Fiscal Impacts of Immigration to the US

1. Fiscal Impacts of Immigration to the US Berkeley Formal Demography Workshop Friday, June 7, 2019 Gretchen Donehower UC Berkeley Demography, CEDA

2. 1997 2016

3. Fiscal impacts • Net fiscal impact • Change in taxes (revenue)? • Change in benefits (expenditures)? • Two types of impacts from immigration: 1. Economic impacts affect aggregates • Growth, innovation, labor force, wages, employment, prices, returns to capital 2. Taxes versus benefits affect impact of an immigrant • Evidence suggests that first type is positive overall • We will now focus on second type, so a lower bound for total fiscal impact

4. Taxes by Age, 2012 (but looks pretty much the same in 2017) • What does this include? • Income, property, corporate, and sales taxes • FICA taxes and other social contributions • How was it computed? • Age profile concept from National Transfer Accounts (ntaccounts.org) • Aggregate amounts from national accounts data and gov’t expenditure reports • Age shapes from surveys (CPS)

5. Benefits by Age, 2012 (similar in 2017 except S/L line is a little higher relative to Fed) • What does this include? • Targeted: programs for poor, elderly, kids, workers, etc • Non-targeted: public administration and services, national defense, interest payments, subsidies, foreign aid • How was it computed? • Aggregate amounts come from national accounts and gov’t expenditure reports • Age shapes come from surveys (Current Population Survey)

6. Net Fiscal Impacts by Age, 2012 (broadly similar in 2017) • This is benefits minus taxes • Note how important AGE is to overall fiscal impact • Kids are expensive to state/local government • Elderly are expensive to federal government • Fiscal impact of immigrants very dependent on age distribution • Not much surplus? • Nope! Overall US runs deficits - on average the government spends more on each of us than it gets from us in taxes

7. How to think about “benefits” an immigrant gets and how much they cost? • Increases cost of targeted programs - Social welfare programs, edu, etc. rise with enrollment • Increases cost of congestible public goods - Police, fire dept., libraries, etc., rise with population • Does not increase cost of pure public goods • Defense, payment of interest on past debts, subsidies… • Defense and interest are a big chunk of government spending • Results very sensitive to how you treat these expenditures

8. Fiscal Impacts by Generation • 1st generation is poorer on average but is excluded from many public programs • Some of this is compositional: more 2nd gen with highest education and earnings, more 1st generation with lowest • Some not: even within BA and BA+ education groups, 2nd generation earns more and pays more taxes For taxes, picture in 2017 is pretty close for ages 40-60, but 2nd and 3rd+ are now closer at older and younger ages. For benefits, picture in 2017 is very similar.

9. Fiscal Impacts by Generation, 2012 Note: benefits do not include defense, interest payments, or subsidies

10. Fiscal Impacts by Generation, 2017

11. Population by Generation Solid lines for 2012, dotted for 2017

12. How to summarize all this data? • Static approach • This year, is the average immigrant paying more or less in taxes than he receives in benefits? • Dynamic approach • If another immigrant arrives, will she and her descendants pay more in taxes over time than they cost the government in benefits?

13. Taxes – benefits (net fiscal impact), for 1st generation, by age and time Pure public goods not included.

14. Taxes – benefits (net fiscal impact), for 1st generation, by age and time Pure public goods not included. STATIC APPROACH: Multiply each of these net impacts by the population in each age group to find the total impact (or do this with taxes and benefits separately, compare the size of flows)

15. Static approach In year “y”, sum population-weighted taxes (T) and benefits (B) over all ages “a” and take the ratio:

16. Taxes – benefits (net fiscal impact), for 1st generation, by age and time Pure public goods not included. DYNAMIC APPROACH: For a particular age at arrival, follow an immigrant forward in time, tracking his projected taxes and benefits, and those of his descendants

17. Taxes – benefits (net fiscal impact), for 1st generation, by age and time Pure public goods not included. DYNAMIC APPROACH: For a particular age at arrival, follow an immigrant forward in time, tracking his projected taxes and benefits, and those of his descendants For dynamic approach, have to make assumptions about future probability of survival, emigration, and changing fiscal policy. Also about education, because that is so important in determining tax and benefit flows.

