The development of parliamentary constraints on the executive was critical in Western European political history. External war is the most widely discussed stimulus for the development of states and political institutions.Footnote 1 Yet the consequences of war varied for parliamentary institutions. The most common argument in the literature is that war stimulated parliamentary development, which finds support in the late medieval period. Weak monarchs throughout Western Europe negotiated with elites to gain funds to fight external wars. The standard bargain was that parliaments would gain the right to levy specific extraordinary taxes, in return for granting the resultant revenues to the Crown.Footnote 2 However, centuries later, most European parliaments lost these powers. Scholars again highlight the importance of external wars during the Age of Absolutism, but posit the opposite effect. This time, wars stimulated a demand for professional standing armies. This provided a means for rulers to coerce revenues from elites instead of negotiating with parliaments.Footnote 3
These contrasting empirical experiences raise foundational questions about wars and parliaments. Theoretically, what are the mechanisms by which external wars influence the adoption of political constraints? And empirically, why did external wars vary in their effects on parliamentary delegation in Western Europe? In this article, we develop a general theoretical framework for the war–parliament nexus. Our formal model explains the conditions under which war can either promote or undermine parliamentary development. The empirical implications of the theory provide strategic foundations for waves and reversals of historical European parliaments.
We analyze a strategic interaction between a ruler and a domestic elite actor. They jointly face the prospect of invasion. They can reduce this external threat with a well-funded war effort, which entails the elite contributing taxes that the ruler spends on security. However, the ruler cannot fully commit to using the elite-contributed revenues to boost security, as opposed to expropriating the funds as private rents. The ruler can submit to parliamentary constraints that partially tie her hands against exploiting elites, but this choice creates a trade-off. The ruler benefits from her hands’ being tied if that induces elites to contribute taxes. However, sharing power not only limits the ruler's discretion to spend tax monies for purposes besides security, but also bolsters the bargaining position of the elite actor relative to the ruler. These constraining effects may dissuade the ruler from sharing power. Given this trade-off, we examine the equilibrium conditions under which a ruler delegates authority to parliament. In essence, delegation requires the ruler to value the potential revenue stream enough to outweigh the loss of bargaining power; meanwhile, the elite must provide tax funds if and only if the ruler concedes powers to parliament. We focus primarily on how the strength of the external threat affects both sides of this calculus.
We first recover a conventional implication: stronger external pressure makes the ruler more willing to delegate privileges to parliament. A well-funded war effort is always more likely to succeed than a poorly funded one, but a stronger external threat widens the gap in the corresponding probabilities of surviving. Thus, when facing a strong external threat, the greater value-added of gaining war funding makes the ruler more tolerant of the sacrifices needed to induce a tax contribution from the elite actor. Thus, if we hold all else fixed, war threats stimulate parliament by enhancing the ruler's willingness to delegate privileges.
The problem for the conventional logic, though, is that we cannot hold all else fixed. War threats affect not only the ruler's demand for revenue but also the bargaining position of the elite actor. Incorporating this consideration into the model yields a previously unrecognized tension. Stronger outsider threats increase pressure either on elites to fund the ruler or on the ruler to introduce constraints—but not both simultaneously. This sheds light on various ways in which external wars can undermine parliament.
In one scenario, war threats can undermine parliamentary development by increasing the bargaining leverage of the ruler vis-à-vis elites. For elites whose wealth is concentrated in immobile assets (like landed nobles), a stronger war threat increases their dependence on security provided by the ruler. For such elites, the outside option is to refuse funding. However, this option is toothless in the face of a strong invasion threat that imperils their immobile assets. They are compelled to provide funding regardless of whether the ruler delegates parliamentary powers. Thus, war threats can undercut the credibility of elite demands for parliamentary privileges in return for war funding. Facing a strong threat, the ruler may bypass fiscal constraints, knowing the elite's best option is to pay taxes anyway.
In another scenario, war threats undermine parliamentary development by creating the opposite tension: increasing the bargaining leverage of elites vis-à-vis the ruler. Elites whose wealth is concentrated in mobile assets (merchants, for example) have an outside option to exit, that is, to move, hide, or self-protect their assets in response to an invasion threat. A strong war threat boosts their leverage to demand parliament, but potentially by too much. Elites with a viable exit option will not fund a hopeless war effort. Thus, external war can undercut the willingness of elites to fund the ruler even when they gain parliamentary privileges.
After analyzing this baseline, we extend the model in two ways. First, we allow the ruler to use a standing military to coerce revenue from elites, rather than negotiate. Second, we allow wars to be offensive (seizing wealth from another country) rather than defending the homeland. These components do not fundamentally alter the core logic, but do shift the parameter values under which the ruler delegates to parliament. The coercion option provides the ruler with an alternative if elites are unwilling to fund, and offensive wars enhance the credibility of elites to refuse funding if they are not granted parliamentary privileges.
Overall, three conditions are needed for the ruler to delegate to parliament in equilibrium: ruler willingness, elite credibility, and elite willingness. Stronger external war threats make the ruler more willing to delegate to parliament in return for war funding. However, too strong an invasion threat breaks at least one of the elites’ incentive conditions. A strong threat either reduces the credibility of the demand for parliamentary powers in return for funding (if elite wealth is immobile), or diminishes the willingness of elites to fund a constrained ruler rather than exit (if elite wealth is mobile). This implication arises from a tension between the elite's credibility in rejecting funds for an absolutist ruler and the elite's willingness to fund a parliamentary ruler. When stronger outside threats promote one of these conditions, it tends to come at the expense of the other. This makes delegation to parliament difficult to sustain if the outside threat is very weak or very strong. Even between these extremes, the effect of a marginal increase in external war threats on parliamentary delegation is ambiguous. The relationship depends on additional parameters such as the credibility of parliamentary constraints, the cost of effective war efforts, and geostrategic position (which affects the offensive versus defensive nature of warfare).
We apply the mechanisms of the model to offer strategic foundations for waves and reversals in European parliamentary history. The first wave occurred between the thirteenth and sixteenth centuries, when most major states developed parliaments with veto power over extraordinary taxes. Often, elite willingness to supply war funding was relatively high, for two reasons. First, impoverished rulers lacked viable alternative options to gain funds if not from parliament, which bolstered the de facto credibility of parliamentary constraints. Although most parliaments lacked formal control over expenditures, rulers who misappropriated tax funds jeopardized their access to future parliamentary revenues. Second, relatively inexpensive wars lowered the amount that monarchs asked from elites. The model also helps account for aberrant cases during this period. In France, elite credibility failed because landed elites faced a strong defensive threat. In parts of Germany, elite willingness failed. Merchants in cities sought protection from expropriation along long trade routes, but territorial princes offered only weak protection. The merchants’ control of trade and naval expertise gave them a viable exit option in the form of trading leagues, most notably the Hanseatic League.
Between the sixteenth century and the French Revolution, Western Europe experienced a wave of parliamentary reversals. We focus on how the Military Revolution eroded elite willingness in many European countries. The leading states began to adopt professional standing armies, which replaced feudal levies supplemented with mercenaries. This and related changes in military technology made wars more expensive, which increased the amount that rulers demanded from parliament. Furthermore, standing armies could be used for domestic coercion, which expanded the range of ways in which a ruler could renege on their promises to parliament. These inauspicious conditions engendered gridlock and, subsequently, the permanent weakening of parliament in places like Brandenburg-Prussia. However, England avoided a similar fate, in part because of a more favorable geostrategic position that reduced the threat of defensive wars.
Overall, we adopt Levi's core premise that rulers “maximize [state revenue] subject to the constraints of their relative bargaining power vis-à-vis agents and constituents.”Footnote 4 Our core innovation is to consider how war threats affect the relative bargaining power between ruler and elites. We do so by assuming that war threats affect the elite's preference for funding the government relative to either of two outside options, which correspond to different types of elite wealth. By distinguishing types of wealth, we draw from Tilly's analysis of capital-intensive (mobile wealth) and coercion-intensive (immobile wealth) paths of European state formation. For each, however, we uncover a distinct logic for how external wars can undermine parliamentary authority. This contrasts with Tilly's overarching thesis that wars promoted militarization and centralization, which in turn produced bargaining over taxes and constitutional rule.Footnote 5
Key Concepts and Trade-Offs
We model a strategic interaction between two players, a ruler and a distinct elite actor.Footnote 6 These two actors bargain over the division of domestic wealth, which is threatened by an external actor. The ruler and elite can mitigate this threat by coming to an agreement on funding the government and spending the funds on security. Prospects for a successful deal depend on the institutional environment. The ruler can grant parliamentary representation to the elite actor, which (imperfectly) ties the ruler's hands against reneging on a promised division of spoils. The series of moves is: (1) The ruler chooses institutional constraints: ruling absolutely or delegating to parliament. (2) The ruler makes a tax demand, with an implicit promise to spend it on security; the elite actor can accept this demand or exercise an outside option. (3) If the elite actor accepts, the ruler may have a chance to renege. (4) Nature determines whether the external threat succeeds. Here we provide more motivation for the first two moves: the ruler's decision on parliamentary delegation and the elite actor's outside options.
Ruler's Choice of Parliamentary Institutions
We posit a trade-off for the ruler that reflects core contentions about authoritarian institutions. Power-sharing deals share spoils and reallocate power to make it costly for the ruler to renege.Footnote 7 We incorporate these elements by modeling the following benefits and costs for the ruler.
Delegating privileges to parliament raises the elite actor's incentives to fund the government, which benefits the ruler. In our formal analysis, we let q denote the probability that the ruler is bound to spend tax revenues on security. We assume that in a parliamentary regime, the ruler's hands may be tied against spending the funds irresponsibly. Formally, with probability 
$q = \bar{q} > 0$, the ruler must spend any funds from the elite on security, which reflects the improved bargaining position of elites in a parliamentary regime.Footnote 8 By contrast, when the ruler does not delegate to parliament, we interpret the tax demand as an ad hoc levy. The elite's ability to refuse funding arises entirely from their de facto sources of wealth, rather than from any de jure rights within a parliamentary body. Consequently, in an absolutist regime, the ruler always has leeway to spend funds as she desires, that is, q = q = 0. In sum, delegating to parliament provides greater assurance that the ruler will spend the funds responsibly, which may induce the elite actor to supply funding they would have refused for an absolutist government.
