Time and Preference: Assessing Future Utility

ABSTRACT - In many purchase situations, a delay may exist between the transaction and consumption (e.g., catalog or Internet shopping). We examine consumers’ evaluations of future consumption relative to immediate consumption in situations when the delay is not under their control. The results of a study to assess purchase behavior and the form of discount rates indicate that, in contrast to findings from research in economics about discount rates for monetary amounts, consumers appear to discount product outcomes at an increasing rate over time. We also demonstrate that location influences the pattern of discounting; consumers discount expensive, Web-based purchases more rapidly than comparable, locally consummated purchases.

Citation:

Eloise Coupey and Erin Sandgathe (1999) ,"Time and Preference: Assessing Future Utility", in E - European Advances in Consumer Research Volume 4, eds. Bernard Dubois, Tina M. Lowrey, and L. J. Shrum, Marc Vanhuele, Provo, UT : Association for Consumer Research, Pages: 77-81.

European Advances in Consumer Research Volume 4, 1999      Pages 77-81

TIME AND PREFERENCE: ASSESSING FUTURE UTILITY

Eloise Coupey, Virginia Polytechnic Institute & State University, U.S.A.

Erin Sandgathe, Virginia Polytechnic Institute & State University, U.S.A.

ABSTRACT -

In many purchase situations, a delay may exist between the transaction and consumption (e.g., catalog or Internet shopping). We examine consumers’ evaluations of future consumption relative to immediate consumption in situations when the delay is not under their control. The results of a study to assess purchase behavior and the form of discount rates indicate that, in contrast to findings from research in economics about discount rates for monetary amounts, consumers appear to discount product outcomes at an increasing rate over time. We also demonstrate that location influences the pattern of discounting; consumers discount expensive, Web-based purchases more rapidly than comparable, locally consummated purchases.

INTRODUCTION

One characteristic of many purchases is the delay between the transaction and the delivery and consumption of the product, as for purchases where items are not in stock, catalog orders, and Internet purchases. Consumers’ evaluations of the desirability of such purchases may be affected by their reactions to the delay of consumption, an act which may require consumers to exercise adequate self-control to forego a more immediate opportunity to consume with a potentially more expensive local purchase than with a time-distant purchase.

This characteristic of many shopping situations underscores a need for research to examine how people assess the value of an item that may be purchased, depending on when they actually receive the product. In economic literature, issues of valuation over time have been the focus of research on intertemporal choice. Very simply, intertemporal choice involves the tradeoffs that consumers make between costs and benefits of consumption at different points in time (Loewenstein and Elster 1992). These tradeoffs, and the factors that influence them, may have implications for developing effective pricing strategies for delayed consumption purchases.

CONCEPTUAL BACKGROUND

Decisions that involve weighing the consequences of waiting to consume may be generally characterized as decisions involving tradeoffs. For instance, deciding to order a product from the Internet may invoke economic concerns of trading off the value of having the money available for other consumptionBhence, a cost in terms of liquidityBagainst the anticipated benefits to be experienced when the product is obtained. A large literature on self-control has tackled issues associated with the psychological formulation of these tradeoffs, such as, "If I eat this third piece of cheesecake tonight, I’ll have to jog fifty extra miles next week" (c.f., Ainslie 1975). These researchers have attempted to determine whether there are factors that systematically influence preference for outcomes at different points in time (e.g., Mowen 1992: present v. future outcome framing and risk preference; Wertenbroch 1998: perceptions of products as vices or virtues). For example, research on self-control indicates that the decisions of pigeons and of people exhibit predictable, adaptive tendencies (Mazur and Logue 1978, and Hoch and Loewenstein 1991, respectively).

Intertemporal choice as described by economists often relies on the notion of discounting; that people exponentially discount the utility of consuming at later points in time (Rabin 1998). A normative approach to intertemporal choice (e.g., Fisher 1930; Fishburn 1970) generally assumes that people exhibit positive time preference, underweighting, or discounting, the utility of consumption in the future relative to consumption in the present. The exponential aspect of the models implies that people are time-consistent in their preferences; one would feel just as favorably toward a consumption tradeoff now as in a year from now.

Tradeoffs over time are often reflected as discount rates (e.g., Thaler 1981). Among early descriptions of discounting is the discounted utility model (DU) developed by Samuelson (1937). In DU, a high discount rate indicates rapidly diminishing utility, and a low discount rate indicates slowly diminishing utility. A discount rate is the ratio between one unit of consumption later to one unit of consumption now; that is, the marginal rate of substitution between now and later.

