Long live the doge? Death as a term limit on Venetian chief executives

We simplify “chief executive” to “executive” throughout. While our theoretical framework and case study focus exclusively on chief executives, we believe that it could be extended in future research to other positions.

Electoral costs could include adverse health consequences owing to stress (Chang and Meyerhoefer 2020).

Leguizamon and Crowley (2016) explore the relationship between the effects of the loss of electoral accountability associated with term limits and incumbent’s age; they provide evidence that term limits imposed on older incumbents have less effect on policy choice because such officeholders face already shortened time horizons.

Projected life expectancy may also be relied upon for executive lifetime appointments outside government, such as the Catholic Pope. We leave the exploration of that conjecture for future research.

We refer to the Venetian period studied herein, from 1172 to 1797, as “historic Venice” going forward.

The Venetian governance system was complex and did not provide any separation of powers. Governing councils and offices, such as the doge, held legislative, executive, and judicial authority. For a richer and historically nuanced discussion of historic Venice’s governance system, see Lane (1973), Madden (2012), and Gasparini et al. (2019). The doge, as Venice’s top executive, did not serve defined terms or face binding term limits, unlike other councils and magistracies with executive power.

See North (2005) and Williamson (2009) on the distinction between formal and informal institutions.

Informal mechanisms may become even more important as life expectancy increases.

As an important component of democratic institutions, the imposition of specified terms and term limits as a formal institution dates back to the ancient republics of Athens and Rome (Grofman 1996).

Term limits for legislative bodies likewise may be welfare-enhancing because they mitigate inter-district inequalities in legislative power, thus balancing the redistribution of resources (Dick and Lott 1993; Elhauge 1998; Friedman and Wittman 1995, 1996; Krehibiel et al. 1987; Petracca 1992; Gowrisankaran et al. 2008).

See Müller (2007) for a discussion of inefficiencies in a model with politicians who are either populist or policy success-seekers.

Another downside to formal term limits found in the literature unrelated to our case study is that they necessarily weaken electoral accountability by eliminating the possibility for reelection. Whether those consequences would be positive or negative is disputed. For instance, incumbents may use their last term to pursue policies that are not in the voters’ interests, but instead focus on personal gains (Cohen and Spitzer 1992, 1996). On the other hand, term limits may induce incumbents to focus on the long-run public interest in their final terms, since they will not face the voters again (Maskin and Tirole 2004; Ginsburg et al. 2011; Smart and Sturm 2013; Sieg and Yoon 2017). The loss of electoral accountability in the final term has received considerable scrutiny in the empirical literature, with mixed results (Alt et al. 2011; Bender and Lott 1996; Besley 2006; Besley and Case 1995a, b, 2003; DeBaker 2012; Lott 1987, 1990; Sieg and Yoon 2017). No consistent economic or policy distortion associated with term limits is found, and depends on the incumbents’ future prospects, including running for higher political office (Besley and Case 1995a; Lott and Reed 1989; Laband and Lentz 1985; Crowley and Reece 2013; Leguizamon and Crowley 2016).

The elections of pirate ship captains offer an example of such systems operating historically (Leeson 2007, p. 1067).

A formal model of our conceptualization detailed in this section is provided in Appendix A.

We thank an anonymous referee for detailed comments on the history elaborated in this section.

The people of Venice were referred to as the General Assembly, the popolo, the arengo, or the concio (Lane 1973, p. 91).

The group of 40 electors was retained as the ducal elector process evolved, but following a tied vote in 1229, the council was expanded to 41 electors (Coggins and Perali 1998, p. 711; Madden 2012, p. 170; Lane 1973, p. 94).

Calculations based on data from Gasparini et al. (2019). The shortest term was one day (Pietro Grimani in 1741) and the longest was 51 days (Pietro Ziani in 1205).

The oath can be traced back to at least 1192 with Doge Enrico Dandolo’s oath (Finlay 1980, p. 110; Madden 2012, p. 110). Following the tenure of Doge Jacopo Tiepolo in 1249, the oath of office was reviewed and updated after the death of each doge and enforced by state inquisitors (Bertolizio 2013, p. 205).

In fact, one could argue that informational asymmetries and their cost would both be higher for the dogeship than for other government positions.

Shortly after the Great Council moved into the grand chamber in 1366, it ordered Falier’s portrait to be covered by a curtain. When a fire destroyed the painting in 1577, the portrait was replaced by a painting of the curtain, along with the statement that Falier was “decapitated for his crimes” (Madden 2012, p. 210).

In particular, the Italian city-states of Genoa and Florence were susceptible to instability and tyranny during our period of study. While we are unable to fully explore the distinct histories of those city-states in the present paper, they do provide additional context for our central argument and likely also would serve as fruitful case studies of their own. We thank an anonymous referee for drawing the parallel.

The Venetian patricians likely would have had even more reason to adopt informal term limits if both the threat of internal tyranny and the threat of external military conquest varied over time.

Another interesting question raised by an anonymous referee is why the elected procurators of the charitable trusts of St. Mark, as lifetime appointees, were not subject to fixed terms or term limits. See Finlay (1980, p. xvi), Lane (1973, p. 98), and Mueller (1970) on the procurators’ role. Although it is an interesting question, because the position of procurator falls below the trusts’ top executive positions, the answer is beyond the scope of this paper. It is possible that the rank-order hypothesis discussed in Sect. 5.2 may be a factor.

It is important to note that the governance system of the Catholic Church also created a tendency for the selection of elderly candidates for pope (Tullock 2005, pp. 94-95).

