013/02/venture-capital-returns/ http://avc.com/2013/02/venture-capital-returns/ What percentage of certified investors actually do invest in startups, instead of just index funds or fancy managed funds that invest in established companies?
- mseebach 12y ago
  That's good point, but if that's the effect, the cause isn't inefficient firms, it's the regulation that forbids "regular people" from investing in startups. https://en.wikipedia.org/wiki/Accredited_investor https://en.wikipedia.org/wiki/Accredited_investor Realistically, though, some of your 401k fund is likely invested in startups, it's just highly diversified because it's very high risk, and you don't want to gamble with your pension pot.
- eru 12y ago
  Income equality has been around for longer than the whole certified investor business (or even the US).
vijayboyapati 12y ago
Peter Klein has some very interesting research showing there are limits imposed on organizations by the problem of economic calculation: http://mises.org/document/350/Economic-Calculation-and-the-Limits-of-Organization http://mises.org/document/350/Economic-Calculation-and-the-L... This is an application of Mises' work on economic calculation in socialist economies and why it is so problematic for such economies to allocate capital without squandering it. At scale I think this affects companies like Microsoft which, as they grow, find it harder and harder to allocate resources without an internal price function. E.g., how many engineers do you put on infrastructure versus Xbox? The answer isn't obvious because you can't directly measure the profitability of the infrastructure team. When teams are small (such as in startups) the leader can kind of guess this stuff, and successful leaders do it well, while bad ones go out of business. Some companies try create internal markets to allocate capital but that has problems too, because it causes some level of friction in the process of how capital is internally allocated. This is why I think Ycombinator's model is potentially very useful. Instead of a monolithic structure where capital is essentially moved between divisions based on the judgment of a single person (the CEO), there is a syndicate of companies which tend to produce goods that the others need, while still maintaining a market so that prices can be established and used as a guide for which projects should receive more (or less) capital.
- lifeisstillgood 12y ago
  I may have missed something - are you saying that ycombinator requires all it's startups to use say Dropbox and not pay Box? I had not heard this. If this is just informal is it measurable? I don't doubt it is effective however - it suggests alumni networks or diaspora like family networks (strongly suggested in China and India for example)
  - vijayboyapati 12y ago
    It is not a requirement, but companies within batches, and within Ycombinator in general have a unique insight into the demands and pain-points of their fellow companies. This is why each batch usually has a handful of companies working on products specifically designed for other companies (e.g., MailGun, SendHub, Heroku, Mixpanel etc). Some of the companies pivot toward building tools if/when they discover their original idea isn't getting traction.
JacobAldridge 12y ago
I run a number of Capacity Planning projects for small businesses. It's really not economically feasible for them to determine all of their constraints to the nth degree, so I usually start with two figures: What is their perfect world, 100% Capacity / utilisation rate? And then what happens if we subtract 20-30% of that for 'Uncontrollables' and use that as the Real World target. There's a bit more science behind it than just that, but the end result proves accurate and practical. Losing 20% doesn't sound like much, until you realise that's 1 day per week for every full time employee that we conclude will never be efficient for any decent length of time.
rythie 12y ago
Most of these problems are due to employees looking out for themselves and their own career. Almost all employees stand to gain nothing from a large company they work for, doing well. Even if the company does badly, any work they've done on their career whilst there will help them get a new job (salary, seniority, achivements etc.)
- michaelochurch 12y ago
  Almost all employees stand to gain nothing from a large company they work for, doing well. Even if the company does badly, any work they've done on their career whilst there will help them get a new job (salary, seniority, achivements etc.) No, the employees want their companies to do well. If you're at a place for 3 years or longer, its reputation will affect yours. However, most of them can't really do anything that would have a significant effect on the performance of the company. On the day-to-day microtrades, they're going to do what favors themselves. No one (except an idiot, or someone with 20+ percent in stock) would get fired for his own company's sake. The claim that employees "don't care" is off. It's always better to be at a successful company than a failing one, all else equal. They do favor their own interests, of course, when there is a conflict. It is true that those with substantial power in an organization (who can effect its macroscopic performance) tend to be egregiously selfish and possibly toxic. There are bad actors out there, and the executive overpay of the past 30 years has only made it more profitable for executives to trash their own companies (through externalized costs, so they don't get caught) if they can bolster their careers by doing so. We're way past the point of getting increasing quality per dollar paid to executives.
  - rythie 12y ago
    Performance wise it's still going to be the same company you joined. For example, Microsoft hasn't done well over the last few years, but people are mostly still going to be still impressed you worked there.
exelius 12y ago
