In 2006 a new tax favored savings scheme was introduced in The Netherlands, the so-called Life Course Savings Scheme (‘Levensloopregeling’). Employees can save substantial amounts of their monthly salary in this savings scheme, but are not obliged to do so. In this savings scheme, employees can only save to buy time for parental leave, a sabbatical or early retirement. We employ monthly salary data for two large companies in the financial sector, in total over 30,000 employees. A first finding is that the marginal propensity to save for this scheme out of different salary components (e.g.) gross salary, holiday allowance and bonuses) differs. This is a violation of the principle of fungibility of money. Moreover, one of the two companies paid a small, once-only contribution with the name of the Life Course Savings Scheme communicated. Participants in the savings scheme could choose whether they wanted this contribution either to be deposited in the savings scheme, or to have it paid out on top of their salary. Employees who wanted to deposit this contribution had to fill out a form and specify the amount to deposit. Despite this discouraging default setting, we find evidence that this labeled payment has the largest marginal propensity to deposit compared to other salary categories. We take this as evidence for the so-called labeling effect. We also use semi-parametric methods to test for these labeling effects. Moreover, one of the two firms changed the institutional setting of payments from the year 2006 to the year 2007. Up to 2006, some salary components, like the holiday allowance, used to be paid once a year. In 2007, the total amount of money of these salary components was paid in equal shares every month, leaving the total annual amount of salary unaffected. We investigate the effects of this change in timing on savings contributions. Finally, we use the empirical findings to simulate possible policies to take advantage of the labeling and timing effects by increasing the effectivity of tax favored schemes and/or increasing employee welfare. One policy recommendation is that firms and the government can strongly influence savings behavior by labeling salary components. Another policy recommendation is that the frequency of payments matters for the outcome.

Keywords: labeling effects, timing effects, employee savings, behavioral economics
JEL Classifications: D03, D91