Anda di halaman 1dari 5

Key

characteris,cs of 3 papers
Milligan-Smart (M-S): es,mate behavioural responses
(reported incomes) of top earners, to compute
maximum top marginal tax rates (MTRs) using formula
Other two papers assume NO behavioural responses
Murphy-Veall-Wolfson (M-V-W): compute progressivity
of tax expenditures for top earners (standard assn)
Heisz-Murphy (H-M): decompose redistribu,ve impact of
selected tax and transfer programs into measures of
progressivity and size
The laKer two papers require inquiry into behaviour and
economic incidence for meaningful policy assessment
Findings from all papers require broader economic and
scal impact assessments for useful policy guidance

Issues in compu,ng maximum top MTR (*)


* is revenue-maximizing rate; formula ignores welfare of
top earners; likely that social op,mum rate < *
Issues in es,ma,ng elas,city () of reported incomes (or
income shares) with respect to (1 MTR)
reects both real and tax-planning/avoidance behaviour;
broaden base, ,ghten enforcement: reduce & raise *
Factors sugges,ng op,mal top rate higher than *
Factors sugges,ng op,mal top rate lower than *
Focuses on top 5%, 1%, or even 0.1%; es,mate of is
lower for high incomes below top 1%; op,mal may be
higher for that income bracket than for top bracket;
Assessed total incomes (2010): $100-149K $118 bn;
$150-250K $69 bn; $250K+ $109 bn BOTTOM LINE???

Behavioural eects: incidence maKers!


Relevant for policy inferences from M-V-W and H-M
If personal tax increase falls (mainly) on individual (rather
than employer or client), then eec,ve for redistribu,on
in short run [similarly if benet increase mainly redounds
to individual rather employer]
If personal tax increase falls (mainly) on employer or client
(shioing of incidence), then does not redistribute income
from those taxpayers, but likely redistributes high-paid
jobs, high-value industries, to other jurisdic,ons
If benet increase redounds (mainly) to employers rather
than individual workers, then redistribu,on may arise
mainly through increased employment of lower-paid
workers rather than increased net income per worker

More incidence!
Broadening taxable income base (e.g. employee stock
op,ons as in M-V-W) may reduce poten,al avoidance/
tax-planning (reduces ) but incidence s,ll falls on
employers if interna,onally compe,,ve for top
management or have tax-compensa,on programs
Issues of mobility of labour cross-provincial vs. mobility
between Canada and other countries (esp. U.S.)
Cultural issues (e.g. Quebec, Atlan,c) aect mobility
Hence, interpreta,on of all ndings in M-V-W and H-S
requires deeper analysis of incidence issues, impacts on
broader measures of impacts on total tax revenues from
all sources, secondary impacts on other (beKer-quality
but not highest-paying) industries and jobs

Decomposi,on of redistribu,ve impacts


H-M paper useful for thinking about both structure and size of
programs (both tax and transfer) in redistribu,ve impact
Kakwani progressivity index for tax (Pt), transfer (Pb), measure
departures from propor,onality; what program components?
Reynolds-Smolensky decomposi,ons of redistribu,ve impacts (R):
Taxes: Rt = Pt (t/1t) where t is average tax rate
Benets: Rb = Pb (b/1+b) where b is average benet rate
Country with a heavy avg. tax rate but less progressivity can
redistribute more than country with greater tax progressivity but
lower avg. tax rate; size of welfare state maKers more than how
progressively it is nanced; redistribu,on vs. inequality impacts
Implica,ons for scal policies, redistribu,on, inequality
Issue of how much of t goes to b vs spending on non-redistribu,ve
programs (e.g. highways, defence, research, etc.)
Universal program nanced by propor,onate rate: redistribu,ve
Policy changes: Pt & Pb as well as t & b endogenous via behav. response

Anda mungkin juga menyukai