A Framework for Promoting Retirement Savings

By Wiener, Josh Doescher, Tabitha

This article identifies the constructs that influence an individual’s intention to save for retirement and discusses how and when these factors can be changed by an agent trying to induce an individual to enroll in a retirement plan, increase his or her contribution to a plan, or purchase a particular retirement product. A broad array of psychological theories is used to develop a series of persuasive communications that can encourage a person to save. In addition, the persuasive communication approach is placed in the broader context of all efforts used to promote retirement savings. Many Americans are not saving enough for their retirement. According to the 2007 Retirement Confidence Survey (Helman, VanDerhei, and Copeland 2007), only 66 percent of workers report that they and/or their spouse have saved money for retirement and only 60 percent report that they are currently saving. Even among those who do save, savings can be insufficient. About half of all workers saving for retirement report that the total value of their investments, excluding their home and their defined benefit plan, was less than $25,000.

Because this problem has been well known for many years, both private agents and public policy makers have made numerous efforts to increase the extent to which American workers save for retirement. Basically, two broad approaches, the first structural and the second involving communication, have been used to do this.

Structural approaches attempt to change the conditions under which people save. For example, policy makers can alter the financial conditions associated with retirement saving (e.g., enhance tax benefits or raise retirement plan contribution limits), while employers can rework plan design (e.g., institute matching, enrollment by default, or automatic contribution increases). In contrast, communication approaches focus on changing both workers’ knowledge and their perceptions. The former occurs through education (e.g., teaching the fundamentals of investing); the latter occurs through persuasion (e.g., creating normative pressures or enhancing the perceived importance of one’s retirement years).

Past policy efforts have focused on structural changes and on education. For example, the most recent effort, the Pension Protection Act of 2006, contains several favorable retirement savings measures, including allowing employers to offer automatic enrollment plans as the default and permitting plan providers to offer certain investment advice without running afoul of advisor conflicts of interest. However, 42 percent of all working-age wage and salary employees do not work for an employer or union that sponsors a retirement plan (Copeland 2006) and therefore will not be affected by either of these changes. Moreover, these changes may not be sufficient even among workers who have access to employer- sponsored plans. A recent study by VanDerhei (2007) suggests that automatic enrollment will not be a magic bullet: although replacement rates (the percent of final salary that is replaced in retirement) will increase, they will still be below the traditional minimum recommended replacement level of 70 percent.1 In addition, some experts doubt that automatic enrollment plans (even for new employees) will rapidly diffuse (Laise 2007).2 The effects of the educational change (investment counseling) are also problematic: despite some successes, the impact of past employee educational programs on participation and contribution rates has been disappointing (Benartzi and Thaler 2007). Thus, the evidence suggests that although these structural and educational efforts have been and are valuable, there is a need to do more.

In this article, we focus on the role that can be played by persuasive communications. To depict the role of persuasive communications, we use a policy framework. The framework is constructed from the point of view of an agent who is trying to get an individual to join a retirement savings program (e.g., a 401(k) plan), increase his or her level of contribution to such a program, or make a discrete savings decision (e.g., buy an individual retirement account [IRA]). The likelihood a person will act is strongly influenced by two factors, intentions and inertia/ procrastination. Because there is a very rich literature on the role inertia/procrastination plays in the decision to save (see Appendix 1), the focus in this article, and in our framework, is on intentions.

More specifically, the article identifies the factors that influence an individual’s intent to save and provides guidance regarding how and when these factors can be changed by persuasive communications. In addition, the article places the persuasive communications approach in the broader context of all efforts being used to promote retirement savings.

INTENTIONS OVERVIEW

A person’s intention to save for retirement is a conscious decision to engage in a savings action. In a host of psychological models, for example, social cognitive learning (Bandura 1986), reasoned action/planned behavior/trying/goal directed (Bagozzi 1992), and protection motivation/health belief (Tanner, Hunt, and Eppright 1991), the likelihood a person engages in a particular behavior is an increasing function of the strength of his intention to act. In our model, a person may have a strong transient intention that reflects little cognition. This transient desire is viewed as an intention because it can be influenced by many of the same constructs that can influence a more enduring intention.

An agent seeking to alter a person’s intention to save can take an action that attempts to change a person’s perceptions of the following constructs: his perceived ability to save, the benefits associated with saving, his concerns about the future, and the costs associated with saving (Figure 1). The first construct, a person’s perception of his ability, has both an actual and a subjective dimension. In other words, a person may over- or underestimate his true or actual ability. The second construct is more straightforward: the primary types of benefits are short-term benefits (e.g., matching dollars or normative pressures), having a higher retirement income due to earning a higher real rate of return and having greater hope. The third construct is concerns. Both hopes, which indirectly affect intentions through their impact on benefits, and concerns, which directly affect intentions, are determined by how the agent frames the long-term consequences from saving. These consequences can be framed as either hopes (e.g., how having more money during retirement will make one’s life better) or concerns (e.g., how additional retirement income is needed to stave off dire consequences). When a positive (hope) frame is employed, the agent’s action simply enhances the benefits associated with saving. When a negative (concern) frame is employed, the agent’s action has the potential to set in motion a much more complex set of psychological responses. Finally, the last construct, the cost of saving, depends on both the opportunity cost of the dollars put into a savings instrument and the deprivation associated with the value of these dollars.

Changing these constructs will influence a person’s intention to save. First, increasing one’s ability to save will increase the likelihood of saving as long as the benefits outweigh the costs. Second, increasing the benefits associated with saving will have a positive effect on the likelihood a person will save as long as a person thinks that he has the ability to save. Third, increasing concerns can either increase or decrease the likelihood a person will save: if a person thinks that he is capable of taking actions that will save him from an awful fate, then enhancing concerns will motivate a person to save; however, if he does not feel capable, enhancing concerns can reduce the likelihood of both current and future savings behaviors. In other words, when perceived ability is too low, a persuasive message that emphasizes concerns (e.g., a classic fear appeal) can be counterproductive. Finally, increasing the costs associated with savings will reduce the likelihood a person will act.

