Discussion of

1
Discussion ofService regulation and growth
evidence from OECD countries by Guglielmo
Barone and Federico Cingano

• Francesco Daveri
• Università di Parma and IGIER

2
Paper Motivation

• Starting point is widely held opinion Excess
  regulation is bad for growth
• And indeed economic regulation in OECD countries
  thought to have gone too far by the late 1980s or
  early 1990s
• Hence research focus on deregulation and growth,
  prompted by OECD impressive effort in data
  collection. Yet little theoretical advances in
  parallel (exception Blanchard-Giavazzi, few
  others)
• This paper is latest offspring within this strand
  of literature

3
The paper in short

• WHAT Data from 17 OECD countries and 15
  industries to carry out thought experiment
  adapting methodology of Rajan and Zingales (AER,
  1998)
• Industries most heavily dependent on services
  (energy, TLC, transports, professional services)
  likely taxed disproportionately more from
  regulation
• HOW contrast
• PMR data (entry barriers, vertical integration
  and conduct regulation from the OECD)
• Input-output data -- to evaluate dependence of
  any given downstream industry on specific
  regulated service industries
• with
• Growth of value added per employed person,
  employment and exports, from STAN, as dependent
  variables

4
Empirical specification

• VA(j,c) a ß Ss w(j,s) X(c,s) f SHARE
  µ(c) µ(j) e(c,j)
• c country j industry
• w share of service s in industry j in the US
• X value of service regulation s index in country
  c
• wX overall weight of service regulation on
  downstream industry
• SHARE initial share of industry (j,c) in total
  value added in country c
• VA growth rate of value added, 1996-2002

5
Main results

• Service regulation is bad for growth in
  downstream industries
• Effect is quantitatively sizable
• about 1pp productivity growth gain in average
  downstream industry moving from highly regulated
  to lightly regulated services
• Effect is asymmetric between services
• Stronger for energy and professional services
• Not there for transports
• Results withstand a few changes of specifications

6
General remarks

• Useful and interesting paper
• Being focused on downstream (thus indirect)
  effects of service regulation goes much beyond
  existing literature
• Three open issues to think about

7
Open issues theory

• Why a growth (as opposed to level) effect of
  regulation?
• Possible answer regulation negatively hits scale
  parameter of production function, hence BC
  estimate transitional dynamics effect
• If so, investment rates should be affected as
  well
• Implication to further test theory and learn
  more about mechanisms of transmission, investment
  as well as price deflator dynamics more sensible
  dep variables than employment or exports

8
Open issues interpreting asymmetries

• Heterogeneity in estimated coefficients suggests
  puzzle
• Why is Francesco Giavazzi right when he blames
  regulation for lawyers and notai and not when
  he speaks of taxis?
• Or Why is regulating energy and professional
  services so bad while regulating transports is
  not so bad after all?
• Are transports less useful to firms and more to
  households? So they should hit GDP and
  consumption and not investment. This is testable.

9
Open issues pairwise correlation

• Is pairwise correlation there?
• Can graphical evidence show that convergence in
  regulation (implied by figure 1) has driven to
  convergent productivity levels? Between
  countries, within industries I guess?
• Evidence on regulation and prices in Figure 2 is
  at least contentious
• If pairwise correlation is not there, then
  something else - and not regulation determines
  industry productivity trends. So why bother about
  deregulating services?

10
Some specific remarks

• Why does data set stop in 2002?
• Why I-O coefficients for the US and not country
  specific? Rationale offered is not transparent.
  Not to me at least
• Why STAN and not EUKLEMS? (I guess I know the
  answer EUKLEMS more complete, STAN more reliable)