All models are wrong, but that doesn't mean they're not useful. OP right in that individual firms are often quite inefficient, but the thought is that the market will take care of this in the long run: either a new competitor will arise and disrupt an existing market, or the shareholders will get rid of the management if the inefficiency is especially egregious. The thing about macroeconomics is that you basically get to ignore the blips because on a long enough time scale, the irregularities start to smooth out. Scientific advances happen in other fields that may significantly disrupt another industry and vault progress forward, so even if for a while things were inefficient, the market catches up eventually. Economics isn't about predicting the future with great accuracy; it's about gathering as much information as possible to be able to make decisions about what to do right now. You're playing the odds, and odds are, companies will be efficient. Sometimes they're not, which is the entire reason we have the stock market.
jasode 12y ago
The article's logic is flawed. It lists 19 inefficiencies but such a list is not as persuasive as the author thinks. What we ultimately care about is the net efficiency of the firm. Does the elimination of internal transaction costs outweigh the other inefficiencies the author points out? Internal economic exchanges as explicit transaction costs don't exist within the firm (e.g. acct dept does not send an invoice to the CTO for CPA work relating to budgeting the capital purchases of rack servers. And IT does not charge acct dept for upgrading the desktops web browser from IE6 to IE9.) To prove the author's idea, the comparison would be something like: A firm of 100 employees with those 19 inefficiencies underperforms and is less profitable than a loose confederation of 100 freelancers with their particular inefficiencies. (The 100 freelancers would also have a unique list of inefficiencies -- because nothing is perfect.) We could also itemize 20 "inefficiencies" of the biological family unit. E.g. the child learns the language of his parents -- maybe Portugese -- instead of learning the most marketable and global language such as Chinese or English. The child eats some meals from grandmother's recipes instead of eating a scientifically designed diet for peak growth. And so on. Does such a list of inefficiencies convince society that Plato's idea of taking all children from parents and assigning them government custody is a better idea? That's the type of logic I see in the article.
- ebiester 12y ago
  Except that individual departments do bill each other internally. It's how major corporation accounting works.
  - jasode 12y ago
    Internal cost accounting, activity-based costing, chargebacks to departments for Xerox photocopies, etc isn't what I'm talking about. The CFO & acct dept does not send a literal invoice to the IT department where the CTO has his own mini accounting staff handle the invoice and cut a check back to the CFO. When corporate counsel reviews a Terms of Agreement for the IT dept's web service API, they don't send an invoice for $400/hour for "legal services". Not dealing with thousands of such "transactions" as explicit costs is one of the reasons firms exists (so the theory goes.) Internal operations do not send purchase orders to each other. They don't remit checks to each other. Each sub department doesn't have their own sales team haggling internal prices & discounts with other department heads. Etc, etc. It's the external companies' interactions that have all those costlier transactions.
tikhonj 12y ago
This really reminds me of how Yanis Varoufakis, Valve's economist, describes most firms: > Interestingly, however, there is one last bastion of economic activity that proved remarkably resistant to the triumph of the market: firms, companies and, later, corporations. Think about it: market-societies, or capitalism, are synonymous with firms, companies, corporations. And yet, quite paradoxically, firms can be thought of as market-free zones. Within their realm, firms (like societies) allocate scarce resources (between different productive activities and processes). Nevertheless they do so by means of some non-price, more often than not hierarchical, mechanism! The firm, in this view, operates outside the market; as an island within the market archipelago. Effectively, firms can be seen as oases of planning and command within the vast expanse of the market. In another sense, they are the last remaining vestiges of pre-capitalist organisation within… capitalism. The image of a firm as an island of control within a greater see of capitalism really stuck with me. I think it's a pretty good image to support Valve's style of organization. (Of course, I'm also a bit biased: I prefer that philosophy for reasons of my own besides just increasing efficiency.) The whole article ("Why Valve? Or, what do we needcorporations for and how does Valve’s management structure fit into today’s corporate world?"[1]) is well worth a read. I think it's a very constructive look at the issue that this post is talking about. [1]: http://blogs.valvesoftware.com/economics/why-valve-or-what-do-we-need-corporations-for-and-how-does-valves-management-structure-fit-into-todays-corporate-world/ http://blogs.valvesoftware.com/economics/why-valve-or-what-d...
- jerf 12y ago
  I actually think that's an important aspect of capitalism's success. It only (metaphorically) claims to be a good global organization scheme. Within that framework, experimentation and variation is permitted. Do you want a hippie commune? Sure. Coop? Sure. Conventional corporation? Yup. One dude in a garage? Sure. National-scale non-profit? Go for it. And they're all different organizations that can freely live and die as the situation warrants. It may not be perfect, but it can evolve without breaking. Contrast that with a State-Uber-Alles philosophy like Communism, and you get a situation where you've got what is putatively the same organization trying to organize the strategy for keeping the borders secure with nuclear weapons also trying to manage the local school bake sale. It's a fragile setup if it isn't perfect, and of course it isn't perfect. And lo, Communist governments that fall tend to fall hard and shatter their whole society, economy, country, and all in the process. All the eggs, you might say. And let me highlight the issue of resiliency and fragility as being my main point, rather than any particular local issue of efficiency on this point or that. Capitalist societies hardly blink if a 100-person organizations proves unable to make a profit. It might be locally annoying, even devastating to some people, but globally it tends to work out for the best. (And those 100 people tend to be able to move on to something else too.) By contrast a centralized command-and-control system may keep this net-negative organization going for a long time for any number of reasons, but all of them, in the end, meaning that a net-negative organization lives on, draining the society it lives in. (Bear in mind the meaning of "net negative"... the entire point is that whatever benefits it may have, it is still net negative in my discussion here.) The summation of such decisions is incredibly destructive.
michaelochurch 12y ago
Warning: this is a huge topic. For further reading: Venkatesh Rao's series on the Gervais Principle: http://www.ribbonfarm.com/2009/10/07/the-gervais-principle-or-the-office-according-to-the-office/ http://www.ribbonfarm.com/2009/10/07/the-gervais-principle-o... My series on the same issue (organizational inefficiency): http://michaelochurch.wordpress.com/2013/02/19/gervais-principle-questioned-macleods-hierarchy-the-technocrat-and-vc-startups/ http://michaelochurch.wordpress.com/2013/02/19/gervais-princ... In a picture: http://gapingvoid.com/2004/06/27/company-hierarchy/ http://gapingvoid.com/2004/06/27/company-hierarchy/ Short version: firms are inefficient because organizations are innately prone to a predictable degeneracy of the group into three sets of people, none of which is capable of meeting an organizations trifecta of demands: subordination, dedication, and strategy (working on the right things). A person who is strategic in his personal life is going to either optimize for least-work/most-comfort (Macleod "Loser" strategy) or most-yield/least-subordination (Macleod "Sociopath"). By definition, those who are subordinate and dedicated are not strategic in their work lives, and probably not equipped to run a company for that reason. Strategically subordinate but not dedicated "Losers" do the bare minimum and won't push themselves beyond how far they are managed. Among the strategic and dedicated (but not subordinate) "Sociopaths" there are some good people, but plenty of narcissists and bad actors as well. Cleaning up the messes made by their Sociopath bosses and Loser underlings are the Clueless middle managers, who are subordinate and dedicated but not strategic. For various reasons, these three sets are not good at working together but usually play against each other, with one dominant archetype characterizing the organization. (In VC-funded startups, it's Clueless; at Enron, it was the bad kind of Sociopath. In most checked-out mediocre companies, it's the Loser.) There are "good Sociopaths" out there (e.g. "sheep dogs") and I'm probably one of them, but they're often at a disadvantage against the bad ones. Often they get sick of taking on political risk (for games whose prizes cannot be enjoyed fully by the most ethical people; power of any kind is most valuable to those who will be corrupt with it) and become Losers with age. A few might be able to start their own companies, but VC won't help them because VC-funded startups are all about Cluelessness (which is why VCs have their chicken-hawk fetishism for very young, white, upper-class men). Fixing this problem is possible, even in larger companies, but very few organizations view it as worth the cost, especially in the circa-2014 startup game where companies are built to be sold, not to last more than 5 years (in which case MacLeod-style degeneracy will just be starting to impede performance, but is unlikely to kill the thing). Why? The classical explanation (from MBAs and investors and even founders) for why they keep starting crappy companies is that they can't afford the risk of alternative organization styles (e.g. Valve's open allocation) while also innovating on product, as if there were some finite pool of risk and "all of it" had to be allocated to product innovation. I strongly disagree. I think they're full of shit and trying to justify their own mediocrity. But that's what they say.
porter 12y ago
Yet another blog that I want to subscribe to via email, but can't (easily). This kind of thoughtful analysis is missing from today's obsession with listicles.
- notduncansmith 12y ago
  http://www.overcomingbias.com/feed http://www.overcomingbias.com/feed + https://ifttt.com/recipes/4447-rss-mail https://ifttt.com/recipes/4447-rss-mail I made it easy :)
Pitarou 12y ago
In essence, this is quite a simple matter. Here's Theory of the Firm 101 for you: In the real world, pure free market economies are inefficient. (If you think otherwise, you've never tried to co-ordinate a complex project with multiple independent contractors.) In the real world, pure bureaucracies are inefficient. (Think North Korea, or the old Soviet Union.) But they are inefficient in different ways. It turns out that we can improve efficiency if we have islands of bureaucracy (firms) in a sea of free market competition. And, since these firms get to keep some of that efficiency gain for themselves, that's what happens: one way or another, people organize themselves into firms, and reap the benefits.
mrcactu5 12y ago
why is the author so reluctant to agree these inefficiencies exist? is he afraid of biases in his own list? Also "smart sincere people" is not a notion I am used to hearing in academic discourse. If he is going to be critical, I wonder how he characterizes this notion Is it purely from a game-theory perspective?
- Sniffnoy 12y ago
  See e.g. http://www.overcomingbias.com/2009/12/the-smart-sincere-syndrome.html http://www.overcomingbias.com/2009/12/the-smart-sincere-synd... He also refers to "smart sincere" people as "nerds", so you can search his blog for that as well.
ttty 12y ago
TIL: "Modeling is always a tradeoff between realism and understanding."
mempko 12y ago
I am happy to see that people are slowly rediscovering socialism. However, the "efficiency" argument is one of the poorest arguments because "efficiency" has always been a hustle. There is no way to predict and even measure all the effects of an economic decision to make any judgement based on efficiency.
mathattack 12y ago
There is always a principal/agent problem in large companies. "What's in it for me?" differs from "What's in it for the company?" Much of management and leadership is related to overcoming this issue.