The remainder of this article examines these constructs in detail. Each section begins with a discussion of the basic theories behind the construct and their application to retirement savings and includes empirical evidence supporting the theories’ relevance to retirement savings. The section then turns to the type or types of approaches-structural changes, education, and persuasive communications-that can be used to encourage individuals to save for retirement.

PERCEIVED ABILITY

One version or another of perceived ability plays a critical role in many psychological theories. In each theory, the construct has a somewhat unique definition and, in many cases, is given a theory- specific name. We use the language of Bagozzi (1992) in which perceived ability is synonymous with his term goal efficacy.

Bagozzi created the term goal efficacy to capture two distinct versions of efficacy: the self-efficacy dimension denoted by perceived behavioral control in the theory of planned behavior and the response efficacy dimension denoted by success and failure expectancies in the theory of trying. Goal efficacy is therefore a person’s perception of the likelihood that she will reach a goal if she tries (Bagozzi 1992; Bagozzi and Edwards 1998). It includes a person’s appraisal of both her ability to perform the instrumental act (self-efficacy) and her estimate of the likelihood that if she performs the act she will achieve the goal (response efficacy). There is considerable empirical support for the hypothesis that the strength of a person’s intention to act is a positive function of his degree of self- and/or response efficacy. This holds whether the action is framed as gaining a personal benefit (Bagozzi 1992; Bandura 1986), reducing a threat (Tanner, Hunt, and Eppright 1991), or obeying a norm (Schwartz 1977). In all cases, there is a boundary condition: a person must think that the positive consequences of acting outweigh the costs.

In all theories that include a self-efficacy construct, for example, trying (Bagozzi 1992), normative behavior (Schwartz 1977), and protection motivation (Tanner, Hunt, and Eppright 1991), very low levels of self-efficacy do more than simply reduce the likelihood of acting. When the action is viewed as producing a benefit or obeying a norm, self-efficacy plays a gatekeeping role. If a person’s level of self-efficacy is so low that she does not think she is capable of engaging in an act, then she will not even try. In other words, when self-efficacy is low, increasing the benefits from acting will not result in a higher likelihood of acting. When the action is viewed as reducing a threat, both self- and response efficacy determine whether increasing the threat level leads to a higher or lower likelihood of acting. If either self- efficacy or response efficacy is too low, then increasing the threat level can reduce the likelihood a person will act now or in the future.

Although saving has not been one of the behaviors empirically studied, the argument that efficacy will influence savings behavior is consistent with two empirical findings. First, Farkas and Johnson (1994) and Public Agenda (1995) find a positive association between 401(k) participation and thinking that it is easy to save for one’s retirement. Since assessment of task difficulty is a proxy for self- efficacy (Bandura 1986), this finding suggests that individuals with higher levels of self-efficacy (in terms of saving for retirement) are more likely to participate in 401 (k) plans. Second, Public Agenda (1995) finds a positive relationship between 401(k) plan participation and confidence that one will reach one’s retirement goals. Being confident of reaching one’s retirement goals can be thought of as a proxy for response efficacy.

There are two primary ways of influencing a person’s level of goal efficacy (i.e., perceived ability). The first is to increase the individual’s actual ability to perform and accomplish critical tasks; the second is to increase the individual’s subjective judgment of her ability to perform and accomplish these tasks given her current skill/knowledge base.

Increasing Actual Ability

Behavioral economists argue that most people lack the computational ability and knowledge needed to solve the dynamic optimization problem that the standard neoclassical theory of savings assumes people act as if they solve (Benartzi and Thaler 2007). In particular, behavioral economists recognize it is difficult to (1) calculate how much money one needs to accumulate in order to be able to have a comfortable retirement, (2) calculate how much one must save on a monthly basis in order to reach this financial goal, and (3) decide how to invest this money. Their claims are consistent with numerous studies reporting that most Americans know very little about calculating retirement needs (Yakoboski and Dickemper 1997), investing (Public Agenda 1995), or even when they will retire (Employee Benefit Research Institute 2000). The standard approach is to deal with knowledge problems through education. Several studies have found that providing educational materials and seminars has a positive impact on participation in employer-sponsored retirement plans (Conte 1998; Yakoboski, Ostuw, and Hicks 1998). For example, 73 percent of those who have received educational materials from their employer have begun to save for their retirement as compared to 57 percent of those who have not received such material (Yakoboski, Ostuw, and Hicks 1998). However, as noted by Benartzi and Thaler (2007), many educational efforts have been disappointing.

A second point made by the behavioral economists is that people lack the willpower required to forgo consumption and save on a regular basis (Benartzi and Thaler 2007). Economists have long recognized that a key reason people fail to save is that they lack self-control. For example, in 1930, Fisher argued that people failed to save because they were “impatient”; some of the characteristics that led to impatience were “weak will,””habits of spending freely,” and “slavish following of whims of fashion” (Wameryd 1990). A long-standing structural solution to the self-control problem is to provide plans that allow a person to make a single one-time commitment to enroll in an automatic salary deduction plan, for example, a 401(k) or 403(b) plan. This type of precommitment strategy imposes a constraint on future behavior, thereby enhancing a person’s willpower because the sacrifice is made before the source of short-term gratification is truly encountered. The strongest evidence supporting this view is from empirical studies which find that although providing workers with 401(k) plans increases savings, changing the level of the employer match does not (e.g., Papke and Poterba 1995). In other words, these studies suggest that some workers may be responding to the availability of a 401(k) plan and not simply to the financial benefit.

Recently, there have been a number of innovative structural changes that jointly address the cognitive ability and willpower barriers (see Benartzi and Thaler 2007, for a detailed discussion). The changes include quick enrollment (a person can enroll and accept a preset contribution level and investment strategy by simply checking a box), enrollment by default (if a worker does nothing, she is automatically enrolled), and automatic escalation (the percentage of a person’s income allocated to a retirement fund automatically increases over time). Early evidence finds that these structural changes can be very effective. For example, one study found that implementing quick enrollment for new employees increased plan participation from 5 to 19 percent after one month and from 10 to 35 percent after four months. When all nonparticipating employees were given the quick enrollment option, about 25 percent of these employees joined a plan, and consequently, the overall participation rate rose from 54 to 66 percent (Beshears et al. 2006). Not surprisingly, the power of automatic enrollment has been even more dramatic. For example, Madrian and Shea (2001 ) found that with automatic enrollment, the level of participation by new hires at a large financial services company rose from 49 to 86 percent. While these structural changes appear promising, it should be noted that they are not a panacea: as mentioned in the introduction, not all workers are covered by a pension plan and not all workers can be expected to adopt plans with these features.

