Imperial Politics, English Law, and the StrategicFoundations of Constitutional Review in America

Sean Gailmard∗

December 23, 2017

Abstract

In the colonial period of American history, the royal Privy Council eval-uated acts of colonial assemblies on grounds of both policy and higher law.Assembly acts were sometimes nullified through this process. In this way,Council review established constitutional review of legislation in America.This paper presents a formal model of this review procedure to understandits strategic value to the imperial crown. I argue that Privy Council leg-islative review counteracted political pressure on imperial governors in thecolonies, to approve laws contrary to the empire’s interests. Optimal reviewin the model combines both legal and substantive considerations, becausethis gives governors strong incentives to avoid higher level review by veto-ing bad laws. Thus, review of legislation for consistency with higher lawhelped the crown to grapple with agency problems in imperial governance,and ultimately achieve more favorable policy enactments from colonies.

∗Professor, University of California, Berkeley. E-mail: gailmard@berkeley.edu. Thanks toDaniel Klerman for inspiring my interest in this topic, and to Deborah Beim, Dan Carpenter,Aditya Dasgupta, John Ferejohn, Lindsey Gailmard, Sandy Gordon, Kinch Hoekstra, RyanHubert, Lewis Kornhauser, Dan Lee, Nolan McCarty, Ken Shepsle, and seminar participantsat Harvard University and NYU Law School for helpful comments.

1

Two defining tenets of constitutionalism in the US are that (i) there are rela-

tively entrenched limits on the legislative powers of government,1 and (ii) specific

laws can be voided for violating these limits. Judicial review, a preeminent aspect

of constitutionalism in the US today, rests on these premises. But the applica-

tion of these precepts in US history encompasses other devices. For example,

Federalist-era understanding of the presidential veto was partly as a check on

unconstitutional legislation. At the state level, New York’s constitution empow-

ered a council of review to judge state laws on constitutional grounds until 1821,

without requiring any specific case to challenge them in court.

Scholars have long recognized that, though it affirmed constitutionalism in

the sense of these two tenets, the framing of the US Constitution in 1787 was not

the origin of this doctrine in America (McLaughlin 1932). Though framing and

ratification debates make some references to constitutional review of legislation,2

they make no pretense to having invented it (nor does Justice John Marshall’s

celebrated opinion in Marbury v. Madison (1803)). The colonial governments

prior to American independence also did not develop these two tenets of American

constitutionalism. No colony empowered a court or other body (outside the

assembly itself) to review or invalidate statutes relative to colonial charters before

independence (Treanor 2005, Ciepley 2017).3

Common arguments about the origins of political institutions in America,

and particularly limited government, look to early modern English practice. It

is sometimes argued, perhaps more as shorthand than as a full throated theory,

that the English imported key institutions with them to America (e.g. North

1990). Yet entrenched limits on legislative power and constitutional review do

not and did not hold the same sway in Britain, where parliamentary supremacy

has held since at least 1688, and royal prerogative before that.

1More precisely, the limits on government policy making power cannot themselves bechanged through an act of the government. Thus, while the limits may be amended or evolveover time, they are external to the government itself.

2For example, The Federalist no. 78 avers that “every act of a delegated authority, contraryto the tenor of the commission under which it is exercised, is void. No legislative act, therefore,contrary to the Constitution can be valid.” See also Brutus, Essays XV and XVI, 1788.

3Corporate and religious backgrounds of American colonies engendered written charters assources of entrenched constitutional limits of government structure and power (McLaughlin1932, Lutz 1988), but did not imply external review of government enactments within thecolonies (Ciepley 2017).

2

If external limits and constitutional review of legislation did not originate

with the US Constitution, colonial American constitutions, or the English Con-

stitution, where did they originate? Legal historians have answered this question

by focusing on the British imperial constitution—or what Mary Sarah Bilder

(2004) has called “the Transatlantic Constitution” (cf. Greene 1986). Unlike

Parliament, colonial assemblies exercised powers pursuant to grants by the crown.

Having empowered colonial assemblies to mitigate a more fundamental agency

problem between the crown and governor (Gailmard 2017), the crown’s basic

problem was to ensure that their enactments were beneficial, and not harmful,

to the interests of the crown and empire. To this end, colonial legislation was

reviewed by the Board of Trade and royal Privy Council in London. Review

combined considerations of the substantive merits of colonial policy, and legal

evaluations of consistency with the laws of England and the colony’s charter.

This legislative review habituated colonists to external limits of assembly power

relative to more fundamental or higher level law. Accordingly, historians have

identified review by the king-in-council, a corollary of the imperial origins of the

US, as the forerunner of the contemporary American practice of constitutional

review (Russell 1915; McGovney 1944; Bilder 2004; Bilder 2006; Ciepley 2017).

The purpose of this paper is to understand the incentives of the British crown

to adopt this forerunner of modern constitutionalism and constitutional review as

part of the imperial constitution.4 I argue that, even if the crown had no concern

for legal consistency, it benefitted from creating proto-constitutional hierarchies

for legislative review. That is, the legal institutions at the vanguard of American

constitutionalism served the purely substantive interest of the crown in controlling

colonial policy.

The logic works through the colonial governor, who was the first line of defense

against bad laws (from the empire’s perspective) from colonial assemblies. The

governor was an agent of the crown armed with veto power, and had more infor-

mation than the crown about the merits of colonial enactments (Greene 1898).

Ideally for the crown, the governor would approve good laws and veto bad ones,

4The right of the crown to review colonial acts was implied by their corporate structure(Bilder 2006, Ciepley 2017). The incentive of the crown to exercise that right, and its effects,are another matter, considered here.

3

thereby signaling a law’s quality (from the empire’s point of view) through its

approval decision. But the governor was often subjected to immense political

pressure from the assemblies (Greene 1963), so that their passage of a law was

no guarantee of its benefits to the crown (Russell 1915). Thus, as a reviewer of

legislation, the governor was an imperfect agent of the crown. The crown, too,

had an arsenal of sanctions to inflict on governors for passing bad laws—but these

were only meted out if the crown determined an inappropriate law was passed.

For this purpose, crown review was also necessary, but it carried an opportunity

cost due to the time and attention required.

This is where legislative review by the Privy Council enters the story. Legal

skepticism of colonial laws by the Council acted as a commitment or incentive to

review these laws in both legal and substantive terms—for only high substantive

benefits could outweigh the cost of legal inconsistency. That stronger commit-

ment to review, in turn, made crown sanctions on the governor for passing bad

laws more certain. This neutralized the assembly’s pressure on the governor,

and emboldened the governor to veto laws the crown and Council would find

undesirable. The model in this paper encapsulates and clarifies this logic.

This logic can backfire. If the Council’s weight on legal inconsistency is too

great, it may be so skeptical of laws approved by the governor that it vetoes

all laws without review. In this case, the governor’s best response is to approve

all laws, even undesirable ones—thereby validating the Council’s skepticism, but

also avoiding sanctions from the colonial assembly, and letting the Council do

the dirty work of preventing adoption of undesirable policies. But when the

Council places more moderate (but positive) weight on legal inconsistency, it

believes that some laws from colonial assemblies might have sufficient substantive

benefits to outweigh the cost of legal inconsistency. In this case, the Council’s

legal considerations give it a stronger incentive to review colonial laws, and thus

give the governor a stronger incentive to veto undesirable ones—thereby avoiding

the Council’s enmity, and communicating valuable information about a law’s

quality in the process.