Birth year and age data were unavailable for four of the 82 doges from 1172 to 1797 (see Appendix B).

The members of the Savi Grandi and Council of Ten were considered part of the inner governing circle (Lane 1973, pp. 256–257; Muir 1981, p. 20).

Doge Vitale II Michiel is included in the 742–1172 period; Doge Sebastiano Ziani marks the start of the 1172–1797 period.

Appendix C shows differences in age between the two periods using both linear and nonlinear predictions. A simple test for differences in means is statistically significant (t = 7.04, p = 0.000).

An additional concern that may have contributed to Gritti’s election is the 1522-1524 plague that swept through central and northern Italy, although it ended up not reaching Venice (Cohn 2010, p. 20).

It is worth noting that the accuracy of any data source depends on assumptions, methodologies, and the availability of records. All sources, naturally, attempt to be as accurate as possible. For instance, Phillips and Axelrod (2005) attempt to create a “representation of wars throughout history…”, rather than an “elusive comprehensiveness.…” Dupuy and Dupuy (1993, p. xix) stress that “[n]o work which deals with the activities of men of all nations and all parts of the world since the dawn of history can hope to be either completely accurate or totally comprehensive. Even if the authors were not subject to human error and human bias, they would be forced to rely upon sources which have all too many human shortcomings.…” Kohn (1999, p. vii) defines wars “fairly broadly, to mean an overt, armed conflict carried on between nations or states.…” Differences in what constitutes a party involved would, of course, change what is considered to be a military conflict for Venice.

Appendices D, E, and F provide the data for military conflicts assembled by Phillips and Axelrod (2005), Dupuy and Dupuy (1993), and Kohn (1999), respectively.

The weaker statistical evidence, despite the strong historical anecdotal evidence, may suggest that factors other than warfare may have affected the tenure-length calculus. Pietro Gradenigo, for instance, elected doge in 1289 at the age of 38, took office during a period when the people of Venice were threatening revolution by attempting to reassert their right to have a voice in the selection of the doge. The patricians, concerned with losing their control, knew that they needed “to find a man who should be at once devoted to their cause and yet of commanding power” (Brown 1907, p. 62). Major disease outbreaks likewise had the potential to influence electors. For instance, during the plague outbreak (and famine) of 1556 (Cohn 2010, pp. 20, 145), the patricians elected 67-year-old Lorenzo Priuli, the younger brother of the subsequently elected Girolamo Priuli, who was at the time 70 years old. Given that Girolamo was later elected doge after the untimely death of his younger brother, it can be surmised that the electorate purposefully selected the younger Priuli, owing either to his more youthful age or his experience, during a crisis period of rampant disease and famine (Norwich 1982, p. 460). During the 1595 election, Venice was “reeling from episodes of scarcity and disease” (van Gelder 2018, p. 265), leading to the election of 63-year-old Marino Grimani.

While interesting, we leave the question of why that was the case for future research.

The conjecture would provide an explanation for the above-mentioned shift in the utilization of age as term limit from 1172 to 1422, when a doge was elected with approval of the people, to 1423–1797, when the norm of getting the approval of the people was dropped.

The formal model in Appendix A can be augmented by accounting for governing ability, which itself can vary with age at election and expected tenure. In such a model, the electorate would weigh the benefits of ability alongside the other benefits and costs of executive tenure when choosing the optimal term length.

Dynasties, as defined by the passage of the dogeship within the same family at least twice, were common prior to 1032 (Puga and Trefler 2014).

Being too young also was an acceptable excuse (Queller 1986, p. 130).

We thank an anonymous referee for suggesting this alternative hypothesis.

We thank an anonymous referee for suggesting avoiding incompetent leaders as an alternative explanation.

As with the wisdom hypothesis, controllability can be included in a formal model of term length choice such as that found in Appendix A; in that case, however, ability would be inversely related to age.

Although causation could run in the opposite direction.

It is important to note that what we refer to as political stability is distinct from regime stability. Regime stability refers to the steadiness that comes with having the same top executive in a leadership role for an extended period. Political stability is continuity in governance institutions and legal framework. While a tyrant’s long rule may provide regime stability, it likely will result in political instability. The argument for term limits, both formal and informal, is that by limiting tenure lengths (and thus regime stability), democracies can foster political stability.

De Long and Shleifer classify Southern Italy as having been under absolutist rule during the full period. The Roman Republic, which predates our study period, of course, also succumbed to tyranny.

Pecquet (2017, p. 53) writes that “Roman emperors tended to have short reigns and short life expectancies due to assassination and intrigue. Few emperors had the luxury to even consider the long-run consequences of their policies, which produced institutional drift.” Also see Saleh (2019).

The Venetian Republic also was able to secure low interest rates on its floating debt. Such low interest rates could be explained by the vibrant economy and the republic’s stability. That observation raises the question of whether interest rates correlate with the ages of the elected doges. If the election of elderly doges raised lenders’ confidence, it is possible that one would observe higher interest rates during times when young doges were in power. In appendices H and I, we present data on interest rates on public debt for Venice during the fourteenth century from Mueller (1997) and Homer (1977). Although the data are not exhaustive, they represent the period when the Grain Office became the de facto state bank, which would allow the state to easily borrow money to cover extraordinary expenses, such as war (Mueller 1997). The data are scarce because of loose recording requirements but do show little correlation between interest rates and the doge’s age. Even in instances in which the data indicate otherwise, one must be careful, since younger doges were more frequent during wartime. Money typically was scarcer during wartime, driving up market rates in general.