Saving the Financial System by Buying Troubled Assets

by Arne Andersson, CEO, Trade Extensions

October 08, 2008 - There are many good reasons for a government to save its financial system, even if it will cause a large cost for the tax payers. The essential goal of such a procedure can simply be phrased as:

“Save as many financial institutions as possible at the lowest possible cost”

The typical approach, taken by US Department of Treasury, and probably followed in other countries, is to buy insecure assets at overprice, and then try to recover some of the costs later on by selling the same assets. The question is how much to buy and what to pay. Like in the US case, it is reasonable to assume that the government will appoint an available budget that should be spent as wisely as possible. A very natural way to do this is by an auction procedure, where asset holders act as sellers and the government acts as a buyer.

Essentially, the sellers will compete for shares of the available budget, by offering their assets at some price. The buyer will rank bids by cost per asset value, and the sellers will decrease their prices until the auction ends.

What is the best auction format?

First of all, it needs to be stated that there is no scientifically proven clear answer to what is the best suitable auction design, although there are a number of educated opinions. First of all, as the participating seller, the financial institutions that are to be rescued, will have different needs, the auction must be able to allocate different volumes to different bidders. Also, if we end up in a situation where one or more sellers manage to sell some assets, but not enough to cover their needs, the money spent on these sellers will be spent in vain. Hence, we clearly need an auction mechanism where we can allocate different volumes to different bidders and where a bidder can ensure that it gets the right volume, if any.

Drawbacks with Simultaneous Auctions

A common auction mechanism, used in a number of cases in the public sector throughout the world, is a simultaneous auction, like the so-called “clock auctions”. The basic idea is to split the market into a large number of small “lots” that are then auctioned simultaneously. In order to get proper total volumes bidders will “explore” their way through the bidding rounds; if they win too little, they have to bid more aggressively, if they win too much they will wait for others to bid, etc. The auction proceeds until we hopefully end up in a final suitable allocation. Although this type of auction may often lead to acceptable result, it clearly has a number of disadvantages:

1. There is no real way for a bidder to ensure getting the proper volume. More or less all of the intricate methods to improve this situation e.g. running clocks in parallel, allowing bidders to take back bids, etc, is essentially covering a weak construction. In the end, we are more or less just hoping for a good result.
2. Once bidding has started, it is likely that all available lots will receive bids, which means that all lots will eventually be bought. This means that in a case where 90% of the budget would have been enough to rescue, say, 10 sellers, but rescuing the 11th would require 110%, all 100% will be traded anyway, at a very high extra cost for the government (and the tax payers).
3. The strategic game of how to act as a bidder in order to converge at the proper volume is extremely intricate, and the strategies are not well known. When should I bid more aggressively? When should I take back some bids? What happens if some of my competitors takes back some of their bids? The list of questions is almost endless.

Our Proposal: Volume-Constrained Auctions

Instead of running a simultaneous auction, we propose the use of bids with volume constraints. The method is conceptually simple and clear:

“Declare the available budget to the sellers and let them express bids where
they state price and volume constraints.”

This can be done in several ways:  by a set of volume-constrained bids per bidder, or by each bidder presenting a price-demand curve with upper and volume bounds.

The use of volume-constrained bids has a number of advantages:

1. Sellers will only give bids that actually represent a meaningful volume, so there is less risk of a bidder ending up with the wrong volume (a bidder may end up with zero volume. From a tax payer’s perspective, this is actually better than if the bidder ands up with some insufficient volume.
2. The total traded volume does not necessarily end up at 100% of available budget, which is a true savings for the buyer.

Can a Volume-Constrained Auction be Run in Practice?

An auction with volume-constrained bids requires an optimization engine tailored for this type of market. Fortunately, such and engine exists. The Trade Extensions platform has been used in a large number of auctions where then complexity both in terms of number of active bids and number of used constraints is considerably larger than what is required in this case. It is a piece of cutting-edge technology delivered by leading specialists in Computing, Optimization, and Game Theory.

-ends-

About Arne Andersson
Arne Andersson is the founder and CEO of Trade Extensions.  He is Chair Professor in Computing Science at Uppsala University, Sweden, and is a leading expert on algorithm design. He has developed improved algorithmic techniques for fundamental computational tasks, such as sorting and searching for data and he has published many articles in leading journals and regularly speaks at international conferences.

About Trade Extensions
Trade Extensions produces unique procurement software to help organisations specify tenders, collect offers and, most significantly, optimise the results.  Trade Extensions was set up in 2000 by computer scientists from Uppsala University, Sweden, and the founder and CEO - Arne Andersson - is one of the leaders in the field of algorithm design and sort theory.  On-line procurement projects generate millions of variables and it is the algorithm design that underpins the success of Trade Extensions’ software. Since 2000 it has been used in the sourcing of over €10 billion of goods and services. 

More information

Niklas Pettifor
Pettifor & Pettifor
niklas@pettifor.com
+44 (0)20 7936 9269

Back to News & Press