Skip to main content

Table 3 Descriptive statistics

From: A survey-based estimation of the Swiss franc forward term premium

 

Unit

Mean

t stat

p50

Min

Max

sd

Count

Panel A: Interest rates

        

Libor 3M

%

1.91

 

1.34

− 0.90

9.41

2.27

306

Futures 3M

%

1.92

 

1.59

− 0.99

8.71

2.16

306

Futures 12M

%

2.17

 

2.04

− 1.06

7.86

2.02

306

Survey 3M

%

1.94

 

1.51

− 0.94

8.80

2.22

306

Survey 12M

%

2.19

 

2.04

− 0.79

7.88

2.05

306

Panel B: Excess returns and decomposition

        

Excess return 3M

pp

0.12**

2.55

0.04

− 2.00

2.34

0.48

306

Excess return 12M

pp

0.63***

3.89

0.56

− 2.57

3.56

1.02

306

Term premium 3M

pp

− 0.02

− 0.90

− 0.03

− 1.39

0.88

0.27

306

Term premium 12M

pp

− 0.02

− 0.30

− 0.05

− 1.91

1.29

0.41

306

Forecast error 3M

pp

0.14***

2.83

0.06

− 1.60

2.42

0.50

306

Forecast error 12M

pp

0.64***

3.86

0.53

− 1.88

3.28

1.00

306

Panel C: Explanatory variables

        

dispersion 3M

pp

0.13

 

0.12

0.02

0.51

0.08

219

dispersion 12M

pp

0.27

 

0.27

0.03

0.74

0.13

219

garch 3M

pp

1.07

 

1.02

0.27

3.37

0.60

306

garch 12M

pp

1.54

 

1.33

0.39

4.10

0.80

306

VIX

pp

19.67

 

17.63

10.64

62.47

8.07

306

Δ BCI

index

0.01

 

0.04

− 2.55

1.92

0.63

306

Δ Libor 3M

pp

− 0.03

 

0.00

− 1.19

1.19

0.24

306

Δ CHF real

index

0.00

 

0.00

− 0.10

0.08

0.02

306

  1. The table provides descriptive statistics for interest rates (panel A), excess returns, term premiums, forecast errors (panel B), and independent variables used in the regression analysis (panel C). The statistical significance of mean values for Panel B (column 3) is based on t-statistics (column 4). Heteroscedasticity and autocorrelation robust t-statistics are reported, using the Newey and West (1987) correction. The number of lags used equals the length of the contract (number of months). ***, **, and * denote statistical significance (two-tailed) at the 1%, 5%, and 10% significance levels, respectively. The dataset contains monthly observations from March 1991 through August 2016