Evidence from economics, combined with a growing literature in marketin (e.g., Hoch and Loewenstein 1991; Marshall, Mowen and Stone 1995; Mowen and Mowen 1991; Wathieu 1997), suggests that while DU may often provide a summary view of choices across time, it may not adequately describe the patterns of discount rates often observed in economics (cf. Loewenstein 1992). Several economists have suggested that non-constant changes in discount rates are ubiquitous, and that they reflect psychological rather than economic influences on discounting behaviors. For example, Strotz (1955) speculated that psychological perceptions of time result in the heightened valuation of proximal outcomes relative to distal outcomes, an expectation borne out by the appearance of high implicit discount rates in purchases of durables (e.g., Hausman 1979). Strotz’ intuition has also been empirically supported by research that demonstrates a non-linear discount rate for the valuation of an outcome’s utility at different points in time (e.g., Thaler 1981; Benzion, Rapoport and Yagil 1989).

Outcomes as Products v. Dollars

Consistent with Strotz’ intuition, economists have tended to find that people’s discount rates are larger earlier in a time span and that the rates decrease as time progresses. The tendency to overweight the present in terms of outcome evaluation is indicated by research in self-control across a variety of behaviors (e.g., reward v. punishment effects on choice (Mischel and Grusec 1967) and savings behavior (Thaler and Shefrin 1981)). A similar result has been observed in the health domain, in which both magnitude and delay effects have been observed (Chapman 1996).

For decisions about product outcomes, the emphasis on the present may be particularly pronounced. Although a consumer’s utility for receiving a monetary amount at different points in the future may vary, the desire to have the money is unlikely to disappear. Thus, patterns of discount rates for money tend to similarly exhibit decreasing rates over time, characterized as decreasing impatience (Ainslie 1975), often in the form of hyperbolic discounting. In contrast, a consumer’s utility for a product may be time-dependent; the rate of discounting may increase to reflect the disappearance of the desire or need to consume the product. In essence, we would expect to observe discounting consistent with increasing impatience when product utility diminishes; that is, when the reward effectiveness (Ainslie 1975) declines with time. We can look at willingness to wait for the product as a function of the differential valuation placed on the product via discount rates. In this sense, differential valuation reflects changes to perceptions of reward effectivenessBitself a function of comparisons between the present situation and the anticipated situation.

A second difference between monetary outcomes and product outcomes is that for products, the value of future consumption may be the result of many factors. For example, Loewenstein (1987) suggests that people may place value on outcomes by integrating not only the outcome itself, but also the utility associated with being able to anticipate (or dread) the outcome. Although his reasoning is focused on the added utility of the time until consumption and may also be pertinent to monetary outcomes (e.g., income taxes and refunds), it suggests that outcome assessment is an accumulation of concerns beyond mere asset position evaluation. Hoch and Loewenstein (1991) develop a model of consumer self-control in which they consider factors that may account for time-inconsistent preferences. Several of the factors that they propose may influence choices over time may also lead to an increasing pattern of discount rates for products (e.g., temporal proximity and social comparison). For example, buying a pair of athletic shoes of the brand owned by all of your friends may only retain the status of a prepotent need as long as your friends are still wearing their shoes.

Situational Characteristics and Discount Rates

Nonlinear change in discount rates over time may reflect individual differences due to individual utilities, rather than interest rate computations. These individual differences may reflect perceptions of the situation, perhaps combined with an attitude toward risk. While Fisher (1930) suggested that time introduced risk into a decision, with more distant times regarded as riskier, more tangible aspects of a purchase situation may also be related to risk perception.

Dodds, Monroe and Grewal (1991) examine the use of extrinsic cues to guide behavior in different contexts. The researchers demonstrate that perceptions and behaviors regarding products are influenced by information about price, store, and brand. For example, they present results which indicate that available information about the store and the brand exerts a greater influence on consumers’ perceptions of quality, value, and willingness to buy when the store and brand are viewed as of higher quality, rather than lower quality. Their results suggest that the store and brand characteristics serve as risk-inducing or risk-reducing aspects of the decision.

