CUDAHY, Circuit Judge.
This appeal presents as a sole issue the constitutionality of a provision of the Federal Magistrate Act of 1979, 28 U.S.C. § 636(c), Pub.L. No. 96-82, § 2(2), 93 Stat. 643 (1979), which permits magistrates with the consent of the parties to try civil cases and to enter judgment with respect to them. Jurisdiction for this suit in the district court was based on diversity of citizenship, 28 U.S.C. § 1332. The plaintiff sued on theories of negligence, breach of implied warranties and strict liability in tort. The case was filed on November 30, 1982. On February 3, 1983, the district court referred the matter to the United States magistrate for the Northern District of Indiana for all further proceedings under 28 U.S.C. § 636(c). The plaintiff and defendant consented to trial of the case before the magistrate. A jury trial was held before the magistrate on August 22 and 23, 1983, and a verdict was returned for the defendant. The magistrate ordered the clerk of the court to enter judgment on the jury verdict, and a direct appeal to this court was taken pursuant to 28 U.S.C. § 636(c)(3). The only point raised on appeal is the question whether section 636(c) of the Magistrate Act conflicts with Article III of the United States Constitution and is therefore unconstitutional. We hold that section 636(c) does not conflict with Article III and is therefore constitutional.
I
Section 636(c)(1) of the Federal Magistrate Act provides that, with the consent of both parties, a United States magistrate may conduct all proceedings in a jury or nonjury civil matter and can order the entry of judgment in the case. Section 636(c)(2) provides safeguards to ensure that the consent of the parties is truly voluntary, and, under section 636(c)(6), the reference to the magistrate may be withdrawn at any time by the district court for good cause on its own motion or under extraordinary circumstances shown by any party.
After entry of judgment by a magistrate, a party may appeal directly to the appropriate court of appeals in the same manner and presumably subject to the same standards of review as an appeal is taken from [1039]*1039a judgment of the district court. The consent of the parties allows a magistrate to exercise the civil jurisdiction granted in section 636(c)(1) and to direct the entry of judgment of the district court. The parties may also, however, at the time of the reference to the magistrate, consent to appeal on the record to a judge of the district court in the same manner and subject to the same standards of review as on an appeal from a judgment of the district court to a court of appeals. The district court may then affirm, reverse, modify or remand the magistrate’s judgment. In this latter situation, a case may be reviewed by the appropriate court of appeals at its discretion only upon petition for leave to appeal by a party stating specific objections to the judgment. In either case, there is no limitation on any party’s right to seek review by the Supreme Court.
Six circuits have, to date, considered the constitutionality of section 636(c) of the Magistrate Act. Panels of four circuits have held that the provision is constitutional: the First Circuit in Goldstein v. Kelleher, 728 F.2d 32 (1st Cir.1984); the Second Circuit in Collins v. Foreman, 729 F.2d 108 (2d Cir.1984), petition for cert. filed, 52 U.S.L.W. 3777 (U.S. April 2, 1984) (No. 83-1616); the Third Circuit in Wharton-Thomas v. United States, 721 F.2d 922 (3d Cir.1983), and the Fifth Circuit in Puryear v. Ede’s Ltd., 731 F.2d 1153 (5th Cir.1984). The Eighth Circuit has reached the same conclusion in Lehman Brothers Kuhn Loeb, Inc. v. Clark Oil Refining Corp., 739 F.2d 1313 (8th Cir.1984) (en banc). A panel of the Ninth Circuit held that the provision was unconstitutional in Pacemaker Diagnostic Clinic of America, Inc. v. Instromedix, Inc., 712 F.2d 1305 (9th Cir.1983), but this decision was vacated and the provision was held constitutional by the Ninth Circuit sitting en banc, 725 F.2d 537 (9th Cir.1984), petition for cert. filed, 52 U.S.L.W. 3875 (U.S. May 16, 1984) (No. 83-1873).
