Abstract
This paper models theocracy as a regime where the clergy in power retains knowledge of the cost of political production but which is potentially incompetent, quarrelsome, or corrupt. This is contrasted with a secular regime where government is contracted out to a secular ruler, and hence the church loses the possibility of observing costs and creates for itself a hidden-information agency problem. The church is free to choose between regimes—a make-or-buy choice—and we look for the range of environmental parameters that are most conducive to the superiority of theocracy and therefore to its occurrence and persistence, despite its disabilities. Numerical solution of the model indicates that the optimal environment for a theocracy is one in which the “bad” (high-cost) state is disastrously bad but the probability of its occurrence is not very high. Quantitative evidence on the rise of ancient Israelite theocracy and the current surge of Islamic theocratic fundamentalism provides surprisingly strong support for this prediction. Lastly, supportive evidence is suggested by two rare instances of a theocracy’s peaceful demise.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Many of the contributions in Ferrero and Wintrobe (2009) explore broader definitions, both in theory and historical evidence.
As the list just given also suggests, only some religions, not all, seem to be given to theocratic experiments; these include the main branches of Islam, Judaism, several varieties of Protestantism, and Buddhism. The Catholic contribution is exceptional and that of Orthodox Christianity entirely absent. Other non-European religions do not feature at all. This differential propensity to theocracy is discussed in Ferrero (2009).
By definition, a theocracy will suppress religious competition. Therefore a decentralized, competitive religious tradition like Protestantism can support a theocracy run only by one particular sect or church that happens to hold a monopoly in a particular time and place, as was the case at Muenster and Geneva, for example. By contrast, these conditions were not met in British North America, as discussed in the next section. On the other hand, Sunni Islam—another decentralized, though not competitive, religion—can manage a theocracy either half-way, like a traditional caliphate, or full-fledged, like modern Islamist revolutionary groups. On these and other cases see the discussion in Ferrero (2005, 2009).
Our source here is especially worth listening to as Thubten Norbu was the elder brother of the current Dalai Lama, and himself formerly the abbot of a great Tibetan monastery.
Assuming risk aversion would complicate the algebra but add no new insight. Effort aversion is sufficient to ensure an interior solution.
In principle K could be a parameter that is relevant only in one or the other state of nature. However, it is not clear why a corrupt cleric would not take advantage of the church in any state, nor why clerics’ incompetence or fractiousness should have a different impact on the church in different states. For want of strong intuitions to the contrary, introducing K in all states seems to be the least biased option.
Wages in each state would be determinate if we wrote the problem subject to two separate participation constraints, specifying that the agent’s utility in each state be no lower than his reservation utility plus K, as in problem (5) below. Here the constraints would both bind at the optimum. Otherwise, optimal effort levels would not be affected. Such a formulation would fit an interpretation of the state of nature as, in effect, a type of agent.
This is an ex-ante participation constraint, which in effect assumes that there are two types of agents but the principal is unable to tell the type before drawing up the contract. With an ex-post participation constraint (PC) the problem becomes uninteresting. For if we assume that the agent can take any course of action after observing the state’s realization, then the contract must ensure his participation whatever the state will turn out to be, which yields a single PC weighted by probabilities, as in problem (3) above, instead of the two PCs of problem (5). But then the revelation principle dictates that truth-telling must be ensured in any case, which implies that neither incentive constraint (iii) nor (iv) bind at the optimum. This allows the solution to be first-best while squeezing all the surplus out of the agent, so that the PC binds and the solution mimics (4a) and (4b). Proof is available from the author upon request.
Since in the solution to problem (3) wages are indeterminate within the bounds of the participation constraint, we substitute into expected benefits the wages that exactly equalize utilities across states. This would not be a neutral option but a requirement if wages were determined, as in the variant formulation of note 7 above.
Even though this objective function is not concave, it is continuously differentiable with respect to all variables; hence any solution within the compact set that satisfies the Kuhn-Tucker conditions will be a global maximum.