Yet delegating privileges to parliament also has drawbacks for the ruler. Delegating to parliament shares spoils, increasing the elite's share of domestic wealth from ω to 
$\bar{\omega }$. Concomitantly, this reallocation reduces the ruler's share. Furthermore, when the ruler delegates, she now faces positive probability 
$\bar{q}$ of being forced to spend funds on security even when she would rather do otherwise. Ex ante, the ruler may benefit from delegating, but ex post, she always prefers to have discretion to spend the funds as she sees fit.
Empirically, delegating to parliament could entail varying constraints. One historically common way to curtail absolutist rule was to give parliament veto rights over extraordinary taxes. Some share of domestic wealth was unambiguously in the royal domain (
$1-\bar{\omega }$), and societal elites unambiguously controlled some sources of wealth (ω). However, monarchs and elites disputed other sources of wealth, which comprised the gap 
$\bar{\omega }-\underline{\omega } $. Examples include manorial control following a death, proceeds from taxing peasants, and administrative control over specific taxes. Designating these sources of wealth as extraordinary taxes and giving parliament a veto over them enabled elites to control these disputed shares of wealth. This worsened the ruler's reservation value to not reaching a deal, lowering her share of societal wealth 1 − ω. And with restricted access to resources, the ruler had fewer opportunities to act arbitrarily without facing repercussions.Footnote 9 Regular parliamentary meetings also bolstered the bargaining position of elites by enabling them to act collectively in a centralized body.Footnote 10 These factors raise the probability of the ruler's using the funds responsibly, q.
For example, in Spain in the early seventeenth century, the Castilian Cortes granted the monarch revenues from a direct tax, the millones. The Cortes and councils from its constituent cities administered the tax themselves, which enabled them to profit from the corrupt collection of a tax they did not themselves pay.Footnote 11 Furthermore, elites could leverage their control over the tax to influence, at least somewhat, how the Crown spent it:
[Elites] insisted on retaining administrative control of the new taxes they conceded; they attempted to impose budgets to ensure that public monies were spent in the public good, and they demanded other conditions in the form of legislation and restrictions on the use of the royal prerogative, as their price for approving new subsidies. In effect, they began to distinguish between public revenues and the king's private funds and to exercise, if only in a limited and tentative fashion, the power of the purse.Footnote 12
Yet parliamentary veto powers over extraordinary taxes do not eliminate all the ways a ruler could expropriate funds granted by parliament. The ruler retains some means to raise revenues independent of parliament, including ordinary taxes, external rents, debasing the currency, or issuing new debt. The ruler also retains legal leeway to spend the money differently than promised. Overall, in the case of parliamentary veto powers over extraordinary taxes, 
$\bar{q}$ is positive but less than 1.
A stronger privilege is a parliamentary monopoly over revenue raising and expenditures. In this case, 
$\bar{q}\approx 1$, as the ruler never has an opportunity to allocate revenues in ways not sanctioned by parliament. Credible budget procedures create a favorable reversion option for parliament if the ruler fails to reach a deal (that is, high 
$\bar{\omega }$) and cut off ways to circumvent parliament.Footnote 13 England following the Glorious Revolution is an early case, and other major European states adopted these reforms in the nineteenth century.Footnote 14
After the model analysis, we discuss how 
$\bar{q}$ varied over time even before the rise of parliamentary powers over expenditures. In the medieval era, 
$\bar{q}$ was relatively high because rulers lacked nonparliamentary sources of funding. Later, the Military Revolution eroded the underlying credibility of parliamentary promises because access to a standing professional army gave rulers other ways to renege on deals. Overall, the value of 
$\bar{q}$ in a particular case reflects not only de jure parliamentary prerogatives but also broader factors that influence the de facto credibility of constraints on the executive.
Elite's Outside Options
We assume that the elite actor's possible course of action depends on the nature of its wealth. First, the elite can refuse to fund the ruler. Refusal protects the elite's wealth from government appropriation. However, by not moving or hiding its wealth, it remains at risk from the outsider. An elite with immobile wealth—such as land, which cannot be moved, easily liquidated, or hidden—would have refusal as its outside option.
The other outside option is to exit. By exiting, the elite keeps its wealth safe from both the ruler and the external threat, but at a cost: it retains only a fraction of that wealth. This option corresponds to an elite with mobile wealth, such as merchants, although we interpret asset “mobility” in broad terms. Most obviously, actors can physically move their wealth outside a ruler's domain. Alternatively, they could hide their wealth internally. Bates and Lien discuss how the mobility of capital affected the bargaining leverage of elites when negotiating taxation with kings in early modern Europe. Trade taxes could be “highly lucrative,” but also “they could be easily avoided.”Footnote 15 Tilly highlights that effective taxes on movables required high levels of monetization and surveillance.Footnote 16 These features provided ample opportunities for exit given the low bureaucratic capacity in historical European states.Footnote 17
A more extreme variant of exit was for elites (typically merchants) to form their own protection leagues and guard their wealth against transgressions by princes. Thus, in the broadest sense, what distinguishes the exit option from the refusal option in our model is simply whether the source of wealth leaves the ruler's domain, regardless of the precise mode by which the ruler loses the ability to either protect or expropriate the source of wealth. For example, in a case we discuss later, what made the exit option viable for the cities of the Hanseatic League was not simply that their wealth was mercantile, but also that they had the naval capacity and expertise to organize for self-defense without the protection of existing sovereigns.
Model Setup
1. Ruler's choice of parliamentary delegation
The ruler moves first and chooses whether to delegate privileges to parliament. If the ruler does not delegate, the elite's baseline share of domestic wealth is ω ∈ (0, 1). We normalize total wealth to 1, so the ruler's wealth in this case is 1 − ω. If the ruler delegates, the elite's baseline share of domestic wealth is a greater value, 
$\bar{\omega }\in ( {\underline{\omega } , \;1} ) $, leaving the ruler with 
$1-\bar{\omega }$. To save on notation, we denote the ruler's choice to delegate in terms of the elite's resultant wealth: ω = ω denotes absolutist rule, and 
$\omega = \bar{\omega }$ denotes delegation to parliament. When describing the equilibrium results, we say the state is “limited” when the ruler chooses 
$\omega = \bar{\omega }$.Footnote 18
2. Elite taxation
The ruler demands taxes τ > 0 from the elite to fund a military improvement. The value of τ is exogenously fixed, reflecting the cost of making a discrete improvement in the state's security. Examples include hiring a mercenary company, adopting new naval technology, or building a standing army.
The elite may either accept or reject the demand. If the elite accepts, then the tax is transferred to the ruler, and the game moves to step 3. If the elite rejects, then we say the elite has exercised its outside option. Depending on the nature of the elite's wealth (immobile or mobile), this option is to either refuse or exit, as described earlier.
3. Ruler's opportunity to renege
If the elite accepted the tax demand in the last step, Nature makes a move determining whether the ruler is bound to spend the tax money on security. With probability q ∈ [0, 1], the ruler's hands are tied, and she has no option to expropriate the taxes for her own consumption. We assume an absolutist ruler has no commitment power: q = q = 0.Footnote 19 A ruler who delegates has less freedom of action, and with positive probability will have no option but to spend tax monies on security: 
$q = \bar{q} > 0$.
If the ruler is not bound ex post, then she chooses between using the taxes for security or reneging on the deal with the elite. If the ruler abides by the deal, then the taxes are spent to improve the state's military capacity (see step 4). By contrast, reneging entails expropriating the tax revenue. In this case, the proceeds τ are added to her wealth, and the state's protection against outside threats remains low (see step 4). An elite with mobile wealth is assumed to exercise its outside option in this case, obtaining a share σ of its post-tax wealth, ω − τ.Footnote 20
4. External threat
At the end of the game, regardless of whether the elite paid taxes or exercised its outside option, an external war threat governed by a Nature move determines each player's final consumption. Let θ ≥ 0 denote the external threat's strength. This is the critical parameter in our analysis.
If the elite did not pay taxes, then the probability of surviving the external threat is relatively low: 
$\pi _{\rm L}( \theta ) = {1 \over {1 + \theta }}$.Footnote 21 We assume that the probability of survival remains at π L if the elite paid taxes but the ruler reneged on the promise to provide security.Footnote 22 The provision of security raises the state's power endowment by Δ > 0, resulting in a higher probability of survival: 
$\pi _{\rm H}( \theta ) = {{1 + \Delta } \over {1 + \Delta + \theta }}$. Accordingly, when describing the equilibrium results, we say the state is “strong” when the security boost is provided along the path of play. To ensure that the ruler is willing to provide security voluntarily if the external threat is great enough, we assume that the military benefit is sufficiently high relative to the costs: 
$\Delta > {\tau \over {1-\bar{\omega }}}$.
One important feature of the contest functions is that they satisfy the monotone likelihood ratio principle. As depicted in Figure 1, the ratio 
${{\pi _{\rm H}} \over {\pi _{\rm L}}}$ increases monotonically in θ.
Stronger external threat makes internal cooperation more important
Consumption
Suppose no external takeover occurs. If the elite has not exited the ruler's domain, then its consumption is ω, less τ if it paid taxes. If the elite has exited, it receives σ ⋅ ω if it did not pay taxes, or σ ⋅ (ω − τ) if it did. If the elite paid taxes and the ruler reneged, then the ruler's consumption is 1 − ω + τ. Otherwise, the ruler's consumption is 1 − ω.
If instead an outside takeover occurs, then the ruler receives 0. The elite also receives 0 except if it exited—either in lieu of paying taxes, or because the ruler reneged.
Figure 2 presents the game tree, and Appendix A.1 summarizes all notation in the model.
