Berggren, Niclas; Bjørnskov, Christian

Working Paper
Regulation and government debt

IFN Working Paper, No. 1239

Provided in Cooperation with:

Research Institute of Industrial Economics (IFN), Stockholm

Suggested Citation: Berggren, Niclas; Bjørnskov, Christian (2018) : Regulation and government
debt, IFN Working Paper, No. 1239, Research Institute of Industrial Economics (IFN),
Stockholm

This Version is available at:

[Link]

Standard-Nutzungsbedingungen: Terms of use:

Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Documents in EconStor may be saved and copied for your
Zwecken und zum Privatgebrauch gespeichert und kopiert werden. personal and scholarly purposes.

Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle You are not to copy documents for public or commercial
Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich purposes, to exhibit the documents publicly, to make them
machen, vertreiben oder anderweitig nutzen. publicly available on the internet, or to distribute or otherwise
use the documents in public.
Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen
(insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, If the documents have been made available under an Open
gelten abweichend von diesen Nutzungsbedingungen die in der dort Content Licence (especially Creative Commons Licences), you
genannten Lizenz gewährten Nutzungsrechte. may exercise further usage rights as specified in the indicated
licence.
IFN Working Paper No. 1239, 2018

Regulation and Government Debt

Niclas Berggren and Christian Bjørnskov

Research Institute of Industrial Economics

P.O. Box 55665
SE-102 15 Stockholm, Sweden
info@[Link]
[Link]
Regulation and government debt*

Niclas Berggren
Research Institute of Industrial Economics (IFN), Box 55665, 102 15 Stockholm, Sweden;
Department of Economics (KEKE), University of Economics in Prague, Czechia.
[Link]@[Link]

Christian Bjørnskov
Department of Economics, Aarhus University, Fuglesangs Allé 4, 8210 Aarhus V, Denmark;
Research Institute of Industrial Economics (IFN), Sweden. chbj@[Link].

Abstract
Government debt is high in most developed countries, and while it may reflect short-term
attempts to kick-start the economy in times of crisis through fiscal stimulus, the longer-term
consequences risk being detrimental to investment and growth. This makes it important to
identify factors that are associated with debt. While previous studies have related government
debt to economic and political variables, they have not incorporated the degree to which the
economy is regulated. Using regulatory freedom (absence of detailed regulation of labor,
business and credit) from the Economic Freedom of the World index, we conduct an
empirical analysis covering up to 67 countries in the period 1975–2010. The main finding is
that regulatory freedom, especially for credit, affects debt development negatively. The effect
is more pronounced when the political system is fractionalized and characterized by strong
veto institutions, indicating policy stability and credibility, and when governments have a
right-wing ideology.

*
The authors wish to thank Niklas Potrafke and participants at the 2016 Conference of the Italian Society of
Law and Economics in Torino, especially Enrico Colombatto, the 2017 Research Workshop of the Institute for
Research in Economic and Fiscal Issues (IREF) in Manchester, the 2017 Conference of the European
Association of Law and Economics in London and the 2018 conference “The Importance of Janos Kornai’s
Research for Understanding the Changing Role of the State in the Economy” at Corvinus University in Budapest
for valuable comments and suggestions. Olga Pugatšova provided excellent research assistance. Financial
support from the Swedish Research Council (grant 2103-734, Berggren), Torsten Söderberg’s Foundation (grant
E1-14, Berggren), the Czech Science Foundation (GA ČR) (grant 16-19934S, Berggren) and the Institute for
Research in Economic and Fiscal Issues (IREF) is gratefully acknowledged.

1
Keywords
Debt, Economic freedom, Regulation, Markets, Stimulus. Keynesianism

JEL codes
E02, H63

1. Introduction

Government debt has been at the forefront of economic-policy debates in recent years, not
least as some countries have experienced large increases in indebtedness during and after the
financial crisis. It is not hard to see why such a focus has arisen. As outlined by Elmendorf
and Mankiw (1999), the debt ratio is important because it affects resource usage in an
economy and ultimately economic growth.1 While, according to many economists, increased
short-term debt is able to boost aggregate economic performance during downturns, long-term
debt of a substantial size will tend to dampen growth, e.g., by reduced national saving and
through deadweight losses from taxation needed to service the debt.2 What seems clear is that
there is a negative association between high debt and growth, although the exact magnitude of
the relationship is being discussed.3 That debt is related to growth motivates this study, which
introduces a new predictor of debt development.
We propose that the degree to which an economy is regulated matters, on the basis of
four theoretical links. The first one is that people who hold pro-market attitudes tend to
eschew regulation and hold skeptical views about substantial government debt. The second

1
Debt is also relevant when considering other long-term goals, e.g., inflation, employment, intergenerational
equity and fiscal sustainability (Auerbach, 2008).
2
Indeed, Reinhardt and Rogoff (2010) find that the debt-growth relationship is quite weak at “normal” debt
ratios but that very high debt ratios (above 90% of GDP) tend to reduce growth sharply: from 3% to 1.7% over
the two-century period they study, but even more, from 3% to –0.1% in the post-war sample. These estimates
were criticized by Herndon et al. (2014), who find smaller negative effects of high debt and no particular
threshold at 90%; but Reinhart et al. (2012) provide some further support for the previous findings. See also
Eberhardt and Presbitero (2015) for further evidence of a negative relationship but with no common threshold
across countries; cf. Égert (2015).
3
High debt levels also make it more difficult for stimulus to work: see Nickel and Tudyka (2014).

2
and third are that regulation affects the way the economy functions in a way that influences
debt, and that regulation may function as a signal to lenders regarding contemporaneous or
future problems in the economy or the government, such that they set interest rates (that
influence debt levels) accordingly. Finally, a regulated economy may be comparatively
inflexible and unable to adjust very well to changing macroeconomic circumstances, which
could leave the government with increasing debt to counteract downturns as its only
politically viable option.
Against this background, which leads us to expect a relationship between regulation and
debt, we investigate empirically if regulation is indeed related to debt and, if so, how strongly
and in what direction. As our measure of regulation, we use an area of the Economic Freedom
of the World index – regulatory freedom – which measures how little an economy is regulated
in the areas of labor, credit and business. We employ panel data from up to 67 countries with
Western-style political institutions for which we have data on government debt as well as a
set of policy indicators. We observe these countries in up to seven five-periods in which we
are able to follow the development of their debt levels and policy changes. Our main result is
that regulatory freedom is negatively related to the debt ratio: an increase in the former of one
unit (on a 10-unit scale) is on average associated with a six percentage-point reduction in the
debt ratio. Moving from the regulatory freedom of Greece to that of Denmark would, all other
things being equal, entail a reduction in the debt ratio of ten percentage points within a five-
year period. This negative relationship is primarily driven by regulatory freedom in the credit
market.
Moreover, we interact regulatory freedom with other variables in order to see whether
the relationship depends on various circumstances – and, e.g., find that regulatory freedom is
more negatively related to debt the more right-wing the government is, the stronger the veto
players, above a certain threshold, the stronger is the negative association between regulatory
freedom and debt and the higher the degree of legislative fractionalization, the more
regulatory freedom is able to restrain
This study contributes to a literature which investigates determinants of government
debt by proposing a new factor of importance. Previous studies have looked at a range of
economic, political and institutional variables as explanatory factors of which we now
mention some of the more important ones dealing with the character of government and
governance. Roubini and Sachs (1989) identify short tenure and the need to form government
coalitions as important sources of debt. Grilli et al. (1991) similarly find that governments
who expect to stay a short time in power do not care very much about the long run and

3
therefore tend to run up higher debt. Persson and Tabellini (2003) confirm that majoritarian
electoral systems, with fewer and larger parties, are associated with smaller deficits. Cheibub
(2006) finds support for presidential systems having better fiscal outcomes than parliamentary
systems, especially if the president has a strong role in the budgetary process and if he or she
can veto legislation. Volkerink and de Haan (2001) and Elgie and McMenamin (2008) show
that more fragmented governments have higher deficits, while governments with a large
majority in the legislature have lower deficits.
Furthermore, Koehler and König (2015) find that government debt in the Eurozone
would have been higher without the Growth and Stability Pact, hence indicating an effect of
this institutional setting. Leachman et al. (2007) document that both strict budget institutions
and federalism contribute to lower deficits. Gunzinger and Sturm (2016) test to what degree
political constraints affected the size of stimulus packages in connection with the great
recession. The idea is that the ease with which political decisions in general can be taken
affects the ability to undertake (debt-increasing) decisions – and it is supported by the
empirical findings.4 However, to our knowledge, tests of the hypothesis that regulation plays a
role for debt accumulation have not been undertaken before, and we thereby think we fill an
important gap in the literature.
The rest of the paper is structured as follows. In Section 2, we provide a brief theoretical
framework to structure our analysis. We then turn to the empirics, presenting the data and
empirical method in Section 3 and the results in Section 4. Concluding remarks close the
paper, in Section 5.

2. Some theoretical considerations

We specify four mechanisms that link regulation to government debt and call them 1)
reflection effects, 2) direct economic effects, 3) reputation effects and 4) political economy

4
Political business cycles could provide a further explanatory mechanism for debt accumulation, if voters
reward expansionary fiscal policy and punish more restrictive policy. However, Alesina and Passalacqua (2015:
18) write that such cycles “cannot be the main explanation for large and long lasting accumulation of public
debt”.

