The Price Elasticity of Charitable Giving: Toward a Reconciliation of Disparate Literatures

Transcription

1 The Price Elasticity of Charitable Giving: Toward a Reconciliation of Disparate Literatures Daniel M. Hungerman Mark Ottoni-Wilhelm University of Notre Dame and NBER Department of Economics, IUPUI and Indiana University Lilly Family School of Philanthropy Philanthropy Research Workshop December 1, 2016

2 Overview - the price elasticity of giving Tax-price literature: e [-.50, -1.30] Match-price literature: e checkbook [0, -.40]. e amt received = e checkbook 1 Innovations: First paper to directly bridge the two literatures: a tax-price e and a match-price e for same group of donors, same organization, same time period. Novel kink methods to estimate a tax-price e identification. Tax-price e Match-price e checkbook -.20 Significance Reconciliation first step. Match-price e checkbook informative for tax policy, at least in this setting and perhaps in other settings.

3 Figure 3. Gifts from Indiana residents to a university (joint). Percent Gifts Notes: Sample size N = Gifts were made jointly, indicating a married couple facing a limit on the tax credit at $400.

4 Figure 4. Match-price effect on donations. 0.6 Match Begins Match Ends

5 Previous literature Randolph (1995) Auten, Sieg, Barrett, McGuirk, Bakija Clotfelter & Steinberg (1997) & Heim (2011) (2002) 0 [ ] Karlan & List (2007) Davis (2006) Scharf & Smith (2015) Eckel & Grossman (2003)

6 The rest of the talk Theory: How the credit and cap affects the distribution of giving. Estimation: How we figure out e from the distribution of giving. Environment: The tax credit and the sample. Results: Three elasticities. Discussion: Implications of the results in the context of the previous literature.

7 Indiana education tax credit Married-joint tax filers get 50% of g given to Indiana colleges, universities, seminaries reduced from income tax owed, for g $400. p =.50 for g ($0, $400], but then p = 1.0 for amounts given over $400.

8 Indiana CC-40 Schedule CC-40 State Form (R / 8-01) Your first name and last name Indiana College Credit for the Year Attachment Sequence No. 08 This schedule is for computing credit for contributions to colleges and universities located in Indiana. This schedule, or a statement showing the same information, must be attached to the contributor's income tax return. A Your Social Security Number Spouse's first name and last name (if filing a joint return) Spouse s Social B Security Number Name of Corporation or Fiduciary D Federal ID Number (if applicable) C PART I Itemized Contributions to Eligible Institutions (See reverse side for list and 4-digit code number) 4-Digit Code Date of Amount Given Indicate below the E Name of Eligible Indiana College or University F Number* G Contribution H I type of return filed by the contributor. $ Individual $ Corporate $ Fiduciary *See 4-digit college code listing on back of this schedule. Column A Column B PART II Individual and Fiduciary Computation of Credit Single or Married Married Filing but Filing Separately a Joint Return 1. Enter the total contributions to Indiana colleges and universities listed above Enter 50% of line Limitation ($100 single return or $200 joint return) Enter the lesser of line 2 or line Enter Indiana adjusted gross income tax from line 14 of IT-40, line 12 of IT-40PNR or line 8 of IT Allowable College Credit: Enter line 4 or 5, whichever is less. Enter here and on IT-40 Schedule 2, line 4; on IT-40PNR Schedule E, line 4; or on IT-41 line PART III Corporation's Computation of Credit 1. Enter the total contributions to Indiana colleges and universities (listed above) Enter 50% of line 1, or $1,000, whichever is less Enter the adjusted gross income tax from the appropriate line on IT-20 or IT-20SC Multiply line 3 by 10% (.10) CREDIT: Enter line 2 or line 4, whichever is less. Enter here and on appropriate line of the corporate tax return... Important: The taxpayer will be required to retain the receipts given by the Indiana colleges and universities indicating that a contribution has been made. These receipts should be maintained for a period of three years after the due date of the annual tax return where the credit was taken or three years after the date that return was filed, whichever is later

9 Indiana CC-40 Contributions made by Individuals, Fiduciaries and Corporations An individual or fiduciary is allowed a credit of fifty percent (50%) of the contribution or $100 (on a single or fiduciary return) or $200 (on a jointly filed return), whichever is less. The amount of a corporation's credit is equal to fifty percent (50%) of the total amount given during the tax year. However, the credit may not exceed the lesser of: a) ten percent (10%) of the corporation's adjusted gross income tax liability, or b) the amount of $1,000. For any further information about this credit, contact the Department at: Indiana Department of Revenue, Taxpayer Services Division, 100 N. Senate Avenue IGCN 105, Indianapolis, IN., , telephone #(317) Eligible Institutions with Code Numbers Code Indiana Code Indiana # Institution Location # Institution Location 5000 Ancilla College... Donaldson 4800 Anderson University... Anderson 2000 Associated Mennonite Biblical Seminary... Elkhart 1800 Ball State University Foundation... Muncie 1801 Ball State University... Muncie 7100 Bethel College... Mishawaka 8900 Bethany Theological Seminary... Richmond 4900 Butler University... Indianapolis 4500 Calumet College of St. Joseph... Whiting 4901 Christian Theological Seminary... Indianapolis 0200 Concordia Theological Seminary... Fort Wayne 6700 DePauw University... Greencastle 8901 Earlham College... Richmond 4100 Franklin College of Indiana... Franklin 2001 Goshen College... Goshen 4301 Grace Theological Seminary and Grace College... Winona Lake 3900 Hanover College... Hanover 7101 Holy Cross Junior College... Notre Dame 3500 Huntington College... Huntington 4902 Independent Colleges of Indiana... Indianapolis 0201 Indiana Institute of Technology... Fort Wayne 8400 Indiana State University... Terre Haute 8401 Indiana State University Foundation... Terre Haute 0202 Indiana-Purdue Foundation at Fort Wayne Fort Wayne 5300 Indiana University Foundation... Bloomington 2700 Indiana Wesleyan University... Marion 5301 Indiana University... Bloomington... and Regional... Campuses 4903 Indiana University - Purdue University... Indianapolis 0203 Indiana University - Purdue University... Fort Wayne 4904 IvyTech Foundation... Indianapolis 4905 IvyTech State College... Indianapolis... and Regional... Campuses 8500 Manchester College... North Manchester 4906 Martin University... Indianapolis 4907 Marian College... Indianapolis 2600 Oakland City College... Oakland City 7900 Purdue Academic Facilities Foundation... West Lafayette 7901 Purdue Agricultural Alumni Association... West Lafayette 7902 Purdue Alumni Association... West Lafayette 7903 Purdue Alumni Foundation... West Lafayette 7904 Purdue Foundation... West Lafayette 7905 Purdue Research Foundation... West Lafayette 7906 Purdue Student Housing Corporation... West Lafayette 7907 Purdue University... West Lafayette... and Regional... Campuses 8402 Rose-Hulman Institute of Technology... Terre Haute 7908 Ross-Ade Foundation... West Lafayette 3700 St. Joseph s College... Rensselaer 8403 St. Mary-of-the-Woods College... Terre Haute 7102 St. Mary s College... Notre Dame 7400 Saint Meinrad School of Theology... St. Meinrad 0206 Taylor University... Fort Wayne... Campus 2701 Taylor University... Upland 7600 Tri-State University... Angola 8200 University of Evansville... Evansville 4908 University of Indianapolis... Indianapolis 7103 University of Notre Dame... Notre Dame 0205 University of St. Francis... Fort Wayne 8201 University of S. Indiana Foundation... Evansville 8202 University of Southern Indiana... Evansville 8203 USI-New Harmony Foundation... Evansville 6400 Valparaiso University... Valparaiso 4200 Vincennes University Foundation... Vincennes 4201 Vincennes University... Vincennes 5400 Wabash College... Crawfordsville 1000 Webster University Graduate Center... Jeffersonville

10 Indiana education tax credit (cont.) Example: Currently giving $350, thinking about giving +$50, p =.50 If there were not a cap (counterfactual), currently giving $400, thinking about giving +$50, p =.50 However, because of the cap, currently giving $400, thinking about giving +$50, p = 1.0 Hence, some people (maybe a lot) who in the counterfactual world would give amounts like $450 (somewhat higher than the cap location at $400) will because of the cap, bunch at the cap location a bunch of people end up giving exactly $400.

11 Theory: No tax credit; price = 1.00 Percent A B C D Gifts Quasi-linear utility: g = θ p e g = θ

12 Counterfactual 1: Tax credit =.50, with no cap and e = 1. Percent A A B B C C D D Gifts (g0) g = θ p e g = θ (.50) 1 = θ(2)

13 Policy in effect: Tax credit =.50, but cap at $400 (e = 1). Percent A A B C (all people in B give exactly $400) D Gifts

14 Policy in effect: Tax credit =.50, but cap at $400 (e = 1). Percent A A B C (all people in B give exactly $400) D Gifts g = θ p e = θ (2) = g 0 if g 0 < $400 = $400 if $400 g 0 < $800 = θ if $800 g 0

15 Estimation: (a) bunching at the cap location and (b) bunching interval. Percent A A B C (all people in B give exactly $400) D Gifts g = θ p e = θ (2) = g 0 if g 0 < $400 = $400 if $400 g 0 < $800 = θ if $800 g 0

16 If the response to price (e) is smaller: The percentage in group B of people who bunch at the cap is smaller, and The bunching interval is more narrow: z = $800 z $400 An estimate ˆB and an estimate of the density of gifts just below $400 An estimate of the width of the bunching interval ( z) ê

17 Counterfactual 1: What would happen in Indiana if the cap were eliminated, but the p kept at.50? The people who give at the cap location because of the cap being in place... what is the largest amount such a person would have given in the counterfactual world with p =.50 and no cap? The most generous (θ) guy to bunch at the cap ( kink ) The guy who would have given g 0 = $800

18 Real world Percent A A B C (all people in B give exactly $400) D Gifts

19 Counterfactual 1 Percent A A B B C C D D Gifts (g0)

20 Counterfactual 1: Compensated price elasticity What would happen in Indiana if the cap were eliminated, but the p kept at.50? The most generous (θ) guy to bunch at the kink... that is the guy at the upper edge of the bunching interval. The guy who would have given g 0 = $800 Can show that the elasticity (e) estimated based on this counterfactual is approximately a compensated price response (Hicks).

21 Counterfactual 2: What would happen in Indiana if the credit (and cap) were eliminated, and the p = 1.00? The people who give at the cap location because of the cap being in place... what is the smallest amount such a person would have given in a counterfactual world with p = 1.00 (no credit; no cap)? The least generous (θ) guy to bunch at the kink. The guy with θ = $200.

22 Real world Percent A A B C (all people in B give exactly $400) D Gifts

23 Counterfactual 2 Percent A B C D Gifts

24 Counterfactual 2: Uncompensated price elasticity What would happen in Indiana if the credit itself were eliminated, and the p = 1.00? The least generous (θ) guy to bunch at the kink... that is the guy at the lower edge of the bunching interval. The guy who would have given θ = $200. Can show that the elasticity (e) estimated based on this counterfactual is approximately an uncompensated price response (Marshall). The guy who would be giving $200 in the counterfactual p = 1.00 world, would get a $100 boost in income when the p =.50 credit is introduced.

25 Kink-based estimation: Basic idea Estimate the (counterfactual) percentage of people who would have given an amount equal to the location of the kink the cap: g 0 = $400 anyway, even if the cap were not in place. Like group C ˆB = Observed mass point at $400 Ĉ ˆB is the percentage of excess bunchers at the kink. Then use that ˆB percentage combined with an estimate of the density of gifts, to distribute that ˆB percentage to the right of the kink, and find the upper end of the bunching interval. The larger the ˆB excess bunching at the kink and the less dense is the density of gifts, the larger must have been the price elasticity e.

26 Real world Percent A A B C (all people in B give exactly $400) D Gifts

27 Counterfactual 1 Percent A A B B C C D D Gifts (g0)

28 Estimation: Counterfactual 1 three methods price elasticity) (Compensated Nearest neighbor (Saez, 2010). Polynomial (Kleven & Waseem, 2013). Nearest round neighbor.

29 Counterfactual 1a: Nearest neighbor Percent Gifts Notes: Sample size N = Gifts were made jointly, indicating a married couple facing a limit on the tax credit at $400. Identification assumption: Without the cap, density of gifts at $400 would look like average of density at $375 and $425.

30 Counterfactual 1b: Polynomial... but what about round-number clumping at $400? Percent Gifts Notes: Sample size N = Gifts were made jointly, indicating a married couple facing a limit on the tax credit at $400. Identification assumption: Third-order polynomial and round-number dummies ($25s, $50s, $75s, $100s) capture what the density would have looked like at $400, if the cap had not been in place.

31 Counterfactual 1c: Nearest round neighbor Percent Gifts Notes: Sample size N = Gifts were made jointly, indicating a married couple facing a limit on the tax credit at $400. Identification assumption: Counterfactual density around the kink is monotonic (down or up; relative to the mass points used to bound the elasticity).

32 Estimation: Counterfactual 2 (Uncompensated price elasticity) Density Indiana Other states Identification assumption: Donors in other states can serve as the counterfactual no-tax-credit Indiana.

33 Estimation: Counterfactual 2 New method Pick a range around the kink location: Θ = [$201, $1,000]. Find the Indiana donor just below the kink at $400: percentile = 49.4 th. Find the 49.4 th percentile donor in the other states get that person s giving amount ($335). ê u based on percentage change in giving ($400 - $335) in response to the percentage change in price ( ). The more people in Indiana bunch at the kink, the lower the percentile of the person in Indiana just below the kink, the lower will be the associated amount in the other states, and the larger will be the ê u.

34 Summary: Identifying assumptions Counterfactual 1 (compensated) Counterfactual density at the kink location ($400)... would look like density near the kink.... would follow a poly-3 function with round-number dummies.... would be monotonic, relative to two surrounding mass points. Counterfactual 2 (uncompensated) Donors in control states can stand in for Indiana donors (if Indiana donors did not get the tax credit). Donors in a least one state can stand in for Indiana donors (bounds). Qualitatively different, 1 and 2. Neither rely on income for identification, in any way.

35 Match-price elasticity Donor from class of 1960 $3 million Match gifts from other members of the class of Gifts $250, months: Dec. 1, 2008 June 30, Difference-in-differences estimation The treatment (the match) turns on, but also turns off. Each other class (..., 1958, 1959, 1961,...) has a 19-month placebo match variable. Dummies: class, year, month of year. Identifying assumption: other classes giving behavior, around their 50 th anniversaries, can stand in for the class of 1960 in the counterfactual situation in which the class of 1960 had not been offered the match.

36 The data All gifts to a university in Indiana, from N = 651,747 gifts. Focus on joint-gifts married people; N = 373,994 joint gifts.

37 Donors to the institution: Indiana compared to all states. Number of joint gifts between $200 and $1,000: 7,585 (Indiana), 79,122 (All states). Average joint gifts between $200 and $1,000: $383 (Indiana), $374 (All states). Difference is at the top of the distribution: More gifts of very large size from other states.

38 Compare to the PPS. Table: 2. Educational giving and all giving in the PPS Indiana residents Residents of all states who give to who give > $1,000 education to all purposes combined Age Married Giving to education inst. $466 $237 Giving to charitable orgs. $1,615 $1,673 Giving to congregations $2,787 $2,727 Giving to all purposes $4,403 $4,401 combined

39 Result 1. Table: 3. Compensated tax-price elasticity estimates Lower bound Point estimate Upper bound estimate (linear interpolation) estimate Nearest round neighbor (.021) (.024) (.033) Nearest neighbor (.036) (.019) Polynomial (.022) Nearest round neighbor (Mass pt at $500 $600) (.021) (.025) (.035)

40 Result 1. Table: 3. Compensated tax-price elasticity estimates Lower bound Point estimate Upper bound estimate (linear interpolation) estimate Nearest round neighbor (.021) (.024) (.033) Nearest neighbor (.036) (.019) Polynomial (.022) Nearest round neighbor (Mass pt at $500 $600) (.021) (.025) (.035)

41 Result 1. Table: 3. Compensated tax-price elasticity estimates Lower bound Point estimate Upper bound estimate (linear interpolation) estimate Nearest round neighbor (.021) (.024) (.033) Nearest neighbor (.036) (.019) Polynomial (.022) Nearest round neighbor (Mass pt at $500 $600) (.021) (.025) (.035)

42 Result 1. Table: 3. Compensated tax-price elasticity estimates Lower bound Point estimate Upper bound estimate (linear interpolation) estimate Nearest round neighbor (.021) (.024) (.033) Nearest neighbor (.036) (.019) Polynomial (.022) Nearest round neighbor (Mass pt at $500 $600) (.021) (.025) (.035)

43 Result 1. Table: 3. Compensated tax-price elasticity estimates Lower bound Point estimate Upper bound estimate (linear interpolation) estimate Nearest round neighbor (.021) (.024) (.033) Nearest neighbor (.036) (.019) Polynomial (.022) Nearest round neighbor (Mass pt at $500 $600) (.021) (.025) (.035)

44 Result 1. Table: 3. Compensated tax-price elasticity estimates Lower bound Point estimate Upper bound estimate (linear interpolation) estimate Nearest round neighbor (.021) (.024) (.033) Nearest neighbor (.036) (.019) Polynomial (.022) Nearest round neighbor (Mass pt at $500 $600) (.021) (.025) (.035)

45 Result 2. Table: 6. Uncompensated tax-price elasticities: Percentile-based estimates Baseline Heterogeneity Heterogeneity Lower bound Upper bound (.042) (.082) (.012)

