10Y ACGB Not Modeled By Australian Economic Fundamentals: DB

For modelling 10Y ACGB yield, usage of market rates such as the 10Y UST yield and relative front-end pricing would be preferred rather than a model based on Australian domestic economic fundamentals, believe analysts at Deutsche Bank AG (NYSE:DB) (ETR:DBK) (FRA:DB). David Plank suggests in his July 29 report titled “Australia: Modeling the 10Y ACGB” that the China/U.S. growth gap would be more a reasonable forward indicator for relative AUD/USD front-end pricing.

Some approaches to model 10Y ACGB

Plank points out that though domestic fundamentals could be an obvious starting point to model for the 10Y ACGB, simply looking at nominal GDP growth doesn’t appear to provide much information about its likely level. The analyst argues that a model which combines a number of variables such as real GDP, CPI, the AUD and the RBA cash rate seems to do a pretty good job. By using quarterly data back to 1990, the analyst get an R-squared level of 0.9:

However, the analyst concedes that there are econometric problems with such a reduced form equation, such as model miss-specification and covariance. Plank points out that the OECD and RBA have attempted to model long-term AUD bond yields. However, he also argues that this approach has problems as the data is often published with quite long lags. Plank argues that one could look to solve some of these problems by focusing on more timely data, though it’s difficult to reduce the time frame to less than a month.

Discussing other options, Plank points out that an alternative approach that might suit a shorter-term market focus is to take advantage of the fact that the Australian bond market is simply a small part of a global capital market. The analyst notes there is clearly a high degree of correlation between the Australian bond yields and U.S. and German bond yields.  However, he concedes that there is also a reasonable amount of divergence over time, which is where the concept of the premium between the 10Y ACGB and other markets enters the calculation.

Model based on 10Y UST yield and relative front-end pricing

Plank argues that there are several ways to approach modeling the premium between the 10Y ACGB and its UST and Bund counterparts. He points out that one could attempt to model it on the basis of economic fundamentals such as GDP, fiscal policy, current account deficits and so forth. However, he concedes that such an approach would fail the usefulness test from a market perspective.

Plank points out that the term structure theory of interest rates highlights why relative front-end pricing will matter for relative 10Y bond yields. The analyst argues that such a model for the 10Y ACGB works remarkably well if we use the 10Y UST yield and relative AUD/USD front-end pricing.

The analyst points out that the R-squared on a model for the daily 10Y ACGB/UST spread using just relative front-end pricing has been above 0.9 since the start of 2014. The Deutsche analyst notes for the 10Y ACGB/Bund spread, the R-squared is quite a bit lower at 0.65. Besides, the 10Y ACGB/UST spread vs. relative front-end pricing relationship is pretty stable over time.

Plank points out also that this model is very helpful in terms of explaining day-to-day moves in the 10Y ACGB, and one can reference the move in the 10Y UST and the shift in relative front-end pricing. He notes the China/U.S. growth gap could be a reasonable forward indicator for relative AUD/USD front-end pricing.