The Malinvestment Boom In Coders

Submitted by Pater Tenebrarum of Acting Man blog,

Computer Science Enrollments Explode

Techcrunch had an interesting article about the explosion of freshmen deciding to study computer science. There can be no doubt that this type of knowledge is currently in great demand – however, we do believe that there are some signs that the boom is so to speak 'getting out of hand' and is beginning to reflect the effects of the technology echo bubble on Wall Street.

The give-away is the size of the demand for computer science studies relative to other fields of study. Even if computer technology and everything connected to it is steadily increasing in importance, this looks like 'bubble behavior' to us. Let us not forget, vast increases in the money supply and the concomitant suppression of the natural interest rate are mainly reflected in the higher order stages of the economy's production structure, and investment in 'human capital' is definitely a very 'high order' stage (meaning: it takes a very long time before the investment actually bears fruit and produces an income).

New enrollments in computer science vs. other studies – click to enlarge.

People may be inclined to instinctively judge that it can only be good if so many young people decide to study something 'useful' as opposed to, say, art history or literature. However, as Mish has recently pointed out, the widespread yammering about a 'skills shortage' is actually misguided (he provides a few additional links to recent press reports on the topic that are well worth checking out as well). There is in reality no shortage of workers skilled in science and engineering.

We happen to believe that jumping on a popular bandwagon is almost always a mistake. By the time these freshmen finish with their studies, they may well find out that a shortage has developed in an entirely different field and that their chosen profession is suddenly crowded with job seekers.

Bubble or Not?

One needs to keep in mind here that the last time enrollment in computer science peaked was in the year 2000 – concurrently with the technology mania. This is obviously no coincidence. What is slightly disconcerting is that the current peak in enrollments towers vastly above that previous bubble peak. This can be gleaned from the data published by individual colleges. Here is e.g. Carnegie-Mellon University as a pertinent example:

Carnegie-Mellon University's computer sciences enrollments. The current peak dwarfs the year 2000 peak – click to enlarge.

The researchers quoted by Techcrunch assert that it is 'not a bubble this time' based on what we believe is spurious reasoning:

“Now, when the word “bubble” gets thrown around, many tech industry leaders like to point out that public market activity is nowhere near the irrational exuberance seen during the first tech boom. But in academia, enthusiasm for learning to code is at all time highs, with CS enrollment at some schools far surpassing the numbers that were seen during the late 1990s.”

Our reply to this is that not every bubble is accompanied by the kind of public frenzy seen in 1929 or 2000. After suffering through two major bear markets and a real estate collapse, the still overleveraged public is no longer as enthused about the stock market's get rich quick promise as it once was, that is certainly true. However, that only means that the bubble is driven by 'professionals' nowadays – with the main difference that they are probably even more reckless, as they play with other people's money. In other words, they get all the reward, while incurring very little risk. In fact, the biggest career risk investment professionals as a rule face is being out of the market while it goes up. If they get caught fully invested in a bear market, they have little to fear because they will have plenty of company and even more excuses (of the 'no-one could have seen it coming' type).

Hence it could well be argued that the current echo bubble is in some ways even more worrisome than the frenzy seen in the late 1990s. It has definitely invited a great deal more leverage already:

Investor margin debt balances have ballooned to a far greater extent than at the year 2000 peak – never before have negative credit balances been at such an extreme level – click to enlarge.

Lastly, in order to determine whether or not one is in a bubble era, opinions and observations about the mass-psychological backdrop are one thing, but the data are more important. The most important datum of them all is money supply growth, and with regard to that, the current period definitely stands out. Since 2008, the broad US money supply TMS-2 has increased by about 93% (from $5.3 trn. to $10.2 trn.). Since 2000, the increase in the broad US true money supply amounts to an even more staggering 245%. So there is now almost twice as much money in the economy than in 2008, and nearly three and half times as much as in 2000. This is the datum most relevant to determining if we are in a bubble era or not, and it clearly tells us that we are. It matters little if it 'feels' like we are in a bubble to 'tech industry leaders' (one of whom let us not forget has just spent $19 billion for an 'app', buying profitless 'eyeballs'). What matters in this case is the evidence.

US money supply TMS-2 (without memorandum items) since the 1990s – click to enlarge.

As to the effect of this monetary expansion on asset prices, below is a chart of the Nasdaq Internet Power Shares ETF documenting it quite nicely. As an aside to all this, we are not saying anything regarding the timing of the bubble's eventual demise. The question is only 'is it a bubble?' We say it is one, and it can be shown that this is the case.

The huge increase in computer science enrollment is therefore most likely an example of malinvestment in human capital. Students would do well to take the time to carefully ponder their choices in this respect.

PNQI – one of the market sectors displaying the exponential growth typical of bubbles – click to enlarge.

Lastly, here is a slightly dated chart we have shown before, namely the percentage of money-losing IPOs, which has recently reached the year 2000 peak again: