People v. Allen

Brickley, J.

These cases were consolidated in order to resolve differing interpretations of the application of MRE 609(a) to the practice of impeaching criminal defendants by prior conviction.

We find that MRE 609(a), as presently constituted, leads to an interpretation that encourages the use of "bad man” evidence without a commensurate gain in the evaluation of a witness’ testimony. More specifically, the central element in the exercise of discretion in this area, i.e., the extent to which the prior conviction is probative of credibility, has been greatly de-emphasized. The lower courts have focused instead on other issues such as the need for credibility evidence. The resulting difficulty in interpreting this important rule of evidence more often acts to the detriment of the defendant’s right to testify than to the service of the jury’s ability to evaluate a witness’ credibility.

The approach of the dissenting opinions, in our *564view, does not clarify the present confusion and, in one respect, adds a standard that creates an irreconcilable conflict in the application of MRE 609.

In order to resolve the cases before us, we therefore set forth a more specific procedure for the exercise of discretion under the present MRE 609(a) which we think more fully comports with its original intent. We also take this opportunity to promulgate an amendment to MRE 609(a) which will apply to all cases tried after March 1, 1988. (See appendix a.) It provides for bright-line rules that recognize and exclude certain prior convictions which are inherently more prejudicial than probative and allows admission of those convictions which are inherently more probative than prejudicial. For those crimes that fall in between, because their relationship to credibility is less clear, the amendment adopts the same standard for the exercise of judicial discretion as is set forth herein for the resolution of these cases and for cases yet to be resolved under MRE 609(a) prior to its amendment.

i

The facts of these cases are succinctly set forth in the opinion of Chief Justice Riley.

ii

The use of prior convictions to impeach a witness-accused has long been controversial. This controversy is apparent in the debates surrounding the adoption of Michigan Rule of Evidence 609,1 as *565well as the relevant federal rule.2 The question has been the source of more scholarly commentaries than can be cited here.3 What underlies this controversy is the clash between the fundamental principle defined in MRE 404 that character evidence (including prior conviction evidence) may not be admitted to prove that a defendant acted in conformity therewith and the practice of permit*566ting the admission of prior conviction evidence as contained in MRE 609 for purposes of impeaching a witness-accused’s credibility when he testifies in his own behalf.4

A

There can be little doubt that an individual with a substantial criminal history is more likely to have committed a crime than is an individual free of past criminal activity. Nevertheless, in our system of jurisprudence, we try cases, rather than persons, and thus a jury may look only to the evidence of the events in question, not defendant’s prior acts in reaching its verdict.5 See United *567States v Mitchell, 2 US (2 Dali) 348, 357; 1 L Ed 410 (1795).6

However, when a defendant testifies, he takes on a second role, that of a witness. And the question of the reliability of a witness’ testimony, though a collateral issue, may play an important role in a jury’s decision. We therefore permit the introduction of evidence to prove that a witness is not worthy of belief. For example, the witness’ ability to perceive or understand events may be questioned. (MRE 601.) Similarly, the witness’ testimony may be disallowed if he has no personal knowledge of the events about which he testifies. (MRE 602.) In addition, the witness’ reputation for dishonesty may be introduced or specific instances of conduct demonstrating a lack of veracity elicited on cross-examination. (MRE 608.) These methods of impeachment may be used against all witnesses, and their use against a witness-accused has not been challenged. However, when a defendant takes on the role of witness, the rule permitting impeachment of witnesses by prior conviction presents a special problem, i.e., the danger that evidence admitted to impeach the defendant-as-witness will be used by the jury in evaluating defendant-as-defendant.7

*568A jury should not be allowed to consider the defendant’s guilt of the crime before it on the basis of evidence of his propensity for crime. Finding a person guilty of a crime is not a pleasant or easy assignment for a representative group of twelve people. It is much easier to conclude that a person is bad than that he did something bad. Hence the appetite for more knowledge of the defendant’s background and the slippery slope toward general "bad man” evidence.8

*569This appetite presents three types of impropriety. First, that jurors may determine that although defendant’s guilt in the case before them is in doubt, he is a bad man and should therefore be punished. Second, the character evidence may lead the jury to lower the burden of proof against the defendant, since, even if the guilty verdict is incorrect, no "innocent” man will be forced to endure punishment. Third, the jury may determine that on the basis of his prior actions, the defendant has a propensity to commit crimes, and therefore he probably is guilty of the crime with which he is charged. Beaver & Marques, A proposal to modify the rule on criminal conviction impeachment, 58 Temple LQ 585, 592-593 (1985). All three of these dimensions suggest a likelihood that innocent persons may be convicted.

