The five smallest industries by firm size

For the U.S.:

Fishing, hunting, trapping: 3.1 workers on average

Building construction: 5.5

Real estate: 5.9

Funds, trusts, and other financial vehicles: 6

Repair and maintenance: 6.1

Time for some creative disruption, people…

Those figures are from the new and excellent Big is Beautiful: Debunking the Myth of Small Business, by Robert D. Atkinson and Michael Lind.


> Time for some creative disruption, people…

Yes there are certainly many opportunities to create massive-revenue-generating global platforms and turn more knowledge workers into commoditized serfs.

You mean like Uber creating hundreds of thousands of one man firm's.

Seriously, are there really more hunting/fishing etc guides than Uber "drivers", the best example of a one man firm.

Uber has made a business startup much easier, though they keep increasing the hurdles to starting your one man business, but they have spawned other firms promoting one man businesses, e.g. Lyft, Taskrabbit, etc.

Each of these business have no paid workers.

My new app Worms will totally disrupt the baitshop industry.

Real estate: 5.9

This is an odd one. It means that people know their suburb well, but it also means that they rarely get repeat customers and hence no incentive not to rip everyone else off.

This seems ripe for some sort of on-line electronic platform. Craig's Property List perhaps. Motorcars have done it. Why not houses?

I question the real estate stat, unless Century 21 / Caldwell's etc (real estate brokers in the USA) are in fact small family franchises and not national companies.


while there are large national firms, RE/Max, Century 21, and large state or regionally based firms such as McEnearney Associates in Northern Virginia, firm size may be calculated on the ratio of agents to a given broker. In most states agents cannot operate independently as must be associated with a broker. A large firm will have many brokers each with a number of agents.

For me that number is surprisingly high.
(But maybe my english is badder than I thought and "real estate" doesn't include "landlords"?)

In times of low interest rates home ownership is advertised as a good, secure investment for everyone. Just put on some debt, buy a condo, let it for some years ("pays for itself"). Consequently prices have risen overall, often forcing newer entrants to sub-let (even though they originally wanted to owner-occupy) because banks won't finance it otherwise. In the end we get lots of one-unit-landlords.

I want to chip in on "the myth of the small business":

I've come to doubt that increasing home-ownership rates should be a policy goal (as it usually is), as being a low-unit landlord is usually an inefficient business by itself and a guarantee for low quality services for the renter.

Germany, Austria and Switzerland have - compared to other countries - pretty low ownership rates, but sadly these countries believe in the myth as well and engage in increasing that rate. Effectively giving us more inefficient businesses and more barely stable mortgage-plans.

Apropos of nothing, I saw that German cities made a recent list of "most inflated real estate cities in the world" which shocked me, as the US cities were not on this list (NYC, San Fran were close but not in the 'most inflated' category). The criteria was growth over the last few years compared to salary growth and such, and it was shocking since it's well known that German cities have lots of renters and houses are not that expensive, compared to other cities in the world. I guess if you massage the data enough, you can find any stat you want?!

It depends. Berlin? Cheap and easy to find housing. Munich, Dresden? Very much inflated until you leave the center city, but Germany doesn't have quite the commuter culture that America does. Yet.

Housing in German cities was extremely cheap, but has increased a lot in the past years. At the same germany received a lot of immigrants, but we are not allowed to see the connection here. Also lots of foreign investors.
But still German housing is cheap compared to other western european countries.

It means that people know their suburb well, but it also means that they rarely get repeat customers and hence no incentive not to rip everyone else off.

Which, in turn, ruins their reputation.

Real estate agents before maybe the 80s depended on repeat business and referrals. But IBM was a big part of the business/job culture then. In a decade my job moved me 6 times, and if I had owned a house, that number might have been only 4 or 5, but my employer would have paid for each real estate transaction.

But my working class e.g. factory worker peers and parents often moved once every 5-10 years as they gained seniority and their family changed size and composition. The small starter house replaced by a spare house for a couple of kids, then a suburban house with llots of land for parking the multiole cars, boat, RV, plus yard for recreation.

