Brewer, Saez, and Shephard,

1
Brewer, Saez, and Shephard, Optimal Household
Labor Income Tax and Transfer Programs

• Robert Moffitt
• Johns Hopkins University

2
Summary of the Paper

• Goals
• To review the UK tax and transfer system (MTRs,
  ATRs, etc)
• To review recent developments in optimal tax
  theory
• To propose a tax reform plan for the UK

3
Review of UK tax/transfer system

• Computes MTRs and ATRs including all transfer,
  tax credit, positive tax, etc programs, including
  NI
• Finds very high (100) MTRs at bottom because of
  phaseout of income support
• E.g., up to 3000 for one-earner couple with two
  children
• MTRs and ATRs fall a bit at the 16-hour/week
  introduction of the WFTC

4

• But MTRs remain high in the middle range
• E.g., over 70 for the one-earner couple, up to
  28,000 due to phaseout of the many means-tested
  tax credits, beginning of the positive tax
  system, etc.
• Above that, MTRs fall to 40 and then below 40
• See figure

5
(No Transcript)
6

• Any reasonable person would regard these as very
  high
• Exact MTRs and location of the ranges vary with
  different household structures
• Historically, top rate in the UK has fallen, but
  introduction of means-tested transfers (mainly
  refundable tax credits) has raised the MTR in the
  middle range
• Positive tax system is individual-based,
  transfer/credit system is family/joint based

7
Optimal Tax Theory

• Simple exposition of the Mirrlees model, tradeoff
  between desire for redistribution and labor
  supply disincentives
• Result that MTR as the bottom is positive can be
  reversed if allow responses at the extensive
  rather than the intensive margin (i.e., can get
  earnings subsidies)
• Assumes some labor supply elasticities,
  redistributional preferences, does some sims

8
(No Transcript)
9

• Other issues
• International migration
• Joint vs individual taxation
• Collective model
• Marriage and fertility responses
• Administrative issues (problems with the WFTC)

10
Reform Plan

• Family allowance different guarantees (G) for
  different household structures (e.g., 9000 for
  couple with two children)
• Phaseout at 45 MTR (individual, not joint,
  based)
• Integrate the positive tax system to start right
  when Family allowance ends, start at 45 MTR

11

• Pay for it by increasing other MTRs eliminate
  the 10 bracket and start with 25 increase 22
  to 25 keep 40 have a new 46 bracket and
  increase NI taxes
• Administration have all benefits paid on the
  same periodicity as the paycheck (weekly,
  monthly) update all incomes and household
  structures at same periodicity

12
Hilary Hoynes comments

• US, UK, and some other countries have shifted
  toward in-work, earnings subsidy, work
  requirement, and other programs
• Does this represent a change in the nature of the
  government's preference function? E.g., a
  preference that income be earned rather than
  unearned?
• If not,what explains this shift?

13

• Labor supply elasticities vary by marital status,
  number of children, and skill level/wage how
  does this affect the optimal program?
• Surprising that optimal system is individual
  rather than joint how much of this is due to the
  assumptions? (one-earner vs two-earner, intensive
  vs extensive,etc)
• Not enough attention paid to elasticities at the
  top
• Reform plan just looks like a reduced taper and
  some harmonization. Anything else? In-work
  programs?

14
Moffitt Comments

• Good things about the paper
• Nice, simple, intuitive explanation of the
  Mirrlees model
• (but Mirrlees (1971) was pretty good, too!)
• (but no discussion of the Mirrlees results on
  the top rate of zero, or the subsequent
  literature's discussion of that
• result)
• Willing to stick its neck out on a reform
  plan

15
Some Minor Issues

• No review of the (large) literature on optimal
  tax models since Mirrlees but maybe this is not
  the right paper for that
• Bringing in some international evidence on labor
  supply and household structure elasticities
  wouldnt hurt
• Paper is still in draft/rough form, will be
  cleaned up

16
Elasticities at the top

• Agree with Hoynes, a huge issue
• Started with Feldstein (1985) very high
  elasticities at the top, on the W margin as well
  as H margin
• Even if elasticities are half of what he
  estimated, they are big tail may wag the dog
• Obvious implications for the authors reform plan

17
The Reform Plan

• Looks like a simple NIT idea lower the MTR,
  raise MTRs on the higher up taxpayers
• This type of reform has gone out of style in the
  US
• Maybe different in UK and Europe, where ts are
  still high but so are Gs (US both Gs and ts
  are low) makes it expensive to lower t

18
Changes in Government Preferences?

• Agree with Hoynes on the question
• There is a literature on this e.g. Moffitt (EJ,
  2006) adds work per se to the government
  preference function, generates earnings subsidies
  as well as work requirements, etc
• Is a literature Besley-Coate, Blackorby, Cuff,
  et al. (some welfarist, some non-welfarist
  government preference fns)

19
Categorization

• Perhaps of some interest to note that US has cut
  the Gordian knot of how to keep high G with low t
  by simply eliminating G for many groups
• Ggt0 only for those with special needs (disabled,
  elderly, unemployed, single parents with young
  children) rest have G0
• Means you can afford an EITC because G0 for most
  of those recipients

20

• There is a model for this Akerlof tagging model
  (Diamond, too)
• Have i1,,N observable groups each has a
  different weight in the SWF give a different G
  and t to each, optimize over them allow for
  efficiency losses from incentives to change
  group allow for errors
• Akerlof get an increase in social welfare
  (reduce G on the less needy, work-able groups,
  increase it for the more needy, less-work-able)
• Inevitable that SW rises government has an
  additional tool