MATEO - Mannheimer Texte Online
Abstract
A key element of household behavior, saving is still not satisfactorily understood. The seminal life-cycle theory of saving by Modigliani et al has been augmented by the consideration of liquidity constraints, mortality, morbidity, income, interest, and other uncertainties, and, more recently, by behavioral elements. Each of these amendments explains part of saving behavior. However, contradictions remain widespread. The analysis of different models of life cycle saving decisions departs from the age-saving profiles observed in Germany and compares them with age-saving profiles that were simulated using different variants of augmented life-cycle saving theories based on empirical income processes estimated from German household data. This poses the question in how far observable age-saving profiles can identify competing theories of saving behavior. It shows that age-saving profiles per se cannot separate competing theories of saving behavior and that none of the simulated models can explain the German saving profile satisfactorily
Then I introduce a new model which allows for both life-time and income uncertainty. My main findings are, first, that allowing for mortality risk improves the life-cycle model’s slightly, and second, that simulated saving rates still fail to match their empirical counterparts.
Because the solution of the saving models are quite difficult to compute and people perhaps are not able to solve them, in the final part of this text I investigate how intertemporale optimization is better than any rule of thump.
I analyze life-cycle savings decisions when households use rules of thumb rather than solve the underlying intertemporale optimization problem. The decision rules I use are a simple Keynesian rule where consumption follows income; a simple consumption rule where only a fraction of positive income shocks is saved; a rule that corresponds to the permanent income hypothesis; and two rules that have been found in experimental studies. Using these rules, I simulate life-cycle savings decisions based on empirical income processes. Under a CRRA utility function, I then compute the utility loss incurred when some rule of thumb is used relative to the optimal decision rule given by the backwards solution of the intertemporale optimization problem. My central finding is that the utility losses induced by non-optimal rule-of-thumb behavior are relatively low. This observation might rationalize why non-optimal behavior is persistently observed in many real world situations: Behaving optimally is not only costly, it is also not much better than using simple rules of thumb.
Volltext in pdf-Format. Zum Lesen benötigen Sie den ACROBAT-Reader.
[MATEO Übersicht Dissertationen] [MATEO Homepage]