The key point is that legal skepticism of colonial laws by the Privy Council

changed the bargaining dynamic within the colonies between governor and assem-

bly. Therefore, legal analysis and review in the Council was useful for solving a

4

political problem in the colonies. The optimal review that emerges in equilibrium

is sensitive to matters of both law and policy.

Understanding the development of constitutional limits and review of legisla-

tion is important because these are touchstones of American constitutionalism.

In the guise of judicial review, they are a crucial part of the policy making pro-

cess. While the link from colonial practice to the US Constitution was obviously

mediated by the framers, it is noteworthy all of them experienced legislative re-

view in colonial politics. Having recognized as colonists that their legislatures

were decidedly imperfect stewards of the public interest, Americans as constitu-

tion designers saw the value of external enforcement of limitations on legislative

power (Wood 1969). And, with the substitution of the people for the king as the

sovereign originator of constitutional charters, the theoretical step from colonial

to American legislative review was short (Ciepley 2017).

This paper also builds on the formal literature on legislative review, set mostly

in the contemporary US context. Most formal work on ex post review, judicial

or otherwise, highlights its upstream effects on the policies that get adopted in

the first place (Shipan 2000; Rogers 2001; Bueno de Mesquita and Stephenson

2007; Fox and Stephenson 2011; Fox and Vanberg 2014; Patty and Turner 2018).

Dragu and Board (2015) emphasize that judicial review can enhance information

transmission in policy making; in a similar spirit applied to administrative regu-

lation, Gailmard and Patty (2017) argue that a combination of political and legal

considerations in judicial review is useful for producing more information about

policy consequences. Beim et al. (2014), while not addressing judicial review

per se, also present a hierarchical model of review. A key difference is that the

intermediary can dissent but not veto, and there is one dimension of evaluation.

Cameron and Kornhauser (2012) and Clark and Carrubba (2012) present models

of review with multiple dimensions of evaluation (e.g. “legal” and “substantive

quality”); my approach shares this feature for reasons explained below.

Taken together, this paper and Gailmard (2017) have several implications for

understanding the foundational institutions of American government. First, two

of the institutional linchpins of the US Constitution—separation of powers and

constitutional review—can be foreseen in the strategic response by the British

empire to agency problems created in imperial governance. Second, and more

5

specifically, these aspects of American government co-evolved over 160 years of

imperial rule. The need for the institutions leading to constitutional review arose

only because of the design of institutions leading to separation of powers. Only

the empowerment of local assemblies across the several colonies, responding to

diverse conditions and their own parochial interests, generated the possibility of

incoherence and conflicting interest in colonial laws, and incongruence with the

laws of England. Third, at a broad level, the model implies that government

begets more government. With limited organizational forms and only partial un-

derstanding of their effects, no institutional device turns out to be an unalloyed

good for its designer. As is often the case in complex environments with incom-

plete contracts, the solution to one agency problem generated another, which in

turn necessitated a new institutional device to mitigate it. In this way, experi-

ence leads to a thickening of institutional structure, each layer of which reflects

a strategic response to problems revealed in lower layers.

The rest of the paper is organized as follows. In the next section I review key

institutions and practices of legislative review in the first British empire. Next

I lay out the formal model. Following this I analyze the model and identify key

equilibria. Then I present the key results of the paper, on optimal legislative

review from the crown’s point of view; finally, I conclude.

Crown Review of Colonial Legislation

As England’s New World colonies (and domestic politics) stabilized in the 17th

century, institutions emerged in London to monitor and oversee the colonies pur-

suant to the metropole’s interest. From 1696 to 1791, these duties were shared

by the Board of Trade (formally, “Lords Committee of Trade and Plantations”)

and the royal Privy Council. The Board included both members of Parliament

and the crown’s inner circle of advisors, and focused entirely on colonial matters.

The Council was comprised of the crown’s senior ministers, and in conjunction

with the crown itself, formed the kingdom’s central policy making body.

One of the most important duties of the Board and Council in colonial over-

sight was review of legislation (Dickerson 1912). All laws originating in royal

colonies were subject to this review upon passage. Most of the roughly 8500 laws

6

passed by future US states during the colonial period were subject to legisla-

tive review, with about 500 invalidated by metropolitan authorities in this way

(Simpson 1911).5

In this process, the Board took first-cut review of colonial laws and issued

advisory opinions to the Council, which in turn made a legally binding disposition

in the crown’s name. Formal dispositions were communicated to the colony, along

with reasons for nullification if applicable. The whole process typically took 1-

2 years and sometimes more (Simpson 1911). In practice, the Council almost

always followed the Board’s advice (Dickerson 1912), but most colonial laws were

in fact not formally reviewed at all, and instead were left to “lie by,” thus taking

effect in the colony without formal approval (Russell 1915).6 The reason for

this is that the Board, and especially Council, were frequently occupied with

numerous pressing issues of domestic and imperial politics, and full review of

colonial legislation consumed precious time (Dickerson 1912). In short, review

entailed an opportunity cost for the Board and Privy Council.

When the Board and Council did review colonial legislation, it was based on

both legal and substantive considerations (Russell 1915). The central formula in

legislative review was to prevent “repugnancy to the laws of England”—a con-

dition prohibited in every colonial charter—but also to permit salutary “diver-

gences” to address particular situations within a colony (McGovney 1945, Bilder

2004). This formula required assessment of a colonial law both in terms of its

legal position within English law, and the substantive conditions in the colony

motivating it. Legal inconsistency could be tolerated, if substantive conditions

in the colony rationalizing a law were sufficiently important.

For example, colonies were typically starved for cash as a result of mercantilist

trade policies. Yet colonial legislation issuing bills of credit was repeatedly nulli-

fied by the Privy Council, which viewed trade imbalances as a desirable feature.

5The Privy Council also sat as the highest court of appeals for lawsuits originating in thecolonies based on colonial law (Smith 1950, Bilder 2004), and thus reviewed colonial legislationin purely judicial terms as well. This paper focuses on legislative review because it was muchmore common and more visible to colonists (Smith 1950).

6After the Seven Years’ War, the crown increasingly required colonial laws to include “sus-pending clauses,” which prevented them from taking effect until formal disposition by theCouncil. Given the delays involved, this created a legal incoherence in the colonies, theirfrustration at which is reflected in the list of grievances in the Declaration of Independence.

7

However, in times of war, colonial militias required supply, which was provided

by colonial legislatures from internal taxes, and colonial credit was often the only

way to sustain it. Accordingly, in these circumstances the Council allowed colo-

nial credit bills to lie by (Russell 1915). Another illustration comes from colonial

inheritance law. With land much more plentiful than in England itself, numer-

ous colonies provided for equal division of land upon an estate holder’s death.

Metropolitan authorities repeatedly and summarily rejected these laws as con-

trary to the English custom of primogeniture. After decades of effort by colonial

lawyers and lobbyists, the Board and Council recognized that land availability

in the colonies rationalized equal division inheritance despite its inconsistency

with English law. In these cases, colonial legislation at variance with English law

was accepted by the Board and Council, because of substantive conditions in the

colonies.

