Which democracies prosper? Electoral rules, form of government and economic growth

Abstract

Electoral rules and form of government have important economic effects, for example on taxation and public spending. However, there are no robust results in the literature when it comes to their effect on economic growth. This paper investigates whether electoral rules and form of government affects economic growth by applying panel data techniques on a very extensive dataset. There is no robust effect of presidentialism or parliamentarism on growth. However, there is very robust evidence for a positive, and quite substantial, effect of Proportional Representation (PR) electoral rules on economic growth. This is partly due to PR systems’ propensity to generate broad-interest policies, like universal education spending, property rights protection and free-trade, rather than special interest economic policies. Also semi-proportional systems seem to enhance growth relative to plural-majoritarian systems.

Full text

1
Which democracies prosper? Electoral rules, form of government and economic growth
Carl Henrik Knutsen
Department of Political Science, University of Oslo
ABSTRACT
Electoral rules and form of government have important economic effects, for example on
taxation and public spending. However, there are no robust results in the literature when it
comes to their effect on economic growth. This paper investigates whether electoral rules and
form of government affects economic growth by applying panel data techniques on a very
extensive dataset. There is no robust effect of presidentialism or parliamentarism on growth.
However, there is very robust evidence for a positive, and quite substantial, effect of
Proportional Representation (PR) electoral rules on economic growth. This is partly due to PR
systems’ propensity to generate broad-interest policies, like universal education spending,
property rights protection and free-trade, rather than special interest economic policies. Also,
semi-proportional systems seem to enhance growth relative to plural-majoritarian systems.
RESEARCH HIGHLIGHTS
 There is no effect of parliamentarism or presidentialism on economic growth
 PR and semi-PR systems increase growth relative to plural-majoritarian systems
 These results are based on panel data analysis, using a very extensive dataset
 PR may enhance growth through inducing broad-interest- and credible policies.

2
1. INTRODUCTION
Do electoral rules or form of government (the parliamentarism-presidentialism distinction)
matter for economic growth? Political scientists have long been interested in the potential
political consequences of electoral rules and form of government, such as nature of the party
system and regime stability. However, there has lately been an increasing focus also on
constitutional rules’ economic effects, largely thanks to the great work by Torsten Persson
and Guido Tabellini (e.g. 2003; 2004). This has resulted in several theoretical insights and
empirical results. Among others, presidential systems and plural-majoritarian systems likely
have lower tax rates, less public spending, and allocate a smaller share of public spending to
universal programs (but, see Acemoglu, 2005; Gabel and Hix, 2005).
Constitutional rules may also affect economic growth through systematically affecting
governments’ economic policies (e.g. Persson and Tabellini, 2003; Rodrik, 1996) and
countries’ economic institutions (e.g. North, 1990; Acemoglu et al., 2001; Persson, 2005).
However, there is no consensus on what specific types of constitutional features that improve
growth. Although some studies indicate that PR and parliamentarism may improve growth,
there are no well-established, robust results in the literature. By utilizing a greatly expanded
dataset and panel data methodology, this study is able to more accurately test hypotheses
related to whether electoral rules and form of government affect growth.
Section 2 discusses why electoral rules and form of government may matter for growth.
Section 3 presents the data, and Section 4 the empirical analysis. The analysis finds no robust
effect of parliamentarism or presidentialism on economic growth. However, PR and semi-PR
electoral rules produce substantially higher growth rates than plural-majoritarian rules. This
result is very robust, particularly for PR, and contrasts with the mainly non-robust results in
previous studies based on less extensive data material.

3
2. LITERATURE REVIEW AND ARGUMENTS
2.1 ELECTORAL SYSTEMS AND ECONOMIC GROWTH
As Gabel and Hix (2005: 4-5) put it, “[i]n their purest form, in PR systems voters choose
between lists of candidates presented by parties in multi-member districts and seats are
allocated in proportion to the share of votes received. In majoritarian systems, in contrast,
voters choose between individual politicians in single-member districts and the winning
candidate in each district is the one who receive the most votes in a district”. The simple
dichotomy of PR and plural-majoritarian electoral rules hides substantial variation in electoral
rule design (e.g. Grofman and Lijphart, 2003; Lijphart, 1999; Cox, 1997). In the empirical
section, I add a third category of semi-PR systems, based on Schjølset’s (2008) classification.
As Schjølset notes, this is a somewhat heteregenous category, incorporating different systems
that tend to produce intermediate outcomes in terms of proportionality between votes received
and seats in parliament (p. 80). This category includes systems based on specific electoral
rules, like the Single Nontransferable Vote, which is often considered semi-proportional
(Lijphart, 1994), and systems, like Germany’s, that utilize both plurality- and PR rules when
electing legislators.
1
Political economic literature emphasizes political accountability’s importance for achieving
“good” policies and outcomes, for example economic growth (e.g. Ferejohn, 1986; Benhabib
and Przeworski, 2005). When accountability is high, voters can and do vote poorly
performing politicians out of office (e.g. Powell and Whitten, 1993). This in turn induces
politicians to exert extra effort and opt for better macroeconomic policies. Under plurality rule
“mapping from votes to seats becomes steep if electoral races are close. This connection
ought to create strong incentives for good behavior: a small improvement in the chance of
1
As for example Lijphart (1994) notes, also other aspects, such as election (or legal) treshold,
district magnitude and assembly size affect the proportionality between votes and seats.

