Comparing winners with similar account balances in the same drawing holds baseline wealth constant while isolating randomized variation in winning an odds prize versus a fixed prize.

Download image · 1200 × 630 PNG

Research procedure

Build balance-matched odds-prize cells

  1. Select single-prize winners from the same draw
  2. Match account balances within narrow intervals
  3. Reassign matched winners; retain qualifying cells

Illustrative summary of the source-described odds-prize matching. Match a single odds-prize winner to single fixed- or odds-prize winners. Move matched fixed-prize winners from old cells; exclude cells with winnings below 100,000 SEK.

See source →

How

When to use

Researchers need to isolate exogenous wealth shocks among winners of account-balance-dependent lottery prizes in prize-linked savings programs. Account balances directly increase odds-prize amounts, creating endogeneity between winner wealth and unobserved labor preferences. Matching odds-prize winners to other prize winners with similar account balances eliminates balance-driven selection bias.

Construct comparison cells by matching single odds-prize winners to single prize winners in the same monthly drawing who hold similar account balances.

  1. Identify account holders who won exactly one odds prize in a specific monthly draw during the microfiche observation period.
  2. Select account holders in the exact same monthly draw who won exactly one prize of either fixed or odds type.
  3. Measure account balances of all eligible winners in that draw using microfiche administrative records.
  4. Group each odds-prize winner with winning account holders whose account balances fall within an identical narrow balance interval.
  5. Reassign matched fixed-prize winners from their original fixed-prize cells to the newly created odds-prize cell.
  6. Exclude odds-prize cells where total winnings fall below 100,000 SEK to maintain manageable cell volume.
  7. Assign cell fixed effects to control for account balance baseline variations in panel regressions.

Example

In December 1990, researchers constructed a matching cell containing one winner of a single odds prize and 19 winners of a single fixed prize who all held account balances between 3,000 SEK and 3,200 SEK in the Vinnarkontot prize-linked savings lottery.

About the study

Participants

1,620 odds-prize cells constructed from Swedish prize-linked savings account holders aged 21 to 64 participating between 1986 and 1994.

Study Design

Panel regression controlling for balance-matched odds-prize cell fixed effects to measure five-year labor earnings responses to randomized wealth shocks.

Results

Winning a lottery prize reduces annual pre-tax labor earnings by 1.066 SEK per 100 SEK won, with wage reductions (-0.964 SEK per 100 SEK) driving the response.

Source

See source →