Tuesday, March 16, 2010

GDP Bond Addition To My February 18 Post "Limits Of Econometric Models ..."

The following is an addendum to my February 18, 2010 post, "Limits Of Econometric Models Of The Macro-Economy". It is reprint of a comment I wrote in early 2009 about using GDP bonds to reveal investor expectations of future US GDP growth.
Why not try a different approach to forecast the economy directly based on rational expectations and the securities markets. Along the lines of TIPS securities, the US Treasury (at the insistence and lobbying of economists) could issue securities whose prices are transparently dependent on future GDP. Different, acceptable forms of securities based on future GDP are possible.

One form could be a US Treasury bond that had a semi-annual interest payment that was a single, fixed percentage (determined at issuance time by a Treasury auction process) of published US nominal GDPs (nominal since cash flows are discounted by nominal interest rates) at the time of the scheduled semi-annual interest payments. Issuance of securities with differing maturities and with the aging of seasoned securities would allow computation of a GDP forward rate yield curve and computation of expected future real GDP growth for different periods. Of course, some might want to make corrections for tax effects, risk premiums, and other possibly distorting and biasing effects.

Another possible form of the security could be a discounted single principal payment at maturity bond that paid a fixed percentage (also determined at issuance through an auction) of US published GDP at time of maturity. Obviously, this percentage would be cumulative for the missed interest payments during the life of the bond, similar to principal only US Treasuries.

Whatever final forms this type of security took, as long as it is transparently dependent on future GDP, it would allow the computation of future real GDP expectations for different future periods. Additionally, other securities based on different macro-economic values can be developed.

A GDP security would also allow for event studies of the US economy and for studies of real time price movements due to government, geo-political, Fed and other actions. The volatility of these securities' prices will reflect a measure of the uncertainty of future GDP growth faced by businesses and others that must make decisions in anticipation of future economic growth. The securities could be used to determine a risk premium of the US economy and to see if it is time varying. It would also allow event studies of the effects of governmental actions on different future period GDPs.

Businesses could use options, futures and other derivatives on these securities to cheaply hedge against future GDP slowdowns and downturns. For example, while the auto industry, and other manufacturing industries, can hedge many raw material and commodity input costs, the new security would allow for a hedge for an unexpected economic slowdown that would cause an excessive inventory buildup.

Therefore, if macro-economic modeling has shortcomings, possibly some of them can be overcome by the use of the expectations embedded in a GDP dependent US bond. Of course, the assistance of the US Treasury Department in issuing a bond of this type is essential.
The above is a reprint of a comment I originally published over a year ago, February 13, 2009, on "Why Macro-Econometric Models Don't Work" on the EclectEcon blog.

No comments:

Post a Comment