Quantifying the impact of regulatory reform
17. Things become even more tricky when assessing the effects of changes in the regulatory stance on
growth and government budgets. The OECD has developed a broad range of indicators concerning the
stringency of labour and product market regulations, which are discussed in more detail in Duval (2005)
and feed into the econometric exercise in the next section.2 The OECD has also produced a lot of empirical
work that traces the effects of the regulatory stance on employment and unemployment rates, R&D
intensity and ultimately growth. This work has culminated in the OECD’s Growth Project (OECD, 2001)
and feeds into the ongoing structural surveillance work (OECD, 2005). Substantial methodological progress
has been achieved in constructing structural policy indicators with an econometric link to economic
performance.
18. Such inventories are potentially very useful, but there are also limitations, when assessing the
budgetary implications of structural reforms:
• These inventories leave out a wide variety regulations, such as health and safety or
environmental regulations, which could have a budgetary impact.
• There is in general no direct link between these inventories and budgetary effects and it
would seem difficult to establish the link between reforms and side-payments to get
reforms underway.
19. Moreover, while the body of empirical work on growth is enormous, there are still considerable
disagreements about what reforms can achieve in the short and long term. In particular, the short-term
adjustment costs associated with reforms are under-researched. In addition, results are data-quality, model
and estimator dependent. While cross-country growth regressions have been an extremely popular means of
testing ideas about the sources of growth, many of the variables claimed to be significant have not stood up
to tests for statistical robustness (Ahn and Hemmings, 2000). Another problem is the lack of accepted
formal theoretical models that can accommodate the wide range of variables that are often included as
explanatory variables, despite advances in the theory of economic growth. A related issue is that causal
links between aggregate economic variables and growth are bi-directional, hence most estimates are likely
to suffer from endogeneity problems. We are tempted to quote Blanchard (2005) who, after reviewing the
literature on whether enough is known about unemployment to give advice to policy makers about how to
reduce it, concludes: “I believe we do - with the proper degree of humility”.
“Big bangs” and gradualism: the experiences of New Zealand and Australia
20. Some OECD economies have undergone a radical transformation, with major fiscal implications.
New Zealand is a prime example. It initiated quite radical and wide-ranging reforms in the mid-1980s.
These reforms encompassed both macroeconomic stabilisation and structural change (Evans et al., 1996).
The reforms followed a decade of anaemic growth, which, at less than ½ per cent, was indeed considerably
lower than it is currently in the euro area, while inflation was rampant and the government and current
account deficits exceeded 6 and 8% of GDP, respectively.
21. Key reforms included financial market deregulation and the granting of operational independence
to the Reserve Bank; deep labour market reforms, though they were only enacted from the early 1990s
onwards; and telecommunication and electricity reforms. These reforms did not include any side-payments,
but there were many reforms affecting the budget directly. Support to manufacturing industries and
agriculture was withdrawn, which lowered spending, but also lowered revenues due to hefty tariff cuts on
industrial products. Direct government assistance to industry and agriculture declined form 16% of primary
government spending to just 4% in 1993/94. At the same time, tax policy put a sharp focus on the neutrality
2. The World Bank has also developed inventories of policy measures. Another example is the indicators developed by the Fondazione Rodolfo
Debenedetti.
51