The danger then is that a jury will misuse prior conviction evidence by focusing on the défendant’s general bad character, rather than solely on his character for truthtelling.

B

Judge Weinstein has described the basis for MRE 609 admissions as follows:

The theory that all convictions are relevant to credibility depends on a two-fold assumption: (1) that a person with a criminal past has a bad general character and (2) that a person with a bad *570general character is the sort of person who would disregard the obligation to testify truthfully. [3 Weinstein, Evidence, ¶ 609[02] at 609-59. Emphasis added.]

Similarly, Judge Holmes stated in Gertz v Fitchburg R Co, 137 Mass 77, 78 (1884), that

when it is proved that a witness has been convicted of a crime, the only ground for disbelieving him which such proof affords is the general readiness to do evil which the conviction may be supposed to show. It is from that general disposition alone that the jury is asked to infer a readiness to lie in the particular case, and thence that he has lied in fact. The evidence has no tendency to prove that he was mistaken, but only that he has perjured himself, and it reaches that conclusion solely through the general proposition that he is of bad character and unworthy of credit.

The same analysis was offered in the Advisory Committee notes to the first draft of federal rule 609 which permitted impeachment by all prior felony convictions:

A demonstrated instance of willingness to engage in conduct in disregard of accepted patterns is translatable into willingness to give false testimony. [46 FRD 161, 297.]

Thus, jurors are not directly informed through impeachment evidence, with the exceptions described above, that a witness may not be truthful. That understanding is derived indirectly through the mediation of the belief that the witness has a "general readiness to do evil.” Gertz, supra, p 78. In other words, for example, the commission of an assault does not involve the telling of a lie.

The dissent’s reference to the diagnostic criteria *571of the antisocial personality disorder described in the Diagnostic and Statistical Manual of Mental Disorders (3d ed) (dsm iii k), (Washington, DC: American Psychiatric Association, 1987), further demonstrates this phenomenon. While the manual does not indicate that violent behavior is an example of a lack of veracity, it explains instead that the presence of a certain set of behaviors categorize an individual as antisocial, and that an antisocial person often is a liar, e.g., that an assaultive individual is a bad person or of bad general character and that such people have a tendency to lie.9

Most crimes can, therefore, be seen as evidence of a lack of veracity only when mediated through the belief that the individual has a bad general character. This results in both a low probative value and a strong potential for prejudice, as it will be difficult for jurors to put out of their minds the very step which allowed them to reach their conclusion as to veracity. On the other hand, crimes having an element of dishonesty or false statement are directly probative of a witness’ truthfulness and at the same time present little possibility for prejudice since they can be understood as such, absent the mediation of the conclusion that the witness-accused is of bad general character.

In spite of the likelihood of prejudice, it can be argued, and indeed the dissent does so eloquently, that a limiting instruction will suffice. As the United States Supreme Court stated in Delli Paoli v United States, 352 US 232, 242; 77 S Ct 294; 1 L Ed 2d 278 (1957):

Unless we proceed on the basis that the jury will *572follow the court’s instructions where those instructions are clear and the circumstances are such that the jury can reasonably be expected to follow them, the jury system makes little sense.

However, as the United States Supreme Court later stated in Bruton v United States, 391 US 123, 135; 88 S Ct 1620; 20 L Ed 2d 476 (1968), when overturning the holding in Delli Paoli:

[TJhere are some contexts in which the risk that the jury will not, or cannot, follow instructions is so great, and the consequences of failure so vital to the defendant, that the practical and human limitations of the jury system cannot be ignored. Compare Hopt v Utah, [120 US 430, 438; 7 S Ct 614; 30 L Ed 708 (1887)]; Throckmorton v Holt, 180 US 552, 567[; 21 S Ct 474; 45 L Ed 663 (1901)]; Mora v United States, 190 F2d 749 [CA 5, 1951]; Holt v United States, 94 F2d 90 [CA 10, 1937].[10]

*573We believe that we are now presented with such an exceptional circumstance. Limiting instructions are generally employed where more than one fact may be derived from a given piece of evidence, but not all are permissible considerations. This is not always easy for a jury to do, but we must sometimes rely on limiting instructions if our system is to function. However, as explained above, in the case of most prior conviction evidence the permissible consideration can only be understood by first recognizing the impermissible consideration. Where the two factors are so inextricably linked, we do not believe that a jury can be reasonably expected to follow the instruction.

