The name is absent



interest rates movements also at short horizons. Empirical evidence suggests that the predictive
power of the spread increases with maturity in both the single and in the threshold regimes22.


BETA reg1   95% conf int


BETA reg2 95%> conf int



BETA reg1 abs 95% conf i nt


BETA reg2 abs 95% conf int


Figure 7

In the single regime the slope coefficient never reaches one though. Threshold estimates of the
slope coefficient are quite close to
one; so that, in each regime the predictive power of the yield
spread is substantial. In particular, the 95% confidence interval at medium-long horizons contains
the expectations hypothesis value
one (the horizontal dashed line).

The advantage of employing a threshold model can be shown by a kernel density estimate23 of the
short term rate (
m = 3) distribution. In Figure 8 the shape of the density estimate makes clear the
convenience of adopting a multiple regime model, since the estimated empirical distribution of the
3-month yield in regime one is peaked (central panel); that is short term interest rates are easily
predictable. The left panel plots the kernel density for the 3-month yield in the entire sample, while
the right panel shows the empirical distribution in regime 2, characterized by high values of the
term premium. The standard deviation is much lower in regime one (below the estimated threshold)
than in the other regimes; whilst the kurtosis is much higher. In regime one, characterized by low
values of the term premium the empirical distribution of the short term rate is leptokurtic.

22 In regime two when the threshold is the absolute value of the term premium this result does not hold.

23 The Epanechnikov kernel density estimation is performed with automatic bandwidth selection. These results are
obtained by focusing on the pair of maturities (60, 3); similar results hold focusing on other couples of maturities (
n, 3).
In the entire sample the standard deviation of the 3-month yields is 2.62, the kurtosis is 5. In regime one (below the
estimated threshold) the std. deviation is 2.08 and the kurtosis is 7.18. In regime 2 (above the estimated threshold) the
std. deviation is 2.90 and the kurtosis is 3.80.

21



More intriguing information

1. Are Japanese bureaucrats politically stronger than farmers?: The political economy of Japan's rice set-aside program
2. Disturbing the fiscal theory of the price level: Can it fit the eu-15?
3. The name is absent
4. Reputations, Market Structure, and the Choice of Quality Assurance Systems in the Food Industry
5. ADJUSTMENT TO GLOBALISATION: A STUDY OF THE FOOTWEAR INDUSTRY IN EUROPE
6. Iconic memory or icon?
7. Modeling industrial location decisions in U.S. counties
8. The name is absent
9. Testing Gribat´s Law Across Regions. Evidence from Spain.
10. An Attempt to 2
11. Tariff Escalation and Invasive Species Risk
12. Fiscal Policy Rules in Practice
13. The name is absent
14. BILL 187 - THE AGRICULTURAL EMPLOYEES PROTECTION ACT: A SPECIAL REPORT
15. Secondary stress in Brazilian Portuguese: the interplay between production and perception studies
16. The name is absent
17. The name is absent
18. Evaluating the Success of the School Commodity Food Program
19. The Impact of Minimum Wages on Wage Inequality and Employment in the Formal and Informal Sector in Costa Rica
20. Confusion and Reinforcement Learning in Experimental Public Goods Games