PUBLISHED
UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT
No. 20-2253
RLI INSURANCE COMPANY,
Plaintiff - Appellee,
v.
NEXUS SERVICES, INC.; LIBRE BY NEXUS, INC.; HOMES BY NEXUS, INC.,
Defendants - Appellants.
Appeal from the United States District Court for the Western District of Virginia, at Harrisonburg. Michael F. Urbanski, Chief District Judge. (5:18-cv-00066-MFU-JCH)
Argued: October 28, 2021 Decided: January 27, 2022
Before GREGORY, Chief Judge, QUATTLEBAUM, Circuit Judge, and FLOYD, Senior Circuit Judge.
Affirmed by published opinion. Senior Judge Floyd wrote the opinion in which Judge Gregory joined. Judge Quattlebaum wrote an opinion concurring in parts I, II, III, V, and VI and in the judgment.
ARGUED: Carl August Anderson, ROCK SPRING LAW GROUP, PLLC, Washington, D.C., for Appellants. Vivian Katsantonis, WATT, TIEDER, HOFFAR & FITZGERALD, LLP, McLean, Virginia, for Appellee. ON BRIEF: Mario B. Williams, NDH LLC, Atlanta, Georgia; John M. Shoreman, MCFADDEN & SHOREMAN, Washington, D.C., for Appellants. FLOYD, Senior Circuit Judge:
This case concerns relatively straightforward obligations of a bond surety, RLI
Insurance Company (RLI), and its indemnitor, Nexus Services, Inc. (Nexus), under a
standard Commercial Surety General Indemnity Agreement (the Agreement). Because
surety is a zero-loss industry, the Agreement contains several clauses designed to keep RLI
whole. One obligates Nexus to provide collateral sufficient to cover all of RLI’s exposure,
and the parties task us with resolving what kinds of risk “exposure” means to capture. What
makes our task unique is that, unlike the familiar commercial or construction relationships
that typically contemplate only a handful of guarantees, this Agreement involves nearly
2,500 bonds RLI issued to the U.S. government on behalf of individual immigrant
detainees. Nexus insists that we must nonetheless measure RLI’s exposure on each bond
individually and that RLI is not actually “exposed” to any risk—and Nexus
correspondingly does not need to deposit collateral—until the parties have reason to
believe that RLI will have to pay out that particular bond. The first tangible evidence of
that, Nexus continues, comes about when an immigrant fails to appear in court on the
designated date, breaching the bond. In short, Nexus suggests it should deposit collateral
only up to the sum of the already-breached bonds. RLI objects the Agreement is not so
limited. Although we do not know which particular immigrant will breach, we can be
certain some will. It follows that the Agreement must secure against aggregate risk—that
is, the likelihood Nexus will be able to (timely) indemnify RLI for all future breached
bonds. Because Nexus’s financial condition, its willingness to indemnify RLI so far, and
historical rate of bonds breached all bear on that likelihood, they should likewise inform
2 the collateral calculus. The district court sided with RLI, and after reviewing the plain
terms of the Agreement, we agree. We also affirm the district court’s calculation of the
collateral amount as a sound exercise of its discretion to order equitable relief.
I.
An immigration bond, much like a criminal bond, allows the release of a detained
individual from custody based on a surety’s contractual undertaking to the United States to
either deliver the individual as demanded or forfeit the penal sum specified in the bond.
Nexus runs the bonds program: It screens the immigrants likely to keep their promise to
appear in court and maintains contact with them throughout their release. But Nexus lacks
the Department of Treasury’s commercial-surety certification, and so needs another surety
to take on the liability to the government. RLI agreed to perform that function in exchange
for a set fee upfront, and Nexus agreed to indemnify RLI for all losses. Specifically, Nexus
agreed to pay upon demand:
2(a)(i) all losses, costs, damages, attorneys’ fees and expenses of whatever kind or nature which arise by reason of, or in consequence of, the Surety having executed any Bond on behalf of the Principal, or in enforcing this agreement against any of the Indemnitor(s) . . . .
2(a)(ii) an amount sufficient to discharge any claim made against Surety on any Bond. This sum may be used by Surety to pay such claim or be held by Surety as collateral security against loss on any Bond.
J.A. 53. Nexus also agreed that:
3(c) [u]ntil Surety has been furnished with conclusive evidence of its discharge without loss from any Bonds, and until Surety has been otherwise fully indemnified . . . , Surety shall have the right of access to the books, records and accounts of the Indemnitor(s) . . . .
3 3(d) Surety shall have every right, defense, and remedy allowed by law including the rights of exoneration and subrogation. Indemnitor(s) will, upon the request of the Surety, procure the discharge of Surety from any Bond and all liability by reason thereof. If such discharge is unattainable, Indemnitor(s) will, if requested by Surety, either deposit collateral with Surety, acceptable to Surety, sufficient to cover all exposure under such Bonds or Bonds, or make provisions acceptable to Surety for the funding of the bonded obligations[ ].
Id. at 54. Illinois law governs the Agreement. Id.
While the parties have always differed as to what the Agreement requires, they have
never disputed the basic facts of how their relationship progressed. See RLI Ins. Co. v.
Nexus Servs. Inc., 470 F. Supp. 3d 564, 571 (W.D. Va. 2020). Between February 2016 and
February 2017, RLI issued 2,486 immigration bonds totaling $30 million at Nexus’s
request. From the start, RLI insisted on collateral, and Nexus agreed to deposit $500,000.
