Abstract
This article introduces a novel database that measures governments’ compliance with national constitutions. It combines information on de jure constitutional rules with data on their de facto implementation. The individual compliance indicators can be grouped into four categories that we aggregate into an overall indicator of constitutional compliance: property rights and the rule of law, political rights, civil rights, and basic human rights. The database covers 175 countries over the period 1900 to 2020 and can be used by researchers interested in studying the determinants or the effects of (non)compliance with constitutions. Our investigation of the stylized facts of constitutional compliance reveals a long-term increase in compliance, which occurred primarily around the year 1990. The Americas experienced the steepest increase in compliance, but also Africa and Europe improved particularly at the end of the Cold War. Democracies – particularly those with parliamentary and mixed systems – show more constitutional compliance than nondemocracies, among which military dictatorships perform the worst. Constitutional design also matters: Constitutions that allow for the dismissal of the head of state or government for violating constitutional rules are being complied with more.
Similar content being viewed by others
1 Introduction
While social scientists have been interested in constitutions for several decades, quantitative empirical research on their causes and consequences has only taken off over the last decade. One important reason for the slow development of such scholarship has been the limited availability of comparable data on the characteristics of countries’ constitutions (see, e.g., Elkins et al., 2009). Research shows that constitutional text can, but does not always, have important consequences (Bjørnskov & Mchangama, 2019; Ginsburg & Huq, 2016; Persson & Tabellini, 2003; Voigt & Gutmann, 2013). This has motivated a recent and very active strand of literature dealing with when constitutional text translates into constitutional practice (Chilton & Versteeg, 2020; Voigt, 2021 surveys this literature).
Even though constitutional compliance is observable and highly relevant, quantitative research of this phenomenon has long been neglected. Empiricists have rather focused on measuring and studying related concepts, such as judicial independence (e.g., Feld & Voigt, 2003; Linzer & Staton, 2015; Voigt et al., 2015) or the rule of law (Gutmann & Voigt, 2018; Kaufmann et al., 2011; World Justice Project, 2022). By systematically measuring and studying the (mis)match between constitutional rules and practice, Law and Versteeg (2013) made a seminal contribution to the literature on constitutional compliance. The importance of measuring, studying, and eventually better understanding constitutional compliance is, for example, underlined by the frequent referral to constitutional provisions in reviews of countries’ efforts to implement the Sustainable Development Goals. These discussions usually disregard whether the constitutional provision is part of a constitution that is generally respected by the executive, or whether it is unlikely to have any consequences in practice. A better understanding of the causes and consequences of constitutional compliance promises to lead to concrete policy recommendations and the monitoring of constitutional compliance by citizens is easier than, for example, that of the rule of law (Gutmann et al., 2022). Constitutional compliance is clearly normatively desirable when the constitutional rules complied with are substantively desirable. However, constitutional compliance can also be desirable on a more technical level – and at least to some extent independent of the substance of the law – in that it creates a predictable environment and makes government promises credible, both of which is economically favorable (see, e.g., Dreher et al., 2015; Voigt et al., 2007).
Here, we introduce the Comparative Constitutional Compliance Database, which includes indicators for governments’ compliance with 14 constitutional rules that can be grouped into four legal areas: (1) property rights and the rule of law (encompassing indicators for property rights, judicial independence, equality before the law, and rule of law); (2) political rights (freedom of association, freedom of assembly, and the right to form parties); (3) civil rights (free media, free speech, free movement, and religious freedom); and (4) basic human rights (the right to life, freedom from slavery, and protection from torture). The construction of our indicators involves matching information on de jure constitutional rules from the Comparative Constitutions Project (CCP; see Elkins et al., 2009) to data on these rules’ de facto implementation according to the Varieties of Democracy project (V-Dem; see Coppedge et al., 2022). Individual scores for these 14 compliance indicators, as well as aggregated measures for our four categories of constitutional rules and an overall compliance indicator across all rules, are calculated for 175 countries.
Our database improves on existing indicators for constitutional compliance in various ways. With over 10,000 country-year observations covering 175 countries over the period 1900 to 2020, it offers wide country and time coverage. One major advantage of our indicator construction is that all de jure and all de facto data, respectively, come from the same data source, avoiding problems of comparability due to differing coding rules and measurement strategies for compliance with different rights. Law and Versteeg (2013), for example, had to rely on five different data sources to measure de facto constitutional rules. The CIRI dataset (see Cingranelli & Richards, 2010) as their main source of de facto information covers various rights, but only for the years 1981 to 2010, and the unadjusted comparison of CIRI data over time can be problematic (Fariss, 2019). Although the World Bank data on literacy and life expectancy used by Law and Versteeg (2013) covers many countries and years, this information is a rather crude proxy for de facto constitutional rights and requires that the authors come up with coding rules for when constitutional promises are (not) being complied with based on a society’s socio-economic development. The V-Dem dataset allows us to avoid these problems.
Moreover, our database offers indicators for constitutional compliance in four different categories, which are based on the same data sources and follow the same coding rule, ensuring their comparability. Finally, by including V-Dem’s own expert coding of executive compliance with the constitution in our database, we make it possible to cross-validate results based on our indicators, using another indicator of constitutional compliance with an even larger country and time coverage. A serious downside of V-Dem’s compliance indicator is that it is constructed as a black box. No one knows the criteria used by experts to evaluate a government’s constitutional compliance and whether experts coding different countries share a common definition of what compliance is. Our new compliance indicators do not suffer from a comparable weakness.
