I just stumbled across a link to an article which made me cry a little bit. The article is The Start-Ups We Don’t Need, and basically argues that start-ups aren’t beneficial to the economy. It’s written by a Professor of Entrepreneurial Studies at Case Western, but it’s just…bad.
The basic argument of the article runs like this: New start-ups are less efficient than larger businesses; therefore, start-ups don’t drive an economy. What the author misses is that the inefficiency of start-ups is exactly what makes them great for an ailing economy.
Let’s say we’re a business and we employ a hundred people. If we become more efficient, that means we can accomplish the same thing with fewer people, which means we can cut down on the number of people. Which means that high efficiency in business leads to unemployment from those businesses. Even if we grow, our increased efficiency means we have to hire fewer people to keep up with that growth. On the other hand, if we’re less efficient, that means we can employ more people, and growth means employing even more. Granting that a low employment rate is a desirable thing for the economy, then start-ups and their inefficiencies are exactly what the doctor called for.
Small businesses also lead to a diversified economy, which mediates economic risk. We’re just starting to recover from an economic crisis caused by Too Big to Fail businesses, and there’s serious economic damage possible when a tiny number of huckster-executives decide to screw the system (some cites). Small businesses, on the other hand, don’t place the same kind of risk in the economy, and they are easier for agencies to successfully regulate.
The article also makes a random point about subsistence farming, and sweeps the crazy differences between an agrarian society and information-economy entrepreneurship under the rug. It seems to exist solely to make a vaguely racist play: “So if you want to find countries where there are a lot of entrepreneurs, go to Africa or South America.”1 and then to draw some weird and apparently arbitrary correlation between declining rates of self-employment and economic growth. Granted: people were moving off the farms and back into cities in France, West Germany, and Italy in 1953. Probably because they were bombed out of the cities and back to the farms a few years earlier.
1 Notably, there’s a Nobel prize in economics given to the man who drove the adoption of microlending, primarily as a vehicle for increasing entrepreneurship in places like Africa and South America. Which was obviously a bad thing to do, since that’s just hurting their economy worse. Stupid Nobel prize committee.
Other arguments are based on equally bogus statistics. Stats like “By contrast, companies with at least one employee that are more than ten years old account for 60 percent of all employment in this country.” really just mean that entrepreneurship isn’t common, and the big boys employ more people overall. Granted. How is this a knock on entrepreneurship? Another gem is “the cohort of new employer firms founded in the United States in 1998 employed 798,066 people in its first year but employed only 670,111 people in 2002″, which just shows that companies are growing in efficiency when they hit the 6 year mark. Oh, and 2002 might not have been a stellar economic year compared to 1998.
There is one statistic that does seem rather damning: “To get one business employing at least one person ten years from now, we need 43 entrepreneurs to begin the process of starting a company. And how many jobs will that startup have, on average, ten years after it was founded? The answer is nine.” Of course, that’s the one statistic without any evidence, cite, or explanation to back it up. Even if it is valid (and I obviously have my doubts), the whole statistic misses the point: those 9 employees are pure gravy. Imagine I could kick off an endeavor which would employ millions of people, but would close up shop in 9 years2. This would be one of those “failed” enterprises in that statistic, but does it mean it’s a bad thing to do? Of course not. Over the course of the lifespan of the “failed” entrepreneurial enterprises, we’ve relieved unemployment temporarily and have provided a unique brand of on-the-job training to people with a high tolerance for commercial risk. We’ve also dumped disproportionately more into R&D and tried out innovative business practices that couldn’t fly in a more established business. And as long as another endeavor waits in the wings to pick up the more experienced team from the failed ones, we can roll the dice again. Maybe next time we’ll be the lucky endeavor.
2 Such endeavors in our past have included the CCC, WW2, and the drive to the moon.
About half-way through the article, the tone suddenly changes and it starts to be a love letter to job-generating entrepreneurship. Which I get and appreciate, and I even support the author’s point to focus more heavily on small businesses with growth potential. The underlying tone of love-letters for venture capital firms is a bit strange and totally unsupported with argumentation (What’s the success rate and cost of the average VC firm endeavor? How do you support entrepreneurs and companies before they’re ready for VC? How do you cope with the disproportionate collapse of the VC market in a down economy?), but final point—that we need to back growing and innovative businesses instead of shotgunning over all businesses—is a fair one. The opening argument that independent businesses are somehow bad for the economy…? That’s the argument which fails to be convincing if you even scratch the surface. Probably because it’s wrong.
