Context

Traditional models of saving have been found to be inconsistent with empirical findings, leading researchers to develop alternative models and this study is based on one of these alternative models: prospect theory (Kahneman& Tversky, 1979; Tversky & Kahneman, 1992). Prospect theory states that individuals are not always rational in the face of uncertainty and that consumption and saving decisions are based on a reference point, rather than on lifetime income, thereby allowing saving horizons to vary.

The study aims to contribute to gaps in the current research base as :

• Relatively few researchers have studied motives for saving and the effect of saving horizon empirically;
• Most studies focus on only one savings motive and relatively few allow for several saving motives to coexist;
• There is a lack of knowledge on how different saving motives are related to saving behaviours.

The study

The study uses data from a large, nationally representative dataset in the US, the 2007 Survey of Consumer Finances (SCF). The sample for this analysis comprises 3,822 working age respondents. It considers three hypotheses:

1. Holding other things constant, holding a specific saving motive will significantly affect the likelihood of saving;
2. Holding other things constant, a longer saving horizon will have a significant positive effect on the likelihood of saving;
3. Holding other things constant, saving motives will affect the likelihood of saving differently among the saving horizon groups.

To test the hypotheses, logistic regression and likelihood ratio tests are used

Key findings

• The findings support the idea from prospect theory that a household’s saving is based on a reference point (or what the household is used to), since having income above the reference level significantly increases the likelihood of saving over the past year (p<0.01), while having income below the reference level significantly reduces the likelihood of saving over the past year.
• The findings provide mixed support for the relationship between saving motives and saving behaviour:
• Only having a motive to save for the education of children or grandchildren was found to be significant in explaining the likelihood of saving, with such households being 13% less likely to have saved over the past year (p = 0.05); - Having an emergency or retirement saving motive significantly increases the likelihood of saving regularly: those with a motive to save for emergencies are 21% more likely to be regular savers;
• Having a motive to save for retirement increases the odds of being a regular saver by 44% (both (p = 0.001). However, the retirement motive does not significantly increase the likelihood of having saved over the past year, indicating that making people aware of the need to save for retirement saving may not be leading individuals to actually save for retirement. Another possibility is that individuals do not view putting money aside for retirement as saving, since they will not access the funds for a number of years;
• The down payment and bequest motives were not significant in predicting the likelihood of saving over the past year, or of being a regular saver;
• Saving horizon has a significant effect on the likelihood of saving and the likelihood of being a regular saver, even after controlling for other variables. Having a medium or long saving horizon was found to significantly increase the likelihood of saving relative to having a short saving horizon, in line with previous studies. Households with a medium saving horizon are 45% more likely to be regular savers while households with a long saving horizon are 85% more likely (p < 0.001);
• Policy makers should strive to change the focus of consumers from the present or immediate future to a longer time period;
• Saving motives vary by saving horizon, but more research is required to explore this further.

Points to consider

• • The sample is limited to non-retired respondents since retirees have been shown to have different saving behaviours to those not retired;
  • Measures of saving are self-reported, so respondents may not respond honestly or accurately;
  • Analysis was limited to saving/not saving rather than being able to consider the amount saved;
  • Prospect theory has components that can be applied to saving behaviours, but was not designed solely to model saving behaviours and lacks features that have been shown to affect saving, such as motivation;
  • The saving motive variables used were based on respondents’ reasons for saving, regardless of whether the individual was actually saving for that goal at the time of the survey. The approach was limited to exploring the relation between the saving motive variables and whether a household spent less than income over the previous year or saves regularly.
• • A number of findings are in line with those from other studies, as indicated above;
  • The financial policy context and economic context has changed since the study was undertaken and differs between the US and UK, which may place some limits on transferability of findings.
• • The study implies that getting consumers to establish a certain saving goal does not necessarily mean that savings will occur, and additional steps are required. For example, offering automatic saving plans could help those who plan to start saving but fail to do so;
  • It is important to include a saving horizon in any saving model. The strong positive effect of having a long horizon on saving is also at odds with the prevailing culture of immediate consumption.