Five Laws of Politics: A Follow-Up

Alfred G. Cuzán

doi:10.1017/S1049096519000519

Five Laws of Politics: A Follow-Up

Published online by Cambridge University Press: 23 April 2019

Alfred G. Cuzán

Show author details

Alfred G. Cuzán*: Affiliation:
University of West Florida

Article contents

Abstract
FINDINGS
DISCUSSION
CONCLUSION
SUPPLEMENTARY MATERIAL
Footnotes
References

Rights & Permissions

Abstract

In this follow-up to “Five Laws of Politics,” I show that the original estimates of electoral outcomes describing the five “laws” hold up with twice the number of elections and three times as many countries as in the original article. The analysis yields long-run equilibrium values for several electoral parameters that set limits to incumbent support in a democracy. The evidence lends additional support to the notion that elections in democracies appear as if governed by “laws of politics.”

Information

Type: Article
Information: PS: Political Science & Politics , Volume 52 , Issue 3 , July 2019, pp. 457 - 464
DOI: https://doi.org/10.1017/S1049096519000519 [Opens in a new window]
Copyright: Copyright © American Political Science Association 2019

In “Five Laws of Politics” (Cuzán Reference Cuzán and Cuzán2015), an analysis of 426 elections in 23 presidential and parliamentary democracies drawn from Organisation for Economic Co-operation and Development members and Latin America revealed the following invariant or nearly invariant patterns:

(1) Law of Minority Rule. On average, 75% of the electorate turned out to vote, of whom 42% marked their ballots for the incumbents—the president’s or prime minister’s party—which yields a support rate of less than one third of the electorate.
(2) Law of Incumbent Advantage. Incumbents win reelection 60% of the time.
(3) Law of Shrinking Support, also known as “the cost of ruling” (Budge et al. Reference Budge2012; Nannestad and Paldam Reference Nannestad and Paldam1999; Wlezien Reference Wlezien2017). The incumbents incur a loss of support between elections averaging 4 percentage points.
(4) Law of the 60% Maximum. The incumbent party candidate succeeded at crashing that ceiling in fewer than 3% of cases.
(5) Law of Partials. No single party or coalition of parties can harmonize the diversity of interests and opinions of the electorate. The democratic synthesis is produced by competing parties taking turns at governing (Budge et al. Reference Budge2012). On average, any one party serves two terms for a total of eight years in office.

The purpose of this follow-up article is twofold: (1) to present additional confirmatory evidence in support of the contention that there are, indeed, “laws” of politics^{Footnote 1}; and (2) to extract theoretical insights from the empirical findings. I argue that democracies appear to exhibit a general tendency toward equilibrium values in electoral outcomes. The findings derive from a dataset composed of twice as many elections (i.e., 971 versus 426)^{Footnote 2} and three times as many countries (i.e., 74 versus 23) as in the original article.^{Footnote 3} To be added to the original dataset, elections generally must have been held in countries in which for at least two decades ending in the most recent year their Freedom House (2018) ratings on political rights and civil liberties were no higher than 4 (i.e., Partly Free). However, in some instances—usually in Africa, where democracies are scarce—this rule was relaxed if no more than in one or two nonelection years one of the ratings reached no higher than five. Figure 1 displays the distribution of most countries across space and region.^{Footnote 4} Also, to avoid losing data on current administrations or governments, if an incumbent party was currently in its third or later term at the time of the analysis, its values on Terms, Outcome, Reign, and Maximum Vote were “right censored.” That is, its values on Terms and Reign are counted as of the time they were observed for the analysis because they will not be any lower.

The purpose of this follow-up article is twofold: (1) to present additional confirmatory evidence in support of the contention that there are, indeed, “laws” of politics; and (2) to extract theoretical insights from the empirical findings.

Figure 1

Map of Elections

FINDINGS

I now proceed to analyze the data, first as averages over the entire period and then longitudinally, across time.

Statics

The descriptive statistics of the electoral variables are shown in table 1. The first column of data presents values for the entire dataset. These should be compared to those displayed in table 1 in Cuzán (2015, 417). The averages have budged only slightly, although the variation around the mean values has widened—an understandable difference given the range of democratic development in the countries covered in the more inclusive series. The mean values showing so little difference demonstrates that the original findings are robust with respect to the number and diversity of cases. This suggests that adding more countries or elections is unlikely to alter the parameter estimates. One benefit is that it may bring into relief regional patterns and variations with greater confidence than can be justified at present. Only the few that appear unlikely to be diluted with additional data are mentioned in this article.