But it never did. Over the course of the year, Nexus’s performance only continued to
deteriorate. It repeatedly allowed several invoices from the government to become past
due, forcing RLI to pay hundreds of thousands of dollars from its own pocket to avert
referral to the Departments of Treasury and Justice. At one point, the unpaid invoices
totaled $709,789.37. When RLI reached out to Nexus to resolve this crisis, Nexus refused
to answer for weeks at a time, denied access to most of its financial records, misrepresented
when checks were sent to the government, and failed to indemnify RLI until RLI brought
several enforcement actions—which Nexus protracted by “cloak[ing] two of its affiliate
companies” and obstructing discovery. RLI Ins. Co. v. Nexus Servs. Inc., No. 5:18CV66,
2020 WL 6262967, at *12 (W.D. Va. Oct. 23, 2020).
4 Nexus also misrepresented the risk RLI was undertaking. During contract
negotiations, it assured RLI that its proactive screening and tracking techniques would
result in only about 2% of the bonds being breached. By the start of this suit, however,
immigrants had breached about 48% of the discharged bonds. 1 Meanwhile, Nexus has
been investigated by states, the federal government, and insurance companies. And has
had several liens placed on its assets for failing to satisfy creditors. All of this led RLI to
finally invoke ¶ 3(d) of the Agreement and request Nexus to discharge its liability on all
outstanding bonds or deposit $10 million in collateral. Nexus refused to do either, and RLI
turned to the courts. Nexus counterclaimed that RLI requested the $10 million in bad faith.
Before the district court, RLI argued it has absolute discretion to request any amount
of collateral up to its then-current liability, $20 million. Nexus countered that ¶ 3(d)
secures only existing government claims or, at most, the penal value of the already-
breached bonds, even if the government has yet to send a final claim. The district court
rejected all of those readings as inconsistent with the plain text of ¶ 3(d). RLI Ins., 470 F.
Supp. 3d at 583–84. The clause, it reasoned, mentions neither the sole discretion RLI
seeks, nor the specific limitations Nexus proposes. Id. Rather, “exposure” encompasses
all sources of risk. Some of that risk stems from the number of bonds breached, yes; but
that is not the only consideration. Nexus’s poor accounting and questionable financial
1 Discharged bonds refer to finally resolved bonds. Some bonds are resolved when an immigrant appears in court as directed, is granted legal status, or is deported—all of which result in the bond’s cancellation. Alternatively, RLI may resolve a bond by satisfying the government claim on a breached bond. So RLI calculates the 48% rate as 𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏ℎ𝑒𝑒𝑒𝑒 𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 . 𝑎𝑎𝑎𝑎𝑎𝑎 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑ℎ𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏
5 health endanger RLI, too, and the amount of collateral must account for those
shortcomings. The court also rejected Nexus’s bad-faith counterclaim, holding that RLI
merely sought to enforce the Agreement as RLI understood it.
The court then held a separate evidentiary hearing to determine the amount of
collateral that reasonably reflects RLI’s anticipated losses, ultimately directing Nexus to
deposit only $2.4 million. And it awarded RLI $3.4 million for litigation costs as damages
Nexus must indemnify under ¶ 2(a)(i).
Nexus now appeals. It has walked back its most extreme position as to the meaning
of “exposure”—that ¶ 3(d) encompasses only final claims—but maintains that collateral is
appropriate only up to the value of the already-breached bonds. Alternatively, Nexus
contends it did not have to deposit collateral at all because RLI did not fulfill the condition
precedent of requesting it in good faith. Puzzlingly, Nexus does not appeal the district
court’s on-point finding that RLI acted in good faith; it asks instead that we take the district
court’s paring down of the collateral to $2.4 million to imply the court found RLI’s original
demand for $10 million “unreasonable as a matter of law.” Opening Br. 36. Nexus finally
disputes the district court’s method of calculating the collateral and cursorily requests we
reconsider the district court’s costs determination. RLI defends the district court’s
judgment; it does not argue for its original interpretation of “exposure” or seek a higher
deposit. 2 We take these arguments in turn.
2 Apart from requesting collateral, RLI sought several injunctions to compel Nexus to indemnify its payments on past-due bonds, pay for future invoices as they arise, and open its books and records for inspection. Those requests led, in turn, to protracted discovery (Continued) 6 II.
We begin with the district court’s ruling that “exposure” reaches all sources of risk,
not just risk associated with a particular breached bond. Because this decision comes to us
on summary judgment, we review de novo and draw all inferences in in the light most
favorable to Nexus, Denzler v. Questech, Inc., 80 F.3d 97, 101 (4th Cir. 1996).
To affirm, we must agree not only that the Agreement conveys the meaning
espoused by the district court but that the Agreement does so unambiguously, for
“questions of contractual ambiguity” in Illinois must be “given to the trier of fact.” Cont’l
Cas. Co. v. Nw. Nat. Ins. Co., 427 F.3d 1038, 1041 (7th Cir. 2005) (citation omitted). We
agree and affirm.
A.
In Illinois, as elsewhere, indemnity agreements are governed by the usual principles
of contract interpretation. Hanover Ins. Co v. Smith, 538 N.E.2d 710, 796 (Ill. App. Ct.
1989), aff’d, 561 N.E.2d 14 (Ill. 1990). We thus begin by looking “to the language of [the]
contract” because it provides “the best indication of the parties’ intent.” Minn. Life Ins.