Increasing Subjective Assessment of Ability

A second way of enhancing goal efficacy is to use persuasive communications to increase the individual’s subjective judgment of her ability to perform and accomplish certain tasks given her current skill/knowledge base. A person’s subjective assessment of her ability is strongly influenced by four factors: effective attainment, vicarious experience, persuasive messages, and heuristics and biases. Through using marketing strategies and tactics to influence these factors, it is possible to alter an individual’s assessment of her ability, thereby influencing either self-efficacy or goal efficacy. (Response efficacy will not be independently affected.) The type of efficacy affected will depend on whether the person is led to think in terms of the instrumental act (e.g., can I save) or the goal (e.g., if I try, will I be able maintain my current lifestyle during my retirement years). The remainder of this section discusses the four determinants of subjective judgment and the appropriate marketing strategies for manipulating them.

Because a person’s level of efficacy will be higher if she can recall situations where she successfully performed a similar task, one way of enhancing an individual’s assessment of her ability to perform a task is through effective attainment, that is, past successful performance of a similar task. When applied to the issue of saving for retirement, this suggests a footin-the-door communications strategy. This strategy entails encouraging nonsavers to begin saving small amounts of money on a regular basis and then, over time, encouraging them to increase their level of saving. Because saving at a low amount is very similar to saving at a somewhat higher amount, a foot-in-the-door approach will enhance a person’s perception of her ability to save enough to meet her retirement needs. Note that this strategy is contrary to the current practice of relying on financial planning tools that identify the level at which a person must save to reach her retirement goal. The level of saving identified by a retirement planning tool can be so high so that it may well appear to be an impossible task to a nonsaver.

A second way of increasing a person’s judgment of his ability to accomplish a task (and hence her efficacy) is through vicarious experience, that is, observing similar individuals succeeding at a task. In the context of saving for retirement, this suggests the value of communicating how others with similar jobs, employers, and life circumstances are saving. This positive bandwagon approach is contrary to the current emphasis on how, because most workers are not saving enough to meet their retirement needs, there is a national savings crisis. The emphasis on a current crisis serves to undercut a person’s perception of her efficacy because it is communicating that similar people are not succeeding at the retirement task. A third approach to boosting an individual’s perception of her ability to successfully perform is through persuasive messages. A person’s level of efficacy can be influenced by messages, which communicate that she is capable of meeting her objectives. For example, when applied to retirement savings, this suggests using messages that convey confidence in the individual’s ability to save enough money to reach either her retirement goal (goal efficacy) or some smaller sum, such as the amount matched by her employer (self-efficacy). The positive, optimistic orientation suggested by this tactic is contrary to the crisis orientation exhibited by much of the existing promotional material. However, it is consistent with the current use of persuasive messages that emphasize the magic of compounding.

A fourth approach is to exploit the heuristics and biases people use when making judgments under uncertainty (see Kahneman, Slovic, and Tversky 1982, for a comprehensive review). For example, the availability heuristic means that a person will think an event is more likely if it is easy for her to recall salient examples. The belief in small numbers means that a prediction can be based on very little data. The representativeness heuristic means that irrelevant data are used when predictions are made. Put together, they suggest that if communications can lead a person to easily recall a small number of very salient examples of people “like herself” who successfully saved for their retirement, she will think that she is more able to save.

BENEFITS

The act of saving for retirement may produce benefits that can be enjoyed either immediately and/or during the retirement years. The short-term benefits consist of the short-term monetary rewards from savings (e.g., tax deferral) as well as the short-term nonmonetary rewards (e.g., gaining the approval of others). The retirement period benefits are the future additional utility gained from saving current income, that is, when a person saves a dollar, he augments his future retirement income and gains some level of future additional utility from this augmentation. In our framework, the impact of increased savings on one’s future income is influenced by the rate of return, and the impact of changing one’s future income on the benefits produced by saving is influenced by one’s hopes. In this section, we focus on how a policy maker can try to augment each of these potential benefits. In essence, four types of action can be taken: provide a short-term monetary reward from saving, provide a short-term nonmonetary reward from saving, increase the anticipated rate of return, and increase hopes by increasing the importance of a positively framed future.

Short-Term Financial Benefits

Behavioral economists argue that a person will be motivated to save by the opportunity to gain short-term financial benefits, such as the tax savings and employer matching dollars associated with some retirement plans (Thaler 1994). Prior research clearly supports the idea that a person is more likely to select a savings instrument that provides a short-term reward over one that does not (Thaler 1994). Feenberg and Skinner (1989) find that, holding all else constant, people who have to write a check to the Internal Revenue Service on April 15 are more likely to invest in an IRA than those who do not have to write a check. In addition, a number of companies have reported that the introduction and promotion of a matching program led to a significant increase in participation by younger, less-educated employees in the company’s retirement program (Conte 1998). However, there is no conclusive evidence that these financial incentives will increase a person’s overall level of retirement savings (see the debate between Engen, Gale, and Scholz 1996 and Poterba, Venti, and Wise 1996).

Short-Term Nonmonetary Benefits

The primary potential source of nonmonetary benefits from saving (at least for people who currently are not savers) is norms. There are four types of norms: societal, personal, injunctive social, and descriptive. Each is discussed in more detail below.

Societal Norms

A societal norm is a belief shared by most members of a social system about what constitutes good or bad behavior (Schwartz 1977). A societal norm (such as the belief that one should save for retirement) has an indirect, rather than direct, impact on behavior. It can only influence behavior to the extent that it is internalized, that is, becomes a personal norm.

Personal Norms

A personal norm is a self-based standard of expectation for behavior that flows from one’s internalized values; it is enforced through the anticipation of self-enhancement or self-depreciation (Cialdini, Kallgren, and Reno 1991; Schwartz 1977). The specific personal norm that a person applies in response to a given problem is, in most instances, constructed by drawing on the societal norms he has internalized in the past and on the values that stem from his unique history, for example, the individual’s upbringing (Schwartz 1977).