The list of questions researchers will be able to answer based on our new data is long. To give just a few examples: Do traits of politicians (e.g., Gutmann et al., 2021b) or society’s culture (e.g., Gutmann et al., 2021a) influence a country’s level of constitutional compliance? Are populists more likely to violate constitutional rules (e.g., Gutmann & Rode, 2022)? Are violations promoted by political conflict or polarization (Lewkowicz et al., 2020)? Is constitutional compliance stable during extreme events (e.g., Choutagunta et al., 2022)? Do foreign investors reward constitutional compliance and does it spur economic growth?
The remainder of this article is structured as follows. In Sect. 2, we discuss limitations of existing indicators of constitutional compliance. Section 3 explains the construction of our new indicators of constitutional compliance. In Sect. 4, we use different compliance indicators to illustrate stylized facts of constitutional compliance across the world and its development over time. Finally, we conclude with an outlook on current and prospective research that can benefit from our database.
2 Existing indicators of constitutional compliance
Four approaches to measuring constitutional compliance are proposed in the literature. What separates these studies from the broader and swiftly growing literature that works with de jure and de facto indicators of constitutional rules is that they are explicitly measuring the gap between de jure constitutional promises and what is then implemented de facto. Most of these studies have been published only in recent years. Law and Versteeg (2013) produced the first dataset of what they call constitutional underperformance a decade ago.Footnote 1 They measure constitutional compliance with respect to 15 constitutional rights divided into three categories: personal integrity rights, civil and political freedoms, and socioeconomic and group rights.Footnote 2 Law and Versteeg collect their own de jure data by analyzing the contents of 729 national constitutions adopted by 188 countries between 1946 and 2010. The main source of Law and Versteeg’s de facto data is the CIRI database (Cingranelli & Richards, 2010), which provides data on 195 countries for the period 1981 to 2010.Footnote 3 Law and Versteeg create 15 de jure and 15 de facto indicators, which are then used to calculate one constitutional compliance score (corresponding to a country-year). Countries that protect a right in their constitution score one point for the de jure protection of that right, others get zero points. Countries that protect a right de facto get up to one point, depending on the level of protection, and countries that do not provide relevant protection of that right in practice score zero. The de jure-de facto gap, or “constitutional underperformance”, is calculated based on those rights that a country protects in its constitution. The indicator divides the total number of points a country scores on its de facto protection level of each de jure protected right by the number of points that could have been reached if all de jure promised rights were de facto fully protected. The resulting indicator ranges from zero to one and indicates the share of de jure promises in the constitution that are de facto upheld.
Metelska-Szaniawska (2021) uses a different approach to measure the de jure-de facto gap. Her study focuses on seven civil and political rights in post-socialist countries: freedom of movement, freedom of association, freedom of expression, freedom of the press, freedom of religion, prohibition of torture, and the right to habeas corpus. Its de jure data comes from the Comparative Constitutions Project (Elkins et al., 2009), while the corresponding de facto data is from the CIRI database (Cingranelli & Richards, 2010) and Freedom House (2015). As in Law and Versteeg (2013), all de jure and de facto rights are coded between zero and one, where one indicates (full) protection of a constitutional right. But unlike Law and Versteeg (2013), Metelska-Szaniawska (2021) calculates a combined indicator of constitutional compliance and constitutional overperformance by subtracting the sum of a country’s de facto scores from the sum of its de jure scores and then dividing by seven. Hence, this overall indicator takes on positive or negative values, depending on whether more rights are protected de facto or de jure, and it can range from -1 to +1. Given that we are only interested in measuring compliance with rules guaranteed in the constitution, the de jure-de facto gap as operationalized in Metelska-Szaniawska (2021) is not suitable for our purposes.
Mataic and Finke (2019) propose two versions of an indicator for the de jure-de facto gap in religious freedom. They use de jure data from the Religion and State Project’s constitutions dataset (Fox, 2012), capturing the constitutional protection of 21 religious freedoms, and two de facto indicators from the Religion and State Project’s main dataset (RAS3; see Fox, 2017), which measure 65 restrictions of religious practices imposed on all religions or only on religious minorities. For all constitutions that contain at least one constitutional promise of religious freedom, Mataic and Finke standardize the de jure and the de facto indicators and then subtract one of the de facto indicators from the de jure indicator. The resulting two indicators, one for compliance with religious freedoms of all religions and one specifically for religious freedoms of religious minorities, are only crude proxies for a de jure-de facto gap when compared to the indicators discussed so far. The de jure and de facto rights do not match, meaning that it remains unclear whether restrictions of religious practices are really contradicting constitutional rules. By standardizing their indicators before subtracting, Mataic and Finke (2019) also give individual de jure and de facto rights different weights in the construction of the indicators.