Game tree
Alternative outside options
In our main model analysis, to clarify the exposition of the key mechanisms, we focus one by one on a pure refusal option and a pure exit option for the elites. In supplementary analyses, we consider two alternative outside options that combine elements of each ideal type. Appendix B.1 considers a hybrid outside option in which some proportion of elite wealth is moved out of the country while the rest remains. Appendix B.2 considers the possibility that the value of exiting declines as external threats become stronger, reflecting a lack of safe harbors for mobile assets.
Three Conditions for Parliamentary Delegation
We begin the analysis by deriving three incentive compatibility conditions that are individually necessary and jointly sufficient for the ruler to delegate to parliament in the first stage of the game: ruler willingness, elite credibility, and elite willingness. In the next section, we analyze how the strength of the external threat affects each condition.
Ruler's Willingness to Delegate
We say ruler willingness holds when the ruler would rather delegate to parliament and receive taxes than choose absolutist rule and have the elite exercise its outside option. In other words, the potential revenue stream must provide enough value to the ruler to outweigh the loss of bargaining power.
If the ruler chooses absolutist rule and the elite takes its outside option, then the ruler gets a relatively large share of the domestic pie but with weak defense against the outsider, for an expected utility of π L(θ) ⋅ (1 − ω). If instead the ruler delegates and the elite accepts the tax demand, then there are two possible outcomes. First, the ruler spends the money on security. With probability
$\bar{q}$, the parliamentary constraints bind and this is the only possible choice. Alternatively, if the constraints do not bind, the ruler can nonetheless voluntarily use the funds to bolster security. In either case, the ruler's expected consumption is 
$\pi _{\rm H}( \theta ) \cdot ( {1-\bar{\omega }} ) $. If instead she reneges (when given the opportunity), the ruler's expected utility is 
$\pi _{\rm L}( \theta ) \cdot ( {1-\bar{\omega } + \tau } ) $, reflecting worse security against the outsider but more consumption in case the invasion fails. Ruler willingness holds when the ruler would rather delegate and gain funds—which may entail either always spending on security, or reneging when possible—than not receive funds.
$$\eqalign{& {\bf Ruler{\hbox -}willingness\;condition\colon } \cr & \pi _{\rm L}( \theta ) \cdot ( {1-\underline{\omega } } ) \le \underbrace{{\pi _{\rm H}( \theta ) \cdot ( {1-\bar{\omega }} ) }}_{{{\rm always\;spend\;on\;security}}}\,\,\,\,\,{\rm or} \cr & \pi _{\rm L}( \theta ) \cdot ( {1-\underline{\omega } } ) \le \bar{q}\cdot \!\!\!\!\!\!\! \underbrace{{\pi _{\rm H}( \theta ) \cdot ( {1-\bar{\omega }} ) }}_{{{\rm security\;when\;constraints\;bind}}} \!\!\!\!\!\!\!+ \, ( {1-\bar{q}} ) \cdot \underbrace{{\pi _{\rm L}( \theta ) \cdot ( {1-\bar{\omega } + \tau } ) }}_{{{\rm renege\;when\;possible}}}} $$This condition compares the ruler's best-case scenario under delegation (acceptance of the tax demand) to her worst-case scenario under absolutist rule (rejection). Therefore, if ruler willingness fails, there cannot be an equilibrium in which 
$\omega = \bar{\omega }$.Footnote 23
Lemma 1: If the ruler-willingness condition does not hold, then there is no equilibrium in which the ruler delegates to parliament.
It may appear counterintuitive that ruler willingness would ever fail to hold in the presence of an external threat. Overall expected domestic surplus is greater with an agreement than without because π H > π L, so a standard bargaining logic would seem to suggest that an agreement should be reached. However, imposing greater constraints shifts the share of domestic wealth away from the ruler, reducing the ruler's potential consumption on delegating. If this effect is large enough, the ruler prefers a larger share of a smaller pie. This means that a bargaining surplus result akin to that in Fearon does not apply to the ruler-willingness condition.Footnote 24 In effect, the elite's inability to commit to not exploiting its enhanced bargaining power can cause a deal to unravel.Footnote 25
Besides external threat strength, which we address in detail later, three additional factors shape the ruler's willingness to accept parliamentary constraints. The first is how much of her baseline wealth she gives up by delegating to parliament, formalized as the difference 
$\bar{\omega }-\underline{\omega } $. If the difference is negligible, then ruler willingness is certain to hold. By contrast, if the upfront cost of delegation is very high, then the ruler will prefer absolutist rule even if her tax demand is then sure to fail.
The second factor affecting ruler willingness is 
$\bar{q}$, the likelihood that the institutional constraints bind her decision ex post. We interpret this parameter as a measure of institutional strength. Whenever the ruler would not voluntarily spend on security after delegating, lower institutional strength promotes ruler willingness. The greater chance to expropriate raises the ruler's expected utility.
The third key influence on ruler willingness is Δ, the extent of the improvement in protection against the invasion threat when the ruler spends taxes on security. Naturally, the ruler is more willing to forgo consumption when the benefits of spending those funds on security are greater. We formalize these comparative statics in Remark A.1 in the appendix.Footnote 26
Ruler's choice to provide security
To analyze the elite's decisions, we must specify whether a ruler with funding will voluntarily provide security rather than expropriate. Holding fixed the initial choice of constraints, the ruler's expected utility from spending on security is π H(θ) ⋅ (1 − ω), compared to π L(θ) ⋅ (1 − ω + τ) from reneging. Therefore, the following condition determines whether the ruler will voluntarily spend on security:
$$\displaystyle{{\pi _{\rm H}( \theta ) } \over {\pi _{\rm L}( \theta ) }} \ge \displaystyle{{1-\omega + \tau } \over {1-\omega }}$$The left-hand side of this equation represents the proportional increase in the ruler's expected consumption from the security increase, holding fixed her share of the society's wealth. The right-hand side represents her proportional increase in consumption due to expropriating elite wealth, holding fixed the likelihood of resisting invasion. Thus, the ruler will voluntarily provide security when the proportional security improvement dominates the proportional increase in wealth from expropriation.
Stronger external threats bolster the premium to maximizing security, as illustrated in Figure 1. This raises the ruler's propensity to spend taxes on security even when institutional constraints do not bind. Also, because the security benefit is multiplicative, wealthier rulers (greater 1 − ω) have a greater incentive to spend on security, holding fixed the opportunity cost of lost consumption, τ. The following result summarizes these findings.
Lemma 2 (War threats promote security spending): Let the initial choice of ω be fixed. If institutional constraints do not bind, the ruler prefers to spend taxes on security if and only if 
$\theta \ge \hat{\theta }( \omega ) $, where 
$0 < \hat{\theta }( {\underline{\omega } } ) < \hat{\theta }( {\bar{\omega }} ) < \infty $.
Elite Credibility: Rejecting Demands by an Absolutist Ruler
We now analyze the elite's demand for constraints on the ruler's ability to spend tax monies. When is delegating privileges to parliament pivotal for inducing the elite to accept the ruler's tax demand? As we show later, in order for delegation to occur on the equilibrium path, it is not enough for the ruler to be willing to do so—additional conditions on the elite side must be met as well. Existing theories largely overlook these considerations about elite demand for parliamentary constraints.
We derive the general form of the constraints before distinguishing between different types of elite outside options. Our first consideration is whether the elite can credibly threaten to sanction a ruler—that is, exercise its outside option—who opts to rule absolutely. We call this the elite-credibility condition. By rejecting an absolutist ruler's tax demand, the elite receives expected utility μ(θ) ⋅ ω, where μ(θ) = π L(θ) if the elite's exit option is to refuse and μ(θ) = σ if it is to exit. This payoff is surely better than funding a ruler who will expropriate the funds, which leaves the elite with μ(θ) ⋅ (ω − τ). Thus, applying Lemma 2, elite credibility holds for sure if 
$\theta < \hat{\theta }( {\underline{\omega } } ) $. By contrast, if the external threat is greater, the ruler will spend tax monies on security. This yields an expected payoff of π H(θ) ⋅ (ω − τ) for the elite. Whether this is better or worse than the outside option depends on the external threat strength, the size of the tax demand, and other parameters. Elite credibility holds whenever the elite's best response is to reject a demand by an absolutist ruler.
$$\eqalign{& {\bf Elite{\hbox-}credibility \; condition\colon } \cr & \underbrace{{\theta \le \hat{\theta }( {\underline{\omega } } ) }}_{{{\rm Ruler\;expropriates}}}\quad {\rm or}\quad \underbrace{{\mu ( \theta ) \cdot \underline{\omega } \ge \pi _{\rm H}( \theta ) \cdot ( {\underline{\omega } -\tau } ) }}_{{{\rm Elite\;rejects\;despite\;known\;security\;spending}}}} $$ When elite credibility fails, the elite will agree to the ruler's tax demands even if she does not delegate parliamentary privileges. This is sufficient for the ruler to not empower parliament. Delegation has two direct downsides for the ruler: the loss of initial wealth (from 1 − ω to
$1-\bar{\omega }$) and the loss of autonomy conditional on receiving tax funds (probability 
$\bar{q}$ of being constrained to spend on security).
Lemma 3: If the elite-credibility condition does not hold, then there is no equilibrium in which the ruler delegates to parliament.
Besides external threat strength, which we discuss later, factors shaping elite credibility include the size of the tax demand and the elite's baseline share of wealth. All else equal, a higher tax demand τ facilitates elite credibility: the elite can more easily threaten to reject the demand when the cost of accepting it is higher. Meanwhile, the elite's baseline share of wealth ω has ambiguous effects: it may strengthen elite credibility by making the ruler less likely to provide security voluntarily, or weaken it by increasing the relative benefit of security for the elite. Remark A.2 in the appendix formalizes these comparative statics.