Personal norms must be activated to influence behavior (Cialdini, Kallgren, and Reno 1991). In other words, the person must consciously think about how the norm is relevant to a specific situation before the norm will have an impact. According to Schwartz’s (1977) theory of altruism, the likelihood of this occurring is a positive function of both the individual’s ascription of responsibility and his awareness of the consequences of acting.

Schwartz’s model can be applied to the retirement savings decision. An individual’s upbringing, coupled with the societal belief that people should save for their retirement, creates the following personal norm: “I should save for my retirement.” The individual recognizes that if he does not save for his retirement, he will not have a satisfactory retirement standard of living. He takes on the following ascription of responsibility: “I am responsible for my own retirement standard of living.” He is also aware of the consequences of retirement plan participation: “If I participate, I can have a satisfactory retirement standard of living.” Together his personal norm, his ascription of responsibility, and his awareness of the consequences of acting motivate him to participate in a retirement plan. All three elements must be present. A personal norm will not be activated unless a person has both a feeling of responsibility and an awareness of the consequences.

There is some empirical evidence supporting the potential power of personal norms to motivate saving. A nationally representative survey of working-age individuals found a strong association between saying that the statement “I learned the importance of saving money as a child” described them “very well” and actual retirement savings behavior, that is, having an IRA, being enrolled in a 401(k) plan, or having at least $50,000 in retirement savings (Farkas, Bers, and Duffett 1997). Since norms are learned during childhood, this suggests a positive link between possessing a personal norm that enjoins saving and saving for one’s retirement.

Second, there is a strong relationship between an individual’s level of retirement savings and his or her belief that individuals, not the government or employers, should take the greatest share of responsibility for ensuring that they have an adequate income in their retirement. About 60 percent of those surveyed said that an individual had the greatest level of responsibility; however, 79 percent of those with savings more than $50,000 chose individual responsibility as compared with about 53 percent of those with savings less than $50,000 (Public Agenda 1995).

Because influencing personal norms is a very difficult task, agents seeking to promote retirement savings in the near term should focus their efforts on activating latent norms, that is, on increasing a person’s ascription of responsibility for his own retirement and on enhancing his awareness of the consequences of saving. Statements from public officials and educational materials emphasizing that retirees cannot count on social security will enhance a person’s ascription of responsibility, while the tactics that can enhance a person’s perception of his goal efficacy (e.g., effective attainment, vicarious experience, persuasive messages, and heuristics) will increase his awareness of the consequences.

A reasonable long-term goal would be to significantly increase the proportion of the population holding a strong personal norm that enjoins saving. Current efforts to create and deliver savings programs to elementary and middle school students are a step in this direction.

Injunctive Social Norms

An injunctive social norm refers to a person’s perception of what relevant others expect him to do. These people are most often friends, family, and coworkers. Their influence flows through a person’s desire to avoid social sanctions and gain social rewards (Cialdini, Kallgren, and Reno 1991). Numerous studies have found support for the hypothesis that injunctive social norms can have some impact on a person’s behavioral intentions and behavior (Sheppard, Hartwick, and Warshaw 1988).

Prior research suggests that injunctive social norms have the potential to motivate savings behavior. A survey by Fidelity Institutional Retirement Services found that 31 percent of the employees who chose not to participate in a 401(k) plan cited the outside influence of friends and family as a reason (Williamson 1996). In addition, Yakoboski, Ostuw, and Hicks (1998) report that when workers who save for retirement were asked what had motivated them, 55 percent said that advice from friends or family had provided at least some motivation. For this reason, agents who want to encourage participation in retirement plans may find it fruitful to communicate with the friends and families of their employees. This approach has been used in other contexts by using mail to reach spouses, using school programs to reach children, and encouraging people (including coworkers) who are engaging in desired behavior to talk to those who are not. In a broad sense, a person’s retirement plan participation decision is not purely a private decision, but rather one that is subject to social influences. The final norm expands on this point.

Descriptive Norms

A descriptive norm is the belief that one should imitate the behavior of others. The power of descriptive norms is one of the most widely held tenets in psychology (Cialdini, Kallgren, and Reno 1991; Larimer et al. 2004). Numerous studies have found that people will imitate the behavior of others under a wide range of conditions. The power of descriptive norms was first recognized in situations in which individuals were in personal contact with one another (Cialdini, Kallgren, and Reno 1991). However, the power of descriptive norms has also been confirmed in situations where a person is simply made aware of how others have acted (e.g., Cialdini, Kallgren, and Reno 1991; Wiener and Doescher 1994).

Cialdini, Kallgren, and Reno (1991, p. 263) argue that a key reason for the power of descriptive norms is that people use a simple information processing rule of “if everyone is doing or thinking or believing it, it must be a sensible thing to do, or think, and believe.” The term “everyone” refers to all individuals who are in similar circumstances regarding the behavior in question. Similarity is a matter of degree; it can reflect physical proximity, common resources, common demographics, and so forth. A key point is that because the information processing rule is so simple, a person may imitate the behavior of others without consciously realizing that he is doing so.

A study by Duflo and Saez (2002) provides some evidence to support our contention that descriptive norms can be used to motivate retirement behavior. They studied plan participation at 11 libraries of a major university and found that participation rates ranged from 14 to 73 percent. They conclude that since the effect of the specific library a person works at is significant, even when factors such as age and tenure structure are controlled for, their study supports the hypothesis that peer influences are an important determinant of savings decisions.

One important implication of these propositions is that many of the current print promotional materials distributed by the leading preretirement savings organizations may be counterproductive. In their effort to communicate that there is a national savings problem, the pamphlets emphasize the extent to which people are neither saving nor participating in 401(k) plans. The descriptive norm literature suggests that many workers will respond to this communication by reasoning “If other workers aren’t participating, I won’t either.”

Fortunately, there are many social marketing strategies that exploit the power of descriptive norms. Their common trait is that they take the tactic of continually communicating how many people (like the target market) are planning to engage in, or are already engaging in, the behavior being promoted. For example, a small company could follow this strategy by emphasizing how similar companies have very high 401(k) participation rates. Large companies can begin their promotional efforts by expending a high level of promotional resources on a small segment of workers and then running a less resource-intensive communication effort trumpeting how a particular plant or work unit has achieved a very high participation rate. At a broader level, promotional material could emphasize how the number of workers participating in 401(k) plans has been increasing rapidly over the past decade.