The fourth existing indicator of constitutional compliance, v2exrescon, is part of the V-Dem dataset and follows a very different methodology than the indicators discussed so far (Coppedge et al., 2022; Pemstein et al., 2022). Instead of measuring whether specific de jure constitutional rules are complied with de facto, v2exrescon is based on expert evaluations of whether members of the executive respect the constitution in general. Country experts are asked to rate constitutional compliance in a country-year on a five-point scale and the responses of various experts are then aggregated using an item response theory model (see Pemstein et al., 2022).Footnote 4 The major advantage of this indicator is that it is, in principle, not restricted to measuring compliance with a limited set of constitutional rules for which matching de facto information is available. However, the indicator also has two important disadvantages. First, like most perception-based governance indicators, it may be subject to various biases (see, e.g., Gutmann et al., 2020). Second, since constitutional compliance – without further concretization – is a highly abstract concept, expert evaluations of such a concept are produced in a black box. It remains unclear how the involved experts would define constitutional compliance and on what aspects of constitutional compliance they primarily base their evaluations.Footnote 5 To give just one example, it is unclear if the executive’s respect for the constitution should be judged based on the expert’s reading and interpretation of the constitution or if it only matters that national courts find a violation of constitutional rules. The fact that V-Dem experts do not always understand what they are supposed to code becomes clear when one inspects the time series of this indicator more closely. V-Dem claims to evaluate constitutional compliance even in years in which countries were not independent and had not promulgated a national constitution (e.g., Afghanistan pre-1919, Algeria pre-1962, Angola pre-1975, Bahrain pre-1971, Barbados pre-1966).
Of the four indicators measuring a de jure-de facto gap that we are aware of, the indicators by Metelska-Szaniawska (2021) and Mataic and Finke (2019) are not indicators of constitutional compliance in the narrow sense, because they conflate constitutional over- and underperformance. This leaves the indicator by Law and Versteeg (2013) and V-Dem’s v2exrescon as important reference points for our new constitutional compliance indicators. The V-Dem indicator should be treated with caution due to the subjectivity of expert evaluations of abstract concepts and the black box character of the evaluation process. The main weakness of Law and Versteeg’s (2013) indicator is the limited availability of de facto information at the time of its production, which forced the authors to draw on information from different data sources (making the data more difficult to compare and aggregate) and to come up with coding rules, for example, regarding which life expectancy corresponds to a de facto guaranteed right to health or which literacy rate indicates compliance with a right to education.
3 A new set of indicators of constitutional compliance
To establish a country’s level of constitutional compliance, we draw on information from two datasets. These datasets are ideal for our purposes because of their data quality, their unmatched breadth in terms of institutional characteristics covered, and their almost universal country and time coverage.
The Comparative Constitutions Project (CCP) provides data on the de jure rules in national constitutions. Elkins et al. (2009) collect and hand-code English translations of all national constitutions and most constitutional amendments worldwide since 1789. To minimize errors in coding, each constitutional event is evaluated twice by independently working coders. After that, a reconciler reviews each constitution to check for errors and to resolve coder discrepancies.
Data on the de facto compliance with constitutional promises identified in the CCP data comes from the Varieties of Democracy project (V-Dem; version 12). While CCP measures hundreds of characteristics of national constitutions, V-Dem measures hundreds of political institutions and other traits of the political and legal system. Over 3,500 country experts and dozens of social scientists involved in the project are supposed to ensure that their expert evaluations of de facto institutions are reliable and comparable across countries and over time (Coppedge et al., 2022; Pemstein et al., 2022). The V-Dem dataset is unmatched as a source of information on de facto institutional quality in terms of its coverage and the number of institutional details captured.
Matching the information provided by CCP and V-Dem, we arrive at 14 constitutional rules that are covered consistently by one or more indicators in each dataset. When there is more than one de jure indicator for a specific constitutional rule, it is sufficient that one of them is coded as “present” to establish a constitutional promise (e.g., the guarantee of media freedom or the prohibition of censorship). When there is more than one de facto indicator corresponding to a particular constitutional rule, then each one of them needs to indicate a sufficient level of constitutional compliance, such that the constitutional promise is considered to be kept (e.g., courts may not discriminate against men, women, the poor, and members of other social groups). Our indicators operationalize constitutional compliance as the absence of a “de jure-de facto gap”. Such a gap would imply that representatives of the different government branches are not following the rules specified in the country’s written constitution. What distinguishes our measurement approach from a typical de facto institutional quality indicator is, thus, that compliance or noncompliance with a constitutional rule is only possible if the constitution includes that rule in the first place. This means that the standard to which a country is held varies both over time and across countries and is determined by members of these societies – be it all citizens or only an elite – when they introduce or change their national constitution. We limit our database to the time period after 1900, which V-Dem calls the contemporary as opposed to the historical data, since our dataset would cover only few countries before 1900 and the data appears less reliable. Historical V-Dem data is, for example, based on fewer expert coders.
According to our main coding rule, compliance with a constitutional rule is coded 1 if that rule is protected both de jure and de facto. The compliance indicator is coded 0 if the right is protected de jure, but not de facto. In this case, we speak of a de jure-de facto gap (or, in the words of Law & Versteeg, 2013, constitutional underperformance). If a constitutional right is not protected de jure, we assign a value of 0.5, irrespective of the de facto indicator. The logic behind this coding rule is that countries start from a value of 0.5 and they can deviate from that value upwards or downwards only if they enter constitutional commitments.
De jure is coded based on whether a rule exists in the constitution. Sometimes it is sufficient that one of two alternative rules exists to speak of de jure protection. For example, it is sufficient that either freedom of opinion or freedom of expression is guaranteed to score 1 on de jure free speech. The de facto coding is based on ordered categorical indicators, which are typically based on a five-point scale where the highest category indicates full compliance with the rule. The second highest category indicates weaker enforcement of the rule, but – according to our interpretation – no intentional or systematic disregard by the government. We consider governments that score in one of the top two categories as de facto enforcing the rule and governments rated in a lower category as not complying with the rule. For example, it is sufficient for compliance with the freedom of assembly that “state authorities generally allow peaceful assemblies, but in rare cases arbitrarily deny citizens the right to assemble peacefully.” Sometimes we can rely on more than one V-Dem indicator to measure compliance with a constitutional rule (e.g., judicial independence both at the highest court and lower-tier courts). In these cases, the government must score in the top two categories of each of these indicators to be considered compliant. Table OA1 in the Online Appendix, available on the Review of International Organizations’ webpage, lists the exact criteria for the de jure and de facto coding of each of our 14 constitutional rules.