46 Result 3. Table: 7, 8, 9. Response to a matching grant: Baseline estimates Amount Number Amount Amount (like col. 3) (conditional) of gifts (non-donors) Post-match + donors) control (Table 7) (Table 8) (Table 9) (Table 9) (.074) (.125) (.093) (.045)

47 Result 3. Table: 7, 8, 9. Response to a matching grant: Baseline estimates Amount Number Amount Amount (like col. 3) (conditional) of gifts (non-donors) Post-match + donors) control (Table 7) (Table 8) (Table 9) (Table 9) (.074) (.125) (.093) (.045)

48 Previous literature + new results Randolph (1995) Auten, Sieg, Barrett, McGuirk, Bakija Clotfelter & Steinberg (1997) & Heim (2011) (2002) , [ ] Karlan & List (2007) Davis (2006) Scharf & Smith (2015) Eckel & Grossman (2003)

49 Previous literature + new results Randolph (1995) Auten, Sieg, Barrett, McGuirk, Bakija Clotfelter & Steinberg (1997) & Heim (2011) (2002) , [ ] Karlan & List (2007) Davis (2006) Scharf & Smith (2015) Eckel & Grossman (2003)

50 Discussion: Match-price e and tax policy Tax-price e Match-price e -.20 In this giving environment, match-price e is directly informative for tax policy. Match-price e -.20 similar to inelastic e checkbook in previous match-price literature. Suggests previous e checkbook also may be informative for tax policy.

51 Previous literature + new results + Match papers tax-price e Randolph (1995) Auten, Sieg, Barrett, McGuirk, Bakija Clotfelter & Steinberg (1997) & Heim (2011) (2002) , [ ] Karlan & List (2007) Davis (2006) Scharf & Smith (2015) Eckel & Grossman (2003) (-.11,-.34)

52 Discussion: Match-price e and tax policy (cont.) In previous studies, it is the e checkbook that is similar to tax-like rebate es. Extends checkbook result to high-stakes match and govt tax policy currently in effect. The e amt received = e checkbook 1, is not similar to tax-price e. ê amt received = 1.13 Govt could substantially more funds to educational institutions at no additional cost to itself, by replacing tax credit to donors with a match to the institutions.

53 Tax-price e and tax policy: the CBO Randolph (1995) Auten, Sieg, Barrett, McGuirk, Bakija Clotfelter & Steinberg (1997) & Heim (2011) (2002) , [ ] CBO Karlan & List (2007) Davis (2006) Scharf & Smith (2015) Eckel & Grossman (2003) (-.11,-.34)

54 Tax-price e and tax policy: French tax credit study Randolph (1995) Auten, Sieg, Barrett, McGuirk, Bakija Clotfelter & Steinberg (1997) & Heim (2011) (2002) [ Fack & Landais (2010) ] , [ ] CBO Karlan & List (2007) Davis (2006) Scharf & Smith (2015) Eckel & Grossman (2003) (-.11,-.34)

55 Discussion: Tax-price e and tax policy (cont.) More about Fack & Landais (2010) Tax reform changed price, holding income constant ê [.20,.60] Very different setting: total giving (not just education), France. Similarity to our tax-price ê suggests that inelastic responses to price, like our ê, may apply to settings beyond education in Indiana.

56 Discussion: Tax-price e and tax policy (cont.) Relevant to policies that use credits (Canada, France, New Zealand, Spain; other states in the U.S.). Proposals to cap charitable deductions at the federal level in the U.S., and thereby break tie between tax-price and income relationship (at the margin): Feldstein, Romney in Obama administration, repeatedly. Simpson-Bowles Commission. Bipartisan Policy Center s Debt Reduction Task Force would work like a match.

57 Conclusions 1. First step toward reconciling tax-price and match-price literatures. Tax-price e estimated with weaker identifying assumptions than previous literature. Parallel match-price e checkbook Both approximately the same. 2. First evidence that privately-offered match price response similar to a price response to a tax policy in effect.

58 Conclusions (cont.) 3. Match-price ê checkbook close to estimates from previous match-price literature. To date, all estimates produced using variation independent of income have produced similar, inelastic results. Our s. Previous match-price estimates Previous tax-credit-based estimates. Match-price e informative for tax policy, at least in our setting and perhaps in more. 4. Previous match-price literature, plus present results, suggest match-based credit could more funds to non-profit organizations at no additional cost to government.

59 Thanks go... To the IU Lilly Family School of Philanthropy and... The John Templeton Foundation for financial support. To you, for your participation today.