4
victory would create a large return in terms of seats.” (Persson and Tabellini, 2004: 82).
However, if parties or candidates expect to win or lose by a landslide, the incentives to exert
effort or forego rents are weaker under plurality rule. Nevertheless, PR increases the
frequency of coalition governments (Persson et al., 2003; Powell, 2000). Under coalition
governments, voters are less able to discern which party to blame for bad performance or
reward for good. Thus, retrospective “economic voting” decreases under coalition
governments (Powell and Whitten, 1993). Also, plurality rule is associated with smaller
district magnitudes, which eases voters’ monitoring of candidates, thus improving
accountability at the district level.
Electoral rules may also affect economic growth through affecting public sector size. PR
increases taxation and public spending (Persson, 2003; Persson and Tabellini, 2003, 2004).
However, it is unclear whether balanced increases in tax revenues and public spending
increase or reduce economic growth. Microeconomic theory points to distortionary effects
from high tax-rates. However, more public revenue means better opportunities for investing in
infrastructure, education and public health care, which enhance growth (e.g. Easterly and
Rebello, 1993; Mankiw et al., 1992).
As seen above, plural-majoritarian rules were expected to increase district-level
accountability. However, voters in small districts are perhaps less concerned with policies that
enhance nationwide growth than with obtaining government resources at the rest of the
country’s expense, or protectionist measures for their cornerstone industries. Voters in PR
systems may also be concerned with redistributive policies, but larger districts and more
continuous mapping from voters to seats reduce the political incentives for proposing
narrowly targeted redistributive programs (e.g. Persson and Tabellini, 2004). Rather,
politicians under PR systems may want to propose universal redistributive programs that

5
appeal to large, geographically dispersed groups of voters. Such redistributive programs often
include increased spending on broad-based public education programs and healthcare
spending, both of which boost growth through increasing human capital (e.g. Mankiw et al.,
1992). Indeed, several studies indicate that PR increases the share of public spending going to
universal programs, and that plurality rule increases the share going to special interest groups,
particularly geographically concentrated groups in electoral distrcits with tough competition
(e.g. Persson and Tabellini, 2004; Milesi-Feretti et al., 2002). Such spending-allocation is a
rational political strategy for office-motivated politicians, but narrow programs that affect
only particular regions and groups are likely to generate less economic benefit than broader
programs related to national public goods.
Generally, PR produces “policies better serving the interests of broad majorities than do
majoritarian elections, either directly through incentives of politicians, or indirectly via party
formation and the incidence of coalition government” (Persson, 2005: 7-8). Politicians with
large winning coalitions have stronger incentives for pursuing general property rights
protection (Bueno de Mesquita et al., 2003). Thus, as PR increases the size of government’s
winning coalition, we would expect PR systems to perform better on property rights
protection. PR is also expected to induce open trade policies rather than selective
protectionism for narrow producer interests, as it insulates politicians from protectionist
demands by regional or sector-specific interests (Rogowski, 1987: 207-8). Empirically,
Persson (2005) indeed finds that PR improves property rights protection and induces trade
liberalization (see also Persson and Tabellini, 2006).
PR systems should also enhance growth rates relative to plural-majoritarian systems through
yielding less abrupt changes in parliamentary seat composition before and after elections;
policies by governments elected in PR systems are expected to last longer, and, if changed,

6
policies will be adjusted rather than reversed (e.g. Rogowski, 1987). Conversely, one of the
main problems with “majoritarian democracies” is the sudden and substantive policy-
alterations induced by shifting electoral fortunes (Lijphart, 1999). Policy- and reform
reversals have negative economic effects, as they create instability and render private
investments less productive (Rodrik, 1991). Policy reversals become even more problematic
when foresighted investors, in anticipation, withhold investments altogether. Policies’
economic success hinge upon private actors’ beliefs about their expected stability; success is
negatively related to the probability of reversal (Kydland and Prescott, 1977; North and
Weingast, 1989). In addition to having less dramatic temporal changes in parliamentary seat
composition, PR systems more often have broad coalition governments, implying that many
reforms are political compromises with broad support.
To sum up, the political economic literature has focused on the beneficial accountability
characteristics of plural-majoritarian systems, which may induce office-motivated politicians
to enact growth-promoting policies. However, other aspects with PR systems may counter
such an effect. First, there are stronger incentives for politicians under PR systems to enact
public policies that benefit broad rather than narrow interests, including growth enhancing
policies such as provision of education for the masses, property rights protection and free-
trade policies. Moreover, PR systems should increase investment, as the investment climate is
less risky because of less abrupt policy changes. Persson and Tabellini (2003) find some
evidence for the hypotheses that PR increases productivity growth, but the result is not robust.
Although Persson (2005) in some models finds a positive effect of PR on GDP per capita and
total factor productivity, the effects are not robust. Persson and Tabellini (2006) find no
significant effect of electoral rules on growth, when studying democratizing countries. Older

