I had a discussion recently with the head of research at an investment bank and somewhere along the line we gravitated towards DSGE models. He personally didn’t like them but he agreed that’s where much of economic analysis is headed. Thus, I decided to write my first post here on them.
Economics is evolving. In Microeconomics, the game theoretic approach will gradually replace the classical analysis of individual, firms and markets, while in Macroeconomics, the Keynesian approach which revolutionized economics and birthed macro as we know it will give way to – DGE type analysis.
DGE or DSGE does seem imposing, like some acronym for a powerful highly intellectual word. DGE means Dynamic General Equilibrium and as we live in an uncertain world the ‘S’ in DSGE stands for stochastic. DGE is an attempt to solve one of the numerous paradoxes of macroeconomics and indeed economics – the existence of two seemingly mutually exclusive yet interconnected spheres of analysis.
It’s also historical as in much of the early 70s most of the theoretical positions of Keynesian economics when applied to data had poor predictive powers. They did well in explaining current events but whenever extrapolations were made they failed leading to widespread disillusionment. Now I must state here that I am a Keynesian myself.
Anyway this presented an opportunity for old classical economists now reformed into neo-classicals to seize the initiative. And this they did, beginning with Milton Friedman they struck back, but the most illustrative was Future Nobel Prize winner Robert Lucas. In his famed ‘Lucas Critique’ he essentially tore down Keynesian Macro. Keynesian Macro analyzed the economy in parts, its ‘equilibrium’ concept related to short run efficient allocations in the interactions between economic agents and their environments i.e markets. It was thus ‘static’ in that conclusions gleaned from this partial analysis and strung together using ad hoc theorizations to create models would fail to predict future events if salient parameters often untracked changed in between periods.
However, history held a solution, as circa 1928, an Englishman Frank Ramsey built small general equilibrium models starting with the individual at the centre. He sought to analyze the inter-temporal decision making process of the three economic agents; individuals with regards to consumption, labour supply and asset holdings. Firms determine the supply of goods and services consumed by individuals. In their investment decision making with regards to assets, capital and profits, they determine the quantity of labour demanded with implications for unemployment. Lastly, governments who print money, set taxes and raise debt to finance their expenditure and transfers.
Building on further developments on business cycles by future Nobel Prize winners Finn Kydland and Ed Prescott, DGE models were born. They are not static as they obtain parameters observed from data and use microeconomic postulates to calibrate economies before arriving at solutions. Equilibrium here refers to whether agents (idiosyncratically and in agglomeration) are on a stable path towards long run growth and not a short run demand and supply equilibration. Arising from this micro-foundation they form a more objective basis of explaining discrepancies encountered between macro and micro. They are the future and they are here.