This is part two of a two-part series. Click here to read the first part.
Yesterday, we talked about predictive markets and touched on their applicability to metal markets in general. Without surveying all of the literature, suffice it to say that predictive markets, can be much more accurate than traditional polling techniques or other analytical methods. So given the fact that there are dozens of research firms and analysts covering the metal and mining sector, why are there no predictive markets (at least, we couldn’t find one), with the exceptions noted from the comments in yesterday’s post (e.g. the futures markets for aluminum, copper and the balance of base metals).
To begin, there are several “must-haves” for a predictive market to work. And perhaps not surprisingly, many of these must-haves relate to the requirements of all markets in general. They all need liquidity or people to take positions. One of the primary reasons why the steel long products futures market has not come on stream, in our opinion is due to lack of liquidity on the producer side. See an earlier MetalMiner post on that subject here.
This raises an interesting question. Is the LME and Nymex a predictive market? One could certainly say it is for certain products in that a daily spot price and futures price is established via a formal trading mechanism. But I would argue a predictive market and the formal futures markets are not one and the same for one basic reason…who is participating. In the case of a futures market for say aluminum ingots, the traders, ETF’s, aluminum producers etc are all “participating” in the market. But the ultimate end users, the OEM’s who consume and drive the demand for semis and other finished products have little to say in the price setting market of the LME or Nymex. The purpose is not to predict prices; the purpose is to protect the participants against changes in the price. A predictive market then, if placed closer to actual demand, may generate some predictions correlated but not causal to the underlying LME or Nymex markets.
The second element needed in predictive markets is a traded currency. Markets that solicit opinions but don’t require participants to put a little skin in the game may not yield the “truest” prediction. It comes down to “I say what I mean and do what I say.” According to this blog on crowdsourcing, the money issue is a real problem due to prohibitions on gambling. Another point suggests, “the accuracy of futures markets is in direct proportion to the “thickness” of the market” – back to the liquidity issue stated above. And finally, predictive markets fail (as do others) due to fads, bubbles and other acts of irrational exuberance, to borrow Alan Greenspan’s nomenclature.
As for a fix to these issues, the main one being liquidity, the crowdsourcing article suggests that small groups of market participants undergo a question/answer analysis to evaluate their risk to aversion. Those that don’t take risk management to heart may be given a higher rating whereas cautious market participants are given a lower rating. The ratings are then weighted against the risk index and against their predictive responses.
Hey who knows, maybe we can create a predictive market here some day or develop some clever little gaming site where you can all take your turns being steel avatars in the virtual world. Now wouldn’t that be a hoot!