The previous comment would seem to be implying that there's "a real fundamental value buried in the noise" for investment instruments, but none for models. I would say that there's always real fundamental value in people, models included, and that this value is in some sense more real than anything in any market.

The momentum is the only source of value.

The momentum is the only source of value modulo the market. There is indeed "a real fundamental value" to all personnel assets, but it isn't even "buried in the noise." The problem is that the market is almost totally disconnected from this underlying reality. (The question seems too complex, to ever be fully connected.)

There is a sort of fear and insecurity that overtakes tastemakers faced with overabundant choices with no good analytical tools to help them. I'm not sure what the solution is. An "indy" market is no solution -- these seem to merely become the "bush leagues" for their preexisting mainstream. After awhile in cultural markets, the noise machine takes over the market and disconnects it from the human and cultural perceptual machinery which gave rise to it. (Which I will call "true taste.") Subcultural 'scenes' arise with a renewed connection to true taste, but these scenes eventually get swept up into the same disconnected market.

I suspect that Buddhism talks about this. I suspect that this is just a part of the human condition. Things like Rotten Tomatoes can help, in that rogue tastemakers who are simply succumbing to bribes and other simple manipulations can be left out in the cold. Maybe the 4chan folks have got it right, and informational/social anarchy is the ultimate solution.

In the meantime, a dispassionate analysis of such markets might yield opportunities to make money through trading.

Models may have fundamental value, but wisdom on model valuations is a little sparse at this point in history. Same for people in general -- there just aren't more than a few people (and I'm giving a benefit of a doubt that they exist at all) out there making a living by estimating the future earnings potential of an individual. Valuating corporations, on the other hand, is quite popular and there are many thousands of highly-paid analysts who sit around and talk about how much a company is worth, then sell that information to their clients. Actuaries are sort of in the same ballpark and society already frowns on them enough as it is, so while this might be a market ripe for the taking I suppose there needs to be a really compelling reason to pay for the service first. It would make more sense if one could purchase an individual -- I know I've seen Robin Hanson talk about the concept of a society where parents sell shares in the future earnings of their children to finance their education.

But for the moment, even if there were "model analysts" who published estimates on how much a model is worth, it would have little impact on their popularity. Models don't get popular because they have earning potential so much as the other way around. A model who is popular will command higher rates and more frequent appearances. If she is diagnosed with cancer and has 6 months to live then her potential is greatly diminished, but she can leverage it for charity purposes and get even higher rates. Valuation is interesting an academic sense, but for the time being it is a very useless thing to have in the fashion market. This effectively leaves the idea of a model's fundamental value even less tangible than that of a corporation.

Considering the extent to which bubbles form even with the possibility of valuation, the only way that a model popularity market could be analogous to an equity market is if it lives in a perpetual bubble, effectively negating the relevance of the fundamental value. "Reckoning" in the equity markets might be rare, but it does in fact happen. The career of a model, on the other hand, has zero empirical effect by their fundamental value.

When it comes down to it, there is such a thing as talent. There is also something real called taste. The tragedy is that it's often really difficult for this to be reflected in a market. Perhaps a part of the problem is that talent is actually widespread: there is no distinct rare "superstar" talent, just lots of cute young women with a certain kind of look. Therefore, there is no rarefied taste that can recognize the "superstar."

The valuation of personnel is also greatly complicated by all sorts of other factors. A musician might have amazing chops, but the wrong kind of personality to deal with the rigors of touring or the temptations of stardom.

In essence, there are "model analysts." In the article, they are called tastemakers. They don't publish their conclusions, instead disseminating information at parties. They are not necessarily paid directly for their analysis, but most certainly participate in a reputation market.

The takeaway is not that talent and taste do not exist. They do exist. It's that truly distinct "superstars" and the talent that can recognize them are something of a marketing fiction. I agree that such markets are perpetual bubbles. Because of this fact, they do more harm than good -- they act to cloud information to buyers and reduce access to resources.

Re: Trading -- knowing that certain kinds of markets are always bubbles points to certain trading strategies.

That's what we do in the mortgage and insurance business, when giving long-term loans or underwrite policies.

The last company I worked for (a branch of Equifax) had a full-time Statistician just for that kind of work.

PS: VHS and Global Worming are poor examples for this however they are both somewhat sticky beyond their technical merits.

EX: The amount of GW research relates to its political implications far more than the complexity of the theory. When the implications stop having economic implications the level of interest and resistance will quickly drop. As will the amount of debate by people with limited understanding of the issue. Also, VHS's initial technical advantage was soon less important than its level of adoption.

The problem is that they only do so in the longer run, where as people expect performance on a daily basis from a trader. Here is where social influences come into play. Who do you think will get more commendation from their bosses? A trader who makes an investment based upon intrinsic value that may take years to show? Or, someone who "rode the markets" that day?

The problem is that the free market hypothesis assumes that we are rational beings capable of making cold judgments. Our "rationality" is inexplicably tied to our emotions we may not realize it, but this is something as ubiquitous as our ability to learn language. Ask anyone with a mood disorder and they can tell you just how different the world looks when the balance is tilted. You may argue that I am citing an extreme case, but evidence points out that mood disorders are nothing but abnormal functioning of circuits within our brain. They just happen to expose just how fragile our worlds are.

On the other hand, I've always wondered if it is possible to take advantage of this and make an "infinite money generator". It's a thought experiment really;

Let us say that you have this program running on an impressive machine that scans the entire internet and parses out information related to financial markets, and categorizes it on the basis of stock, the perception attached to that stock and the emotions associated with it (twitter/facebook/the latest fad). It also has data on how this "herd" has behaved before. Let us also assume that AI has born the fruit of accurate simulation of humans at a larger scale.

If we input this data into the program and use it to predict the market do you not think that it will be more successful at making money than a machine designed to work only on the "facts"?

Adopting the "there's no reason" approach is akin to treating the process as random. First advocated by Boltzman to lay the groundwork of statistical physics, this approach basically says that the process is so complex that the only way to deal with it is through probabilistic methods.

This approach in fact is widely used in analyzing stock markets. So in fact, the current wisdom is not to disregard the "noise" in the market to arrive at the "real" model, but to treat the whole market data as noise!

I must respectfully disagree with you and agree with gxti.

Based on the article, it seems that what gets a modelling career started to build up its momentum is virtually random. Once it is started its own momentum can sustain it for some time, but that start is random.

Now, the market on a day to day basis is indeed virtually random. But the overall trends are not. Over the long term with general trends, the stock of strong companies will do well because they generate profits which are largely independent of the stock price. On the other hand, a high stock price might bolster an otherwise failing company for a while, but only for a while. Daily fluctuations are almost entirely random, but long term trend lines will tend to track the actual value of a company, with some exceptions.

As Ken Arrow and others have pointed out, what will have real value in the future can depend heavily on what "noise traders and trend chasers" are doing now. Investment choices have feedback effects.