230 points by a3voices 12y ago

On the inside, firms are usually set up as something like communist autocracies (formally everyone subordinate to the CEO, no internal prices, budgets set by fiat, etc). This has the advantage of reducing transaction costs (which in classical theory of the firm, is why they exist in the first place) but results in some of the same pathologies.

TheOtherHobbes 12y ago
Including a Cult of Personality, in too many cases. Even if it's not always wreathed around individuals, the role has been mythologized to an unhealthy, even a ridiculous extent which has allowed many CEOs to ignore consequences and accountability. But my suspicion is that in the general case, all systems tend towards aristocracy and oligarchy, no matter how they dress on the outside, and that every system evolves its own Theory of Social Privilege to legitimise and promote certain kinds of relationships and transactions and prevent others.
- rhizome 12y ago
  Indeed, we see entire industries rewarding failure.
nhaehnle 12y ago
> communist autocracies That's a pretty crazy misuse of the term "communist". Just say "autocracies", the inside setup of typical firms really has nothing to do whatsoever with communism.
- fiatmoney 12y ago
  Resource allocations are set by fiat and there are no price signals or "private property" per se (the accounting department doesn't "own" their computers, they're allocated them, and if that department is outsourced they'll be allocated to someone else). Departments don't technically control their own investments; any revenues they bring in are the company's as a whole and redistributed as they see fit. Seems apt to me; what would you call it?

holograham 12y ago

A very interesting list to say the least. The major component of the mega-corp money saving justification is the consolidation of overhead functions. e.g. finance, payroll, HR, media At tiny companies, these tasks are either outsourced or done by the c-level employees. As a company scales you need departments to handle these things. A company of 50 people may need 2 HR folks but the tasks for HR can scale (especially with technology) such that if that same company scaled to 1,000 employees it wouldnt need to hire 20 HR folks (maybe just 10 total needed) thus saving 10 heads in HR of salary + benefits + office + IT assets + support. Not insignificant sums.

porlw 12y ago
I think that in most large companies the efficiencies produced by economies of scale are mostly captured by the management hierarchy (in terms of executive compensation and the support apparatus provided by consultants) And why not? - the main shareholders are institutional investors who can't manage investments in hundreds of small companies, so it's not like they're going to take their money elsewhere.

aasarava 12y ago

It's really import to look for ways to make your organization more efficient (and these are good starting points) -- but I think there's a common misunderstanding between this sort of efficiency and the efficiency that economists talk about in relation to firms. Correct me if I'm wrong, but economists don't mean firms are "smooth flowing". They are saying that firms are more efficient relative to other firms and one-on-one marketplaces. E.g., it's inefficient for individuals to go directly to a marketplace to buy parts for a printer and then to find someone to assemble those parts. It's much more efficient for a firm to step in and handle the sourcing of the parts and to create a single place where people can buy them. That doesn't mean the firm has to be completely without internal friction.

staticshock 12y ago

The market pressures faced by firms (which, I'm assuming, are responsible for anything you might view as "firm efficiency") are similar to evolutionary pressures faced by any living creature. Evolution doesn't create "perfect" or "efficient" creatures, but predominantly creatures that are just "good enough." If you can live long enough to reproduce and pass on your genes, you're in the club. An animal with no natural predators, for instance, may get bigger and less adept at fighting off potential threats (after some generations,) which is kind of similar to the stagnation of a monopoly.

rayiner 12y ago

> My working hypothesis to explain these inefficiencies is that the people and supporting coalitions closest to them tend to gain from them, and that selection pressures on political coalitions are often much stronger than selection pressures on firms. Economists have a term for this, the principal-agent problem: http://en.wikipedia.org/wiki/Principal%E2%80%93agent_problem http://en.wikipedia.org/wiki/Principal%E2%80%93agent_problem. It is a key issue in understanding why corporations deviate from the efficient ideal. Most of the examples listed boil down to principal-agent problems. Consider, for example, poison pills. A company might adopt a poison pill because its executives and board feel that they might get replaced in a takeover. On the flip side, they may realize that the acquisition will be disastrous, and the other company's executives may be pushing for it because it feathers their own nests. I don't think any economist thinks that firms are actually efficient. They use it as an idealized model for analysis, just as an engineer might use an idealized linear model of a more complex physical system.

