Safeco Insurance Company of America (Safeco) petitioned this court for review of a decision of the Court of Appeals, Division One. Safeco brought an action for declaratory judgment and summary judgment against Jan V. Hirschmann and Evelyn F. McElmeel, its insureds, in the King County Superior Court based on a clause in its all-risk homeowners policy purporting to exclude damage resulting from landslides. The trial court granted judgment *623for Safeco. The Court of Appeals reversed and entered partial summary judgment for the insureds, finding the exclusionary language in Safeco's contract to be the functional equivalent of a clause in a policy of Public Employees Mutual Insurance Company (Perneo) at issue in Villella v. Public Employees Mut. Ins. Co., 106 Wn.2d 806, 725 P.2d 957 (1986). We agree and affirm the decision of the Court of Appeals.
The sole issue in this case is whether the language of Safeco's exclusionary clause circumvents the "efficient proximate cause" rule of all-risk homeowners insurance coverage.
The facts in this matter are not in dispute. On January 16, 1986, severe winds of 50 miles per hour struck the Seattle area. This was followed the next day by heavy rains totaling 4.2 inches for the 24-hour period of January 18. The rainfall exceeded that of a "100-year" storm and caused a number of landslides in the greater Seattle area.
Late in the afternoon of January 18, 1986, Dr. Jan V. Hirschmann had to vacate his family home because the embankment behind it had caused the kitchen walls and windows to give way. The weight of the hillside pushed the house from its foundation. By 9 p.m. the second story of the house had broken down into the first floor. The house was destroyed, for a total loss estimated at $250,000.
Three adjacent residences were affected by the same landslide. The insureds' residence at 635 38th Avenue, Seattle, was built in 1920. It had stood in that location for over 65 years without significant movement of the house or its foundation. Safeco insured the house under an "all-risk" policy.
William L. Shannon, an engineer, investigated the cause of the destruction. He determined that the type of soil in the hillside behind the Hirschmann residence is very slowly permeable to water and loses strength when saturated. He stated that mechanical action of the strong winds on trees and bushes loosened the surface soil and increased infiltration of rainwater into the soil. In his opinion, the "primary *624cause" of the collapse of the hillside was the heavy and prolonged rainfall which saturated the ground.
In bringing this action for declaratory judgment, Safeco claimed that its "all-risk" policy did not cover the loss. The trial court granted Safeco's motion for summary judgment. The Court of Appeals, Division One, reversed, and directed the trial court to enter an order granting partial summary judgment to Hirschmann on the issue of coverage. We granted review on January 10, 1989.
This court first announced the "efficient proximate cause" rule in Graham v. Public Employees Mut. Ins. Co., 98 Wn.2d 533, 656 P.2d 1077 (1983). Subsequently, we were asked to consider whether a difference in exclusionary clause language could circumvent the rule. In Villella v. Public Employees Mut. Ins. Co., 106 Wn.2d 806, 725 P.2d 957 (1986) we answered in the negative. We are once again asked to consider whether another variation in exclusionary clause language can circumvent the rule. We again answer in the negative.
At issue is the following language in Safeco's all-risk policy, crafted by Safeco in an effort to circumvent the "efficient proximate cause" rule:
We do not cover loss caused by any of the following excluded perils, whether occurring alone or in any sequence with a covered peril:
2. Earth Movement, meaning: a. earthquake; landslide; mudflow; earth sinking, rising or shifting; . . .
(Italics ours.)
The clause, "We do hot cover loss caused by . . . excluded perils whether occurring alone or in any sequence with a covered peril", replaced the previous clause, "We do not cover loss resulting directly or indirectly from [listed excluded perils].'1
Safeco implicitly concedes that the earlier language would have accorded coverage to Hirschmann by arguing that its newly crafted language commands a different result *625than that in Villella. The Court of Appeals held Safeco's clause and the clause in Villella to be "functionally similar." Safeco Ins. Co. of Am. v. Hirschmann, 52 Wn. App. 469, 475, 760 P.2d 969 (1988). Thus, the question now before this court is whether the difference in contractual language effects a different legal result. We hold that it does not.
