I don't think a global wealth tax is practical, so questions of its desirability are moot. I don't think that an endlessly progressive system of taxation is that desirable either, but on the other hand I think your idea of market constraints are also a bit facile. Of course some people misjudge a market and go bust spectacularly, but most people with a lot of wealth tend to accumulate more.
ISTM that Piketty is reaching for a global wealth tax because he doesn't have anything in his economic modeling toolbox that would do the trick instead, so proposes reducing inequality by fiat. He's right about the problem, without having a good solution.
the problem is not the existence of inequality per se. I am fine with some inequality and consider it is natural for Pareto distributions to emerge in economics, just as in nature the existence of whales and elephants is not the result of a conspiracy against plankton or ants. Indeed, there are abundant examples of unequal distributions in nature, from the lengths of rivers to numbers themselves, ie Benford's law.
However, the distribution of capital is not an entirely natural process, and most people who think it is are handwaving simplistic economic models of perfect competition, infinitely substitutable goods, and so on, and forgetting that these perfect models only exist ceteris parabus, ie 'all other things being held equal.' In reality capital flows through a complex contractual, regulatory, and legislative economic environment that treats different participants differently, interferes with default supply and demand (often for perfectly good reasons), and which in turn can be influenced by the infusion of more capital through political lobbying or plain old corruption. What we often end up with instead of a 'free market' is a 'superstar economy' where capital flows increase disproportionately towards people in the top 1%* of the economy, resulting in much greater inequalities than we would expect in a Pareto model.
There's an excellent book on this trend ion the Entertainment industry called Blockbusters, written by Anita Elberse last year. Here's a salient excerpt taken from http://lefsetz.com/wordpress/index.php/archives/2013/10/16/t... (~1000 wrods, worth your time) with this particular bit pointing out that Chris Anderson's 'long tail' model doesn't seem to be panning out as he predicted or as a lot of digerati and creative types (including myself) hoped that it would:
The same patterns are visible in album sales. …out of a total of 870,000 albums that sold at least one copy in 2011, 13 album titles sold more than a million copies each, together accounting for 19 million copies sold. That’s 0.001 percent of all titles accounting for 7 percent of sales. The top 1,000 albums generated about half of all the sales, and the top 10,000 albums around 80 percent of sales. Deep in the tail, 513,000 titles or nearly 60 percent of the assortment, sold fewer than 10 copies each, together making up half a percent of total sales.
tl;dr while we expected a power-law distribution in economics because we're so used to them from nature, the way our economic system is actually calibrated results in a much more extreme distribution that almost disappears into the axes of the demand graph.
Now this sort of thing is a problem for a lot of creative types, because while we can all leverage the incredibly low marginal costs of digital distribution (hooray), there are often significant fixed costs too, and if you can't spread those out over a sufficiently large number of copies then You Are Fucked because you won't be able to get enough capital to pay those fixed costs.
On the positive side, this works out well for writers: writing is of course time-consuming, but since computers are so cheap these days and most people in the developed world have one anyway, the non-labor cost of producing a literary work is very low. So you can have a runaway hit like 50 Shades of Grey where the author typed up a bunch of sexual fantasies, released the result as an e-book (which turned out to be a great move because it's hard to tell what people are reading on a Kindle, compared to a peperback novel with a lurid cover), and made a fortune. For musicians, it's a bit different; some people are blessed with musical talent, others work hard at it as a craft. But then musical performance has to be recorded or brought to the public, and while that's far cheaper and easier than it used to be thanks to digital technologies, it's not free or super-easy and so you see a lot of people raising money from fans on Kickstarter to record albums or launch a tour (depending on whether the artists are better skilled/equipped to play live or record themselves). These artists need an extra capital infusion because launching a new musical product has some significant fixed costs which are otherwise tough to accumulate from normal operating revenues.
Then there's movies, the field I work in. This too has been affected by digital technology; in that there's no need to shoot on film any more thanks to digital cameras, and you can also do amazing special effects that would either not have been achievable with practical (physical) methods or which would have been prohibitively expensive. Yay digital. However, the cost of film and the inconvenience of film cameras are only a small part of a typical film budget. You still have to pay for things like actors, propers, locations, food, transport and so on, and where effects are concerned people's expectations of spectacle increase to match the processing power available, so the public wants ever more elaborate effects and ever-higher production quality (eg color-grading that requires a much greater degree of visual consistency through the whole film than used to be the case a few decades ago). So the cost of manufacturing a film product that meets normal commercial expectation is still in the hundreds of thousands, if you get lucky; if your business strategy is to sell it rather than hope for awards and then sell it, you need to spend at least $1 million, a sum which does not grow on trees.
(There are many famous instances of hit films being made on tiny shoestring budgets of only a few thousand $. And occasionally that happens, but: a) they're vanishingly rare; b) they usually take many years to produce; c) the actual cost to get into the theater is always far higher - the low number is what the producers had spent before a movie executive wrote a check to pay for commercial-quality post-production; and d) the shoestring film almost always involves a ton of freebies from friends/relatives/industry insiders. The odds of making a hit film for a few thousand are about as good as the odds of you outdoing Flappy Bird this weekend, ie virtually nil.)
Now getting back to economics, this is a problem for filmmakers because unless you start out rich or have good social capital, it's not that easy to get your hands on $1 million in the first place, and it's more difficult to do so when more and more consumer spending on tickets/DVDs/rentals goes towards blockbuster movies, as there are fewer unit sales over which you can average your fixed costs. This is not to suggest that the pie of consumer movie spending is completely fixed such that if someone who watches my film will necessarily skip yours; but the pie isn't infinite and (more to the point) the acceleration of revenue concentration around blockbusters is greater than the overall sales growth, so blockbusters get larger and larger shares of a slowly-increasing pie. The overall increase is not sufficient to maintain sales growth or even parity for less popular/populist offerings. B-movie film god Roger Corman summed up the problem succinctly here: http://blogs.wsj.com/speakeasy/2010/09/18/film-school-roger-...
What happens in practice is a race to the bottom; because it's increasingly hard to raise funding for a venture of uncertain commercial prospects as the demand tail shrinks and the (externally) perceived costs of production and distribution fall, that exerts downward pressure on wages as a factor of production for films that are not well-budgeted, such that the extremely bimodal distribution of film sales in reflected in wage scales. Rather than resulting in equilibrium (because goods in this market are not easily substitutable), it's far more likely that less successful participants will simply exit the market, which further exacerbates and accelerates concentration.