The crucial assumption behind review for “repugnancy and divergence” was

that several organs of law—the laws of England, broadly construed; the charter

of the colony itself; the Board of Trade’s instructions to the colony’s governor;

and the policy of the empire with respect to trade—took precedence over colonial

legislative acts, and thus limited the powers of colonial assemblies. The Council

sometimes found these limits (substantive and procedural) in colonial charters,

imbuing them with a similar function as a modern American constitution. For

example, a Connecticut law against “Hereticks,” which caused issue for Quakers,

was invalidated as “contrary to the Liberty of Conscience Indulged to Dissenters

by the Charter granted to the Colony” (Russell 1915, p. 148). A New York law

granting residents preferential access to fisheries was invalidated as inconsistent

with its charter provision that “no subject of England should be debarred from

fishing on the sea coast” (Russell 1915, p. 148). Several Maryland laws, passed

in the absence of a governor or acting governor, were invalidated as contrary

to the procedural forms prescribed in its charter (Russell 1915, p. 145). Other

limits were found in a broad construction of English law and custom. Thus

various colonial legislation prescribing excessive and unusual punishments (such

as physical dismemberment or death) for minor crimes or crimes against Mosaic

law was nullified in Council on the grounds that such severe penalties “were never

inflicted by any law in His Majesty’s dominions” and were “contrary to reason

8

and the custom of the Kingdom” (Russell 1915, p. 144).7

Thus, unlike England’s Parliament, the American colonial assemblies were

limited and subject to external legal review from the beginning. In this way, a

limit on legislative authority provided by specific, entrenched sources (including

written documents) was built in to the American colonists’ legislative tradition.

This limit is the cornerstone of constitutional review in America, and is unlike

any limit operating on the parliament of Britain then or now.

The colonial governor, who was supervised by the Board of Trade, also played

an important role in review of colonial legislation. Governors exercised veto power

over enactments of colonial assemblies. The crown depended on governors not so

much for authority in law, which was in abundant supply on the Board and Coun-

cil (Simpson 1911). Rather, governors were crucial because they were informed

about substantive conditions within the colonies that might rationalize colonial

laws (Greene 1898). For this process to work in the crown’s interest, it was nec-

essary for the governor to pass laws only when substantive conditions supported

them, and to veto them otherwise. Crown authorities could then weigh sub-

stantive considerations, communicated by the governor’s decision, against legal

considerations.

However, the colonial assemblies often had their own designs, and not much

concern for repugnancy to the laws of England when colonial conditions called for

divergence (Bilder 2004). The assemblies thus exerted great pressure on gover-

nors to pass laws they desired, irrespective of the instructions (Greene 1898). For

example, assemblies would withhold the governor’s salary (which it was their obli-

gation to pay) or refuse to pass military supply bills until the governor complied

with their wishes (Greene 1898, Greene 1963). The Council could apply more

significant sanctions, up to and including recall (and loss of the entire stream

of future rents from office), for governors that passed laws clearly detrimental

to the empire’s interests (Greene 1898). But applying the sanctions required the

Board and Council to incur the opportunity cost of review to discover a violation.

Simpson (1911) summarizes:

7It is also clear from some of these examples that colonial legislative review was sometimesintended to protect colonists from intemperate acts of colonial assemblies. As fervently asthe assemblies might believe these acts were desirable, English authorities sometimes took adifferent perspective.

9

The effect which the royal disallowance had upon the colonial gover-

nors was very great. The closeness of the scrutiny with which they

inspected colonial laws before signing must have been considerably in-

creased by the knowledge that the matter would be carefully thrashed

out by the Board of Trade and they held responsible. Thus a very

considerable indirect influence was exercised over colonial legislation

(p. 56).

That “indirect influence” on colonial legislation, operating through the governors,

is the essence of the strategic model of legislative review presented below.

A Model of Legislation and Review

This section lays out the model of Council legislative review as an extensive form

game. The players are the governor G, and the Council C. The colonial assembly

A is left implicit and its strategic behavior is not considered here. Assume that

A has enacted a policy a “ 1; the issue is whether G and C should maintain it,

or revert to a status quo a “ 0.

Suppose there are two dimensions or attributes of policy a “ 1, denoted

∆1 P t´1, 1u and ∆2 P t´1, 1u. Assume that Prr∆1 “ 1s “ δ P p0, 1q; its

realization is observed by G but only its distribution is known by C. Assume

that ∆2 is drawn and fixed at the start of the game and observed by all players.

Implicitly, a “ 1 is judged relative to a “ 0.

Here ∆1 represents the substantive merits of the policy; it works as a “state

of the world” in typical models of policy making with uncertainty. I assume that,

as actors on the ground in the colony, the assembly and governor were better

informed than the Board and Council about local conditions and the importance

of a given assembly act in light of them (Greene 1898, Olson 1992). The Council

could acquire information on such factors, but it would take some time, and

being filtered through the perspective of other actors such as colonial agents and

merchant interest groups (Dickerson 1912, Russell 1915), the information was of

uncertain quality.

On the other hand, ∆2 captures broader, imperial-level legal considerations,

10

such as “repugnancy to the laws of England.” I assume that the Board of Trade

and Privy Council had relatively easy access to legal judgments on such consid-

erations (Simpson 1911, Dickerson 1912), and could share them with governors

through instructions and circulars (Labaree 1930). Thus, this attribute is com-

mon knowledge.

In the following, capital Greek letters are random variables; lower case Greek

letters are probabilities or exogenous (scalar) parameters; and Roman letters are

endogenous choices. Assume that all exogenous random variables are statistically

independent.

The sequence of play is as follows; the full extensive form is depicted in figure

1 at the end of the paper.

0. Nature draws ∆1,∆2 for a “ 1 and reveals the results as specified above.

1. G upholds (uG “ 1) or vetoes (uG “ 0).

(a) If G upholds, on to next step.

(b) If G vetoes, a “ 1 is rejected, game ends, G is sanctioned κA by A.

2. If G upholds, C can review (r “ 1) or not (r “ 0) at cost γ, and learns ∆1

with probability λ.

(a) If ∆2 “ ´1, or review is informative (Λ “ 1), on to next step.

(b) If ∆2 “ 1, and either review is uninformative (Λ “ 0), or no review is

performed (r “ 0), a “ 1 stands, game ends.

3. C upholds (uC “ 1) or vetoes (uC “ 0). If review is performed (r “ 1), and

is informative (Λ “ 1), and ∆1 “ ´1, G is sanctioned κC by C.

C’s review is a function of ∆2, which C observes; it is informative about ∆1.

Let Λ P t0, 1u denote whether C learns ∆1, given that review is performed; then

PrrΛ “ 1s “ λ. Also note that a veto by C requires either information about

∆1, or legal inconsistency (∆2 “ ´1). That is, C cannot reject a “ 1 without

compelling information. The Board of Trade and Privy Council were law-bound

institutions and required reasons for their decisions (Russell 1915).

Let x P t0, 1u denote the colonial policy chosen in this extensive form. If

11

mintuG, uCu “ 0, then x “ 0; otherwise, x “ 1. Ex post utilities are:

UC “

ˆ

∆1 ` σ

ˆ

∆2 ´ 1

2

˙˙

x´ rγ,

UG “ ∆1x´ uGrΛ

ˆ

1´∆1

2

˙

κC ´ p1´ uGqκA.

Thus, the Council’s utility is composed of policy payoffs and review costs. When

a “ 1 is upheld, C obtains utility ∆1 ´ rγ if ∆2 “ 1 (legal consistency of a “ 1);

and obtains ∆1 ´ σ ´ rγ if ∆2 “ ´1 (legal inconsistency of a “ 1).8 Here σ ą 0

captures the salience of legal relative to substantive issues for C. For example,

in time of war, a colonial credit bill that allows purchase of militia supplies may

be very important—even if inconsistent with the laws of England. In this case

the salience of ∆2 is low. But in time of peace, the inconsistency of the same bill

with English law regulating currency looms larger in imperial calculations.