Table 1

Descriptive Statistics, All and by Type

Notes: ^aBoth parliamentary and presidential elections are included for France, Poland, and Portugal. Also, three times the Israeli prime minister was elected on a nation-wide vote, and those elections are classified as “presidential.”

On average, 75% of eligible voters turn out to vote, casting 40% of their ballots for the incumbents, which amounts to a “support rate” of not quite one third of the electorate (i.e., Law of Minority Rule). When victorious, the in-party typically (1) wins approximately 45% of the vote, which is around 4 percentage points higher than the opposition’s share when it prevails; (2) is reelected nearly 60% of the time; and (3) serves two terms lasting about eight years (i.e., Law of Incumbent Advantage). This advantage notwithstanding, incumbents lose about 4 percentage points per term (i.e., Law of Shrinking Support),^{Footnote 5} and in only about 3% of cases does the in-party top the Law of the 60% Maximum.^{Footnote 6}

Compared to their counterparts in presidential systems, incumbents in parliamentary democracies face a larger electorate (i.e., turnout is higher); are reelected more often but with a smaller share of the vote; are constrained by a lower ceiling on that vote (i.e., presidential systems break the 60% ceiling twice as often as parliamentary systems); but incur a smaller loss between elections. Moreover, when the opposition wins, it also takes in a lower fraction of the vote than those in presidential systems.^{Footnote 7} However, there is no difference in the incumbents’ share of the vote (win or lose), the number of consecutive terms they serve, or the length of their reigns.

As shown in figure 2, in almost two thirds of the cases, the vote share of incumbents is highest when they first are elected to office; another 25% peak in their first reelection; and only 10% crest in their second or later reelection. Figure 3 paints an even more dramatic picture, depicting a leftward-skewed distribution in which 80% of incumbents manage to obtain no more than 50% of the vote. Beyond that, their numbers fall precipitously. In fact, fewer than 10% take in more than 55% and, as mentioned previously, no more than 3% top 60% of the vote—with almost all of those outliers falling outside the developed democracies, in Africa, Asia, or Latin America. Actually, some in the top 3% reflect opposition boycotts; at the very next contested election, the incumbents’ share shrank drastically. Furthermore, in only nine countries does the Incumbent Vote average exceed 50%; all but one (i.e., Australia, 50.2%) are found outside of the developed democracies.^{Footnote 8}

Figure 2

Incumbents’ Highest Vote Share within Reign, by Election

Note: 0=first elected; 1=first reelection; 2=second; 3=later.

Figure 3

Incumbent Vote Frequency Distribution

Dynamics

The longitudinal behavior of these variables, shown in figure 4 (Turnout and Incumbent Vote) and figure 5 (Incumbent Winning Vote and Opposition Winning Vote), all slightly drift downward across time, although the R-sq. in all is small (≤0.05). Conversely, Incumbent Loss, Terms, and Reign are flat, as shown in figure 6. Turning to the developed democracies only, as a group they show no relationship between time and Turnout, Incumbent Loss, Terms, or Reign. However, figure 7 shows a secular decline among them in the Winning Vote for both the Incumbents and the Opposition, with the more pronounced decrease occurring in the former (R-sq.=0.19 and 0.07, respectively).^{Footnote 9}

Figure 4

Turnout and Incumbent Vote

Figure 5

Winning Vote: Incumbent and Opposition

Figure 6

Incumbent Loss, Terms, and Reign

Figure 7

Winning Vote in the Developed Democracies: Incumbent and Opposition

Although modest, the secular decline in Incumbent Vote in the developed democracies, where elections have been held continuously for three-quarters of a century or longer, may reflect a combination of a breakdown of old-party cohesion, erosion of an old-party brand, and an increasingly divided electorate, which makes it more difficult for erstwhile catchall parties to aggregate interests—probably not a good thing. Or, it may signal a more positive development because in most of these democracies one party was dominant for many years. Taken together, these changes may mean that the political arena in the developed democracies has become more competitive. This may be regarded as a good thing because vigorous competition for votes among parties is a hallmark of modern mass democracy (Budge et al. Reference Budge2012). Which of the two interpretations is closer to or encompasses a larger share of the truth of the matter remains to be seen.