7
empirical studies using smaller samples generally find a small, positive or no effect of
“Consensus democracy” or PR more specifically on GDP growth (Lijphart, 1999: 263-270).
Carey and Hix (2009) show that ‘intermediate’ electoral systems to a large extent generate
one of the beneficial outcomes of PR systems, relatively accurate representation of various
voters’ preferences, without reducing the accountability-benfit of plural-majoritarian systems
too much. Above, plurality systems were hypothezised to enhance growth mainly through
enhancing accountability. Moreover, representational issues are at least indirectly related to
the argument on PR enhancing growth through inducing politicians to follow broad-interest
policies. If the representation-accountability trade-off is concave, as argued by Carey and Hix,
semi-PR systems may be expected to produce relatively high growth rates. The reason is
simply that semi-PR systems may to a large degree possess the different advantages of both
PR and plurality systems, without having the respective systems’ drawbacks.
2.2 FORM OF GOVERNMENT AND ECONOMIC GROWTH
There is considerable institutional variation within the classes of presidential and
parliamentary systems (e.g. Shugart and Carey, 1992; Strøm, 1990). However, in a
parliamentary system “the executive … is chosen by and responsible to an elective body (the
legislature), thus creating a single locus of sovereignty”, whereas in a presidential system
“policy-making power is divided between two separately elected bodies: the legislature and
the president” (Gerring et al., 2009: 337). It is common to operate with a third category, semi-
presidential systems; also denoted “mixed systems” (Cheibub, 2007). Cheibub classifies these
as systems where either the assembly or the directly elected president can remove the
government (2007: 34).
Presidential systems are more fragmented, institutionally, than parliamentary systems,
because of weaker ties between legislature and executive. Separation of executive and

8
legislative power is argued to facilitate credible political commitments (Keefer and Stasavage,
2003) and increase information revelation about political processes to the public (Persson et
al., 1997), thus improving political accountability, which may again enhance growth (e.g.
Benhabib and Przeworski, 2005). Moreover, in a presidential system, both executive and
legislature are “institutional veto players” (Tsebelis, 2002). Presidential systems are thus
assumed to have greater checks and balances on executive power, which mitigates
opportunities for power abuse. Power concentration may reduce growth, as the incentives of
those in power are often not aligned with growth-conducive policies (e.g. Olson, 1993).
However, separation of powers under presidentialism may also amplify rigidity in the political
system (Linz, 1990). The possibility of different parties controlling the legislature and the
executive under presidentialism could result in political gridlock, thereby complicating
passing of efficiency-enhancing economic reforms.
2
Presidential systems may also increase
special interest groups’ opportunity to affect political processes, which may reduce economic
dynamism (Olson, 1982). Industrial groups or other lobbies can tie their fortunes with
important groups in parliament or the president (e.g. Cox and McCubbins, 2001).
Presidential systems reduce public spending and taxation (see e.g. Persson et al., 2000;
Persson and Tabellini 2004). But as there is no clear link between public sector size and
economic growth, it is unclear whether presidentialism hurts or helps growth through this
channel. However, according to Gerring et al. (2009: 353), there are several other reasons
why parliamentarism improves governance, and thus economic outcomes. One underlying
factor is parliamentarism’s propensity to better solve political coordination problems, since it
2
Several contributions have either nuanced (e.g. Shugart & Carey, 1992) or disputed (e.g.
Cheibub, 2007) Linz’ argument (1990) that presidentialism increases the probability of
democratic breakdown.

9
institutionalizes debate and negotiations, and reduces the number of veto players (pp. 354-5).
Parliamentarism also strengthens party discipline (e.g. Gerring et al., 2009), among others
because of the vote of confidence procedure (e.g. Huber, 1996). High party discipline
channels public spending towards broad, national programs, instead of narrow, targeted
programs to region-specific interests. Indeed, parliamentary regimes allocate more of their
public spending to broad, universal programs (Persson and Tabellini, 2004).
Gerring et al. (2009) find that parliamentarism outperforms presidentialism on among others
economic growth, bureaucratic quality, investment environment and life expectancy. But,
Persson and Tabellini (2003) find no robust effect on corruption of form of government.
However, presidential systems inhibit economic institutional factors like bureaucratic quality
and property rights, and may reduce labor- and total factor productivity growth in low-quality
democracies (Persson and Tabellini, 2003). Persson (2005) finds that reform towards
parliamentary democracy may generate higher economic growth than reforms towards
presidential democracy, but the results are not robust. Reforms into parliamentarism
strengthen property rights and enhance trade openness (Persson, 2005). But, Persson and
Tabellini (2006: 321) find that a “new parliamentary democracy grows 1.5 percentage points
less than a new presidential democracy”, when studying democratizing countries.
3. DATA AND METHODOLOGICAL ISSUES
Several models below include more than 3500 country-year observations from over 100
countries. Some countries have time-series from 1820 to 2002, although most have shorter
time series. This is the most extensive dataset used for analyzing constitutional rules’
economic effects. The long time-series used here dramatically expands the number of data
points, and panel data analysis becomes an option. This allows utilizing within-nation-, as
well as cross-sectional, variation. Constitutional structures seldom change (Persson and

