top of page

Behind The Curtain v2.0

The Great Financial Crisis (GFC) shook the economist profession since no macroeconomic model in the aughts (or ‘noughties’ in British English) came close to anticipating the crisis. It prompted a profound introspection process (see Behind The Curtainin 2018), which is still ongoing.


The stakes are high. Macro projections influence monetary and fiscal policies as well as investment preferences and decisions. Their quality affects nothing less than global capital allocation efficiency.


Given the world's complexity, it would be naïve to expect macro models to perform with a high degree of accuracy. In addition to the issue posed by geoeconomics, feedback loops (economic agent behavior > economic activity > economic agent behavior) represent a significant challenge. This complication sets macro forecasting apart from weather forecasting, which it is often (unflatteringly) compared to.


Despite the deep flaws identified after the GFC, economists decided to stick to so-called Dynamic Stochastic General Equilibrium’ or ‘DSGE’ models with a commitment to improving them. DSGE models rely on the aggregation of microeconomic models or ‘microfoundations’ based on the presumed rational behavior of representative agents (households, firms, governments).


One of the critical issues in DSGE models is the assumption that representative agents behave homogeneously. In response, economists have been refining their modeling approach by assuming heterogeneous representative agents.* Depending upon their wealth, some household categories are more sensitive to unemployment, interest rates, or inflation than others. Finally, macro forecasting is taking inequality into account.


But many economists have criticized this type of fix to DSGE models. Continuously adding sub-models to ‘enrich’ DSGE models only leads to Frankensteinesque tools with poor forecasting power. A damning observation was made by economists at the ECB: ‘Most [DSGE] models fail to coherently explain up to 80% of key macroeconomic variables.’


What is the alternative? Large-scale macroeconomic forecasting models based on past correlations of economic variables. Regrettably, they cannot be reconciled with microeconomic theory!


To sum up, macroeconomists are pedaling in yogurt. The entire economic system relies on forecasts generated by flawed macroeconomic models. That is even before internalizing the cost of externalities (pollution, biodiversity loss, social fabric deterioration, etc.) and before accounting for the impact of weather forecasts on economic forecasts (and, potentially, vice versa.)


The upshot is that the economic consequences of human actions, starting with monetary and fiscal policies, cannot be determined. It follows that the world is unmanageable.


In this humbling context, macroeconomic forecasts can only be one source of information to be taken with a truckload of salt when making executive decisions. The ability to read a firm’s revenue trends with granularity and conviction about long-term business trends are much more reliable allies for executives.


*This gave birth to ‘Heterogeneous Agent New Keynesian’ or ‘HANK’ models

Recommended readings:

114 views0 comments

Recent Posts

See All

12 Layers Of Complexity v3.0

Here is an updated version of the proposed ’12 Layers of Complexity’, which executive teams must contend with. Since the last version in 2022, many trends have hardened while themes have crystallized.


Genius has long been associated with the idea of a superpower that transcends human nature and may reveal the existence of a divine or, when operating on the dark side, evil force. But the definition

(Product) Portfolio Management

In ‘The Brussels Effect’ (2020), Columbia Law professor Anu Bradford convincingly presents the European Union as a ‘peaceful and quiet’ regulatory global superpower. If the United States relies on pri


bottom of page