Dual Formulation of the Optimal Consumption Problem with Multiplicative Habit Formation
Thijs Kamma, Antoon Pelsser · 2025
Read Dual Formulation of the Optimal Consumption Problem with Multiplicative Habit Formation by Thijs Kamma, Antoon Pelsser (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-13678.txt.
Introduction
Thijs Kamma and Antoon Pelsser reformulated a classic savings problem in 2025. An agent chooses how much to consume over time, but satisfaction depends on consumption relative to an internal habit built from past consumption, and here the habit enters multiplicatively, as a ratio. That ratio form destroys the problem's concavity and makes the present depend unavoidably on the whole path, so the usual Lagrangian route to a dual problem fails.
Using a duality theorem of Fenchel, the authors identify a candidate dual formulation and prove strong duality, meaning the two problems share the same optimal value. From that they derive relations converting optimal choices in one problem into optimal choices in the other. The dual view also yields an analytical way to bound how accurate any approximate solution to the original problem really is.
Catalog record
| CREATOR | Thijs Kamma, Antoon Pelsser |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 116 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-13678
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Dual Formulation of the Optimal Consumption Problem with Multiplicative Habit Formation Thijs Kamma, Antoon Pelsser Abstract This paper provides a dual formulation of the optimal consumption problem with internal multiplicative habit formation. In this problem, the agent derives utility from the ratio of consumption to the internal habit component. Due to this multiplicative specification of the habit model, the optimal consumption problem is not strictly concave and incorporates irremovable path-dependency. As a consequence, standard Lagrangian techniques fail to supply a candidate for the corresponding dual formulation. Using Fenchel’s Duality Theorem, we manage to identify a candidate formulation and prove that it satisfies strong duality. On the basis of this strong duality result, we are able to derive duality relations that stipulate how the optimal primal controls depend on the optimal dual controls and vice versa. Moreover, using the dual formulation, we develop an analytical evaluation mechanism to bound the accuracy of approximations to the optimal solutions. We are very grateful to An Chen, Tahir Choulli, Michel Vellekoop, Bas Werker, Ralf Werner and to the conference and seminar participants at the Netspar Pension Day (2021), Tilburg University (2021), Ulm University (2021), the International Pension Workshop (2022), the Winter Seminar on Mathematical Finance (2022), the 11^{\mathrm{th}} World Congress of the Bachelier Finance Society (2022), the 25^{\mathrm{th}} International Congress on Insurance: Mathematics and Economics (2022), the 5^{\mathrm{th}} European Actuarial Journal Conference (2022), the 14^{\mathrm{th}} Actuarial and Financial Mathematics Conference (2023) and the Doctoral Seminar in Sankt Gallenkirch (2023), for their helpful comments and suggestions. Keywords: Fenchel Duality, habit formation, life-cycle investment, stochastic optimal control, utility maximisation JEL Classification: C61, D15, D53, D81, G11 1 Introduction Habit formation describes the phenomenon of an individual growing accustomed to a certain standard of living. In a financial context, this standard of living is dependent on a person’s past decisions with regard to saving and consumption. Consuming more or less than a person-specific living standard may impact the utility levels of an individual, cf. Kahneman and Tversky, (1979) . It is therefore plausible that habit formation affects the current consumption behaviour of a person. To model and analyse the impact of habit-forming tendencies on this behaviour, a wide variety of studies have investigated optimal consumption problems that incorporate a habit level, representing the agent’s living standard. These studies can be distinguished into two categories: (i) those that focus on additive habits, and (ii) those that concentrate on multiplicative habits. We start by discussing the additive habits. In optimal consumption problems with additive habits, the utility-maximising individual draws utility from the difference between consumption and a habit level. The literature on these habits is pioneered by Constantinides, (1990) and has been studied by e.g. Detemple and Zapatero, (1991) , Campbell and Cochrane, (1999) , Munk, (2008) , Muraviev, (2011) , and Yu, (2015) . Additive habit models typically employ arithmetic habit levels, which