The essence of the claimed constitutional conflict is that the statutory provision permits magistrates to exercise the judicial power of the United States although they do not conform to the requirements of Article III of the Constitution which provides
The judicial power of the United States, shall be vested in one Supreme Court, and in such inferior Courts as the Congress may from time to time ordain and establish. The Judges, both of the supreme and inferior Courts, shall hold their offices during good behavior, and shall, at stated times, receive for their services, a compensation, which shall not be diminished during their continuance in office.
The constitutional guarantees of tenure during good behavior and of protection against reduction in compensation are the bulwarks of independence of the federal judiciary against reprisal, fear of reprisal or undue influence from any quarter and particularly from the other branches of the federal government. Judicial independence is crucial to the preservation of our system of government as has been demonstrated throughout the history of the Republic. Despite the current pressure exerted by unprecedented docket burdens in a litigious society, no "work emergency” is adequate grounds for undermining the constitutional guarantees of an independent judiciary.
In contrast to the constitutional guarantees of life tenure afforded to federal judges by Article III, magistrates serve for only eight-year terms but may be reappointed. Magistrates are appointed by vote of a majority of the district judges sitting in a particular district or, lacking such a majority, by the chief judge. A magistrate may be removed by the district judge or judges during his or her term only for incompetency, misconduct, neglect of duty or physical or mental disability. 28 U.S.C. § 631. The compensation of magistrates may not be reduced during a term below the compensation fixed at the beginning of the term, 28 U.S.C. § 634(b), although Congress could presumably effect such a reduction by repealing or overriding this statutory provision.
[1040]*1040II
The case before us concerns the exercise of the federal judicial power in adjudicating private, state-created rights under the fed-, eral diversity jurisdiction granted in Article III. This case therefore specifically concerns the same type of rights as were at issue in Northern Pipeline Construction Co. v. Marathon Pipe Line Co., 458 U.S. 50, 71, 102 S.Ct. 2858, 2871, 73 L.Ed.2d 598 (1982) — in Marathon, the right to contract damages and, in the case before us, the right to tort damages. None of the exceptions delineated by the plurality in Marathon as exempt from Article III strictures is therefore applicable: the “territorial courts” including the District of Columbia courts, the courts-martial and the administrative tribunals, which primarily adjudicate “public rights.” Id. at 63-70, 102 S.Ct. at 2867-71. Cf. Wharton-Thomas, supra, 721 F.2d at 930 (reaching the constitutional issue although case, based on the Federal Tort Claims Act, was arguably limited to a federally-created right).
The central issue in determining the constitutionality of section 636(c) is therefore whether the magistrates, in exercising the powers granted to them in this section, are performing the functions of federal judges. Because they clearly do not enjoy the Article III judicial protections, if they are acting as judges, then the provision is unconstitutional. If, however, they are in fact acting as “adjuncts” and are not exercising the federal judicial power, then this provision is constitutional.1 In order to make this determination, we must examine several facets of the functioning of the magistrates. In Marathon, the Supreme éourt declared the Bankruptcy Act of 1978 unconstitutional because it conferred federal judicial powers on bankruptcy judges who were not accorded Article III protections. This decision, therefore, provides a model for determining whether the magistrates are to be considered as adjuncts to the district court or as judges in fact, if not in name.
In addition, it is possible to identify two preeminent values which the Article III protections are intended to safeguard, much as did the Ninth Circuit in Pacemaker, 725 F.2d at 541. An examination of these values illustrates whether the powers which are given to the magistrates under section 636(c) constitute an excessive delegation of authority to the magistrates so as to negate their status as adjuncts of the district judges. These two values implicated by Article III are the right of the litigant to exercise of the federal judicial power by an independent federal judiciary and the preservation of the separation of powers among the three branches of government.
