Founded on English common law, "attorney-client privilege" is a well-established and robust protection that is recognized in all courts in the United States. Where the privilege applies, it protects from compelled disclosure an oral communication or a document in any form. In addition, courts in the United States recognize a "work product doctrine," often called a "work product privilege," which protects from disclosure those documents, in any form, that were prepared by the attorney or by an agent of the attorney, "in anticipation of litigation."

These protections are durable and, assuming they are not waived, will apply not only to the case that occasioned the establishment of the protection, but also to any related or unrelated legal proceedings thereafter. But waiver, too, is enduring. Once waived, or deemed waived, the privilege or protection cannot be recaptured.

The importance of these protections in the context of litigation in the United States cannot be overemphasized. Unlike the vast majority of nations, whose courts do not permit litigants to engage in extensive formal discovery, the judicial systems in the United States provide for the liberal production and examination of documents and other information in preparation for trial, as well as for the testimony of, and production of documents by, witnesses who are not even parties to the case. Balancing the right of the litigant to engage in such "US-style" discovery by seeking information from the opponent is the right of the opponent to these protections. Thus, generally, litigants may preserve communications that are privileged because they were made between an attorney and their client for the purpose of seeking or giving legal advice or that are "work product," because they are documents prepared by the attorney in order to defend the client or to advocate on their behalf.

The two protections are considered separately below:

Attorney-client privilege

The attorney-client privilege is recognized under federal common law and is also codified in state statutes. Additionally, under the American Bar Association Model Rules of Professional Conduct, attorneys in the United States are bound by an ethical duty to keep confidential any privileged communications or other confidential information, the disclosure of which would likely be detrimental to the client. This obligation is subject to specific exceptions, as discussed below. The attorney-client privilege is based on the societal goal of fostering the relationship between the lawyer and the client by facilitating the free exchange of information.

General criteria for the attorney-client privilege

Iterations of the attorney-client privilege by scholars and courts are innumerable, but four basic criteria can be said to define the privilege:

• A communication
• Made between privileged persons (attorney and client)
• In confidence
• For the purpose of seeking, obtaining, or providing legal assistance

The privilege applies to both oral communications and documents and other records that reflect communications between privileged persons, including emails. Importantly, a pre-existing document conveyed by a client to an attorney does not become privileged as a result of the transfer. The content within the document itself must be privileged for the document to be protected by the privilege. Similarly, with regard to emails, parties cannot assert the privilege over communications where they have merely cc'd an attorney. Rather, emails, like all other communications, are only protected by the privilege if they reflect a confidential request for legal advice.

In the corporate context, it is often difficult to determine whether the purpose of a particular communication was to seek, obtain, or provide legal assistance because communications often have mixed business and legal purposes. Under these circumstances, courts uphold the privilege if the primary purpose of the communication was to obtain or provide legal advice.

To be recognized in legal proceedings, the attorney-client privilege must be affirmatively asserted or invoked by the client. A further condition for the application of the privilege is that it has not been "waived" by the client with respect to the communication at issue, such as by disclosing the content of the communication to, or making an oral communication in the presence of, a third party.

Although the client is generally the holder of the attorney-client privilege, an exception is that, in some jurisdictions, communications intended only for internal review by a law firm that represents the client are privileged and need not be disclosed to the client — even though the client presumptively has access to the file maintained by the attorney. Thus, in the securities litigation In re Refco Securities Litigation, where investors sued a law firm for legal malpractice with respect to a bankruptcy and for aiding and abetting a breach of fiduciary duty, a federal court ruled, under New York law, that the law firm's internal email about the client was not discoverable absent a showing of a clear need. Permitting a law firm to withhold such internal communications is in the client's interest, allowing attorneys to privately record their thoughts in order to ensure effective representation.

Choice of law for attorney-client privilege

The judicial system in the United States is marked by two "sovereign" systems, state and federal, that are governed by different procedural rules and by different substantive law of privilege and work product. Every court in the United States, state or federal, recognizes the attorney-client privilege where the basic criteria enumerated above are found to apply. Although the basic privilege precept remains consistent from jurisdiction to jurisdiction, the scope of the privilege may vary considerably under particular circumstances, as may the conditions that constitute its waiver. Given the variations in the privilege laws across jurisdictions, an important consideration in current or prospective litigation is the choice of law, i.e., whose privilege law applies.

The procedural aspects of privilege in federal courts are governed by the Federal Rules of Evidence and the Federal Rules of Civil Procedure, and in state courts by similar rules enacted by the legislatures of those states. As to the substantive law, the federal courts usually apply either the federal common law of privilege, or else the substantive privilege law of the US state in which they sit.

