Economic forecasting

Building a macroeconomic model shouldn’t be a one-off process. But for some countries, investing in complex macroeconomic models proves to be just that: they simply don’t have the skills or institutional capacity to ensure that the carefully constructed model can remain at the heart of policy-making. Ensuring that models are used by their recipients requires a deeper understanding of the developing country context.

The simple truth is that economic forecasting is not just about the computing power of the model and crunching the numbers. It’s equally about ensuring that the processes and principles of forecasting are embedded into policy formulation, to help support intra-departmental working and transparent decision-making.

That means ensuring the country has a cadre of economists and analysts that understand how the forecasting model operates and what data it requires, as well as securing senior ministerial support for the model. In Pakistan the development of the forecasting model took in place in parallel with the establishment of a macroeconomic policy committee, bringing together high level officials from different departments involved in economic policy. By understanding the institutional context in this way, the model is now at the heart of robust discussions within the committee about macroeconomic direction.

One vital step in securing this buy-in is through a process of co-development: building the model with those who will be expected to use it, not for them. This is at the heart of OPM’s approach and ensures that the models created will be used by the government well beyond the completion of the project. Working side-by-side with government also provides a far more detailed insight into the economy of the country and improves the model’s accuracy and usefulness.

For example, instead of relying on generic assumptions around different variables, the co-development process - drawing on local knowledge and OPM’s experience and expertise - allows the model to include complex country-specific behavioural factors and examine the socioeconomic impact of macro changes: the potential effect on income distribution or poverty levels in rural areas.

Not only does this mean the model is more valuable in the immediate term, but because those tasked with using it have a comprehensive understanding of how it was developed, it makes the forecasting model more sustainable in the longer term.