monotonically increase over time, cf. Detemple and Karatzas, (2003) , Bodie et al., (2004) , and Polkovnichenko, (2007) . Furthermore, as most standard utility functions only admit strictly positive arguments, additive habit specifications force the agent to maintain consumption above the habit level. For this reason, the habit component is sometimes interpreted as a subsistence level, see e.g. Yogo, (2008) . This interpretation is sensible for exogenous habits. However, if we assume that habits are endogenous, the habit level depends on the individual’s past decisions and becomes person-specific. Consequently, for endogenous habits, it is hard not to consider the habit component as a standard of living that increases over time. Although individuals have a natural incentive to maintain consumption at least above their living standard, it is clear that additive habit models are too restrictive to be realistic. We attribute this restrictiveness to two main reasons. First of all, in practice, adverse changes in the financial circumstances can urge people to scale down consumption below the level to which they have become accustomed. Second, because of the latter phenomenon, an individual’s standard of living may decrease over the course of a lifetime. To arrive at a more realistic model setup that manages to deal with the preceding two situations, the following two modifications can be made. As for the possibility of a declining standard of living, one can employ a geometric specification of the habit level, cf. Kozicki and Tinsley, (2002) , Corrado and Holly, (2011) , and van Bilsen et al., 2020a () . Unlike the arithmetic habit levels, this geometric specification relies on the logarithmic transformation of consumption, and can therefore decrease over time. As for the possibility of scaling down consumption below the habit level, one can make use of multiplcative habit models. We now continue with a discussion of the multiplicative habits. Optimal consumption problems with multiplicative habit formation assume that the utility-maximising individual derives utility from the ratio of consumption to a habit level. The specification of these habits dates back to Abel, (1990) , and has been economically advocated by Carroll, (2000) and Carroll et al., (2000) . Contrary to the additive case, consumption is in this multiplicative setup not constrained to achieve values above the habit level. Namely, since the ratio of consumption to the habit level is always strictly positive, it can be included as an argument in all standard utility functions. The multiplicative habit model consequently allows the agent to reduce consumption levels below the habit component. Furthermore, the multiplicative habit model endows the utility-maximising agent with a strong incentive to fix consumption near/above the habit level. This incentive is due to the fact that the utility function of the agent increases with the magnitude of the ratio. When the habit level is endogenously determined (internal), standard solution techniques generally fail to solve optimal consumption problems with multiplicative habit formation in closed-form. Because of its dependence on past consumption decisions, the habit component gives rise to path-dependency in the objective function. This path-dependency is irremovable and cannot be handled in an analytical manner. 1 1 1 This analytical intractability is unique to problems involving multiplicative habits. In case of additive habits, the path-dependency can be eliminated from the problem, cf. Schroder and Skiadas, (2002) . Due to the structure of multiplicative habits, the optimal consumption problem is not strictly concave. In general, non-concave optimisation problems are more difficult to solve than concave ones, see e.g. Chen et al., (2019) . To be able to analyse the corresponding optimal solutions, the general approach is to fall back on (i) numerical routines, (ii) approximations or (iii) duality techniques. In a discrete-time setup, Fuhrer, (2000) and Gomes and Michaelides, (2003) employ numerical methods to analyse the internal multiplicative habit model. More recently, in a continuous-time setup, van Bilsen et al., 2020a () , Li et al., (2021) and Wang et al., (2024) have made use of both approximation and numerical routines. 2 2 2 We exclusively mention studies that focus on the consumption problem with internal multiplicative habits. Problems involving external habit formation, see e.g. Carroll et al., (1997) , Chan and Kogan, (2002) and Gómez et al., (2009) , do not pose issues when it comes to deriving optimal (duality) results. Martingale duality techniques, developed in the seminal contributions by Pliska, (1986) , Karatzas et al., (1987) ,