18 Comments
In essence, I agree with your article – startups are beneficial to the economy.
However, I do disagree with you on one thing – that somehow larger companies are more efficient than smaller companies. I’ve spent three years working in Siemens, and was loaned to several other large companies in telecoms and aviation business. The amount of HR overhead in those places is scary beyond belief.
In none of those places have I seen the efficiency of a typical startup, where one employee/founder will customarily be doing the work of at least two people employed by larger companies.
What does your experience tell you?
Completely off-topic:
“Granted: people were moving off the farms and back into cities in France, West Germany, and Italy in 1953. Probably because they were bombed out of the cities and back to the farms a few years earlier.”
You forgot the “… by us”.
@Ivan
There’s a reality to the economy of scale which more than compensates for the lower individual drive and initiative in corporate settings. Remember that productivity is measured in terms of output, and while start-ups might be driven by a John Henry, the steam hammer wins over the long haul.
Robert: start ups are not only driven by John Henry, they’re also much more likely to adopt the steam hammer. Paul Graham had it right (at least in this): if you do what everyone else is doing, you more or less get what everyone else gets. So large companies can afford to invest in “inefficient” technologies and languages like Java and C++, because what they’re likely to get is a slow, steady growth (the default result for large, stable companies). But if a start up gets their default result, that means they go out of business. Startups are by nature a “go big or go home” sort of undertaking. Thus start ups (and other small companies which are in much the same boat) are much more likely to adopt the new, productivity increasing, technology. How many large corporations are using Ruby or Grails as central technologies (not pilot projects, not someone sneaking it in, adopting these languages as the main languages for the business)?
So I disagree that startups are innately more inefficient than large corporations for those things startups can do.
I’m going to quote of of the responses on Hacker News, which reflects exactly what I thought when I read this article:
“The entire argument [..] is the broken window fallacy.”
Now I agree the article you linked to is also wrong, but not for the reasons you give here. I think it is wrong because there usually isn’t a large business that does the same thing as a startup and hence the startup isn’t being inefficient… you can’t be inefficient compared to [nothing].
The special pleading for start-ups re: efficiency isn’t really valid. I do agree that start-ups do novel things in the area of R&D/cutting edge technologies and business process innovation. I disagree that this somehow makes them exempt from productivity/efficiency consideration: at the end of the day, they’re still looking for revenue somewhere, which means they’re producing something. Or, I suppose, they’re not looking to produce anything and they are simply parasitic on the VCs, which puts them as nothing but overhead in this particular metric.
People’s gut about inefficiency in larger corporations is simply false: find some statistics/studies that back up that gut feeling and then we’ll talk. I’ve seen the US Census statistics akin to those in the original argument (I think in the Atlantic, although IIRC, it’s firm size vs. productivity, not age vs. productivity), and a quick glance through some Googleable research implies that the positive correlation between firm size and productivity is still true: cite, cite.
And it’s no doubt that I’m building the same case as the broken window fallacy. Except that it’s not a fallacy in this case: I’m arguing that if your problem is unemployment (as it is in the US right now), then increased employment is good. To create the same number of widgets, efficiency reduces employment, inefficiency increases employment. Therefore, inefficiency is good if your problem is unemployment. Where in that argumentation am I wrong? Where’s the fallacy?
I’m also arguing large businesses are more of a threat to the economy than small ones. While small businesses may not be able to leverage themselves to increased output with the same ease as big businesses, small businesses also don’t become a “Too Big to Fail” cornerstone of the market, and a crisis within a small business doesn’t drag the whole economy down with it.
The fallacy is that when startups have a higher efficiency, the GDP gets larger, creating more (financial) room for new startups, etc. That’s the point of the broken window fallacy: you forget the alternative ways the money could be spent. The other way around: you forget where the money to keep the inefficiency (assuming there is any) in existence is coming from and how it could be spent in another way.
Where am I making the start-up -> GDP -> more financial room argument, Ivo?
And if you’d like to argue that alternative ways of spending that money are beneficial, that’s fine. I was directly responding to the argument in the original post, which said that inefficiency is innately bad for the economy, therefore start-ups are innately bad for the economy. In the article I’m responding to, I don’t see a direct case for a particular alternative—quite the contrary: it ends with a sales pitch for certain start-ups.
