Data Flash (New Zealand)
Modeling the NZ/US 10Y bond spread
Summary
-------
* In modelling the 10Y NZ/US bond spread, NZ's current account deficit is a significant explanatory variable. Each
percent of the deficit adds some 16 bp to the spread (assuming the parameter estimates are unbiased).
* In the past 12 months, the negative impact of a larger current account deficit on long-term interest rates has been
more than offset by lower short-term interest rates.
* Perhaps of more interest is the fact that the model currently predicts a spread in excess of 150 bp. The actual spread
was just over 80 bp at the time of writing.
* The model's output is consistent with our recommendation that investors not have an exposure to the NZ market at
present. However, we are not as bearish about the prospects for the NZ market as our model.
NZ's current account deficit
----------------------------
* In the year to December 1999, NZ's current account deficit reached 8% of GDP. As the above chart shows, this is the
largest deficit since the mid-1980s.
* The deterioration in the current account position over the last two years reflects the impact of the Asian crisis, two
years of drought and a boom in imports brought about by strong consumption growth and a reduction in tariffs. The
deficit in the year to December 1999 also includes the import of the second ANZAC frigate. If this is excluded the
deficit would "only" have been 7.4% of GDP.
* The current account deficit has been identified as one of the key reasons for the poor performance of the NZD in
recent times. The deficit will also have an impact on the bond market via its impact on the risk premium for NZ assets.
* Interestingly, while the NZ current account deficit is at its widest for more than a decade the risk premium on NZ
bonds does not look all that high. The 10Y NZGB/UST spread presently sits close to 80 bp. Since NZ achieved low
inflation in 1991, the 10Y spread against the US has averaged slightly more than 100 bp. This seems to suggest the
widening of the current account deficit has had no impact on the risk premium built into NZ interest rates.
* Of course, the current account deficit is only one of many influences on the bond market. In the following section we
look at the impact of the current account on the bond market in more depth.
Modeling the impact of the current account on the NZ bond market
* In attempting to discern the impact of the current account deficit on the NZ bond market, we have modelled the 10Y
NZGB/UST spread as a function of the current account deficit and a number of other variables.
* The explanatory variables used are: the annual current account deficit as a percentage of GDP; the spread between
short term interest rates in NZ and the US; the percentage of bonds held by non- residents as per the RBNZ survey; and
the 10Y US swap spread. This last variable has been included as a proxy for the relative attractiveness of instruments
that investors might consider close substitutes to NZ bonds. The equation was estimated using quarterly observations and
averages and has an adjusted R-squared of 0.73. The spread predicted by the model is plotted against the actual spread
in the chart below.
* According to the model, each percent of GDP on the current account deficit adds 16 bp to the NZ/US 10Y bond spread.
This indicates that the current account deficit is a major contributor to the present spread of just over 80 bp.
* The fact the 10Y NZ/US spread has been able to trend lower over the past 12 months despite a larger current account
deficit reflects lower short term interest rates.
* Note that in looking at the current account deficit, we used both the absolute level of NZ's deficit and the
difference between the current account deficits of NZ and the US. The choice of variable had virtually no impact on the
results of the regression.
Model suggests the NZ bond market is extremely overvalued
---------------------------------------------------------
* Perhaps of greater immediate interest than the contribution of the current account deficit is the gap between the
actual 10Y bond spread and the predicted spread. This gap is presently at its widest ever level. Based on the current
values of the exogenous variables (including an assumed 8% of GDP current account deficit for the March quarter), the
model predicts a 10Y NZGB/UST spread of more than 150 bp. At the time of writing the actual spread was just over 80 bp.
* Much of the widening in the predicted spread has been driven by US swap spreads. At 115 bp, the 10Y US swap spread is
more than 30 bp wider than the average level over 1999. This adds some 50 bp to the predicted spread.
* To the extent that the widening of US swap spreads is related to specific US issues it may be of less relevance than
usual for other markets. However, at the margin it will still have an impact.
* The issue is whether the current level of swap spreads is expected to persist. As indicated, if swap spreads return to
the average level for 1999 then the predicted NZ/US bond spread narrows by some 50 bp. Coupled with the narrowing in the
current account deficit expected over the course of this year, this brings the predicted spread back towards 100 bp even
after taking into account the extent to which the RBNZ is expected to tighten relative to the Fed.
* Of course, even if all the NZ "friendly" events transpire the model predicts a spread in excess of 100 bp. From this
standpoint it's hard to argue the NZ bond market represents outstanding value at current levels. This is consistent with
our long standing recommendation that investors have no exposure to the NZ market.
* Despite this, we are nowhere near as bearish on the prospects for the NZ market as the model implies we should. While
in part this is due to our expectation that there will be NZ "friendly" events like an improvement in the current
account deficit, it also reflects our view that the absolute level of interest rates will attract investors. We have
already seen behaviour of this type in Australia and Canada. As rates rise to levels that domestic investors find
attractive, spreads to other markets may become less important in driving the NZ market.
David Plank
Fixed Income Strategist
This, along with an extensive range of other publications, is available on our web site http://research.gm.db.com
In order to read our research you will require the Adobe Acrobat Reader which can be obtained from their website
http://www.adobe.com for free.
For answers to your EMU questions, check Deutsche Bank's EMU web site http://www.db.com/emu or email our helpline
business.emu@db.com.