Now that we have taken the initiative to start this blog, I cannot resist the opportunity to explain how economic modelers worked in the old days, so that all of us realize how close we believe we are to an innovative breakthrough. By developing iSimulate @ World Bank, and more importantly, by pressing for a free public release on the Internet, our young team has loaded a big responsibility on its shoulders. iSimulate is a big step into new uncharted territory. Global macroeconomic modeling will never be the same again.

In the old days, global macroeconomic modeling was a highly specialized activity. Multi-year projects in international organizations or sizeable economic institutions would produce large economic models that were clearly identified with the institution that created them: Multimod at the IMF, Interlink at the OECD, MCM at the Federal Reserve to name a few. Creating these models comprised more than just formalizing the underlying economic theory or estimating the equations or testing simulation properties. A lot of effort also went into constructing data systems with a logical nomenclature, improving solution algorithms and designing interfaces for input and output.

Most often all these organizations went through a development cycle of sorts:

Interestingly, every institution would struggle through this cycle independently!

So then, why do I say that iSimulate @ World Bank is a breakthrough? Not because I feel that there is finally a model that will last forever. And not because everybody can now use the same model with the same data. On the contrary, iSimulate will provide the opportunity to create new models much more frequently than ever before and will make it possible to work with many different models at the same time. I feel the real breakthrough is in how models will be created and how they will be used in the future.

iSimulate @ World Bank already has three key characteristics that define this next generation of macroeconomic modeling:

1. Models are being built backwards: With a strong focus on user interaction, models are now being built with easy-to-use interfaces with quick access to up-to-date information. Underlying mechanisms are being gradually improved with the cost-benefit analysis for the user in mind.
2. Models are shared before they are ready (and in my opinion, they actually will never be ready). Triggering interest from others and sharing initial ideas is now considered to be more important than internally perfecting systems that are clearly identified with the institution.
3. And most importantly, the modeling can now be done in cooperation across institutions and with people from different parts of the world. Technology can now provide a common environment for simulations with the ability to share results easily.

Gauresh and David might argue that I don’t have enough imagination; that I am thinking too narrowly. In their dreams they are already beyond macro-economic models and see applications in climate models or other areas. As my team’s responsibility is mainly global macroeconomic analysis, I will tend focus their efforts in this area. I feel there is much to be done even in this seemingly narrow application.

The ultimate success though will not depend on what we in the World Bank can do (that will always be limited), but on the reactions this initiative will trigger outside the Bank. I have no idea about the latter, but I am very curious to find out.

Hans Timmer is a Manager at the World Bank and Project Lead for iSimulate @ World Bank