In federal court, the choice depends on the basis for the federal court's subject-matter jurisdiction to hear the claims at issue. Where subject-matter jurisdiction is based on a "federal question" — that is, where the claim itself arises under federal law — the federal common law of privilege applies to determine the substantive law of privilege. Where, however, the federal court's subject-matter jurisdiction is based on the amount in controversy and complete "diversity of citizenship" — that is, where the alleged damages exceed USD 75,000 and the claims are founded on state law, but the plaintiff and defendant hail from different nations or from different US states — the court, by default, will apply the substantive privilege law of the state in which the federal court sits ("forum state").

Parties may occasionally attempt to gain a tactical advantage by invoking the privilege law of a jurisdiction that is more favorable to them than the privilege law that would normally apply — either to protect their own documents from disclosure or to convince the court that an adversary's documents should be disclosed. Under these circumstances, the party arguing that the court should apply a foreign jurisdiction's privilege must satisfy two criteria: (i) it must show that the foreign privilege law conflicts with the law that would ordinarily apply (i.e., the communications are privileged under the laws of only one of the jurisdictions); and (ii) if the court agrees that the two jurisdictions' privilege laws are in conflict, it must show that the foreign jurisdiction has the "most significant relationship" with the communications at issue. Under this test, the jurisdiction with the most significant relationship with the communication at issue will usually be the jurisdiction in which that communication took place. The most significant relationship test, which most state courts have adopted, is liberal and often favors disclosure, permitting the communication to be entered into evidence if either it would be admissible under the privilege law of the forum state (absent a special reason that weighs in favor of nondisclosure) or it would be admissible under the privilege law of the jurisdiction with the most significant relationship with the communication (absent a strong policy of the forum state indicating that nondisclosure is appropriate).

In the international context, application of the most significant relationship test may put a non-US litigant at a disadvantage because privilege laws in other nations are generally not as robust. For example, a US resident who files a lawsuit in the United States against a corporation in another nation may seek to invoke the privilege law of that nation so as to require the corporation to produce its communications with its in-house counsel abroad. If the communications occurred in the non-US jurisdiction, the court may determine that the non-US jurisdiction has the most significant relationship with the communications and apply the privilege law of that jurisdiction. If the court determines that the communication would not be privileged under the foreign law but would be privileged under the applicable US law, the communication will be admitted absent a strong public policy of the forum.

And even if the reverse is true — that is, the communication would be privileged under the foreign law but is discoverable under US law — the court likely will deem the communication discoverable provided that no special reason barred admissibility.

Some courts also apply a "touch base" analysis to determine whether a communication that occurred in a foreign country is protected by the privilege. Under this test, the communications will be protected by the privilege provided that they touch base with the United States, in that they have had more than an incidental connection with the United States. For example, in Gucci Am., Inc. v. Guess?, Inc., 271 F.R.D. 58 (S.D.N.Y. 2010), the court ruled that communications with Italian counsel, and maintained on an Italian server, touched base with the United States because they related to a company's legal strategy to prosecute trademark infringement actions in both Italy and the US. Accordingly, it applied US law to determine whether the communications were privileged.

In short, the choice-of-privilege-law question in diversity-based suits in federal court may be of particular consequence where the communications at issue occurred or were produced outside of the United States, particularly where they were between in-house counsel and their corporate clients in jurisdictions that do not recognize a privilege with respect to communications between corporate representatives and in-house counsel.

Other types of privilege

In addition to attorney-client privilege, some US jurisdictions recognize other types of privileges that protect communications. For example, some US states recognize a privilege with respect to communications or confidences between physicians and patients, or between priests or ministers and their confessors (although these privileges are not recognized under federal common law or the Federal Rules of Evidence). Some states have recognized an accountant-client privilege. The Federally Authorized Tax Practitioner Privilege extends common law privileges to certain communications between a taxpayer and a federally authorized tax practitioner for the purposes of noncriminal tax matters before the US Internal Revenue Service. Communications of foreign patent agents may also be entitled to a form of privilege in courts in the United States.

Work product protection, also called work product privilege, differs from attorney-client privilege in respects that are best understood in the context of the differing goals of the two protections. While the attorney-client privilege exists to encourage free communication between attorneys and their clients so as to protect and foster that important relationship, the work product doctrine exists to encourage careful and thorough preparation by the attorney. Work product protection thus serves the adversarial system.

Additionally, work product protection is invoked and waived by the attorney, in contrast to attorney-client privilege, which belongs to and is invoked by the client. Because of this difference, a waiver of work product protection does not constitute a waiver of attorney-client privilege, and vice-versa. Another major difference is that while the attorney-client privilege protects communications, the work product doctrine protects from disclosure documents (in any form) and tangible things. And while the attorney-client privilege is nearly absolute — the finding that a communication or document is privileged bars compelled disclosure — a court may compel the disclosure of work product under certain conditions, depending on the nature of the documents or things and the opposing party's ability to demonstrate a need for the information therein.

Like attorney-client privilege, the work product doctrine varies from state to state. However, federal courts generally apply the federal work product law, regardless of the basis of their subject-matter jurisdiction with respect to the claims at issue. This section considers the federal rules governing work product protection, although (as with the attorney-client privilege) state statutes codifying work product protection, and used by the state courts, vary from state to state.

