Cues to Save Money

Americans have a notoriously low savings rate, a problem we explored in a podcast about prize-linked savings plans. In another podcast, “A Mouse in the Salad,” Richard Thaler (author of Nudge) discussed “anchoring,” a cognitive shortcut whereby we make decisions based on an anchoring number even if it is randomly generated.

A new NBER paper (ungated version here) by Yale’s James J. Choi and Cade Massey, along with Emily Haisley of Barclays Bank and Jennifer Kurkoski of Google, shows that anchoring very much affects how people save (or don’t save) their money.

The researchers conducted a field experiment in which employees at a technology company received an e-mail about their 401(k) savings plans. The control group received an email with just the basics: a reminder of employer matching contributions and of how much the employees had put away in their 401(k). Another group received the same email with a short anchoring message: “For example, you could increase your contribution rate by 1% of your income and get more of the match money for which you’re eligible.” The experiment also tested two other cues: suggesting a goal ($7,000, for example), and highlighting the company’s match incentive or the maximum possible contribution rate.

The researchers found that anchors increase or decrease 401(k) contribution rates by up to 1.4 percent of income. For all cues tested, low cues decrease contribution rates by up to 1.5 percent of income, while high cues increase contribution rates by up to 2.9 percent of income. Based on their findings, the authors reckon that a few carefully selected words could be an easy way to help people save:

Our findings provide both an opportunity and a warning for organizations and policy makers. The kinds of cues we investigate could be intentionally used to influence saving behavior more efficiently than through the use of more costly interventions, such as financial education or increases in matching incentives. But unintentional cues buried in mundane communications can also affect behavior. Thus, organizations and policymakers should take responsibility for the cues they disseminate and wield them mindfully.

Cyril Morong

What is the optimal savings rate and how do we know that all this nudging will achieve it? What if everyone at my company is nudged to save more? Then we don't have the extra money to go buy sandwiches for lunch from the place across the street. That owner goes out of business. Is that optimal?

Mike H

Wow, what a weird perspective out from left field.

A great follow up post for this story would be: what is the benefit to the company to encourage their employees to contribute more to their retirement plan? Especially, if it costs the company more money as more employees would be eligible for the employer-funded match. Is it some sort of tax break for x number of dollars saved?


Is not one of the problems with lack of saving in America that the tax code discriminates against saving(and why we have these convoluted 401k and IRA plans)?

Earn income> Pay tax.
You can use after tax income to either buy goods or invest, but if you invest...
Investment Gains value: Capital gains tax.
Company you invested has profits: Corporate profits taxed.
Dividends issued: Dividends tax.


I wonder if those cues do, or do not, include communication about how funds in 401(k) plans technically belong to the U.S. government, and can therefore be confiscated if the government determines that is necessary. That might be a bit of a drag on getting people to invest in 'em!

Eric M. Jones.

"Saving Money" or even investing money has been shown to be a fool's game. Whilst I had a comfortable IRA for retirement, powers far more clever than I planned and DID make it vanish.

"Saving?" I say invest in baked beans and bullets. Lots and lots of bullets. And torches and pitchforks.

7.62x39mm Ammo - 122gr FMJ: $1600/10,000 Rounds

Canned Bake Beans In Tomato Sauce: $750 / 1000 kg

Torches and Pitchforks: $/4.25 each in Q of 10,000 sets.

This would have out-performed my Schwab IRA handily.


The reason Americans have a low savings rate?

Americans have no #*#*$ money.

When your bills outstrip your income--and you're living sensibly with a modest mortgage, given your region, not spending stupid, not going out to dinner, not buying things you don't need-need (as opposed to want-need)--how are you supposed to save? Costs keep going up; wages keep stagnating, and the kids still need shoes.

Enter your name...

From my perspective, as a person living in a very expensive real estate market, I'd guess that your problem is the "modest mortgage". Just because the mortgage cost is modest for your area doesn't mean that your family can actually afford it. Rather than buying a home, you might need to consider renting a smaller apartment.