DISCUSSION

The previous analysis suggests that in countries where there is sufficient freedom to contest them, election outcomes gravitate toward a set of parameters describing electoral equilibrium.^{Footnote 10} At equilibrium, the variables measuring incumbent support generally tend toward the following values: Support Rate (33%); Incumbent Vote (40%); Incumbent Winning Vote (45% in parliamentary, 50% in presidential); the ceiling they are highly unlikely to break (60%); Incumbent Loss in percentage points per term (-3 in parliamentary, -6 in presidential); their reelection rate (60% in parliamentary, 50% in presidential); Terms (2); and Reign (8 years). These parameters could be interpreted as the “natural” values toward which democracies incline, fluctuating up and down in the long run.^{Footnote 11}

If indeed there are natural or equilibrium values in electoral outcomes, then we would expect democracies that have been around a long time to be closer to them than those that have not. As shown in figure 8, this appears to be the case. The range of values around the overall average among the newer democracies (i.e., those with fewer elections) is 1.5 to 2.0 times as large as that of the older democracies. (Outliers above the mean are mostly African and those below it are a mix of Latin American and post-Communist European.) Furthermore, as shown in figure 9, all regional trend lines in Incumbent Vote and Incumbent Winning Vote converge. It is as if time acts as a funnel, drawing toward the equilibrium level what initially had been widely dispersed country outcomes. From figure 8, we can surmise that the threshold for reaching the equilibrium values appears to be somewhere between 10 and 15 consecutive elections. Held on average every four years, it may take about 40 years for the average values of at least some of the new democracies to approach equilibrium. Nevertheless, the process may proceed faster in some and slower in other democracies. It depends on the strength of friction encountered from electoral rules (Carey Reference Carey, Herron, Pekkanen and Shugart2018; Shugart and Taagepera Reference Shugart and Taagepera2017) and from country-specific contexts including political culture, recent history, and exceptional leaders (Colomer Reference Colomer2010).

Figure 8

Incumbent Vote by Number of Elections

Figure 9

Convergence among Regions

Be that as it may, the evolution might proceed more or less as follows. The initial elections may favor one party above all others due to, say, the prestige of its founding leader or its history in the fight for independence or founding of democracy. Following the death or retirement of the leader or particular failures in policy, infighting among rival claims to the presidency or the premiership rooted in ideas, interests, and personal ambition would ensue, causing a split.^{Footnote 12} Starting at the other end, political entrepreneurs would seek to merge ideologically close parties that divide the vote of a large minority or plurality of the electorate into a larger, more competitive organization or alliance. We do occasionally observe party splits, mergers, and short-term as well as enduring alliances in various countries from time to time.^{Footnote 13}

CONCLUSION

This analysis appears to uphold the claim that elections in democracies are governed by a number of “laws” of politics constraining the share of support and, therefore, the time that an incumbent party can expect to exercise power in a single stretch. Although these “laws” may be largely a function of mechanical or statistical features of democratic institutions (Lebo and Norpoth Reference Lebo and Norpoth2007; Norpoth Reference Norpoth2014; Shugart and Taagepera Reference Shugart and Taagepera2017; Stokes and Iverson Reference Stokes and Iverson1962), their operation in specific cases is mediated by particular electoral rules and country-specific contexts, which are the product of history, culture, and political leadership. From the analysis, we may draw a specific recommendation for institutional reform: more frequent elections should aid the equilibrium-seeking processes of democracies.^{Footnote 14} Parliamentary democracies should consider holding elections no less often than every three years and presidential democracies no more than every four years, including off-year elections for the legislature halfway through the presidential term, as in the United States.

SUPPLEMENTARY MATERIAL

To view supplementary material for this article, please visit https://doi.org/10.1017/S1049096519000519

ACKNOWLEDGMENTS

This article is much improved from earlier drafts thanks to comments and suggestions from anonymous reviewers and from Ian Budge, John Carey, Cal Clark, Jay Cost, Andrew Gelman, Adreas Graefe, Richard J. Heggen, Randall J. Jones, Jr., Charles Kroncke, Arend Lijphart, Vello Pettai, Stephen and Elaine Snyder, William B. Tankersley, Michelle H. Williams, and especially Josep M. Colomer and Rein Taagepera. Thanks also to Yamisle Roca for her assistance in gathering much of the data collected during 2018. I also am grateful to The University of West Florida for its generous support and benign academic environment. The usual caveat applies: any errors of fact or interpretation or deficiencies in presentation are entirely my own.

Footnotes

1. I put the word “law” inside quotation marks in deference to those who are skeptical that a relationship as strong as a scientific law is operative in politics. As in the original article, Wikipedia was the source of election data. On the use of this resource for academic purposes, see Brown (Reference Brown2011), Cassel (Reference Cassel2018), and Tomaszewski and Macdonald (Reference Tomaszewski and Macdonald2016).

