I want to investigate whether post-crisis saving levels are the same as pre-crisis. I understand how I can perform the chow test for two subsamples, but how does it work for three? 
In other words, I know that I can test whether 2008 is a breakpoint for period 1 & 2. I can also test whether 2010 is a breakpoint between period 2 & 3.
But how do I test whether the crisis temporally or permanently changed saving behaviour? Thus did savings (as a dependent variable) only change in period 2, or are there effects in period 3 as well?