The federal definition of "work product," codified in the Federal Rules of Civil Procedure, is typical, providing a qualified protection from discovery when the items sought are:

• Documents and tangible things that are otherwise discoverable
• Prepared in anticipation of litigation or for trial
• Prepared by or for another party or by or for that other party's representative

Key to identifying protected work product is the finding that the documents or things were prepared "in anticipation of litigation." Actual litigation need not have commenced at the time the attorney prepared the documents or things at issue. However, the court, in applying work product protection, must be satisfied that they were prepared at a time when there was a risk of litigation and that they were not merely prepared in the ordinary course of business. Generally, the federal courts will follow either of two tests for determining whether a document was prepared "in anticipation of litigation":

• The "primary purpose" test, in which the court must, in order for the work product doctrine to apply, find that the primary motivating purpose behind the creation of the documents or things at issue was to aid in possible future litigation
• The "because of test," in which the court must, in order for the work product doctrine to apply, find that the documents or things at issue can fairly be said to have been prepared or obtained because of the prospect of litigation in light of the nature of the document and the factual situation in the particular case

The latter test does not consider the motivation for creating the document and, therefore, is arguably the more protective and more inclusive of the two tests — especially with respect to documents or things prepared both for a regular business need and for the purposes of current or future litigation.

However, even if the documents or things at issue are found to have been prepared in anticipation of litigation under the relevant test (and all other requirements for the work product doctrine are met), the party seeking disclosure may overcome the work product protection. In the federal courts, under the Federal Rules of Civil Procedure, work product privilege is overcome when the party seeking disclosure demonstrates both of the following:

• A "substantial need" for the documents or materials in order to prepare the party's case
• The inability of the party seeking discovery to obtain the substantial equivalent of the information by other means without "undue hardship"

Federal courts have determined that a "substantial need" for the information must be specifically articulated by a litigant who argues for the disclosure; general statements by the party seeking the discovery will not suffice. Courts may conduct an in camera inspection of the documents or things to determine whether they serve a "substantial need." For example, the fact that an accident site or other key locale has fundamentally altered, thus depriving one party of the ability to collect evidence from the site, may give rise to a "substantial need" that entitles that litigant to the work product that would otherwise be protected. A showing that a key witness is deceased, mentally ill, or unavailable may constitute a "substantial need" for witness statements or other work product of an attorney. Consideration of financial cost alone does not, however, constitute a "substantial need."

Cost may, however, be a consideration in the analysis of "undue hardship," the second requirement a litigant must meet in order to overcome its adversary's right to work product protection. Courts have found "undue hardship" where re-creating the information contained in the work product would be time-consuming, impracticable because of the volume of information at issue, and costly. In determining whether the "undue hardship" criterion is met, the court will likely weigh the abilities and resources of the parties to undertake these efforts.

Despite these exceptions to the work product doctrine, an attorney's mental impressions, conclusions, opinions, and legal theories remain protected as "core" or "opinion" work product. An example of core work product is prior drafts of a brief, letter, or answers to discovery requests. If prepared with the client's assistance, such documents would also be subject to attorney-client privilege and would not be discoverable. Documents and things that constitute core or opinion work product of one party's attorney are almost never discoverable under the Federal Rules of Civil Procedure, even if the opponent can show a "substantial need" for such information and "undue hardship" resulting from nondisclosure.

The fact that a communication is protected by the attorney-client privilege does not mean that a document containing that communication is protected work product, or vice versa. Generally, courts in the United States, whether state or federal, require parties to identify on a "privilege log" documents that are withheld during discovery on the basis of the attorney-client privilege (or other privilege) or on the grounds of work product protection. Failure to note that a privileged document is also work product may foreclose a party's ability to invoke work product protection should the claim of privilege fail. Therefore, the privilege log must be carefully drafted.

Changes to Rule 26 of the Federal Rules of Civil Procedure ("Rule") have expressly extended work product protection to all drafts of expert reports in any form and exempt from discovery most communications between the trial counsel for a party and those of its testifying experts whom the Rule requires to submit an expert report. Under this "attorney-expert" privilege, all communications, whether oral, written, or electronic, between trial counsel and the expert are protected unless they relate to the testifying expert's compensation; they identify facts or data that the party's attorney provided, and that the expert considered, in forming the opinions expressed; or they identify assumptions that the party's attorney provided to the expert, and upon which the expert relied in forming their opinions. Since the enactment of these changes, courts have grappled with questions left unanswered by the Rule's new language. For instance, district courts have come to different conclusions as to whether an expert's notes (as opposed to their drafts of an expert report) related to a matter are protected. Similarly, there is some uncertainty as to whether communications between attorneys and experts who are not required by the Rule to submit an expert report are privileged or discoverable.