Having produced 14 individual rule indicators of de jure-de facto gaps, the challenging question is how to form broader indicators of constitutional compliance. We start by grouping the 14 indicators into four legal areas. “Property rights and the rule of law” includes four rules: private property rights, judicial independence, equality before the law, and rule of law. The category of property rights and the rule of law is not commonly used by legal scholars, but it is frequently employed by empirical social scientists (see, e.g., Gwartney et al., 2022; Miller et al., 2022). Moreover, Blume and Voigt (2007) identify it in their exploratory data analysis as one central dimension of human rights (see also Gutmann & Voigt, 2017). Mukand and Rodrik (2020) provide a political-economic theoretical underpinning for such a distinct category of rights. The second category, “political rights”, includes three rules: freedom of association, freedom of assembly, and the right to form parties. “Civil rights” (or “civil liberties”) includes four rules: free media, free speech, free movement, and religious freedom. Finally, the category “basic human rights” (or physical integrity and autonomy) covers three rules: right to life, freedom from slavery, and protection from torture. To aggregate the individual indicators within each of the four categories, we apply factor analysis and extract the first factor as our indicator of constitutional compliance in that legal category.Footnote 6 Factor analysis reveals that there is exactly one dominating factor representing a latent variable of constitutional compliance in each category. The advantage of factor analysis over, for example, calculating the mean value is that variables are weighted according to how representative they are of the indicators in that category. The four newly created compliance indicators are called cc_prop, cc_polit, cc_civil, and cc_basic.
Our preferred coding rule differs from that applied by Law and Versteeg (2013) who measure compliance as the share of the de jure protected constitutional rules that is enforced de facto (see Sect. 2). For comparison, we construct additional indicators and apply the same coding rule as Law and Versteeg to aggregate our 14 gap indicators in the four legal areas named above. To that end, we simply recode all 14 indicators as missing (instead of 0.5) when the constitution does not guarantee the respective rule de jure. Then we take the mean value over the non-missing gap indicators within each of the four categories. The resulting indicators are called cc_prop_lv, cc_polit_lv, cc_civil_lv, and cc_basic_lv.
Finally, we take the mean value of the four legal area-specific compliance indicators according to each coding scheme to generate two overall constitutional compliance indicators, cc_total and cc_total_lv. For comparison, we add to our database the expert evaluation of constitutional compliance, v2exrescon, produced by V-Dem. All indicators are then standardized to have a mean of zero and a standard deviation of one across the entire dataset, which spans the time period 1900 to 2020. This is particularly important for comparing the indicators in the following section. Higher values on an indicator reflect higher constitutional compliance.
Although based on the same data, cc_total and cc_total_lv, as well as their respective subindicators, can lead to very different evaluations of constitutional compliance in individual country-years, if a constitution includes only very few of the rules for which we measure compliance. This is because cc_total_lv ignores how many constitutional rules exist and it only measures the share of the existing ones that are complied with. cc_total, in contrast, ensures that compliance is scored lower if one right is promised and complied with, compared to a situation where five rights are promised and all of them are complied with. Vice versa, cc_total gives a higher compliance score if one de jure rule exists and that rule is not complied with, compared to a situation where five constitutional rules exist but none of them are complied with. As the number of relevant de jure rules included in the constitution approaches 14 (i.e., the maximum), cc_total and cc_total_lv converge to the same score.
Descriptive statistics are shown in Table 1 and bivariate correlations in Table 2. The high correlation between cc_total and cc_total_lv (\(r=0.92\)) shows that the coding rule employed by Law and Versteeg (2013) does not lead to a very different evaluation of constitutional compliance than our preferred coding rule. The correlations of both indicators with the expert rating by V-Dem (v2exrescon, \(r\approx 0.7\)) indicate that they are far from providing identical information to that collected from the country experts.Footnote 7 Table 2 also shows the correlations with a de jure index that counts how many of the 14 rules of interest are guaranteed in the constitution. Since none of the constitutional compliance indicators are significantly correlated with the de jure index, it appears that a high number of constitutional promises as such is not detrimental to constitutional compliance.
4 Stylized facts
In this section, we present stylized facts regarding correlates of and time trends in constitutional compliance. Furthermore, we compare the different compliance indicators in our dataset to highlight their differences, while also demonstrating their validity.
Figure 1 illustrates the global time trend in constitutional compliance based on the three indicators of overall compliance and our four subindicators for compliance in specific legal areas. First, we observe that the trends of cc_total and cc_total_lv are almost identical at the global aggregate level, in spite of the different aggregation rules they are based on. Both indicators suggest an increase in global constitutional compliance over time, especially around the year 1990, i.e., the end of the Cold War. Interestingly, v2exrescon captures similar short-run fluctuations, but it does not show the same long-term trend. The trends depicted in the diagram on the right show that compliance with different types of constitutional rules follows a similar global trend. Figure OA1 in the Online Appendix illustrates that not only constitutional compliance, but also the number of rights protected by the constitution has increased over time.