2. Actually, the calculation of Incumbent Vote or Outcome omits 89 elections because they were the first in a country’s series (85) or because the previous winner was an independent candidate or head of a party that disappeared or did not contest the election. In a few cases, it was clear that the incumbent party had lost; this was noted in the Outcome, but the vote could not be determined. As discussed in the original article, some countries presented special challenges due to highly fractured party systems and changes of government between elections.

3. For the most part, the rules for including countries or elections in the dataset in Cuzán (Reference Cuzán and Cuzán2015) were retained in this follow-up article. On publication, a detailed summary of the rules and the data will be available at uwf.edu/acuzan, with a standing request for corrections of any errors found in the data. See the online appendix in which elections are grouped by country, region, and type.

4. The countries in figure 1 are colored by region, a feature that will appear in the online but not the printed edition of the article. The regions are Africa (10 countries), Asia (6), British Caribbean (9), Developed Democracies (19), European Post-Communist (14), and Latin America (16). (Incidentally, all figures were made using Tableau. Many thanks to the company for making the program available free of charge for educational purposes.)

5. Explanations for this “law” range from an instance of the well-known statistical law of regression to the mean Budge (Reference Budge2012) to Wlezien’s thermostatic model of the relationship between policy and public opinion (Soroka and Wlezien Reference Soroka and Wlezien2010).

6. Although these empirical estimates are inductively arrived at, Colomer (Reference Colomer2016) constructed a mathematical model that proceeds from a country’s population to assembly size. Then, taking into account district magnitude, he arrived at the number of parliamentary parties and the size of the largest party—usually but not always the prime minister’s party—which is 40%. Taagepera (Reference Taagepera2018) suggested a simple logical way for arriving at other estimates. Theoretically, the maximum value of Turnout is 100% (the actual in the data is 97% in Austria in 1949 and 1953). At the opposite end, it can be argued that no democracy could sustain itself for long if turnout fell below 50%. Taking the mean between the two extremes yields 75%, which is almost the same as the actual overall value shown in table 1 (Taagepera 2018). Indeed, if less than half of the electorate repeatedly showed up at the polls, it can be said that the body politic was suffering from something analogous to anemia. Happily, however, fewer than 5% of the observations of this variable fell below 50%, spread across 15 countries—most of which did so only once or twice, Colombia and Poland excepted. Therefore, although it is not an absolute or theoretical minimum, a 50% floor on Turnout seems about right. A similar calculation applied to Outcome yields a mean value of 0.5—again, this is close to the actual overall value. Regarding Terms and Reign, the theoretical minimum would be 1; however, Reign could be only a few days, weeks, or months—but there is no theoretical maximum for either variable. Taking the actual maxima (i.e., 13 for Terms and 54 for Reign), the geometric mean for each yields 3.6 (versus actual overall of 2.2) and 7.3 (versus 8.4), respectively. For his part, Gelman (Reference Gelman2018) suggested that the 60% maximum reflects the fact that any attempt to achieve more than that would take the party so far from the median voter that one or more rival parties would move in to its left or right, as the case may be. In fact, as shown herein, most incumbents do not receive much more than 50% of the vote; therefore, this represents a hard constraint and is in line with “the law of partials.”

7. All of these differences are statistically significant at 0.05 or less. Technically, statistical significance has no meaning because the elections analyzed constitute a population, not a sample. Nevertheless, in conformity with convention, it is reported here.

8. The opposite end, averaging less than 25% of the vote, is occupied almost exclusively (Guatemala is the only exception) by new, post-Communist European democracies characterized by highly fractured party systems (Cabada, Hlousek, and Jurek Reference Cabada, Hlousek and Jurek2014).

9. That the incumbents’ victory margin shrunk relative to that of the opposition does not negate the Law of Incumbent Advantage because the gap has not been closed and Terms and Reign show no change over the same period. Time may tell whether Colomer’s (Reference Colomer2012) pronouncement that the incumbent advantage had vanished turns out to have been prescient or in error.

10. On the need to establish constants in the social sciences, see Taagepera (Reference Taagepera2008).

11. Budge et al. (Reference Budge2012) argued that evaluating democratic performance requires tracking its behavior over time.

12. Budge et al. (Reference Budge2012) discussed factional infighting within political parties. The United States is an example: George Washington was the only president in the country’s history to have been unanimously elected (by the Electoral College). Even within his administration, struggles between Treasury Secretary Alexander Hamilton and Thomas Jefferson at State (in cahoots with James Madison in the House of Representatives) were a constant irritant. After Washington’s retirement, the splits between President John Adams and Vice President Thomas Jefferson and between Madison and Hamilton only worsened.