In conclusion, because descriptive norms are so powerful, they are an unavoidable element of any strategy to increase retirement plan participation. If the strategy focuses on the empty glass (the coverage problems faced by American workers), the strategy will inhibit participation. If the strategy focuses on the full glass (examples of how workers are participating), the strategy will motivate. Finally, if the strategy is silent, then workers will imitate their friends and coworkers: if participation among their friends and coworkers is high, they will participate; if participation is low, they will not.

Rate of Return

Both Browning and Lusardi (1996) and Thaler (1994) conclude that increasing the real after-tax rate of return can have an ambiguous effect on savings. The theoretical reason in the neoclassical lifecycle savings model for the ambiguity is that while a higher rate of return increases the returns from saving, it reduces the amount of saving required to achieve any given level of retirement income (Thaler 1994). Thaler’s (1994, p. 186) carefully worded summary of the empirical evidence is that most studies are “unable to reject the hypothesis that the elasticity of personal saving with respect to the interest rate is zero.”

Although neither theory nor empirical work provides a sound basis for predicting how a change in the rate of return will influence a person’s intention to save (Browning and Lusardi 1996; Thaler 1994), there is widespread agreement that many savers invest their money poorly (Benartzi and Thaler 2007). In other words, the rate of return these savers earn is lower than the rate of return they could earn if they invested their money differently. To address this problem, employers can provide investment education, allow plan providers to offer face-to-face personally tailored investment advice, limit investment options to a small number of truly appropriate plans, and offer a truly appropriate default plan.

Hopes

A person’s hopes for retirement will be influenced by the extent to which he is aware of the positive aspects of leading a good retirement and the length of time he must wait before he enjoys his golden years. A person’s awareness reflects both how tangible retirement is to him and how connected the idea of retirement is with his basic psychological needs; these ideas, and their associated marketing strategies, are discussed in the next two sections. In the third section, we discuss the role of time and, in particular, how marketing strategies can lessen the degree to which one’s hopes for retirement are discounted.

Tangibility

The tangibility, to an individual, of his retirement refers to that individual’s perceptions of the tangible, concrete outcomes he may experience during his retirement years. (Outcomes that are not tangible or concrete are conceptualized as fulfilling basic psychological needs or being tied to abstract goals and are discussed later.)

The concept of tangibility is based on both basic attitude theory and information processing theories of attribute importance and memory. Specifically, the tangibility of one’s retirement is an increasing function of the number of concrete outcomes a person associates with retirement, his subjective evaluation of these outcomes, the degree to which he has allocated attention to these outcomes, and his assessment of the likelihood of experiencing these outcomes. These outcomes may be based on anything from sober analysis to dreams or fantasies. Greater tangibility leads to greater hopes and so increases the value of any increase in one’s retirement income.

The characteristics of both retirement and retirement plan products (e.g., 401(k) plans and IRAs) are such that we can presuppose that many people will have problems identifying the tangible outcomes. This is because, for many people, both retirement lifestyle and retirement products are innovative products: they cannot be tried, are very complex, do not fit one’s current lifestyle, and are not associated with a well-developed product- class knowledge structure (see Gatignon and Robertson 1985 for a discussion of innovative products).

The supposition that many people cannot identify the tangible outcomes associated with retirement is consistent with empirical findings. Based on a series of focus groups and surveys, Farkas and Johnson (1994, p. 13) conclude that a key reason many people do not prepare for their retirement is that most people just do not think about it very often, that is, “it may be on their radar screen, but just barely.” For example, in one survey, 35 percent said they never or rarely thought about retirement (and this question was ninth in a series of questions about retirement). Not surprisingly, there is a connection between thinking about retirement and savings behavior; for example, 41 percent of all individuals who said they never thought about retirement have no retirement investments or savings.

When taken together, the concept of tangibility, the product innovation literature, and the empirical studies of retirement combine to identify a key marketing task: agents seeking to encourage retirement savings must effectively communicate how saving for retirement will lead to positive tangible outcomes. The most straightforward (and generic) approach to enhancing the value of an unknown outcome is to communicate concrete consequences.

Although for marketers this task is obvious, its implications for current policy are significant. A major current public education effort is the publication and distribution of millions of pamphlets such as “Choose to Save” and “Ballpark Estimate” by the Department of Labor and American Savings Education Council (see American Savings Education Council 2007). These efforts presume that people are motivated to have enough funds to support a good retirement and focus on informing people (1) that to have a good retirement, they will need 70 to 90 percent of their preretirement income level, (2) that most people are not on track to reach this goal, and (3) how much they will need to save to reach this goal. These pamphlets reflect the underlying assumption of the prosavings community that the key to motivating people is to teach them what they need to do to save enough for their retirement. In other words, there is no recognition that marketing stimuli must be used to transform a vague notion that retirement is important into a set of tangible outcomes. Basic Psychological Needs

Both the means-end (Reynolds and Gutman 1988) and the hierarchy of goals (Bagozzi and Edwards 1998) literatures reach a common conclusion that a person is more likely to engage in an action if the outcomes of the action are linked to basic psychological needs, such as self-esteem, security, and social approval. In the context of retirement savings, this implies that the hopes an individual has for his retirement will be influenced by the extent to which he links reaching his retirement goal with fulfilling these needs.

Both theories postulate, and empirical work confirms, that typically the links between basic psychological needs and product or services are latent. In other words, people are unaware of the links unless a researcher probes or a persuasive communication activates. Persuasive communications activate latent links by making the link very salient. For this type of persuasive communication to work, the receiver must be involved, that is, have a high degree of interest in either the product being promoted or the promotional message itself.

Due to the low utilitarian and hedonic benefits associated with the act of saving for retirement (by nonsavers), it can be expected that, despite the actual importance of retirement, some people will process as if it were a low-involvement product. However, research by Maclnnis, Moorman, and Jaworski (1991) suggests that properly executed communications should be able to stimulate the processing of a retirement message and thus overcome the problem of low involvement.

