CLARK, Circuit Judge:

First National Life Insurance Company (FNL) sued on an employee fidelity bond issued by Fidelity Deposit Company of Maryland (F D) for losses incurred when purchasers of controlling stock in FNL's parent company raided the assets of FNL. The case was tried to a jury, which answered special interrogatories in a manner generally favorable to FNL. The district court entered a judgment for F D notwithstanding the verdict, based upon an opinion which alternatively upheld three policy defenses: (1) the bond's definition of "employee" excluded the defrauders who had named themselves officers and directors; (2) the capture amounted to a proscribed "takeover" of FNL; and (3) litigation settlements amounted to a release of "principals." The perpetrators of this fraud were not covered by the bond's undertaking to insure against the fraudulent or dishonest acts of employees. We therefore affirm the judgment of the district court without reaching the merits of its alternative supports.

The loss for which coverage was claimed arose from a 1972 stock and asset manipulation. FNL, an Alabama insurance company headquartered in Texas, is a wholly-owned subsidiary of First National Corporation, which had its principal place of business in Houston, Texas. First National Corporation was capitalized with two classes of stock, A B, identical in all respects except that the holders of Class B stock had the right to elect a majority of the Board of Directors of First National Corporation, which in turn controlled FNL. All of the B stock and 15% of the A stock was held by the Three R Trust (3R).

The undisputed proof showed that in early 1972 the trustees of 3R decided that it should divest itself of all its holdings in First National Corporation and set a tentative price of 2 million dollars, with a minimum down payment of one-fourth. A group of promoters, using an almost simultaneous series of shell-game transactions, borrowed funds for the down payment, obtained control of FNC, elected themselves officers and directors of both the holding company and FNL, sold 1 million dollars of corporate bonds owned by FNL, and used half of the proceeds to pay the obligation they had incurred in "raising" the down payment and split the rest. In effect, the purchasing group both looted FNL to buy it and vice versa.

The choice of law in a diversity case is determined by application of the conflicts rules of the forum. Klaxon Company v. Stentor Electric Manufacturing Company, 313 U.S. 487, 61 S.Ct. 1020, 85 L.Ed. 1477 (1941). Alabama, the forum state, follows a place-of-performance rule where a contract is to be performed in a place other than where it is made. See J. R. Watkins Company v. Hill, 214 Ala. 507, 108 So. 244 (1926). See also Ideal Structures Corporation v. Levine Huntsville Development Corporation, 396 F.2d 917, 921-25 (5th Cir. 1968). Because performance on the bond would be due to FNL in Texas, the district court correctly ruled that Texas law controls the interpretation of the bond's provisions.

However, contrary to most conflicts questions, ascertainment of the Texas rule is a more difficult matter. We do not confront the ideal situation where state law has been "declared by the legislature or by decision of the highest court of the state." Erie Co. v. Tompkins, 304 U.S. 64, 78, 58 S.Ct. 817, 822, 82 L.Ed. 1188, 1194 (1938). Where no such declaration exists, a federal court must utilize the same methodology that would be employed by the state courts themselves when they are confronted with a question for which no authoritative answer can be found. Our function in diversity actions is "`in effect, [as] only another court of the State.'" Angel v. Bullington, 330 U.S. 183, 187, 67 S.Ct. 657, 659, 91 L.Ed. 832, 835 (1947); Mississippi Power Co. v. Roubicek, 462 F.2d 412 (5th Cir. 1972).

The district court sustained F D's first contention that these corporate raiders had not become covered "employees" of FNL as defined in the Bond. This crucial definition reads:

Wherever used in this bond, Employee and Employees shall be deemed to mean, respectively, one or more of the officers, clerks and other natural persons in the service of the Insured while employed in, at or by any of the Insured's offices while covered under this bond during the currency of this bond and who are compensated by salary, wages or commissions, and whom the Insured has the right to govern and direct at all times in the performance of such service but, not to mean brokers, general agents, sub-agents, loan agents, fiscal agents, property management agents, real estate agents, or other representatives of the same general character.

Apparently, only one Texas case has considered the scope of the term "employee" in the policy language of a fidelity bond. In First State Bank of Temple v. Metropolitan Casualty Insurance Co., 125 Tex. 113, 79 S.W.2d 835 (1935), the officers and directors of a recently reorganized bank voted themselves a "dividend" to defray their personal obligations. Payments went only to those who participated in declaring it. Other directors and shareholders testified at trial that they had no knowledge of the dividend for months after it was declared. 125 Tex. at 117, 79 S.W.2d at 837. The officers and directors were covered by a fidelity bond that defined employees as embracing "all officers and employees, except inactive vice-presidents and directors other than officers." Id. at 118, 79 S.W.2d at 838. Without articulating reasoning pertinent here, the Supreme Court of Texas held that the self-serving corrupt acts of the directors "did not become the acts of the bank itself, and the insurance company was liable for their misdeeds under the terms of the bond." Id. at 124, 79 S.W.2d at 841.

