Paul Graham “Clarifies” Again.

Paul Graham has added a note to the bottom of his original inequality essay. This is different than his “shorter version.” In this new note, Graham denied making a straw man argument, but then helpfully removed the straw man argument with a new clarification. It’s great he removed it, though, because it was pretty frustrating. But Graham is right. There’s something larger worth talking about.

More to the point, his clarified point is still… well, to be polite about it, confusing. To be less charitable, intentionally so.

Allow me to explain not just the logical fallacy inherent in this note, but the deeper problem it masks, and the ramifications of it.

The Logical Fallacy

Recall, again, that the phrase “economic inequality” is a term of measurement.

Then look at what Mr. Graham has to say.

Because in fact startups do not merely cause some economic inequality, they cause great economic inequality. The median US household net worth is about $80k. It’s common for the stock of a successful startup founder to be worth a hundred times as much, and not unheard of for it to be worth ten thousand times as much.
So startups present a problem not just for those who want to eliminate economic inequality completely, but for anyone opposed to high levels of it. Which is why they make a good counterexample to the idea that economic inequality per se is what we should be worried about, rather than poverty or lack of social mobility.

For a man so renowned for logic, this is a fairly obvious slip. To continue with my global warming analogy, no one gives a damn about one or two fires. Or even a thousand. They care about the planet’s temperature. Just because some (very few) founders can make a lot of money, this does not prove, or even lend logical credence to the notion that startups “cause great economic inequality.”

The fallacy is easy to spot: the aggregate measurement of “economic inequality” is no more influenced by wealth of any individual tech millionaire — even Bill Gates — than global warming is by one really, really hot fire somewhere. Yes, it has an impact, but it is hardly a “great cause,” as Graham says.

If can look like graham is arguing that the group — startup founders — are making an impact. And he is arguing that but his defense of it is about the individual — and a rare one at that.

He’s just proving that some things cause big temperature changes — a fire is a lot hotter than the much more ample air around it — not that that thing influences the planet’s temperature as a whole.

There’s one other obfuscation in there. Yes, “it’s common for the stock of a successful startup founder to be worth a hundred times as much.” This is tautological. The successful ones do well. It is completely, radically, insanely uncommon for any tech founder’s stock to be worth a hundred times as much. The successful ones are, yes, successful. Most of them are not. Just like every other job.

What we have here is an obfuscating argument not backed up by data, resting on a logical fallacy.

You, the reader, can stop right there if you want.

But maybe you’re inclined to give him the benefit of the doubt. After all, just because his argument is flawed doesn’t mean the underlying proposition is untrue. Maybe you think Tech is good, and therefore tech millionaires are A-OK in your book (I would agree with you). And maybe you believe Graham when he says that they’re a big part of the wealth of the world.

Or maybe you’re wondering why he keeps saying such a weird thing. Like, “okay, tech founders aren’t so bad. We still need to fix economic inequality. Why does he keep talking about tech cofounders? Is it just because he works in tech? Is someone, somewhere, advocating abolishing startup founders?”

Well read on.

The Actual Math

On the Forbes 400 richest people in the US list — the very one Graham mentioned: thirteen percent of the richest people (53) in America made their money in tech. The share of people who made their money via investments (aka hedge funds and the like), by comparison, is 24%. Yes, Tech is the second largest category, so I suppose Graham’s not incorrect by calling it significant. But it’s not, like, a the majority of it.* Not even close.

To illustrate: those tech billionaires on the Forbes 400 control $515 billion between them. The top 1% of the United States controls just shy of twenty eight trillion dollars. The Forbes 400 as a whole controls around $4 trillion.

Okay, damn. yeah. That’s a lot of money for fifty three people to control.

This can seem to lend credence to Graham’s arguments, yes. Not as much as he implies, but sure. There is a lot of wealth in the fifty-three tech founders (and VCs) who have made the Forbes list. But there are other ways to look at the numbers:

Yes, successful startup founders make a lot of money. And the stock of a few — a very few — can be worth “hundreds times as much” as the median US household net worth. But this is not really a problem because there aren’t very many of them. There aren’t enough tech millionaires out there to make much of a dent in economic inequality. I’ve worked in it for more than a decade and I probably only know a dozen or so people who’ve hit a net worth of $8 million (100 times median).

Let’s go all mathematical on it.