The governor’s utility is composed of policy payoffs and sanction costs from

the Board/Council or assembly. If G upholds a “ 1 (so uG “ 1), G incurs a

sanction κC if the Council reviews the policy, the review is informative, and the

policy is substantively undesirable (∆1 “ ´1). If G vetoes a “ 1 (so uG “ 0), G

incurs a sanction κA from the colonial assembly with certainty.

Note that governor and Council share common interests in terms of the sub-

stantive attributes of the policy, but G does not internalize C’s considerations of

legal consistency. Substantively, the legal aspects in ∆2 are broader, imperial-

level considerations that were not the province of governors, who focused primar-

ily within their own colonies (Greene 1898).9 Additionally, both players obtain

policy utility 0 from the status quo policy a “ 0; this simply reflects that the

utility of a “ 1 is judged relative to the known utility of a “ 0.

8Thus, legal consistency confers no utility in itself, whereas legal inconsistency is costly initself. Of course, the level of utility is only a normalization; the important point is the differencecreated when ∆2 “ ´1 vs. ∆2 “ 1.

9The key results below would hold if G did internalize the cost of legal inconsistency, as longas assembly sanctions are large enough relative to legal salience.

12

Analysis

The governor is privately informed about substantive attributes ∆1 of interest to

the Council. A key question is whether G’s veto decision uG signals this informa-

tion. Ideally for the Council, the governor would uphold if and only if ∆1 “ 1,

i.e., would fully reveal its information through its veto. The natural equilibrium

concept to investigate this issue is perfect Bayesian equilibrium, which preserves

sequential rationality under incomplete information.

In the following, φ∆1 ” PrruG “ 1|∆1s denotes G’s probability of upholding

action a “ 1 as a function of ∆1; ρ∆2 ” Prrr “ 1|∆2, uG “ 1s denotes C’s

probability of review as a function of ∆2. Let χ denote G’s belief that x “ 1 is

the final policy, given that G itself upholds (uG “ 1).

Lemma 1 If ∆1 “ 1, G upholds pφ1 “ 1q.

The logic is simple, so formal proof of this (and all lemmas) is omitted for brevity.

Upholding when ∆1 “ 1 gives G expected utility χ ě 0, while vetoing gives

´κA ă 0. Upholding a good bill gives a lottery over favorable policy benefits,

and avoids political pressure from A. So G always upholds a “ 1 when it is

beneficial on the merits in the colony. The only question about G’s strategy is

φ´1. Say that G separates if φ´1 “ 0 and G pools if φ´1 “ 1.10

Privy Council Disposition

Consider the final node of the game, C’s decision to uphold or invalidate the

colonial legislation. By the time this decision is made, review is already done,

and the decision is made on the policy and legal merits given the information at

the time.

If C is informed of ∆1, then C has full information about both dimensions

∆1,∆2. If ∆2 “ 1 (legal consistency), C will uphold if and only if ∆1 “ 1 (high

substantive merit). If ∆2 “ ´1 (legal inconsistency), C will uphold if and only

if ∆1 ą σ. This is only possible if σ ă 1, i.e., the salience of the legal dimension

10Lemma 1 also implies that there is no information set for C that is off the equilibrium path,which simplifies the analysis of beliefs in PBE.

13

is lower than the policy dimension. That is, if σ ą 1 and ∆2 “ ´1, then C will

invalidate the statute regardless of the substantive merits ∆1.

To evaluate C’s decision in case it is not informed of ∆1, define C’s belief

Prr∆1 “ 1|uG “ 1s ” d. Recalling that G always upholds when ∆1 “ 1 (lemma

1), Bayes’s rule yields

d “δ

δ ` p1´ δqφ´1

, (1)

and Er∆1|uG “ 1s “ 2d ´ 1. These values hold both when C reviews but it is

uninformative, and when C does not review; that is, failure of review to reveal

∆1 is not informative about the value of ∆1.

Recall that when C is uncertain of the substantive merits ∆1, crown veto

uC “ 0 is possible only with legal inconsistency (∆2 “ ´1). C’s utility from

overturning the policy is 0. Its expected payoff from upholding the policy in this

case is 2d´1´σ. So in this case, C upholds a “ 1 if and only if φ´1 ď`

δ1´δ

˘ `

1´σ1`σ

˘

.

If the right hand side is at least 1, then C upholds when uncertain about ∆1 for

any strategy by G—even full pooling. This occurs when

σ ď 2δ ´ 1. (2)

Say that C is legally flexible if inequality (2) holds, and C is legally fastidious

otherwise. When C is legally flexible, it is willing to overlook legal inconsistency

and uphold a “ 1 even without hard information about ∆1. If C is legally

fastidious, it requires hard information of policy benefits (∆1 “ 1) in order to

overcome legal inconsistency.

Several polar cases establish some intuition. If σ ą 1, so salience is very

high, even certainty of high substantive merit (∆1 “ 1) is not enough to save

a “ 1 when it is legally problematic for C (∆2 “ ´1). On the other hand,

when σ “ 0, so the legal dimension does not matter, the Council would support

a “ 1, even with full pooling by G, so long as ∆1 “ 1 is more likely than not

(δ ą 1{2). In general, the greater the salience of the legal dimensions (σ), the

larger the expected policy benefits must be to ensure Council approval under a

given strategy by G.

Gathering these results together, the Council’s optimal disposition, as a func-

tion of ∆2, its information about ∆1, and G’s strategy φ´1, is as follows. When

14

C has hard information about ∆1 from review,

u˚Cp∆2 “ 1q “

$

&

%

1 if ∆1 “ 1

0 otherwise(3)

u˚Cp∆2 “ ´1q “

$

&

%

1 if ∆1 “ 1 and σ ď 1

0 otherwise.(4)

When, instead, C believes Prr∆1 “ 1s “ d,

u˚Cp∆2 “ ´1q “

$

&

%

1 if φ´1 ď`

δ1´δ

˘ `

1´σ1`σ

˘

0 otherwise.(5)

Privy Council Review of Policy Merits

Moving one stage up the game tree, to the decision of the Privy Council to review

substantive merit ∆1, C faces different tradeoffs about review depending on legal

consistency ∆2.

First, if ∆2 “ 1, repeal of the colony’s policy (uC “ 0) is only possible if

r “ Λ “ 1. C’s expected utility of not reviewing is Er∆1|uG “ 1s “ 2d´ 1. If C

does review, its expected utility is λd`p1´λqp2d´1q´γ. This is the probability

that review is informative times the (updated given uG “ 1) probability that

∆1 “ 1; plus the probability that review is uninformative times the expected

utility of a “ 1 given that C has no grounds to reverse.11

Review is beneficial if γ ď λp1 ´ dq, or its cost is less than its probability-

weighted marginal benefit in policy terms. From the definition of d in equa-

tion 1, and provided γ ă λ, this condition can be further expressed as φ´1 ě`

δ1´δ

˘

´

γλ´γ

¯

. So, as G’s approval decision becomes less informative about ∆1, it

is easier to satisfy C’s constraint for review. At the same time, if the review cost

is very low (γ small), or the probability of beneficial policy is very low (δ small),

C prefers to review even if G’s approval is very informative. Given ∆2 “ 1,

informative review is C’s only ticket to reverse a “ 1. When δ is very small, C

11The λp1´dq probability event that review is informative, and ∆1 “ ´1, receives no weight,because C will reverse to x “ 0 in this case and obtain 0 policy utility.

15

believes this reversal is likely to be beneficial.