A
In considering the constitutional values to be maintained, we must first examine the proposition that the required voluntary consent of the parties to trial of civil matters before magistrates obviates the need for Article III protections. The fact that magistrates can exercise their authority only with the litigants’ consent is a significant distinction from the nature of the authority exercised by the district judges. The statute attempts to ensure the voluntary character of this consent and specifically provides that the district court is authorized only to inform the parties of the alternative of a reference to a magistrate and is then forbidden to persuade or induce the parties to accept the reference. In addition, litigants retain the option of asking the district court, albeit only under [1041]*1041extraordinary circumstances, to withdraw the reference.2
The element of required consent in section 636(c) serves to distinguish this provision from both the Bankruptcy Act at issue in Marathon and the section 636(b) provision at issue in United States v. Raddatz, 447 U.S. 667, 100 S.Ct. 2406, 65 L.Ed.2d 424 (1980). In fact, in Marathon, the Court used the absence of any requirement for the parties’ consent as one of the three primary characteristics distinguishing the 1978 Bankruptcy Act from the earlier system of bankruptcy adjudication. Under the prior law, the consent of the parties was needed to authorize bankruptcy referees to exercise jurisdiction over controversies beyond those involving property in the possession of the court. 458 U.S. at 79-80 n. 31, 102 S.Ct. at 2875-76 n. 31. See also 458 U.S. at 91, 102 S.Ct. at 2881 (Rehnquist, J., concurring).
The magistrate reference system is evidently intended to provide a quicker and less costly alternative to the usually more delayed adjudication in a district court. Congress explicitly stated that one of the purposes of the 1979 revisions was to “improve access to the federal courts for the less-advantaged.” S.Rep. No. 74, 96th Cong. 1st Sess. 1, reprinted in 1979 U.S. Code Cong. & Ad.News 1469, 1469. While this objective may seem to be desirable, some opponents of the system have argued that its workings may result in a double system of justice — a second-class magistrate system for the impecunious and a first-class district court system for the wealthy and for large corporations. Pacemaker, 725 F.2d at 553-54. Such a possibility, however, seems to be little more than speculation, at least at this time. In fact, the unlikelihood of this kind of development is apparently bolstered by the fact that parties, regardless of their wealth, often voluntarily choose to settle their disputes through recourse to the extra-judicial avenue of arbitration rather than through the costlier and often much lengthier judicial process. Arbitrated decisions may, in some instances, end up in court for final resolution, but the parties fully realize that in subsequent court litigation the arbitration award is ordinarily denied de novo consideration; instead, the court in its usual practice accords great deference to the decision of the arbitrator. See, e.g., DeCosta v. Columbia Broadcasting System, Inc., 520 F.2d 499, 505-06 (1st Cir.1975), cert. denied, 423 U.S. 1073, 96 S.Ct. 856, 47 L.Ed.2d 83 (1976) (equating consensual reference to magistrate with arbitration, fact-finding of both being subject to “clearly erroneous” standard).
Of course, a major objection to trial of civil matters by reference to magistrates is that the exercise of the judicial power of the United States cannot depend upon stipulations of the litigants. See Pacemaker, supra, 725 F.2d at 548-52. The Constitution requires that the judicial power be exercised by Article III judges, and the objectors contend that .this is a requirement of subject matter jurisdiction which is not waivable by the parties in favor of a non-Article III judge. The right of access to Article III judges, however, may be viewed more correctly as personal to the litigants and therefore subject to waiver. Pacemaker, 725 F.2d at 541-43. This characterization is also more realistic in that the issue is not an expansion of the subject matter jurisdiction of the federal courts but rather the parties’ choice of the forum in which the suit, for which federal jurisdic[1042]*1042tion clearly exists, is to be litigated. Once the parties have waived their right to Article III protections, they should not be allowed to challenge the constitutionality of the provisions under which they voluntarily chose to proceed. The Pacemaker majority, in fact, pointed out that criminal defendants are permitted to waive fundamental personal rights including the right to counsel, the right to trial by jury and even the right to any trial at all. Id. at 543. See also Goldstein v. Kelleher, 728 F.2d at 35.