10
Tabellini, 2003: 88). However, there are democracies with different constitutional
experiences, particularly on electoral rules. To name some, Argentina, Belgium, Chile,
Colombia, Greece, Italy, Macedonia, New Zealand, Norway, The Philippines, Spain and Sri
Lanka have all experimented with different electoral rules over the time period covered here.
Thus, the dramatic expansion of time series and application of panel data models contribute to
a literature where empirical tests have found ambiguous and mainly non-robust results.
Data on PPP-adjusted GDP per capita and population are from Maddison (2006). Maddison
has estimated GDP and population using a variety of sources, and some estimates rely on less
valid proxies than others, especially before WWII (Maddison, 2006). Some of the population
and GDP time-series are interpolated, assuming constant GDP- and population growth rates
(REFERENCE REMOVED). Data on democracy level and regime duration are from Polity
IV (Marshall and Jagger, 2002). I use Schjølset’s (2008: 135-42) classification and data on
constitutional structures. Only country-years with a Polity-score ≥3 are classified as
democratic by Schjølset (the Polity-index goes from -10 to 10), and thus included in the
analysis. After 1820, 38 percent of classified country-years are presidential systems, 26
percent are semi-presidential, and 37 percent are parliamentarian. Also, 44 percent are PR
systems, 10 percent are semi-PR and 46 percent are plural-majoritarian.
I furthermore use Schjølset’s (2008) data on federal, unitary and mixed federal-unitary
systems. Colonizer-, region- and plurality religion dummies are from REFERENCE
REMOVED. Data on latitude, the Frankel-Romer instrument for trade openness (FRAROM),
and fraction of English (ENGFRAC) and other major European language speakers
(EURFRAC) are from Hall and Jones (1999). Ethnic fractionalization data are from Alesina et
al. (2003). This index covers more countries than the older ELF-index used by Persson and
Tabellini (2003).

11
On average, presidential systems had lower economic growth rates than semi-presidential-
and parliamentarian systems from 1820 to 2002. Whereas presidential systems’ GDP per
capita on average grew 1.5 percent annually, parliamentarian systems grew with 2.2 percent
and semi-presidential with 2.5 percent. Especially after the oil-crisis in 1972, presidential
systems grew slowly relative to other systems. On average, PR systems grew with 2.5 percent,
whereas semi-PR and plural-majoritarian systems grew with 1.9 and 1.6 percent respectively.
There are few observed semi-PR and PR country-years in the 19th century. Differences
between electoral systems are therefore primarily based on data from the 20th century. PR
systems outperformed plural-majoritarian systems throughout the 20th century, especially
before 1972. However, countries with PR rules and parliamentarism may grow faster because
they are systematically related to other relevant variables, like initial income level or
geographic factors. We therefore control for such factors below.
There is no single, overarching theoretical framework on this issue, which complicates
statistical modeling. However, the solid empirical work by Persson and Tabellini (2003), and
later methodical criticisms (e.g. Acemoglu, 2005), has helped identify plausible control
variables. I follow Persson and Tabellini (2003) closely, in terms of control variables used.
Since particular institutional structures have been relatively popular in different time periods,
and these periods are associated with different global economic growth rates, I enter decade
dummies as controls in all models (1990s extend to 2002), to control for time-specific effects.
In all models, the independent variables are lagged with two years, to reduce endogeneity
problems.
I first employ OLS with Panel Corrected Standard Errors (PCSE) (Beck and Katz, 1995).
OLS with PCSE utilizes both between- and within-nation variation. The interpretation of
coefficients is similar to OLS, but the models take into account the cross-sectional – time-

12
series structure of the data (standard errors take into account heteroskedasticity and
contemporaneous correlation between panels, and AR1-autocorrelation within panels). GDP
per capita growth is the dependent variable. In Model 1A, I enter dummies for PR and semi-
PR systems and dummies for parliamentary and semi-presidential systems. Plural-
majoritarian and presidential systems are reference categories. In addition, following Persson
and Tabellini (2003), I add dummies for British (plus American) colonies, a dummy for
Spanish and Portuguese colonies, and a dummy for other former colonies. I also add region
dummies, dummies for federal and mixed federal-unitary systems, the ethnic fractionalization
index, log GDP per capita, log population, log regime duration and decade dummies.
4. EMPIRICAL ANALYSIS
4. 1 MAIN RESULTS
The results from Model 1A in Table 1, based on 3701 country-years from 107 countries,
indicate a strong, significant effect of electoral rules; both PR and semi-PR systems produce
higher growth than plural-majoritarian. The effect of PR is significant at the 0.1%-level and
the effect of semi-PR at the 1%-level. The estimates indicate that PR elections increase annual
GDP per capita growth with almost 1 percentage point relative to plural-majoritarian.
However, 1A shows no significant effects of form of government on growth.
Model 1B controls also for fraction of English speakers, fraction of other major European
language speakers, the Frankel Romer trade instrument and latitude. In 1B, several
observations drop out because of missing data. The results are however strikingly similar to
those from 1A; the positive effects of PR and semi-PR are quite large, and significant at the
1%- and 5%-levels respectively. There is still no significant effect of form of government.
Model 1C also controls for degree of democracy (Polity) and plurality religion. As seen in
Table 1, the results are relatively similar to those from 1A and 1B. The positive effects of PR
and semi-PR are quite robust. Form of government, however, does not matter for growth.

