6 ms·

Start-ups account for all of net new jobs creation

15 points by c1sc0 16y ago

cperciva 16y ago
Alternative title: How to coerce statistics by defining variables creatively. The author defines a "startup" as a company which has existed for less than 1 year and an "existing firm" as a company which has existed for 1 year or more, and compares the net jobs created by startups against the net jobs created by existing firms. Consider a startup company which has 6 employees and goes out of business after 20 months. It has provided 10 person-years of employment, but its net job creation is zero -- two years after it was founded, it isn't contributing anything to the employment numbers. Under the definitions used in this report, however, that company is counted as a startup with a net 6 jobs created... AND as an existing firm with a net -6 jobs created. This failed company which has no long-term effect on job creation is counted not once but twice towards the job-creation superiority of startups over existing firms -- and the more startups fail, the more skewed the numbers will be! Come on guys... startups are important enough in job creation without fudging the statistics like this.
- krschultz 16y ago
  Very true. To build on this: -They define any business under 1 year old as a "startup". Is a new gas station a startup? A new cleaning business? A new home healthcare franchise? These are all new businesses run by entrepreneurs, but they are not "startups" in the way we generally refer to them on HN. I know several family members that have these kinds of franchise businesses and they never refer to themselves as owners of startups, but as owners of small businesses. Small distinction but lest the VCs and hackers say they create all the jobs, realize that the companies doing "established business" they so look down on are counted amongst startups in this data. 2) Common sense defeats their numbers. Who starts a startup with less than a year runway? That would be ridiculous. If they change their numbers to count startups up to age 3, what would this look like?
  - c1sc0 16y ago
    "These are all new businesses run by entrepreneurs, but they are not 'startups' in the way we generally refer to them on HN." Isn't that just a matter of semantics though? Would s/startup/newly incorporated business/g do? I agree that their definition of startup certainly skews the statistics. More interesting would be "Sum of total job-years by startups vs. established companies".
- notahacker 16y ago
  Agree that the data itself doesn't really support the implicit policy recommendation of encouraging startups rather than established companies, but the basic finding that firms in their first year account for ten times the gross new jobs created by firms in any subsequent year and thus after a year companies on average lay off more staff than they hire is pretty interesting. It would be much more interesting to see what the stats looked like excluding firms that fail in a given year (ie. how many net jobs do firms that survive into year five create in their fourth year)
  - cperciva 16y ago
    It would be much more interesting to see what the stats looked like excluding firms that fail in a given year (ie. how many net jobs do firms that survive into year five create in their fourth year) I downloaded the BDS data from the US census website and arbitrarily picked companies which were age 0 in 2000 to look at. Of the 486k "establishments" (approximately equal to companies), roughly 50% survived 5 years. These 486k establishments created 3.1M jobs in 2000, and 1.6M jobs -- roughly 50% -- were lost over the following 5 years due to establishments closing, for a net gain of 1.5M jobs. Meanwhile, 2.4M jobs were created over years 1-5 while 1.7M were lost (at establishments which did not close), for a net gain of 0.7M jobs. Over a 5 year timescale, it seems to me that startups (in the general sense of "new companies") outperform existing companies in job creation by a 2:1 ratio -- still impressive, but also rather more credible than the article's claim.
    - c1sc0 16y ago
      This is why I love hacker news! Thanks for crunching the numbers!
- trevelyan 16y ago
  It's not actually clear, Mark Thoma just blogged about this w/r/t an NBER paper released earlier this month on this same question: http://economistsview.typepad.com/economistsview/2010/08/who-creates-jobs-small-vs-large-vs-young.html http://economistsview.typepad.com/economistsview/2010/08/who... Conclusion seems to be that job creation happens most heavily in young firms which manage to survive a few years, but that the high rates of job destruction pretty much wash out small businesses gains otherwise. Would be interesting to hear how VC plays into the equation.
robertgaal 16y ago
Argh PDF alert.
- c1sc0 16y ago
  Or do you mean scribd alert? ;-)
terra_t 16y ago
This kind of study provides all sorts of opportunities for shell games, since the classification of "startup" vs "established", or "new" vs "old" or "big" vs "small" often obscures the real issues. Just to take an example, imagine the person who cleans the lobby of a Fortune 500 headquarters lobby. This person could be a W2 employee of the big co, or could be an independent contractor, or could be an employee of a small firm (say 5-10 people) that does the cleaning, or could be an individual or small firm that owns a franchise for a cleaning services business -- and thus be something that's some kind of symbiotic association of a big co and a little co. The work that gets done is substantially the same, and it's done for the same reason, but the benefits, tax consequences and how it shows up in economic statistics is entirely different.
ynniv 16y ago
Better title: Companies that don't go out of business create the most jobs. An important caveat is that existing firms are so diverse that it can be misleading to think of a “typical” firm. For example, there are two simple categories of existing firms: those that go out of business (Deaths) and those that continue (Survivors). On balance, existing firms lose more jobs than they create. But once Deaths are set aside, Survivors usually create more net jobs than startups do. I think they just proved that their own title is creative accounting. If the goal is a policy change, this report shows that we can reduce unemployment most effectively by keeping established companies open.
zacharyvoase 16y ago
I’m reminded of James Tauber’s blog post on ‘job creation’: http://jtauber.com/blog/2008/11/12/x_will_cost_y_jobs/ http://jtauber.com/blog/2008/11/12/x_will_cost_y_jobs/

