1: \begin{abstract}

2: {\footnotesize{}For a partial structural change in a linear regression

3: model with a single break, we develop a continuous record asymptotic

4: framework to build inference methods for the break date. We have $T$

5: observations with a sampling frequency $h$ over a fixed time horizon

6: $\left[0,\,N\right],$ and let $T\rightarrow\infty$ with $h\downarrow0$

7: while keeping the time span $N$ fixed. We impose very mild regularity

8: conditions on an underlying continuous-time model assumed to generate

9: the data. We consider the least-squares estimate of the break date

10: and establish consistency and convergence rate. We provide a limit

11: theory for shrinking magnitudes of shifts and locally increasing variances.

12: The asymptotic distribution corresponds to the location of the extremum

13: of a function of the quadratic variation of the regressors and of

14: a Gaussian centered martingale process over a certain time interval.

15: We can account for the asymmetric informational content provided by

16: the pre- and post-break regimes and show how the location of the break

17: and shift magnitude are key ingredients in shaping the distribution.

18: We consider a feasible version based on plug-in estimates, which provides

19: a very good approximation to the finite sample distribution. We use

20: the concept of Highest Density Region to construct confidence sets.

21: Overall, our method is reliable and delivers accurate coverage probabilities

22: and relatively short average length of the confidence sets. Importantly,

23: it does so irrespective of the size of the break.}{\footnotesize\par}

24: \end{abstract}