Redfin comes close; their agents are salaried, not commissioned.

How are these averages being calculated? Just taking the number of workers total and dividing by number of firms total?

Because in a lot of these industries, I can imagine that the average firm is quite small, but that most of the work and output is due to huge, well-known firms; you don't even need to assume that huge firms are more efficient than small ones to get this. It's totally reasonable that the average ice-cream company is like 10 people, but a handful of gigantic companies produce 95% of the ice-cream anyway.

(The average family is much smaller than the expected number of siblings that a random person has - because a 6-kid family is 1 sample in the first calculation, but 6 samples in the second)

If we look at total wealth under management - if I select one dollar of managed money at random - what is the expected size of the firm managing it? Same for the others - if we select randomly from the output of the industry (weighted by exchange value) and look at the size of the firm that produced it, rather than select a random firm and compute its size - how would this list change?

+1. Exactly. Is this the mean or the median? And does it control for market share?

Having the distribution would be more useful.

I think that it doesn’t control for market share is the point, or at least one of them.

I’d need to look closer at what was included in each category but the quoted firm sizes look reasonable.

If hunting is mostly guides you’re looking at a lot of sole proprietor no employee set ups. There’s some big boats mostly in the North Pacific but few east coast fishing boats (Gloucester, New Bedford) have more than five or six people on them. There’s some multi boat companies but still a lot of one boat owner operators too.

The retail real estate places are franchises and the agents are all or almost all contractors.

Something like this I would expect to be Pareto.

If for creative destruction you imply an accompanying increase of the size of the firm, not at all, Tyler. The growth of the large corporations starting during the Reconstruction and continuing in the Gilded Era is a major reason of the political dumbness of the American citizens. Until the War between States Americans were mostly small enterpreneurs, farmers, “mechanics”, traders. They had to compete in the marketplace everyday, and that bred a healthy dose of individualism and skepticicism of the State. Once they had a fixed salary, most stopped thinking, they became sheeple, and accepted first TR imperialism, and later FDR welfare, and it has been downhill ever since. The proles is now free, just like the animals.

No, let the web destroys the boundaries of the firm, let’s revert to a real world of personal responsibility, where you eat what you hunt, not what your politician manages to steal from the hunters. Time of creative destruction for the huge, chronist, large corporations.

Right you are Massimo. Also Progressivism (political movement) at the time of the Gilded Era set the stage for TR Roosevelt anti-trust busting and FDR Big Government later. However, what you speak of so well is a function of efficiency: since machines are more efficient when bigger (read V. Smil's book on diesel engines), and factories with economies of scale get more efficient over time, then indeed creative destruction accompanies an increase in size, with the negative consequences that you mention. John D. Rockefeller was a monopolist who engaged in dirty tricks, but, at the same time, Standard Oil really was the lowest cost producer. It's an age old problem, akin to the "tyranny of silicon" in Silicon Valley process manufacturing.

Software is eating the world, Ray. In software we don’t know how to get economies of scale.

To be sure, the large companies: Microsoft, Google, probably others, have internal tools which give their employees a leg up, but the “world” generally copies those technologies after a few years.

Further, software problems today require enormous amounts of labor (staff) to ship, but there is “no reason” other than trust why this couldn’t be farmed out to independent contractors.

One note: some problems today are becoming so computationally complex that there are economies of scale (eg: AI problems do not converge sufficiently quickly without a staggering amount of capital), so in a few areas (which may be the most important) we may be heading towards a parallel of the industrial revolution. Cloud computing, however does democratize this effect, somewhat.

Excellent comment.


+1. Insightful.

The book Tyler linked is basically a disagreement with your argument "American democracy does not depend on the existence of brave bands of self-employed citizens. [...] The idea that self-employed citizens are the foundation of democracy is a relic of Jeffersonian dreams of an agrarian society." It may be worth reading if you want a counterpoint.

Good point, Adam, when I read the teaser in Amazon I realized it and I decided to read the book.

Firms are not equal. Perhaps a 5.5 average people building firm is optimal while a 100 people bank or airline are very inefficient.

