Abstract:

The thesis attempts to figure out if a simple yield curve can be used to predict
the stock market.
The yield curve is made up of two variables; the 10year treasury bond and the 3
month treasury bill. Due to the fact that the Norwegian 3month treasury bill was
only established in 2003 the regression analysis was split into two. A short and
a longterm. The annualized quarterly return on seasonally adjusted, chained
GDP was used as the dependent variable and the spread was the independent
variable in the regression analysis.
The results of the regression analysis, plotting GDP against the spread, indicated
that the yield curve could be predicting seven quarters ahead for the long and
ten quarters for the shortterm in Norway. In the US the yield curve seemed to
predict ten quarters in the long and sixteen in the shortterm. These
assumptions were based on the strength of the models.
The S&P 500 and OSEBX indexes were used as proxies, representing their
respective countries stock markets. The correlation test and the regression
analysis was performed to see if the stock market is able to predict future
changes in GDP. The stock market seems to be leading GDP by three quarters in
Norway and one quarter in the US. The tests were only performed using the longterm
scope.
In order to test the overall research question, a comparison test was designed.
The cumulative return on a risky, riskfree and Alternative portfolio were
compared. The predictive horizon of the yield curve and stock market was
decided on the basis of the results of the regression analysis.
At least one Alternative portfolio outperformed the stock market proxy in all
tests. It was, however, outperformed by the riskfree portfolio in the longterm
Norway comparison.
The only Alternative strategy portfolios that outperformed the stock market in
both the short and longterm were k3 and k9 in Norway and the US respectively.
The null hypothesis was rejected and the yield curve is able to predict the stock
market to a certain extent. 