4
effects.5 If they emerge, they do so in a political setting, the characteristics of which we
present first.
2.1. The political process

The political process, and the factors therein that affect government debt, are illustrated in
Figure 1.6

Figure 1 about here

Beginning with the role and behavior of politicians, they adhere to political platforms
containing a number of positions: they present them during campaigns, and some get elected
with the stated purpose to implement them. In devising their policy positions, they are, on the
one hand, “internally motivated”: they have certain basic values and they have certain beliefs
about how the world works, i.e., certain ideas. The former identify the political ends and the
latter clarifies (what the politicians perceive to be) the means to achieve those ends. The ideas
can concern all kinds of issues, but we focus on the basic belief of the relative ability of
government and markets to function in a way that satisfies some conception of the public
interest. However, policy positions are not only determined by the internal workings of the
politicians – since the aim is to get elected, there is a need to pay attention to what voters and
interest groups want as well. These may, of course, desire different things, in turn. If we
assume that politicians are vote-maximizers (since no matter what they want to accomplish,
they need to be in power), this could make their policy positions the result of trade-offs on the
margin: between the politicians’ own preferred policy positions, between the policy positions
of voters, between the policy positions of necessary coalition partners and between the policy
positions of interest groups (to the extent that these go in different directions). Depending on
the weights put on these four factors, and beliefs about how the support of voters and interest
groups varies with the policy positions taken by the politicians, a platform of policy positions

5
In Section 4.4, we provide an explorative empirical analysis to see if some of them can be shown to function as
actual channels between regulation and debt.
6
This exposition concerns democracies. For non-democratic settings, voters and the legislature are not relevant
as such – authoritarian regimes often have legislatures that only have nominal power; and if there are voters,
they do not de facto have alternatives to vote for.

5
emerges.7 For example, if a politician embraces the basic value that the happiness of the
people is what matters and if he is very government-friendly, then he might be inclined to
propose to regulate corporations by consumer-protection legislation. If he thinks that a large
segment of the electorate (especially important marginal voters) will support such a position,
he is reinforced in advocating it. However, important corporate interest groups probably
oppose the position and would give resources to other politicians, which is a cost to consider.
This overall reasoning can be extended to political parties, whose politicians are united behind
a platform of policy positions after a collective process with the individual positions as inputs.
The next step is the political process. It is one thing, as an elected politician or a party,
to have a set of policy positions; it is another to see those positions become enacted in the
legislature. Policy positions are necessary but not sufficient for political decisions. The latter
are determined by governments and legislatures, whose characteristics in turn are determined
by the rules of the political game (the political institutions) and whose composition are
determined in elections. Many political decisions affect economic outcomes of various kinds,
such as debt, and some of those outcomes in turn influence debt.

2.2. Reflection effects

The first proposed link assumes that ideas (in this case regarding how well markets and
government function) affect policy and debt – illustrated by “1” in Figure 1. Keynes (1936:
383–384) famously stated:

The ideas of economists and political philosophers, both when they are right and when they are wrong,
are more powerful than is commonly understood. Indeed the world is ruled by little else.

Undoubtedly, one of the most important economists in this regard is Keynes himself, whose
ideas about the causes of and solutions to economic downturns have been powerful in
affecting the mind-sets of both policymakers and economists (Hall, 1989). In connection with

7
We here follow the general approach of Peltzman (1976). However, the literature on public choice has
documented a number of mechanisms through which politicians’ stated ideological preferences could either be
distorted or entirely reversed: Hillman (2009) and Holcombe (2016) provide compendia of such mechanisms.

6
the great recession of 2007–2009, Keynesian-type policy prescriptions were widely adopted
in many countries, often advocated by economists8, and resulted in higher government debt.
However, there have always been scholars who were skeptical of the Keynesian
prescriptions. Such skepticism arguably rests on a different assessment of the relative abilities
of markets and governments to function well. Where the Keynesians tend to see markets as
prone to malfunction and in need of government intervention, at least temporarily, the
skeptics are much more optimistic about the long-term ability of markets and more
pessimistic, due to knowledge and incentive problems, about what government can do to
improve matters (Pennington, 2011). This speaks in favor of a negative relationship between
pro-market/government-skeptical ideas and debt.
Even though our main hypothesis is one of a negative relationship between market-
friendliness and debt, it cannot be ruled out, on theoretical grounds, that the relationship is
positive. One reason could be that some market-friendly people advocate an approach to debt
called “starving the beast”, the purpose of which is to use tax cuts as a method to reduce
government expenditures (see, e.g., Mulligan, 2008). The presumption is that debt will not be
tolerated and that the size of government will therefore have to be cut. However, empirical
analysis seems to undermine this proposed mechanism (see Romer and Romer, 2009), and
hence our main hypothesis is still that market-friendliness, as proxied by the relative absence
of substantial regulation, is negatively related to debt.

2.3. Direct economic effects

The reflection effects described in Section 2.2 explain the regulation–debt relationship by
regarding regulation as an indicator of certain ideas that affect policy positions both on
regulation and debt. Here, we propose another mechanism that can explain the relationship:
direct economic effects. Regulation is taken as the starting point: for some reason – a
combination of politicians’ values and ideas, the influence of interest groups and voters and
the political decision-making process – it is there, and it does not matter for what reason. In

8
The IGM Economic Experts Panel leaned towards the Keynesian position in 2014 (IGM, 2014). When asked
whether the benefits of the U.S. stimulus will end up exceeding its costs, 20% strongly agreed, 36% agreed, 23%
were uncertain and 5% disagreed. An even stronger agreement with Keynesian ideas was found in a large survey
of economists some decades earlier (Frey et al., 1984).

7
and of itself, it can affect debt. These effects are illustrated by the number 2 in Figure 1, going
from regulation via economic outcomes to debt.
These effects mean that regulation influences how the economy works in a manner that
in turn has repercussions on the debt ratio. If the business sector is heavily regulated, this may
impede competition, dynamism and flexibility, with fewer innovations and, as a result, slower
and maybe negative economic growth.9 If growth slows down, this can affect the debt ratio
both through the numerator (through lower tax revenues, due to smaller tax bases, or higher
government expenditure, e.g., to subsidize rigid sectors and fund unemployment benefits) and
the denominator (which is GDP, the growth of which is retarded and maybe even become
negative). One particular market that tends to suffer from heavy regulation is the labor
market, with adverse employment outcomes (Skedinger, 2011). Increased unemployment
reduces the tax base and tax revenues, and it increases government expenditure through
financial support to the unemployed, thus increasing debt. However, although these negative
effects of regulation are plausible, there are some types of regulation that can have the
opposite effect, e.g., certain regulation of the financial sector. That sector can, as was evident
in large parts of the world in 2007–2009, bring about crises with very serious fiscal
consequences. Thus, it needs to function within a well-crafted set of rules (Zingales, 2015)
that, among other things, make debt build-up improbable and credibly prevent bail-out
guarantees. Still, what those optimal rules are is difficult to say (Cochrane, 2014).

2.4. Reputation effects

Reputation effects refer to how regulation is perceived by lenders. The idea is that increased
regulation is considered as a signal that influences interest rates (cf. Afonso et al., 2011;
Biglaiser and Staats, 2012). The effect could go in both directions. On the one hand,
regulation may be interpreted as a sign of problems – either in the economy or in
government’s handling of things – that market actors observe. They might react rationally by
increasing interest rates, which increases debt over time by making it more costly to roll it

9
There are a number of studies linking stricter regulation to lower adaptability and dynamism. For example,
Alesina et al. (2005) show that regulatory reform of product markets in OECD countries is associated with an
increase in investment. Djankov et al. (2006), Jalilian et al. (2007) and Justesen (2008) find that countries with
less strict regulations grow faster. Haltiwanger et al. (2014) find strong and robust evidence that stringent hiring
and firing regulations tend to reduce the pace of job reallocation. Bjørnskov (2016) shows that countries with
more regulated markets tend to experience substantially longer and deeper economic crises.

8
over and by increasing the costs of servicing it. On the other hand, regulation may be seen as
an indicator of political strength and a determination to solve problems in the economy, which
may lead lenders to decrease interest rates, to the extent they believe that the regulation will
help. This mechanism is illustrated by the number 3 in Figure 1.

2.5. Political economy effects

The fourth type of effects hinges on political considerations. If an economy is heavily

regulated, adjustments by economic actors in the face of macroeconomic downturns may not
be forthcoming, thus worsening or prolonging a crisis. 10 If politicians are committed to
keeping regulations in place, they may nevertheless feel obliged to “do something”, and fiscal
stimulus is then an instrument they can apply. While this is usually a short-term phenomenon,
we suggest that if the political logic (in accordance with Buchanan and Wagner, 1977) is such
that budget surpluses tend not to be run when the economy is doing well, this will entail long-
term debt increases as well. As shown by “4” in Figure 1, this reflects that political decisions
that increase debt are taken given strong regulation.
In summary, several mechanisms may create a link between regulatory activity and debt
development. Most, but not all, of these mechanisms entail a positive association, and may
depend on regulatory content and political circumstance.

3. Data and empirical method

3.1. The data

Our main data are total general government debt as a share of GDP, which we get from the
IMF (2016) and the World Bank (2016). Government debt includes all liabilities that require
some form of payment of interest and/or the principal by the debtor (a government) to the
creditor at some or more dates in the future. It is this not restricted to certain forms of
liabilities and both include debt to domestic as well as international creditors. We start by

10
In a very general way, Alchian (1950) describes how adaptability is a central feature for economic
development. Feldman (2009) documents how labor-market regulations tend to affect unemployment, which, in
certain situations, gives rise to an inclination to increase public spending.

