By Pascal Michaillat and Emmanuel Saez
http://d.repec.org/n?u=RePEc:nbr:nberwo:24971&r=dge
This paper extends the New Keynesian model by introducing wealth, in the form of government bonds, into the utility function. The extension modifies the Euler equation: in steady state the real interest rate is negatively related to consumption instead of being constant, equal to the time discount rate. Thus, when the marginal utility of wealth is large enough, the dynamical system representing the equilibrium is a source not only in normal times but also at the zero lower bound. This property eliminates the zero-lower-bound anomalies of the New Keynesian model, such as explosive output and inflation, and forward-guidance puzzle.
I have to confess that I have a hard time with the innovative assumption of this paper. Do people really really care about their wealth beyond what it brings in terms of future consumption? OK, wealth brings status, but does this matter this much? And is this status wealth in government bonds? I would have thought that to rather be in oversized housed, expensive cars and golden toilets.
NEP-DGE Blog 
Spot on. This ranks as low as the money-in-the-utility-function models where holding cash is somehow valued over what you can do with the cash (and forgetting that money can be held in other ways, too).
If people overvalue housing wealth, then at least model housing wealth. Nobody overvalues government bonds. Financial institutions even dislike holding T-bills, seeing that they do not have them of their own will, but for regulatory reasons!
I am still upset over this paper.
Do the authors offer any good justification for this assumption? The paper itself has only some handwaving: a splatter of old references, some of their own. More importantly, there is no word on why specifically government bonds should be more relevant than any or wealth.
I hope they will provide an answer on this forum.