Elite Willingness: Funding a Parliamentary Ruler
The other side of elite incentives is whether the elite prefers to accept the tax demand of a constrained ruler. We call this the elite-willingness condition. As with elite credibility, we evaluate elite willingness by comparing the elite's outside option value to the expected payoff from funding the ruler. The difference is that we now make the comparison assuming the ruler has delegated parliamentary privileges. By taking the outside option, the elite receives 
$\mu ( \theta ) \cdot \bar{\omega }$. Meanwhile, the value to the elite of funding the ruler depends on how tightly the constraints bind, as well as how the ruler will respond when the constraints fail ex post. If the ruler spends on security (either voluntarily or because the constraints bind), then the elite's expected utility from accepting the tax demand is 
$\pi _{\rm H}( \theta ) \cdot ( {\bar{\omega }-\tau } ) $. On the other hand, if the constraints do not bind (which occurs with probability 
$1-\bar{q}$) and the ruler opts to expropriate, the elite's expected utility is 
$\mu ( \theta ) \cdot ( {\bar{\omega }-\tau } ) $. We put these together to define the elite-willingness condition.
$$\eqalign{& {\bf Elite{\hbox -}willingness\;condition\colon } \cr & \mu ( \theta ) \cdot \bar{\omega } \le \left\{{\matrix{ {\underbrace{{[ {\bar{q}\cdot \pi_{\rm H}( \theta ) + ( {1-\bar{q}} ) \cdot \mu ( \theta ) } ] \cdot ( {\bar{\omega }-\tau } ) }}_{{{\rm Ruler\;expropriates\;when\;able}}}} & {{\rm if\;}\theta < \hat{\theta }( {\bar{\omega }} ) , \;} \cr {\!\!\underbrace{{\pi_{\rm H}( \theta ) }}_{{{\rm Ruler\;always\;provides\;security}}}\cdot \!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\! \cdot\;( {\bar{\omega }-\tau } ) } & {{\rm if\;}\theta \ge \hat{\theta }( {\bar{\omega }} ) .} \cr } } \right.} $$Like the elite-credibility condition, elite willingness is critical for there to be an equilibrium in which the ruler delegates to parliament. When elite willingness fails, the elite rejects the ruler's tax demand regardless of her initial choice of constraints. Because delegation reduces the ruler's initial wealth, absolutist rule is clearly better for her if the tax demand will be rejected either way.
Lemma 4: If the elite-willingness condition does not hold, then there is no equilibrium in which the ruler delegates to parliament.
In addition to the strength of the external threat, which we address later, elite willingness is determined by the size of the tax demand and the strength of institutional constraints. All else equal, a greater tax demand τ threatens elite willingness for the same reason that it strengthens elite credibility: even if the ruler were sure to uphold the deal, the elite must forgo more wealth to achieve the same proportional increase in security (see Remark A.2). In circumstances where the ruler would be tempted to renege on the deal given the opportunity, greater institutional strength 
$\bar{q}$ has the opposite effect by reinforcing elite willingness.Footnote 27 These comparative statics are formalized in Remark A.3 in the appendix.
Combining the Three Conditions
The combination of the three conditions we have introduced—ruler willingness, elite credibility, and elite willingness—determines whether the ruler delegates parliamentary privileges in equilibrium. If ruler willingness fails, then the ruler prefers absolutist rule with low security against outside invasion even if a taxes-for-security deal would in principle be possible. If elite credibility fails, then the elite has too little bargaining power to extract meaningful fiscal concessions from the ruler. Finally, if elite willingness fails, then the elite has too much bargaining power: the elite would reject even the most favorable deal, driving the ruler away from the bargaining table. Yet when all three conditions hold, the equilibrium outcome is for the ruler to delegate privileges to parliament.
Proposition 1: There is an equilibrium in which the ruler delegates to parliament if and only if ruler willingness, elite credibility, and elite willingness hold. If all three conditions hold strictly, parliamentary delegation is the unique equilibrium outcome.
This result, in essence, shows that two broad requirements must both be met for delegation to occur. First, the ruler's gain from the potential increase in revenue must outweigh her losses from the reduction in bargaining power and opportunity to renege that come with additional constraints (ruler willingness). Second, the elite must prefer to transfer tax funds to the ruler if and only if parliament is empowered (elite willingness and credibility).
In the remainder of the analysis, our goal is to determine how external war threats affect the possibility of an equilibrium in which the ruler delegates to parliament. Proposition 1 gives us a roadmap for this analysis: we must characterize the effect of external threats on each condition. Because the elite conditions are functions of the elite's outside option, which in turn depends on whether the elite's wealth is mobile or immobile, we analyze these cases separately. In each case, however, the broad story is similar: external threats have countervailing effects on the equilibrium conditions, so fiscal constraints are sustainable only when external threats are neither too weak nor too strong. Stated differently, the values of other parameters determine the overall effect of external threats on parliamentary delegation.
Effects of External War Threats
We now analyze how the magnitude of the external threat affects each of the three conditions for an equilibrium in which the ruler bestows parliamentary privileges. We uncover novel channels through which war threats undercut parliamentary delegation in equilibrium, contrary to the conventional expectation. These channels differ depending on the source of the elite's wealth. An elite with immobile wealth depends on the ruler for security even if it refuses to fund her. Thus, a strong external threat makes delegation unnecessary because the ruler's tax demand will be accepted even if she does not submit to institutional constraints. For an elite with a viable exit option, however, a strong war threat makes parliamentary delegation insufficient to induce elite funding. The elite prefers to move, hide, or self-protect its wealth rather than risk losing everything in war. Our new results follow from a previously unrealized tension: war threats tend to push elite credibility and elite willingness in opposite directions. Consequently, strong external threats either induce elite funding or compel parliamentary constraints, but not both. These results demonstrate the necessity of considering how external threats affect the bargaining leverage of actors with whom the ruler interacts, as opposed to the standard assumption that war threats affect only the ruler's calculus.
Ruler Willingness to Delegate
We begin by analyzing the effects of external threats on the ruler's initial decision to delegate privileges to parliament. When we considered the ruler's final decision of how to spend tax monies in Lemma 2, we found that stronger war threats increased the ruler's incentive to spend on security voluntarily. Now, moving up the game tree to consider the ruler's delegation decision, we recover the conventional logic that connects wars to parliament. Stronger threats increase the ruler's incentive to trade personal wealth for security, which is a precondition for the ruler to submit to fiscal constraints in equilibrium.
To understand why war threats promote ruler willingness, consider the ruler's comparison between (a) absolutist rule without tax funding and (b) constrained rule with tax funding. Accepting constraints reduces the ruler's slice of the domestic pie. The potential benefit to the ruler is that taxes will allow her to provide security, thereby increasing the probability of keeping her wealth safe from external invasion. The decrease in the ruler's slice of the domestic pie resulting from parliamentary delegation is independent of the external threat, whereas the potential security benefit increases as the outsider grows stronger. Consequently, ruler willingness fails if the war threat is weak enough and holds if the threat is strong enough.
Lemma 5 (War threats promote ruler willingness): The ruler-willingness condition holds if and only if θ ≥ θ rw, where θ rw > 0 if and only if 
$\tau < {{\bar{\omega }-\underline{\omega } } \over {1-\bar{q}}}$ and where θ rw < ∞ if and only if 
$\Delta > {{\bar{\omega }-\underline{\omega } } \over {1-\bar{\omega }}}$.
Figure 3 presents a region plot for the ruler-willingness condition. In this and all the following region plots, the outsider's power is on the x-axis and the strength of institutional constraints is on the y-axis, with all other parameters fixed at values listed in the note accompanying the figure; the shaded region indicates parameter values in which the condition is met.
Ruler-willingness condition
As we move from left to right along the x-axis, representing stronger external threats, we begin with the ruler possibly being unwilling to delegate, but eventually reach a point at which ruler willingness holds. This finding comports with the bellicist logic of state formation: a looming invasion threat increases the ruler's demand for security, driving her to submit to parliamentary constraints on her spending in order to receive needed taxes for military improvements.
Figure 3 also shows how the ruler is less willing to submit to constraints that have a higher probability of binding her spending decisions. Moving up the y-axis on the graph, representing stronger institutional constraints, we see that ruler willingness initially holds at low levels, but eventually fails unless the external threat is large. In the absence of a strong war threat, an increase in the constraining power of parliamentary institutions may work against the formation of limited executive power, insofar as it decreases the ruler's willingness to submit to fiscal constraints.
Elite Constraints with Immobile Wealth
When the elite's wealth cannot be easily moved out of the country, hidden within the ruler's domain, or self-protected, its outside option is to refuse the tax demand. By refusing to fund the ruler without safeguarding its wealth, the elite runs the risk of expropriation by the outsider, but otherwise consumes its full endowment. Because the ruler can provide security improvements only with tax funding, the elite expects to retain a fraction π L(θ) of its wealth after refusing:
$${\rm {\opf E}}[ {U_{\rm E}( {{\rm outside\;option}} ) } ] = \pi _{\rm L}( \theta ) \cdot \omega $$Elite credibility holds when the elite would refuse to pay taxes to a ruler who is unconstrained by parliament. There are two mechanisms by which a stronger outside threat hampers elite credibility when the elite's wealth is immobile. First, the ruler will spend funds on security if the war threat is sufficiently ominous (see Lemma 2). In this case, the ruler's incentives are aligned with those of the elite, and thus it does not matter whether the ruler is formally constrained.
Second, outsider threats undercut the elite's refusal option. A stronger threat drives down the value of both possible choices, either refusing or funding, because in either scenario the invasion is more likely to succeed. However, by increasing the value-added of enhanced security, a stronger external threat drives down the expected utility of refusal by a larger magnitude. This follows from the assumption that the ratio 
${{\pi _{\rm L}} \over {\pi _{\rm H}}}$ decreases with external threat strength.
Altogether, a stronger external threat increases the elite's incentive to yield to an absolutist ruler's tax demand, which undermines elite credibility.
Lemma 6 (War threats reduce immobile elite credibility): Assume the elite's outside option is to refuse. The elite-credibility condition holds if and only if the external threat is weak enough: 
$\theta \le \theta _{{\rm refuse}}^{{\rm ec}} $, where 
$\theta _{{\rm refuse}}^{{\rm ec}} \ge \hat{\theta }( {\underline{\omega } } ) > 0$.
Figure 4 illustrates when the elite-credibility condition is satisfied as a function of the invader's power and the strength of institutional constraints. Clearly, a strong outside threat undermines elite credibility. Meanwhile, the strength of institutional constraints does not affect elite credibility, as this condition depends only on the elite's optimal choice conditional on the ruler having chosen absolutism.