9
illustrating the structure of these data by plotting the average debt rates in 2010–2014 for all
countries in our sample; the data are depicted in Figure 2. The figure illustrates the vast
diversity that stretches from a debt level of eight percent of GDP in Estonia and 11 percent in
Chile to 162 percent in Greece and 231 percent in Japan. To illustrate developments over time
of our central variables, we provide Figure OA1 in the Online Appendix, which shows the
average debt rates and the average value of regulatory freedom since 1980.

Figure 2 about here

The remaining data are drawn from a number of different sources. Our main variable of
interest is economic freedom, as measured by the Economic Freedom of the World dataset
(Gwartney et al., 2016), where we primarily use the index of regulatory freedom, which
measures the burden of regulations in credit, product and labor markets on a scale from 0 to
10; the overall index is composed of these three subindices. These data are available on a
five-year basis from 1970 for an increasing number of countries. In an extension, we apply
the four other elements of the overall economic freedom index, which also entails government
size, legal quality, monetary freedom, and the freedom to trade internationally.
We use the following control variables, based on previous studies as well as our own
theoretical considerations in the text above. We first include a lagged dependent variable,
such that all effects in the following can be interpreted as influences on the development of
debt and not associations with an equilibrium level.11 Second, we include the five-year
average growth rate of real PPP GDP as well as a count measure of the number of years
within a five-year period that growth was negative, i.e., that the country was in recession. The
recession measure captures the additional effects of problems and political pressure during
recessions and crisis. We also control for the (log to the) size of the population, as larger
countries may be better able to service a larger debt burden (as indicated by better credit
ratings; see Fuchs and Gehring, 2017).
In addition, we control for a number of political features that are potentially relevant.
We include dummies for political instability, captured by failed coups during a given period,

11
Including a lagged dependent variable is equivalent to estimating changes in debt levels during five-year
periods. Empirically, either using the change as a left-hand side variable or the level is the same as long as both
specifications include a lagged dependent. The only real difference is the estimate on the lagged dependent
variable.

10
electoral democracy, presidential political institutions (democratic as well as autocratic) and
proportional voting systems (cf. Persson and Tabellini, 2003). We also include a measure of
the strength of political veto players, which may both serve to prevent some decisions from
being made and lock in specific policy decisions (Justesen, 2008). Finally, we include two
measures of the political situation: the ideological position of the incumbent government and
the legislative fractionalization of parliament, measured as the Herfindahl-Hirschmann index.
We note that substantial fractionalization is likely to have similar effects as strong veto
institutions.
In line with the literature presented in the Introduction, our expectations for the political
variables are as follows. Presidential systems may be better at avoiding log-rolling problems
that cause larger and potentially under-funded expenditures (cf. Tullock, 1981). Conversely,
however, presidents may also have their own agenda and discretionary spending decisions,
which could cause larger expenditures and debt burdens. Proportional voting may also
contribute to such problems by creating more parliamentary instability or more fractionalized
legislatures in which larger coalitions are necessary before reaching policy decisions. Our
theoretical expectation would thus be that countries with proportional voting and
fractionalized legislatures lead to larger debt burdens while the consequences of presidential
institutions are ambiguous. Similarly, the association with government ideology is also
ambiguous as left-wing governments are ideologically more prone to larger government
spending but may also be better able to implement the necessary funding (Cukierman and
Tommasi, 1998).
Summary statistics are given in Table 1, while we provide definitions and sources in
Appendix Table A1. In Table A2 in the Appendix, the included countries are listed.

Table 1 about here

3.2. The empirical method

Our data consist of an unbalanced panel of 67 countries observed in five-year periods between
1975 and 2010. In the following, we estimate all results using OLS with panel-corrected
standard errors (Beck and Katz, 1995). We do so since it is reasonable to expect cross-section
contemporaneous error correlation due to similar shocks across developed countries and
similar developments in the international financial system. The Beck and Katz estimator
allows for such correlations while also offering standard errors that are robust to

11
heteroscedasticity in a panel setting. The estimates in the following can therefore best be
interpreted as medium- to long-run effects on debt development following policy and
institutional changes.12

4. Results

4.1. Baseline findings

We present our baseline findings in Table 2, by gradually expanding the empirical model.
While the point estimates of regulatory freedom, i.e., the absence of regulation, are relatively
small in the first three specifications, as well as weakly or not at all significant, these simple
estimates are likely to be subject to substantial omitted variables bias. In particular, much of
the previous literature has focused on the sizeable effects of institutional differences, which
we add to the model in column 4. In addition, regulatory activity is likely to differ across
types of political institutions, which further exacerbates the bias.13
The full model points at a sizeable negative and statistically significant association with
the debt ratio.14 Taking this finding as the basis, if regulatory freedom increases by one unit
on the ten-point scale, for example going from the level of Austria to that of Canada, this
comes with a lower debt ratio of six percentage points within a five-year period. This suggests

12
The results in the following are robust to applying either a standard random effects estimator or adding
country fixed effects. We do both but prefer an OLS estimator with panel-corrected standard errors, as the use of
fixed effects in particular more clearly identifies short to medium run effects. We note that our choice in the
present context produces the most conservative estimates.
13
Estimating the association between our control variables and regulation, and using the latter as the dependent
variable, illustrates the potential problems. Regulation is strongly associated with both veto player strength,
proportional voting, presidential democracy and government ideology. Not including these characteristics in the
specification must give rise to omitted variable bias in our main estimate.
14
Inspired by Gørgens et al. (2005), we have also tried non-linear modelling, but it did not produce evidence of
curvilinear effects. Including a squared term yields a worse fit and no statistical significance. Furthermore,
categorizing the regulatory freedom variable in four equal categories suggests that the effects are approximately
linear.

12
that a relatively strong market orientation, in the form of a freer, less regulated economy, is
related to lower debt as a share of GDP.15

Table 2 about here

When looking at the control variables in column 4, we see that the lagged debt ratio is
positively related to the subsequent one. By including this we can effectively be said to study
the development of the debt ratio over our five-year periods. Growth is negative for the debt
ratio, plausibly because if affects both the nominator (negatively, by indicating more
economic activity and higher tax revenues) and the denominator (positively). Recessions have
a positive effect, and larger populations are related to higher debt ratios. Turning to political
variables, having experienced a coup that failed does not seem to affect debt ratios in a
statistically significant way. Democracy, presidential systems and proportional electoral
systems are all associated with lower debt ratios than their alternatives, while veto players
(i.e., blocking features of the political institutions) are related to higher debt. On the one hand,
one could have expected the opposite sign, as strong veto players indicate that it is more
difficult to agree on various spending measures; but on the other hand, it could be that veto
players agree by granting each other favors or that veto players effectively block certain, but
not all, types of political decisions (maybe it is easier to pass tax cuts than expenditure
increases). Government ideology does not appear to matter; nor does the estimate for
legislative fractionalization attain statistical significance.16

4.2. Conditional findings

In order to get a more granular picture of the relation between market-friendliness, as

captured by regulatory freedom, and debt, we now interact our freedom variable with five
other variables: recession, ideology, veto players, fractionalization and debt ratio above 90%.
The idea is that these factors might affect the way in regulation affects debt – and maybe

15
We have also tried including four additional indicators of economic freedom, indicators that together with
regulatory freedom constitute the Economic Freedom of the World Index (Gwartney et al., 2016). As can be seen
in Table A3 in the Appendix, they are not related to the debt ratio in a statistically significant way.
16
We have also tested whether an ideological change of government within each five-year is related to debt. We
find that this dummy variable is statistically insignificant and that its inclusion does not change any of our main
findings.

13
differently in democracies, which is why we, in presenting the interaction results in Table 3,
separate results for democracies and for all countries in the sample. Throughout the table, we
report only the central estimate although we apply the full specification as in Table 2, column
4.
The addition of interaction terms also alleviates another problem. We have so far
interpreted our findings as causal, i.e., as evidence of an effect of changes to regulatory policy
on the development of government debt. However, it also remains plausible that increasing
government debt could give rise to regulatory reforms. For example, governments may react
to increasing debt levels, or increasing costs of servicing a growing debt level, by attempting
to regulate labor markets to avoid increasing public wage bills, or financial markets to directly
regulate the interest on their domestic debt. If such mechanisms are prominent, our estimates
are subject to endogeneity bias. Yet, as shown by Nizalova and Murtazashvili (2016), even in
the presence of endogeneity bias, interactions can be interpreted causally as long as one of the
interacting variables is approximately exogenous. In addition to allowing for different
reactions under different economic and political conditions, the results in Table 3 therefore
have the benefit of being partially causally interpretable. We believe this is a more viable
solution to the potential endogeneity problem than more standard solutions.17

Table 3 about here

We find that the relation of regulatory freedom to debt does not depend on whether
there is a recession or not. Turning to government ideology, the more complete picture of how
it affects the regulatory freedom-debt relationship can be found in Figure 3, where the
estimated coefficients for regulatory freedom, in its relation to debt, have been plotted against
all values of ideology.18 As can be seen, regulatory freedom is more negatively related to debt
the more right-wing the government is – if the government is sufficiently far to the left that no
center or right-wing party partakes in government, regulatory freedom is not affected in its

17
The standard solution is an IV approach, but the most obvious candidates as identified in previous studies –
ideological differences and legal origins – are also candidates for direct influences. We have been unable to find
any other theoretically valid variables that provide any real identification and therefore cannot include any IV
estimates.
18
Figures 3–5 have been calculated using the delta method (Brambor et al., 2006). They cover the full sample.

