主题：【原创】谈谈时间序列的平稳性（1） -- 万里风中虎

It's a misspecified model. The simple model is ADF. And your ADF already demonstrated there is no trend or drift. It's just a pure random walk. I'm not sure if the conclusion changes if you add in cycle. More importantly, ADF clearly says you should not use OLS. If you do, results are not valid. Regressing a random walk on time t very likely will give significant coefficient. But it's spurious and misleading.