LOTTERY Scott Hankins Mark Hoekstra Paige Marta Skiba University - - PowerPoint PPT Presentation

THE TICKET TO EASY STREET? THE FINANCIAL CONSEQUENCES OF WINNING THE LOTTERY

Scott Hankins University of Kentucky Paige Marta Skiba Vanderbilt University Mark Hoekstra University of Pittsburgh June 2010

Primary Question

Does giving cash to people in financial distress help them avert bankruptcy?

Motivation

1.

The Lottery as the Golden Ticket  Does winning the lottery reduce the future financial distress of lottery winners?  The introduction of a state lottery is associated with an average decline of $46 per month on non-gambling expenditures. (Kearney 2005)  e.g. Lotto lout Michael Carroll going back to being a binman after blowing £9.7m “The party has ended and it's back to reality. I haven't got two pennies to rub together and that's the way I like it. I find it easier to live off £42 dole than a million.”

Motivation

1.

The Lottery as the Golden Ticket  Does winning the lottery reduce the future financial distress of lottery winners?  The introduction of a state lottery is associated with an average decline of $46 per month on non-gambling expenditures. (Kearney 2005)  e.g. Lotto lout Michael Carroll going back to being a binman after blowing £9.7m “The party has ended and it's back to reality. I haven't got two pennies to rub together and that's the way I like it. I find it easier to live off £42 dole than a million.”

Motivation

1.

The Lottery as the Golden Ticket  Does winning the lottery reduce the future financial distress of lottery winners?  The introduction of a state lottery is associated with an average decline of $46 per month on non-gambling expenditures. (Kearney 2005)  e.g. Lotto lout Michael Carroll going back to being a binman after blowing £9.7m “The party has ended and it's back to reality. I haven't got two pennies to rub together and that's the way I like it. I find it easier to live off £42 dole than a million.”

Motivation

2.

The Desirability of Lump-sum Tort Settlements  Concern about lump-sum settlements => move toward structured settlements  “Lump sum payments all too often are improvidently invested or squandered by unsophisticated recipients and so fail to provide for the lifetime of medical bills and unemployment faced by victims of serious injury.” Judge Joseph Weis, Third Circuit 3. The Lottery as a Bailout  Does giving indebted individuals additional resources help them avert bankruptcy?  Relevant to recent debates on loan modifications?

Considerations:

 Rational  Strategic gaming of bankruptcy system  Irrational  High discount rates

 Rational?  Short-run impatience vs. long-run selves’ preferences

 Consumption Commitments  Habit formation  Money may not be fungible

 Mental Accounting  “Treat your found money with the same respect as earned money.”

SuddenMoney.com

 Gambling with house money  Financial illiteracy

Identification

 Problem: endogeneity of large transfers

 e.g., Tort claimants, bailout recipients

 Our solution: Exploit variation in amount won in Florida’s Fantasy 5 lottery

game

 Identifying assumption:

 The magnitude of cash prizes is random conditional on winning.

 Identification Strategy: Compare bankruptcy filing rates of large winners

($50,000 - $150,000) to small winners (<$10,000) from Florida’s Fantasy 5 game.

 Similar to Imbens, Rubin, and Sacerdote (2001) and Lindahl (2005)

Florida Fantasy 5

 Players pick 5 numbers є ¡(1, 36)  Ticket costs $1  Manually choose numbers or have computer choose

randomly

 IRS withholds 25%

Data

Total Fantasy 5 Winners: 56,160 Unique in 2008 Phone Book: 38,836 First-time Winners: 34,987 Amount Won Frequency $600 - $10,000 14,668 $10,000 - $50,000 19,107 $50,000 - $150,000 1,212

Identification Strategy

 Identifying Assumption:

Conditional on winning at least $600 one time, the amount won is uncorrelated with one’s underlying propensity to file for bankruptcy Bankruptcyi = YearFE + β0 (After Game Change) + β1 ($10,000 ≤ Amount < $50,000)i + β2 ($50,000 ≤ Amount< $150,000)i + εi