14
relationship with debt by the ideology of the government.19 In a right-wing setting, the
market-friendly/government-skeptical attitudes seem more potent in hindering debt.
Continuing with veto players, the point estimates reveal that when veto players are
anything but very weak, the effect of regulatory freedom is again negative, and more so the
stronger the veto players. This holds both in democracies and in the full sample. Looking at
Figure 4, we see the latter effect illustrated, with confidence intervals: above a certain
threshold, the more veto players there are, and thus the less likely it is that regulatory reforms
will be reversed, the stronger the negative association between regulatory freedom and debt.
This indicates that veto players appear to make the consequences of regulation more
powerful, by providing a formal-institutions constraint on decision-making, making it harder
to change policy ahead.
Next, let us look at legislative fractionalization, which to some extent proxies for log-
rolling problems as hypothesized above, but also has similar effects as strong veto
institutions. To the extent that there is a significant relationship, the more split the party
landscape in the legislature is, the higher the debt, in line with earlier studies. As can be seen
in Figure 5, where the full relationship between the regulatory freedom-debt estimates and
legislative fractionalization is shown (with relatively narrow confidence intervals),
fractionalization (above low levels) strengthens the negative impact of market-friendly
attitudes or policies, as proxied by regulatory freedom, on debt. One way to interpret this last
finding is that substantial fractionalization primarily has similar political functions as
institutionalized veto players in the sense that it makes any policy decision, including reform
reversals, less likely. However, another realistic interpretation may be that log-rolling allows
governments to agree on regulatory changes with supermajorities, such that more parties
publicly commit to any policy decision.
Lastly, we interact a dummy indicating whether the debt ratio is very high (above 90%).
Our findings indicate that this factor, which reflects a particularly dire fiscal situation, is not
related to the subsequent debt ratio; nor does it affect how regulatory freedom is related to the
debt ratio.
Although not direct evidence, we note that if the reverse causal direction would be
dominant, we would expect regulatory reactions to increasing debt levels to be stronger when
debt levels surpass some level, and when the economy is in a recession. We find no

19
We suggest that this finding can be explained by the fact that left- and right-wing governments can be
associated with different types of regulation (Potrafke, 2010; Bjørnskov and Potrafke, 2012, 2013).

15
heterogeneity in the initial debt level and the reverse pattern when we interact with a
recession dummy. As we also find rather strongly significant interactions, we take these to
imply that at least across some range of ideology, veto player strength and legislative
fractionalization, our causal interpretation fits the actual data.

Figure 3 about here

Figure 4 about here
Figure 5 about here

4.3. Sensitivity analysis

In addition to the evidence above, we have performed a number of sensitivity analyses, two of
which are reported in Table A4 in the Appendix. First, we apply country fixed effects and
report the findings in columns 1–3 (which should be compared to columns 4, 6 and 8 in Table
3). The main idea behind applying fixed effects is to effectively control for the potential
influence of approximately time-invariant factors such as, e.g., culture, long-term debt history
and constitutional influences. These estimates are therefore arguably more robust to omitted
variables bias and econometric reflection problems. When comparing the fixed effect
estimates in Table A4 with the corresponding ones in Table 3, we nonetheless find that they
are all in general larger, but the overall pattern is very much the same (in columns 6 and 8 of
Table 3 and columns 2 and 3 in Table A4).20 Yet, we remain agnostic of which results are
closer to a “true” effect, as the fixed effects estimates are more likely to capture pure short- to
medium-run effects than those in previous tables.
Second, we exclude countries with less than ten years of democratic experience to avoid
specific cases of transition – not least the post-communist transition in Central and Eastern
Europe – that can influence the results. These results are reported in columns 4–5 of Table A4
(which should be compared to columns 6 and 8 in Table 3). Reassuringly, we obtain
quantitatively similar findings for regulatory freedom as well as the interactions. Excluding
particular countries with very specific history or development, such as Argentina, Chile or
Singapore, or single time periods (not shown), likewise has no consequence for our results,
nor does extending the period of having been a democracy to 20 years.

20
However, the ideology interaction loses significance in the fixed effects estimates.

16
 Third, we have excluded observations with the 10% highest and 10% lowest debt levels,
in order to ensure that our results are not driven by observations with uncharacteristically high
or low debt burdens. If so, the results could in principle be robust, but the findings would not
generalize to most countries or situations. We nevertheless find only very small and
insignificant changes in our estimates of regulatory impact. Similarly, we have excluded
observations with the 10% highest and 10% lowest regulation scores, in order to ascertain that
the finding also apply to variation within a “normal” regulatory scope. We again find
quantitatively very small and insignificant changes in our main estimates, which all remain
significant at conventional levels. These results are available on request.
Fourth, we have so far included the lagged dependent variable among the explanatory
variables in order to specify a dynamic model. However, there is a potential risk of bias
stemming from correlation between the lagged variable and the idiosyncratic error term. In
order to investigate this possible problem, we have run a regression in which we remove the
lagged variable from the regression of Table 2, column 4, with results being reported in Table
4, column 1. As can be seen, we find that the main results are relatively unchanged, which
should alleviate concerns for serious bias stemming from the lagged dependent variable.
We conclude that our main results are robust to the most intuitively important
robustness tests.

4.4. Exploring transmission mechanisms

While we have established a clear and robust connection between regulatory freedom and
debt development, which type of transmission mechanism that is more likely remains an open
question. In Table 4, column 2, we therefore repeat the main estimates from Table 2, column
4, with a reduced sample for which all three sub-indices are available, and subsequently
explore the potential effects of freedom from credit, labor and business regulation within the
same sample.21

Table 4 about here

21
This is a way of recognizing that the regulatory environment is a complex one, and that any aggregation is
bound to hide a great deal of heterogeneity, e.g., in terms of what the effects on some outcome variable are. The
point estimates we produce are to be seen as “net effects” of all the underlying variables of a particular index,
where both sizes and signs can vary between those variables.

17
First, we present a comparison between the original and restricted samples in columns 2 and
3, which reveals a slightly, but not significantly, larger point estimate of regulation in the
restricted sample, such that results can be compared although the restricted sample is about 25
percent smaller. Second, separating the three sub-indices in columns 4–6 reveals that the main
result is driven almost entirely by freedom from credit regulation: neither regulatory freedom
for labor nor for business attain significance, and the sign of the association even changes for
the former. The main effects in Tables 2 and 3 thus seem to be driven by regulatory activity
specific to credit markets and the financial sector.
In further, very tentative tests (available in an Online Appendix), we explore three of
the four channels through which regulatory freedom could affect debt, as outlined in our
theoretical section. We have only been able to do this with cross-sectional data for quite
limited samples; and we should stress that the specific indicators used are not necessarily
clear-cut for testing the channels (and for two of the channels, we have been unable to
identify indicators to conduct any test). Still, to test reflection effects as a channel whereby
regulatory freedom and debt development are both the result of attitudes towards market and
government, we explore the cross-sectional association between debt development and
measures of confidence in major companies, people’s preferences for government
responsibility versus the use of markets and of whether people think that market competition
is a force for good or bad.22 Then, to test direct economic effects, we use the Grant Thornton
(2018) global dynamism index to see whether the flexibility of an economy with regulatory
freedom relates to debt development. Lastly, to test reputation effects, we use the association
between debt development and credit ratings from Standard and Poor’s. Throughout, we look
at the association of these indicators with overall regulatory freedom and freedom from credit
regulation, and with debt development.
We first find that countries’ credit ratings are positively and significantly associated
with overall regulatory freedom, and the specific measure of freedom from credit regulation is
also associated with confidence in major companies. These results indicate that reflection
effects and reputation effects may explain the link between the regulatory measures and debt
development. However, we also find that their inclusion does not change the cross-sectional

22
As in Aghion et al. (2010) and Pitlik and Kouba (2015), these measures all derive from the most recent waves
of the World Values Survey and the European Values Study. In the case a country is included in both surveys,
we take the average.

18
correlation between regulation and debt development, which we would have expected if these
were channels at work. We do not observe any indications of such changes and therefore do
not find direct evidence of either reflection, reputation or direct economic effects, although
we cannot rule them out with this limited analysis. Nor can we, of course, rule out political-
economy effects, since we could not provide any test of this potential channel.
Hence, while we can state with some certainty that the development of sovereign debt is
strongly and significantly related to the specific regulation of credit markets, no survey with
sufficient coverage across countries and over time provides information that would allow us
to conclusively test the mechanisms at work. Clearly, more research should be undertaken
regarding mechanisms as panel data become available.

5. Concluding remarks

Government debt is of classic concern to policymakers and economists alike, not least
because of the perceived benefits and costs associated with it. Whether the benefits dominate
the costs, or vice versa, is a matter of contention, but it is quite clear that many are in favor of
short-term deficits in order to combat recessions – often supposing both that such stimulus
works and that the political incentives are such as to run subsequent budget surpluses in better
times. Others are less prone to accept Keynesian stabilization policy, on the basis of a
skepticism regarding the ability and incentives of government to use its fiscal tools in a way
that produces better outcomes than if markets are left alone.
In this study, we focus on a particular type of policy, regulation, and its effects on
government debt, a relation that to our knowledge has not been analyzed before. Arguably,
the degree to which an economy is regulated stems from ideas regarding how comparatively
well markets and government work, just like in the case of debt. One possible link between
regulation and debt is therefore that of pro-market/government-skeptical ideas that cause both
little regulation and high debt. But there are at least three other potential links: direct
economic effects (such as regulation affecting growth and unemployment in a way that in turn
influences the debt), reputation effects (when lenders regard regulation as a signal of either
economic problems or as solutions to economic problems and set interest rates accordingly)
and political economy effects (a regulated economy may induce politicians to stimulate it
since it does not adapt easily on its own).