RESEARCH HYPOTHESES

The preceding discussion may be summarized as a series of hypotheses intended to provide insights into consumers’ perceptions of the utility of product consumption, and how these perceptions, reflected as discount rates, are predicted to change as a function of time, cost, and location. For instance, similar to the findings of numerous researchers of intertemporal choice, we anticipate that discount rates for products will exhibit some similarities to discount rate patterns for money outcomes. More specifically, we expect to observe non-constant discounting, with a higher mean rate of discounting for cheaper products than for expensive products.

Based on our reasoning that reward effectiveness for products diminishes as the time horizon increases, we also predictCcontrary to typical findings using outcomes in terms of moneyCthat discount rates for product outcomes will increase over time. That is, the products will lose value to consumers at a faster rate as time passes.

Tying the previous research on time discounting to marketing research on extrinsic cues, we propose that differences in the patterns and sizes of discount rates may be observed as a function of location. In choice over time, location may reflect consumers’ concerns about implicit risk; that is, some locations may increase the probability of undesirable outcomes (e.g., product not received; product not as anticipated).

METHOD

Subjects and Procedure

One hundred and thirty-eight business students in a large, southeastern university completed the pencil and paper questionnaire for course extra credit.

The questionnaire contained a training problem to familiarize subjects with the concept of discount rates, two scenarios with accompanying requests for discount rate information, and several manipulation checks. Subjects completed the study, on average, in fewer than thirty minutes.

Stimuli

The stimuli were brief descriptions of hypothetical purchases. Subjects were asked to assume that they were contemplating the purchase of one of two products and to provide information about how much they would pay for thefocal product, given that they would not receive it until one of seven points in the future. They were given a present value for the product, and they were told that the payment was immediate, although delivery was not.

Design

The study was a mixed design, with two factors manipulated within-subjects and two factors manipulated between-subjects.

Independent Variables. The within-subjects factors were cost and time. Each subject received two scenarios. In one scenario the product was valued at $15, while in the other scenario the product was valued at $300. The selection of these values was based on previous research conducted to examine the effect of cost on discounting (e.g., Thaler (1981); Benzion, Rapoport and Yagil (1989). The second within-subjects factor, time, was included to assess the pattern of discount rates. Based on previous studies of rates and on the types of products used in this study, a time span of 64 weeks was used. In addition to the present time, subjects were asked about the value of the product to them at each of seven points within the time span (1 week, 2 weeks, 4 weeks, 8 weeks, 16 weeks, 32 weeks, and 64 weeks).

The between-subjects factors were location (WWW or local) and product (wine or compact disc (cd)). Internet purchases may be viewed as risky, in the sense that the Internet provides a novel context for consumption (Coupey 1999). Costs to set up shop may be lower, thus decreasing barriers to entering the market and creating consumer fear of here-today, gone-tomorrow businesses. Additional concerns include security and privacy, each of which may heighten implicit risk in purchasing via the Internet.

Subjects were randomly assigned to one of the four experimental conditions. Within each condition, the order of the cost (i.e., cheap or expensive) was counterbalanced across subjects.

Dependent measures. The primary dependent measures are the discount rates for each time period. Seven discount rates were calculated for each scenario, for each subject (i.e., 14 rates per subject). The discount rate was inferred from the value data provided by each subject, using the following formula:

R=(F/P)^1/t B 1,

where R is the inferred discount rate, and F and P are the future and present values, respectively. In the study, t is measured in weeks.

RESULTS

To assess sample characteristics, we obtained information from each subject about his/her age, experience (number of purchases), and omfort with the Internet (-5: very uncomfortable to +5: very comfortable.

Ages ranged from 19 to 40, with a modal age of 20. Sixty-two percent of the subjects were under 21. Fifty-five percent of the subjects were male. Subjects indicated greater levels of comfort with the Internet than discomfort, with an overall mean of 2.6 on a B5 to +5 scale. Over half of the sample claimed to have made purchases from the Internet (53%).

In the following section, we tested hypotheses about the outcome valuation over time, casting time as one factor that may influence the decision to purchase (or not to purchase) products through the Internet.

Hypothesis Testing

Manipulation checks. Several analyses were conducted in order to determine whether the manipulations were effective. For the within-subjects factors, cost and time, subjects were asked to indicate the longest amount of time they would be willing to wait before they would not be willing to pay anything. If the waiting times differ significantly, we can infer that the perceptions of costs also differ.