Second, if ∆2 “ ´1, C can reverse a “ 1 even without review. As noted

above, if φ´1 ą`

δ1´δ

˘ `

1´σ1`σ

˘

, C will reject a “ 1 absent hard information that

∆1 = 1. Thus C’s expected utility given review is λdp1 ´ σq ´ γ. C’s expected

utility given no review is 0. This case is possible (but not assured) if C is legally

fastidious (σ ą 2δ ´ 1: inequality 2 fails).

If ∆2 “ ´1 and φ´1 ď`

δ1´δ

˘ `

1´σ1`σ

˘

, C will uphold a “ 1 in the absence of

hard information about ∆1. In this case, C’s expected utility given review is

λdp1 ´ σq ` p1 ´ λqp2d ´ 1 ´ σq ´ γ. C’s expected utility given no review is

2d ´ 1 ´ σ. This case is assured if C is legally flexible (σ ď 2δ ´ 1: inequality 2

holds).

Gathering these results together and using the posterior d (equation 1), C’s

best response review R˚∆2is characterized as follows.

If ∆2 “ 1,

R˚1 “

$

&

%

1 if φ´1 ě`

δ1´δ

˘

´

γλ´γ

¯

ą 0

0 otherwise,(6)

If ∆2 “ ´1 and φ´1 ą`

δ1´δ

˘ `

1´σ1`σ

˘

,

R˚´1 “

$

&

%

1 if φ´1 ď`

δ1´δ

˘

´

λp1´σq´γγ

¯

0 otherwise.(7)

If ∆2 “ ´1 and φ´1 ď`

δ1´δ

˘ `

1´σ1`σ

˘

,

R˚´1 “

$

&

%

1 if φ´1 ě`

δ1´δ

˘

´

γλp1`σq´γ

¯

ą 0

0 otherwise.(8)

In equation (6), λ ą γ defines the non-negativity constraint; in equation (8),

λp1 ` σq ą γ does. Informative review precludes upholding a policy that is

undesirable on one or both dimension; thus, it adds utility 1 or 1`σ, respectively,

with probability λ. This probability-weighted gain must exceed the cost of review,

for review to be worthwhile in this case.

16

Some special cases are straightforward.

Lemma 2 If G separates and γ ą 0, C never reviews pρ∆2 “ 0q.

This is apparent from inequalities (6)-(8) for R˚. Review is useful to C for two

reasons. First, it gives C information. Second, if ∆2 “ 1, it gives C a chance to

repeal a “ 1. If G separates, review is not informative. Given u˚C , separation also

implies C is certain to prefer upholding a “ 1, so review is also not necessary to

enable repeal. Since review is costly (γ ą 0) and it confers no benefit when G

separates, C does not review.

Lemma 3 If φ´1 ą 0 and γ is sufficiently small, then C reviews with positive

probability.

This is apparent from inequalities (6)-(8) for R˚; in each case, for a given φ´1 ą 0,

γ Ñ 0 triggers R˚ “ 1. If information about ∆1 and/or the right to invalidate

a “ 1 are sufficiently cheap, C always prefers to purchase.

Governor’s Review

G’s decision to uphold, uG, balances the expected cost of sanctions from the

Assembly below, sanctions from the Board/Council above, and bad policy. Recall

that uG “ 1 with certainty when ∆1 “ 1 (lemma 1). The remaining question

is uG when ∆1 “ ´1. Let φ´1 denote C’s belief about PrruG “ 1|∆1 “ ´1s at

the time it acts, which then maps into the posterior d “ Prr∆1 “ 1|uG “ 1s

via equation (1). Then G believes that C will review with probability ρpφ´1q, as

specified in equations (6)-(8). Further, G believes x “ 1 is the final policy with

probability χ, as specified in equations (3)-(5). In particular, if ∆2 “ ´1 and

φ´1 ď p δ1´δqp1´σ

1`σq, or if ∆2 “ 1, then χ “ 1 ´ ρλ. In these cases, informative

review would reveal ∆1 “ ´1 and lead to a veto, but otherwise C will uphold.

If ∆2 “ ´1 and φ´1 ą pδ

1´δqp1´σ

1`σq, then χ “ 0. In this case, informative review

would again lead to a veto; but since the Council is sufficiently skeptical given

φ´1, so would uninformative review or no review at all.

Putting these elements together, given ∆1 “ ´1, ∆2, and ρ, G’s expected

17

utility from uG “ 1 is ´χ´ ρλκC , or

EU1G “

$

’

’

’

&

’

’

’

%

´p1´ ρλq ´ ρλκC if ∆2 “ 1

´p1´ ρλq ´ ρλκC if ∆2 “ ´1 and φ´1 ď`

δ1´δ

˘ `

1´σ1`σ

˘

´ρλκC if ∆2 “ ´1 and φ´1 ą`

δ1´δ

˘ `

1´σ1`σ

˘

.

(9)

G’s expected utility from uG “ 0 is EU0G “ ´κA. Thus, G separates if ρλpκC ´

1q ě κA ´ 1 in the first two cases of equation (9), and if ρλκC ě κA in the third.

Note that the third case is only possible if C is legally fastidious (σ ą 2δ ´ 1).

Equilibria

The historically interesting cases involve κC ą κA ą 1.12 In this case, Assembly

and Council sanctions are strong enough to affect the Governor’s veto decisions,

which occurred repeatedly. If Assembly and Council sanctions meet these condi-

tions, there is no equilibrium in which G separates (i.e., approves if and only if

∆1 “ 1.)

Lemma 4 If κA ą 1 and γ ą 0, there is no separating perfect Bayesian equilib-

rium.

Recall that (by lemma 2) ρ “ 0 in a separating equilibrium. So κA ą 1 implies

ρλpκC ´ 1q ă κA ´ 1, and thus φ´1 “ 1. If there were a separating equilibrium,

lemma 2 would imply that C never reviews (and upholds whenever possible).

But if C never reviews and κA is large, G is tempted to approve laws even when

∆1 “ ´1, since the cost of bad policy is less than that of sanctions from the

Assembly. Intuitively, lemma 4 means that if political pressure from the Assembly

on the Governor is relatively high, the Council must put countervailing pressure

on the Governor (in the form of review and possible sanctions) to obtain any

12A separating PBE is possible with κA ă 1; the Governor would rather face the cost ofassembly sanctions than the cost of bad policy. But if κC ă 1 and Privy Council review is bothvery informative and very cheap, a pooling equilbrium is also possible. In this case G prefersto “pass the buck” to C, avoid the cost of assembly sanctions, and let C clean up bad policychoices through its review. This is the pathological effect of judicial review often highlightedin formal models on the topic.

18

information from the Governor’s decision. If C rubber stamps colonial legislation,

that rubber stamp will propagate down through the legislative review process.

While separation is not possible with high political pressure from the Assem-

bly, partial information may still be conveyed to C by G’s disposition. In the

remainder I focus on the most informative possible PBE for given parameter con-

figurations. The most interesting case is when ∆2 “ ´1, so a “ 1 is inconsistent

with English law and imperial policy—here the Council is predisposed against

approval because of its legal consequences.

Recall that C can reverse the policy even without review in this case. Note

from equation (9) that as long as κC ą 1, G’s expected utility from upholding

a “ 1 is decreasing in C’s review probability ρ. Ideally, the needle G wishes to

thread is upholding the policy (thus avoiding κA) while also ensuring that C does

not review (thus avoiding κC). While this would potentially expose G to the cost

of bad policy, it would avoid the larger (given κA ą 1) cost of A’s sanctions. If C

is legally fastidious (inequality (2) fails), it is possible for G to thread this needle

by pooling.