We do not believe, however, that jurors are by any means deficient because they are unable to *574ignore general character evidence. To look to such evidence is natural, and many "experts” have done the same.

Our own Court of Appeals has at times been unable to resist doing so, and, indeed, it is sliding further along the slope to complete elimination of MRE 404 protection. In People v Jones, 98 Mich App 421, 428-429; 296 NW2d 268 (1980), the Court opined:

[Frequently the probative value of a conviction, or series of convictions, is increased, proportionally or greater, on the issue of credibility because of [the] similarity [to the present charge].
[I\t is more probable that a person has committed a crime if he has done it before, maybe several times. . . . When it comes to whom to believe, it should not benefit the defendant that he is a repeater, perhaps specializing in this kind of crime, or that his record is so bad that it will weigh heavily against him. After all, he committed the previous crimes and they tell a great deal about him and about whether he is lying now. [Emphasis added.]

A similar view of prior conviction evidence was expressed by Congressman Hogan in his dissent to the House Judiciary Committee’s version of the rule which would have allowed impeachment only by crimes involving dishonesty or false statement. He argued that "it would be misleading to permit the accused to appear as a witness of blameless life on those occasions when the accused chooses to take the stand.”11

The notion that a limiting instruction can prevent misuse of prior conviction evidence is simply mistaken, as it asks the jury to do what it cannot. *575This view is supported by a number of empirical studies. A Canadian study concluded that

present research leaves little doubt that knowledge of a previous conviction biases a case against the defendant. The likelihood that a jury will convict the defendant is significantly higher if the defendant’s record is made known to the jury. The fact that the defendant has a record permeates the entire discussion of the case, and appears to affect the juror’s perception and interpretation of the evidence in the case.
. . . These effects are present in spite of the fact that jurors were given speciñc instructions to ignore the fact of record while assessing the defendant’s guilt. [Hans & Doob, Section 12 of the Canada Evidence Act and the deliberations of simulated juries, 18 Crim L Q 235, 251 (1975-76). Emphasis added.]

Perhaps most disturbing was the study’s conclusion that

jurors used [prior conviction evidence] only minimally in considering the issue of credibility. [Id. at 247.]

Thus, the intended use of the evidence was of little actual importance.

Another study observed that the notion of a successful limiting instruction, with respect to character evidence, runs counter to the basic evaluative mechanisms discussed above.

The assumption that people can differentiate the use of information goes directly against what is known as the "halo” effect: the phenomenon by which a person will infer positive characteristics about a person where favourable information has been received and will infer negative characteristics about someone where unfavourable informa*576tion has been received. As Rosenberg and Olshan (1970) [citing Rosenberg & Olshan, Evaluative and descriptive aspects in personality perception, J of Personality and Soc Psyc, 16, 619-626 (1970)] have pointed out, "Studies using [the technique by which a subject is asked to infer characteristics about an individual on the basis of information about other characteristics] . . . have consistently shown that subjects infer favourable traits from one or more favourable stimulus traits and infer unfavourable traits from unfavourable stimulus traits” (p 619). Thus, if a person is told one negative thing about another person, he is going to assume other negative things. With respect to s. 12 of the Canada Evidence Act, it follows that we would expect that if a juror hears that an accused has one unfavourable characteristic (e.g., that he has a criminal record) he will then think of that accused as a generally bad person. If this same juror is then asked to judge the guilt of the accused, he will be likely to infer that the accused is guilty of the crime in question. [Doob & Kirshenbaum, Some empirical evidence on the effect of s. 12 of the Canada Evidence Act upon an accused, 15 Crim L Q 88, 89-90 (1972-1973).]

One of the researchers in the University of Chicago report leading to publication of The American Jury stated that the jury examinations revealed

"almost universal inability and/or unwillingness either to understand or follow the court’s instruction on the use of defendant’s prior criminal record for impeachment purposes. The jurors almost universally used defendant’s record to conclude that he was a bad man and hence was more likely than not guilty of the crime for which he was then standing trial.”
Letter from Dale W. Broeder, Associate Professor, the University of Nebraska College of Law, who conducted intensive jury interviews, to Yale *577Law Journal, dated March 14, 1960, on file in Yale Law Library. [Note, Other crimes evidence at trial: Of balancing and other matters, 70 Yale L J 763, 777, n 89 (1961).]