13
Table 1 Results from OLS with PCSE and Fixed Effects modelsa
OLS1A
OLS1B
OLS1C
FE2A
FE2B
FE2C
b/(t)
b/(t)
b/(t)
b/(t)
b/(t)
b/(t)
PR
0.93*** (3.31)
0.95** (3.20)
1.04*** (3.36)
1.22** (3.14)
0.95* (2.45)
0.81* (2.11)
Semi-PR
1.17** (2.86)
1.02* (2.53)
1.02* (2.39)
1.38* (2.30)
0.53 (0.91)
0.44 (0.75)
Parliam.
-0.15 (-0.45)
-0.19 (-0.57)
-0.21 (-0.62)
0.30 (0.54)
0.41 (0.74)
0.58 (1.05)
Semi-pres.
-0.04 (-0.12)
-0.25 (-0.68)
-0.43 (-1.07)
0.02 (0.03)
0.15 (0.28)
0.43 (0.82)
Ethnic fract.
-1.67*** (-3.86)
-1.55*** (-3.45)
-1.34* (-2.54)
British col.
0.26 (0.69)
-0.24 (-0.47)
-0.16 (-0.29)
Iberian col.
-1.61* (-2.16)
-2.62*** (-3.47)
-2.46** (-2.96)
Other col.
-2.03* (-2.52)
-2.651** (-3.10)
-2.23^ (-1.78)
Federalist
0.09 (0.32)
-0.02 (-0.07)
0.01 (0.04)
0.11 (0.11)
0.19 (0.20)
0.63 (0.65)
Hyb. fed-un.
0.33 (0.76)
0.40 (0.97)
0.40 (0.98)
-0.22 (-0.37)
-0.00 (-0.00)
0.27 (0.44)
E.Eur.-Sov.
-0.73 (-0.64)
-0.26 (-0.27)
-0.41 (-0.42)
Africa SS
-1.28* (-2.02)
-1.23 (-1.50)
-0.83 (-1.00)
Asia-Pacific
0.61 (0.99)
1.05 (1.34)
1.25 (1.27)
M.E. - N.A.
0.18 (0.32)
0.22 (0.35)
0.78 (0.64)
Latin Am.
0.27 (0.36)
0.14 (0.19)
-0.12 (-0.17)
Ln GDP pc
-0.41 (-1.55)
-0.35 (-1.36)
-0.35 (-1.30)
-4.99*** (-9.22)
-4.21*** (-8.28)
-4.30*** (-8.42)
Ln Pop.
-0.08 (-0.95)
-0.27* (-2.33)
-0.37** (-2.89)
-1.26*** (-3.36)
-1.04** (-2.84)
-1.06** (-2.90)
Ln Reg. dur.
0.02 (0.21)
-0.08 (-0.74)
-0.12 (-1.06)
0.13 (0.95)
-0.12 (-0.92)
-0.17 (-1.33)
ENGFRAC
-0.55 (-0.93)
-0.66 (-1.01)
EURFRAC
0.64 (1.43)
1.10* (2.12)
HJFRAROM
-0.50* (-2.08)
-0.58* (-2.21)
Latitude
-0.00 (-0.46)
0.00 (0.47)
Polity
-0.07 (-1.00)
-0.17** (-2.61)
Sunni
-1.44 (-1.46)
Shia
1.57 (1.29)
Catholic
-0.60 (-0.47)
Protestant
-0.45 (-0.33)
Orthodox
0.63 (0.39)
Hindu
0.22 (0.17)
Buddhist+
-0.54 (-0.43)
Indigeneous
-1.50 (-1.11)
N
3701
3521
3520
3701
3521
3520
The results above may be driven by omitted country-specific factors. There is relatively little
within-nation variation in electoral systems and form of government over time (Persson and
Tabellini, 2003: 88). However, I have more within-nation variation to draw on than Persson
and Tabellini, because of much longer time-series. I therefore estimate Fixed Effects (FE)
models, which control for country-specific factors by adding country-dummies. FE models
thus estimate effects based only on within-nation variation over time. The FE results are
strikingly similar to the OLS with PCSE results. There is no significant effect of form of
government. PR significantly improves growth in all models (at least 5%-level). Even if we
a *** indicates p<0.001, ** indicates p<0.01, * indicates p<0.05 and ^ indicates p<0.10.
Decade dummies and constant are omitted.