Persuasive communications should therefore attempt to activate the latent links that exist between basic psychological needs and retirement by using the tactics suggested by Maclnnis, Moorman, and Jaworski, for example, the use of unusual formats, attention- attracting pictures, strong associations between the product and the affect-laden visual images, emotional appeals, and spokespeople who are either celebrities or someone with whom the message receiver can identify. For example, showing grandparents having a good time with their grandchildren is an example of using an affect-laden visual image to activate the latent psychological link between the retirement and the importance of family.

Time and Discounting

Finally, a person’s hopes for retirement will be influenced by time. The longer one must wait for a future benefit, the less it will be worth in the present, that is, future outcomes are discounted. This idea is a key element of the neoclassical theory of saving and is consistent with numerous studies that have found a strong positive relationship between an employee’s age and the degree to which he or she saves for retirement (Andrews 1992; Yakoboski and VanDerhei 1996).

Two approaches can be used to understand why people discount the future. The first is that there are personality differences, that is, different people have different internal rates of discount. Although empirically supported (Lin and Mowen 1994), this idea does not lead to direct policy recommendations since policies cannot alter personality traits. A second approach is more diffuse, but more promising in terms of identifying actions a policy maker can take.

Specifically, marketing scholars have investigated related issues, such as the effect of self-referencing anticipatory visualization on the formation of mental representations (Krishnamurthy and Sujan 1999), the link between mental representations of future outcomes and both brand attitudes and purchase intentions (Krishnamurthy and Sujan 1999), the traits associated with ability to delay gratification (Lin and Mowen 1994), and the relationship between visualization and purchase intentions (Hoch and Loewenstein 1991; Krishnamurthy and Sujan 1999; Maclnnis and Price 1987). These lines of research suggest a reason why future events are discounted.

Because a person’s vision of the future is constructed from available data, mental representations will become less and less detailed as one moves farther into the future and toward life situations more removed from one’s present state (Krishnamurthy and Sujan 1999). Moreover, numerous studies find that there is a positive association between the richness of one’s mental representation and the attitudes and intentions. Based on these ideas, we conclude that future events are discounted due to the inability of a person to form a rich mental representation.

A person’s mental representation can include both verbal propositions, such as those that directly contribute to tangibility, and sensory associations, such as visual images (Maclnnis and Price 1987). Two streams of research suggest that a dearth of sensory associations will lead to extensive discounting. One line consistently finds an association between a person’s ability to envision the future and his ability to delay gratification (Lin and Mowen 1994). The second line of research reports that the attractiveness of a future experience, measured by attitudes and intentions, is associated with the extent to which a person can form a concrete, self-referencing, visual image of the experience (Hoch and Loewenstein 1991; Krishnamurthy and Sujan 1999; MacInnis and Price 1987).

Both early economic reasoning and current focus group research support the argument that an inability to envision the future leads to excessive discounting and discourages people from saving. Over 100 years ago, von Bohm-Bawerk argued that people do not save because “Future goods are less clearly perceived than present goods … [in part due to a] … lack of imagination and ability to understand” (Warneryd 1990, p. 50). In addition, one of the more robust findings of a series of focus groups was that savers were far more likely to imagine retirement life than nonsavers. For example, a young Atlanta woman said, “It’s a real fun thing for us to imagine our retirement-we don’t sit around and bite our nails. We know exactly what we’re gonna do, how much we’re gonna have, what a typical day’s going to be” (Farkas and Johnson 1994, p. 17).

The above findings suggest that the degree to which retirement is discounted can be influenced by marketing communications. This stands in sharp contrast to the economists’ assumption that the degree to which a person discounts retirement income is solely a function of time and his internal rate of discount. Specifically, the extent to which people discount retirement can be limited by communications that are self-referencing and that are capable of inducing the receiver to imagine rich (i.e., visualized, concrete) retirement-related experiences. Since almost all potential savers will neither naturally allocate very high levels of attention to processing the retirement communications nor have strong pre- existing associations with retirement, we would recommend the use of communications that are both visual and filled with contextual detail (see Keller and Block 1997).

CONCERNS

A person’s concerns capture all the negative thoughts she might have about the retirement years. Concerns come into play when a policy maker tries to motivate a person to save by emphasizing the dire consequences of not having sufficient income during the retirement years. A person’s concerns can be enhanced by messages that make the future bad state more tangible, more connected to psychological needs, and more sensory rich. In other words, the tactics for increasing the salience of a future state work whether the future state is good (hope) or bad (concern). The critical issue for a policy maker is to identify when a strategy emphasizing concerns should be employed.

Tanner, Hunt, and Eppright’s (1991) protection motivation model depicts how people respond to threats. A threat can evoke two responses. First, a person can engage in a coping response that will reduce the feeling of threat. Second, she can feel fear.

There are two types of coping responses: adaptive (i.e., an action that reduces the actual threat level) and maladaptive (i.e., a response, which can be cognitive, that lessens the feeling of fear without reducing the actual threat level). When the concern is associated with having too little income during one’s retirement, the adaptive response is to save money. The maladaptive response is to come up with an excuse, for example, “why worry about retirement- everyone in my family dies young.”

Whether or not an adaptive or maladaptive response is followed depends on the individual’s level of coping appraisal, that is, the extent to which a person (1) is aware of an action she can take that will remove the threat and (2) feels capable of engaging in that action. If the answer to both is yes, there will be an adaptive response. If the answer to one or both is no, there will be a maladaptive response. As a result, the effect of threats on willingness to save will be positive or negative depending on coping appraisal.

Consequently, messages raising concerns should be used only under certain conditions: (1) when the target market has a pre-existing high level of coping appraisal or when the message also contains a sufficiently strong efficacy message to concurrently create a high enough level of coping appraisal, (2) when the message will not significantly spill over into low coping appraisal markets, and (3) when the individuals who are being exposed to the message do not have a history of rejecting similar messages. In other words, fear appeals should not be used to encourage retirement plan participation for middle-aged and older people who have very few retirement assets. COSTS

People can incur two types of costs when saving for retirement. First is the conventional opportunity cost, which is the loss of utility equal to the level of utility that would have been gained if the money had been used to purchase the next best alternative. This cost depends on factors such as one’s tastes, income, number of dependents, and stage of the lifecycle. The second cost is deprivation, that is, the cost generated when saving leads to a negative discrepancy between one’s current position and a reference value. Because the factors that influence opportunity costs are not themselves policy levers, this section deals solely with deprivation.