Tests of the Identification Strategy

 Regress amount won on 13 demographic variables

(income, education, gender, and interactions)

 Explain only 0.1% of total variation

 Compare the bankruptcy filing rates of large and

small winners before they won

 Compare 4-number matchers to 5-number matchers

(for sample after July 2001)

Flows into Bankruptcy (after removing year effects)

Robustness Checks

 Probit  Redefine control group  <$2,500  <$1,000 (includes both Florida Lotto and Fantasy 5)  Include controls for total payout and maximum prize from the previous drawing  Include controls for exposure to effects of the Bankruptcy Abuse Prevention and Consumer Protection Act  # Months from March 1, 2005 – October 16, 2005  # Months from October 17, 2005 – November 26, 2007  Compare players who matched 5 numbers to players who matched 4 numbers ($80,000 versus $1,000)

Attrition

 Attrition could cause bias under two conditions:

1.

Amount won is correlated with propensity to move out of the county

2.

At least some of the individuals who moved out of the county filed for bankruptcy elsewhere

 Cause for concern?



Counties are large

 Average (by population) county in Florida is 1,866 square miles  > 6x New York City

 Test:



Are large winners more likely to be observed in the phone book 1 to 6 years later?

Additional Questions

 Are winners filing for bankruptcy due to consumption commitments?

Large Small Homeowner 60% 50% Real Estate Market Value $89,521 $82,427

 Are recipients gaming the bankruptcy system?

Large Small Real Estate Equity $21,582 ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡$24,412 ¡

 Do cash transfers improve the debt profile of bankrupt individuals even if

they don’t help them avoid bankruptcy? Large Small Net Assets $-­‑15,265 ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡$-­‑23,468

Conclusions

 Recipients of $50,000 - $150,000 are 1.6 percentage points (50%) less

likely to file for bankruptcy in the two years after winning

 BUT: Large winners are 1.2 percentage points MORE likely to file for

bankruptcy 3 to 5 years after winning

 No evidence that large winners gamed the unlimited homestead

exemption in Florida bankruptcy law; only some evidence of increased housing consumption commitments

⇒ Bottom Line: ⇒ Receiving large amounts of cash appears to delay bankruptcy rather than

prevent it

Extensions

 Effect of lump sums on home foreclosures

 Collecting foreclosure data (Lis Pendens) for three

largest counties in FL.

 Effect of small lump sums on payday loan debt,

repayment and default.

 Exploit distribution of tax rebates in 2001

 Mail date of $300 check based on last two digits of Social

Security Number

Florida Fantasy 5

 April 1993 through June 2001

 Individuals who matched 5 of 5 numbers won an average of about $20,000

 Prize for matching 5 numbers varied from $1,300 to $132,000 depending

• n the number of winners/players.

 July 2001 through November 2002

 Matching 5 numbers won an average of $120,000  Matching 4 numbers when no one matched 5 won an average of $900

 Implications:

 Many more winners of <$10,000 after 2001, fewer winners of between

$10,000 and $50,000

 Controlling for change in game structure or year played is important

Data

 34,987 winners of the Fantasy 5 game in the Florida Lottery from April

29, 1993 to November 26, 2002

 1.4 million bankruptcy records filed from 1988 to November 26, 2007

(Source: PACER)

 Linked using first and last name and county of residence  Variables

• Date Won
• Zip Code
• Amount won (adjusted for CPI)
• Date of Bankruptcy
• First and last name
• Chapter of Bankruptcy
• More detailed bankruptcy information for a subset of cases filed after

2004 including total assets, total debt, monthly income, and monthly expenditures.