Start-ups account for *all* of net new jobs creation

15 points by c1sc0 16y ago

Alternative title: How to coerce statistics by defining variables creatively. The author defines a "startup" as a company which has existed for less than 1 year and an "existing firm" as a company which has existed for 1 year or more, and compares the net jobs created by startups against the net jobs created by existing firms. Consider a startup company which has 6 employees and goes out of business after 20 months. It has provided 10 person-years of employment, but its net job creation is zero -- two years after it was founded, it isn't contributing anything to the employment numbers. Under the definitions used in this report, however, that company is counted as a startup with a net 6 jobs created... AND as an existing firm with a net -6 jobs created. This failed company which has no long-term effect on job creation is counted not once but twice towards the job-creation superiority of startups over existing firms -- and the more startups fail, the more skewed the numbers will be! Come on guys... startups are important enough in job creation without fudging the statistics like this.

krschultz 16y ago
Very true. To build on this: -They define any business under 1 year old as a "startup". Is a new gas station a startup? A new cleaning business? A new home healthcare franchise? These are all new businesses run by entrepreneurs, but they are not "startups" in the way we generally refer to them on HN. I know several family members that have these kinds of franchise businesses and they never refer to themselves as owners of startups, but as owners of small businesses. Small distinction but lest the VCs and hackers say they create all the jobs, realize that the companies doing "established business" they so look down on are counted amongst startups in this data. 2) Common sense defeats their numbers. Who starts a startup with less than a year runway? That would be ridiculous. If they change their numbers to count startups up to age 3, what would this look like?
- c1sc0 16y ago
  "These are all new businesses run by entrepreneurs, but they are not 'startups' in the way we generally refer to them on HN." Isn't that just a matter of semantics though? Would s/startup/newly incorporated business/g do? I agree that their definition of startup certainly skews the statistics. More interesting would be "Sum of total job-years by startups vs. established companies".
notahacker 16y ago
Agree that the data itself doesn't really support the implicit policy recommendation of encouraging startups rather than established companies, but the basic finding that firms in their first year account for ten times the gross new jobs created by firms in any subsequent year and thus after a year companies on average lay off more staff than they hire is pretty interesting. It would be much more interesting to see what the stats looked like excluding firms that fail in a given year (ie. how many net jobs do firms that survive into year five create in their fourth year)
- cperciva 16y ago
  It would be much more interesting to see what the stats looked like excluding firms that fail in a given year (ie. how many net jobs do firms that survive into year five create in their fourth year) I downloaded the BDS data from the US census website and arbitrarily picked companies which were age 0 in 2000 to look at. Of the 486k "establishments" (approximately equal to companies), roughly 50% survived 5 years. These 486k establishments created 3.1M jobs in 2000, and 1.6M jobs -- roughly 50% -- were lost over the following 5 years due to establishments closing, for a net gain of 1.5M jobs. Meanwhile, 2.4M jobs were created over years 1-5 while 1.7M were lost (at establishments which did not close), for a net gain of 0.7M jobs. Over a 5 year timescale, it seems to me that startups (in the general sense of "new companies") outperform existing companies in job creation by a 2:1 ratio -- still impressive, but also rather more credible than the article's claim.
  - c1sc0 16y ago
    This is why I love hacker news! Thanks for crunching the numbers!
trevelyan 16y ago
It's not actually clear, Mark Thoma just blogged about this w/r/t an NBER paper released earlier this month on this same question: http://economistsview.typepad.com/economistsview/2010/08/who-creates-jobs-small-vs-large-vs-young.html http://economistsview.typepad.com/economistsview/2010/08/who... Conclusion seems to be that job creation happens most heavily in young firms which manage to survive a few years, but that the high rates of job destruction pretty much wash out small businesses gains otherwise. Would be interesting to hear how VC plays into the equation.

robertgaal 16y ago

Argh PDF alert.

c1sc0 16y ago
Or do you mean scribd alert? ;-)

terra_t 16y ago

This kind of study provides all sorts of opportunities for shell games, since the classification of "startup" vs "established", or "new" vs "old" or "big" vs "small" often obscures the real issues. Just to take an example, imagine the person who cleans the lobby of a Fortune 500 headquarters lobby. This person could be a W2 employee of the big co, or could be an independent contractor, or could be an employee of a small firm (say 5-10 people) that does the cleaning, or could be an individual or small firm that owns a franchise for a cleaning services business -- and thus be something that's some kind of symbiotic association of a big co and a little co. The work that gets done is substantially the same, and it's done for the same reason, but the benefits, tax consequences and how it shows up in economic statistics is entirely different.

ynniv 16y ago

Better title: Companies that don't go out of business create the most jobs. An important caveat is that existing firms are so diverse that it can be misleading to think of a “typical” firm. For example, there are two simple categories of existing firms: those that go out of business (Deaths) and those that continue (Survivors). On balance, existing firms lose more jobs than they create. But once Deaths are set aside, Survivors usually create more net jobs than startups do. I think they just proved that their own title is creative accounting. If the goal is a policy change, this report shows that we can reduce unemployment most effectively by keeping established companies open.

zacharyvoase 16y ago

I’m reminded of James Tauber’s blog post on ‘job creation’: http://jtauber.com/blog/2008/11/12/x_will_cost_y_jobs/ http://jtauber.com/blog/2008/11/12/x_will_cost_y_jobs/