Proposition 1 Suppose κA ą 1, κC ą 1, and ∆2 “ ´1. If C is legally fastidious

pσ ą 2δ ´ 1q, there is a perfect Bayesian equilibrium in which G pools pφ˚´1 “

φ˚1 “ 1q, C never reviews pρ˚ “ 0q, and the policy is always reversed pu˚C “ 0q.

Formal proof of this and all propositions is in the appendix. Intuitively, C is

already skeptical of a “ 1, and will reject it absent hard evidence proving ∆1 “ 1.

But the value of review to C is therefore decreasing in φ´1, as a greater chance

of pooling by G only makes C more skeptical of an upheld policy.

When C is legally fastidious, ∆2 is very important relative to ∆1. Then

∆2 “ ´1 is sufficient for rejection of a “ 1, regardless of ∆1 and without any

further review. This reflects the hierarchy of rationales sometimes observed by the

Council in review of colonial legislation. Sometimes, inconsistency with English

law or imperial policy was enough to get a colonial statute invalidated in the

Council—irrespective of its merits for the colony. When these breaches of English

law were especially clear and important to the Board or Council, review on the

merits may not even take place; the inconsistency was enough to invalidate the

law. What proposition 1 reveals is that this strategy depends on the governor’s

19

behavior. It is a reasonable response by the Council to both high importance of

legal issues, and low confidence that the governor was reversing bad policies rather

than sending them for review. In other words, summary rejection of colonial

policy on legal grounds alone, without substantive review, is partly a recognition

by the Council of the political interaction between governor and assembly that

often led to approval of questionable laws that were not worth the Board or

Council’s time to investigate fully.

When C is legally flexible (σ ď 2δ´ 1), G is not able to obfuscate its way out

of review by C. When G pools, the value of information to C is high in this case,

and so G is likely to incur sanctions from C. G can only reduce the probability

of C’s review by reducing the probability of passing laws when ∆1 “ ´1.

Proposition 2 Suppose κA ą 1 and ∆2 “ ´1. If C is legally flexible pσ ď

2δ ´ 1q, λpκC ´ 1q ě κA ´ 1, and γ ď λp1 ` σqp1 ´ δq, there is a mixed strategy

perfect Bayesian equilibrium in which G upholds a “ 1 with probability

φ˚∆1“

$

&

%

`

δ1´δ

˘

´

γλp1`σq´γ

¯

if ∆1 “ ´1

1 if ∆1 “ 1

and C reviews an upheld policy with probability

ρ˚∆2“´1 “κA ´ 1

λpκC ´ 1q.

Therefore, by applying countervailing pressure from C to G (to balance the pres-

sure from A on G), the Council can extract some information about ∆1 from G’s

review. When κA ą 1, the governor faces high pressure from the assembly. With-

out second stage review by the Council, the governor would pass all laws, despite

complete agreement between G and C on policy terms. Second stage review is

necessary to expose G to a chance of even greater sanctions at the hands of the

Board and Council.

Note that in this equilibrium, increasing the Council’s weight on legal issues

σ decreases the governor’s probability of upholding a bad law φ˚´1, but does not

affect the Council’s frequency of review ρ˚. This is exactly like a mixed Nash

equilibrium in simple 2ˆ2 auditing games. There, the auditor’s gain from detect-

20

ing malfeasance affects the probability the perpetrator commits it, but not the

probability of an audit. In a mixed equilibrium, the auditor is indifferent between

auditing and not, which is determined by the probability of a violation by the

perpetrator. As the auditor’s gain from detecting malfeasance rises, the perpe-

trator’s chance of committing it must fall to preserve the auditor’s indifference.

Other comparative statics in this equilibrium have the natural interpretation.

But there are conditions for it to hold. Not only must C be legally flexible,

but the Council’s sanction and informativeness of review must be large enough.

Corollary 1 Suppose κA ą 1, ∆2 “ ´1, and C is legally flexible. Then if

λpκC ´ 1q ă κA ´ 1 or γ ą λp1` σqp1´ δq, G pools pφ˚1 “ φ˚´1 “ 1q in all perfect

Bayesian equilibria.

Given κA ą 1, λpκC ´ 1q ă κA ´ 1 implies C must review with probability

greater than 1 in order to make G indifferent about approving and vetoing a bad

policy. Since C’s sanction on G is diluted by the probability review is informative

(λ), C’s sanction must be relatively greater to compensate for this and remain

effective. Otherwise, G prefers to pool. G also prefers to pool when there is no

chance that C will review, and this is the case when γ ą λp1 ` σqp1 ´ δq. Note

that this γ threshold is greater than when ∆2 “ 1; because legal inconsistency

lowers the value of a “ 1 to the Council, it will pay greater costs to investigate.

Now consider the case of ∆2 “ 1, so the colonial policy is not inconsistent

with English law and imperial policy. Without informative review, C has no hard

evidence to justify reversing the policy.

Proposition 3 Suppose κA ą 1 and ∆2 “ 1. Then if λpκC ´ 1q ě κA ´ 1 and

γ ď λp1 ´ δq, there is a mixed strategy perfect Bayesian equilibrium in which G

upholds a “ 1 with probability

φ˚∆1“

$

&

%

`

δ1´δ

˘

´

γλ´γ

¯

if ∆1 “ ´1

1 if ∆1 “ 1

21

and C reviews an upheld policy with probability

ρ˚∆2“1 “κA ´ 1

λpκC ´ 1q.

Corollary 2 Suppose κA ą 1 and ∆2 “ 1. Then if λpκC ´ 1q ă κA ´ 1 or

γ ą λp1´ δq, G pools pφ˚1 “ φ˚´1 “ 1q in all perfect Bayesian equilibria.

If λpκC ´ 1q ă κA´ 1, the Council’s sanctions, when weighted by the probability

they will be meted out, are not very potent relative to the Assembly’s. The

governor would rather face the chance of Council sanctions, and bad policy, than

the certainty of Assembly sanctions. The governor therefore pools even if the

Council reviews with certainty. The Council reviews with certainty if γ is small

enough, and not at all otherwise.

If γ ě λp1´ δq, review is too costly for C, relative to its expected benefits, to

ever engage in it—even if G is pooling. With ρ “ 0, G in turn passes everything

to avoid Assembly sanctions—which are worse for G than the cost of bad policy.

Law and Policy in Optimal Review

The previous analysis takes σ, the reviewer’s salience of legal consistency, as

given. Examples discussed above suggest that σ could vary over time for a given

issue, depending on conditions in the colony. But σ can also be understood in

general as a function of the institutional design of the reviewing body, which was

influenced by the Crown. For example, stacking the Board of Trade or Privy

Council with jurists, or charging a subset such as the Judicial Committee with

review, would tend to increase σ. As it happened, legislative review in the Board

and Council was dominated by politicians and statesmen, not jurists per se. Their

foremost concern with legal consistency was about ensuring intelligibility of the

colonies to the metropole (thus facilitating their oversight), and their viability as

settlements for Britons (Dickerson 1912).

In view of this institutional interpretation, this section considers the Crown’s

preferences over σ for its reviewing agent. It asks, if the Crown could choose the

reviewer’s σ before the play of the extensive form (and its choice were common

22

knowledge), what σ would it choose? The central finding is that, even if the

Crown has no intrinsic concern for legal consistency, it benefits from injecting

a legal dimension into review of colonial legislation. Crown support for legal

structures follows entirely from Crown focus on substantive concerns.13

To see this formally, let K denote the Crown and C denote the Council,

the reviewing agent. Further, let σK denote the Crown’s actual concern for legal

consistency, and σC the concern of the reviewing agent baked into its institutional

structure. The Crown is able to choose σC , and will do so to maximize its utility

given σK . Note that σC has no effect on a law’s disposition (or, thus, K’s utility)

when it is consistent with English law (∆2 “ 1). Therefore, assume ∆2 “ ´1.