Figure 2 shows regional trends in the three indicators of overall constitutional compliance based on the United Nations’ classification of world regions. The different indicators show very similar levels and time trends across world regions. However, V-Dem experts judge constitutional compliance in Asia significantly more positively than our indicators. At the same time, our indicators show a more sizable increase in compliance in the Americas. Africa, the Americas, and Europe experienced a significant uptick in compliance around the year 1990. As of 2020, European countries show the highest levels of constitutional compliance, whereas constitutional compliance in Africa and Asia is lower than in other parts of the world. One might attribute this pattern to differences in economic development across continents, but the bivariate correlation between income per capita and compliance is rather small (\(r=0.26\)). This means that constitutional compliance is not a privilege enjoyed only in developed countries. Figure 3 shows regional trends for post-socialist countries and Middle Eastern countries, as well as for landlocked versus non-landlocked countries, as defined by Raciborski (2008). Notable is the substantial disagreement between our compliance indicators and v2exrescon concerning the Middle East. V-Dem experts rate constitutional compliance in the Middle East only slightly below the global average, whereas our indicators rate the Middle East as having some of the lowest compliance levels in the world. One explanation could be the prevalence of Islamic constitutions in the region (Gouda & Gutmann, 2021; Gutmann & Voigt, 2015). Figures OA2 to OA4 in the Online Appendix graphically illustrate the global distribution of cc_total and its subindicators across time and space. Whereas the 1950s and 1970s saw predominantly countries with low compliance scores, the distribution today appears bimodal and includes both, countries with low and high compliance scores.
Table 3 shows the five countries with the highest and lowest constitutional compliance according to cc_total during the same decades as in Figure OA2. The best-performing countries in all decades are democracies as classified by Bjørnskov and Rode (2020). The worst-performing countries are, with only two exceptions, not democratic. Not surprisingly, the bivariate correlation between democracy and cc_total (\(r=0.72\)) is much higher than that with income per capita.
Figure 4 shows constitutional compliance and its evolution over time for six selected country cases. China, India, and the United States are the three most populous countries in the world. The cases of Hungary, Poland, and Turkey have been intensively discussed in recent studies of democratic transition and backsliding. The three indicators of constitutional compliance draw very similar patterns for Poland and the United States, both of which experienced significant declines in constitutional compliance in recent years. However, the decline in compliance in the U.S. measured by cc_total is smaller than that measured by the other indicators. Regarding Hungary and India, there are some differences between cc_total and v2exrescon. cc_total indicates a dramatic decline in constitutional compliance in India over the last decade, which v2exrescon does not show. Hungary under communist rule is evaluated significantly better according to v2exrescon than cc_total would suggest and also the decline under Viktor Orbán’s government is less pronounced as measured by v2exrescon. Unlike v2exrescon, cc_total indicates an increase in Hungary’s constitutional compliance in 1990 that is comparable to that of Poland. However, all indicators agree that constitutional compliance dropped significantly over the last decade in both Poland and Hungary. Finally, cc_total and v2excrescon disagree considerably on how to rate constitutional compliance in China and Turkey. cc_total indicates a decline in constitutional compliance in China over the past century, whereas v2excrescon suggests dramatic improvements over time. In Turkey, both indicators capture the slumps in constitutional compliance following coups as well as its erosion under ongoing AKP (Justice and Development Party) rule. However, v2excrescon indicates high levels of constitutional compliance starting in the mid-1980s, whereas cc_total captures the dramatic positive effect of the democratic reforms in the early 2000s starting from a very low compliance level. As a general conclusion, it seems that V-Dem country experts sometimes rate the constitutional compliance of nondemocratic regimes more generously than V-Dem’s de facto data used in the construction of our compliance indictor seem to justify. This might be because V-Dem coders focus on different constitutional rules than the ones that enter cc_total or they might hold these countries to a different standard. It appears difficult to justify why, according to v2exrescon, China today should score higher on constitutional compliance than the United States.
Figure 5 shows constitutional compliance plotted for six different regime types according to Bjørnskov and Rode (2020) who adopt their classification from Cheibub et al. (2010). All three indicators of constitutional compliance show similar ratings for parliamentary, mixed, and presidential democracies, although v2exrescon does not show an improvement in presidential democracies after 1990. The indicators show less agreement on how to rate nondemocracies. v2exrescon indicates that civilian dictatorships before 1990 and monarchies from 1950 onwards exhibit higher constitutional compliance levels than measured by cc_total. Overall, parliamentary and mixed democracies show the highest levels of constitutional compliance among all regime types, whereas military dictatorships perform the worst; and, consistent with the cases presented in Table 3, democracies clearly outperform nondemocratic regimes.
Figures 6 and 7 show constitutional compliance for countries categorized by their legal origin according to La Porta et al. (2008) and their colonial history as classified by Hadenius and Teorell (2007). French civil law countries show the lowest levels of constitutional compliance. Since 1990, German and Scandinavian civil law countries clearly show the highest levels of constitutional compliance. It can also be seen that countries that were never colonized exhibit higher levels of constitutional compliance than any former colonies. Former Spanish colonies have made the greatest improvements in constitutional compliance between 1950 and 2000, whereas former British and French colonies appear to have stagnated.