In this connection, however, we emphasize a very important point. We do not believe that reference of civil matters to magistrates can be sustained against constitutional challenge unless an alternative of trial before an Article III judge is maintained for all litigants as a realistic and viable alternative. If a litigant were required to wait ten years for a trial before an Article III judge in lieu of a prompt trial before a magistrate, we would have little difficulty finding that the constitutional grant of jurisdiction had been frustrated. The dissenters in Pacemaker point out the very real objection that
In the Magistrates Act Congress has delegated to judicial councils the power to create magistrate positions and thereby has abdicated [the] constitutional responsibility [to create judgeships]. In practical terms, this abdication reduces the pressure on Congress to create more Article III judgeships, and increases the pressure on district courts to escalate the use of magistrates.
725 F.2d at 549.
It is entirely possible that the agencies of judgeship creation will operate as the Pacemaker dissenters fear. If this is the case, we can foresee the balance of constitutionality moving against the magistrate system. As long, however, as an appropriate balance is maintained between magistrates and Article III judges so that a real option of trial by an Article III judge is preserved, we cannot say that the Constitution has been violated.
Of course, a related and fundamental concern affecting the magistrate reference system is whether the consent of the litigants can ever be said to be truly voluntary. According to the dissent in Pacemaker, 725 F.2d at 553-54, such consent is not voluntary but, rather, coerced by the very pressures which led to the creation of the magistrate system — the cost and delay of litigation in the federal district courts. See also Phillips v. TV.A. Engineering Assoc., Inc., petition for cert. filed, 52 U.S.L.W. 3910 (U.S. April 23, 1984) (No. 83-1732) (one question presented concerns coerced consent to reference to magistrate), petitioning from 725 F.2d 684 (6th Cir.1983) (mem.). These pressures might seem to render the consent of those litigants who can least afford the expense, and for whom the magistrate system was ostensibly created, of somewhat questionable voluntariness. But this concern is almost entirely speculative; there is no hard evidence of economic, or other systemic, coercion. In addition, the statute itself bends over backward to preclude undue influence by either judges or magistrates. To invalidate the magistrate reference system would only be to increase further the cost and delay of district court litigation for those — whether poor or wealthy — who prefer to litigate in district court. Such an invalidation would also deny to those who would choose a magistrate a reasonably quick and less costly alternative. Again the result of invalidation might well be to deny to less advantaged litigants access to the federal courts altogether. At the same time those who would persevere with the existing system of district courts, despite its manifest inadequacies, would be severely burdened.
B
Consideration of the second value — the preservation of the independence of the judiciary in relation to the other branches of the government — requires an examination of several features of the magistrate system as embodied in section 636(c). We should reiterate that the use of magistrates in other contexts and for other purposes has been held to be constitutional, as in [1043]*1043Raddatz. The use of other district court “adjuncts,” such as the referees in the former bankruptcy system, has also passed the test of constitutionality.
The crucial question is whether section 636(c) presents a magistrate system which is truly an “adjunct” of the district court or one which constitutes an independent (and extra-constitutional) court system. The bankruptcy courts established by the Bankruptcy Act of 1978 have been explicitly characterized as being independent of the district courts. Marathon, 458 U.S. at 80 n. 31, 102 S.Ct. at 2876 n. 31. The bankruptcy judges were to be appointed by the President with the advice and consent of the Senate and these judges were subject to removal by the judicial council of the circuit. Not only was the jurisdiction of the bankruptcy courts independent of the parties’ consent but there was no procedure by which the district court could review the proceedings of the bankruptcy court while they were being carried on or could remove a case from the bankruptcy court’s jurisdiction on its own initiative or upon request of the parties.
The magistrates, on the other hand, are subject to appointment and removal only by the district judges (and, in some circumstances, to removal by the circuit judicial council). While such a dependent relationship may, to some extent, implicate the independence with which the magistrates exercise their judgment, it does not implicate either the judiciary’s or the magistrates’ independence vis a vis the Executive or Legislative Branches. In addition, the district court retains supervisory authority over the proceedings of the magistrate, while they are in process, through its ability to withdraw the reference.