Elite credibility with immobile wealth
For virtually the same reasons that elite credibility holds only when the outsider is weak, the willingness condition for an elite with immobile wealth holds only when the outside threat is strong enough. To see why, first consider the extreme case with no outside threat: θ = 0, so that π L(θ) = π H(θ) = 1. Then even if the ruler spends tax revenues on security, the elite would have reduced its endowment by τ without increasing its chance of prevailing against invasion. Clearly, it would be better to not pay taxes in this case. This calculus changes once we introduce an outside threat and increase its magnitude. The same two factors discussed before push the elite toward funding a ruler who has submitted to fiscal constraints—alignment of the ruler's spending incentives with those of the elite, and the elite faces higher incentives to pay taxes because the relative benefit of security is higher. Altogether, then, outside threats increase the elite's willingness to accept the tax demands of a ruler who has accepted fiscal constraints.
Lemma 7 (War threats promote immobile-elite willingness): Assume the elite's outside option is to refuse. The elite-willingness condition holds if and only if the external threat is strong enough: 
$\theta \ge \theta _{{\rm refuse}}^{{\rm ew}} $, where 
$\theta _{{\rm refuse}}^{{\rm ew}} > 0$.
Figure 5 illustrates the willingness condition for an elite with immobile wealth. Stronger external threats and tighter institutional constraints both promote elite willingness in this case. Compared to the illustration of ruler willingness (Figure 3), we see a tension in the effects of institutional strength on the ruler's and elite's willingness conditions. The commitment effect pushes the elite toward funding a ruler who has delegated, but it also pushes the ruler away from delegating in the first place.
Elite willingness with immobile wealth
Elite Constraints with Mobile Wealth
Elites with mobile wealth, which can be moved, hidden, or self-protected, have a distinct outside option—and thus a distinct strategic calculus. For these elites, exercising the outside option protects their wealth against outside invasion because exiting takes it outside (or hides it within) the territory under threat.Footnote 28 The cost of exiting is that an elite with mobile wealth leaves behind a fraction 1 − σ of what it would otherwise consume. This represents the costs associated with liquidation, transportation, and evasion:
$${\rm {\opf E}}[ {U_{\rm E}( {{\rm outside\;option}} ) } ] = \sigma \cdot \omega .$$In contrast to the case with immobile wealth, the value of the elite's outside option in this case is independent of the outsider's strength, θ. This difference yields the distinct mechanism that a stronger outside threat hampers elite willingness when the elite's wealth is mobile.
In the immobile-wealth case, we saw the elite's willingness to accept taxes increase with external threat strength. An elite with immobile wealth depends on the ruler for security against invasion, so its willingness to fund the ruler in hopes of receiving a security benefit rises as the threat increases. The strategic calculation is reversed for an elite with mobile wealth, who does not depend on the ruler for security. Even if the ruler always spends taxes on security, the elite's expected utility from funding the ruler approaches 0 as the outside threat increases in magnitude. Exiting, by contrast, yields positive consumption. In this case, a stronger outside threat reduces the expected utility of paying taxes to the ruler and depending on her for security, but does not affect the value of the outside option. This is the key difference between the outside options of refusal and exit. An elite with mobile wealth will therefore accept a constrained ruler's tax demand only if the outside threat is not too strong.Footnote 29
Lemma 8 (War threats reduce mobile elite willingness): Assume the elite's outside option is to exit. The elite-willingness condition holds only if the external threat is weak enough: 
$\theta \le \theta _{{\rm exit}}^{{\rm ew}} $, where 
$\theta _{{\rm exit}}^{{\rm ew}} < \infty $.
Figure 6 illustrates the relationship between outside threats and elite willingness for an elite with mobile wealth. The pattern is the opposite of that for an elite with immobile wealth (Figure 5). Now the elite is willing to fund a constrained ruler even in the absence of an external threat, which avoids the costs associated with exit. Once the threat is large enough, however, the elite would rather exit than depend on the ruler for security. We also see that institutional strength again promotes elite willingness, for the same reasons as with immobile wealth.
Elite willingness with mobile wealth
The exit option also changes the nature of the elite-credibility condition (Equation 3), which holds whenever the elite would refuse the tax demand of an absolutist ruler. In the immobile-wealth case, we identified two ways that external threats undermine elite credibility: first by increasing the likelihood that the ruler would choose security over consumption if taxes were collected, and second by reducing the relative value of the outside option for the elite. The first of these effects is still operative for an elite whose outside option is to exit. However, as noted in the discussion of elite willingness, the second effect now goes in the opposite direction, as a stronger threat increases the relative value of exiting. This creates a channel through which war threats enhance elite credibility.
Due to these countervailing effects for elites with mobile wealth, there is potentially a nonmonotonic relationship between external threat strength and elite credibility. If the threat is low, 
$\theta < \hat{\theta }( {\underline{\omega } } ) $, then elite credibility holds because an unconstrained ruler would expropriate tax revenue. If the threat is high, such that π H(θ) ≤ σ, then elite credibility again holds. In this case, the expected losses from invasion outweigh the costs of exit even when the ruler is certain to provide security (the same mechanism that affects willingness for a mobile elite). In between these extremes, elite credibility may fail. In this case, the ruler does not delegate to parliament in equilibrium.Footnote 30
Lemma 9 (War threats and mobile elite credibility): Assume the elite's outside option is to exit. If 
$\sigma \ge \hat{\sigma }\equiv \pi _{\rm H}( {\hat{\theta }( {\underline{\omega } } ) } ) \cdot \boldsymbolg({1-{\tau \over {\underline{\omega } }}} \boldsymbolg)$, then the elite-credibility condition holds for all θ. Otherwise, if 
$\sigma < \hat{\sigma }$, then the elite-credibility condition holds if and only if 
$\theta \notin ( {\hat{\theta }( {\underline{\omega } } ) , \;\theta_{{\rm exit}}^{{\rm ec}} } ) $, where 
$\hat{\theta }( {\underline{\omega } } ) < \theta _{{\rm exit}}^{{\rm ec}} < \infty $.
Equilibrium Delegation to Parliament
How do external threats shape the ruler's ultimate decision to delegate privileges to parliament? To characterize the relationship between war threats and the choice to delegate, we bring together the ruler and elite conditions analyzed earlier. Surprisingly, given the different mechanisms for elites with immobile versus mobile wealth, the implications for how wars affect parliaments are qualitatively similar. In both cases, the countervailing effects of war threats on the elite-credibility and elite-willingness conditions yield the divergence from the conventional logic.
Immobile wealth
When the elite's wealth is immobile, there is a noteworthy tension between the elite's credibility and willingness conditions. This can undermine parliamentary delegation as an equilibrium outcome. If war threats are too weak, then the elite has no demand for security and thus no incentive to pay taxes even if the ruler has accepted fiscal constraints. But if the threat is too strong, then the elite's demand for security is too great to pressure the ruler for institutional concessions; the elite will provide taxes to an unconstrained ruler. For both conditions to be met—which is necessary to induce parliamentary delegation—the outside threat must be moderate. This requirement illustrates how our understanding of the relationship between war threats and parliamentary privileges changes once we account for elite incentives to fund the government.
Recall that delegation to parliament is the equilibrium outcome when ruler willingness, elite credibility, and elite willingness all hold (Proposition 1). In the case of an elite with immobile wealth, the threat must be large enough that ruler willingness and elite willingness hold, but small enough that elite credibility still holds. The following proposition summarizes these requirements.
Proposition 2 (Parliamentary equilibrium with immobile elite wealth): Assume the elite's outside option is to refuse. There is an equilibrium in which the ruler delegates to parliament if and only if the external threat is moderate: 
$\max { {\theta^{{\rm rw}}, \;\theta_{{\rm refuse}}^{{\rm ew}} } } \le \theta \le \theta _{{\rm refuse}}^{{\rm ec}} $. (θ rwis defined in Lemma 1, 
$\theta _{{\rm refuse}}^{{\rm ec}} $ in Lemma 6, and 
$\theta _{{\rm refuse}}^{{\rm ew}} $ in Lemma 7.)
For an elite with immobile wealth, Figure 7 illustrates the parameter values for which the ruler delegates to parliament in equilibrium. The shaded region represents the intersection of the three equilibrium conditions (previously illustrated in Figures 3–5), making clear the conditionality of the relationship between war threats and parliament. In the presence of a weak threat, the ruler will choose absolutism and the elite will respond by refusing the tax demand, resulting in a government that is neither strong nor limited. From there, a stronger invasion threat may make parliamentary delegation more likely by promoting ruler willingness or elite willingness. However, too strong a threat breaks the possibility of parliamentary privileges by undermining elite credibility. In this case, the elite will accept the tax demand even without parliamentary constraints. This results in a government that is strong but not limited.
Parliamentary equilibrium with immobile elite wealth
Because of the tension between ruler willingness and elite willingness, stronger institutional constraints also have conditional effects. The question is which constraint binds most tightly. If elite willingness just barely holds (θ close to 
$\theta _{{\rm refuse}}^{{\rm ew}} $), then a slight weakening of institutional constraints might foreclose the possibility of an equilibrium with delegation. But if ruler willingness just barely fails to hold (θ close to θ rw), a similar weakening might actually push the ruler toward delegation. As the external threat grows stronger, the set of sustainable levels of institutional strength expands, until we reach the point of elite credibility failing.
Mobile wealth
Turning to the case of an elite with mobile wealth, once again we see that pressing external threats do not lead to states that are strong and limited. High fiscal constraints are sustainable as an equilibrium for only an intermediate range of war threats. The war threat must be great enough that the ruler is willing to make fiscal concessions in order to receive tax monies, yet not so great that the elite would rather exit than depend on a constrained ruler for security. And if the elite's losses from exiting are high enough, the magnitude of the threat must be outside the range where elite credibility fails.
