1: \begin{abstract}

2: In this paper, we study the distribution of the so-called "Yule's nonsense 

3: correlation statistic" on a time interval $[0,T]$ for a time horizon $T>0$ ,

4: when $T$ is large, for a pair $(X_{1},X_{2})$ of independent 

5: Ornstein-Uhlenbeck processes. This statistic is by definition equal to :  

6: \begin{equation*}

7: \rho (T):=\frac{Y_{12}(T)}{\sqrt{Y_{11}(T)}\sqrt{Y_{22}(T)}},

8: \end{equation*}

9: where the random variables $Y_{ij}(T)$, $i,j=1,2$ are defined as  

10: \begin{equation*}

11: Y_{ij}(T):=\int_{0}^{T}X_{i}(u)X_{j}(u)du-T\bar{X}_{i}\bar{X_{j}},\text{ \ } 

12: \text{ \ }\bar{X}_{i}:=\frac{1}{T}\int_{0}^{T}X_{i}(u)du.

13: \end{equation*}

14: We assume $X_{1}$ and $X_{2}$ have the same drift parameter $\theta >0$. We 

15: also study the asymptotic law of a discrete-type version of $\rho (T)$, 

16: where $Y_{ij}(T)$ above are replaced by their Riemann-sum discretizations. 

17: In this case, conditions are provided for how the discretization (in-fill) 

18: step relates to the long horizon $T$. We establish identical normal 

19: asymptotics for standardized $\rho (T)$ and its discrete-data version. The 

20: asymptotic variance of $\rho (T)T^{1/2}$ is $\theta ^{-1}$. We also 

21: establish speeds of convergence in the Kolmogorov distance, which are of 

22: Berry-Esséen-type (constant*$T^{-1/2}$) except for a $\ln T$ factor. Our 

23: method is to use the properties of Wiener-chaos variables, since $\rho (T)$ 

24: and its discrete version are comprised of ratios involving three such 

25: variables in the 2nd Wiener chaos. This methodology accesses the Kolmogorov 

26: distance thanks to a relation which stems from the connection between the 

27: Malliavin calculus and Stein's method on Wiener space.

28: \end{abstract}