13. See Shugart and Taagepera (2017, 80–84) on the strategic uses of alliances under different electoral rules and number of parties in the legislature.

14. In “A Constitutional Framework for a Free Cuba” (Cuzán [Reference Cuzán2000], 2013), I recommended a three-year presidential term and that one third of the national congress be up for reelection every year. As it turns out, the latter idea was proposed by John Dickinson at the Constitutional Convention that met in Philadelphia in 1787. See Madison’s Notes on June 21 (available at Gutenberg.org.EBook #40861).

References

REFERENCES

Brown, Adam R. 2011. “Wikipedia as a Data Source for Political Scientists: Accuracy and Completeness of Coverage.” PS: Political Science & Politics 44 (2): 339–43.Google Scholar

Budge, Ian, et al. 2012. Organizing Democratic Choice: Party Representation over Time . Oxford: Oxford University Press.CrossRef Google Scholar

Cabada, Ladislav, Hlousek, Vit, and Jurek, Petr. 2014. Party Systems in East Central Europe. Lanham, MD: Lexington Books.Google Scholar

Carey, John M. 2018. “Electoral System Design in New Democracies.” In Oxford Handbook of Electoral Systems, eds. Herron, Erik, Pekkanen, Robert, and Shugart, Matthew Soberg, 85–111. Oxford: Oxford University Press.Google Scholar

Cassel, Mark K. 2018. “When the World Helps Teach Your Class: Using Wikipedia to Teach Controversial Issues.” PS: Political Science & Politics 51 (2): 427–33.Google Scholar

Colomer, Josep M. 2010. The Science of Politics: An Introduction. Oxford: Oxford University Press.Google Scholar

Colomer, Josep M. 2012. “Firing the Coach: How Governments Are Losing Elections in Europe.” Democracy and Society 10 (1): 1–6.Google Scholar

Colomer, Josep M. 2016. Personal communication with author, September 29.Google Scholar

Cuzán, Alfred G. [2000] 2013. “A Constitutional Framework for a Free Cuba.” Available at https://ssrn.com/abstract=2305141. This is a revised and extended version of a paper presented at the 2000 meeting of the Association for the Study of the Cuban Economy.CrossRef Google Scholar

Cuzán, Alfred G.Cuzán, Alfred G. 2015. “Five Laws of Politics.” PS: Political Science & Politics 48 (3): 415–19.Google Scholar

Freedom House. 2018. “Country and Territory Ratings and Statuses, 1973–2018.” Available at https://freedomhouse.org/sites/default/files/Country%20and%20Territory%20Ratings%20and%20Statuses%20FIW1973-2018.xlsx.Google Scholar

Gelman, Andrew. 2018. Personal communication with author, May 24–25.Google Scholar

Lebo, Matthew J., and Norpoth, Helmut. 2007. “The PM and the Pendulum: Dynamic Forecasting of British Elections.” British Journal of Political Science 37 (1): 71–87.CrossRef Google Scholar

Nannestad, Peter, and Paldam, Martin. 1999. “The Cost of Ruling: A Foundation Stone for Two Theories.” Aarhus, Denmark: University of Aarhus Department of Economics, Working Paper No. 1999-9. Available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=165709.Google Scholar

Norpoth, Helmut. 2014. “The Electoral Cycle.” PS: Political Science & Politics 47 (2): 332–35.Google Scholar

Shugart, Matthew S., and Taagepera, Rein. 2017. Votes from Seats: Logical Models of Electoral Systems. New York: Cambridge University Press.CrossRef Google Scholar

Soroka, Stuart N., and Wlezien, Christopher. 2010. Degrees of Democracy: Politics, Public Opinion, and Policy. Cambridge: Cambridge University Press.CrossRef Google Scholar

Stokes, Donald, and Iverson, Gudmund. 1962. “On the Existence of Forces Restoring Party Competition.” Public Opinion Quarterly 26 (2): 159–71.CrossRef Google Scholar

Taagepera, Rein. 2008. Making Social Sciences More Scientific: The Need for Predictive Models. Oxford: Oxford University Press.CrossRef Google Scholar

Taagepera, Rein. 2018. Personal communication with author, June 19–21.Google Scholar

Tomaszewski, Robert, and Macdonald, Karen I.. 2016. “A Study of Citations to Wikipedia in Scholarly Publications.” Science & Technology Libraries 35 (3): 246–61.CrossRef Google Scholar

Wlezien, Christopher. 2017. “Policy (Mis)Representation and the Cost of Ruling: US Presidential Elections in Comparative Perspective.” Comparative Political Studies 50 (6): 711–38.CrossRef Google Scholar