Proposition 3 (Parliamentary equilibrium with mobile elite wealth): Assume the elite's outside option is to exit. If 
$\sigma \ge \hat{\sigma }$ (see Lemma 9), then there is an equilibrium in which the ruler delegates to parliament only if 
$\theta ^{{\rm rw}} \le \theta \le \theta _{{\rm exit}}^{{\rm ew}} $. Otherwise, if 
$\sigma < \hat{\sigma }$, then there is an equilibrium in which the ruler delegates to parliament only if 
$\theta ^{{\rm rw}} \le \theta \le \min \boldsymbolg\{{\mathop \theta^{{\rm ew}}_{{\rm exit}} , \;\hat{\theta }( {\underline{\omega } } ) } \boldsymbolg\}$ or 
$\max { {\theta^{{\rm rw}}, \;\theta_{{\rm exit}}^{{\rm ec}} } } \le \theta \le \theta _{{\rm exit}}^{{\rm ew}} $. (θ rwis defined in Lemma 1, 
$\hat{\theta }( {\underline{\omega } } ) $ in Lemma 2, 
$\theta _{{\rm exit}}^{{\rm ew}} $ in Lemma 8, and 
$\theta _{{\rm exit}}^{{\rm ec}} $ in Lemma 9.)
As in the prior case, we find that parliamentary delegation tends to be sustainable only when external threats are not too large. Although the topline predictions are similar for elites with either type of wealth, there is a sharp difference in the mechanism that drives the result. For an elite with immobile wealth, strong external threats discourage parliamentary delegation by reducing the elite's bargaining power. In the presence of an overwhelming threat, we would expect an elite with immobile wealth to agree to rulers’ tax demands even without meaningful constraints on the ruler's ability to spend. The state will be strong but not limited. If instead the elite's wealth is mobile, we would expect the same behavior by the ruler—no submission to fiscal constraints—but because of a distinct elite response. An elite with mobile wealth would exit rather than fund a hopeless effort against invasion, even if the ruler could promise to spend taxes on security. In this case, the state will be neither strong nor limited.
Figure 8 illustrates equilibrium delegation of parliament for an elite with mobile wealth. The figure is the intersection of the willingness conditions shown in Figures 3 and 6.Footnote 31 As before, we see that marginal increases in war threats or in institutional strength can enable or hinder delegation, depending on which constraint is closest to binding. However, there is one substantive difference from the case of immobile wealth: now parliamentary delegation is sometimes sustainable as an equilibrium outcome even when the outsider is weak. This is because of the fixed cost of the exit option: the elite may prefer spending τ on a low-value security improvement over losing 1 − σ of its wealth to exit.Footnote 32
Parliamentary equilibrium with mobile elite wealth
Alternative outside options
As noted in the model setup, we consider two outside options that combine elements of immobile and mobile wealth; full details are in the appendix. These are a hybrid outside option where some wealth is moved while the remainder stays in the ruler's domain (Appendix B.1), and an alternative exit option where the value of exiting decreases with external threat strength due to shrinking safe harbors for wealth (Appendix B.2). In both cases, strong outside threats have the same effects as in the baseline model with mobile wealth—strengthening elite credibility, but weakening elite willingness. However, marginal increases in initially low threat levels may sometimes have opposing effects.
Extensions
Coercion
An army does not merely defend against external threats—the ruler may also use it for internal coercion.Footnote 33 This consideration is particularly relevant when “funding the ruler” entails building a professional standing army, as opposed to supplementing a feudal levy with mercenary battalions. To analyze the new trade-offs that emerge when rulers seek standing armies, we extend the baseline model with immobile elite wealth in two ways. First, we assume that the ruler can use a standing army both to defend against external threats and, if unconstrained, to expropriate additional wealth from the domestic elite. Second, we allow the ruler to pay an upfront cost to obtain the standing army through coercion, rather than through negotiating with the elite.
The main new insight from the extension is that the ruler gains funding under circumstances she would not in the baseline model. Figure 9 illustrates the possible outcomes. Before, if elite willingness failed, the ruler was unable to gain funding. Now, the ruler can coerce funding from a recalcitrant elite. Another factor mitigating against consensual bargaining in this extension follows from a downstream effect: the possibility of coercion makes it even riskier for the elite to fund the government, which weakens elite willingness—therefore widening the range of parameter values in which the ruler needs to coerce to gain funding in the first place. Later we illustrate this mechanism in the case of Brandenburg-Prussia. Yet overall, the core takeaways from the baseline model are qualitatively unchanged, in particular that parliamentary delegation occurs only if the outsider threat is moderate in strength. Formal details appear in Appendix B.3.
Equilibrium choices in the model with coercion
Offensive Wars
We next extend the model to allow for offensive wars. In this case, military success is not necessarily a public good, as the ruler may monopolize access to the spoils of victory abroad. The ruler is more motivated to fight than the elite because she expects an outsized share of the benefits, which introduces problems of political bias and moral hazard.Footnote 34 Delegating powers to parliament can mitigate these issues, but only to the extent that institutional constraints are likely to bind how the ruler allocates the spoils of war.
The key takeaway from this extension is that offensive war dramatically strengthens elite credibility, which now holds regardless of the outsider's strength. An absolutist ruler has no ex post incentive to share the spoils of war with the elite. Knowing this, the elite has no incentive to accept an absolutist ruler's tax demand, as we illustrate below in the case of England. Meanwhile, elite willingness is now inversely related to the strength of the outsider, as the expected spoils of conflict are greater against a weaker opponent. Nevertheless, the core findings are the same as in the baseline model: delegation occurs in equilibrium only if the outsider's strength is moderate. Formal details appear in Appendix B.4.
Application to Historical European Parliaments
The mechanisms in our model provide strategic foundations that help explain episodes of parliamentary delegation and reversals in Western European history. Drawing from the vast literature on war, state formation, and parliaments in Europe, we distinguish three main periods for territorial states: (1) the emergence of parliaments with taxation powers through the sixteenth century, (2) the decline of parliaments over the next two centuries, and (3) a resurgence of parliaments, with new-found power over expenditures, in the nineteenth century and afterward. In this section, we discuss the first two periods. Varying values of three parameters help explain differences across polities in the elite credibility and willingness constraints: the magnitude of war threats, θ; the credibility of parliamentary constraints, 
$\bar{q}$; and the size of parliamentary grants needed to fund effective war efforts, τ. In Appendix C.2, we discuss ruler willingness in the context of nineteenth-century reforms. We argue that external wars were less important than domestic conditions for the rise of parliaments with spending powers.
Figure 10 graphically summarizes these three periods using data described in Appendix C.1. Both panels use a sample of ten major territorial states between 1200 and 1900 with annual observations averaged over decade-long intervals. In Panel A, we plot the fraction of states that had parliaments with power over taxation (solid line) and with power over expenditures (dashed line). In Panel B, we plot the fraction of states with at least one parliamentary meeting in a given decade (solid line) and this fraction by year (dashed line).
Western European parliaments over time
Emergence of Medieval Parliaments
Participation in external wars was one stimulus for the rise of parliaments in Western Europe between the thirteenth and fifteenth centuries. As Figure 10 shows, during the fifteenth century, parliaments with powers over taxes met frequently in major European states. Most bodies gained the right of approval for extraordinary taxes, although monarchs retained legal prerogatives over expenditures. Historians broadly concur that warfare contributed to the early rise of European parliaments by inducing rulers to seek taxes.Footnote 35 Analyzing data between 1000 and 1600, Cox, Dincecco, and Onorato show that recent participation in war correlates positively with a ruler calling their first parliament;Footnote 36 and Blaydes and Paik demonstrate that areas with more Holy Land crusaders were more likely to have at least one parliamentary meeting between 1100 and 1400.Footnote 37
Our model provides strategic foundations both for the emergence of parliaments with tax powers in this period and for exceptions to the overall trend.Footnote 38 During this period, we contend that the credibility of parliamentary constraints, 
$\bar{q}$, tended to be relatively high; and the size of parliamentary grants needed to fund an effective war effort, τ, tended to be relatively low. Monarchs were weak both financially and militarily, which undercut their ability to raise revenue independent of parliament. Furthermore, wars were inexpensive in this period (compared to the later Military Revolution), so even small grants from parliament could significantly bolster a war effort. These conditions promote elite willingness (Remark A.3 in the appendix), meaning this condition is likely to hold in the presence of moderate war threats, θ. Yet despite generally favorable scope conditions for war to promote parliament, several counterexamples from this period illustrate why war often fails to promote states that are both funded and limited, as in France and the Hanseatic League.
Factors promoting elite willingness
Monarchs were financially impoverished across medieval Europe, which bolstered the de facto credibility of parliamentary constraints, 
$\bar{q}$. Although the Carolingian dynasty temporarily created a large and centralized state in the ninth century, a core element of its strategy for consolidating power was to offer land grants to knights in return for their service. Over time, these fiefs became hereditary and created a class of landed nobles that governed largely autonomous regional kingdoms and duchies.Footnote 39 Later, the resumption of long-distance trade amid a Commercial Revolution coincided with a boom in urban population and wealth,Footnote 40 which also lay beyond the direct control of monarchs.
Monarchs controlled piecemeal sources of revenue. These included royal domains (or Crown lands) and profits from the administration of justice. Rulers could sell off parts of the state through tax farming, selling offices, allowing cities to purchase charter rights, and selling immunities and pardons. Other acts were essentially confiscation: debasing the currency, defaulting on the debt, English purveyance, and the French chambre de justice. Many rulers also had access to certain ordinary taxes and to royal monopolies on goods such as salt, although these reflected bargains with elites. These conditions contrasted with those in contemporaneous empires in China and the Middle East, which had bureaucracies and standing armies better positioned to collect central revenues.Footnote 41
For impoverished monarchs wanting to fund a war effort, calling domestic elites such as landed nobles, clergy, and merchants into an organized assembly yielded clear benefits. In fact, some monarchs mandated attendance at parliamentary meetings for elites otherwise reluctant to endure the costs and time of traveling to the capital.Footnote 42 Conversely, monarchs had few viable sources of revenue if bargaining with elites failed. This reality constrained rulers against acting irresponsibly without facing serious repercussions, as we discussed earlier. Consequently, parliaments could gain relatively strong de facto powers, reflected in the parameter
$\bar{q}$, despite more limited de jure prerogatives stemming from the lack of direct parliamentary control over how the ruler spent funds.