Comparing the mean numbers of weeks that subjects were willing to wait for both a cheap and an expensive product, the results of a paired samples t-test suggest that subjects did view the costs differently. The comparison for the times before the product was valueless did differ significantly according to cost (t₍₁₃₇₎=-5.51, p<.000; X_cheap=14.13, X_expensive=24.65).

The waiting time data were also used to assess the appropriateness of the time span used in the study. Our goal was to determine whether the 64-week span was adequate to capture the preponderance of the variance in a sequence of discount rates. Using the mean numbers of weeks from the paired samples t-test, and adding one standard deviation to each mean results in 14.13 + 20.63=34.76 for a $15 product, and 24.65 + 30.57=55.22 for a $300 product. Thus, the 64-week span appears appropriate for the costs.

We also examined whether the Internet was perceived differently from a local shopping venue. To accomplish this, we analyzed subjects’ responses to six questions about the similarity between four shopping venues: the Internet, catalogs, local shopping, and distant shopping (described as 100 miles away). The responses were analyzed with a multidimensional scaling procedure, Alscal, using an Euclidean distance model. The analysis resulted in a two-dimensional solution, with RSQ=.997. The results indicated that subjects do perceive Internet shopping as different from local shopping, although not as different as catalog shopping or shopping in a distant location.

Order and Product Effects. A MANOVA was conducted to determine whether the order of presentation of the cheap and expensive products influenced the discount rates. There was no significant main effect of the cost order variable (p<.97), and it did not interact significantly either separately or jointly with cost, time, or location. A similar result was obtained for the product factor (p<.48).

HYPOTHESIS TESTS

First, we tested the replicatory prediction that subjects’ discount rates over time would not follow the constant, exponential discounting form suggested by discounted utility theory. To assess this hypothesis, we completed a multivariate analysis of variance that tested for differences in the inferred discount rates. Fourteen rates were computed for each subject, seven for each scenario. These rates were used as the dependent measures. The general linear model included the effects of time and cost as within-subjects factors, and location (WWW or local) and product (wine or compact disc) as between-subjects factors.

The effect of time was significant (F_{(6, 129)}=22.3, p<.0001). This result indicates that discount rates vary across time, and that the rate of discounting is not constant. Thus, we replicate the findings of previous research (e.g., Benzion, Rapoport and Yagil 1989), demonstrating their applicability in a product consumption context.

Next, we examined the effect of cost on the expressed discount rates. Recall that previous research has demonstrated higher discount rates for lower monetary amounts than for higher monetary amounts. We sought to determine whether this result would be replicated using products, rather than monetary outcomes. We anticipated that for outcome magnitude, products would be similar to money; that is, cost differences would result in greater discounting for cheaper products, relative to more expensive products, suggesting a significant main effect of cost.

Using the analysis and model described previously, we obtained a significant effect of cost (F_{(1, 134)}=84.38, p<.0001). Subjects tended to discount the cheaper product more than they discounted the more expensive product. The average discount rate for the cheap product was -.35. For the expensive product, the mean rate was -.17.

We also examined the interaction effect of cost and time to determine whether the effect of cost was manifest at different points in time. The interaction was significant (F_{(6, 129)}=3.80, p<.002). Expensive products were discounted at a lower rate than cheaper products, thus providing a replication of previous research (e.g., Thaler 1981) with a statistical test of the differences in discount rates as a function of cost and time. The means and standard deviations for the cheap and expensive products are provided for each time period in Table 1.

TABLE 1

MEAN DISCOUNT RATE BY COST AND TIME

TABLE 2

MEAN DISCOUNT RATES BY LOCATION, COST AND TIME

We also tested the prediction that while outcome valuation might be significantly affected by cost and time factorsCconsistent with previous research in economics (e.g., Benzion, Rapoport and Yagil 1989)Cthe pattern of discounting would differ. More specifically, while economists predict a pattern of discount rates that diminish at a decreasing rate (i.e., the discount rate varies inversely with the time to be waited), we proposed that for products, discounting would occur at an increasing rate (i.e., the longer consumers wait, the greater the discount rate). The mean discount rates for each time period, presented in Table 1, clearly demonstrate increases in the rate of discounting with increases in the consumption horizon.