The language of the earlier Safeco policy ("directly or indirectly") is identical to that of the Perneo policy reviewed in Villella:
We do not cover loss resulting directly or indirectly from:
2. Earth Movement. Meaning any loss caused by, resulting from, contributed to or aggravated by:
a. earthquake, landslide, mudflow, earth sinking, rising or shifting . . .
Villella, at 809.
In Villella the insured claimed that a negligent failure to install a proper drainage system set in motion a continuous process of soil destabilization that caused one side of the foundation of the house to drop 8 inches. Perneo countered that the house was damaged by "earth movement." We reversed and remanded the case, reciting the "efficient proximate cause" rule adopted in Graham v. Public Employees Mut. Ins. Co., 98 Wn.2d 533, 656 P.2d 1077 (1983):
Where a peril specifically insured against sets other causes in motion which, in an unbroken sequence and connection between the act and final loss, produce the result for which recovery is sought, the insured peril is regarded as the "proximate cause" of the entire loss.
It is the efficient or predominant cause which sets into motion the chain of events producing the loss which is regarded as the proximate cause, not necessarily the last act in a chain of events.
(Citations omitted.) Graham, at 538, cited in Villella, at 815.
*626Further, in Villella, Justice Dore, writing for the court elaborated that, "Stated in another fashion, where an insured risk itself sets into operation a chain of causation in which the last step may have been an excepted risk, the excepted risk will not defeat recovery." (Italics ours.) Villella, at 815 (citing 5 J. Appleman, Insurance § 3083, at 311 (1970); 18 R. Anderson, Couch on Insurance § 74:711, at 1020-22 (2d ed. 1983)). This is so even though there might be an excluded risk which also contributed to the loss or damage. Villella, at 816.
The Villella decision is dispositive of this case. Villella is not only factually parallel, but also procedurally and conceptually identical, to it.
In Villella, as in this case, the trial court granted summary judgment in favor of the insurer and denied coverage under the terms of the policy. The summary judgment was reversed by this court on appeal. In this case, Safeco obtained an order granting summary judgment in the trial court. The Court of Appeals reversed.
In Villella, as in this case, at issue was a change in language that sought to circumvent the "efficient proximate cause" rule announced in Graham. In Villella, as in this case, the insurer was attempting to avoid liability for a loss resulting from a causal chain in which the final "cause" was an excluded peril.
Further, in Villella, as in the present case, the insurer contended that the exclusionary language it crafted circumvented coverage for any loss resulting from a causal chain including any excluded peril regardless how insignificant the contribution from the excluded peril may have been. In Villella, based upon its post-Graham clause, "Perneo argue[d] that if earth movement contributed to the loss, regardless how slight in degree, coverage is precluded." (Italics ours.) Villella, at 817. We rejected that argument. Presented with a record containing testimony from a Safeco vice-president that in his mind the Safeco and Perneo clauses are "functionally the same," the Court *627of Appeals rejected the identical contention in the present case, noting that:
Safeco's purpose in modifying the exclusion is clear; the language prefacing the exclusions is an attempt to exclude from coverage losses connected with certain perils no matter how insignificant those perils may have been to the loss.
(Italics ours.) Safeco Ins. Co. of Am. v. Hirschmahn, 52 Wn. App. 469, 474, 760 P.2d 969 (1988).1
In Villella, as in this case, the contract language had been changed by the insurer in response to the Graham decision.2 Thus, we are presented with the identical question in Villella: whether by drafting variations in exclusionary clause language an insurer may circumvent the "efficient proximate cause" rule of Graham and deny coverage when a covered peril sets in motion a causal chain the last link of which is an excluded peril. The lesson of Villella is that the rule may not be circumvented.