Why did parliaments not arise before the thirteenth century? The earlier medieval period violated the scope condition of our model: security improvements, which cost τ, must be sufficiently expensive that the ruler cannot self-fund them (see note 21). During the later medieval period, τ grew large enough to meet this scope condition, although it was still very low by the standards of the Military Revolution period, which we discuss later. For centuries, emergent European monarchs could get by without an institutionalized means of extracting revenues from elites because they lived off their royal domains and relied on feudal levies to fight wars. However, by the thirteenth century, rulers increasingly supplemented their feudal levies with mercenaries. The associated rising costs disabled kings from funding wars solely from their royal domains. For example, Richard I of England (1189–99) generally fought wars using the feudal levy and ad hoc requests for additional revenue. By contrast, his successor John raised more taxes to pay additional troops. These changes reflected not only the general unreliability of troops supplied by the feudal levy, but also the growing vulnerability of mounted knights to attacks by armed infantry.Footnote 43 By the end of the thirteenth century, England had largely eliminated knight service in kind.Footnote 44 The frequency of requests for additional funds from nobles and knights in their council, which over time transformed into the English Parliament, increased sharply during this century.Footnote 45
Although conditions during the late medieval period were generally favorable for external war to promote parliaments, even this period contains some counterexamples that correspond to mechanisms from the model.
Failure of elite credibility: France
France diverged from many other Western European states during the medieval period. Despite transforming its tax system during the Hundred Years’ War with England (1337–1453), its Estates-General met infrequently and lacked the power to veto tax requests.Footnote 46 We contend that this large defensive war undermined elite credibility (see Lemma 6). That is, θ was simply too large to sustain parliamentary delegation (see Proposition 2).
The Crown lacked any permanent tax streams before the Hundred Years’ War. This changed in the 1350s when new taxes were needed to pay King John II's ransom after he was captured by the English at the Battle of Poitiers. Renewed efforts at taxation in the 1420s occurred amid new and successful English offensives, which engendered a permanent land tax (the taille) and standing army.Footnote 47 Throughout, landed elites paid these taxes either directly (as with the fouage) or indirectly because taxation of peasants cut into the amount elites could extract from them (as with the taille). The Estates-General played an important role in both of these critical decades but never became a regular body. Excepting the 1350s (five meetings) and the 1420s (four), the Estates-General convened on average once every eleven years between 1302 (its first meeting) and 1453 (the end of the Hundred Years’ War).Footnote 48
In the 1350s the Estates-General failed its first major test by being unable to raise sufficient taxes to pay John's ransom.Footnote 49 Instead, elites and future kings turned to regional estates and ad hoc meetings of the Estates-General to raise taxes. These taxes were the heaviest in French history, but compelled by the external war threat: “The return of the king and a treaty to end the war with England easily met the criterion of ‘evident necessity’ that the legists claimed could justify extraordinary taxation.”Footnote 50 In the 1360s the Estates-General sanctioned another tax, the fouage. Although nobles were subject to this tax, they did not raise much opposition because its sole purpose was to pay military salaries, which overcame France's “chronic lack of military preparedness.”Footnote 51 In the absence of institutionalized concessions by the monarch, these high ad hoc revenues indicate a failure of elite credibility.
The ad hoc structure of tax bargains persisted in the 1420s and 1430s as “the English gathered their strength for an assault to the south,”Footnote 52 which once again hindered elite credibility. King Charles VII regularly turned to the national and regional estates for taxes, and they often complied. In 1428, the Estates-General consented to provide Charles with customs taxes. He continued to impose customs taxes afterward despite not reconvening the Estates-General for another decade, indicating the body's lack of veto power. In 1439, the Estates-General consented to raising large sums via the taille to reform the army. These reforms were recognized as “essential” to repel the English occupation,Footnote 53 and “the French were not in a mood to resist the king's demands for what he claimed he needed to end the war.”Footnote 54 Afterward, Charles VII did not call the Estates-General for the remainder of his reign yet continued to levy the taille. Following more minor reforms earlier in the decade, in 1445, Charles used funds (primarily from the taille) to fund the first regular paid professional army in medieval Europe, the Compagnie d'ordonnance.Footnote 55 The new army ended the Hundred Years’ War.
Overall, “Under Charles V and Charles VII, order, stability and strong government seemed highly desirable and something for which it was worth paying.”Footnote 56 During this period of invasion and occupation, the Crown gained control over permanent sources of taxes without empowering a national parliament, consistent with a failure of elite credibility.Footnote 57
Failure of elite willingness: Hanseatic League
Elsewhere in Europe, other forms of political organization displaced territorial states. Strong external threats cause the elite-willingness condition to fail for elites with high-value exit options, which can entail moving, hiding, or self-protecting their wealth (see Lemma 8). The Hanseatic League provides an example of exit and state fragmentation in a high-threat environment. Rather than fund territorial states governed by German princes, merchants in numerous towns exited by forming a trading league, which provided better protection for their wealth (see Proposition 3).
The fall of the Carolingian empire enabled many central European towns to operate outside the reach of princely control.Footnote 58 Coalescing into a broader political unit offered certain benefits for town dwellers. However, joining a territorial state was not necessarily their best option, even if they gained charter or parliamentary protection by a particular prince. Given the fragmented nature of the European state system in the late Middle Ages, protecting mobile assets (in the sense that they needed to move across space to realize their value) was perilous. “Towns that transacted business across these feudal units were faced with a variety of different legal codes, local tolls, differences in weights and measures, variation in coinage, and sometimes outright robbery, all to the detriment of the burghers’ business.”Footnote 59 Militarily weak princes with limited domains could not solve the large external threat that confronted merchants: numerous points along a trading route where a rival state could expropriate or extort them. Therefore, even responsible use of tax funds by such princes yielded only a small security boost, that is, Δ was low. This factor distinguished central Europe from areas like England:
If the monarchy was strong, as in England, the leading towns might look to the king for charters of liberties for individual towns . . . If the monarchy was weak, as in thirteenth- and fourteenth-century Scotland and Germany, then towns might have to join together in leagues and sometimes wage war to gain or to defend the privileges they valued.Footnote 60
The Hanseatic League was the most prominent trading league, and illustrates a case of viable exit, or high σ. It formed in the thirteenth century after the town of Lübeck threw off Danish rule and became an imperial city within the Holy Roman Empire. Elites in Lübeck struck deals with neighboring towns to jointly resist highway robbers along high-volume river routes and to share a currency.Footnote 61 These towns did not form a new territorial state with centralized governance institutions, such as financial organizations or a military. Instead, consistent with exiting, the towns remained loosely aligned to pursue their “shared desire for peace, for economic properity.”Footnote 62 The Hanseatic League remained powerful for roughly two centuries, including a successful war against Denmark in 1370.Footnote 63 Their main method of coercion was the economic blockade. “Since the Hansa had a virtual monopoly on many bulk goods, they could bring recalcitrant towns, or lords who refused to grant concessions, to their knees.”Footnote 64 Thus this organization enabled merchants to protect their wealth while exiting from the tax demands of territorial princes.
Parliamentary Reversals During the Age of Absolutism
In subsequent centuries, Western Europe experienced a wave of parliamentary reversals. As shown in Figure 10, many parliaments lost their taxation powers, and the frequency of their meetings plummeted between the sixteenth century and 1789. During the Military Revolution and the Age of Absolutism, scholarship has linked wars to parliamentary reversals. Cox, Dincecco, and Onorato study 103 polities within nine composite monarchies that, in 1500, had parliamentary control over taxation and royal control over expenditures.Footnote 65 By 1789, only 39 percent of these polities retained this constitutional arrangement. The majority, 52 percent, transitioned to fiscally absolute regimes in which the crown controlled both; and the remaining 9 percent transitioned to regimes with parliamentary supremacy. In their sample, recent participation in an international war correlates with a higher probability of transitioning to absolutism. Qualitative observations elsewhere in the literature support this relationship.Footnote 66
We contend that in many countries, elite willingness dropped relative to the medieval period. Wars became more expensive (higher τ), and the credibility of parliamentary constraints diminished (lower 
$\bar{q}$). Changes in military technology meant that the most effective mode of military organization empowered the ruler over domestic rivals. When funded, standing armies gave rulers an alternative to negotiating with parliament: coercing revenues from elites. These changes lowered elite willingness (Remark A.3 in the appendix), which can also stimulate the ruler to use coercion (as demonstrated in our extension in Appendix B.3). After elaborating on these factors, we discuss the contrasting examples of Brandenburg-Prussia and England.
Factors undermining elite willingness
For many European countries, the Military Revolution (beginning around 1500) changed the conditions that earlier had underpinned a positive relationship between war and parliament. Improvements in applications of gunpowder technology necessitated building new fortresses that could absorb cannon fire as well as hiring professional and armed infantry troops, which supplanted feudal levies and mounted knights. Effective armies became much larger and more expensive. In the seventeenth century the largest armies of major states were 2.8 times (England), 3.8 times (France), or 15 times (Spain) as large as in the 1470s; and the annual cost of Spain's wars in the 1590s was 4.5 times what it had been in the 1540s.Footnote 67
The Military Revolution undermined elite willingness in two ways for many states. First, changes in military technology made wars more expensive, which dramatically raised the outlay typically needed to fund an effective war effort, τ. Even if elites were willing to provide some war financing, the amount they contributed was usually smaller than what the ruler desired, amid a period in which financing became a stronger predictor of success in battle than in the past.Footnote 68 Consequently, monarchs’ relationships with their parliaments were often characterized by gridlock. Parliaments that lacked control over expenditures were wary of granting large sums to the ruler that might be used in outright irresponsible ways, or for wars in which they were uninterested.
Second, monarchs gained more viable means of reneging on deals with parliament. The shift to standing professional armies reduced monarchs’ ability to commit to promises, which lowered 
$\bar{q}$. One way a ruler could use funds irresponsibly, from the perspective of elites, was to build a professional military and then use the army to coerce taxes from recalcitrant elites. This latent source of monarchical strength contributed to parliamentary gridlock by making elites less willing to provide funds, yet some rulers were able to coerce parliament if negotiations broke down.