csdrane 12y ago
In case you weren't aware, the author of that post, Robin Hanson, is indeed an economist. So I imagine he is well aware of the principal-agent problem.

dj-wonk 12y ago

To give this some context, Dr. Hanson lists a number of ways why firms go not behave as suggested by traditional economic theories. See also: https://en.wikipedia.org/wiki/Theory_of_the_firm https://en.wikipedia.org/wiki/Theory_of_the_firm

clairity 12y ago

one major inefficiency hinted at but not directly named is the coordination problem (a la mythical man month). the tech startup ecosystem addresses that by having each startup "do one thing well" and having a simple way to coordinate efficiently (monetary payments in exchange for products/services). there are a number of business school case studies that talk about addressing coordination inefficiencies in large firms -- toyota (TPS/kanban) and haier (OEC/SST - http://zaipul.wikispaces.com/file/view/The+Haier%27s+Tao+of+innovation+-+a+case+study+of+the+emerging+total+innovation+management+%28TIM%29.pdf http://zaipul.wikispaces.com/file/view/The+Haier%27s+Tao+of+...) come to mind.

duckingtest 12y ago
This doesn't solve the problem, it just makes it invisible. When you buy something complex from other company, time spend communicating is not seen as bureaucracy, but productive work - on your company's end, communication may well be as efficient as possible. The same can be true at other company. When you sum the time/money spent on both sides, it's often worse than 'inefficient bureaucracy' at large company (which does both things internally). It may seem more inefficient just because people are able to see the entire process. If small companies really were more efficient, they wouldn't be able to grow too much, as they would be outcompeted by already existing small companies. The reverse actually happens.
- clairity 12y ago
  not sure i get your point. i'm talking about the services ecosystem around tech. when i need servers, i just sign up at heroku. when i need source control, i get github. when i need agile tools, jira. need an html framework? bootstrap. analytics? email marketing? sales automation? bookkeeping? transportation? lodging? etc. etc. it's super efficient--just go to a website and pay. i spend no time coordinating my efforts with those companies, and the price i pay tends to compare very favorably to building it ourselves. the coordination problem is wrung out of the equation because teams are small and must compete. haier replicated this within the company with their STS model: each group in the supply chain effectively "paid" the upstream group and that became a simple metric for efficiency of coordination.

Htsthbjig 12y ago

There is not such a thing as a perfect company. There are companies that are more efficient than others. Not that big companies for example are "more efficient" because they are big enough. For example, Apple or Google pay very low taxes, because they could move, leaving lots of highly qualified people and technology behind. No state wants that, so they have negotiation power with gobertment, and they could pay lower taxes.

mynegation 12y ago

Many things (and not just in a corporate world) seem to be inefficient until you understand the undercurrents, hidden factors and incentive structure. Take for example hiring consultants. Why would you want to hire someone to get the watch off of your hand and tell you the time, as the joke goes? One factor is that consultants are used as the way to spread responsibility for important decisions. Consulting firms know that they may take (part of the) fall for someone and everyone plays this game. Another thing is that department or division may need to spend the budget, lest next year they are allocated less money.

eru 12y ago
The behaviour is understandable. But it is still inefficient in the economically sense of the term.

josephpmay 12y ago

I have an interesting hypothesis: firm inefficiency leads to greater income inequality. Think about it. Generally, smaller, newer firms are more efficient than older, larger firms. Greater firm inefficiency means that more older firms will be replaced by newer firms (whether by market position or outright bankruptcy). Average consumers can only invest in older, established firms, and, indeed, many 401k's are tied-up in stocks of 50+ year old companies. Millionaires, however, by being certified investors, can invest in many startups, a small portion of which will eventually become large, established firms in the future. This causes the rich to gain money, and others to loose it. Let me give a (contrived) example. Let's say you wanted to invest in a car company in 2005. If you were an average investor, you may have put your money in (old) GM, whose stock is now worthless. If you were a billionaire, you could have split your money between Tesla, ZAP, and Project Better Place. ZAP and PBP are now defunct, but you would have made a boatload of money off of Tesla.

dnr 12y ago
How does that square with the observation that venture capital returns aren't actually that good? http://pando.com/2013/07/30/ouch-ten-year-venture-returns-still-lag-the-broader-markets/ http://pando.com/2013/07/30/ouch-ten-year-venture-returns-st... http://avc.com/2013/02/venture-capital-returns/ http://avc.com/2013/02/venture-capital-returns/ What percentage of certified investors actually do invest in startups, instead of just index funds or fancy managed funds that invest in established companies?
mseebach 12y ago
That's good point, but if that's the effect, the cause isn't inefficient firms, it's the regulation that forbids "regular people" from investing in startups. https://en.wikipedia.org/wiki/Accredited_investor https://en.wikipedia.org/wiki/Accredited_investor Realistically, though, some of your 401k fund is likely invested in startups, it's just highly diversified because it's very high risk, and you don't want to gamble with your pension pot.
eru 12y ago
Income equality has been around for longer than the whole certified investor business (or even the US).