We also predicted an effect of location, such that the patterns of rates observed as a function of cost and time would differ, depending on the source of the purchase. The pattern of rates for expensive Internet purchases was expected to resemble the pattern of rates for cheap purchases, regardless of their location. A three-way interaction of cost, time, and location was observed (F_{(6, 129)}=2.63, p<.02). Again, inspection of the mean discount rates as a function of cost and location, over time, indicates a steeper rate of discounting earlier in the time frame for expensive Internet purchases, a tendency less evident in expensive, but local, purchases. The means are presented in Table 2.

CONCLUSION

Summary and Discussion

A study was conducted to examine how consumers discount product outcomes as the time to consumption increases. We applied theories of intertemporal choice to examine how subjects discounted several types of consumption opportunities as a function of cost, time to delivery, and location.

We demonstrate that outcome valuation for products differs across time, consistent with the results of research on intertemporal choice in behavioral economics. This conceptual result adds to the extant literature by providing an application of discounting using a non-monetary outcome. More importantly, however, we prediced that the pattern of discount rates for product would differ from the pattern typically observed in studies with monetary outcomes as stimuli. Discount rates for money amounts tend to decrease over time, suggesting decreasing impatience. In contrast, discount rates for the products we examined tended to increase over time, indicating increasing impatience.

The rate of increase in the discount rates has implications for policy makers and for marketers. For example, the rapid devaluation of expensive products purchased through the Internet suggests that consumers are generally unwilling to trade off the anticipated gratification of consumption against the perceived risks associated with a distant consumption horizon of unknown reliability.

Our results also suggest that attempts to influence purchase by manipulating price should take into account consumers’ discount rates whenever there is a likely delay between payment and delivery/consumption. For example, in developing an effective Internet pricing strategy, marketers should be cautious about the effect of the Internet on consumers’ willingness to accept a price. The results depicted in Figure 3 suggest that consumers will be willing to accept a product at a particular price for a shorter length of time to delivery when the product is expensive andBregardless of priceBsold through an Internet source.

Limitations

As with most studies of temporal discounting, we are limited in generalizability by the use of hypothetical scenarios. This limitation is tempered, however, by the use of products readily available and presumably familiar to the majority of our subjects.

The use of only two products, however, also limits our ability to generalize these results. Although the effect of the product factor did not significantly influence the other effects of interest (i.e., cost, time, or location), the products do share several characteristics. For example, the products are comparable in cost and are often used for entertainment. A wider variety of products would be desirable for assessing the external validity of our findings.

Future Research

These results provide insights into how consumers evaluate the utility of product outcomes over time, making tradeoffs between the present value of the product and the time to be waited before consumption. Additional research is needed to elucidate the factors that influence these tradeoffs. For example, the product outcomes we examine are discounted quite differently from money outcomes. While we demonstrate the form of the discount rates, further research is needed to determine why product value decreases at such high rates. Although economists have traditionally speculated that time introduces an element of risk that may influence rates (e.g., Strotz 1955), a pure risk explanation does not explain why products should be discounted differently from money.

Hoch and Loewenstein (1991) develop a reference point model of deprivation, in which consumers who adapt to the idea of owning a product, but who do not yet possess it, may experience deprivation. The desire to alleviate deprivation may result in time-inconsistent choices, in which the value of the outcome decreases at an increasing rate. Product outcomes, in contrast to monetary outcomes, entail a transaction of money for an item. When the payment is made, the consumer must trade off the "loss" of the money against the future "gain" of the product. If the consumer is experiencing deprivation, then the time the consumer is willing to wait may be relatively short. Whether such a reference point approach adequately reflects the reasons for which tradeoffs differ over time is a matter for future research.

On an applied level, marketers who sell through the Internet would be benefited by research to delimit factors and conditions under which time will influence the amount and rate of discounting. Thus, research on situational characteristic that affect consumers’ perceptions of transaction tradeoffs in Internet environments is needed.

Pricing is a fundamental concern for marketers and for consumers. Although issues of price are sometimes overshadowed by other concerns, such as brand or self-image, the answer to the basic question of "what does it cost and how much am I willing to pay for it?" influences many consumer decisions. When consumers make purchases in different market environments (e.g., traditional versus Internet) other factors, such as time, may interact with price in ways that are peculiar to that particular environment. When new market environments evolve, public policy concerns may provide the impetus to examine old relationships within new contexts, such as that found with the use of the Internet and World Wide Web for consumer-oriented retailing.