Co. v. Kagan, 724 F.3d 843, 849 (7th Cir. 2013) (quotation marks and emphasis omitted).
Recall that ¶ 3(d) obligates Nexus, “upon the request of the Surety, [to] procure the
discharge of Surety from any Bond and all liability by reason thereof.” J.A. 54 ¶ 3(d). And
if “such discharge is unattainable,” Nexus must “either deposit collateral with Surety,
disputes and litigation over what companies should be considered Nexus’s alter egos. The parties do not appeal the district court’s decisions on any of those issues.
7 acceptable to Surety, sufficient to cover all exposure under such Bonds or Bonds or make
provisions acceptable to Surety for the funding of the bonded obligations.” Id. The clause
does not define “exposure,” but Black’s Law Dictionary instructs the term refers to “[t]he
amount of liability or other risk to which a person is subject.” Exposure, Black’s Law
Dictionary (11th ed. 2019). Commonly understood, then, “exposure” broadly comprises
all sources of risk that open RLI up to liability.
The surrounding text of ¶ 3(d) further confirms this common understanding. Again,
¶ 3(d) directs Nexus to secure collateral “sufficient to cover all exposure under such Bonds
or Bonds.” J.A. 54. We take “such Bonds” to refer to the bonds mentioned earlier in the
provision—bonds that RLI specifically requested Nexus to discharge. The second,
unqualified “Bonds” must then reference bonds generally as defined in ¶ 1 of the
Agreement: “Any contractual obligation, and any modification(s) or amendment(s)
thereto, undertaken by Surety for Principal, before or after the date of this Agreement and
any renewal or extension of said obligation.” Id. at 53. Read together, ¶¶ 1 and 3(d) thus
mandate Nexus to provide collateral “sufficient to cover all exposure under” “[a]ny
contractual obligation . . . undertaken by” RLI to the U.S. government. Id. at 53–54
(emphases added). That language is as broad as it can be. It does not limit exposure in
time or scope. It does not condition the obligation on occurrence of any triggering event.
It makes no mention whatever of breached bonds, non-appearance, or non-performance.
Our interpretation does not change if we focus only on “such Bonds,” as Nexus asks
us to do, because RLI indeed requested Nexus to discharge all outstanding bonds. Either
8 way, Nexus must still secure collateral “sufficient to cover all exposure under [all
outstanding] Bonds.” Id. at 54.
Had the parties intended to narrow the scope of exposure solely to breached bonds,
they had many tools in their belt. They could have limited RLI to collateral on a specific
bond and only after an immigrant fails to attend court. They could have tethered the
amount of collateral to the penal sum of currently breached bonds. Or even condition
collateral on receipt of some sort of notice from the Department of Homeland Security.
Compare Am. Motorists Ins. Co. v. United Furnace Co., 876 F.2d 293, 301 (2d Cir. 1989)
(discussing a “collateral security provision,” which “provided that the triggering event
would be . . . the making of a ‘demand’ by the United States against” a surety); Hanover
Ins. Co. v. Clark, No. 05-C-2162, 2006 WL 2375428, at *6 (N.D. Ill. Aug. 15, 2006)
(interpreting an agreement that required payments “as soon as liability exists or is asserted
against the Surety”). The parties did none of that.
The neighboring provisions point in the same direction. As the district court
observed, ¶ 2(a)(ii) already directs Nexus to pay “an amount sufficient to discharge any
claim made against Surety on any Bond.” RLI Ins., 470 F. Supp. 3d at 585. Cabining
¶ 3(d) to breached bonds would “nullify” and “render . . . meaningless” that claim-based
obligation in violation of the basic tenets of contractual interpretation. See Smith v. Burkitt,
795 N.E.2d 385, 389 (Ill. App. Ct. 2003). Nexus maintains that “claim” differs from “an
initial determination that the bond has been breached” because “the bond obligor(s) . . .
have the right to appeal” that determination. Opening Br. 24. But the district court did not
mean to imply that every initial determination will necessarily result in a final claim. It
9 merely reasoned that, if Nexus provided collateral for each and every breached bond in the
amount of the full penal sum upon initial notice, there would be nothing left for Nexus to
pay when the final claim arrives. So ¶ 3(d) must contemplate a different kind of risk.
Paragraph 2(a)(ii) helps elucidate the parties’ intentions in yet another way.
Because it calls for “an amount sufficient to discharge any claim,” J.A. 53, it confirms the
parties understood how to draft a provision that chains payments to a specific bond. Had
they intended to condition ¶ 3(d)’s “exposure” solely on breached bonds, they could have
easily mirrored ¶ 2(a)(ii) to require collateral in the “amount sufficient to discharge any
[breached Bond].” Instead, Nexus agreed to deposit collateral “sufficient to cover all
exposure.” Id. at 54.
That is also why Nexus agreed, in ¶ 3(c), to furnish RLI with books, records, and
accounts “until [RLI] has been otherwise fully indemnified.” Id. RLI would have no need
for this extended financial information had ¶ 3(d) reached only breached bonds. And ¶ 3(e)
further decouples collateral from specific bonds by allowing RLI to use “all collateral
deposited” as “security on any or all Bonds.” Id. In the end, surety bonds are not insurance;
they aim to prevent all loss, not allocate risk. Absent any indication that the parties sought
to narrow collateral obligations, ¶ 3(d) plainly and unambiguously reaches all risk, whether
it stems from the number of bonds currently in breach, Nexus’s historic unwillingness to
pay, or Nexus’s troubling financial records.