Deprivation

Numerous social science theories postulate that individuals feel deprived if their current level of material well-being declines (see Hoch and Loewenstein 1991, for detailed references). In the context of retirement savings, there are two major sources of deprivation: a reduction in one’s standard of living and the framing of saving so that it appears to be a loss.

Deprivation Due to a Reduction of Current Standard of Living

The unwillingness of individuals to reduce their current standard of living has long been recognized by economists concerned with savings behavior (Wameryd 1990). Duesenberry (1967, p. 85), for example, argues that the reason savings rates are so much lower for families that have experienced a recent decline in income relative to those who have not is that there is a fundamental psychological postulate that “it is harder for a family to reduce its expenditures from a high level than for a family to refrain from making these expenditures in the first place.”

Because individuals will resist reducing their current standard of living, the timing of retirement plan participation requests may play a critical role in determining the extent to which people will participate in a plan. For example, if the opportunity to participate in a 401(k) plan is presented when a salary increase is being received, then it is less likely the person will have to actually reduce his current standard of living to participate. Likewise, the SMART plan feature, which automatically escalates a person’s contribution whenever he receives a raise, is perfectly designed to minimize the individual’s sense of deprivation.

Deprivation Due to Framing

A person’s sense of deprivation and hence willingness to save can be influenced by the way in which the saving opportunity is phrased. One of the most pervasive findings in the psychological decision- making literature is that if an outcome is labeled a loss, a person will think of it as a loss, associate it with deprivation, and seek to avoid it (see Kahneman, Slovic, and Tversky 1982).

The issue of phrasing is relevant because currently some 401(k) participants are asked to sign up for a “salary reduction plan.” Because this phrasing forces them to think of the contribution as a loss, they will be less likely to participate. It would be far better to avoid the term “salary reduction” and refer to the plans using either neutral or positive phrasing, for example, 401(k) plan or tax deferral plan.

A second application of framing draws on the mental accounting literature (Thaler 1994). If a person is asked to take money out of his income stream or another savings account, then he is clearly making a sacrifice. However, if the money is a “windfall” and has not yet been assigned a mental account, then putting it into saving does not require reducing another account. By allowing individuals to purchase an IRA with their tax refund, the 2006 Pension Protection Act allows people to save without incurring deprivation. In addition, since this is additional money, there will not be any deprivation due to a reduction in one’s standard of living.

SUMMARY AND CONCLUSIONS

This article identifies the constructs that influence an individual’s intention to save and discusses how and when these factors can be changed by an agent trying to induce an individual to enroll in a retirement plan, increase his or her contribution to a plan, or purchase a particular retirement product. In addition to placing existing efforts (primarily structural and educational approaches) into context, the article explores a third approach- persuasive communications-that has received far less attention from private agents, public policy makers, and scholars. These persuasive communication strategies encompass both approaches that agents should follow and approaches that agents should not follow to encourage retirement savings.

For the most part, the recommended persuasive communications influence a person’s subjective assessment of his ability to save, his perception of the short-term nonmonetary benefits associated with the act of saving, and the degree of importance or utility associated with consequences experienced during his retirement years (i.e., hopes and concerns). They can be roughly categorized into five main themes.

Theme 1: Be Positive

The overall theme of preretirement communications should change from one emphasizing crisis/doom (focusing on the failure of people to save) to one emphasizing bandwagon/hope (focusing on how people are saving). Emphasizing the positive will enhance perceived ability and exploit the power of descriptive norms. Note that crisis/doom is the frame typically used by organizations attempting to influence public policy by influencing public opinion. What prosavings advocates need to recognize is that communications that may motivate a person to support government policies supportive of retirement may well diminish the same person’s personal willingness to save. In addition to reducing perceived ability and reversing the impact of descriptive norms, the “government action is needed due to the crisis” message can both create fear and reduce an individual’s sense of personal responsibility.

Theme 2: Be Rich in Imagery and Detail

Prosavings communications need to be rich in imagery and focus on the concrete needs and script-based outcomes of meeting one’s retirement objectives. seeing the value of how one can meet abstract higher-order goals in concrete terms will enhance the value of having additional income during retirement. To the extent that messages can present positive pictures and stories of people similar to the target market, it is likely that future outcomes will be more highly valued. In addition to employing this approach in prosavings communications, prosavings advocates should consider imitating the public health advocates. Over the past decade, public health advocates have been working with the television networks to insert prohealthy behavior messages into story lines. The analogy is for prosavings advocates to begin to work with the mass media to insert the theme that one who has saved can live a good life.

Theme 3: Do Not Ask People to Do Too Much

Savings opportunities need to be structured (in terms of initial requested action, financial information provided, wording, and timing) to maximize the likelihood of enrollment, even if the plan is not optimal. For many people, moving from their current level to an optimal level is too daunting a task. If too much is required to reach a goal and failure to reach the goal is framed as failure, then a person may simply do nothing. Automatic escalation may overcome this problem. However, if this feature is not present, promoting savings becomes a continuous process. Once a worker is enrolled in a plan, the task becomes one of continually seeking to influence him to increase his contribution level. Continuous marketing is consistent with the argument that a person should be led to think about the retirement period consequences of saving in terms of positive, concrete images, rather than in terms of achieving a given replacement ratio. Because imagery is open ended, a person who is saving will not be able to assume that since he enrolled in a plan that promises to produce a given replacement level, he now needs to do nothing more.

Theme 4: Use Social and Normative Pressures

The social/normative benefits of saving (e.g., what others are doing, peer pressure, impact on one’s family) should be emphasized, as opposed to casting the decision to save as a purely individual decision. This approach may be well suited for trying to get young people to save. Retirement is so far off in the future that strong, positive imagery; financial incentives; and perceived goal efficacy may not be enough to overcome the negative effect of discounting. Because the benefits produced by norms come from engaging in the behavior rather than the material consequences of engaging, they cannot be discounted and at the same time closely fit the experiential orientation of modern young people. In essence, we are suggesting that agents trying to promote retirement savings use messages that psychologically mirror the messages used by other marketers who are targeting young adults. Specific messages could include: do what others are doing (descriptive norms), gain approval from one’s friends (injunctive social norm), and feel good about one’s self for just doing it (personal norm).