BTW, people are aware that I talk about efficiency in only two paragraphs, right? There’s a whole rest of the post kicking around, too. So to say “The entire argument [..] is the broken window fallacy.” sounds a lot like someone came up with a pithy snap judgment after paragraph 3 and stopped reading.
I think the failure in understanding is on your part. Re-read the first paragraph of “The Start-Ups We Don’t Need”. The whole article is an argument to not just throw government money at anyone who wants to start their own business.
Specifically, “…the typical entrepreneur is very bad at picking industries and chooses the ones that are easiest to enter, not the ones that are best for start-ups. Rather than picking industries in which new companies are most successful, most entrepreneurs pick industries in which most start-ups fail. So by providing incentives for people to start businesses in general, we provide incentives for people to start the typical business, which is gone in five years.”
Also, the Africa comment was clearly saying that Africa is a place with lots of unemployment and low wages, which is the type of environment in which lots of startups are found. Contrast this to America with relatively high employment and wages where there are fewer startups due to workers having more to lose if the startup fails (they lose the steady income they had when they were working for someone else).
There are many more data points in the article that make me think that, all in all, you need to re-read the article slower and understand what it’s actually saying. I’ll admit that in a few places it seems to cherry pick numbers or leave gaps (e.g. if companies more than 9 years old supply 60% of the jobs and companies 2 years old supply 1%, what about the ones in the middle supplying the other 39%? More specifically, what about the ones that survive the dreaded 5-year fail point? Is a better number something like ‘companies over 5 years old provide 85% of all jobs’?). And yes, his choice of years 1998 and 2002 are probably not the best to use for a trend line.
In any case, I found his point to be very clear: don’t waste money on startups that have a very high chance of failure. This is nothing like saying that startups are bad and to not spend any money on them. This whole argument is a straw man and I hope it was unintentional (especially the claim of racist remarks).
I actually agree with the point he says he is making: reread my last paragraph. It’s the statistics and his analysis of those statistics that I’m not a fan of. Which it sounds like you aren’t, either.
If the article’s content and its frame were in agreement, I would have written a very different blog post. But it seems like it opens and closes with one set of assertions and has a very different argument in the payload.
Robert,
I’ll agree with Timmay’s points, but not necessarily his conclusion. My problem with the original cite is that it recommends putting the government in the position of picking winners. As I understand Dr. Shane’s point, he is arguing that the government should eliminate general aid to small and start-up business and focus only on businesses that are likely to succeed. I won’t quibble that this would be a more efficient investment strategy, but I will strenously argue that this is not the purpose of government assistance. In this sense, I agree with you (and Keynes). Fair government has a moral imperative to aid all its citizens, not only the ones who don’t really need the help.
The argument in the initial referenced article seems a little off. It’s almost like saying “babies are bad because very few of them grow up to be executives, and we know that executives really drive the economy, so we shouldn’t support just any baby”. Realistically since there is no established path for growth, you could never actually know if one attempted startup was going to fail or if it was going to spawn the next Walmart. Business is irrational, primarily because if it wasn’t everyone would be rich (and thus not be rich anymore).
In our rush to create larger entities that are meaner and leaner with niffty money saving tricks like just-in-time inventories and other economically sensitive attributes, we’ve been gradually eating away at our overall stability. We’re bending towards short-term efficiencies at the cost of long-term frailties. Any economy with a mass of little companies is likely way more stable than one with three great hulking entities, particularly in this day and age when the executives bend so easily towards downsizing in order to save their own bonuses. More companies, more evenly distributed over a wider range of markets is inherently more stable. And “stable” is a long-term thing that can easily be obscured by bad statistics gathering.
Larger companies are *much* more efficient when staying the course.
Smaller companies are *much* more efficient when changing directions.
It’s a simple matter of economies of scale. A large company, is, well, large; you can’t turn a supertanker on a dime, but it’s hugely wasteful to send its cargo on motorboats instead. To call one universally better than the other is just silly.
OTOH, I wonder how knowledge work vs manufacture affects this. Consider the size of Craigslist based on various metrics, f.ex.
@Timmay
Just re-reading comments, and the Africa thing finally got my goat enough to say something.