And that is exactly what other courts have concluded when interpreting similar
contractual provisions on motions for summary judgment. Of particular interest is Safeco
Ins. Co. of Am. v. M.E.S. Inc., No. 09-CV-3312, 2010 WL 3928606 (E.D.N.Y. Oct. 4,
10 2010), a case both parties urged the district court to consider because it involved a virtually
identical indemnity agreement, obligating the indemnitor, “on request of Surety, [to]
procure discharge of Surety from any Bond” and, “[i]f such discharge is unattainable,” to
“deposit collateral with Surety, acceptable to Surety, sufficient to cover all exposure under
such bond or bonds.” Id. at *2. As here, the Safeco agreement did not define “exposure.”
Id. But the court did not think that lack of definition “render[ed] the term ambiguous”
because there was “no reasonable basis for a difference of opinion”; the agreement plainly
entitled Safeco to “specific performance to enforce a collateral security provision” for all
“losses under a bond that are uncertain but anticipated at some point in the future.” Id. at
*2–3. And when the Safeco court came back to ascertain the amount of collateral
appropriate, it required the indemnitor to deposit $875,000 to cover projected legal and
consulting fees based on the parties’ history of protracted enforcement litigation—risk
arising from the surety arrangement generally, not any bond in particular. Safeco Ins. Co.
of Am. v. M.E.S., Inc., No. 09-CV-3312 ARR ALC, 2010 WL 4828103, at *9 (E.D.N.Y.
Nov. 22, 2010), aff’d sub nom. Safeco Ins. Co. of Am. v. Hirani/MES, JV, 480 F. App’x
606 (2d Cir. 2012).
Reading ¶ 3(d) to condition collateral on an immigrant’s non-appearance would also
create an unwritten condition precedent to Nexus’s obligations. But Illinois law generally
disfavors conditions precedent and requires them to be expressly spelled out in the contract.
See Liu v. T&H Mach., Inc., 191 F.3d 790, 798 (7th Cir. 1999) (“A condition precedent
must be stated expressly and unambiguously.”); Premier Elec. Const. Co. v. Am. Nat. Bank
11 of Chic., 658 N.E.2d 877, 885 (Ill. App. Ct. 1995) (“conditions precedent are not generally
favored”). We decline to imply one here.
At bottom, the text and context of ¶ 3(d) agree that Nexus must deposit sufficient
collateral to secure RLI against all anticipated losses, not just losses on the already-
breached bonds.
B.
Nexus acknowledges that, generally speaking, “exposure” encompasses anticipated
losses, but insists that, in the immigration industry, losses become anticipated only when
an immigrant fails to appear. Nexus accordingly urges us to consider trade-usage evidence
in determining the scope of “exposure” under the Agreement. RLI strenuously objects,
contending Illinois follows the four-corner contract interpretation principle. RLI is correct
in that Illinois will not consider subjective evidence like “the testimony of the parties
themselves as to what they believe the contract means” when a contract is unambiguous on
its face. AM Int’l, Inc. v. Graphic Mgmt. Assoc., Inc., 44 F.3d 572, 574–75 (7th Cir. 1995)
(discussing Illinois law). But as Judge Posner prudently observed, “an ordinary reader of
English would not know about . . . special trade usage, and so [could mistakenly] suppose
the contract unambiguous.” Id. There must therefore exist “a means by which the law
allows these surfaces to be penetrated.” Id. That is why “objective” evidence, like that
“there was more than one ship called Peerless,” “is admissible to demonstrate that
apparently clear contract language means something different from what it seems to mean.”
Id.
12 That Illinois law permits us to look at trade usage, however, does not help Nexus
here. For one, the Agreement was RLI’s first foray into immigration bonds, so we decline
to ascribe to it any immigration-specific knowledge of “exposure.” More fundamentally,
Nexus did not (and still does not) offer any evidence of trade usage. Nexus leans on just
one expert who testified that “[t]here is no risk of financial loss on these bonds absent a
nonappearance notice.” J.A. 2523 (emphasis added). Setting aside any admissibility
issues—for the expert was a lawyer with only one prior surety contracts experience—that
testimony merely reiterates Nexus’s arguments about this Agreement and these bonds. The
expert does not survey, for example, how other industry agreements use the term
“exposure” or what kinds of losses other agreements deem anticipated. He does not offer
any industry dictionaries or templates. He does not cite any court cases. Yet that is what
Illinois law requires: For a term to be “established as part of contract by custom and usage,”
it “must be well-settled and uniformly acted upon, and must be established by several
witnesses”; “it must have existed for a sufficient length of time to become generally
known.” Gord Indus. Plastics, Inc. v. Aubrey Mfg., Inc., 469 N.E.2d 389, 392 (Ill. App.
Ct. 1984) (citation omitted). 3 The expert opinion therefore adds nothing for us to consider
3 The expert opinion is wrong even on its terms. Like Nexus, the expert believes collateral obligations reach only “bonds on which nonappearance notices have been received” because ¶ 3(d) “does not say all bonds, but expressly limits collateral security to exposure under ‘such’ bonds under which discharge is unattainable.” Opening Br. 15 (quoting J.A. 2523). As explained, that reading ignores ¶ 3(d)’s plain text, which requires collateral “sufficient to cover all exposure under such Bonds or Bonds,” meaning ¶ 3(d) encompasses all bonds RLI has issued on Nexus’s behalf. J.A. 54 (emphasis added); see supra pp. 8–9.