Coercing parliament in Brandenburg-Prussia
Brandenburg-Prussia exemplifies the general conditions that fostered European absolutism, in particular low 
$\bar{q}$ and high τ. Elector Frederick William initially delegated to parliament, which yielded military funding. However, later, the ruler used his standing army to renege on the deal with parliament while fighting an external war. The ruler used coercion because elite willingness failed for a war deemed outside the elites’ interests.
Initially, a contractual relationship held between ruler and parliament. Like many European rulers emerging from the medieval period, in the sixteenth century the margrave of Brandenburg was merely “the largest landowner in a society dominated by landowners’ interests,” with a weak state military: “the feudal services had fallen into desuetude during a longer period of peace, and the hiring of mercenaries was prevented by the margrave's impecuniousness.”Footnote 69 Before the “Great Elector” Frederick William gained strong coercive means, he had no alternative but to negotiate with parliament for revenues. Thus, he initially delegated privileges to parliament in return for elite funding for the government. In 1653, Frederick William struck a deal with the Brandenburg Estates that yielded the following economic privileges for large landowners:
Eliminating the legal restrictions which had bound the Junkers in the past, [Frederick William] recognized them as the only class authorized to acquire estates and confirmed and stabilized privileges they had extorted from his predecessors, like exemption from taxation and the right to control the lives of their peasants. Finally, he specifically recognized their authority in local affairs and their right to be regarded as the governing class in all matters that concerned the state as a whole.Footnote 70
This bargain yielded clear benefits for both sides:
Frederick William had got his money grant, but he had to buy it with concessions which not only confirmed the social privileges of the nobility but also the political power of the Estates. From their point of view the granting of supply, although for six years,Footnote 71 probably weighted lightly compared with the important points which they had scored, in particular as the noblemen themselves did not contribute towards the grant.Footnote 72
Later, Frederick William reneged on the deal. The six-year tax grant enabled him to build a stronger standing army, which lowered 
$\bar{q}$ in subsequent interactions with parliament. An unexpected opportunity arose in 1655 when a war broke out between nearby Sweden and Poland.Footnote 73 This enabled the Elector to put his standing army in the field and to demand more from the Brandenburg Estates.
The Brandenburg Estates refused to countenance new demands to prosecute an external war they considered beyond their interests. Thus, if we consider the interaction between the Elector and the Estates in 1655 in its own right, this episode would exemplify a failure of elite willingness in response to high τ. The problem is that the Estates were in a weaker position than before, which enabled the Elector to renege on the original deal. He reacted to parliamentary intransigence by using the standing army to enforce collection of new taxes, hence reneging and gaining additional consumption from elites (as proposed in our coercion extension). “At the end of the war he had gained great strength and a standing army, capable of breaking any resistance against the collection of taxes required for its maintenance. Thus the Estates were never able to regain the ground which they had lost.”Footnote 74
Failed coercion in England
England is an exception to the general pattern of absolutism in the centuries preceding the French Revolution. The English Parliament instead defended, and eventually expanded, its traditional prerogatives. We highlight how England's distinct geopolitical position helped to counteract pro-absolutist pressures unleashed by the Military Revolution. For centuries, minimal defensive war threats hardened elite credibility, as explained in our extension with offensive wars (Appendix B.4). Later, James II almost gained a similar opportunity as Frederick William to overrun Parliament, but the absence of opportunities to deploy his standing army meant that 
$\bar{q}$ was higher than in the Brandenburg-Prussia case.
Throughout English history, its island location hardened elite credibility. Elites in Parliament could credibly refuse to fund wars that solely furthered royal interests, such as regaining lost continental lands.Footnote 75 In the seventeenth century, the Stuart kings repeatedly ran up against the limits of parliamentary tolerance. “On five occasions—in 1625 (twice), 1626, 1628 and 1629—Charles called Parliaments to fund his wars, but only twice did he obtain subsidies. Unlike the Dutch, the English were not fighting for independence and so there was less enthusiasm for war, especially for the French war which seemed to have little to do with the national interest.”Footnote 76 The king attempted to raise revenues despite parliamentary opposition to his foreign adventures through ship money: the right of the Crown to require the maritime towns and counties to furnish ships in time of war, or commute the duty with a monetary payment. However, after ship money proved insufficient, Charles I returned to Parliament in 1641 and made concessions such as the Triennial Act.
In the 1680s, James II nearly escaped the constraints faced by his father, Charles I. Like Frederick William's experience in Prussia, James's initial interaction with Parliament was favorable. He gained lifetime control over ordinary revenue sources: customs, excise, and hearth taxes. Although customary, this grant represented a conscious choice by Parliament to fund the government in anticipation that James would uphold basic civil liberties and promote good governance. The problem for Parliament was that these ordinary taxes yielded unexpectedly generous revenues because of booming colonial trade and plunder.Footnote 77 James then attempted to renege on his initial bargain with Parliament by building a 40,000-strong standing army without parliamentary approval. Members of Parliament were well aware of the dangers of a standing army, but powerless to stop it: “Many MPs complained vociferously in the parliamentary session that autumn that England now had an army which threatened ‘to give up all our liberties at once.’”Footnote 78 In response, James attempted to pack the Parliament with pliant supporters.Footnote 79
Yet ultimately, James's attempt to renege on Parliament failed. James's opportunity to coerce Parliament was less viable than Frederick William's in Brandenburg-Prussia; that is, the de facto credibility of parliamentary constraints, 
$\bar{q}$, was higher in England.Footnote 80 The sprawling Hohenzollern domains on the continent gave Frederick William his opportunity to intervene in a war, in contrast to England's insulated defensive position. This geopolitical reality also undermined domestic popular support for James's army, which he forcibly garrisoned across the country, producing broad grievances.Footnote 81 James's attempts to use coercion to renege on his bargain with Parliament yielded a divergent outcome: a popularly supported invasion by William III and the end of the Stuart dynasty.Footnote 82
Conclusion
We analyze a key question in historical European political development: did external wars promote parliamentary privileges? We provide a new theory to explain the varied consequences of international wars. Consistent with conventional implications, external threats make rulers more willing to offer parliamentary privileges in return for revenue. But wars often do not induce elites to trade taxes for privileges. Elites with immobile assets (like landed elites) fear invasion, which hinders their credibility to refuse funding for an unconstrained ruler. Elites with mobile assets (like merchants) can exit rather than fund a hopeless war effort, which makes them unwilling to fund even a constrained ruler. Only under specific conditions do external threats promote parliament, and under other conditions, the effect goes in the opposite direction. We apply these theoretical conditions to understand waves and reversals in European parliamentary development.
We conclude by considering how our findings interact with three different literatures. First, within Europe, we focus on cases before the nineteenth century. A central bargaining friction we highlight in the model, which war does not resolve, is that even after delegating some powers to parliament, historical European rulers still had problems committing to promises about expenditures and divisions of spoils. In larger territorial states, long traveling distances impeded frequent parliamentary meetings.Footnote 83 This explains why states with limited government and credible budget procedures tended not to arise before the nineteenth century in Europe. However, in the nineteenth century, parliaments not only re-emerged across Western Europe, they also gained greater de facto powers—typically controlling expenditures rather than only having veto rights over extraordinary taxes.Footnote 84 In Appendix C.2, we contend that external wars were relatively unimportant for these later changes. In our model, there are indeed narrow scope conditions under which war threats make rulers willing to countenance heavy restrictions on their royal prerogatives. However, it is more empirically plausible that changes in domestic conditions were pivotal for these later episodes. These cases thus create a historical segue from our theoretical framework—in which we analyze how external pressure creates top-down pressures for parliament—to the predominant focus in the formal democratization literature. These models analyze the consequences of bottom-up pressure based on domestic threats from below or intra-elite splits. Theoretical elaboration on divided elites could also be helpful for explaining aspects of earlier parliaments. Countries like England, with a unitary structure of government, differed from those like France, which had fragmented regional institutions in which monarchs could more easily divide and rule.Footnote 85
Second, we contribute to a broad, and largely recent, rethinking of the bellicose model of European state formation. To explain the rise of European parliaments, some focus on alternative factors such as urban development or the Catholic Church.Footnote 86 Nor are these critiques limited to Europe. Recent research establishes that external wars were unimportant in the formation of major states and regional order in East Asia. Japan and Korea emulated key components of China's bureaucratic institutions to ward off domestic competitors, rather than primarily to guard against an invasion threat from China.Footnote 87 The major states rarely fought each other, as wars tended to be confined to domestic challengers in the periphery.Footnote 88 Like the resurgence of European parliaments in the nineteenth century, these cases help us understand alternative factors that are important for developing state institutions.
Finally, we also contribute to a broader formal-theoretic literature on whether external wars lead to gains in fiscal capacity. Among formal theories, Besley and Persson posit a positive relationship.Footnote 89 Subsequent scholarship contends that the relationship depends on the nature of the wars, their expense, or bureaucratic capacity.Footnote 90 By highlighting additional factors that mitigate the relationship among external wars, fiscal capacity, and parliaments, we propose new considerations for understanding the bellicose origins of states.
Supplementary Material
Supplementary material for this article is available at <https://doi.org/10.1017/S0020818322000303>.
Acknowledgments
Earlier versions of this paper were presented at the University of Chicago; the University of Rochester; the Conflict and Peace, Research and Development workshop at the University of Michigan; the 2020 Formal Models of International Relations conference at Vanderbilt University; the 2020 American Political Science Association annual meeting; and the 2020 Southern Political Science Association annual meeting. For helpful comments and discussions, we also thank Scott Abramson, Brett Benson, Christian Davenport, Mark Dincecco, Chris Fariss, Bobby Gulotty, Jenny Jun, Mayya Komisarchik, Alex Lee, Jim Morrow, Kris Ramsay, Jim Robinson, Peter Schram, Brad Smith, Scott Tyson, Congyi Zhou, three anonymous reviewers, and the IO editors.
Open access