Theme 5: Do Different Things to Different People

Segmentation should be used to identify both the messages that certain groups should and should not get. The groups can be defined in three potential ways. First, a group can be defined, as in the previous example, in demographic terms. Next, a group can be defined in terms of its current retirement savings status. For example, because both fear appeals and retirement worksheets can either encourage more saving by people who see themselves as being clearly capable reaching a goal set by a financial expert or discourage saving by people who do not see themselves as being capable, sending different messages to each segment may be appropriate. Finally, a group can be defined in terms of the structural features of its retirement plan, for example, automatic enrollment. It may well be that employees who are automatically enrolled in a plan should receive very little in the way of retirement education or promotion. This seemingly odd suggestion is based on the contrast in enrollment rates between plans offering automatic enrollment (do nothing) and quick enrollment (check a box). The finding that participation rates are so much higher in automatic enrollment plans suggests that many employees are enrolling without consciously thinking about retirement. If this is the case, an employer using an automatic enrollment plan might be wise to avoid actions that could encourage a prospective enrollee to think about saving for his retirement. In conclusion, persuasive messages can be an effective means of encouraging individuals to save for their retirement. Using persuasive messages complements, both past and current structural and education approaches. Certainly, both the old (e.g., matching, precommitment, and tax deferral) and the new (e.g., automatic enrollment and quick enrollment) structural approaches are beneficial and hopefully will rapidly diffuse. In addition, conventional investor education programs, which help make workers more knowledgeable and more effective savers, are valuable tools in the fight to bring retirement income security to Americans. However, as long as the structural programs are neither fully subsidized by the government nor mandated at either the individual level (like social security) or the employer level (like health care in Massachusetts) and as long as workers lack confidence in their ability to save, there will be a need to encourage individuals to willingly choose to put money aside for their retirement. Therefore, one arrow in the quiver of those seeking to induce people to save should be persuasive communications.

1. VanDerhei estimates all that if employers who currently offer 401(k) plans offered automatic enrollment plans with a six percent contribution level and a lifecycle investment feature as the defaults, the median replacement rates would range from 52 percent for the lowest-income quartile of workers to 67 percent for the highest.

2. If more people participate in a plan which features employer matching, then employers must pay more in matching contributions. Some experts argue that, for this reason, the ability of automatic enrollment plans to achieve very high participation levels will work against their adoption.

REFERENCES

Akerlof, George. 2001. Procrastination and Obedience. American Economic Review, 81 (May): 1-18.

American Savings Education Council. 2007. http:// www.choosetosave.org.

Andrews, Emily S. 1992. The Growth and Distribution of 401(k) Plans. In Trends in Pensions 1992, edited by John A. Turner and Daniel J. Beller (149-176). Washington, DC: U.S. Government Printing Office.

Angeletos, George-Marios, David Laibson, Andrea Repetto, Jeremy Tobacman, and Stephen Weinberg. 2001. The Hyperbolic Consumption Model: Calibration, Simulation, and Empirical Evaluation. Journal of Economic Perspectives, 15 (Summer): 47-68.

Bagozzi, Richard P. 1992. The Self-Regulation of Attitudes, Intentions, and Behavior. Social Psychology Quarterly, 55 (June): 178-204.

Bagozzi, Richard P. and Elizabeth A. Edwards. 1998. Goal Setting and Goal Pursuit in the Regulation of Body Weight. Psychology and Health, 13 (4): 593-621.

Bandura, Albert. 1986. Social Foundations of Thought and Action: A Social Cognitive Theory. Englewood Cliffs, NJ: Prentice-Hall.

Benartzi, Shlomo and Richard Thaler. 2007. Heuristics and Biases in Retirement Savings Behavior. The Journal of Economic Perspectives, 21 (Summer): 81-104.

Beshears, John, James Choi, David Laibson, and Brigitte Madrian. 2006. Simplification and Savings. NBER Working Paper Series No. 12659 (October). Cambridge, MA: National Bureau of Economic Research.

Browning, Martin and Annamaria Lusardi. 1996. Household Saving: Micro Theories and Micro Facts. Journal of Economic Literature, 34 (December): 1797-1855.

Cialdini, Robert B., Carl A. Kallgren, and Raymond R. Reno. 1991. A Focus Theory of Normative Conduct: A Theoretical Refinement and Reevaluation of the Role of Norms in Human Behavior. In Advances in Experimental Social Psychology, edited by Leonard Berkowitz, vol. 24 (201-234). New York: Academic Press.

Conte, Christopher. 1998. The National Summit on Retirement Savings: Agenda Background Materials. Washington, DC: American Savings Education Council.

Copeland, Craig. 2006. Employment-Based Retirement Plan Participation: Geographic Differences and Trends, 2005. EBRI Issue Brief No. 299 (November). Washington, DC: Employee Benefit Research Institute.

Duesenberry, James S. 1967. Income, Saving, and the Theory of Consumer Behavior. New York: Oxford University Press.

Duflo, Esther and Emmanuel Saez. 2002. Participation and Investment Decisions in a Retirement Plan: The Influence of Colleagues’ Choices. Journal of Public Economics, 85 (1): 121-148.

Employee Benefit Research Institute. 2000. 1996 Retirement Confidence Survey www.ebri.org/RCS/1996results.html.

Engen, Eric, William Gale, and John Scholz. 1996. The Illusory Effects of Saving Incentives on Saving. The Journal of Economic Perspectives, 10(Fall): 113-138.

Farkas, Steve and Jean Johnson. 1994. Promises to Keep: How Leaders and the Public Respond to Saving and Retirement. New York: Public Agenda, in collaboration with the Employee Benefit Research Institute.

Farkas, Steve, Ali Bers, and Ann Duffett. 1997. Miles to Go: A Status Report on Americans’ Plans for Retirement. New York: Public Agenda.

Feenberg, Daniel and Jonathan Skinner. 1989. Sources o