The current stimulus spending package offers plenty of opportunities for pork-barrel, in fact one could even argue that it should be encouraged given the Keynesian policy to spend whatever the usefulness. This situation is, however, rather unique. Under "normal" times, do we observe periods of important pork-barrel spending, while there is little of it in others?
Marco Battaglini and Stephen Coate argue that there is a cycle of pork barrel spending that is correlated with public debt changes. Suppose there is a legislature with representative who care only about their district. When spending discipline is loose, they obtain a lot of pork. But once government debt reaches some level, fiscal discipline becomes predominant and pork spending is significantly reduced. Once the debt has become bearable again, porcine spending is in again. This argument seems rather simple and straightforward, so why would it be worthy of a publication in the American Economic Review?
It is in fact not straightforward to obtain such results. For example, it could be optimal for a government to accumulate assets to live off their interest. To avoid this scenario, one needs to have a legislature that does not behave like a social planner. Ways to do so include making representatives self-interested, or in the case of this paper, district-centered. Another reason to abandon districting (the first being gerrymandering).