The "efficient proximate cause" rule was first adopted in Washington in Graham v. Public Employees Mut. Ins. Co., 98 Wn.2d 533, 656 P.2d 1077 (1983). In Graham several homeowners sought recovery for losses caused by mudflows resulting from the volcanic eruption of Mount St. Helens. The insurers rejected the claims on the basis of a clause excluding coverage for losses resulting directly or indirectly from earth movement. An exception within the clause granted coverage for damage caused by "explosion . . . resulting from earth movement". Graham, at 535.
The court reasoned that after a jury determined whether the eruption of the volcano was an "explosion" within the *628meaning of the policy, the question would be whether the loss to the properties was a direct result of the explosion. Graham, at 536. If it was a direct result, then coverage would be provided.
In Graham we fashioned a rule that resolves issues of coverage by looking only to the event that results in a loss, which may involve a series of intervening, but not superseding, events. The characterization of subsequent consequential events within the causal chain is irrelevant to the inquiry. If the initial event, the "efficient proximate cause," is a covered peril, then there is coverage under the policy regardless whether subsequent events within the chain, which may be causes-in-fact of the loss, are excluded by the policy.
Applying this rule to the facts of the present case supports the decision of the Court of Appeals. Safeco does not deny that its policy provided coverage for losses resulting from wind or rain. Similarly, Hirschmann does not deny that the destruction of his home was a result of "earth movement." However, expert testimony in the trial court indicated that the "primary cause" of the landslide was a combination of heavy winds followed by record rains. Thus the wind and rain combined to produce the "proximate cause" of the entire loss under the rule in Graham.
Safeco argues that the rule in Graham does not apply to Hirschmann's policy because its exclusionary clause circumvents the rule. However, comparable language in Villella did not. There the exclusionary clause purported to exclude any loss "contributed to or aggravated by" earth movement. Villella did not deny that his loss was "contributed to or aggravated by" earth movement. Hirschmann does not deny that his loss resulted from a sequence including, as its final step, a landslide. Safeco further does not deny that the intent of its language is the same as the intent of Pemco's language in Villella, namely, to avoid coverage of any loss to which earth movement contributes, no matter how slightly. Nor does Safeco deny its language *629is "functionally the same" as Pemco's. We rejected the insurer's arguments in Villella, noting that ” [t]his same argument has been raised and rejected in those jurisdictions which have adopted the 'efficient proximate cause' rule of insurance coverage." Villella, at 817. We again reject the identical arguments here.
In Villella we noted the specific language in Pemco's policy that purported to exclude coverage. It was uncontroverted that an excluded peril "contributed to or aggravated" the loss. Nevertheless, we declined to allow the insurer to circumvent the rule by use of its exclusionary clause. Similarly, we decline to allow Safeco to circumvent the rule by use of its exclusionary clause.
The Graham rule suggests that whenever the term "cause" appears in an exclusionary clause it must be read as "efficient proximate cause." This interpretation is confirmed by Villella. When an insured risk sets into operation a chain of causation in which the last step may be an excluded risk, the exclusion will not defeat recovery. See Villella, at 815-16.
Restated under the Graham formula, Safeco's exclusionary clause might read this way:
We do not cover any loss where the efficient proximate cause is a landslide, whether occurring alone or in any sequence with a covered peril.
It is well settled that in considering summary judgment the court must view the facts and all reasonable inferences from the perspective most favorable to the party opposing the motion. See, e.g., Turngren v. King Cy., 104 Wn.2d 293, 705 P.2d 258 (1985); Wilber Dev. Corp. v. Les Rowland Constr., Inc., 83 Wn.2d 871, 523 P.2d 186 (1974). In this case there was expert testimony from which a finder of fact could conclude that the efficient proximate cause of the destruction of Hirschmann's home was the covered perils of wind and rain, and that the excluded peril of landslide was merely the final link in the causal chain. The summary judgment in favor of Safeco therefore cannot stand.