13 on top of the arguments Nexus already made about the structure and purpose of the
Agreement.
But even had we recognized some limited value in the expert opinion, Nexus gives
the game away when it insists that “[g]eneral knowledge further confirms this trade
interpretation of exposure.” Opening Br. 16; see id. at 17, 33–34. By definition, trade-
usage evidence can alter the interpretation of an agreement only when it “give[s] particular
meaning to” or “qualif[ies] terms of an agreement.” 810 Ill. Comp. Stat. § 5/1-303(d). But
where the trade meaning does no work, where it merely recites what lay persons can already
learn from the text, we have no reason to examine trade usage at all; we can determine
whether an agreement is ambiguous on its face.
That Nexus does not muster any evidence of a different meaning in the trade is no
surprise. Nexus props its argument—that there can be no risk to RLI until an immigrant
defaults—on the unique structure of immigration bonds. Unlike in other industries, Nexus
reasons, RLI does not deposit any money when it “issues” a bond, it merely promises to
pay if the immigrant fails to appear. But that difference in structure does not translate into
a difference of risk. Even though RLI’s money is not tied up in a government account
today, RLI must eventually pay for all breached bonds just the same.
If anything, the immigration context confirms the parties intended to evaluate risks
more broadly. Unlike in a typical construction or commercial surety agreement, RLI here
issued thousands of bonds. So while RLI cannot predict that a particular immigrant will
breach a particular bond, there is no question some immigrants will default and RLI will
have to pay those claims. Nexus’s collateral must therefore secure against that future
14 aggregate loss. Consider a simple scenario where RLI issues ten bonds. A month in, two
immigrants default and Nexus considers filing for bankruptcy. Fast forward another
month, Nexus files for bankruptcy and four more immigrants default. On Nexus’s view,
RLI’s only risk at the end of the first month flows from the two defaulted bonds. RLI can
consequently request collateral only on those two bonds and must wait to request further
collateral until the four other immigrants default. But by the time they do, RLI may receive
nothing at all, or at best may need to wait months for bankruptcy proceedings to conclude.
That defeats the very point of collateral: to secure RLI’s obligations before it incurs them.
Nexus itself appears to realize that exposure “is a function of total coverage.” Opening Br.
17. Yet it fails to carry that principle to its logical conclusion: financial resources and
ability to pay bear directly on RLI’s exposure.
That Nexus must secure RLI against the loss on all outstanding bonds of course does
not mean Nexus must deposit collateral totaling the penal sum on all those bonds. In that
sense, Nexus is correct that we must measure collateral against some anticipated loss, and
the district court appropriately rejected RLI’s initial position that it could demand collateral
up to $20 million. But we see no principled reason to hold that losses do not become
anticipated until an immigrant fails to appear. We hold instead that “exposure” comprises
all risks, such as Nexus’s poor financial records and its historical failure to timely
indemnify RLI.
15 III.
Apart from challenging the district court’s interpretation of “exposure,” Nexus also
argues the district court erred in ordering Nexus to deposit collateral when RLI had not
properly requested it. See J.A. 54 (directing Nexus, in ¶ 3(d), to deposit collateral “if
requested by Surety”). To be sure, Nexus does not contest that RLI asked it to pay—RLI’s
$10 million demand is the reason the parties have come before this Court. Rather, Nexus
maintains that RLI requested $10 million in bad faith and so did not fulfill a condition
precedent to Nexus’s collateral obligations. As proof, Nexus observes the district court
ruled the amount of collateral must be reasonable and then disallowed RLI’s demand for
$10 million as unreasonable, directing Nexus to deposit $2.4 million instead.
Unreasonableness is not “the test of good faith” in Illinois, Original Great Am.
Chocolate Chip Cookie Co. v. River Valley Cookies, Ltd., 970 F.2d 273, 280 (7th Cir.
1992), and this case helps understand why. After several rounds of briefing, and with the
help of multiple experts, the district court concluded, in hindsight, that $10 million was
unreasonably high. But RLI did not have the benefit of that guidance when it made its
initial request, for Nexus repeatedly refused to disclose its books, carry on a meaningful
dialogue, or propose a lower amount. In those circumstances, there was nothing sinister
about requesting $10 million to secure a $20 million liability. Indeed, while the district
court disapproved of the precise figure RLI requested, it rejected Nexus’s bad faith
counterclaim because Nexus failed to show that RLI “both intended to and did actually
attempt to contravene the purpose of the agreement”—especially “because any collateral
requested merely represents funds held in trust for Nexus, and not for RLI’s profit.” RLI
16 Ins., 470 F. Supp. 3d at 592–93. RLI, the district court concluded, merely sought to enforce
the Agreement as it understood it. Id. And Nexus’s failure to appeal the district court’s
good-faith ruling may alone foreclose this argument.
But Nexus also misapprehends the nature of this action. This is not a breach-of-
contract claim, where RLI seeks to collect damages for Nexus’s failure to provide collateral
and Nexus defends by demonstrating RLI did not fulfill the condition precedent. This is a
request for an equitable remedy. The only consequence of RLI’s unreasonable request is
that the district court must adjust the collateral amount consistent with equitable principles,
as the court did here.