19
 Our main finding is that regulatory freedom – especially in credit markets – is
negatively related to the debt ratio. That is to say, stricter regulation of credit tends to go hand
in hand with higher indebtedness. The effect of changes to regulatory policy may be sizeable.
Increasing regulatory freedom by a standard deviation – approximately 1.5 points on the ten-
point scale or going from the level of Germany (6.6 in 2010) to that of Canada (8.3 in 2010) –
is associated with a lower debt ratio of nine percentage points within a five-year period. In
other words, such a regulatory change on average induces a 1.8 percentage point reduction of
the debt level every year within a five-year period.
In a more detailed analysis, where we interact regulatory freedom with various features
of the political setting, we find that ideology tends to matter such that regulatory freedom is
more negatively related to debt the more right-wing the government is (in democracies), that
the stronger veto players are, i.e., the more difficult it is to agree on decisions, the stronger the
negative association between regulatory freedom and debt, and that the more fractionalized
the legislature is, the more regulatory freedom is able to restrain debt.
When interpreting the negative finding between regulatory freedom and government
debt, one possibility is that policymakers who adhere to a market-friendly/government-
skeptical type of idea tend not only to not regulate the economy but also to refrain from
increasing indebtedness. They reject interventions on the micro level (regulation) and on the
macro level (debt-increasing stimulus or many of the large expenditure programs). But the
result is also consistent with regulatory freedom entailing a well-functioning economy that is
associated with lower debt, with regulatory freedom signaling a well-functioning economy
that merits low interest rates from lenders and with regulatory freedom enabling political
decision-makers to use other methods than debt to stimulate the economy in downturns.
However, an exploratory empirical analysis did not generate clear support for any particular
mechanism, which points at the need for more granular work on this matter as more panel
data become available in the future.

Appendix

Table A1 about here

Table A2 about here
Table A3 about here
Table A4 about here

20
References

Afonso, António, Pedro Gomes and Philipp Rother. 2011. Short- and long-run determinants of
sovereign debt credit ratings. International Journal of Finance and Economics, 16(1), 1–15.
Aghion, Philippe, Yann Algan, Pierre Cahuc and Andrei Shleifer. 2010. Regulations and distrust.
Quarterly Journal of Economics, 125(3), 1015–1049.
Alchian, Armen A. 1950. Uncertainty, evolution, and economic theory. Journal of Political Economy,
58(3), 211–221.
Alesina, Alberto, Silvia Ardagna, Giuseppe Nicoletti and Fabio Schiantarelli. 2005. Regulation and
investment. Journal of the European Economic Association, 3(4), 791–825.
Alesina, Alberto and Andrea Passalacqua. 2015. The political economy of government debt. NBER
working paper 21821.
Auerbach, Alan J. 2008. Long-term objectives for government debt. Report 2008/1, Swedish Fiscal
Policy Council, Stockholm.
Beck Nathaniel and Jonathan N. Katz. 1995. Nuisance vs. substance: specifying and estimating time-
series cross section models. Political Analysis, 6(1), 1–36.
Beck, Thorsten, George Clarke, Alberto Groff, Philip Keefer and Patrick Walsh. 2001. New tools and
new tests in comparative political economy: the database of political institutions. World Bank
Economic Review, 15(1), 165–176.
Biglaiser, Glen and Joseph L. Staats. 2012. Finding the 'democratic advantage' in sovereign bond
ratings: The importance of strong courts, property rights protection and the rule of law.
International Organization, 66(3), 515–535.
Bjørnskov, Christian. 2015. Institutional shocks and economic crises. Paper presented at the annual
meetings of the Southern Economic Association, New Orleans, November 20–23, 2015.
Bjørnskov, Christian. 2016. Economic freedom and economic crises. European Journal of Political
Economy, 45(December), S11–S23.
Bjørnskov, Christian and Niklas Potrafke. 2012. Political ideology and economic freedom across
Canadian provinces. Eastern Economic Journal, 38(2), 143–166.
Bjørnskov, Christian and Niklas Potrafke. 2013. The size and scope of government in the US states:
does party ideology matter? International Tax and Public Finance, 20(4), 687–714.
Bjørnskov, Christian and Martin Rode. 2018. Regime types and regime change: a new dataset.
Working paper, Aarhus University. Data available at
[Link]

21
Brambor, Thomas, William Roberts Clark and Matt Golder. 2006. Understanding interaction models:
Improving empirical analysis. Political Analysis, 14(1), 63–82.
Buchanan, James M. and Richard E. Wagner. 1977. Democracy in Deficit: The Political Legacy of
Lord Keynes.
Cochrane, John H. 2014. Challenges for cost-benefit analysis of financial regulation. Journal of Legal
Studies, 43(S2), S63–S105.
Cukierman, Alex and Mariano Tommasi. 1998. When does it take a Nixon to go to China? American
Economic Review, 88(1), 180–197.
Djankov, Simeon, Caralee McLiesh and Rita Maria Ramalho. 2006. Regulation and growth.
Economics Letters, 92(3), 395–401.
Eberhardt, Markus and Andrea F. Presbitero. 2015. Public debt and growth: Heterogeneity and non-
linearity. Journal of International Economics, 97(1), 45–58.
Égert, Balázs. 2015. Public debt, economic growth and nonlinear effects: Myth or reality? Journal of
Macroeconomics, 43(March), 226–238.
Elmendorf, Douglas W. and N. Gregory Mankiw. 1999. Government debt. In: Taylor, John B. and
Michael Woodford (Eds.), Handbook of Macroeconomics Vol. 1C, Amsterdam: Elsevier, 1615–
1669.
Feldmann, Horst (2009). The unemployment effects of labor regulation around the world. Journal of
Comparative Economics, 37(1), 76–90.
Frey, Bruno S., Werner W. Pommerehne, Friedrich Schneider and Guy Gilbert. 1984. Consensus and
Dissension among Economists: An Empirical Inquiry. American Economic Review, 74(5), 986–
994.
Fuchs, Andreas and Kai Gehring. 2017. The home bias in sovereign ratings. Journal of the European
Economic Association, 15(6), 1386–1423.
Gørgens, Tue, Martin Paldam and Allan Würtz. 2005. Growth, income and regulation: a non-linear
approach. Working paper 2015-12, Centre for Applied Microeconometrics, Department of
Economics, University of Copenhagen.
Grant Thornton. 2018. Grant Thornton Global Dynamism Index. Chicago: Grant Thornton.
Grilli, Vittorio, Donato Masciandaro, Guido Tabellini, Edmond Malinvaud and Marco Pagano. 1991.
Political and monetary institutions and public financial policies in the industrial countries.
Economic Policy, 6(13), 342–392.
Gunzinger, Fabian and Jan-Egbert Sturm. 2016. It’s politics, stupid! Political constraints determined
governments’ reactions to the great recession. Kyklos, 69(4): 584–603.
Gwartney, James D., Robert A. Lawson and Joshua C. Hall. 2016. Economic Freedom of the World:
2016 Annual Report. Vancouver, BC: The Fraser Institute.
Hall, Peter A. (Ed.). 1989. The Political Power of Economic Ideas: Keynesianism across Nations.
Princeton: Princeton University Press.

22
Haltiwanger, John, Stefano Scarpetta and Helena Schweiger. 2014. Cross country differences in job
reallocation: The role of industry, firm size and regulations. Labour Economics, 26(January),
11–25.
Henisz, Witold. 2002. The institutional environment for infrastructure investment. Industrial and
Corporate Change, 11(2), 355-389.
Herndon, Thomas, Michael Ash and Robert Pollin. 2014. Does high public debt consistently stifle
economic growth? A critique of Reinhart and Rogoff. Cambridge Journal of Economics, 38(2),
257–279.
Hillman, Arye L. 2009. Public Finance and Public Policy. Cambridge: Cambridge University Press.
Holcombe, Randall G. 2016. Advanced Introduction to Public Choice. London: Edward Elgar.
IGM. 2014. Economic stimulus (revisited). [Link]
panel/poll-results?SurveyID=SV_5bfARfqluG9VYrP
Jalilian, Hossein, Colin Kirkpatrick and David Parker. 2007. The impact of regulation on economic
growth in developing countries: A cross-country analysis. World Development, 35(1), 87–103.
Justesen, Mogens Kamp. 2008. The effect of economic freedom on growth revisited: New evidence on
causality from a panel of countries 1970–1999. European Journal of Political Economy, 24(3),
642–660.
Justesen, Mogens Kamp. 2014. Better safe than sorry: how property rights and veto players jointly
affect economic growth. Comparative Politics, 46(2), 147–167.
Keynes, John Maynard. 1936. The General Theory of Employment, Interest and Money. London:
Palgrave Macmillan.
Koehler, Sebastian and Thomas König. 2015. Fiscal governance in the Eurozone: How effectively
does the Stability and Growth Pact limit governmental debt in the Euro countries? Political
Science Research and Methods, 3(2), 329–351.
Leachman, Lori L, Guillermo Rosas, Peter Lange and Alan Bester. 2007. The political economy of
budget deficits. Economics & Politics, 19(3), 369–420.
Mulligan, Casey. 2008. Professor Krugman is right: “Now is not the time to worry about the deficit”.
Blog post, [Link]
[Link] (accessed 10.11.2016).
Nickel, Christiane and Andreas Tudyka. 2014. Fiscal stimulus in times of high debt: Reconsidering
multipliers and twin deficits. Journal of Money, Credit and Banking, 46(7), 1313–1344.
Nizalova, Olena Y. and Irina Murtazashvili. 2016. Exogenous treatment and endogenous factors:
vanishing of omitted variable bias on the interaction term. Journal of Econometric Methods,
5(1), 71–77.
Peltzman, Sam. 1976. Toward a more general theory of regulation. Journal of Law and Economics,
19(2), 211–240.