The Supreme Court in Marathon identified the most significant difference between the old and the new bankruptcy systems as being the role of the old bankruptcy referees, in fact as well as in name, as “subordinate adjuncts of the district court” in contrast to the independence of the new bankruptcy judges. Two of the relevant factors identified by the Court were that the referees were appointed and removable only by the district court and that the district court retained control by its power to withdraw a case from a referee. As the Court said, “[t]hus, even at the trial stage, the parties had access to an independent judicial officer.” The dependence of the referees and of the magistrates was primarily on the judiciary, not the political branches of government. And as the Marathon plurality paraphrased Justice Black-mun’s comment in Raddatz, the primary “danger of a ‘threat’ to the ‘independence’ of the [adjunct came] from within, rather than without, the judicial department.” 458 U.S. at 80 n. 31, 102 S.Ct. at 2876 n. 31 (quoting 447 U.S. at 685, 100 S.Ct. at 2417 (Blackmun, J., concurring)). In other words, “subordinate adjuncts,” by reason of their close dependence on independent judicial officers, are shielded against influence from the political branches of government.
In this connection, perhaps the most troubling aspect of the magistrate system with respect to its status as a mere adjunct of the district court is the fact that a magistrate can order entry of final judgment. Only the new bankruptcy system, declared unconstitutional in Marathon, also permitted a non-Article III officer to enter final judgment. But the power to enter final judgment which, in the case of the magistrates under section 636(c), can only be exercised with the consent of the parties, may be viewed more as a technical than as a substantial encroachment on the independence of the judiciary. The reason for regarding this power to enter final judgment as “technical” is that, for example, even under section 636(b) — approved in Raddatz — the judge in reality does not generally give full de novo consideration to a magistrate’s recommendation. Instead, the judge proceeds in fact in a fashion not drastically different from that utilized in reviewing a section 636(c) judgment by either the district or appellate court.3
[1044]*1044The First Circuit in Goldstein v. Kelleher, 728 F.2d at 35-36, pointed out that the Supreme Court has approved consensual grants of authority to masters and referees even to the point where their decisions “have the same force and effect as a judgment of [a federal district] court.” Beckers v. Fowler, 69 U.S. (2 Wall.) 123, 127, 17 L.Ed. 759 (1864). In addition, judgment was entered on the referee’s report pursuant to the district court’s initial order of reference, and the Court essentially applied a waiver theory to prevent the litigants from subsequently challenging this authority. Id. Similarly, in Kimberly v. Arms, 129 U.S. 512, 524-25, 9 S.Ct. 355, 359, 32 L.Ed. 764 (1889), the Supreme Court held that the ruling of a master should not be disturbed “unless clearly in conflict with the weight of the evidence.”
This raises the perhaps more significant question, with respect to the status of magistrates, of the type of review which a district or appellate court may give to the magistrate’s determinations. This inquiry, however, does not lead to an objection of constitutional magnitude. District courts and, occasionally, appellate courts are accustomed to deferring to non-Article III officials acting as factfinders. Examples include not only administrative agencies (which fall under the administrative tribunal exception previously mentioned) but also special masters and magistrates operating under section 636(b). In Raddatz, the Court pointed out that, although the district court was to make a de novo determination with respect to the suppression of evidence, it was not required to hold a new hearing. A requirement of a new hearing would obviously defeat the purpose of the reference to a magistrate. That purpose would also be defeated by de novo consideration of a magistrate’s actions under section 636(c). Deference to a magistrate with respect to fact-finding thus does not seem to implicate questions of judicial independence because the reviewing court— whether a district court or a court of appeals — would, of course, give any determinations of law full independent consideration. Thus, entry of final judgment by a magistrate is not of constitutional significance if the result is deference only to fact-finding and with full review of legal questions in the Article III courts.