IV.
That brings us to Nexus’s objections over the process the district court used to arrive
at $2.4 million. Nexus contends the district court performed “an incredibly specific
calculation” inappropriate for summary judgment and that “any determination of
reasonableness” should instead “have been reserved for a trier of fact.” Opening Br. 37–
38. Nexus also resists the court’s authority to impose any reasonable collateral when the
Agreement called for an “objective” measure by tying collateral to RLI’s exposure. Id. at
31–32.
Here, too, Nexus confuses actions for breach of contract with requests for specific
performance. Because Nexus failed to provide any collateral, it breached the Agreement.
17 But RLI did not sue for damages flowing from that breach, such as compensation for some
“certain” losses it suffered in paying past-due bonds out of its own pocket. Safeco, 2010
WL 3928606, at *3. Instead, RLI requested a specific-performance decree obligating
Nexus to provide collateral going forward. In making that request, RLI had “no legal
remedy,” only “an equitable” one. Am. Motorists Ins. Co. v. United Furnace Co., 876 F.2d
293, 300 (2d Cir. 1989), cited with approval in, Hanover, 2006 WL 2375428, at *5; accord
Eakin v. Cont’l Ill. Nat. Bank & Tr. Co. of Chi., 121 F.R.D. 363, 366 (N.D. Ill. 1988)
(explaining that collateral decrees comprise equitable remedies because “to date, there
simply are no damages to award”), aff’d, 875 F.2d 114 (7th Cir. 1989); cf. Com. Ins. Co.
of Newark v. Pac.-Peru Constr. Corp., 558 F.2d 948, 954 (9th Cir. 1977) (holding specific
performance “not available” when a plaintiff has “an adequate remedy at law”).
Like any grant of specific performance, then, the district court’s order to deposit
collateral constitutes “a matter of sound judicial discretion controlled by established
principles of equity and exercised upon a consideration of all the facts and circumstances
of a particular case.” Schwinder v. Austin Bank of Chi., 809 N.E.2d 180, 196 (Ill. App. Ct.
2004) (citation omitted). In that capacity, Illinois courts “balance the equities between the
parties” and “may refuse to grant specific performance where the remedy would cause a
peculiar hardship or inequitable result.” Id. (citation omitted). Outside Illinois, courts, too,
uniformly understand their “function” in such actions as “determin[ing], in light of the
circumstances of the case and the supporting documentation submitted by the parties,
whether [a surety]’s demand is a reasonable estimate of its anticipated losses.” Safeco,
2010 WL 4828103, at *4; accord Am. Motorists, 876 F.2d at 303 (“leav[ing] to the district
18 court the question of how much collateral security [the indemnitor] is obligated to
provide”); Travelers Cas. & Sur. Co. of Am. v. Sw. Contracting, Inc., No. 4:05CV99-DJS,
2006 WL 276942, at *3 (E.D. Mo. Feb. 2, 2006) (granting specific performance where
collateral sought was “reasonable and not excessive”); U.S. Fid. & Guar. Co. v. J. United
Elec. Contracting Corp., 62 F. Supp. 2d 915, 922 (E.D.N.Y. 1999) (“a collateral security
clause is enforceable as long as the amount demanded by a surety is reasonable”); U.S. Fid.
& Guar. Co. v. Feibus, 15 F. Supp. 2d 579, 588 (M.D. Pa. 1998), aff’d, 185 F.3d 864 (3d
Cir. 1999) (same, because “there is no windfall for the surety”). What courts do not do is
assess “demand[s] for collateral security” “under the summary judgment standard.”
Safeco, 2010 WL 4828103, at *4.
Disputed facts and conflicting assessments of risk therefore do not defeat the district
court’s authority to order specific performance. After all, “[i]f it were a prerequisite for
the surety and indemnitors to agree upon all the material facts pertaining to future
payments,” “no surety would ever succeed in enforcing its interim right to collateral.” Id.
(internal quotation marks and citation omitted). And, if taken seriously, the logic of
Nexus’s argument would require us to send any factual disputes inherent in requests for
injunctions or restraining orders to the jury. We decline to so undermine the very
foundation of the law-equity divide and instead hold that the district court correctly
approached the task before it by scrutinizing the reasonableness of the collateral amount.
19 B.
Because the district court applied the principles of equity, we review its
determination of the collateral amount for abuse of discretion. Klein v. PepsiCo, Inc., 845
F.2d 76, 78 (4th Cir. 1988). Finding no abuse, we affirm.
After ruling on the contractual-interpretation question above, the district court
dedicated a separate round of briefing to ascertaining the appropriate amount of collateral.
It then held a two-day hearing, where both parties had an opportunity to present fact and
expert witnesses. RLI called an actuary with experience in the surety industry, a forensic
accountant, and a more general expert on surety and insurance contracts. The actuary
identified four characteristics predictive of bond breach: bond amount, date of the
immigrant’s last contact with Nexus, the immigrant’s last payment to Nexus, and the
immigrant’s country of origin. RLI Ins., 2020 WL 6262967, at *3 & n.5. Based on those
characteristics, she calculated a future breach risk to RLI on the outstanding bonds to
constitute approximately $10 million—the amount RLI first requested as collateral. Id.