23
Pennington, Mark. 2011. Robust Political Economy: Classical Liberalism and the Future of Public
Policy. Cheltenham: Edward Elgar.
Persson, Torsten and Guido Tabellini. 2003. The Economic Effects of Constitutions. Cambridge, MA:
The MIT Press.
Pinkovskiy, Maxim and Xavier Sala-I-Martin. 2016. New need not be better: Evaluating the Penn
World Tables and the World Development Indicators using nighttime lights. NBER working
paper 22216.
Pitlik, Hans and Luděk Kouba. 2015. Does social distrust always lead to a stronger support for
government intervention? Public Choice, 163(3), 355–377.
Potrafke, Niklas. 2010. Does government ideology influence deregulation of product markets?
Empirical evidence from OECD countries. Public Choice, 143(1–2), 135–155.
Reinhart, Carmen M. and Kenneth S. Rogoff. 2010. Growth in a time of debt. American Economic
Review, 100(2): 573–578.
Reinhart, Carmen M., Vincent R. Reinhart and Kenneth S. Rogoff. 2012. Public debt overhangs:
Advanced-economy episodes since 1800. Journal of Economic Perspectives, 26(3), 69–86.
Romer, Christina D. and David H. Romer. 2009. Do tax cuts starve the beast? The effect of tax
changes on government spending. Brookings Papers on Economic Activity, 40(1), 139–214.
Roubini, Nouriel and Jeffrey Sachs. 1989. Political and economic determinants of budget deficits in
the industrial economies. European Economic Review, 33(5), 903–938.
Skedinger, Per. 2011. Consequences of employment protection legislation. Nordic Economic Policy
Review, 2(1), 45–83.
Tullock, Gordon. 1981. Why so much stability? Public Choice, 37(2), 189–204.
Volkerink, Bjørn and Jakob de Haan. 2001. Fragmented government effects on fiscal policy: New
evidence. Public Choice, 109(3–4), 221–242.
Zingales, Luigi. 2015. Does finance benefit society? Journal of Finance, 70(4), 1327–1363.

24
Tables and figures

Table 1. Descriptive statistics

Variable Mean Standard deviation Observations
Government debt 51.238 32.998 329
Growth 1.807 2.586 443
Recession 1.174 1.195 476
Log population 15.783 1.581 533
Failed coup .011 .059 519
Democracy .797 .375 544
Presidential .411 .493 544
Proportional .751 .433 498
Veto players .366 .164 515
Government ideology .170 .437 495
Legislative fractionalization .161 .123 495
Debt over 90 % .131 .337 216
Size of government 5.706 1.586 482
Legal quality 6.236 1.853 461
Sound money 7.425 2.344 493
Freedom to trade 7.054 1.957 475
Regulatory freedom 6.321 1.313 463

25
Table 2. Debt-ratio predictors
1 2 3 4
Lagged dependent .655*** .613*** .604*** .573***
(.081) (.076) (.076) (.074)
Growth -2.399*** -2.988*** -2.805***
(.776) (.828) (.799)
Recession 4.889*** 4.787*** 4.475***
(1.401) (1.399) (1.321)
Log population .564 .529 1.220*
(.686) (.688) (.742)
Failed coup -24.010
(31.384)
Democracy -13.844*
(8.136)
Presidential -8.959***
(2.867)
Proportional -9.399***
(3.305)
Veto players 23.852**
(9.959)
Government -4.223 -1.311
ideology (3.456) (3.299)
Legislative -8.551 3.529
fractionalization (11.062) (11.156)
Regulatory freedom -2.974** -2.109* -1.453 -5.091***
(1.268) (1.208) (1.322) (1.509)
Regional FE Yes Yes Yes Yes
Observations 258 256 255 255
Countries 68 67 67 67
R squared .605 .669 .675 .703
Wald Chi sq. 256.51 351.06 363.89 500.82
Note: *** (**) [*] denote significance at p<.01 (p<.05) [p<.10]. All regressions include a constant term. Regions
include Asia, Latin America and the Caribbean, and the formerly communist countries in Central and Eastern
Europe.

26
Table 3. Conditional findings
1 2 3 4 5 6 7 8 9 10
All Democracies All Democracies All Democracies All Democracies All Democracies
Lagged .569*** .566*** .572*** .569*** .575*** .571*** .565*** .559*** .428*** .417***
dependent (.073) (.074) (.074) (.075) (.074) (.074) (.073) (.074) (.063) (.062)
Recession -9.294 -8.664 4.533*** 4.435*** 4.468*** 4.366*** 4.726*** 4.677*** 3.448*** 3.517***
(8.700) (9.062) (1.318) (1.325) (1.313) (1.315) (1.293) (1.293) (1.130) (1.133)
Veto players 24.221** 23.473** 22.694** 22.249** 130.112* 193.653** 21.399** 16.007 10.442 7.774
(9.943) (10.315) (9.929) (10.187) (74.606) (81.922) (9.808) (9.888) (8.057) (8.273)
Government -1.375 -.139 16.475 23.204 -2.121 -.672 -.028 1.294 .326 1.785
ideology (3.243) (3.429) (15.353) (15.321) (3.319) (3.406) (3.235) (3.394) (2.723) (2.858)
Legislative 2.616 .056 4.809 2.021 2.479 -2.389 257.477*** 296.565*** 2.126 -.387
fractionalization (11.029) (11.369) (11.229) (11.468) (11.174) (11.407) (72.037) (76.495) (9.392) (9.516)
Debt over 90 % 28.084 10.473
(18.820) (19.139)
Regulatory -7.494*** -7.264*** -4.376*** -4.049*** .869 4.771 2.423 3.828 -3.782*** -4.033***
freedom (1.983) (2.037) (1.564) (1.543) (4.405) (4.889) (2.639) (2.778) (1.328) (1.331)
Reg. * 2.045 1.940
Recession (1.282) (1.341)
Reg. * ideology -2.643 -3.397
(2.251) (2.229)
Reg. * veto -14.887 -24.194**
players (10.458) (11.602)
Reg. * -37.182*** -43.369***
fractionalization (10.385) (11.009)

27
 Reg. * debt 1.294 4.216
over 90 % (2.709) (2.770)
Regional FE Yes Yes . Yes Yes Yes Yes Yes Yes Yes Yes
Observations 255 247 255 247 255 247 255 247 255 247
Countries 67 64 67 64 67 64 67 64 67 64
R squared .706 .700 .704 .699 .705 .703 .716 .714 .793 .795
Wald Chi sq. 505.81 437.26 501.30 429.09 522.94 459.64 548.69 493.77 934.08 873.12
Note: *** (**) [*] denote significance at p<.01 (p<.05) [p<.10]. All regressions include a constant term and the full specification from Table 2, column 4. Regions include
Asia, Latin America and the Caribbean, and the formerly communist countries in Central and Eastern Europe.

28
Table 4. Debt-ratio predictors with specific regulatory indicators
1 2 3 4 5 6
Lagged .573*** .648*** .643*** .653*** .642***
dependent (.074) (.090) (.086) (.092) (.094)
Growth -2.778*** -2.805*** -3.147*** -3.605*** -3.282*** -3.294***
(1.077) (.799) (.745) (.676) (.785) (.757)
Recession 4.842** 4.475*** 5.691*** 3.046** 6.343*** 6.290***
(2.026) (1.321) (1.389) (1.351) (1.439) (1.421)
Log population 2.539** 1.220* 1.319 1.439* 1.886* 1.695*
(.999) (.742) (.939) (.872) (.972) (.945)
Failed coup -55.825 -24.010 -51.231 -6.264 -34.242 -45.802
(36.895) (31.384) (39.042) (34.917) (28.893) (33.976)
Democracy -31.398*** -13.844* -10.293 -12.188 -6.043 -8.115
(10.586) (8.136) (9.654) (9.739) (9.417) (10.157)
Presidential -16.529*** -8.959*** -8.811*** -10.946*** -7.700** -8.493**
(4.015) (2.867) (3.143) (3.055) (3.275) (3.447)
Proportional -4.536 -9.399*** -10.206*** -8.039*** -2.705 -4.667
(4.792) (3.305) (3.330) (2.801) (3.799) (3.063)
Veto players 37.883*** 23.852** 28.953*** 30.561*** 25.104** 26.541**
(14.281) (9.959) (11.123) (10.639) (11.062) (11.455)
Government 1.432 -1.311 3.711 2.865 -.092 1.459
ideology (4.271) (3.299) (3.499) (3.178) (3.571) (3.645)
Legislative -4.099 3.529 -6.335 -1.342 -9.778 -9.846
fractionalization (16.259) (11.156) (11.364) (10.779) (11.879) (11.930)
Regulatory -4.629** -5.091*** -5.823***
freedom (1.962) (1.509) (1.664)
Credit -6.845***
regulation (1.152)
Labor .616
regulation (1.192)
Business -1.513
regulation (1.464)
Regional FE Yes Yes Yes Yes Yes Yes
Observations 305 255 196 196 196 196
Countries 67 67 67 67 67 67
R squared .285 .703 .785 .809 .772 .773
Wald Chi sq. 129.95 500.82 597.04 728.48 569.80 556.36
Note: *** (**) [*] denote significance at p<.01 (p<.05) [p<.10]. All regressions include a constant term.
Regions include Asia, Latin America and the Caribbean, and the formerly communist countries in Central and
Eastern Europe.