It is easy to rule out solutions on the upper borders of the parameter space. Inspection of conditions (4b) through (8b) shows that P=1, with any θ L , yields corner solutions to problems (3) and (5), i.e., \(e^{*}_{L} = \hat{e}_{L} =0\), while for θ L =θ H =1 (6b) collapses into (4b) and this violates (8b). In effect, in either case the choice problem itself vanishes.
I am indebted to Fabio Privileggi for this suggestion.
In this case, (8b′) implicitly defines a function P ∗(θ L ), which is the locus of all (P,θ L ) pairs that satisfy the equation. Taking (4b′′), (6b′′), and (8b′) and solving them numerically for \(e^{*}_{L}\), \(\hat{e}_{L}\), and P ∗ for given values of θ L , we find a steeply increasing, strictly concave P ∗(θ L ) curve that starts at P ∗=0 for θ L =0, rises to 0.64 for θ L =1/4, and to 0.88 for θ L =1/2. The theocratic optimum obtains at the crossing of the two curves \(\bar{P}( \theta_{L})\) and P ∗(θ L ), where P≈0.64 and θ L =0.25 (allowing for rounding).
An alternative source that covers the same period, 2 Chronicles, is considered much later and less reliable. It overlaps a great deal with the record in Kings but differs from it in places. If we used this source, we would have two bad kings according to Kings (Abijah and Manasseh) who were not so bad or changed their wicked ways, and two good kings according to Kings (Jehoash and Amaziah) who turned wicked at some point. The former two kings total 57 years of reign, the latter two total 52 years, hence using this alternative evaluation would not materially change the P values found in the text.
The territories may have had other names in colonial times and their borders may not exactly match today’s borders. The list includes the West Bank and Gaza due to its relevance but excludes Muslim-majority territories that are not independent or whose status is contested at the present time, such as Kosovo, Chechnya, and Kashmir.
I am indebted to Donald Wittman for bringing these two cases to my attention.
References
Ardussi, J. (2005). Formation of the state of Bhutan (‘Brug gzhung) in the 17th century and its Tibetan antecedents. In C. Cuppers (Ed.), Proceedings of the seminar on The relationship between religion and state in traditional Tibet. Lumbini international research institute, monograph series (Vol. 4). Accessible online through Wikipedia, “History of Bhutan”.
Baron, D., & Myerson, R. (1982). Regulating a monopolist with unknown costs. Econometrica, 50, 911–930.
Bawden, C. R. (1968). The modern history of Mongolia. New York: Praeger.
Cohen, S. J. D. (2006). From the Maccabees to the Mishnah (2nd ed.). Louisville: Westminster John Knox Press.
Eickelman, D. F. (1985). From theocracy to monarchy: authority and legitimacy in inner Oman, 1935–1957. International Journal of Middle East Studies, 17, 3–24.
Ferrero, M. (2005). Radicalization as a reaction to failure: an economic model of Islamic extremism. Public Choice, 122, 199–220.
Ferrero, M. (2009). The economics of theocracy. In M. Ferrero & R. Wintrobe (Eds.), The political economy of theocracy (pp. 31–55). New York: Palgrave Macmillan.
Ferrero, M., & Wintrobe, R. (Eds.) (2009). The political economy of theocracy. New York: Palgrave Macmillan.
Lane Fox, R. (1992). The unauthorized version. Truth and fiction in the Bible. New York: Knopf.
Lapidus, I. M. (1988). A history of Islamic societies. Cambridge: Cambridge University Press.
Myerson, R. (1979). Incentive compatibility and the bargaining problem. Econometrica, 47, 61–74.
Norbu, T. J., & Turnbull, C. (1972). Tibet. Its history, religion and people. Harmondsworth: Penguin.
Rose, L. (1977). The politics of Bhutan. Ithaca and London: Cornell University Press.
Shanks, H. (Ed.) (1999). Ancient Israel. From Abraham to the Roman destruction of the Temple (revised and expanded edition). Washington: Biblical Archaeology Society
Wintrobe, R., & Padovano, F. (2009). Theocracy, natural spiritual monopoly, and dictatorship. In M. Ferrero & R. Wintrobe (Eds.), The political economy of theocracy (pp. 83–118). New York: Palgrave Macmillan.