Proposition 4 There is a critical value σK ă 1 such that, whenever σK ď σK,

the Crown’s optimal reviewer is legally flexible with σ˚C “ 2δ ´ 1 ą 0. Whenever

σK ą σK, the Crown’s optimal reviewer is legally fastidious.

Here, then, is a formal exposition of the rationale for imperial development

of colonial constitutionalism. A Crown preference for substance alone generates

a legal structure to enforce the consistency of laws. Even if the crown itself has

little or no concern for legal consistency, it prefers agents like the Board of Trade

and Privy Council that review colonial enactments in part on the basis of legal

consistency within the empire.14

Intuitively, the reviewer’s legal skepticism benefits the crown through an in-

direct effect on governors. Up to σC “ 2δ´ 1, legal skepticism raises the value of

information about ∆1 to the reviewer. In turn, to keep down the expected cost

of sanctions meted out by the Board in case he is caught passing bad laws, the

governor responds by vetoing more of them. However, if some legal skepticism is

good for the crown, more is not necessarily better. Fastidious insistence on legal

13There is a natural analogy here with modern executive politics in the US. The Office ofInformation and Regulatory Affairs (OIRA), on behalf of the President, reviews regulationsfrom federal agencies in both substantive and legal terms. The analogy puts the President inthe role of the Crown; OIRA in the role of the Council; agency heads in the role of the governor;and agency staff in the role of the Assembly. Further analysis of this connection is left for futureresearch.

14Ruling on this basis is of course beneficial if the crown has a sincere preference for legalconsistency. But it is more interesting that ruling on this basis is in the crown’s interest even ifit has no intrinsic concern for legal consistency at all, and cares only about substantive policyquality on the merits.

23

consistency (σC ą 2δ ´ 1) makes the Council highly skeptical of colonial legisla-

tion ex ante. Its value of information from review is very low. In its desire to

evade both assembly sanctions and Council review, the governor plays into this

skepticism by obfuscating completely through its approval decision. But that

very obfuscation underlies the skepticism of C which makes review irrational in

the first place. This logic underlies the rubber stamp approval of all laws, good

and bad, by the governor when the Council is legally fastidious (proposition 1).

In short, the optimal review for the Crown blends considerations of both law

and substantive policy. This is so even when the Crown has no intrinsic concern

for legal consistency.

Conclusion

In empowering colonial assemblies, the crown mitigated one agency problem with

colonial governors (Gailmard 2017), but raised another. How could the crown

induce the governor to resist assembly enactments that were contrary to the

crown’s interests, in the face of immense political pressure from the assemblies

on the governor to pass them? In this paper, I argue that legislative review

by royal agents in the Board of Trade and Privy Council helped to solve this

problem.

The Board and Council’s concern for consistency with higher English and

imperial law added a layer of skepticism to review of colonial legislation. When

this concern is moderate, but not exceedingly high, it implies that widespread

rubber stamping of colonial laws by the governor will result in scrutiny from

the Board of Trade. In the case of an undesirable enactment, that scrutiny will

bring the governor sanctions from above. When sufficiently potent and likely,

the governor knew he would be better off accepting sanctions from below. The

governor’s best response was to veto more bad laws, making his approval per se

informative to the Board and Council. However, the value of this skepticism had

its limits: extreme concern for legal consistency inclines the Council to veto all

colonial laws, for even meritorious ones would not confer sufficient benefits to

overcome possible legal downsides. In this case, the governor escapes assembly

sanctions—and the cost of enacting bad laws—by rubber stamping colonial laws,

24

good and bad.

Privy Council review presupposed that there was a higher body of law that

could take precedence over colonial law (Russell 1915). Colonial lawmakers ac-

knowledged this and leveraged these arguments in defense of their enactments

(Bilder 2004). Thus, colonial legislatures came of age alongside review of their

enactments for consistency with, and possible invalidation from, more fundamen-

tal law. For the entire colonial period, legislative power in America never existed

without this proto-constitutional review. At the same time, I argue above that

the crown’s incentive to institute this review stemmed from its creation and sup-

port for separation of powers as a check on colonial governors. Thus, separation of

powers and constitutional review coevolved in America as the crown strategically

pursued its interests in imperial governance.

25

References

Beim, D., A. V. Hirsch, and J. P. Kastellec (2014). Whistleblowing and com-

pliance in the judicial hierarchy. American Journal of Political Science 58 (4),

904–918.

Bilder, M. S. (2004). The Transatlantic Constitution. Harvard University Press.

Bilder, M. S. (2006). The corporate origins of judicial review. The Yale Law

Journal , 502–566.

Bueno de Mesquita, E. and M. C. Stephenson (2007). Regulatory Quality under

Imperfect Oversight. American Political Science Review 101 (3), 605–620.

Cameron, C. and L. Kornhauser (2012). Modeling collegial courts (3): Adjudi-

cation equilibria. SSRN working paper .

Ciepley, D. (2017). Is the us government a corporation? the corporate origins of

modern constitutionalism. American Political Science Review 111 (2), 418–435.

Clark, T. S. and C. J. Carrubba (2012). A theory of opinion writing in a political

hierarchy. The Journal of Politics 74 (2), 584–603.

Dickerson, O. M. (1912). American Colonial Government 1696-1765: A Study of

the British Board of Trade in its Relation to the American Colonies. Cleveland,

OH: Arthur H. Clark Company.

Dragu, T. and O. Board (2015). On judicial review in a separation of powers

system. Political Science Research and Methods 3 (3), 473–492.

Fox, J. and M. C. Stephenson (2011). Judicial review as a response to political

posturing. American Political Science Review 105 (2), 397–414.

Fox, J. and G. Vanberg (2014). Narrow versus broad judicial decisions. Journal

of Theoretical Politics 26 (3), 355–383.

Gailmard, S. (2017). Building a new imperial state: The strategic foundations

of separation of powers in america. American Political Science Review 111 (4),

668–685.

26

Gailmard, S. and J. W. Patty (2017). Participation, process, and policy: The in-

formational value of politicized judicial review. Journal of Public Policy 37 (3),

233–260.

Greene, E. B. (1898). The Provincial Governor in the English Colonies of North

America. New York: Longmans, Green and Co.

Greene, J. P. (1963). The Quest for Power: The Lower Houses of Assembly in

the Southern Royal Colonies, 1689-1776. New York: W. W. Norton.

Greene, J. P. (1986). Peripheries and Center: Constitutional Development in

the Extended Polities of the British Empire and the United States, 1607-1788.

Athens, GA: University of Georgia Press.

Labaree, L. W. (1930). Royal Government in America: A Study of the British

Colonial System before 1783. New Haven: Yale University Press.

Lutz, D. S. (1988). The origins of American Constitutionalism. Baton Rouge:

Louisiana State University Press.

McGovney, D. O. (1944). The british origin of judicial review of legislation.

University of Pennsylvania Law Review 93 (1), 1–49.

McGovney, D. O. (1945). The british privy council’s power to restrain the leg-

islatures of colonial america: Power to veto. University of Pennsylvania Law

Review 94 (1), 59–93.

McLaughlin, A. C. (1932). The Foundations of American Constitutionalism. New

York.

North, D. C. (1990). Institutions, Institutional Change, and Economic Perfor-

mance. New York: Cambridge Univ Press.