REFERENCES

Ainslie, George (1975), "Specious Reward: A Behavioral Theory of Impulsiveness and Impulse Control," Psychological Bulletin, Vol. 82, No. 4, July, 463-496.

Benzion, Uri, Amnon Rapoport, and Joseph Yagil (1989), "Discount Rates Inferred from Decisions: An Experimental Study," Management Science, 35 (March) 3, 270-284.

Chapman, Gretchen B. (1996), "Temporal Discounting and Utility for Health and Money," Journal of Experimental Psychology: Learning, Memory, and Cognition, Vol. 22., No. 3, 771-791.

Coupey, Eloise (1999), "Advertising in an Interactive Environment: A Research Agenda," in David Schumann and Esther Thorson (Eds.), Advertising and the World Wide Web. Princeton, NJ: Erlbaum & Associates, 193-211.

Dodds, William B., Kent B. Monroe, and Dhruv Grewal (1991), "Effects of Price, Brand, and Store Information on Buyers’ Product Evaluations," Journal of Marketing Research, Vol. XXVII (August), pp. 307-319.

Fishburn Peter C. (1970), Utility Theory and Decision Making, Wiley: New York.

Fisher, Irving (1930), The Theory of Interest, Macmillan: New York.

Hausman, Jerry C. (1979), "Individual Discount Rates and the Purchase and Utilization of Energy-Using Durables," Bell Journal of Economics, 10, 33-54.

Hoch, Stephen J. and George F. Loewenstein (1991), "Time-Inconsistent Preferences and Consumer Self-Control, Journal of Consumer Research, 17 (March) 4, 492-507.

Loewenstein, George (1987), "Anticipation and the Valuation of Delayed Consumption," The Economic Journal, 97, 666-684.

Loewenstein, George (1992), "The Rise and Fall of Psychological Explanations in the Economics of Intertemporal Choice," in Choice Over Time, eds. George Loewenstein and Jon Elster, New York: Russell Sage Foundation, pp. 3-34.

Loewenstein, George, and Jon Elster (1992), "Utility from Memory and Anticipation," in Choice Over Time, eds. George Loewenstein and Jon Elster, New York: Russell Sage Foundation, pp. 213-234.

Marshall, Greg W., John C. Mowen, and Thomas H. Stone (1995), "Risk Taking in Sales-Force Selection Decisions: The Impact of Decision Time and Frame," Psychology and Marketing, 12 (July), 265-285.

Mazur, James E., and A. W. Logue (1978), "Choice in a "Self-control" Paradigm: Effects of a Fading Procedure," Journal of the Experimental Analysis of Behavior, 30, (July), pp. 11-17.

Mischel, Walter, and Joan Grusec (1967), "Waiting for Rewards and Punishments: Effects of Time and Probability on Choice," Journal of Personality and Social Psychology, 5 (1), 2-31.

Mowen, John C. (1992), "The Time and Outcome Valuation Model: Implications for Understanding Reactance and Risky Choice in Consumer Decision Making," in J. Sherry and B. Sternthal (Eds.), Advances in Consumer Research, Vol. XIX. Provo, UT: Association for Consumer Research, pp. 182-189.

Mowen, John C. and Maryanne M. Mowen (1991), "Time and Outcome Valuation: Implications for Marketing Decision Making," Journal of Marketing, 55 (October) 4, 54-62.

Rabin, Matthew (1998), "Psychology and Economics," Journal of Economic Literature, 36 (March), 11-46.

Samuelson, Paul (1937), "A Note on the Measurement of Utility," Review of Economic Studies, 4, pp. 155-161.

Strotz, Robert H., (1955-56), "Myopia and Inconsistency in Dynamic Utility Maximization," Review of Economic Studies, 23, pp. 165-180.

Thaler, Richard H (1981), "Some Empirical Evidence on Dynamic Inconsistency," Economic Letters, 8, 201-207.

Thaler, Richard H and Hersh M. Shefrin (1981), "An Economic Theory of Self-Control," Journal of Political Economy, 89 (April) 2, 392-406.

Wathieu, Luc (1997), "Habits and the Anomalies in Intertemporal Choice," Management Science, Vol. 43, No. 11, November, 1552-1563.

Wertenbroch, Klaus (1998), "Consumption Self-control by Rationing Purchase Quantities of Virtue and Vice," Marketing Science, Vol. 17, No. 4, 317-337.

----------------------------------------