Despite our view of the authority to enter final judgment as having technical significance only, perhaps some clear line of demarcation between the power of an Article III judicial officer and a magistrate is required. Such a line of distinction may be found in the allocation of the contempt power, because, under no aspect of the Magistrate Act, can a magistrate punish for contempt. 28 U.S.C. § 636(e). According to section 636(e), if an individual commits an act constituting contempt of court, the magistrate must certify the facts of the incident to a district judge. The judge, after holding a hearing and evaluating the allegedly contemptuous conduct, may determine the nature and severity of appropriate punishment, if indicated. Unlike the relatively mechanical entry of judgment, the power to punish for contempt of court is the means by which many court judgments, not including the collection of money judgments, are enforced. Despite the effort of the dissent to trivialize the significance of the contempt power, it remains the primary means of enforcing many court judgments, particularly injunctions. The vesting of this power exclusively in the hands of Article III judicial officers would seem, for present purposes at least, to provide an adequate distinction between such judges and non-Article III officers. This clear line also serves to limit the ultimate exercise of judicial power to persons enjoying the constitutional guarantee of independence.
The dissent raises a number of serious arguments which we must consider carefully since the independence and integrity of [1045]*1045the federal judiciary is too important a matter to be sacrificed to pride of opinion. As the dissent itself concedes, however, many of its arguments are quite speculative. It seems to assume that the 4% of federal trials now conducted by magistrates may grow to 40% in a process that will eventually drown the federal courts. This need not be the case; in fact the dangers are so clear and so fearsome that they are likely to be avoided. As we have said, a radical shift to trial by magistrate could easily result in a finding of unconstitutionality on the new facts. Under all the circumstances, this is not a situation where some experimentation with traditional Article III arrangements is necessarily some sort of faustian bargain.
In any event, it is ironic that, to avoid recourse to such expedients as the trial of some diversity cases by magistrates, the dissent would consider the abolition of federal diversity jurisdiction altogether. The loss to the federal scheme by ending diversity jurisdiction might far exceed any speculative loss in empowering magistrates to try these cases. We would, in effect, throw out the baby with the bath water. The point is that heavy caseloads cause losses somewhere and it is our task not to deny that losses exist but to hold them to a minimum.
In the best of all possible constitutional worlds, there would perhaps be no non-Article III judicial officers. Judicial independence and the purity of the constitutional grant of judicial power might be best assured by barring the door entirely to those unclothed with the constitutional protections. But the pressure to decide cases calls for reasonable measures in response. This does not necessarily mean a flood of new Article III judges, a drastic narrowing of the federal jurisdiction or a higher price of access to the federal courts. It may reasonably mean such programs as the Magistrate Act which are carefully designed to protect the essential values of Article III according to well accepted principles.
As a practical matter, we do not know whether the extremely rapid growth of federal litigation since the early 1960’s is a relatively permanent feature of the legal environment or whether this will appear as an exceptional departure from the long-term trend. The provisions of section 636(c) seem to us to be an appropriate reaction to a crisis, the future course of which is unknown. Presumably, the creation of permanent district judgeships is something that should be keyed primarily to long-term developments as best we can discern them. The system of magistrate reference of civil cases is a flexible mechanism, which seems well-tailored to helping to absorb the surge of litigation which has caused the crisis with which we are now coping — provided, of course, that the key constitutional values can be maintained and preserved.
In light of these considerations, we conclude that magistrates continue to function as adjuncts of the district courts and that the right of individual litigants to access to Article III judicial officers, the independence of the judiciary from the executive and legislative branches and the separation of powers among the three branches of the federal government are preserved in section 636(c) of the Magistrate Act. This conclusion is valid despite the authority granted to the magistrates to hear and enter judgment in civil cases. We therefore hold that section 636(c) of the Magistrate Act is constitutional and we affirm the judgment below.