We are convinced, however, that a Texas court would not apply First State Bank as controlling the interpretation of the employee clause in the case at bar. We base this conclusion on three considerations.

First, the policy language construed in First State Bank was far less precise than the language of the bond in issue here. In effect, the bond in First State Bank covered everyone who worked for the bank except inactive vice-presidents and outside directors. It did not specify that the covered employee be "in the service of the Insured while employed in, at or by any of the Insured's offices," or that they be "compensated by salary, wages or commissions," or that the bank have "the right to govern and direct [the covered employees] at all times in the performance of [their] service . . ."

These narrowing provisions make this bond markedly different from that construed in the First State Bank case. On this ground alone, there is no doubt that a Texas court would not consider the holding in First State Bank as controlling in today's case. See Employers Casualty Co. v. Rockwall County, 300 S.W. 148 (Tex.Civ.App. 1928), modified on other grounds, 120 Tex. 441, 35 S.W.2d 690 (1931); 13 J. Appleman Insurance Law and Practice § 7502 (1943).

Second, the officers and directors who voted themselves the improper dividend in First State Bank were, for the most part, the same individuals who controlled the bank before its reorganization. Unlike the defrauders in today's case who captured control and instantaneously pirated the company's portfolio, the First State Bank culprits had been entrusted by its stockholders to perform everyday business functions for the bank before and after the nominal change in corporate form. The Supreme Court of Texas did not intimate that the officers and directors acquired control of First State Bank solely to profit themselves. Instead, the situation reveals itself to be one in which those who already had occupied positions of trust in the corporation misused them for their benefit — a situation that more nearly corresponds to the customary purposes of fidelity bonding. See, generally 9 J. Appleman, supra, §§ 5661 et seq.

Third, even though we do not find it necessary to reach the merits of the district court's alternative ruling based upon the "takeover" provision of this bond, it is appropriate to look to this and another term in ascertaining the intent of the issuing company and the insured. The function of a contract of insurance is to specify the risks assumed and to define the extent of coverage provided for them. The indemnity contract in this case is typically one of adhesion, which requires that any doubt or ambiguity be resolved in favor of coverage. Nevertheless, the intent of the parties remains the lodestar.

To determine the extent of employee action intended to be covered, we should look primarily to the definition clause but not to the exclusion of related integral parts. Here it is pertinent to note both the clause that denied coverage for acts of directors who were not "employees" and the clause that terminated all coverage in the event ownership of FNL was "taken over" by another concern.

The director limitation confirms that the meaning of "employee" adopted by the district court was proper. By limiting coverage to unfaithful activity that breaches the confidence reposed in regular officials of FNL who also serve as directors, while excluding the acts of outside directors who might have mixed loyalties or primary allegiances elsewhere, the clause complements the limits placed on coverage of officers, clerks and other persons "in the service of the Insured." In a parallel vein, the clause which terminates all coverage in the event of a "taking over . . . by another concern," makes liability for the risks assumed by the bond depend upon a continuity of the same general management of FNL that controlled the company when the underwriting occurred.

Those who conceived and executed this plan to gain personal enrichment at FNL's expense elected themselves as officers and directors. The moment their official status gave them access to FNL's assets, they appropriated them as booty for the capture. Their momentary status as officials never placed them "in the service" of FNL, for they served only themselves. Their ill-gotten gains did not "compensate" them for they had done nothing compensable. The funds they took under the guise of fees were neither "salary, wages or commissions." They were always and in all ways usurpers; not even the victim's loss can make their false status real.

Even giving the bond's language the construction most favorable to coverage can raise no doubt that F D did not contract to bear fraud losses caused by persons who were never regularly employed by FNL. The fact that the bond did not expressly exclude coverage of a scheme as complex and as devious as the one carried out by this capture of the company's parent, and thus its management, by strangers who placed themselves in the positions of power they used to gain access to its assets, cannot afford the basis for a claim. It is sufficient support for F D's defense that the policy made clear that such self-dealing persons were not covered employees. The district court correctly entered judgment notwithstanding the verdict.

Affirmed.

. . . . .

"(d) Any loss resulting from any act or acts of any director or trustee of the Insured other than an Employee of the Insured, except when performing acts coming within the scope of the usual duties of an Employee, or while acting as a member of any committee duly elected or appointed by resolution of the board of directors or trustees of the Insured to perform specific, as distinguished from general, directorial acts on behalf of the Insured."