The Influence of Confidence and Social Networks on an Agent-Based Model of Stock Exchange

Table 3

Statistical summary.

Returns (r)	N	Standard deviation	Excess kurtosis	Skewness	Jarque–Bera normality test
No network effect	1000	0.012	0.005	−0.033	0.178 (ns)
Barabási	1000	0.024	3.624	−0.324	564.690
Random	1000	0.027	3.767	−0.624	656.368
Small-world	1000	0.028	2.670	−0.583	353.662
Real data
Dow Jones	3320	0.010	23.170	−0.934	74727.3
Small caps (Dow Jones	3409	0.013	13.802	−0.959	27572.3
Large caps (Dow Jones	3409	0.010	17.718	−0.899	45039.4

The null hypothesis of normality is rejected at 1% level; ns represents that the null hypothesis is not rejected.