So far, we have focused on the relationship between domestic factors and constitutional compliance. Dreher et al. (2012) argue that governments’ respect for physical integrity rights is higher in more open societies. Indeed, the KOF Globalisation Index (Dreher, 2006; Gygli et al., 2019) is positively correlated with constitutional compliance (\(r=0.54\)). Finally, we study whether four constitutional traits coded by CCP (Elkins et al., 2009) are associated with higher or lower compliance levels with the respective constitutions. The first trait is that the constitution allows for the dismissal of either the head of state or the head of government in case they violate constitutional rules. This rule is clearly designed with the intention to enhance constitutional compliance by making it easier to hold political leaders accountable for violating other constitutional rules. The second constitutional trait is that political office holders are obliged to swear an oath to abide by the constitution. The third trait is that a constitution identifies a model from a foreign government on which the constitution is based. As legal transplants tend to be detrimental to the effectiveness of law (Berkowitz et al., 2003a, 2003b), it can be expected that constitutions that borrow heavily from other countries might also be less complied with. Obviously, this indicator does not identify all legal transplants in constitutions and the identified legal transplants may not be representative of constitutional legal transplants in general. The fourth trait is the log-length of the constitution in words. It has been argued that constitutions in societies with high levels of generalized trust do not need to spell out as many details as those in low-trust societies (Bjørnskov & Voigt, 2014; Voigt, 2009).
Table 4 shows regression model estimates for these four constitutional traits, controlling for country- and year-fixed effects and our de jure index. We find that constitutional compliance is significantly higher if the constitution allows for the dismissal of the head of state or the head of government for violating constitutional rules. In contrast, the obligation to swear an oath to uphold the constitution does not show a statistically significant coefficient estimate. These results indicate that constitutional design can increase constitutional compliance when costly sanctions for violating a constitutional rule are threatened. We find no evidence that constitutions that are longer or based on a model from a foreign government are complied with less. Although we already account for country- and year-fixed effects, future research might reevaluate these results based on a more elaborate identification strategy.
5 Conclusion
This article introduces the Comparative Constitutional Compliance Database, a resource for researchers interested in studying the causes and consequences of constitutional compliance. Our analysis reveals that constitutional compliance improved significantly over the last 120 years, but most of this change took place around the year 1990, i.e., at the end of the Cold War. Consistent with that observation, there is no statistical association between a country’s income per capita and compliance, but democracies exhibit significantly higher compliance levels. Although it is often claimed that common law countries have superior judicial institutions, these countries are clearly outperformed by German and Scandinavian legal origin countries since the fall of the Iron Curtain.
Our results also indicate that former colonies are associated with less constitutional compliance, as the literature on legal transplants (Berkowitz et al., 2003a, 2003b) would suggest. It seems that constitutions reflecting the mores, values, and norms of the societies they are supposed to structure have a higher chance of being implemented. Choutagunta et al. (2022) show that CIA interventions, but not KGB interventions, reduce the level of constitutional compliance in a country. This kind of foreign influence, however, typically does not lead to a new constitution. It might be insightful to study the effect of foreign influence in the process of constitutional design in future research.
Cultural factors might also influence constitutional compliance. Gutmann et al. (2021a) ask whether specific cultural traits are conducive or detrimental to constitutional compliance. They find that countries inhabited by citizens that are more individualistic and have less respect for social hierarchies are more likely to experience constitutional compliance. It is also likely that constitutional compliance does not only depend on the general traits of society, but also on those of political leaders. Gutmann et al. (2021b) show that leaders with military experience and those who came into office irregularly are less likely to comply with the constitution. Gutmann & Rode (2022) show that populist governments comply less with the constitution.
Recent studies and the stylized facts in this article have focused on the causes of constitutional compliance. Yet, questions concerning the effects of constitutional compliance largely remain unanswered. Future research will use our database to study questions such as whether governments that comply with their constitution are also more likely to comply with international agreements, and whether this allows them to conclude more international agreements. Another important question would be if constitutional compliance affects the stability of political regimes. According to an influential argument by Weingast (1997), violations of constitutional rules could help citizens to coordinate on resistance against an exploitative government. Finally, the economic consequences of constitutional compliance need to be better understood. Do foreign investors, for example, reward constitutional compliance in general (see, e.g., Farber, 2002) or only compliance with property rights and the rule of law?
While the Comparative Constitutional Compliance Database in its current form will allow researchers to produce numerous insights on the causes and consequences of constitutional compliance, it is still only based on 14 types of constitutional rules and efforts should be undertaken in the future to create broader indicators of constitutional compliance.
Data availability
The database introduced in this article (version 2.0) and its future updates can be downloaded at: https://dataverse.harvard.edu/dataverse/CCCD. Replication materials for the figures and tables in this article are available at: www.jerg-gutmann.de/data.
Notes
Law and Versteeg (2013) also develop an indicator of constitutional overperformance, i.e., the level of de facto protection of rights that are not guaranteed in the constitution. While this phenomenon is not relevant to the question of constitutional compliance, its determinants have been studied elsewhere, e.g., by Metelska-Szaniawska and Lewczuk (2022).
These are prohibition of arbitrary arrest or detention, prohibition of torture, right to habeas corpus, fair trial rights, prohibition of the death penalty, freedom of assembly or association, freedom of movement, freedom of religion, right to vote, freedom of the press or expression, right to health, right to education, gender equality in marriage, gender equality in labor relations, and protection of minority rights.
Law and Versteeg draw on four more sources of de facto data: the Minorities at Risk Project, Amnesty International, Hathaway (2002), and the World Bank’s World Development Indicators.
The response categories range from “members of the executive violate the constitution whenever they want to, without legal consequences” to “members of the executive never violate the constitution”.
This is not a criticism of all expert evaluations of policy and institutions (and therefore of V-Dem indicators in general), but specifically of expert ratings based on very broad, unspecific questions that could be interpreted in myriad ways, as is the case for the discussed indicator. Our criticism is, thus, not inconsistent with relying on other, more reliable indicators from the V-Dem project.