*630Safeco's references to cases from other jurisdictions are not persuasive. It cites State Farm Fire & Cas. Co. v. Martin, 668 F. Supp. 1379 (C.D. Cal. 1987) as approving language to circumvent California's efficient proximate cause ("dependent concurrent causation") rule. That case in turn relied in part upon Safeco Ins. Co. of Am. v. Guyton, 471 F. Supp. 1126 (C.D. Cal. 1979), aff’d in part, rev'd in part, 692 F.2d 551 (9th Cir. 1982). However, the California Court of Appeals, First District, has stated that the federal court in Guyton misinterpreted California law. See Garvey v. State Farm Fire & Cas. Co.,_Cal. App. 3d_, 227 Cal. Rptr. 209, review granted, 723 P.2d 1248, 229 Cal. Rptr. 663 (1986). Nor can Safeco fail to be aware that the United States Court of Appeals, Ninth Circuit, directly overruled Guyton on the point for which Martin was cited. Safeco Ins. Co. of Am. v. Guyton, 692 F.2d 551, 555 (9th Cir. 1982). Further, Martin is factually distinguishable because in that case the insureds provided "no independent evidence of their own which in any way negates the evidence presented by [State Farm] that excluded perils, including earth movement . . . were the primary causes of the damages." Martin, 668 F. Supp. at 1383. The failure to provide evidence would by itself have been sufficient to sustain the trial court's order of summary judgment in favor of the insurer. Thus, the immediately preceding statement, "even if [a covered peril] was a potential cause, the broad exclusionary language of the contract would disallow coverage" (italics omitted), is gratuitous surplusage of no precedential value.
Safeco also cites in its petition, but does not discuss, State Farm Fire & Cas. Co. v. Paulson, 756 P.2d 764 (Wyo. 1988). The Wyoming court did refuse to find coverage for a loss where the covered peril of "rain" led to damage caused by the excluded peril of "surface water" based upon exclusionary language similar to that in the present case. However, Wyoming has not adopted an "efficient proximate cause" rule such as ours. Indeed, the Paulson court's most *631pointed reference to the "sequential exclusion" clause noted that:
If a policy did not contain a sequential exclusion, as did this one, coverage would exist if an otherwise excluded peril resulted in the occurrence of a covered peril, such as non-covered peril of vandalism resulting in breakage of water pipes which caused covered peril of water damage.
Paulson, at 769 n.2. Paulson, therefore, is not persuasive authority.
Safeco's exclusionary clause may operate to circumvent coverage for losses resulting from causal chains in which excluded perils are the only proximate causes, or chains in which an excluded peril is the efficient proximate cause. However, those questions are not now before this court.
This case is an appeal from an order granting summary judgment. To sustain the Court of Appeals, it is sufficient for this court to find that no question of material fact exists concerning the "efficient proximate cause" of Hirschmann's loss. Such a question, when it exists, is a matter for the trier of fact. See, e.g., Morris v. McNicol, 83 Wn.2d 491, 519 P.2d 7 (1974); Barrie v. Hosts of Am., Inc., 94 Wn.2d 640, 618 P.2d 96 (1980). Here, the record unequivocally indicates that the "primary" cause of Hirschmann's loss was the covered perils of wind and rain.
We affirm the Court of Appeals and remand this case to the trial court for further proceedings.
Utter, Brachtenbach, Dore, Pearson, Andersen, and Durham, JJ., concur.This refers to the deposition of Ronald E. Schlattman, Safeco "vice president of personal lines-underwriting." The following exchange took place:
"Q. Doesn't [the language of your lead-in exclusionary clause] attempt to exclude a cause no matter how slight that cause may contribute to the loss?
"A. Yes, that's the intent."
Safeco proposed an effective date of August 1, 1983, for the language change. The new policy language was in effect at the time of appellants' loss in this case in January 1986. This court filed the Villella opinion October 2, 1986.