Olson, A. G. (1992). Eighteenth-century colonial legislatures and their con-

stituents. The Journal of American History 79 (2), 543–567.

Patty, J. and I. Turner (2018). Ex post review and expert policymaking: When

does oversight reduce accountability? Journal of Politics .

27

Rogers, J. R. (2001). Information and judicial review: A signaling game of

legislative-judicial interaction. American Journal of Political Science, 84–99.

Russell, E. B. (1915). The Review of American Colonial Legislation by the King

in Council. New York: Columbia University Press.

Shipan, C. R. (2000). The legislative design of judicial review: a formal analysis.

Journal of Theoretical Politics 12 (3), 269–304.

Simpson, H. L. (1911). A Study of the Royal Disallowance of Colonial Laws,

1680-1765. Madison: University of Wisconsin Press.

Smith, J. H. (1950). Appeals to the Privy Council from the American Plantations.

New York: Octagon Books.

Treanor, W. M. (2005). Judicial review before Marbury. Stanford Law Review ,

455–562.

Wood, G. S. (1969). The Creation of the American Republic, 1776-1787. Chapel

Hill: University of North Carolina Press.

28

Appendix: Formal Proofs

Proof of Proposition 1.

• C’s review: If σ ą 2δ ´ 1 and ∆2 “ ´1, inequality (7) indicates that for

pφ˚1 , φ˚´1q “ p1, 1q, C’s best response is R´1 “ 0. Therefore ρ˚´1 “ 0.

• G’s approval: If σ ą 2δ´1, ∆2 “ ´1, and C believes G is pooling, equation

(9) indicates EU1G “ ´ρ

˚´1λκC “ 0 ą ´κA “ EU0

G. Given this and lemma

1, pφ˚1 , φ˚´1q “ p1, 1q is sequentially rational, which validates C’s belief that

G pools.

• C’s approval: If σ ą 2δ ´ 1 and pφ˚1 , φ˚´1q “ p1, 1q, equation (5) indicates

that C never approves when R´1 “ 0. Therefore, u˚C “ 0. �

Proof of Proposition 2.

• C’s review: If σ ď 2δ ´ 1 and ∆2 “ ´1, inequality (8) indicates that for

pφ˚1 , φ˚´1q “

´

1,`

δ1´δ

˘

´

γλp1`σq´γ

¯¯

, C is indifferent between R´1 “ 1 and

R´1 “ 0. Therefore, ρ˚´1 “κA´1

λpκC´1qis a best response.

• G’s approval: If σ ď 2δ ´ 1 and ∆2 “ ´1, equation (9) indicates EU1G “

´p1 ´ ρλq ´ ρλκC . When C reviews with probability ρ˚´1, EU1G “ EU0

G.

Given this and lemma 1, pφ˚1 , φ˚´1q “

´

1,`

δ1´δ

˘

´

γλp1`σq´γ

¯¯

is sequentially

rational. �

Proof of Proposition 3.

• C’s review: If ∆2 “ 1, inequality (6) indicates that for pφ˚1 , φ˚´1q “

´

1,`

δ1´δ

˘

´

γλ´γ

¯¯

,

C is indifferent between R1 “ 1 and R1 “ 0. Therefore, ρ˚1 “κA´1

λpκC´1qis a

best response.

• G’s approval: If ∆2 “ 1, equation (9) indicates EU1G “ ´p1´ ρλq ´ ρλκC .

When C reviews with probability ρ˚1 , EU1G “ EU0

G. Given this and lemma

1, pφ˚1 , φ˚´1q “

´

1,`

δ1´δ

˘

´

γλ´γ

¯¯

is sequentially rational. �

29

Proof of Proposition 4. Given propositions (1) and (2), and ∆2 “ ´1, the

Crown’s ex ante expected utility is

EU˚K “

$

&

%

δp1´ σKq ` p1´ δqφ˚´1

`

1´ ρ˚´1λ˘

p´1´ σKq , if σ ď 2δ ´ 1

0, if σ ą 2δ ´ 1(10)

For σC ą 2δ ´ 1 (legally fastidious reviewer C), proposition (1) indicates that C

always vetoes, so EU˚K “ 0 in this case. For σC ď 2δ ´ 1 (legally flexible C),

proposition (2) indicates that a good law (∆1 “ 1) is always upheld by C. This

confers utility p1 ´ σKq on the Crown K. A bad law (∆1 “ ´1) is upheld if (i)

upheld by the governor, and (ii) there is not informative Council review. This

confers utility p´1´ σKq on K. Therefore, in this case, EU˚K is the probability-

weighted utility of upholding a good law, plus the probability-weighted utility

of upholding a bad law. Note that K does not directly internalize C’s cost of

review, since this cost is incurred by C and not K.

The only effect of σC on EU˚K is through φ˚´1, the chance that G upholds a

bad law. From proposition (2), this is decreasing in σC provided σC ď 2δ ´ 1.

Since EU˚K is (weakly) decreasing in φ˚´1, it is clear that the Crown’s optimal σC

is never less than 2δ ´ 1.

The Crown prefers σC “ 2δ ´ 1 to σC ą 2δ ´ 1 if and only if the top line of

equation (10) exceeds the bottom:

δp1´ σKq ` p1´ δqφ˚´1

`

1´ ρ˚´1λ˘

p´1´ σKq ě 0. (11)

Inserting φ˚´1 and ρ˚´1 from proposition (2) and taking σC “ 2δ´1, equation (11)

is equivalent to

30

δp1´ σKq ě p1´ δq

ˆ

δ

1´ δ

˙ˆ

γ

2δλ´ γ

˙

`

1´ ρ˚´1λ˘

p1` σKq ô

1´ σK ě

ˆ

γ

2δλ´ γ

˙

`

1´ ρ˚´1λ˘

p1` σKq ô

1´ σK ě

ˆ

γ

2δλ´ γ

˙ˆ

1´

ˆ

κA ´ 1

λpκC ´ 1q

˙

λ

˙

p1` σKq ô

1´ σK ě

ˆ

γ

2δλ´ γ

˙ˆ

κC ´ κAκC ´ 1

˙

p1` σKq .

Defining Q ”´

γ2δλ´γ

¯´

κC´κAκC´1

¯

, it follows that σC “ 2δ ´ 1 is the Crown’s most

preferred value if and only if

σK ď1´Q

1`Q” σK .

To bound Q (and thus σK), note that´

κC´κAκC´1

¯

P p0, 1q by the assumption that

κC ą κA ą 1. Further,´

γ2δλ´γ

¯

ą 0 by the assumption of proposition (2) that

λp1 ` σq ą γ for all σ ď 2δ ´ 1. Thus, Q ą 0, which implies σK ă 1. If, in

addition, δ ąγ´

1`´

κC´κAκC´1

¯¯

2λ, then Q ă 1 and hence σK ą 0. Note in particular

that if σK ą 0, then the Crown’s optimal skepticism for the reviewer is σ˚C ą 0

even if σK “ 0. �

31

N

G

veto

C

C

veto

uphold

don’t

N

C

veto

uphold

inform.

C

veto

uphold

uninf.

review

uphold

p´1,´1q

G

veto

C

C

veto

upholddon’t

N

C

veto

uphold

uninf.

C

veto

uphold

inform.

review

uphold

p1,´1q

G

veto

C

don’t

N

uninf.

C

veto

uphold

inform.

review

uphold

p´1, 1q

G

veto

C

don’t

N

uninf.

C

veto

uphold

inform.

review

uphold

p1, 1q

Figure 1: Extensive form. Nature’s moves from root node denote p∆1,∆2q.

32