It's complicated. But a good place to start is to dispel the persistent and pernicious myth that a free market does a good job at objectively measuring individual productivity. The reason this myth is pernicious is that people conclude that it is good policy to let the market distribute wealth because it's good to reward productivity. I do not dispute that it is good to reward productivity (of course it is!) What I dispute is that the market reliably measures productivity. Hence, letting the market distribute wealth does not necessarily reward productivity. It can (and often does) reward productivity, but it can reward a lot of other things as well (like luck, or gaming the system), and so we have to be more careful when making policy than simply trusting the market.
Take this example from Paul Graham in a footnote to one of his essays:
[I]t is certainly not impossible for a CEO to make 200x as much difference to a company's revenues as the average employee. Look at what Steve Jobs did for Apple when he came back as CEO. It would have been a good deal for the board to give him 95% of the company. Apple's market cap the day Steve came back in July 1997 was 1.73 billion. 5% of Apple now (January 2016) would be worth about 30 billion. And it would not be if Steve hadn't come back; Apple probably wouldn't even exist anymore.This reasoning is flawed. Just because the value of Apple + Steve Jobs is 200x the value of Apple - Steve Jobs, it does not follow that Steve's value is 95% of Apple + Steve. Why? Because companies are not linear systems. In fact, it is trivial to debunk this: just take away the rest of Apple. By PG's theory, the value of the result (i.e. Steve Jobs working alone) should be 95% of Steve + the rest of Apple. That's obviously not the case. Steve was only able to do what he did because the rest of Apple existed. (But, one might argue, the rest of Apple existed because of Steve! No, it didn't. Steve did not build Apple single-handedly. When Steve built a company single-handedly, the result was not Apple, it was NeXT.)
Here's an analogy: the value of a car lies in its ability to transport you from place to place. If you, say, remove the spark plugs, the car becomes useless. If we stipulate that a working car is worth 20 times as much as a non-working car, then by PG's reasoning it would follow that the spark plugs ought to be worth 95% of the value of the car.
Another counter-argument might be that yes, Steve may have just been a spark plug, but he was unique. Spark plugs are cheap because they are plentiful, but if they were scarce then they would be worth more.
But how do we really know that Steve was unique? It's possible that there are hundreds — maybe thousands, maybe more — people out there who are capable of doing what Steve did, but we don't know about it simply because Apple never ran those experiments.
In fact, it is extremely unlikely that Steve was actually unique, in the sense that he was the only human being with the innate ability and drive to do what he did. Why? Because Steve was born in San Francisco in 1955, which was the exact right place and time to be born to come of age during the semiconductor revolution that made Apple possible. It was the exact right place and time to be born that let him meet Woz, without whom Apple probably would not have happened. The odds that the one human being in all history capable of doing these things just happened to be born under these auspicious circumstances are very, very low.
Much more likely is that there are lots of potential Steve Jobs's out there, but they never get the opportunity to demonstrate their abilities because their life circumstances don't allow it. Maybe they have to work to pay off student loans. Maybe they are refugees (Steve's biological father was Syrian). Maybe the reason that Steve Jobs seems like a scarce resource is simply because the world hasn't bothered to try to cultivate more people like him.
Let me be clear: I am not making the Marxist argument that everyone is equal. Steve was clearly exceptionally talented and motivated. But what made Apple possible was not just Steve's exceptional talent and motivation, but a whole host of other circumstances and people that happened to put him (and them!) at the right place at the right time. And if any one of those circumstances had been different, the outcome for Steve and Apple could have been very different. Steve had a lot of help from people like Woz and Jony Ive. (But, I hear you saying, Steve hired those people! Sure, but if you're going to give Steve credit for that, why not give the Apple board credit for hiring Steve?)
My pont here is just this: simply because Apple + Steve was worth 20x what Apple - Steve was worth, it does not follow that paying 95% of Apple + Steve for Steve is a good deal for anyone but Steve.
So we can't measure individual productivity simply by looking at the value of an organization with and without them. It's a non-linear system. The whole is greater than the sum of its parts.
So how can we measure individual productivity for creative endeavors like running companies? It's really hard. It may actually be impossible. Certainly we don't have any way to do it today. The market is clearly not efficient in this regard. We have CEOs getting paid tens of millions of dollars for running companies into the ground. (Attention, boards of directors: I will bankrupt your company for half of whatever you're paying your current CEO!)
This is only a tiny sliver of the real problem, but it's an important one because so many people base their reasoning on the false assumption that a free market is a reliable measure of productivity, and that rich people are rich because they are so much more productive than non-rich people. No. Rich people are rich because they happened to find themselves in a confluence of many, many circumstances that ended up with their becoming rich. One of those many circumstances may have been (usually is, but often isn't) that they worked hard and took risks. But working hard and taking risks is no guarantee of success (that's why it's called "risk").
Note that this observation says absolutely nothing about policy or quality metrics. That will have to wait for another installment. Like I said at the beginning, it's complicated. But if we're going to deal with this situation rationally we have to start by being honest without ourselves about the ground truth. And part of the ground truth is that there's a lot of noise in the system. That's one of the things that makes it resistant to simple solutions.