14
estimate the effect of PR without utilizing cross-country comparisons, we find that it
improves growth. Only the FE model with the fewest controls yields a significant effect (5%-
level) for semi-PR. However, there are relatively few observations to draw inferences from
for semi-PR, and the non-robust result is thus perhaps not very surprising, particularly given
that the FE models do not utilize cross-country vatiation.
4. 2 ROBUSTNESS CHECKS
The results above may however be sensitive to different specifications. I therefore conducted
additional robustness checks. OLS with PCSE results are reported, but results are quite similar
when using FE models. First, I conducted sensitivity checks, leaving out one independent
variable or set of dummies at a time from 1A. All models showed significant PR and semi-PR
coefficients at least at the 5%-level. The forms of government dummies were always
insignificant. Second, I tested models with annual dummies rather than decade dummies. PR
was still significant (1%-level) in all models, and form of government was still insignificant.
Third, I tested different lags on the independent variables. Models where independent
variables were lagged with one or three years had significant PR-coefficients at the 1%-level;
and models with four or five years had significant coefficients at the 5%-level. Also the semi-
PR dummy was robust. There were no significant forms of government dummies in any
model. Fourth, the criterion for being included in the samples above was a Polity-score ≥3.
Polity cut-offs of 4, 5, 6 and 7 were tested, and the results remain stable. The results above are
thus not driven by the inclusion of “semi-democracies”.
Fifth, I estimated models using only data from after 1899, 1945 and 1979 respectively. All
models still showed a positive, significant effect of PR (at least 1%-level) and semi-PR (at
least 5%-level). Parliamentarism and semi-presidentialism were still insignificant (10%-
level), but with an exception for the post-1979 sample: In 1A and 1B, parliamentarism had a
positive, significant (5%-level) effect. This will be discussed further below.

15
Sixh, I tested whether electoral rules had different effects in Western countries (Western
Europe, US, Canada, Australia and New Zealand) than in other countries. The effect of PR in
the Western sample was somewhat weaker, but still significant at the 5%-level in 1A, 10%-
level in 1B, and insignificant in 1C (p=0.16). This non-robust result is not very dramatic for
the conclusion above, given the relatively few observations in the sample. Interestingly, PR
systems had a significant positive effect at the 1%-level in all models in “the rest” of the
world. The effect of PR is thus relatively robust over time and space.
Seventh, there may be outliers driving the results. The five Nordic countries have been
economically successful, and have mostly used PR. However, the PR coefficient was still
significant (1%-level), when excluding these countries from the global sample. The lack of
positive effect of parliamentarism could have been due to the US’ high growth. But, the effect
of parliamentarism on growth was still insignificant when excluding the US. Table A2 in the
Internet Appendix presents some of the results described above.
I also applied 2SLS to check whether the results above are due to economic growth affecting
constitutional structures (see Persson and Tabellini, 2003; Acemoglu, 2005). These results
yield weaker results for the hypothesis that PR enhances growth, and, like above, find no
effect of form of government. These results, and a methodical discussion, are provided in the
Internet Appendix together with other robustness checks, for example random effects- and
non-parametric matching models. The latter models find quite similar results to those reported
above; a positive effect of PR and semi-PR and no effect of form of government.
4. 3 A BRIEF DISCUSSION
Section 2.2 presented arguments indicating that presidentialism had both positive and
negative effects on growth; the results above may be due to these effects cancelling each other
out. However, the arguments in 2.2 may have questionable relevance in the first place. The

16
number of veto players may affect growth. However, as Tsebelis (2002) points out, this
number depends on several other institutional structures and political factors. There is also
large institutional variation within the categories of parliamentary and presidential regimes.
These variations matter for different political processes and outcomes (e.g. Shugart and
Carey, 1992), and may also matter for growth.
Moreover, formal checks on executive powers are at best moderately correlated with de facto
checks. The negative effect of presidentialism in the post-1980 sample, Gerring et al.’s (2009)
results based on data from after 1960, and Persson and Tabellini’s result (2003) that
presidentialism is economically harmful in low-quality democracies, may all be explained by
personal power concentration in new, low-quality presidential democracies. Parliamentarism
may concentrate power to parties controlling the majority of seats in parliament. However,
personal power concentration is likely worse for an economy (e.g. Olson, 1993), and
presidents may abuse their positions for personal gains. Presidentialism has been linked to
power concentration, personification of politics and weak parties in Africa (van de Walle,
2003) and Eastern European and ex-Soviet countries (Fesnic, 2006). There might thus be a
selection effect in the post-1980 sample, as elites in young democracies have substituted low-
quality presidential democracy for autocracy. However, if elites in democratizing African or
ex-Soviet countries opt for presidentialism because of its (correctly) perceived “advantages”
in allowing for personal power concentration, it is unclear whether this selection generates
pure bias or is indirectly part of a negative effect of presidentialism on growth.
Section 2.1 presented an argument pointing out that plural-majoritarian systems enhanced
growth because of high political accountability. Empirically, this seems to be trumped by
other factors. The theoretical discussion pointed out PR systems’ ability to reduce economic
risk through generating more stable and credible policies. Risk reduction and stability could