29
30
Table A1. Definitions of variables
Variable Definition and measurement
Government debt Full government debt, at central government and lower levels, as a share of
GDP; from IMF (2016) and World Bank (2016)
Growth Five-year average percentage growth rate of real PPP GDP per capita, based on
Heston et al. (2012)
Recession Number of years within a five-year period that yearly growth is negative
Log population The logarithm to the size of population at the beginning of a five-year period
Failed coup The share of years within a five-year period in which a failed coup occurred
Democracy Dummy for the existence of competitive electoral democracy from Bjørnskov
and Rode (2018)
Presidential Dummy for the presidential political system from Bjørnskov and Rode (2018)
Proportional Dummy for proportional voting system from last update of Beck et al. (2001)
Veto players Index of veto player strength from Henisz (2001)
Government ideology Five-step index between -1 (communist) and 1 (classical liberal) of party
ideology; in coalition governments, weights are parties’ seats in parliament;
from Bjørnskov (2015)
Legislative Herfindahl-Hirschmann index of (lower house) legislature; based on data in
fractionalization Bjørnskov (2015)
Debt over x % Dummy for whether end-of-period government debt is above x % of GDP
Size of government Index of the size of government, measured on scale from 0 (maximum
government) to 10 (minimum government), from Gwartney et al. (2016)
Legal quality Index of quality and independence of legal system, from Gwartney et al.
(2016)
Sound money Index of sound money – low and stable inflation and the freedom to hold bank
accounts in foreign currency, from Gwartney et al. (2016)
Freedom to trade Index of the freedom to trade and invest internationally, from Gwartney et al.
(2016)
Regulatory freedom Index of regulatory burden in credit, labor and product markets, from
Gwartney et al. (2016)
Table A2. The countries included in the study
Country Country Country
Albania Germany Norway
Argentina Greece Panama
Australia Guatemala Paraguay
Austria Guyana Peru
Belgium Honduras Poland
Belize Hungary Portugal
Bolivia Iceland Romania

31
Brazil Ireland Serbia
Bulgaria Israel Singapore
Canada Italy Slovakia
Chile Jamaica Slovenia
Colombia Japan Spain
Costa Rica Korea Suriname
Croatia Latvia Sweden
Cyprus Lithuania Switzerland
Czech Republic Luxembourg Trinidad and Tobago
Denmark Macedonia Turkey
Dominican Republic Malta United Kingdom
Ecuador Mexico United States
El Salvador Montenegro Uruguay
Estonia The Netherlands Venezuela
Finland New Zealand
France Nicaragua

32
Table A3. Debt-ratio predictors, including all five areas of the Economic Freedom Index
1 2 3 4 5
Lagged dependent .579*** .566*** .581*** .576*** .573***
(.075) (.076) (.071) (.075) (.074)
Growth -2.822*** -3.086*** -3.098*** -2.946*** -2.805***
(.836) (.820) (.818) (.811) (.799)
Recession 5.189*** 4.919*** 5.733*** 4.869*** 4.475***
(1.350) (1.369) (1.407) (1.349) (1.321)
Log population 1.341* 1.364* 1.109 1.678** 1.220
(.776) (.796) (.793) (.780) (.742)
Failed coup -15.383 -24.368 -15.399 -17.639 -24.010
(28.239) (31.545) (29.649) (28.461) (31.384)
Democracy -9.337 -11.478 -8.676 -12.268 -13.844*
(8.021) (8.337) (7.922) (8.327) (8.136)
Presidential -6.849** -9.179*** -7.164** -9.223*** -8.959***
(3.305) (3.068) (2.971) (3.129) (2.867)
Proportional -4.269 -3.950 -3.047 -3.248 -9.399***
(2.907) (2.747) (2.729) (2.692) (3.305)
Veto players 19.135* 22.405** 15.608 21.389** 23.852**
(9.912) (10.206) (9.708) (10.121) (9.959)
Government -4.062 -3.892 -6.723** -3.734 -1.311
ideology (3.272) (3.291) (3.301) (3.298) (3.299)
Legislative .228 .893 -.126 1.183 3.529
fractionalization (11.603) (11.659) (11.872) (11.545) (11.156)
Size of -1.049
government (1.071)
Legal quality -1.737
(1.414)
Sound money 2.197*
(1.138)
Freedom to trade -1.477
(1.191)
Regulatory -5.091***
freedom (1.509)
Regional FE Yes Yes . Yes Yes Yes
Observations 255 255 255 255 255
Countries 67 67 67 67 67
R squared .692 .692 .696 .692 .703
Wald Chi sq. 475.74 481.11 484.10 471.05 500.82

33
Note: *** (**) [*] denote significance at p<.01 (p<.05) [p<.10]. All regressions include a constant term.
Regions include Asia, Latin America and the Caribbean, and the formerly communist countries in Central and
Eastern Europe.

34
Table A4. Additional tests
1 2 3 4 5
Fixed effects Fixed effects Fixed effects OLS OLS
Democracies Democracies Democracies Stable Stable
democracy democracy
Lagged dependent .271*** .275*** .267*** .655*** .648***
(.047) (.046) (.045) (.068) (.068)
Veto players 15.441 178.547** 9.678 231.366*** 11.279
(15.847) (70.437) (15.221) (78.151) (9.416)
Legislative 2.837 -1.300 357.137*** -8.968 253.203***
fractionalization (3.891) (22.569) (89.143) (12.112) (79.291)
Regulatory freedom -11.386*** -1.457 .181 6.815 2.344
(2.249) (4.733) (3.568) (4.742) (2.652)
Reg. * veto players -23.712** -29.882***
(9.985) (11.001)
Reg. * -53.236*** -37.459***
fractionalization (13.102) (11.299)
Regional FE Yes Yes Yes Yes Yes
Observations 247 247 247 235 235
Countries 64 64 64 63 63
R squared .608 .621 .645 .765 .758
Wald Chi / F 15.13 15.00 16.54 533.89 580.34
Note: *** (**) [*] denote significance at p<.01 (p<.05) [p<.10]. All regressions include a constant term and the
full specification in Table 2, column 4. Regions include Asia, Latin America and the Caribbean, and the
formerly communist countries in Central and Eastern Europe. Stable democracy denotes countries that have
been democratic for at least ten years at the time of observation. Columns 1–3 should be compared to columns
4, 6 and 8 in Table 3, while columns 4–5 should be compared to columns 6 and 8 in Table 3.

35
Figure 1. Theoretical framework

36
 100
150
200
250

0
50
EST
CHL
PAR
BGR
LUX
PER
ECU
SUR
GTM
AUL
NIC
DOM
NEW
MKD
ROK
COS
NOR
BOL
HON
ROM
SWD
LAT
TTO
COL
LIT
ARG
TUR
CZR
PAN
VEN
MEX
Figure 2. Debt rates, all countries in sample

DEN
SWZ
SLO

37
MNT
FIN
SLV
SER
SAL
POL
URU
GUY
BRA
NTH
ALB
CRO
MLT
ISR
SPN
GMY
HUN
BLZ
AUS
CYP
UKG
CAN
FRN
ISL
USA
SIN
BEL
IRE
POR
ITA
JAM
GRC
JPN
Figure 3. Effects of regulatory freedom, conditional on government ideology

Figure 4. Effects of regulatory freedom, conditional on veto players

Note: The figure is based on results obtained with the democratic sample.

38
Figure 5. Effects of regulatory freedom, conditional on legislative fractionalization

Note: The figure is based on results obtained with the democratic sample.

39
Online appendix

In this Online Appendix, we provide Figure OA1, showing the average development of the
dependent variable (government debt/GDP and of regulatory freedom, in both cases from
1980 onwards. Both curves are shown for the full but changing sample (in black) and for the
sample of 18 countries for which data exist for 1980 (in grey). We refer to this figure in the
article, on p. XX.
We also document the cross-sectional test referred to in Section 4.4. In Tables OA1
and OA2, we make use of cross-sectional data from 2010, moving away from the panel
approach in the paper due to the very limited availability of data.
To test reflection effects as a channel, we use three indicators from the World Values
Survey (2014) and the European Values Study (2011): 1) average confidence in major
companies; 2) a rating of whether “government should take more responsibility to ensure that
everyone is provided for” versus “people should take more responsibility to provide for
themselves”; and 3) a rating of whether “competitions is good” versus “competition is
harmful”. The former is explored in Pitlik and Kouba (2015), while Aghion et al. (2010)
associate the two latter with regulatory activity.
To test reputation effects as a channel, we make use of the fact that Standard and
Poor’s (2018) rates 66 of the 67 countries in our sample. We use the credit rating in 2010–
2011, which we turn into numerical values with a perfect AAA rating counted as 1 and
adding one for each subsequent rating step. In case the ratings changed over the period, we
use the lowest rating as almost all cases with rating changes during this particular period
where downwards.
Lastly, the third channel for which we could identify a potential indicator is direct
economic effects. The indicator is the Grant Thornton (2018) measure of Global Dynamism,
which we include since regulatory freedom may affect debt development through a more
dynamic and entrepreneurial business environment.
As noted in the paper, we first of all associate regulation with confidence in major
companies as a proxy for the business or market friendliness of society. We here find a small
but significant positive association with freedom from credit regulation, but only weak
evidence of an association with debt development and again no change in the regulatory
freedom estimate.

