In 2015, US output per capita was 12.5% below its 1950–2007 trend. This paper uses an estimated New Keynesian model to decompose the sources of this gap. The model features demographic trends, real and monetary shocks, and the occasionally binding zero lower bound on nominal interest rates. I calibrate demographic trends to observed mortality and fertility rates between 1940 and 2015, and estimate the model’s business cycle processes on quarterly data from 1984 to 2015. The model is successful in accounting for the post-1990 trends in the real interest rate, the employment-population ratio, and labor productivity growth. I extract the model’s structural shocks over the zero lower bound period and find that about half of the gap between output per capita and its long-run trend is due to the aging of the population, one fifth is due to real factors, one fifth to monetary factors, and roughly one tenth to the binding zero lower bound.
That is the job market paper from Callum Jones of NYU (pdf). Here is his broader portfolio. I would not model “secular stagnation” using New Keynesian models, but that granted it is remarkable how low is the contribution of the zero lower bound to the problem.