The forensic accountant identified several areas of Nexus’s financial records which caused
him concern, such as multiple six- and seven-figure debts, hundreds of thousands of dollars
in bounced checks, and repeated failures to conform financial records to standard industry
practice—most troublingly, underreporting liabilities. Id. at *4–5. Building upon that
testimony, RLI’s industry expert explained that often times, when “the books are a mess
and they are inaccurate,” “that is a sure sign that . . . down the road, . . . the indemnitors
lose their ability to pay.” Id. at *6. That means even a zero-loss situation can quickly turn
20 into losses on “some 80 percent of the claims.” Id. So the appropriate amount of collateral
must account for “what the losses could be in the future.” Id. at *7 (emphasis added).
RLI also offered a factual witness, its Assistant Vice President of Claims. He began
by outlining the basic facts about the bonds. RLI, he explained, has issued a total of 2,486
bonds on Nexus’s behalf. Id. at *3. By the date of the hearing, 393 of those bonds had
been breached, causing RLI to pay $4,460,000 to the U.S. government. 4 Id. Comparing
that figure to the 424 bonds that have been cancelled—bonds for immigrants who appeared
in court or have been deported, relieving RLI of liability—yields a historic breach rate of
about 48%. Id. And 48% of the current outstanding bond amount again yields about $10
million. Id.
The same witness also reviewed Nexus’s financial and operating documents and
identified eight risk factors flowing from Nexus’s financial decisions that affect its ability
to pay, including Nexus’s historic failure to timely indemnify RLI for breached bonds, its
unreliable and inconsistent financial records, a recent—and unrecorded—real estate sell-
off, and ongoing investigations and enforcement actions. Id.
Nexus called no expert witnesses during this evidentiary hearing. Its Vice President
of Operations conceded that “its books and records were a mess” but proffered that Nexus
“was taking affirmative steps to bringing them in order.” Id. at *6. He also admitted to
various investigations and lawsuits, offering only that there was no “money Nexus owed
4 As discussed, the parties do not contest that Nexus eventually reimbursed RLI the full sum of the breached bonds.
21 on other ongoing investigations.” Id. at *7. Nexus’s Vice President of Risk Management
testified that Nexus “made efforts to contact its RLI-bonded program participants during
the past few weeks and obtained declarations from 219 participants stating that they fully
intended to appear.” Id. But Nexus’s CEO conceded that Nexus no longer uses GPS
monitoring because the GPS company terminated its services in light of Nexus’s $7 million
debt. Id. at *5. Nexus’s only quarrel with RLI’s testimony was the formula used to
calculate the historic breach rate—and the end result of 48%. Rather than compare
breached bonds to bonds cancelled so far, Nexus insisted that the court should compare
breached bonds to all pending bonds because “RLI faces no risk” on “bonds for which
principals have demonstrated a history of compliance.” Id. at *3. That analysis would
yield a breach rate of roughly 5%. Id.
The district court carefully considered this testimony in a 13-page opinion,
ultimately ordering Nexus to deposit just $2.4 million, a quarter of RLI’s originally-
requested $10 million. Id. at *9. The court explained that it based that sum on the amount
of outstanding bonds, Nexus’s “historic failure to timely pay breached bonds,” Nexus’s
continued refusal “[t]o this day . . . to make available accurate financial records,” and
Nexus’s deteriorating financial condition, evidenced by “the myriad investigations into
[its] business practices by various state attorneys general and bureaus of insurance, claims
made by large creditors, its recent sell-off of real estate[,] and diminished cash flow.” Id.
We find no abuse in that detailed, well-reasoned order.
22 C.
Nexus tries one last argument. It observes ¶ 3(d) does not leave collateral amount
to RLI’s sole discretion but ties collateral to RLI’s “exposure.” That language, Nexus
insists, invites “objective” assessments of collateral due, precluding the kind of open-ended
reasonableness evaluations the court conducted below. Opening Br. 32. And because the
only “objective” measure happens to be the penal sums on the bonds breached, the district
court should have limited collateral to that amount. As discussed, that is simply not how
courts approach requests for collateral—courts uniformly apply a reasonableness standard,
including in contracts identical to the one here that tie collateral to a surety’s “exposure.”
See Safeco, 2010 WL 4828103, at *4; see supra pp. 17–19. And with good reason: RLI
will not retain the $2.4 million “as its property” but will hold it “in trust” for Nexus. Am.
Motorists, 876 F.2d at 300. If Nexus’s “misgivings regarding the amount of money held
in collateral prove to be correct, [it] will be entitled to have [its] money returned.”
Hanover, 2006 WL 2375428, at *6 (internal quotation marks omitted).
Cases that calculate damages for breach of contract are thus beside the point. See
Opening Br. 31 (citing Boyd v. Tornier, Inc., No. 07-cv-0751-MJR, 2009 WL 1657900, at
*7 (S.D. Ill. June 12, 2009), which refused to calculate quotas under a “reasonability”
principle because the contract provided specific guidelines for quota calculations). Same
with cases that decline to imply a duty of good faith and fair dealing to override express
contractual language. See id. at 31–32 (citing Fields v. Thompson Printing Co., 363 F.3d
259, 271 (3d Cir. 2004); Paulus Sokolowski & Sartor, LLC v. Cont’l Cas. Co., No. Civ.A.
12-7172 MASTJ, 2013 WL 11084770, at *7 (D.N.J. Aug. 30, 2013)). Those cases apply
23 to courts sitting in law, deciding damages for breach of contract on summary judgment.