Also, in this day and age of outsourcing and independent contractors, do these numbers convey any significant information?

On a second thought, this book is between 20 and 30 years late. Today, outsourcing provides lower operating costs and increased flexibility to companies, and revenue per employee charts look awesome. In the case of air carriers, the new low-cost ones outsource a lot of functions and are more efficient than legacy carriers that have employees for every function. Call-centers anyone? "Smaller" is more efficient.

So, what's going on here?

Absolutely. I did not read the book, but it does not make a lot of sense, if it just says that “big is beautiful”. The globalization has increased the diversity of comparative advantages, and the web has drastically reduced transaction costs. The result has been a decrease in the average size of the firms, that now focus on the product/services and specific value added step where they enjoy a competitive advantage, letting the market coordinate the “ecosystem” to bring the finished product/service to consumers.

The only place I see an advantage in being big, and even an increase of the value of the advantage in the last few decades, is in political connections. I think “big is beautiful” does indeed apply, if your business model is to corrupt politicians through political donations to get protection, favourable regulations and state contracts.

" Today, outsourcing provides lower operating costs and increased flexibility to companies, and revenue per employee charts look awesome."

Of course the flip side of that is a lot of low skilled employees working cheaper at small service firms. Like international trade it's a total gain, but more than the total gain goes to a fraction of the workers and the other fraction has a small loss.

The current system is having adjustment pains with that type of bifurcation.

In a previous post, Nassim Taleb was described as urging people to start a business, but here it seems that most small businesses fail and are not particularly important for American prosperity anyway. Perhaps it is more sensible for a great many people to work as employees to large businesses than to start a small business that will probably fail; and better for society if they work as employees too.

All large businesses were once small businesses... indeed Google, Amazon etc were small businesses just 15 years back!

So we should be fostering small business youre saying? Because they grey into the much more important large businesses?

Well, they's not clear from the data here... even if most fail, the expected return could be positive?

I really dislike concluding that one shouldn't start a small business based off the data that says most new small businesses fail. I'd bet a large number of the failed businesses had incompetent owners. You'd be surprised how many business owners have a lack of knowledge in basic accounting, marketing, and micro economics.

Pizza shops probably account for half the failures. Everyone wants to own one.

The end of retail has meant the end of entrepreneurship for most people. It's the small business owner who is the most devoted member of the Chamber of Commerce. At the beginning of this month, a family member closed the clothing store he has owned and operated for decades because the mall where the store is located is dying. And this is not Michigan. Cowen must appreciate that the "disruption" he worships is changing America in ways that could forever alter American entrepreneurship, and not in a positive way. I used to comment that not everyone can sell insurance or real estate. Well, they can't, not with the "disruption" taking place in those two occupations. What, exactly, are the so-called innovators in Silicon Valley doing? Facilitate. No, they haven't created new products or services, rather they facilitate the use of existing products and services. It's not as though the boy wonders invented advertising or the car. What the boy wonders have done is contribute to the mental health crisis in America, not to mention the political crisis. I've commented before that one day people will look back on this era of "disruption" and ask, "What were those people thinking?"

Etsy has turned any number of people into small retailers, without the need for a space in a mall.

If your business is based on one single company platform. You are not a retailer, you are a serf.

That's like saying mail order firms really work for the Post Office.

I agree with Tom that platforms have opened up a wide array of small business. Many are just (is that unfair?) jobbers breaking late lots into small and reselling, but there is a lot of original work going on as well.


If your business is based on one single location. You are not a retailer, you are a serf.

It's generally a worse experience though. Those sellers are incredibly vulnerable to things like chargebacks, shipping fraud, or competitors gaming reviews/causing listings to be removed. The amazon problem-unscrupulous sellers diluting results with mass produced Chinese crap- is an issue too. And platform vulnerability is a real thing; Patreon recently had a big issue which they tried to raise fees on the lowest tier of rewards, causing a huge loss in profits for many creators. They recanted, but you can never guarantee they won't try to screw you over again.

The net is increasingly becoming hostile to those small owners.