40
 Turning to the measures from Aghion et al. (2010), we find even less evidence of any
association with regulation. The direct associations in Table OA1, columns 4 and 7, are small
and insignificant, and neither the measure of preferred government responsibility nor the
assessment of competition is associated with debt. Exploring the simple cross-sectional
associations, as in Aghion et al. (2010), thus does not provide evidence of reflection effects,
but with the limited analysis at hand, neither can we rule out that they play an actual role.
As for reputation effects, if regulatory freedom was important for debt development and
debt dynamics in most countries, one would believe that credit rating agencies ought to take it
into account when rating sovereign bonds. In columns 1–3 of Table OA2, we observe that
presidential institutions matter and that ratings are sensitive to recessions, as known from
previous studies (e.g., Afonso, 2003). Most importantly for our purpose, we also observe a
significant negative association between the rating score and regulatory freedom. For each
point of regulatory freedom, we find that credit ratings on average improve two steps. Yet,
when including the lagged debt level, credit ratings entirely lose significance. It thus seems
more likely that credit ratings and the assessment of dynamism are affected by debt
development instead of the opposite causality.
Finally, to test direct economic effects, we add the measure of global dynamism in
columns 4–6 in Table OA2. We again find that regulatory freedom is associated with
dynamism, although only weakly so for freedom from credit regulation. Yet, as in other
applications, we find no clear evidence of an association with debt development once a
lagged dependent is added to the specification. As in other cases here, the significant
association in column 8 appears to most likely be a cause of debt being reflected in the
dynamism index, i.e. taking into account when the dynamism index was coded, instead of an
actual effect of dynamism.
In total, while we cannot claim that these cross-sectional associations are more than
indications, none of the factors tested here seem to be valid candidates for transmission
mechanisms, but we cannot rule out that they do play a role either. For this, more detailed
analysis using future panel data is needed.

References
Afonso, António. 2003. Understanding the determinants of sovereign debt ratings: evidence
for the two leading agencies. Journal of Economics and Finance, 27(1), 56–74.

41
Aghion, Philippe, Yann Algan, Pierre Cahuc and Andrei Shleifer. 2010. Regulations and
distrust. Quarterly Journal of Economics, 125(3), 1015–1049.
European Values Study. 2011. European Values Study 1981–2008, Longitudinal Data File.
GESIS Data Archive, Cologne, ZA4804 Data File Version 2.0.0, doi:10.4232/1.11005.
Grant Thornton. 2018. Grant Thornton Global Dynamism Index. Chicago: Grant Thornton.
Inglehart, Ronald, Christian Haerpfer, Alejandro Moreno, Christian Welzel, Kseniya
Kizilova, Jaime Diez-Medrano, Marta Lagos, Pippa Norris, Eduard Ponarin and Bi Puranen
(eds.). 2014. World Values Survey: Round Six – Country-Pooled Datafile. Madrid: JD
Systems Institute.
Standard and Poor’s. 2018. S&P Global Ratings. New York: Standard and Poor’s.

42
Table OA1. Cross-sectional test of channels
Confidence Debt Debt Government Debt Debt Competition Debt Debt
in comp. responsibility
1 2 3 4 5 6 7 8 9
Lagged dependent .818*** .852*** .859***
(.078) (.079) (.078)
Growth -.026 -9.265** -5.623** .049 -9.206** -4.899** -.038 -8.219** -4.709**
(.023) (4.174) (2.106) (.063) (4.552) (1.998) (.063) (4.141) (1.981)
Recession -.004 8.102 12.276*** -.083 6.449 12.198*** -.038 6.041 12.021***
(.051) (7.043) (3.690) (.142) (6.867) (3.905) (.131) (6.982) (3.975)
Log population -.018 3.131 .047 .022 4.168 .163 .045 3.846 .177
(.015) (2.288) (2.164) (.054) (2.710) (2.169) (.080) (3.299) (2.151)
Democracy .115 -66.148*** -30.105** -.472 -57.042** -24.811*** -1.012*** -49.561* -25.474**
(.179) (19.69) (11.324) (.386) (27.709) (9.069) (.379) (26.798) (9.691)
Presidential .122 -11.564 -2.095 -.395 -14.242 -3.192 .201 -20.914 -4.332
(.090) (12.359) (6.099) (.204) (13.456) (8.078) (.186) (12.894) (7.463)
Proportional -.121** -15.995 -15.057* -.223 9.512 -5.308 .282 3.665 -6.188
(.058) (11.192) (8.244) (.298) (12.067) (9.024) (.409) (13.121) (9.151)
Veto players -.294 68.241 54.560** -1.424* 72.157 48.759*** 1.292 41.079 43.909**
(.373) (55.693) (22.652) (.816) (51.299) (18.236) (.797) (51.093) (19.417)
Government .122 44.069*** 18.809** -.315 37.046** 12.588 .357 29.296* 11.336
ideology (.138) (15.245) (8.663) (327) (17.842) (9.681) (.329) (15.943) (8.865)
Legislative -.267 -94.367 -62.589* .981 -65.658 -45.944 -1.101 -41.639 -42.481
fractionalization (.454) (59.589) (34.816) (.897) (48.100) (32.575) (1.069) (44.772) (30.803)
Regulatory freedom .051 -17.774*** -9.467*** .124 -20.561*** -8.570*** -.058 -18.522*** -8.095**

43
 (.061) (6.342) (3.472) (.146) (6.471) (4.041) (.124) (6.469) (3.866)
Confidence in major -62.348** -22.928*
companies (28.383) (12.749)
Government 10.613 2.936
responsibility (8.297) (4.721)
Competition 12.353 .444
(8.242) (5.958)
Regional FE Yes Yes Yes Yes Yes Yes Yes Yes Yes
Observations 48 48 48 49 49 49 49 49 49
R squared .308 .651 .906 .429 .599 .889 .292 .604 .888
F statistic - - - 579.44 4.64 31.79 2.63 3.25 30.24
Specific result
Credit regulation .051** -11.417*** -8.750*** .033 -13.740*** -8.842*** -.033 -13.070*** -8.688***
(.025) (3.505) (1.644) (.056) (3.059) (1.577) (.072) (2.979) (1.587)
Note: *** (**) [*] denote significance at p<.01 (p<.05) [p<.10]. All regressions include a constant term. Regions include Asia, Latin America and the Caribbean, and the
formerly communist countries in Central and Eastern Europe.

44
Table OA2. Cross-sectional test of channels
Credit Debt Debt Global Debt Debt
rating dynamism
1 2 3 4 5 6
Lagged dependent .778*** .882***
(.092) (.083)
Growth -.036 -5.540** -2.286 -1.212* -13.167** -6.919**
(.323) (2.631) (1.555) (.647) (4.798) (2.573)
Recession 1.413** 4.195 8.778*** -3.112*** -1.255 8.116
(.604) (4.644) (3.098) (1.162) (8.955) (5.031)
Log population -.716*** 6.984*** 2.509 -1.094 6.073 -.133
(.282) (2.536) (1.598) (.861) (4.184) (2.264)
Democracy .866 -45.739* -21.911** -11.426 -78.345 -15.547
(1.913) (24.239) (8.723) (6.319) (59.397) (16.419)
Presidential 2.962*** -31.487** -7.572 1.555 -23.351 -1.948
(1.092) (13.474) (7.638) (2.362) (16.598) (8.536)
Proportional .455 -13.174 -7.467 -3.105 -2.413 -11.514
(1.057) (10.862) (6.048) (2.681) (16.952) (8.549)
Veto players -4.759 83.406** 42.764*** 14.819* 79.617 31.596
(3.789) (37.494) (15.744) (8.732) (60.878) (21.273)
Government .665 15.618* 11.478* .371 55.286** 21.939*
ideology (1.216) (8.382) (5.979) (3.459) (22.972) (12.329)
Legislative -2.558 -4.903 -14.846 11.443 -64.394 -69.356*
fractionalization (3.141) (23.396) (21.500) (10.938) (62.999) (37.523)
Regulatory freedom -2.058** -5.431 -6.507** 2.813** -10.522 -6.188
(.842) (4.9130) (3.201) (1.224) (8.139) (4.435)
Competition
Credit rating 3.691*** .514
(1.028) (.754)
Global dynamism -2.405** -.306
(.999) (.652)
Regional FE Yes Yes Yes Yes Yes Yes
Observations 66 66 66 40 40 40
Countries
R squared .718 .623 .865 .766 .656 .918
F statistic 21.45 6.51 23.40 - - -
Specific result
Credit regulation -.936** -8.465*** -8.936*** 1.420* -13.173*** -8.194***
(.373) (2.549) (1.710) (.818) (3.663) (2.492)

45
Note: *** (**) [*] denote significance at p<.01 (p<.05) [p<.10]. All regressions include a constant term. Regions
include Asia, Latin America and the Caribbean, and the formerly communist countries in Central and Eastern
Europe.

46
Figure OA1. Average development of the government debt ratio and regulatory freedom
80 7,8
75 7,6
70 7,4
65
7,2
60
7
55
6,8
50
6,6
45
40 6,4

35 6,2

30 6
1980 1985 1990 1995 2000 2005 2010

Debt/GDP (left axis); full but changing sample

Debt/GDP (left axis); sample of 18 countries for which data exist from 1980
Regulatory freedom (right axis); full but changing sample
Regulatory freedom (right axis); sample of 18 countries for which data exist from 1980

47