An Experiment: Bet on the Price Paid for Toxic Assets

alanholland

One of the key problems faced by NAMA is estimating a fair price for toxic assets so that a break-even return can be made in 5-15 years time. We invite you to participate in a prediction market that can help to identify fair market prices for properties. Once registered (using a pseudonym if you wish) you receive EUR5000 in virtual currency to bet on whether you believe that the price paid by NAMA for a certain set of property will be above or below the quoted figure.
See https://nama.inklingmarkets.com/.

The list of fictional properties includes
* a car showroom in west Dublin,
* a hotel on 15 acres in Dublin 4
* an apartment block in Cork city and
* an unfinished estate in Mullingar.

We assume that properties are tied to impaired loans and NAMA wishes to sell the property for a break-even return in 5-15 years time. A property expert will value these fictional properties in accordance with these guidelines under which NAMA is supposed to operate.

Event deadline: Fri next, June 12th @ 2pm.
This forms part of a social experiment being conducted by Alan Holland of UCC.

Économiste Monétaire

alanholland wrote: »

Cork city

5 Euro.

Time Magazine

Welcome Alan,

Nice to see computer programmers getting involved in a more scientific discipline :pac:

As a citizen, thanks for setting this up and playing your part in averting disaster. Quick question: are these games for real money? How incentive compatible (and thus truly reflective of market wisdom) is the experiment if they're not?

alanholland

This game is for 'virtual currency', and once registered you receive $5000. The idea is to show that the concept can be used in an alternative setting where real currency is used and invited participants (perhaps members of the IAVI) are given real money to wager on what NAMA will determine the fair price to be according to a set of guidelines designed to provide a break-even return to the tax payer.

Prediction markets are, in general, incentive compatibile when the final event is an exogenous variable. If, however, the outcome of the predicted event is conditioned on the prediction market it can harm IC. For example, banks may have an incentive to exaggerate values. So if the prediction market is to be truly useful and the predicted price is used as input to NAMA's pricing, it is necessary to somewhat restrict participant entry but have a sufficiently large number of valuers so that robustness against manipulation and liquidity is maintained.

Some research has recently shown that virtual currency prediction markets perform almost as well as real currency markets, so IC is not destroyed. However, in order to overcome perceived costs of entry and IR constraints a cash incentive would help generate interest. In my opinion, it may be a cost effective means of valuing an extensive portfolio in a scalable manner.

Some more details are available in a position paper here http://www.4c.ucc.ie/~aholland/publications/holland09predMkt.pdf

Time Magazine

alanholland wrote: »

Prediction markets are, in general, incentive compatibile when the final event is an exogenous variable.

My concern was that you'd have players not taking it seriously, but

Some research has recently shown that virtual currency prediction markets perform almost as well as real currency markets, so IC is not destroyed.

answers that.

(Aside: if it is the case that virtual currency prediction markets are not subjected to much messing about/noise, this may have non-trivial implications for survey design. It could also be fun to compare noise traders in Irish banking shares to traders in virtual currency prediction markets and see who's more rational.)

In my opinion, it may be a cost effective means of valuing an extensive portfolio in a scalable manner.

Definitely.

eamonnm79

Are people not likely to wager that the asset is worth =
Their percieved value x 1.* (where * is the percentage the government is going to overpay)

alanholland

If NAMA chooses to overpay; then yes the prediction market will predict this and it's value is lost. However, if NAMA wishes to have a break-even return and does its best to determine such a value using a transparent pricing process then the prediction market can accelerate the arrival at this figure and scale this process up to a larger portfolio.

So, it's only as accurate as the intentions of NAMA.