I pre-ordered this awhile back when I saw one of the authors on Twitter making claims that he was selling as original but are actually quite standard in the literature. I will be interested to see if he is more careful in the book.

@Ryan - I think you're onto something: a sort of "Great Stagnation" in original ideas!? I've noticed this too: authors making bold claims for publicity that are not warranted by the facts. Many examples in my mind. In fact, the real novel ideas have yet to be explored, like IP as a lever for increasing Total Factor Productivity, the elasticity of invention (whether it's easy to retrain a hamburger flipper to become a rocket scientist--Gladwell's Outliers book was a step in the right direction, see here:, whether we really have a Great Stagnation and so on.

Large size firms are not inherently more efficient than a small firm. A large firm gains from comparative advantage and economies of scale.

However, a small firm can offset that by judicious use of contracting out the services it doesn't need to do internally, janitorial work, accounting, legal, security, advertising, etc. So, what's left is economies of scale. I suspect that the industries with lots of small business don't currently benefit from economies of scale.

With a small firm, it is entirely possible for one owner / manager to exercise supervisory authority over the whole. There are far fewer places for slackers and deadwood to hide, and the principal agent problems are less pathological.

Yes! Organizations leverage human potential. Up and down.

Stansbury and Summers, On the Link Between US Pay and Productivity (February 20, 2018): Some surprising and some not so surprising findings.

There is plenty of opportunity for disruption, but I don't think firm size is the best metric... in many instances the disruption will involve moving production to another sector.
For instance, in building construction, engineered/modular/prefab firms will replace many construction workers with a manufacturing process. So manufacturing firm size will grow, while a much smaller construction sector will have many small firms with specialized craft skills.
In repair & maintenance, technology will continue to reduce labor demand. For instance, electric vehicles have many fewer moving parts, at lower temperatures and forces, with smarter control systems than internal combustion vehicles. When your car drives itself to the dealer at 2AM so a robot can swap out a defective module while you sleep, that will likely be included in your subscription price, so will be counted as part of the auto sales sector. Auto repair shops will be boutique firms catering to collectors of classic vehicles.

It is already happening.

How many small engine repair shops do you have in your neighborhood? Probably none. At one time there was one every few blocks, busy keeping mowers and other small equipment working. Their competition wasn't a better way to fix things, but that the cost of the equipment decreased. We have a rule of thumb that a repair more than 60% of the cost of new doesn't get done. When the diagnosis costs more than that in time, better buy a new one.

We are seeing that in surprising ways. Commercial restaurant equipment, those stainless steel prep tables are not worth repairing. A repair is over 60% of new, so we tell them to throw them away and buy another cheap one. Budget for it every 3-5 years.

Creative destruction in this case means someone will gather them up and ship them to some low labor country where they are torn down and rebuilt.

Same with appliance repair. Auto repair as well. You get a 5 year warranty, 5 year finance. A transmission or engine repair isn't worth doing. This is already the case with the cheap buzz boxes. There are a few general auto repair shops about, simply because vehicles are too expensive to throw away, but there are failures where you will be more than 60% of value.

I think the market is telling you that manufacturing is massively more efficient than repair.

Maybe 1/100th of an hour of labor in a new lawnmower, vs 1 hour for a repair (plus very inefficient redundant storage of spare parts all over the place).

As an example, when I shattered my phone the repair was quoted at $100. Probably a fair price for their labor, rent, inventory .. but also a strong signal to buy a new phone. Which I did.

Even within repair, re-manufacturing has shifted much of the work. I had a bad rear, non-drive wheel bearing. Instead of taking the hub apart and replacing worn balls or races, I just bought a new hub assembly. A rebuilt cylinder head for that car complete with the valves and overhead cams costs $450.

Vend, don't mend.

When labour was cheap and machines were simple (and simply unreliable), then repair was king. Television Repairman was a thing.

Ain't the world we're in now.

It seems like this is another indicator that change is happening faster than we inherently realize. There's been a push for a Universal Income, which I think is a poor idea.

