In Carl Hommes 2015 book on Expectations, it seems he considers DSGE models (being it either RBC DSGE or New Keynesian DSGE) to be linear, or (log)linearized models, on page 3 of the introduction. He is making a distinction between the non-linear view and the random exogeneous shocks view (with a linear economic model), with both DSGE models and old Tinbergen models being examples of the latter, if I read his paragraph correctly. What are your thoughs about this?
Linear macro-economic models: examples and implications of the system being linear(ized)? Is the model non-linear from the moment one equation contains a non-linearity?
extract from a book by Hommes, 2013:
1.1 Economic dynamics, nonlinearity and complexity Economic dynamics is concerned with modeling fluctuations in economic and financial variables, such as commodity prices, output growth, unemployment, interest rates, exchange rates and stock prices. Broadly speaking, there are two contrasting views concerning the main sources of economic fluctuations. According to the first, business cycles are mainly driven by “news” about economic fundamentals, that is, by random exogenous shocks to preferences, endowments, technology, firms’ future earnings or dividends, etc. These random shocks typically act on an inherently stable (linearized) economic system. This view dates back to the 1930s, to Frisch, Slutsky and Tinbergen, who showed that a stable linear system subject to an irregular sequence of external, random shocks may produce fluctuations very similar to those observed in real business cycles. The linear, stable view was criticized in the 1940s and 1950s, mainly because it did not offer an economic explanation of observed fluctuations, but rather attributed those fluctuations to external, non-economic forces. As an alternative, Goodwin, Hicks and Kaldor developed nonlinear, endogenous business cycle models, with the savingsinvestment mechanism as the main economic force generating business fluctuations. According to this nonlinear view, the economy may be intrinsically unstable and, even in the absence of external shocks, fluctuations in economic variables can arise.
early Keynesian nonlinear business cycle models, however, were criticized for at least three reasons. Firstly, the limit cycles generated by these models were much too regular to explain the sometimes highly irregular movements in economic and financial time series data. Secondly, the “laws of motion” were considered to be “ad hoc,” since they had not been derived from micro foundations, i.e., from utility and profit maximization principles.Athird important critique was that agents’ behavior was considered as irrational, since their expectations were systematically wrong along the regular business cycles. Smart, rational traders would learn from experience to anticipate these cyclic movements and revise their expectations accordingly, and, so the story goes, this would cause the cycles to disappear. These shortcomings triggered the rational expectations revolution in the 1960s and 1970s, inspired by the seminal papers of Muth (1961) and Lucas (1972a and b). New classical economists developed an alternative within the exogenous approach, the stochastic real business cycle (RBC) models, pioneered by Kydland and Prescott (1982). RBC models fit into the general equilibrium framework, characterized by utility-maximizing consumers, profit-maximizing firms, market clearing for all goods at all dates and all traders having rational expectations. More recently, New Keynesian Dynamic Stochastic General Equilibrium (DSGE) models have moved to the forefront of macroeconomic modeling and policy analysis (Clarida et al., 1999;Woodford 2003). Typically these DSGE models are log linearized and assume a representative rational agent framework.Arepresentative, perfectly rational agent nicely fits into a linear view of a globally stable, and hence predictable, economy. By the late 1970s and early 1980s, the debate concerning the main source of business cycles seemed to have been settled in favor of the exogenous shock hypothesis, culminating in the currently dominating DSGE macro models for policy analysis. "
I am a bit confused here, as "real" business cycles are highly irrigular and unstable? Or do they mean the RBC modelling movement?
Where can I read more about this?