Generally speaking, the Ben-Stein-o-sphere is one corner of the financial internet from which I'm delighted to absent myself. Yves Smith is more courageous than I am in that regard, and while strolling the mean streets, he offers the following conjecture:

It is one thing to move the prices of single securities, quite another to move entire markets, particularly ones as big as the global equity markets and the US credit markets. We must have a simply staggering number of traders all conspiring together.

That's a perfectly commonsensical thing to say. But is it true? I don't know, but I think it's a fascinating question. Here's the pro and con as I see it:

Pro: By analogy with price impact behavior on individual securities, market indices should be well-nigh impossible to move. Generally speaking, the price impact of transactions is inversely related to the dollar trading volume of the traded issue and positively related to its volatility. It's much easier to move a thinly traded small-cap whose value has recently ranged all over the map than to move, say GE. Broad indices are by definition diversified (which reduces volatility), and the dollar trading volume of index components is gargantuan. Ergo, it should be very difficult to move indices.

Con: Perhaps the analogy between indices and issues is misleading. Indices can be traded as though they were single issues, via ETFs and futures, for example. But the market for such instruments is fragmented, and trading in any one is orders of magnitude thinner than the volume of the overall market it mimcs. An ETF or index future would be hard to move not so much by virtue of its own depth, but because it is bound by arbitrage to the price of the market as a whole. But the market as a whole has a great many degrees of freedom, and many stocks whose "true" values are uncertain. If one wants to materially move the price of a single issue, one probably has to push against "informed valuation" by investors who specialize in the stock's industry and are willing to take bets on relative pricing. But if one pushes against a whole index, arbitrage constraints can be resolved my moving many stocks only slightly, each issue remaining within the bounds of what would be considered noise by those who might trade a deeper mispricing. There's an elegance to this approach, in that the market itself determines an optimal path to resolve the disconnect created by the manipulator. Stocks for which there is a great deal of valuation uncertainty would move more than those whose prospects are clear, and the market manipulator avoids the transaction costs of trading these tens or hundreds of relatively illiquid issues herself. (Price-weighted or equal-weighted indices would be more vulnerable than value-weighted indices to this sort of attack, as smaller / less liquid / more volatile stocks have disproportionate sway.)

That's a nice theory, but would it work? I have no idea. If anyone out there has any insight into the question, hypothetically speaking of course, please do comment. It is frequently suggested that, while individual stock prices may be manipulable in the short-term, broad markets are immune. Is that right?