I believe that a Universal Income will increase the number of non-workers. The group of new, non-workers, particularly the young men will tend to boredom and then unrest. Furthermore, even the apathetic non-workers will tend to less exercise and less in-person socialization. Work fulfill more than just an economic function in our society. It's a productive creative outlet and key to life fulfillment. It forces social interactions and pushes people out of their comfort zone. It rewards both hard work and ingenuity.

In addition to the effects on the non-workers, I believe that the workers and income earners will resent the large tax burden they will bear. They'll resent all of the people living "care free" while they push the boulder forward every day. I think this will result in a highly divided society. The privileged and hard working "tax payers" and the restless non-working.

On the other hand an expand EITC, which only rewarded some relatively large amount of work (1,000 hour per year or so), would largely perform the same function without quite the same divide.

@JWatts- Re Universal Income--it all depends on the price. Amazingly, a guy at the conservative Hoover Institute a while back came out in favor of UI (as did TC, who later recanted) because if you make UI low enough, and if the people who clamor for UI promise not to clamor for more money once UI is implemented (a big 'if', I admit) then UI is actually *cheaper* than today's welfare state. The Hoover guy showed that at a price of around $6k a year, UI is actually cheaper than today's Welfare State, and thus he was in favor of it. I personally am in favor of UI, since as a Trustee Baby (and also I inherited a bunch of money from my going senile Greek uncle) I would get money from the US government whereas today I get nothing since to get welfare, free cheese, etc, I have to declare I have no assets, which I cannot presently do.

Also I have to add that my anecdote contradicts JWatts' well-known observation that 'work makes you better not bitter' since I am personally happier now that I am without a job and partly dependent on others generosity for money rather than independent and working all day, despite the fact back then I was making six-figures in salary. Turns out I take work too seriously and I'm work allergic. It all depends on the person I guess, but like Tim Ferris I prefer a four hour rather than 40 to 80 hour workweek. Your mileage may vary, but I bet most people would rather take a lump sum gift of the equivalent of working and saving for the next twenty to forty years rather than actually working and saving for the next twenty to forty years. Also, for you guys and gals that are single, working for The Man and climbing the corporate ladder will not get you Darwinian biological success, aka laid, if that's what you want. Living in the Third World as a trustee baby will, trust me, baby.

All this makes me desire an expanded EITC over UI to an even greater extent. Rich non-workers shouldn't get money from the working class.

Work is neither a creative outlet nor a key to life fulfillment for anyone but a relatively tiny knowledge class, and even then they suffer ennui.

The EITC approach would be worse.

1. It ends up the government subsidizing business, who will lower wages to match the EITC contribution.
2. There may not be enough businesses at all to cover the work requirement. Part of the reason UBI is discussed is the expectation that total job creation will decline because productivity will not create enough jobs to balance the ones destroyed.
3. There will be a strong urge to create mandatory make-work programs to make sure people earn their credit. However, the government will have no real market pressures nor desire to make these beneficial jobs, and it could become eerily similar to work in prisons.

In my industry it is either 3-5 people or 100 people. The upper limit is usually determined by the size of the market.

The creative destruction happened already; Wall Street money was showing up buying these firms, but they didn't do well.

There is a reason. The business is very complex. For every 5 people you need a 6th that is the most experienced with business acumen to keep the crew running.

One worker can put you out of business. My competition is often large firms who hire mechanics, and have a sales staff going after national accounts. They do ok, but they can't get the best people. They do best with close proximity, otherwise the travel and management costs get out of hand.

The management effort required is extraordinarily demanding. I've worked with large firms, and they specialize in a narrow slice of the market. They have people qualified for that work, and do it well.

Large companies who need our services cut costs by simplifying the requirements. They train their own staff for the routine maintenance, and for major repairs have tight control and easy time constraints.

I don't know about those other markets, but I suspect there are similar exigencies. If you look at markets where there is high complexity and little rote repeated tasks and large firms serving a wide market, they either have extraordinarily high prices or sub out to smaller firms. Think IT consultancies. Vertical markets prevail, and the costs are very very high. They need to be.

Residential construction in large markets is defined by large developers who hire multitudes of small contractors. Depending on the task, the contractors are 3-6 people. They typically provide labor, tools and fasteners, the bulk of the material is provided by the developer.