Wittman, D. (2010). Strategic behaviour and organizational structure in religious bodies: how it impacts on the age of leaders. Unpublished, University of California Santa Cruz.
Worden, R. L. (1993). Bhutan. In A. M. Savada (Ed.), Nepal and Bhutan: country studies (2nd ed.). Washington: Federal Research Division, Library of Congress, US Government Printing Office. Accessible online through Wikipedia, “History of Bhutan”.
Wikipedia, entries: “Islam by country”; “List of Muslim majority countries”.
Acknowledgements
An earlier draft of this paper was presented at the 10th annual conference of the Association for Public Economic Theory, Galway, June 17–20, 2009, and at the 17th conference of the Italian Association for the Study of Comparative Economic Systems, Perugia, June 25–27, 2009. I am indebted to Fabio Privileggi and Gregory Pavlov for mathematical advice, to Donald Wittman for helpful references, to a referee and especially to the editor of this journal for very useful comments.
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
1.1 A.1 Derivation of the optimal contract with hidden information
Recall problem (5). First, we can ignore constraint (ii) because when constraints (i) and (iii) are satisfied it will be satisfied as well, as follows:
The first inequality is due to constraint (iii) while the last is due to constraint (i). The strict inequality in the middle is due to our assumption (2) that c θ <0 for all e. It follows that constraint (ii) will hold with strict inequality, i.e., the agent will earn a surplus in state θ H .
Secondly, we will proceed to solve the problem ignoring constraint (iv) and later show that any solution to problem (5) that ignores constraint (iv) also will satisfy it. Therefore by dropping constraints (ii) and (iv) problem (5) reduces to the following:
Letting (λ,μ)≥0 be the multipliers on constraints (i) and (iii) respectively, and assuming (w L ,w H )>0, the Kuhn-Tucker conditions for this problem can be written:
Conditions (A.2.1) and (A.2.2) together imply that μ=P>0 and λ=1. Hence, both conditions (A.2.5) and (A.2.6) hold with equality, i.e., both constraints (i) and (iii) must bind at an optimal solution.
Because of our assumptions that b′(0)>0 and that c e =0 for e=0 (2), conditions (A.2.3) and (A.2.4) cannot hold at e=0. Hence, both e L and e H are strictly positive at an optimal solution, which implies that both (A.2.3) and (A.2.4) hold with equality. Then, substituting μ=P and λ=1 into these conditions yields (6a) and (6b) in the text, which characterize the optimal values of e H and e L . Then, w L and w H are determined by constraints (i) and (iii), which hold with equality at the solution.
We now show that constraint (iv) is also satisfied at the optimal solution. The binding constraint (iii) yields
Since in the solution e H >e L , the assumption that c eθ <0 (2) implies
These two together yield constraint (iv) as a strict inequality.
The second-order conditions for this problem are cumbersome but straightforward and will not be reported.
1.2 A.2 Comparative statics
Implicit differentiation of (6b) in the text yields:
Implicit differentiation of (4b) and (6b) in the text, respectively, yields:
The signs of these derivatives are established using the second-order conditions for problems (3) and (5) (not reported), which require all of their denominators to be negative. In particular, (A.4) and (A.5) measure the change in effort in each model as θ L changes, fulfilling the respective FOCs throughout. Assuming (neutrally) that all third-order partials equal zero, yields \(c_{e\theta} ( e_{L}^{ *} \), \(\theta_{L} ) = c_{e\theta} (\hat{e}_{L},\theta_{L} )\), \(b''( e_{L}^{ *} ) = b''( \hat{e}_{L} )\), \(c_{ee}( e_{L}^{ *} ,\theta_{L} ) = c_{ee}( \hat{e}_{L},\theta_{L} )\). It immediately follows that \(d\hat{e}_{L} / d\theta_{L} > de_{L}^{ *} /d\theta_{L} >0\).
Rights and permissions
About this article
Cite this article
Ferrero, M. The rise and demise of theocracy: theory and some evidence. Public Choice 156, 723–750 (2013). https://doi.org/10.1007/s11127-012-9924-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11127-012-9924-1