18. Dynamic approach For an immigrant entering the US at age “x” with education “e” in year “y” : Future net fiscal impact of education group “e” at age “a” in time “y+a-x” Probability of surviving from age x to age a Probability of not emigrating from arrival at age x to age a Discounted at rate r, (a-x) years in the future • And then, we add the descendants!

19. The Descendants! For each year, immigrant has some chance of having a child, who has some chance of surviving and not emigrating with a parent. Also need to guess at child’s future level of education, future fiscal policy, etc. Children grow through age and time.

20. Dynamic approach requires a lot of data and assumptions • Different scenarios: • Projected future taxes and benefits (2 scenarios) • Follow CBO projections for totals • Assume all flows grow at a fixed rate • Characteristics of “average” immigrant arrival (2 scenarios) • Like arrivals in last 5 years • Like current stock of 1st generation immigrants • Treatment of pure public goods (2 scenarios) • Excluded as a benefit to immigrants • Non-interest pure public goods included (defense, subsidies, rest-of-world payments) • Demographic rates stay fixed at current levels • Survivorship, fertility, emigration (probability declines with duration in US) • Projected future education • Arrivals age 25+ maintain level of arrival education • Arrivals age <25 and descendants achieve predicted level of education at age 25, based on parental education and birth region • Other assumptions • 75-year forward-looking horizon • 3% discount rate

21. Static approach results (“snapshot”) Ratio of taxes/benefits, by immigration generation, with differing treatment of public goods For this analysis, dependents age <18 are included with the parental generation.

22. Static approach results over time Ratio of taxes/benefits over time, by immigration generation, for average cost public goods scenario For this analysis, dependents age <18 are included with the parental generation. All public goods included here at average cost to all persons.

23. Dynamic approach results Net present value of an additional immigrant and her descendants, by characteristics of “average” immigrant, future budget scenario, and treatment of non-interest public goods

24. Dynamic approach results • Differ a great deal at Federal versus State/local level • Federal mostly positive, State/local negative • Scenarios generate wide range of results depending on public goods, immigrant characteristics, and future government budgets • Only the more unlikely scenarios generate large negative impacts • New immigration causes defense budgets to rise? • Debt/GDP >200% in the next 20 years? • Future immigrants have less education?

25. Which approach makes sense? • Static approach • Pro – straightforward, easy to understand • Con – only tells you about the situation today, limited policy relevance? • Dynamic approach • Pro – forward-looking analysis more relevant for policy, includes second generation dynamics • Con –complex, requires more assumptions about the future • Issues common to both approaches • Role of public goods • Health and education benefits to children are treated as a cost, not an investment

26. What to remember from all of this? • Age and education drive fiscal impacts! • The exact nature of the question drives the answer. • In nation with deficits, we are all net negative impact on government budgets; in nation paying off deficits, every taxpayer is a boon • 1st generation worse than natives on average, 2nd generation better than natives on average • Impacts very different at Federal versus State/local level

28. Discussion Issue: Baseline versus Scenarios • 1997 report had a baseline estimate of +80k plus many variations from baseline, 2016 report had… well, you saw THAT mess! • “Baseline” versus “scenario” approach • Baseline is overly confident way of putting it • But scenarios ended up with wishy washy results • Which is better? More accurate? More vulnerable to misinterpretation/manipulation? • Scenario-based estimates cherry picked • Center for Immigration Studies and similar • Presidential tweets • This American Life • Other news outlets

29. Why are we fighting about this so much right now? Source: NAS 2016, page. 36, from Haines (2006); U.S. Bureau of the Census (2018).

30. Net change in working-age population each decade, by immigrant generation (in millions) SOURCE: NAS 2016, page 51. Table 2-5 data: Pew Research Center. (2015a, September). Modern Immigration Wave Brings 59 Million to U.S., Driving Population Growth and Change Through 2065: Views of Immigration’s Impact on U.S. Society Mixed. Washington, DC. Available: http://www.pewhispanic.org/files/2015/09/2015-09-28_modern-immigration-wave_REPORT.pdf [March 2016]., Pew Research Center, 2015a.