Lots of reasons. Employment laws and costs are cut, inherent production incentives, costs follow demand, ie. if it is slow no layoffs, severances, paperwork for UI, etc. Simply no one to pay.

No need for the expensive purchase and maintenance of tools and vehicles.

The work is divided into narrow sub trades, very efficient. Insulators, vapor barrier installers, drywall boarders, mudders, then painters, then finish carpenters, then cleaners and touchup.

As efficient and cheap as built in place construction could be with the current methods. The creative destruction has already happened.

Not to mention, from some contractors I've used recently, if one guy is not banging, or is a screw up, they can blow all the profit on the job in one afternoon. Managing more than 5 or 6 people on a team just runs up the costs and chances of screw ups. It's the "Band of the Hand" concept used by the military. If you grow, you have to grow the number of teams rather than the size of the team.

Similarly with auto repair, giant facilities cost money. Not to mention, a lot of mechanics are not salaried employees anymore. They are contractors. And the one guy roaming diagnostician is also a growing field. It takes a lot of expensive scan tools, oscilloscopes, connector kits, etc. these days to ferret out the real problems. But the guy gets call, shows up, finds the problem then the other shop guys do the repair.

I don't see where most of these firms are really prime targets for creative destruction. How much is gained by wiping out the local bait shops?

Reading the Amazon blurb, I 'm not sure how "excellent" this book is. This line stuck out to me, " Small businesses are not systematically discriminated against by government policy makers." No, it's not systematic. It's just that small businesses lack the resources of big businesses to meet these requirements.

Likely the reason why these segments are as described is tax law.

There is no inherent advantage or efficiency of scale to be had. The skills are local and transferable.

What I see happening is simple arithmetic. In Canada we hit a 50% marginal tax rate somewhere around $60-70k. A small business corporate rate is with all fees and taxes in somewhere around 20%. With skills, a bit of acumen someone can set up and increase their wages.

The hunting and fishing workers are mostly guides and outfitters. They are essentially selling an experience. Aside from a few lodges, attempts to do that on a bigger scale have mostly failed because it changes the very experience that is the product.

It should also be noted that you can increase the pump, but not the well. A giant efficiency boost in commercial or outfitted fishing doesn't create more fish...

Very different from developing countries, where the mom and pop shops in the retail sector dominate. While large as a percentage of employment, in developing countries the average firm size in the retail sector is small. Stores tend to be smaller, not the CVS, Wholefoods type of stores...

In construction 5 full time staff makes sense even for firms that can scale up. owner, admin (often the owner's wife), a few full time project managers/superintendents and then hourly labor when needed. For big projects you'll have a large GC with several layers of subcontractors, with many small companies doing a lot of the work in the bottom layer, and most importantly filling in the gaps in a pinch. Command and control doesn't work if you need something immediately, GCs are usually reluctant to take on the risk of direct work, and if they do have laborers their scheduling is rarely flexible enough to fill in at a moment's notice. Being able to call a bunch of small guys can save your ass if you find someone to take care of something unexpected that's critical path.
Pre-fab will take some of this away but we'll never get rid of the unexpected.
Also, are we ever not going to need handymen for little stuff?

Some people do not want to work for "big" companies, even if they make a lot less money.
But, how do you measure that?

If you want an education in people and government, start and operate a small business.

"Irish democracy" may be more prevalent than you think.

Another example:

Here are the largest industries by firm size: 1) Hospitals - 1736 2) Central Bank - 1599 3) General Merchandise Stores - 357 4) Pipeline Transportation - 220 5) Air Transportation - 173.

I didn't know what to expect, but Central Bank is a little surprising.

The blurb is a little misleading. Big firms have been progressively reducing headcount for years. Plenty of small businesses are in sectors where there is intense competition and the founders have limited skills. The most valuable businesses in terms of growing jobs are usually set up by skilled managers who spot a niche that their large employers neglect to target because the profit potential appears to be too small for them. Amar Bhide is the authority on this.

Comments for this post are closed