Guest Post: Information and Planning in Saving Decisions

I am pleased to present this guest contribution by Meryl Motika, an economics job market candidate at UC Irvine. She describes a cleverly-designed laboratory experiment about the factors that influence people's decisions about spending and saving.

Economists have an elegant model of how people might save called the Life Cycle Hypothesis. The basic idea is that people want to maintain a consistent standard of living over time. According to this model, we should see people go into debt when they are young, pay off their debts and save as their wages rise, and retire with enough wealth to maintain their lifestyle until they die.

The Life Cycle Hypothesis seems to fit some people’s behavior pretty well. However, a significant portion of the population either saves much more than they need or fails to save even during their prime earning years. For these individuals, consumption (how much money they spend) tends to go up and down with current income instead of being smoothed over the life cycle.

A fascinating paper by Erik Hurst called Grasshoppers, Ants and Pre-Retirement Wealth: A Test of Permanent Income Consumers shows that the same individuals who fail to save enough to cover temporary layoffs also fail to save enough for retirement. It’s tempting to imagine that they just don’t earn enough to save, but in practice it’s not so easily explained. At any given income level, some people seem to follow the Life Cycle Hypothesis model while others save too much or don’t save.

So why do some people save while others don’t? A number of reasons have been suggested including that people tend to do blindly whatever is suggested to them, failures to plan ahead, differences in thinking about the future, ability to resist temptation, and differences in financial knowledge. There is evidence in favor of each of these explanations. However, they are difficult to study because it’s hard to get good information about wealth, income, probable future income, life expectancy, and other factors in saving decisions. Worse, someone who saves a lot might also get a job with a good retirement plan, resist temptation, and learn about finance just because they’re that type of person.

I decided to take this puzzle to the laboratory and designed an experiment in which subjects decide whether to “spend” or “save” tokens to watch video clips. Since the alternative is looking at a still screen, boredom provides a strong incentive to spend the tokens early. The incentive to save is that the video clips are longer towards the end of the experiment; savers get to watch more video. The great thing about this design is that it is very open-ended. I observe actual spending decisions made over a span of time, without any particular decision being obviously “correct.”

In my paper Information and Planning in Saving Decisions, I find that even in a very short (50 minute) session subjects make various saving choices; some watch all their videos immediately, some save for the end of the session, and some spread the videos out over the session. I was also able to examine the effects of “financial information” (extra instructions with examples and screenshots to make people more comfortable with their options) and being asked to plan ahead. In my experiment, information causes many subjects to save rather than watching the videos immediately. Planning ahead appears to make subjects more likely to smooth their vide-watching over the session, although this point requires further testing.

What can we learn from this? First, under-saving isn’t all about being susceptible to temptation. How the saving problem is presented affects subjects’ decisions even though the temptation to watch videos early is always the same. Second, I was able to induce something like normal saving behavior in the laboratory, meaning that saving decisions can be studied in a laboratory setting. This opens doors for future research into the effects of risk, life expectancy, learning from others, and so on. Validation of my results through replication (especially in sessions spread over several weeks) and examining the relationships between laboratory behavior and actual saving behavior must come first, though!