vijayboyapati 12y ago

Peter Klein has some very interesting research showing there are limits imposed on organizations by the problem of economic calculation: http://mises.org/document/350/Economic-Calculation-and-the-Limits-of-Organization http://mises.org/document/350/Economic-Calculation-and-the-L... This is an application of Mises' work on economic calculation in socialist economies and why it is so problematic for such economies to allocate capital without squandering it. At scale I think this affects companies like Microsoft which, as they grow, find it harder and harder to allocate resources without an internal price function. E.g., how many engineers do you put on infrastructure versus Xbox? The answer isn't obvious because you can't directly measure the profitability of the infrastructure team. When teams are small (such as in startups) the leader can kind of guess this stuff, and successful leaders do it well, while bad ones go out of business. Some companies try create internal markets to allocate capital but that has problems too, because it causes some level of friction in the process of how capital is internally allocated. This is why I think Ycombinator's model is potentially very useful. Instead of a monolithic structure where capital is essentially moved between divisions based on the judgment of a single person (the CEO), there is a syndicate of companies which tend to produce goods that the others need, while still maintaining a market so that prices can be established and used as a guide for which projects should receive more (or less) capital.

lifeisstillgood 12y ago
I may have missed something - are you saying that ycombinator requires all it's startups to use say Dropbox and not pay Box? I had not heard this. If this is just informal is it measurable? I don't doubt it is effective however - it suggests alumni networks or diaspora like family networks (strongly suggested in China and India for example)
- vijayboyapati 12y ago
  It is not a requirement, but companies within batches, and within Ycombinator in general have a unique insight into the demands and pain-points of their fellow companies. This is why each batch usually has a handful of companies working on products specifically designed for other companies (e.g., MailGun, SendHub, Heroku, Mixpanel etc). Some of the companies pivot toward building tools if/when they discover their original idea isn't getting traction.

JacobAldridge 12y ago

I run a number of Capacity Planning projects for small businesses. It's really not economically feasible for them to determine all of their constraints to the nth degree, so I usually start with two figures: What is their perfect world, 100% Capacity / utilisation rate? And then what happens if we subtract 20-30% of that for 'Uncontrollables' and use that as the Real World target. There's a bit more science behind it than just that, but the end result proves accurate and practical. Losing 20% doesn't sound like much, until you realise that's 1 day per week for every full time employee that we conclude will never be efficient for any decent length of time.

rythie 12y ago

Most of these problems are due to employees looking out for themselves and their own career. Almost all employees stand to gain nothing from a large company they work for, doing well. Even if the company does badly, any work they've done on their career whilst there will help them get a new job (salary, seniority, achivements etc.)

michaelochurch 12y ago
Almost all employees stand to gain nothing from a large company they work for, doing well. Even if the company does badly, any work they've done on their career whilst there will help them get a new job (salary, seniority, achivements etc.) No, the employees want their companies to do well. If you're at a place for 3 years or longer, its reputation will affect yours. However, most of them can't really do anything that would have a significant effect on the performance of the company. On the day-to-day microtrades, they're going to do what favors themselves. No one (except an idiot, or someone with 20+ percent in stock) would get fired for his own company's sake. The claim that employees "don't care" is off. It's always better to be at a successful company than a failing one, all else equal. They do favor their own interests, of course, when there is a conflict. It is true that those with substantial power in an organization (who can effect its macroscopic performance) tend to be egregiously selfish and possibly toxic. There are bad actors out there, and the executive overpay of the past 30 years has only made it more profitable for executives to trash their own companies (through externalized costs, so they don't get caught) if they can bolster their careers by doing so. We're way past the point of getting increasing quality per dollar paid to executives.
- rythie 12y ago
  Performance wise it's still going to be the same company you joined. For example, Microsoft hasn't done well over the last few years, but people are mostly still going to be still impressed you worked there.

exelius 12y ago

All models are wrong, but that doesn't mean they're not useful. OP right in that individual firms are often quite inefficient, but the thought is that the market will take care of this in the long run: either a new competitor will arise and disrupt an existing market, or the shareholders will get rid of the management if the inefficiency is especially egregious. The thing about macroeconomics is that you basically get to ignore the blips because on a long enough time scale, the irregularities start to smooth out. Scientific advances happen in other fields that may significantly disrupt another industry and vault progress forward, so even if for a while things were inefficient, the market catches up eventually. Economics isn't about predicting the future with great accuracy; it's about gathering as much information as possible to be able to make decisions about what to do right now. You're playing the odds, and odds are, companies will be efficient. Sometimes they're not, which is the entire reason we have the stock market.

