The main exogenous assumptions concern economic growth, demographics, international fossil fuel prices and technological developments.... Demand for primary energy serves as input for the supply modules.
As a modeller myself, I've always complained bitterly about this structure. In effect, it allows the people using the model to: (1) assume a politically acceptable level of growth; (2) work backwards to the supplies and prices of energy necessary to produce that growth; and (3) assign production levels to various sources in order to produce the necessary supplies and prices. In past years the IEA assigned large amounts of supply growth to the OPEC countries. Now that OPEC has suggested that they won't be providing large increases in production, the IEA forecasts that tight oil in the US (plus natural gas liquids) will provide the needed increases. What's missing in this picture? There should be a feedback loop that links primary energy production costs to supplies and prices. Producing a million barrels per day over decades from tight US formations such as the Bakken require lots of new wells to be drilled essentially forever. Money spent on drilling is not available to the rest of the economy. Energy supplies and prices need to be a part of the economic model, not specified outside of it.
Over the last few years, the IEA forecasts have shown a lot of change from year to year. This year there's a big swing in oil (and near oil) production away from OPEC to the US. Big swings don't give me much confidence in the underlying models.
 Credit: Steven Kopits of Douglas-Westwood, testifying before the US House Subcommittee on Energy.