But the good news is this: in the long term, buying that house is itself a form of savings. Someday (if you pay off the mortgage, rather than endlessly refinance it), you'll be able to live there rent-free. So figure out how many years of rent-free living you expect, and buying your home is the equivalent in savings.


I wonder what fraction of savings is intentional, and what is almost by default. In my case, I make more than I need to spend to live comfortably, have no interest in spending for status (and avoid TV and most other popular media, so usually don't have much clue as to what possessions are culturally supposed to confer status), and dislike shopping. So what else am I supposed to do with my excess income?


I don't get it. Why would you need a sales pitch to get free money. Contributing up to the match gets you 100% return your first year. In any case you are going to have to save much more than that to retire. So why not take the free money and then you can decide how to save the rest of what you need for retirement.

Erik - Dallas

Not sure where to even begin.
Most other studies of employee savings behavior lament the difficulty eliciting any response out of 401(k) plan participants. Communications help but are less than optimally effective at eliciting participation and contribution rates that would fund adequate retirement savings. The 401(k) plans that get high participation and high contribution rates having automatic enrolment in the plan and automatic escalation of contributions. The only real danger is that since most participants do not opt to actively elect alternative plan contribution rate choices, are the automatic rates set too low and thus are actually reducing contributions and missing potential employer matching contributions... The other potential danger (that I really have not encountered) would be if automatic contribution rates were set too high, employees might elect out of the plan.
To paternalistically take care of employees and to encourage older workers to retire and not work past their productive working lifetime, employers have provided some form of pension plans. Historically this would have been a defined benefit (DB) plan with a final average pay times service formula. Over the last 30 years companies have shifted away from defined benefit plans and moved to defined contribution plans (DC) - what you know as the 401(k)s. Usually in the process the company has also shifted a substantial portion of the responsibility for funding the retirement plan to the employees. In the modern financial markets of quarterly profit reporting, companies prefer the regular predictability of contribution to defined contribution plans over the variable less predictable contributions needed to fund defined benefit plans. However, companies still want employees to be able to retire (and not be stuck with a 90 year old work force) and thus they are interested in promoting 401(k) plan participation and contributions. Even though providing matching contributions “increases” employer cost; this is the way that employers encourage employees to participate in 401(k) plans. For an employer paying 50% of DC plan contributions is better than paying 100%. DB cost or being stuck with 90 year old employees. Further, the standard 6% match is usually much lower than what a DB plan would have cost. In fact most employees should probably be contributing between 10% and 20% of pay to their 401(k) plan, but most are contributing less than 10%.
What should saving be and why? Generally the idea is to allow for consumption over a lifetime to maintain the same (or a slightly better) quality of life. This requires substantial savings. Modeling observed human preferences usually results in some kind of immediacy or present-bias quasi-hyperbolic discounting combined with a habit index utility function that rewards level or increasing consumption but has great adversity to any decrease in consumption. Because of the immediacy bias, a return on investment is needed to incentive savings. These incentives can be in the form of interest, dividends, investment return, tax breaks, matching contributions, other or some combination of these; just so long as the expected return is sufficiently large to overcome the discounting of the immediacy bias. Some retirement models allow for a slight reduction in consumption during retirement (also known as a retirement income “replacement ratio” of employment income); the optimal consumption would not reduce the quality of life, but some expense are no longer needed such as business related formal clothing and transportation expenses and potentially housing expenses can be lower if the mortgage is paid off or if once the kids have left and a smaller place is sufficient. But other models in accounting for general inflation or even higher medical inflation actually result in optimal replacement ratios above 100%. With life expectancy at age 65 being around 20 years and longer if one retires earlier, and with some living past 100, saving adequately for retirement could easily require having investments worth more than 10 times ones salary (but probably not more than 20 times salary). To achieve this level of savings usually requires annual savings in excess of 10% of salary.
As we shift away from DB plans to DC (401(k)) plans most US employees only have one to five times salary in savings at retirement age. This would appear to be less than optimal under habit based utility functions that reward maintenance of quality of living throughout ones life. Further I would expect that in the future a substantial portion of our growing elderly population will be living in or near poverty levels.