17
affect a wide variety of economic activities and transactions (North, 1990). One plausible
candidate is physical capital investment. However, when I substitute real investment as share
of GDP for economic growth as dependent variable, there is no significant effect of PR. The
investment-data from the World Development Indicators (WDI) are not as extensive as the
GDP data from Maddison. However, these results may indicate that PR does not enhance
investment, which casts doubts on the “PR and credible policies” argument’s validity.
However, another likely factor behind PR systems’ higher growth is their tendency to produce
broad-interest policies. There is already strong empirical evidence that PR enhances trade
openness and property rights protection (Persson, 2005). Human capital is another vital source
of economic growth (e.g. Mankiw et al., 1992), and provision of universal education is a
typical broad-interest policy (Bueno de Mesquita et al., 2003). Indeed, models where primary
or secondary school enrollment ratios (from WDI) are substituted for economic growth,
typically find a positive, significant effect of PR. The 1A-model, for example, finds a
significant effect at the 1%-level for primary- and 0.1%-level for secondary schooling. This is
a likely important pathway through which PR increases growth.
5. CONCLUSION
This paper estimated the effects of electoral rules and form of government on economic
growth. The results were based on a much larger dataset than earlier studies draw on.
Pesidentialism does not seem to reduce economic growth. Cheibub (2007) contested Linz’
(1990) perils-of-presidentialism thesis on democratic survival. This paper contests the
economic-perils-of-presidentialism results recently recorded by Gerring et al. (2009).
However, also this paper finds a negative effect of presidentialism on growth after 1979.
Persson and Tabellini (2003) find that presidentialism only hurts growth in low-quality

18
democracies, and Gerring et al.’s results may stem from many new, low-quality democracies
opting for presidentialism.
Electoral systems, however, seem to matter systematically for growth. Plural-majoritarian
systems may generate higher political accountability than PR systems, but not necessarily
much higher than semi-PR systems (Carey and Hix, 2009). However, plural-majoritarian
systems produce lower economic growth than both PR- and semi-PR systems. PR systems
likely produce higher growth because they promote broad interest-, rather than special
interest, policies, and perhaps because PR systems produce more stable and thus credible
economic policies. The effect is astonishingly robust, and point-estimates indicate that PR
increases growth rates with 1 percentage point relative to plural-majoritarian systems. PR and
semi-PR systems are thus not only beneficial for representation of diverse groups in politics
(Lijphart, 1999; Carey and Hix, 2009) and for decreasing the distance between median voter
and median government member (Huber and Powell, 1994; Carey and Hix, 2009). PR and
semi-PR systems also generate more prosperity than plural-majoritarian systems.

19
6. LITERATURE
Acemoglu, D., 2005. Constitutions, Politics and Economics: A Review Essay on Persson and
Tabellini’s "The Economic Effect of Constitutions". Cambridge, MA. NBER Working Paper
11235.
Acemoglu, D., Johnson, S., Robinson, J.A., 2001. The Colonial Origins of Comparative
Development: An Empirical Investigation. American Economic Review 91 (5), 1369-1401.
Alesina, A., Devleeschauwer, A., Easterly, W., Kurlat, S., Wacziarg, R., 2003.
Fractionalization. Journal of Economic Growth 8 (2), 155-194.
Beck, N., Katz, J.N., 1995. What to do (and not to do) with Time-Series – Cross-Section Data.
The American Political Science Review 89 (3), 634-647.
Beck, N., Katz, J.N., 2001. Throwing Out the Baby with the Bath Water: A Comment on
Green, Kim, and Yoon. International Organization 55 (2), 487–495.
Benhabib, J., Przeworski, A., 2005. Economic Growth under Political Accountability. New
York University. Working Paper.
Bueno de Mesquita, B., Smith, A., Siverson, R.M., and Morrow, J.D., 2003. The Logic of
Political Survival. The MIT Press, Cambridge, MA.
Carey, J.M., Hix, S., 2009. The Electoral Sweet Spot: Low-Magnitude Proportional Electoral
Systems. Working Paper.
Cheibub, J.A., 2007. Presidentialism, Parliamentarism and Democracy. Cambridge:
Cambridge University Press.
Cox, G., 1997. Making votes count: Strategic coordination in the world’s electoral systems.
Cambridge: Cambridge University Press.
Cox, G., McCubbins, M., 2001. The Institutional Determinants of Economic Policy
Outcomes. In: Haggard, S., McCubbins, M., (Eds.), Presidents and Parliaments. Cambridge
University Press, Cambridge, pp. 21-63.
Ferejohn, J., 1986. Incumbent performance and electoral control. Public Choice 50 (1-3), 5-
25.
Fesnic, F., 2006. Institutional Design and Democracy in Post-Communist Countries.
University of Illinois at Urbana-Champaign. Paper.
Gabel, M., Hix, S., 2005. Reconsidering the Economic Effects of Constitutions. Paper.
Gerring, J., Thacker, S.C., Moreno, C., 2009. Are Parliamentary Systems Better?
Comparative Political Studies 42 (3), 327-359.
Hall, R.E., Jones, C.I., 1999. Why Do Some Countries Produce So Much More Output Per
Worker Than Others? Quarterly Journal of Economics 114 (1), 83-116.

