Jonathan Baron (Pennsylvania, Wharton School) & Edward J. McCaffery (USC) have posted Starving the Beast: The Psychology of Budget Deficits on SSRN. Here is the abstract:

Many opponents of big government favor a strategy of "starving the beast," cutting taxes today with the expectation that spending cuts will follow tomorrow. Why might such a strategy work? Various heuristics and biases help to explain how it can. In two experiments conducted on the World Wide Web, subjects chose general levels of taxation and public spending from various hypothetical starting points. Subjects wanted to reduce both taxes and spending, preferring balanced budgets and even surpluses to deficits. When asked about specific spending cuts, however, subjects showed a marked reluctance to make cuts, leading to deficits. Subjects also showed an anchor and underadjustment bias, changing their responses in light of various baselines, and failing to completely close existing deficits. The "starving the beast" phenomenon, by pairing specific tax cuts with the general, abstract idea of spending cuts, can thus succeed in a population preferring fiscal balance. Once the deficit is created, it will likely persist, influencing future policy preferences.