EVIDENCE ON THE ECONOMIC EFFECTS OF TAXES

JON BAKIJAPROFESSOR OF ECONOMICS, WILLIAMS COLLEGE

JUNE 21, 2012

1

CROSS-COUNTRY EVIDENCE ON TAXES AND ECONOMIC GROWTH: REASONS

FOR SKEPTICISM • Not even close to an experiment

• Reverse causality– Business cycle fluctuations cause tax revenue to change– Demand for government increases with income– Tax administration improves with higher income

• Effect of government size on growth should differ depending on quality of government– Scandinavian countries have:

• Large progressive governments • Well-designed, well-administered, economically efficient

policies

2

• So many other factors affect economic growth– We don’t have good control variables for them all– Econometric estimates are very sensitive to reasonable

changes in the set of control variables– Controlling for a variable that is a channel through which

government affects growth makes estimates worse instead of better!

• Selection bias– If you select only countries that are rich today, you leave

out all the countries with small governments that did not grow

CROSS-COUNTRY EVIDENCE ON TAXES AND ECONOMIC GROWTH: REASONS

FOR SKEPTICISM

3

4

5

6

7

8Source: author's calculations based on data from Piketty and Saez (2003, updated in 2012 at http://elsa.berkeley.edu/~saez/TabFig2010.xls), Saez (2004), and IRS Statistics of Income Bulletin, Spring 2012.

• Skill-biased technical change

• Globalization

• “Superstar” theory

• Executive compensation issues

• Pay (e.g. stock options) tied to booming financial market asset prices

• Shifting of income from corporate to personal tax base

• Changes in tax avoidance and evasion– Costly, but can be addressed by tax reform and enforcement

IS RISING INEQUALITY EVIDENCE OF THE RICH RESPONDING TO BETTER INCENTIVES CAUSED

BY TAX CUTS? REASONS FOR SKEPTICISM

9

10

11

12

13

If entirely a real

behavioral response

to incentives

If mostly due to other

factors (Piketty, Saez, &

Stantcheva’s position)

Percentage change in

income caused by a one

percent increase in net-

of-tax share

0.44 0.2

Deadweight loss per

dollar of additional

revenue raised in top

bracket

$0.95 $0.28

Revenue-maximizing top

marginal tax rate60% 79%

IMPLICATIONS OF CROSS-COUNTRY RELATIONSHIP BETWEEN CHANGE IN TOP PERSONAL INCOME TAX RATE AND CHANGE IN TOP 1% INCOME SHARE

Source: author’s calculations based on Piketty, Saez, and Stantcheva (2012)Note: “net-of-tax-share” is one minus the marginal tax rate.

14