The comment I called out as racist (or at least hemisphere-ist) is not any more valid because you repeat it. Comparing two kinds of American businesses on analogy with two entirely different economic environments is just stupid, and conflating subsistence farming in a truly desolate economy and entrepreneurship in a first-world context is either actively or passively misleading.
Assuming that a Case Western professor should know and behave better than to actively distorting the conversation with bogus arguments, we have to assume it’s a passive play. The play is (actively) a kind of hyperbole, which contrasts America with Africa and South America (rhetorically conflated into a single “other”). Hyperbole works by taking a slight difference and blowing it up into a huge one. So the slight difference is the difference between big companies and start-ups: the huge one is the difference between America and Africa and South America. So the difference between America and Africa and South America is huge: America GOOOOOOOOD, Africa and South America BAAAAAAAAAAAAD.
And that’s just plain ol’ structural racism. The author probably didn’t even realize he was playing off of racism to make his point.
Robert,
I think you’re reading more into his words than what’s there. The way I read it, he claims that A) entrepreneurship happens most in a certain type of economic environment, and that B) Africa and South America are in that type of economic environment. Thus, C) if you want to see high percentages of entrepreneurship, look at Africa and South America. (Technically he claims A and presents C, leaving B as an obvious “between the lines” statement. Although, he does this sort of thing throughout the article, and I must trust that the “between the lines” items are true; I wish he was less sloppy and had more citations.) I see nothing -ist here.
He does go on to use the terms “rich” and “poor”, but I don’t see him ascribing any “goodness” or “badness” with any country. If it’s a fact that one country has historically had more economic growth than another and has a higher GNP, then that is simply a fact. This does not make one bad and one good, they are simply different. This is an economics professor talking about economics. He is using data to make and support claims. In a pure economics discussion morality should be (and has mostly been) left out of it entirely. Another example of this is his discussion of the fact that the government throws away a lot of money by letting it be invested in startups with a high chance of failure. He doesn’t bother discussing good and bad, God and Satan, or democracy. He does briefly mention fairness (and seems to indicate that what he feels is more fair is to give money to those best prepared to use it well). He talks economics and he’s trying to maximize the ROI on money the government invests in startups. How well a program maximizes ROI is his only measure of “good” or “bad”. His only bias is toward those who have a better chance of success.
So now I’ll point out again that I think his article is mostly crap. His aim is to remove money from very small startups and instead give it to well established startups that look like they will do well and to those who already have a track record of making successful startups; that’s hardly a group of people that needs assistance, now is it? He also doesn’t seem to fathom that maybe the answer is to put more money into all startups; maybe the current level of assistance isn’t sufficient to get startups over their major obstacles. Maybe the answer is to keep total funding the same, but only allow X slots per year, so that each funded startup gets more help and you don’t flood the market with too many of the same thing.
The whole article smacks of using data to support a pre-conceived notion rather than let the data show you where the problems are and trying to address those problems. My gut feel is that startups are like other organisms and they require many bad ones and false starts in order to create a healthy and diverse population of corporations. But, I’d find data and let it show me the truth, instead of the other way around, before submitting an article about it. :)
@Timmay
So you’re granting that it’s legit to conflate South American and African subsistence farming and Silicon Valley entrepreneurship?
@Robert F.
I think this is another example of people in this thread being self-centric. The article is not about Silicon Valley or tech start-ups at all. If you think that, then the conflation is happening in your head with the definition of ’start-up’. Here is how the article defines start-up: “The typical start-up is a company capitalized with about $25,000 of the founder’s savings that operates in retail or personal services.”
Realizing that fact should make it easier to see that there is no conflation occurring between Africa and Silicon Valley. Here is the Africa & South America quote:
“…as the source of economic value shifts from activities where self-employment is more common, such as agriculture, toward activities where self-employment is less common, such as manufacturing, the proportion of people running their own businesses drops.
So if you want to find countries where there are a lot of entrepreneurs, go to Africa or South America. ”
All he is claiming is that Africa and South America are currently in the agriculture stage, thus you can find lots of start-ups there.
Can you argue that he is conflating Africa’s economy with America’s? I don’t think so, because he points out that they are in different stages of economic development. I think he brings up the whole thing in the first place to try and discredit the ’small business as a cornerstone of recovery’ idea; after all, if all it took to have a large GNP was lots of start-ups, then Africa and South America would be in a different position. But I honestly don’t know. His arguments are too schizophrenic for me take seriously.