Florida Fantasy 5

 Prize size depends on:

 Structure of the game (pre-July 2001 prizes smaller)  Number of winners relative to number of players  How many numbers you matched (for post-2001 prizes)  How many times you played the identical number on a card (But: no one

played same 5 numbers multiple times)

 Prize size is large relative to debt and income:

 Median large winner: $65,000  Average unsecured debt of the most distressed players at time of winning (i.e.,

those who just filed for bankruptcy): $49,000

 Average annual household income of most distressed players at time of

winning: $17,200

The Effect of Winning the Lottery on Bankruptcy Rates within 2 Years

Bankruptcy ¡Rate ¡within ¡2 ¡Years ¡a4er ¡Winning ¡

Won ¡$10,000 ¡-­‑ ¡$50,000 ¡

• ­‑0.0166*** ¡ -­‑0.0086** ¡
• ­‑0.0106*** ¡
• ­‑0.0087** ¡

(0.0016) ¡ (0.0038) ¡ (0.0033) ¡ (0.0038) ¡ Won ¡$50,000 ¡-­‑ ¡$150,000 ¡

• ­‑0.0215*** ¡ -­‑0.0160*** ¡ -­‑0.0176*** ¡
• ­‑0.0163*** ¡

(0.0043) ¡ (0.0050) ¡ (0.0048) ¡ (0.0050) ¡

N ¡ 34,987 ¡ 34,987 ¡ 34,987 ¡ 34,987 ¡ Controls ¡for ¡change ¡in ¡game ¡ structure? ¡ No ¡ Yes ¡ No ¡ Yes ¡ Includes ¡year ¡fixed ¡effects? ¡ No ¡ No ¡ Yes ¡ Yes ¡

The Effect of Winning the Lottery on Bankruptcy Rates 3-5 Years Later

Bankruptcy ¡Rate ¡within ¡3-­‑5 ¡Years ¡a4er ¡Winning ¡

Won ¡$10,000 ¡-­‑ ¡$50,000 ¡ 0.0084*** ¡ 0.0040 ¡ 0.0081** ¡ 0.0050 ¡ (0.0020) ¡ (0.0047) ¡ (0.0041) ¡ (0.0047) ¡ Won ¡$50,000 ¡-­‑ ¡$150,000 ¡ 0.0143*** ¡ 0.0113* ¡ 0.0143** ¡ 0.0121** ¡ (0.0054) ¡ (0.0062) ¡ (0.0059) ¡ (0.0062) ¡

N ¡ 34,987 ¡ 34,987 ¡ 34,987 ¡ 34,987 ¡ Controls ¡for ¡change ¡in ¡game ¡ structure? ¡ No ¡ Yes ¡ No ¡ Yes ¡ Includes ¡year ¡fixed ¡effects? ¡ No ¡ No ¡ Yes ¡ Yes ¡

The Effect of Winning the Lottery on Bankruptcy Rates 5 Years Later

Bankruptcy ¡Rate ¡within ¡0-­‑5 ¡Years ¡a4er ¡Winning ¡

Won ¡$10,000 ¡-­‑ ¡$50,000 ¡

• ­‑0.0082*** ¡
• ­‑0.0046 ¡
• ­‑0.0025 ¡
• ­‑0.0036 ¡

(0.0025) ¡ (0.0059) ¡ (0.0051) ¡ (0.0060) ¡ Won ¡$50,000 ¡-­‑ ¡$150,000 ¡

• ­‑0.0072 ¡
• ­‑0.0047 ¡
• ­‑0.0034 ¡
• ­‑0.0042 ¡

(0.0068) ¡ (0.0078) ¡ (0.0075) ¡ (0.0078) ¡

N ¡ 34,987 ¡ 34,987 ¡ 34,987 ¡ 34,987 ¡ Controls ¡for ¡change ¡in ¡game ¡ structure? ¡ No ¡ Yes ¡ No ¡ Yes ¡ Includes ¡year ¡fixed ¡effects? ¡ No ¡ No ¡ Yes ¡ Yes ¡