In a far less scientific study, the Columbia Journal of Law and Social Problems conducted a random national survey of trial judges and criminal defense attorneys. When asked whether they believed jurors were able to follow limiting instructions concerning the use of prior conviction evidence, ninety-eight percent of the responding attorneys answered negatively, and, even more disturbingly, forty-three percent of the responding judges agreed. Note, To take the stand or not to take the stand: The dilemma of the defendant with a criminal record, 4 Colum J of Law & Social Problems 215, 218 (1968).

The most recent study of this problem concluded that mock jurors used prior conviction evidence to "help them judge the likelihood that the defendant committed the crime charged” in spite of limiting instructions. Wissler & Saks, On the inefficacy of limiting instructions: When jurors use prior conviction evidence to decide on guilt, 9 Law & Human Behavior 37, 44 (1985).12 Most telling was the fact that a higher conviction rate was found where, all else being the same, the impeaching crime was murder than where the impeaching crime was perjury. Id. at 43. The only explanation for this last result is that the prior conviction evidence was not used exclusively to evaluate credibility. This is emphasized by the fact that the *578researchers found that there was no significant difference between the mock jurors’ ratings of defendant’s credibility when a prior conviction was introduced and when one was not. Id. at 41. They concluded that "[t]he credibility ratings of defendant did not vary as a function of prior conviction,” while "[cjonviction rates [did vary] as a function of prior conviction . . . .” Id.

In State v McAboy, 160 W Va 497; 236 SE2d 431 (1977), the Supreme Court of West Virginia barred impeachment of a defendant’s credibility by prior conviction except where the prior convictions were for perjury or false swearing as "[cjonviction of these crimes goes directly to the credibility of the defendant and . . . their relevancy has a priority over their prejudicial effect.” Id. at 508. Responding to the argument that we must rely on jurors’ ability to distinguish between impeachment of a defendant’s credibility as a witness and impeachment of his character in general, the court stated:

We are not unaware of the contention that such a rationale is pure sophistry and that the State, under the guise of testing credibility, relies upon the hope that the jury, in spite of being instructed by the court as to the purpose for which the evidence is admissible, will nonetheless consider the revelation of other convictions as indicating guilt in the case being tried. [Id. at 502 (quoting State v McGee, 160 W Va 1, 9; 230 SE2d 832 (1976).]

The McAboy court also pointed out that most prior convictions were of dubious relevancy to credibility, stating that the relevancy proposition "is based on the premise that all persons convicted of a crime are disposed not to tell the truth . . . .” Id. at 504. The court footnoted a hypothetical situation which bears repeating:

*579The ultimate paradox of the relevancy premise can be illustrated by altering slightly the example given in Bentham, Rationale of Judicial Evidence (Bowring’s ed. 1827), and quoted in 2 Wigmore, Evidence, § 519 at 610 (3rd ed, 1940). It concerns a man so proud of his truthfulness that he challenges a person who called him a liar to a duel. He wins the duel but is subsequently convicted of murder. Several years later, after release from prison, he is indicted on another crime. If he elects to testify, he knows that his propensity to lie will be evidenced by his prior conviction. [Id. at 504, n 6.]

Even if relevancy was assumed, the court continued, the chilling effect on defendant’s right to testify is overarching. In addition,

the presumption of innocence is impaired in the eyes of the jury, which conceives that his failure to testify, notwithstanding instructions to the contrary, is an indication of guilt. [Id. at 505.]

In light of the overwhelming tendency for jurors, and even trial and appellate judges to misuse prior conviction evidence, it is our view that there is an "overwhelming probability” that most prior conviction evidence introduced for the purpose of impeachment will be considered as if it had been introduced to show that the defendant acted in conformity with his criminal past. See Richardson v Marsh, 481 US —; 107 S Ct 1702; 95 L Ed 2d 176 (1987).

The dissenters criticize us for making behavioral judgments and relying on studies in reaching this determination. They critique at length the conclusions and methodologies of the studies we have cited. We do not agree with their conclusions, but welcome their critique of these materials as social science experiments cannot serve as the primary *580basis for judicial decision. Many fundamental principles of our jurisprudence are based on assumptions of human behavior that have never been, and in most cases cannot be, scientifically tested. We also note that the rule we modify today, and that the dissent urges us to leave as is, was the result of assumptions about jury behavior and the effectiveness of limiting instructions that were accompanied by relatively little analysis or study.