jasode 12y ago

The article's logic is flawed. It lists 19 inefficiencies but such a list is not as persuasive as the author thinks. What we ultimately care about is the net efficiency of the firm. Does the elimination of internal transaction costs outweigh the other inefficiencies the author points out? Internal economic exchanges as explicit transaction costs don't exist within the firm (e.g. acct dept does not send an invoice to the CTO for CPA work relating to budgeting the capital purchases of rack servers. And IT does not charge acct dept for upgrading the desktops web browser from IE6 to IE9.) To prove the author's idea, the comparison would be something like: A firm of 100 employees with those 19 inefficiencies underperforms and is less profitable than a loose confederation of 100 freelancers with their particular inefficiencies. (The 100 freelancers would also have a unique list of inefficiencies -- because nothing is perfect.) We could also itemize 20 "inefficiencies" of the biological family unit. E.g. the child learns the language of his parents -- maybe Portugese -- instead of learning the most marketable and global language such as Chinese or English. The child eats some meals from grandmother's recipes instead of eating a scientifically designed diet for peak growth. And so on. Does such a list of inefficiencies convince society that Plato's idea of taking all children from parents and assigning them government custody is a better idea? That's the type of logic I see in the article.

ebiester 12y ago
Except that individual departments do bill each other internally. It's how major corporation accounting works.
- jasode 12y ago
  Internal cost accounting, activity-based costing, chargebacks to departments for Xerox photocopies, etc isn't what I'm talking about. The CFO & acct dept does not send a literal invoice to the IT department where the CTO has his own mini accounting staff handle the invoice and cut a check back to the CFO. When corporate counsel reviews a Terms of Agreement for the IT dept's web service API, they don't send an invoice for $400/hour for "legal services". Not dealing with thousands of such "transactions" as explicit costs is one of the reasons firms exists (so the theory goes.) Internal operations do not send purchase orders to each other. They don't remit checks to each other. Each sub department doesn't have their own sales team haggling internal prices & discounts with other department heads. Etc, etc. It's the external companies' interactions that have all those costlier transactions.

tikhonj 12y ago

This really reminds me of how Yanis Varoufakis, Valve's economist, describes most firms: > Interestingly, however, there is one last bastion of economic activity that proved remarkably resistant to the triumph of the market: firms, companies and, later, corporations. Think about it: market-societies, or capitalism, are synonymous with firms, companies, corporations. And yet, quite paradoxically, firms can be thought of as market-free zones. Within their realm, firms (like societies) allocate scarce resources (between different productive activities and processes). Nevertheless they do so by means of some non-price, more often than not hierarchical, mechanism! The firm, in this view, operates outside the market; as an island within the market archipelago. Effectively, firms can be seen as oases of planning and command within the vast expanse of the market. In another sense, they are the last remaining vestiges of pre-capitalist organisation within… capitalism. The image of a firm as an island of control within a greater see of capitalism really stuck with me. I think it's a pretty good image to support Valve's style of organization. (Of course, I'm also a bit biased: I prefer that philosophy for reasons of my own besides just increasing efficiency.) The whole article ("Why Valve? Or, what do we needcorporations for and how does Valve’s management structure fit into today’s corporate world?"[1]) is well worth a read. I think it's a very constructive look at the issue that this post is talking about. [1]: http://blogs.valvesoftware.com/economics/why-valve-or-what-do-we-need-corporations-for-and-how-does-valves-management-structure-fit-into-todays-corporate-world/ http://blogs.valvesoftware.com/economics/why-valve-or-what-d...

jerf 12y ago
I actually think that's an important aspect of capitalism's success. It only (metaphorically) claims to be a good global organization scheme. Within that framework, experimentation and variation is permitted. Do you want a hippie commune? Sure. Coop? Sure. Conventional corporation? Yup. One dude in a garage? Sure. National-scale non-profit? Go for it. And they're all different organizations that can freely live and die as the situation warrants. It may not be perfect, but it can evolve without breaking. Contrast that with a State-Uber-Alles philosophy like Communism, and you get a situation where you've got what is putatively the same organization trying to organize the strategy for keeping the borders secure with nuclear weapons also trying to manage the local school bake sale. It's a fragile setup if it isn't perfect, and of course it isn't perfect. And lo, Communist governments that fall tend to fall hard and shatter their whole society, economy, country, and all in the process. All the eggs, you might say. And let me highlight the issue of resiliency and fragility as being my main point, rather than any particular local issue of efficiency on this point or that. Capitalist societies hardly blink if a 100-person organizations proves unable to make a profit. It might be locally annoying, even devastating to some people, but globally it tends to work out for the best. (And those 100 people tend to be able to move on to something else too.) By contrast a centralized command-and-control system may keep this net-negative organization going for a long time for any number of reasons, but all of them, in the end, meaning that a net-negative organization lives on, draining the society it lives in. (Bear in mind the meaning of "net negative"... the entire point is that whatever benefits it may have, it is still net negative in my discussion here.) The summation of such decisions is incredibly destructive.