Sitting in equity, the court below was not only permitted but obligated to conduct a
reasonableness inquiry, so as to ensure that its grant of specific performance would not
“cause a peculiar hardship or inequitable result.” Schwinder, 809 N.E.2d at 196 (citation
omitted).
That is presumably why RLI does not appeal the district court’s winnowing of the
collateral to $2.4 million, even though ¶ 3(d) directs Nexus to “deposit collateral with
Surety, acceptable to Surety, sufficient to cover all exposure.” J.A. 54 (emphasis added).
Even sureties that contractually have “sole discretion” over the amount of the collateral
must, at day’s end, satisfy the courts “the sum demanded is reasonable.” BIB Constr. Co.,
Inc., v. Fireman’s Ins. Co. of Newark, 214 A.D.2d 521, 523 (N.Y. App. Div. 1995). That
is, a surety can sue for damages when an indemnitor breaches its contractual obligation to
deposit the amount the parties have agreed to—here, an amount “acceptable to” RLI. J.A.
54. But, when it comes to specific performance, courts can no more order an indemnitor
to deposit an unreasonable amount of collateral than order an employee to stay in a job she
wishes to leave.
Even setting those foundational principles aside, Nexus’s argument proves too
much. That the Agreement anchors the collateral amount in RLI’s “exposure” rather than
allow RLI to singlehandedly name the price does suggest some objective evaluation of how
much risk RLI faces. But it does not follow that the only objective measure of that risk is
the value of breached bonds. Outside experts’ evaluations, financial documents, and
historic rate of default offer objective gauges, as well. Nothing in such a reasonableness
24 analysis requires—or allows—the court to take RLI at its word. We accordingly decline
Nexus’s invitation to circumscribe the court’s traditional equity function, affirm the district
court’s interpretation of the contract as a matter of law, and affirm its collateral decree as
a sound exercise of discretion.
Nexus finally asks us to reconsider the district court’s award of litigation costs
“assuming that Nexus’s interpretation of its obligations under the Indemnity Agreement
prevails before this court, or if the District Court’s summary judgment rulings are
remanded for further proceedings.” Opening Br. 39. Because we affirm, we have no need
to reach this question. 5
VI.
For the foregoing reasons, the district court’s judgment is
AFFIRMED.
5 We clarify, however, that where an indemnitor expressly agrees to compensate for “all losses, costs, damages, attorneys’ fees and expenses of whatever kind or nature which arise . . . in enforcing this agreement,” J.A. 53, the award of costs and fees constitutes direct damages under the agreement, not fee shifting. E.g., Lamp, Inc. v. Int’l Fid. Ins. Co., 493 N.E.2d 146,149 (Ill. App. Ct. 1986). So we would ask whether RLI incurred its costs and fees “in enforcing this agreement,” J.A. 53—not whether RLI is entitled to them as a prevailing party. See Hanover Ins. Co. v. Smith, 561 N.E.2d 14, 18 (Ill. 1990) (assessing whether a surety sustained attorney’s fees “in consequence” of the execution of a bond agreement); Fid. & Deposit Co. of Md. v. Rosenmutter, 614 F. Supp. 348, 352 (N.D. Ill. 1985) (declining to indemnify litigation expenses where an action “was unnecessary to enforce the contract”).
25 QUATTLEBAUM, Circuit Judge concurring:
I am pleased to join in sections I, II, III, V and VI of Judge Floyd’s excellent opinion.
I write separately only concerning section IV.
There, the majority concludes the district court did not err in examining the amount
of collateral RLI requests from Nexus for “reasonableness.” According to the majority,
RLI sought equitable relief which the district court is afforded discretion to provide. I agree
that, in equity, a court has discretion in fashioning an appropriate remedy. But that
discretion does not include remedies that are inconsistent with the express terms of the
contract. “Courts have traditionally analyzed the plain language of the indemnity
agreement to determine a surety’s right to obtain specific performance of collateral
security.” Hanover Ins. Co. v. Clark, No. 05 C 2162, 2006 WL 2375428, at *5 (N.D. Ill.
Aug. 15, 2006).
Here, the express terms allowed RLI to seek collateral “acceptable to [RLI].” Given
that language, I would not have imposed an objective standard of reasonableness. To me,
that judicially modifies a term on which the parties agreed and reduced to writing. Words
have meaning and, as I understand Illinois law, that meaning does not evaporate just
because a party brings a claim in equity. Butler v. Kent, 655 N.E.2d 1120, 1127 (Ill. App.
Ct. 1995) (“The province of a court in a specific performance action is to enforce the
contract which the parties have made.”); Snyder v. Spaulding, 57 Ill. 480, 484 (1870)
(“Equity does not propose to relieve against the express contract of parties.”).
For those reasons, I agree with Nexus that the district court erred in imposing a
reasonableness standard to the amount of collateral RLI could request when the contract
26 permitted collateral “acceptable to [RLI].” However, despite my agreement with Nexus on
this point, the relief Nexus seeks for this error—as ably explained by the majority in section
II of its opinion—contravenes the plain language of the contract. Ironically, the party
prejudiced by the district court’s imposition of a non-contractual objective standard of
reasonableness was RLI, not Nexus. But RLI did not appeal the district court’s
reasonableness inquiry and finding. Therefore, I concur in the majority’s ultimate
conclusion that Nexus’ arguments on appeal must be rejected.