This aggregation strategy is similar to that of Gygli et al. (2019). However, we do not allow the weights of individual variables to vary over time. For some research questions, a distinction between civil and political rights is not helpful. Thus, our online database also contains an indicator that combines both.
The bivariate correlation between Law and Versteeg’s (2013) constitutional underperformance indicator in 2010 and either cc_total or v2exrescon is 0.71, but that with cc_total_lv is 0.77.
References
Berkowitz, D., Pistor, K., & Richard, J.-F. (2003a). Economic development, legality, and the transplant effect. European Economic Review, 47(1), 165–195.
Berkowitz, D., Pistor, K., & Richard, J.-F. (2003b). The transplant effect. American Journal of Comparative Law, 51(1), 163–203.
Bjørnskov, C., & Mchangama, J. (2019). Do social rights affect social outcomes? American Journal of Political Science, 63(2), 452–466.
Bjørnskov, C., & Rode, M. (2020). Regime types and regime change: A new dataset on democracy, coups, and political institutions. The Review of International Organizations, 15(2), 531–551.
Bjørnskov, C., & Voigt, S. (2014). Constitutional verbosity and social trust. Public Choice, 161(1–2), 91–112.
Blume, L., & Voigt, S. (2007). The economic effects of human rights. Kyklos, 60(4), 509–538.
Cheibub, J. A., Gandhi, J., & Vreeland, J. R. (2010). Democracy and dictatorship revisited. Public Choice, 143(1–2), 67–101.
Chilton, A., & Versteeg, M. (2020). How constitutional rights matter. Oxford University Press.
Choutagunta, A., Gutmann, J., & Voigt, S. (2022). Shocking noncompliance? Effects of extreme events on constitutional compliance, mimeo.
Cingranelli, D. L., & Richards, D. L. (2010). The Cingranelli and Richards (CIRI) human rights data project. Human Rights Quarterly, 32(2), 401–424.
Coppedge, M., Gerring, J., Knutsen, C. H., Lindberg, S. I., Teorell, J., Alizada, N., Altman, D., Bernhard, M., Cornell, A., Fish, M. S., Gastaldi, L., Gjerløw, H., Glynn, A., Grahn, S., Hicken, A., Hindle, G., Ilchenko, N., Kinzelbach, K., Krusell, J., Marquardt, K. L., McMann, K., Mechkova, V., Medzihorsky, J., Paxton, P., Pemstein, D., Pernes, J., Rydén, O., von Römer, J., Seim, B., Sigman, R., Skaaning, S.-E., Staton, J. K., Sundström, A., Tzelgov, E., Wang, Y., Wig, T., Wilson, S., & Ziblatt, D. (2022). V-Dem country-year dataset v12. Varieties of Democracy Project.
Dreher, A. (2006). Does globalization affect growth? Evidence from a new index of globalization. Applied Economics, 38(10), 1091–1110.
Dreher, A., Gassebner, M., & Siemers, L.-H. R. (2012). Globalization, economic freedom, and human rights. Journal of Conflict Resolution, 56(3), 516–546.
Dreher, A., Mikosch, H., & Voigt, S. (2015). Membership has its privileges: The effect of membership in international organizations on FDI. World Development, 66, 346–358.
Elkins, Z., Ginsburg, T., & Melton, J. (2009). The endurance of national constitutions. Cambridge University Press.
Farber, D. A. (2002). Rights as signals. Journal of Legal Studies, 31(1), 83–98.
Fariss, C. J. (2019). Yes, human rights practices are improving over time. American Political Science Review, 113(3), 868–881.
Feld, L. P., & Voigt, S. (2003). Economic growth and judicial independence: Cross-country evidence using a new set of indicators. European Journal of Political Economy, 19(3), 497–527.
Fox, J. (2012). Constitutions dataset, 1990–2008. The Religion and State Project.
Fox, J. (2017). Main dataset and societal module, round 3. The Religion and State Project.
Freedom House (2015). Freedom of the press. Freedom House.
Ginsburg, T., & Huq, A. (Eds.). (2016). Assessing constitutional performance. Cambridge University Press.
Gouda, M., & Gutmann, J. (2021). Islamic constitutions and religious minorities. Public Choice, 186(3–4), 243–265.
Gutmann, J., & Rode, M. (2022). Are populists constitutionalists? An empirical assessment of populist constitutional compliance, mimeo.
Gutmann, J., & Voigt, S. (2015). The rule of law and constitutionalism in Muslim countries. Public Choice, 162(3–4), 351–380.
Gutmann, J., & Voigt, S. (2017). The heterogeneous effects of natural disasters on human rights, mimeo.
Gutmann, J., & Voigt, S. (2018). The rule of law: Measurement and deep roots. European Journal of Political Economy, 54, 68–82.
Gutmann, J., Lewczuk, A., Lewkowicz, J., & Voigt, S. (2021a). Culture and constitutional compliance, mimeo.
Gutmann, J., Metelska-Szaniawska, K., & Voigt, S. (2021b). Leader characteristics and constitutional compliance, mimeo.
Gutmann, J., Padovano, F., & Voigt, S. (2020). Perception vs. experience: Explaining differences in corruption measures using microdata. European Journal of Political Economy, 65, 101925.
Gutmann, J., Sarel, R., & Voigt, S. (2022). Measuring constitutional loyalty: Evidence from the Covid-19 pandemic, mimeo.
