Banks are pitching various semi-automatic tools but questions remain.
Banks have been pitching the idea of outsourcing corporate liability management for some time. But optimizing the “LM” side of the ALM equation has become more urgent as market volatility has increased. Now, the optimization is being offered as what amounts to an overlay strategy, using derivatives to manage duration. If done properly, its proponents say, it reduces risk and lowers the cost of funding.
The idea is to rebalance a company’s fixed-floating ratio on a regular, or if possible, constant basis, using several basic market or internal factors as triggers. These might be long or short-term interest rates, yield curve shifts or rolls, and funding need expectations.
Many companies see dynamic rebalancing as a form of market timing, and frown on treasurers trying to call the market. As such, their treasuries tend to have set fixed-floating ratio mandates from the CFO or board.
But banks argue that using the set factors to trigger ratio changes removes the temptation to make treasury a profit center, and instead enhances risk management by allowing a company to change its mix on the fly within pre-approved parameters.
A treasury consultant who has reviewed several of these programs points out two possible issues. The first is the use of derivatives to rebalance, and the premia this requires. Since most of these schemes use vanilla swaps, that’s not often an issue, and it doesn’t cause hedge accounting headaches.
The second is the theoretical engine behind the rebalancing. Usually this is some form of Modern Portfolio Theory efficient frontier analysis. The problem is, MPT is by now so full of holes that putting faith in it to find the perfect mix of fixed and floating liabilities is not warranted. Although various kludges can be used, dynamic optimization relying on MPT could lead to a worse outcome than picking a ratio and sticking with it blindly, especially if market technicals push yields around sharply, as is happening now in Europe. Such breaks in the theoretical relationships between risk and return renders MPT toothless; they could render a dynamic optimization program useless.