20
Huber, J.D., 1996. The vote of confidence in parliamentary democracies. American Political
Science Review 90 (2), 269-282.
Huber, J.D., Powell, G.B., 1994. Congruence between citizens and policymakers in two
visions of liberal democracy. World Politics 46 (3), 291-326.
Kydland, F.E., Prescott, E.C., 1977. Rules rather than discretion: The inconsistency of optimal
plans. Journal of political economy 85 (3), 473-493.
Lijphart, A., 1994. Electoral Systems and Party Systems: A Study of Twenty-Seven
Democracies 1945-1990. Oxford: Oxford University Press.
Lijphart, A., 1999. Patterns of Democracy. Yale University-Press, New Haven.
Lijphart, A., Grofman, B., 2003. Introduction. In: Lijphart, A., Grofman, B., (Eds.), Electoral
Rules and Their Political Consequences. Agathon Press, New York, pp.1-15.
Linz, J.J., 1990. The Perils of Presidentialism. Journal of Democracy 1 (1), 51-69.
Maddison, A., 2006. The World Economy. OECD Publishing, Paris.
Mankiw, G.N., Romer, D., Weil, D.N., 1992. A Contribution to the Empirics of Economic
Growth. Quarterly Journal of Economics 107 (2), 407-437.
Marshall, M.G., Jaggers, K., 2002. Polity IV Project – Dataset Users’ Manual. College Park:
Program Center for International Development and Conflict Management, University of
Maryland. Manual.
Milesi-Ferretti, G.M., Perotti, R., Rostagno, M., 2002. Electoral Systems and Public
Spending. The Quarterly Journal of Economics 117 (2), 609-657.
North, D.C., 1990. Institutions, Institutional Change and Economic Performance. Cambridge
University Press, New York.
North, D.C., Weingast, B.R., 1989. Constitutions and Commitment: The Evolution of
Institutions Governing Public Choice in Seventeenth-Century England. The Journal of
Economic History 49 (4), 803-832.
Olson, M., 1982. The Rise and Decline of Nations. Economic Growth, Stagflation and Social
Rigidities. Yale University Press, New Haven.
Olson, M., 1993. Dictatorship, Democracy and Development. The American Political Science
Review 87 (3), 567-576.
Persson, T., 2005. Forms of democracy, policy and economic development. Working Paper.
Persson, T., Tabellini, G., 2003. The Economic Effects of Constitutions. The MIT Press,
Cambridge, MA.

21
Persson, T., Tabellini, G., 2004. Constitutions and Economic Policy. The Journal of
Economic Perspectives 18 (1), 75-98.
Persson, T., Tabellini, G., 2006. Democracy and development: The devil in the details.
American Economic Review 96 (2), 319-324.
Persson, T., Roland, G., Tabellini, G., 1997. Separation of powers and political accountability.
Quarterly Journal of Economics 112 (4), 1163-1202.
Persson, T., Roland, G., Tabellini, G., 2000. Comparative politics and public finance. Journal
of Political Economy 108 (6), 1121-1161.
Persson, T., Roland, G., Tabellini, G., 2003. How do electoral rules shape party structures,
government coalitions, and economic policies? Working Paper.
Powell, G.B., 2000. Elections as instruments of democracy. Yale University Press, New
Haven.
Powell, G.B., Whitten, G.D., 1993. A Cross-National Analysis of Economic Voting: Taking
Account of the Political Context. American Journal of Political Science 37 (2), 391-414.
Rodrik, D., 1991. Policy Uncertainty and Private Investment in Developing Countries. Journal
of Development Economics 36 (2), 229-242.
Rodrik, D., 1996. Understanding Economic Policy Reform. Journal of Economic Literature
34 (1), 9-41.
Rogowski, R., 1987. Trade and the Variety of Democratic Institutions. International
Organization 41 (2), 203-23.
Schjølset, A., 2008. Institutional Variance of the Democratic Peace, 1816-2002: Electoral,
Executive, and Federal Systems in Time and Space. University of Southern California. PhD
thesis.
Shugart, M.S., Carey, J.M., 1992. Presidents and Assemblies: Constitutional Design and
Electoral Dynamics. Cambridge University Press, Cambridge.
Strøm, K., 1990. Minority Government and Majority Rule. Cambridge University Press,
Cambridge.
Tsebelis, G., 2002. Veto Players – How Political Institutions Work. Princeton University
Press, Princeton.
van de Walle, N., 1993. Presidentialism and clientelism in Africa’s emerging party systems.
Journal of Modern African Studies 41 (2), 297-321.
TWO REFERENCES ARE REMOVED FOR ANONYMITY