Gwartney, J., Lawson, R., Hall, J., & Murphy, R. (2022). Economic freedom of the world 2022 annual report. Fraser Institute.
Gygli, S., Haelg, F., Potrafke, N., & Sturm, J.-E. (2019). The KOF globalisation index – revisited. The Review of International Organizations, 14(3), 543–574.
Hadenius, A., & Teorell, J. (2007). Pathways from authoritarianism. Journal of Democracy, 18(1), 143–157.
Hathaway, O. A. (2002). Do human rights treaties make a difference? Yale Law Journal, 111(8), 1935–2042.
Kaufmann, D., Kraay, A., & Mastruzzi, M. (2011). The worldwide governance indicators: Methodology and analytical issues. Hague Journal on the Rule of Law, 3(2), 220–246.
La Porta, R., Lopez-de-Silanes, F., & Shleifer, A. (2008). The economic consequences of legal origins. Journal of Economic Literature, 46(2), 285–332.
Law, D. S., & Versteeg, M. (2013). Sham constitutions. California Law Review, 101(4), 863–952.
Lewkowicz, J., Metelska-Szaniawska, K., & Falkowski, J. (2020). Political conflict, political polarization, and constitutional compliance, mimeo.
Linzer, D., & Staton, J. K. (2015). A global measure of judicial independence. Journal of Law and Courts, 3(2), 223–256.
Mataic, D. R., & Finke, R. (2019). Compliance gaps and the failed promises of religious freedoms. Religion, State & Society, 47(1), 124–150.
Metelska-Szaniawska, K. (2021). Post-socialist constitutions: The de jure-de facto gap, its effects and determinants. Economics of Transition and Institutional Change, 29(2), 175–196.
Metelska-Szaniawska, K., & Lewczuk, A. (2022). Constitutional overperformance: An empirical study of de facto protection of rights with no de jure equivalents. European Journal of Law and Economics, 53(2), 289–317.
Miller, T., Kim, A. B., & Roberts, J. M., (2022) 2022 index of economic freedom. Heritage Foundation.
Mukand, S. W., & Rodrik, D. (2020). The political economy of liberal democracy. The Economic Journal, 130(627), 765–792.
Pemstein, D., Marquardt, K. L., Tzelgov, E., Wang, Y., Medzihorsky, J., Krusell, J., Miri, F., & von Römer, J. (2022). The V-Dem measurement model: Latent variable analysis for cross-national and cross-temporal expert-coded data. V-Dem Working Paper 21, 7th edition.
Persson, T., & Tabellini, G. (2003). The economic effects of constitutions. MIT Press.
Raciborski, R. (2008). kountry: A Stata utility for merging cross-country data from multiple sources. The Stata Journal, 8(3), 390–400.
Voigt, S. (2009). Explaining constitutional garrulity. International Review of Law and Economics, 29(4), 290–303.
Voigt, S. (2021). Mind the gap: Analyzing the divergence between constitutional text and constitutional reality. International Journal of Constitutional Law, 19(5), 1778–1809.
Voigt, S., & Gutmann, J. (2013). Turning cheap talk into economic growth: On the relationship between property rights and judicial independence. Journal of Comparative Economics, 41(1), 66–73.
Voigt, S., Ebeling, M., & Blume, L. (2007). Improving credibility by delegating judicial competence: The case of the Judicial Committee of the Privy Council. Journal of Development Economics, 82(2), 348–373.
Voigt, S., Gutmann, J., & Feld, L. P. (2015). The effects of judicial independence 10 years on: Cross-country evidence using an updated set of indicators. European Journal of Political Economy, 38, 197–211.
Weingast, B. R. (1997). The political foundations of democracy and the rule of the law. American Political Science Review, 91(2), 245–263.
World Justice Project (2022). Rule of law index 2022. World Justice Project.
Acknowledgements
Comments and suggestions by Yun-Chien Chang, Abishek Choutagunta, Jan Falkowski, Maria Grigoropoulou, Mazen Hassan, Mahdi Khesali, David Law, Anna Lewczuk, Jacek Lewkowicz, Katharina Luckner, Hashem Nabas, Evangelia Nissioti, Mohamed Rashwan, Roee Sarel, Karol Zdybel, as well as participants of the ICON-S 2021 Mundo Conference, the 39th European Association of Law & Economics annual meeting, the 8th Polish Law & Economics Conference, the 2022 European Public Choice Society annual meeting, and the 2022 Asian Law & Economics Association annual meeting are gratefully acknowledged. The authors thank Tom Ginsburg for sharing the latest data from the Comparative Constitutions Project.
Funding
Open Access funding enabled and organized by Projekt DEAL. This research is part of a Beethoven project funded by the Polish National Science Centre (NCN, #2016/23/G/HS4/04371) and the German Research Foundation (DFG, #381589259).
Author information
Authors and Affiliations
Contributions
Conceptualization – J.G. (33%), K.M. (33%), S.V. (33%); Methodology – J.G. (50%), K.M. (25%), S.V. (25%); Data Curation and Formal analysis – J.G.; Writing – J.G. (33%), K.M. (33%), S.V. (33%). The order of authors is chosen alphabetically.
Corresponding author
Ethics declarations
Competing interests
The authors have no competing interests.
Additional information
Responsible editor: Axel Dreher
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Gutmann, J., Metelska-Szaniawska, K. & Voigt, S. The comparative constitutional compliance database. Rev Int Organ 19, 95–115 (2024). https://doi.org/10.1007/s11558-023-09491-7
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11558-023-09491-7