How many millionaires has tech made? A million? Not even close. How many millionaires are in the United States? More than six million (and I am using the conservative definition, which excludes primary residences). California has the highest percentage of these. You know how many millionaires there are in California? 777,000. But most of these aren’t tech millionaires. Los Angeles, being the 14th largest population of millionaires in the world, has 126,000 of them. San Francisco doesn’t even make the list.**

So, no, Mr. Graham. Startups do not “present a problem not just for those who want to eliminate economic inequality completely, but for anyone opposed to high levels of it.” The startup world could triple the number of millionaires it has made, and it would still be a drop in the bucket when it comes to economic inequality.

But it’s not going to and we both know it. Because…

What’s really going on is that startups are, sadly, just like every other industry: a fewer number of people are getting even richer

Microsoft made 12,000 millionaires. Google made 1,000. Facebook made 1,000 (to be fair,they’ve probably made one or two thousand more since their stock is killing it). Twitter made 1,600. Atlassian has made 100. I worked at a unicorn. It sold for a billion plus. I’d be shocked if it made more than 20. I’d love to find a chart for this — number of tech millionaires made out of employees for every tech IPOs out of time. I would wager, strongly, that the number is declining. (Am I wrong? Please, let me be wrong. I’d love to be wrong. If so, let me know.)

Wonder where all that money went. Oh right. The VCs. And the 2 or 3 co-founders.

Every single person on that Forbes 400 list from tech was a founder, co-founder or a VC. Not a single early employee among them.

San Francisco isn’t in the top 20 cities on earth for millionaires. Or multi-millionaires. But you know what list it is on? Billionaires. San Francisco on its own, ignoring the rest of Silicon Valley, ranks #16 in the world for billionaires.

It’s very, very odd to me to see Graham continually equate “technology” with “wealth.” Tech startups are great, sure. But they have nothing to do with economic inequality. The funding structures of startups do. Why hide behind startups to defend wealth? Oh right. Because people think startups are neat. So by pretending they’re somehow linked with income inequality, you can argue against fixing a problem people want to fix.

The more you think about it, the more you realize that Paul Graham is trying to claim that startups are essential to our society (agreed) and the only way people will build startups is if we do it exactly as we do right now, and the top three people in a startup can get super super rich, like global elite rich, and everyone else in the company doesn’t.

Now, I wish he’d just come out and say that, because I’m super curious. I read his fragmentation essay. It’s pretty great. What I missed in it, though, was any evidence of mention of the US being less technologically innovative at any of those times that weren’t like now. Were we less innovative in the 40's? The 60's? The year of the Mac, 1984, with Apple and Microsoft and Intel and IBM and economic inequality being about half what it is now? No, no, no. The proposition is absurd. It seems to me that people will still start just as many startups with less economic inequality in the world, just like they have in the past. I’ve certainly read no evidence from Graham to the contrary. And all the evidence I can find supports this notion.***

Technology is innovation and innovation is good, yes.

Technology can make people rich, yes.

But does technology influence economic inequality? Does technology cause economic inequality? Graham claims it does, thus arguing economic inequality isn’t all bad. But he presents no evidence. The fallacious argument he does muster (“The median US household net worth is about $80k. It’s common for the stock of a successful startup founder to be worth a hundred times as much, and not unheard of for it to be worth ten thousand times as much.”) is a) logically flawed, and b) a description of how we currently choose to finance technology. We did not always do it this way — Graham himself specifically mentions the Apple/IBM/Microsoft era as a “classic example” of technological disruption.

Personally, I’d love to go back to the economic inequality levels of that period.

And I can’t help but wonder how upset the 99% of the Bay area would be if tech IPOs still made the number of millionaires as they did with Microsoft.

* FWIW, I call this the “big chunk” fallacy. Those words are obfuscating. Is 5% a giant chunk of something? Yes, if, say, it’s your blood alcohol content. No, if, say, it’s the number of questions you got wrong on the SAT. I’m as prone to it as anyone. It’s a great way to make a small number seem bigger.

** SF doesn’t include Silicon Valley in this study, it’s true. Or Oakland, etc. Yet it also doesn’t include Beverly Hills or Orange County in LA, or even Brooklyn in New York. Weird. I am assuming all that comes out in the wash, but even if you give the Bay area the rest of the Cali millionaires (a preposterous proposition given Beverly Hills and the OC), the math is essentially the same.

*** I could use R&D funding as a percentage of GDP or something, but for the reasons Graham outlines so well in his fragmentation essay, I don’t think this would be accurate.