Over the past decade, DSGE models have gained popularity among macroeconomists. In the presence of theoretical interpretation, they can be fitted to data and used, for example, for forecasting or analysing optimal policy rules. To work with the model and fit it to data, it is necessary to solve the model, and there is a vast literature focusing on the solution and estimation of DSGE models (see, for instance, the review by Fernandez-Villaverde et al. (2016)). However, fewer papers consider non-stationary models, while part of the observed data is non-stationary in nature (GDP, investment, consumption, etc.). Models that can be reduced by transforming the variables into stationary ones (for example Fernandez-Villaverde and Rubio-Ramirez (2007)) are standard, but such a class leaves out many interesting models. Models in which the dynamics of trends is predetermined (and/or expectations about the dynamics of these trends are formed) have also been developed. Such models are usually solved using backward recursion, such as that of Kulish and Pagan (2014).