michaelochurch 12y ago

Warning: this is a huge topic. For further reading: Venkatesh Rao's series on the Gervais Principle: http://www.ribbonfarm.com/2009/10/07/the-gervais-principle-or-the-office-according-to-the-office/ http://www.ribbonfarm.com/2009/10/07/the-gervais-principle-o... My series on the same issue (organizational inefficiency): http://michaelochurch.wordpress.com/2013/02/19/gervais-principle-questioned-macleods-hierarchy-the-technocrat-and-vc-startups/ http://michaelochurch.wordpress.com/2013/02/19/gervais-princ... In a picture: http://gapingvoid.com/2004/06/27/company-hierarchy/ http://gapingvoid.com/2004/06/27/company-hierarchy/ Short version: firms are inefficient because organizations are innately prone to a predictable degeneracy of the group into three sets of people, none of which is capable of meeting an organizations trifecta of demands: subordination, dedication, and strategy (working on the right things). A person who is strategic in his personal life is going to either optimize for least-work/most-comfort (Macleod "Loser" strategy) or most-yield/least-subordination (Macleod "Sociopath"). By definition, those who are subordinate and dedicated are not strategic in their work lives, and probably not equipped to run a company for that reason. Strategically subordinate but not dedicated "Losers" do the bare minimum and won't push themselves beyond how far they are managed. Among the strategic and dedicated (but not subordinate) "Sociopaths" there are some good people, but plenty of narcissists and bad actors as well. Cleaning up the messes made by their Sociopath bosses and Loser underlings are the Clueless middle managers, who are subordinate and dedicated but not strategic. For various reasons, these three sets are not good at working together but usually play against each other, with one dominant archetype characterizing the organization. (In VC-funded startups, it's Clueless; at Enron, it was the bad kind of Sociopath. In most checked-out mediocre companies, it's the Loser.) There are "good Sociopaths" out there (e.g. "sheep dogs") and I'm probably one of them, but they're often at a disadvantage against the bad ones. Often they get sick of taking on political risk (for games whose prizes cannot be enjoyed fully by the most ethical people; power of any kind is most valuable to those who will be corrupt with it) and become Losers with age. A few might be able to start their own companies, but VC won't help them because VC-funded startups are all about Cluelessness (which is why VCs have their chicken-hawk fetishism for very young, white, upper-class men). Fixing this problem is possible, even in larger companies, but very few organizations view it as worth the cost, especially in the circa-2014 startup game where companies are built to be sold, not to last more than 5 years (in which case MacLeod-style degeneracy will just be starting to impede performance, but is unlikely to kill the thing). Why? The classical explanation (from MBAs and investors and even founders) for why they keep starting crappy companies is that they can't afford the risk of alternative organization styles (e.g. Valve's open allocation) while also innovating on product, as if there were some finite pool of risk and "all of it" had to be allocated to product innovation. I strongly disagree. I think they're full of shit and trying to justify their own mediocrity. But that's what they say.

porter 12y ago

Yet another blog that I want to subscribe to via email, but can't (easily). This kind of thoughtful analysis is missing from today's obsession with listicles.

notduncansmith 12y ago
http://www.overcomingbias.com/feed http://www.overcomingbias.com/feed + https://ifttt.com/recipes/4447-rss-mail https://ifttt.com/recipes/4447-rss-mail I made it easy :)

Pitarou 12y ago

In essence, this is quite a simple matter. Here's Theory of the Firm 101 for you: In the real world, pure free market economies are inefficient. (If you think otherwise, you've never tried to co-ordinate a complex project with multiple independent contractors.) In the real world, pure bureaucracies are inefficient. (Think North Korea, or the old Soviet Union.) But they are inefficient in different ways. It turns out that we can improve efficiency if we have islands of bureaucracy (firms) in a sea of free market competition. And, since these firms get to keep some of that efficiency gain for themselves, that's what happens: one way or another, people organize themselves into firms, and reap the benefits.

mrcactu5 12y ago

why is the author so reluctant to agree these inefficiencies exist? is he afraid of biases in his own list? Also "smart sincere people" is not a notion I am used to hearing in academic discourse. If he is going to be critical, I wonder how he characterizes this notion Is it purely from a game-theory perspective?

Sniffnoy 12y ago
See e.g. http://www.overcomingbias.com/2009/12/the-smart-sincere-syndrome.html http://www.overcomingbias.com/2009/12/the-smart-sincere-synd... He also refers to "smart sincere" people as "nerds", so you can search his blog for that as well.

ttty 12y ago

TIL: "Modeling is always a tradeoff between realism and understanding."

mempko 12y ago

I am happy to see that people are slowly rediscovering socialism. However, the "efficiency" argument is one of the poorest arguments because "efficiency" has always been a hustle. There is no way to predict and even measure all the effects of an economic decision to make any judgement based on efficiency.

mathattack 12y ago

There is always a principal/agent problem in large companies. "What's in it for me?" differs from "What's in it for the company?" Much of management and leadership is related to overcoming this issue.