Want to be rich? Drop out of school

Billionaires drop out more and get PhDs less adjusted for IQ

I am currently enrolled in a PhD program, but it’s pretty miserable. For one, I don’t want to be a college professor. They are often oppressive, parasitical, and produce little of value. Not all, but the academic system is rotten to the core and being part of it in the long term seems immoral. I mainly signed up for it because I enjoy research, and there is currently a tech job recession. I am 3 months into the program and have not done an inch of actual research. Instead, I have wasted time on useless undergraduate-style courses and being a teaching assistant.

I am surrounded by non-whites as well. Easily 85% of my department are foreign. In conditions prior to this tech job recession, I could easily have been making near $100,000 per year to create websites, potentially remotely, instead of $20,000 per year to grade programming midterms and to be graded myself in obnoxious, irrelevant courses that I don’t attend.

I justified starting a PhD program because I thought it might pay for itself at the long term, which is what prompted me to write this article.

Computer Science PhDs are worthless at the median

When you look at the data, computer science and by extension STEM PhDs in general are worthless for the median worker. Sketchy online statistics say the expected pay increase is about $20,000 per year with a PhD. People who have a PhD in computer science only earn $99,000 on average. Meanwhile, the average CS bachelor degree holder earns $79,000 on average. But a graduate student earns about $20,000 on average. So he is sacrificing as much as $60,000 per year, which over just 4 years (many stay for longer) adds up to $240,000, which takes 12 years to earn back after a $20,000 pay raise! If you graduate from the PhD at the age of 26, equally intelligent peers who did not go to graduate school are richer than you until the age of 38.

This probably understates it, as (before the recession at least) a PhD tier intelligence can usually get an above average paying job out of college.

Computer Science PhDs may be harmful at the right tail

Above covers the case of the median worker. What about ultra-successful entrepreneurs? Might getting a PhD have increased their human capital somehow by getting them to the edge of some valuable field? Probably not.

A billionaire is more likely to have not completed college than he is to have a PhD. About 30% of billionaires have not completed college, while only 10% have a PhD.

But this could be misleading. Compared to a similar non-billionaire population, billionaires may be more likely to have a PhD and less likely to drop out of college. If a set of 1000 random people sampled from an IQ distribution of N(135,10), a good estimate of the billionaire IQ distribution, had a 40% college dropout rate and a 5% PhD rate, then getting a PhD might be a good idea.

We therefore want to know the rates of these things in that high IQ population. This can be done with a simulation. First, generate a sample of 1000 people with IQs on N(135, 10). Next, find the probability that they will drop out of college or get a PhD. The probability they will get a PhD given their IQ for instance is L(their IQ | mean IQ of PhD holders, SD of PhD holders) * P(having a PhD) / L(their IQ | mean=100, SD=15). If we find the information we need on the IQs of PhD holders and drop outs, as well as the general rate of having a PhD and dropping out, we can compute this number for each of the 1000 samples we generated.

Now you can just compute the mean probability.

It turns out P(PhD) = 1.2% and P(Dropout) = 56%.

From this study we have estimations of IQ distributions by educational attainment. The right tails are strage. For PhDs it’s very thick, moreso than a Gaussian. To simulate this, we need to choose Gaussian parameters that look like the empirical data at the right tail.

To do this, I chose the parameters 97, 102, and 110 for the means and 14, 15, and 20 for the SDs.

What is important is that the proportionate areas under the curve between 120 and 150 IQ are very similar, which seems to be the case here.

Running the simulation with these parameters, we get 14.5% of billionaire-like-people should have PhDs, and 22% should have not finished college.

Thus, P(Dropout | Billionaire) / P(Dropout | high IQ) = 30/22 = 1.36 while P(PhD | Billlionaire) / P(PhD | high IQ) = 10/14 = 0.71.

Simulation Example Code

Conclusion: PhDs are probably a waste of time. At best, billionaires don’t have them more than a similarly high IQ population. At worst, they have them about 70% as much and drop out of college 1.36 times more.

I also did a simulation for just good CS PhDs. I found that about 0.038% of the population have these. I assumed their IQ distribution was the same as that of billionaires. I got that .9% of billionaires should have these. In fact, 6/735 do which is about 0.8% of billionaires with CS PhDs. So in this case P(CS PhD | Billionaire) / P(CS PhD | high IQ) = 0.88. This is less certain because I just guessed the IQ distribution of the “good” CS PhD holders. but needless to say there is no evidence that it’s helpful.

In general, these results accord with Bryan Caplan’s finding the university is mostly signaling and hardly increases human capital value, if at all. Education is clearly not making people rich, or even paying for itself at the median. Unlike undergrad, for both the extreme right tail and the median, graduate school is an individual waste of time and money.