The studies cited are nevertheless relevant, support our view, and deserve consideration. In citing them, however, we do not suggest that the amendment of MRE 609 is based solely upon them. The rule change is instead based upon the underlying principles of our legal system and our perception that these principles are not served by the scope of prior-conviction impeachment which is presently permitted.

Some of the most cherished and well-accepted of these principles have been defined by jurists who were willing to state the obvious. The obvious in this area is that criminal defendants are prejudiced by prior conviction evidence as presently allowed and that leaving the question to trial judges has not been successful. As the cases on abeyance (see appendix b) demonstrate, the appellate burden resulting from the present scope of discretion is great; defendants are often impeached with crimes similar or identical to the ones with which they are charged, and in many cases these defendants choose not to testify rather than risk impeachment. Further, in most cases, the standard of appellate review prevents correction of error, and, in those cases where the error amounts to an abuse of discretion, the time and expense of a new trial is required.

We, therefore, in the words of the dissent, act not on the basis of studies, but on the "common*581sense premise” (post, p 678) that some prior convictions are more probative than others, that some are inherently more prejudicial, and that it is absurd to suggest that jurors will be able to avoid improper consideration of a defendant’s criminal character once it has become known to them. As will be explored in the following section, we are not alone in these conclusions.

hi

The foregoing discussion indicates that prior conviction impeachment of a witness-accused presents significant difficulties. Juries will often misuse this evidence and a limiting instruction is of dubious benefit. The amendment of MRE 609 which we adopt today creates a framework which responds to the realities just described and also takes into account the case law defining the purpose and practice of MRE 609 as well as that of the comparable federal rule. A brief overview of both standards will be helpful in determining which types of prior convictions should be excluded from evidence, which types should be included, and what factors should be employed in the exercise of trial court discretion in this area.

A

The traditional rule in the federal courts permitted impeachment of witnesses by prior felony convictions or by any convictions involving dishonesty (crimen falsi). This rule underwent its first major change in 1965 in the District of Columbia Circuit case Luck v United States, 121 US App DC 151; 348 F2d 763 (1965). In that case, the District of Columbia Circuit Court held that evidence of a conviction could not automatically be introduced *582to impeach a witness-accused. The relevant provision of the District of Columbia Code read in pertinent part:

No person shall be incompetent to testify, in either civil or criminal proceedings, by reason of his having been convicted of crime, but such fact may be given in evidence to affect his credit as a witness, either upon the cross-examination of the witness or by evidence aliunde; and the party cross-examining him shall not be concluded by his answers as to such matters. [DC Code Ann, § 14-305 (1961) (amended 1973). Emphasis added.]

The court held that the word "may” provided discretion not only to the prosecutor, but also to the trial court to exclude such evidence

where the trial judge believes the prejudicial effect of impeachment far outweighs the probative relevance of the prior conviction to the issue of credibility. [121 US App DC 156.]

In order to assist trial court judges, the circuit court defined a nonexhaustive list of factors relevant to this exercise of discretion:

the nature of the prior crimes, the length of the criminal record, the age and circumstances of the defendant, and, above all, the extent to which it is more important to the search for truth in a particular case for the jury to hear the defendant’s story than to know of a prior conviction. [Id. at 157.]

Two years later, the District of Columbia Circuit Court further explored this issue in Gordon v United States, 127 US App DC 343; 383 F2d 936 (1967), with Judge Warren Burger writing for the court. First, the court held that the burden of persuasion in such a matter is on the accused. *583Second, Judge Burger expanded upon the passage found in Luck, quoted above, describing the relevant factors:

In common human experience acts of deceit, fraud, cheating, or stealing, for example, are universally regarded as conduct which reflects adversely on a man’s honesty and integrity. Acts of violence on the other hand, which may result from a short temper, a combative nature, extreme provocation, or other causes, generally have little or no direct bearing on honesty and veracity. A "rule of thumb” thus should be that convictions which rest on dishonest conduct relate to credibility whereas those of violent or assaultive crimes generally do not; traffic violations, however serious, are in the same category. The nearness or remoteness of the prior conviction is also a factor of no small importance. Even one involving fraud or stealing, for example, if it occurred long before and has been followed by a legally blameless life, should generally be excluded on the ground of remoteness.
A special and even more difficult problem arises when the prior conviction is for the same or substantially the same conduct for which the accused is on trial. Where multiple convictions of various kinds can be shown, strong reasons arise for excluding those which are for the same crime because of the inevitable pressure on lay jurors to believe that "if he did it before he probably did so this time.” As a general guide, those convictions which are for the same crime should be admitted sparingly; one solution might well be that discretion be exercised to limit the impeachment by way of a similar crime to a single conviction and then only when the circumstances indicate strong reasons for disclosure, and where the conviction directly relates to veracity.
Of course, there are many other factors that may be relevant in deciding whether or not to exclude prior convictions in a particular case. See Luck, supra at 157, 348 F2d at 769. One important consideration is what the effect will be if the *584defendant does not testify out of fear of being prejudiced because of impeachment by prior convictions. Even though a judge might find that the prior convictions are relevant to credibility and the risk of prejudice to the defendant does not warrant their exclusion, he may nevertheless conclude that it is more important that the jury have the benefit of the defendant’s version of the case than to have the defendant remain silent out of fear of impeachment. [127 US App DC 347-348.]

The Gordon factors enumerated above thus include: the vintage of the relevant conviction, the need for defendant’s testimony, the similarity to the crime charged, and the nature of the prior offenses, with assaultive crimes and traffic violations being presumptively inadmissible and acts involving crimen falsi (including stealing) presumptively admissible. While not specifically enumerated as a factor in the balance, the Gordon court also stated that the defendant’s prior record had been received into evidence "because the case had narrowed to the credibility of two persons— the accused and his accuser — and in those circumstances there was greater, not less, compelling reason for exploring all avenues which would shed light on which of the two witnesses was to be believed.” Id. at 348. This statement would later prove to be a cause of confusion and contradiction in Michigan law. See post, pp 587-593.

Gordon also took note of the fundamental problem with this type of impeachment observing that its legitimate purpose

is, of course, not to show that the accused who takes the stand is a "bad” person but rather to show background facts which bear directly on whether jurors ought to believe him rather than other and conflicting witnesses. [Id. at 347.]

*585Finally, Judge Burger expressed his belief that the task of trial judges in this area would be "extremely difficult” and that

it would be much simpler if prior convictions of an accused were totally admissible or totally excludable as impeachment .... [Id. at 348.]

However, he felt that the language of the evidentiary statute (see ante, p 582) foreclosed a bright-line rule barring all such impeachment, while Luck foreclosed a return to the traditional rule allowing all prior-conviction impeachment.

The Luck-Gordon doctrine began to have an effect on other circuits and some states, including our own, in spite of the fact that Congress ultimately amended the relevant statute so as to read "shall be admitted,” rather than "may be admitted” (PL 91-358, § 133[a], DC Code Ann, § 14-305[b][l] [1973]), thus eliminating the doctrine in the jurisdiction in which it had been created.

Meanwhile, Congress was moving toward adopting a federal code of evidence.

The evolution of FRE 609(a) was somewhat tortured, with the text being substantially revised at virtually every level of consideration. For our purposes, it is sufficient to note that the rule ultimately adopted by Congress reads as follows:

(a) General rule. For the purpose of attacking the credibility of a witness, evidence that he has been convicted of a crime shall be admitted if elicited from him or established by public record during cross-examination but only if the crime (1) was punishable by death or imprisonment in excess of one year under the law under which he was convicted, and the court determines that the probative value of admitting this evidence outweighs its prejudicial effect to the defendant, or (2) in*586volved dishonesty or false statement, regardless of the punishment.

Contrary to MRE 609 and the Luck standard, FRE 609 balancing is not engaged in where the prior offense involved dishonesty or false statement and those words, supplying a bright-line rule for admissibility, were intended to be clearly related to credibility:

By the phrase "dishonesty and false statement” the Conference means crimes such as perjury or subornation of perjury, false statement, criminal fraud, embezzlement, or false pretense, or any other offense in the nature of crimen falsi, the commission of which involves some element of deceit, untruthfulness, or falsification bearing on the accused’s propensity to testify truthfully. [Conference Report, HR No 93-1597, reprinted at 120 Cong Rec, H 39939, 39941 (December 14, 1974).]

Since the federal Rules of Evidence have been codified, the federal courts have used those factors described in Gordon in determining the outcome of the balancing test for felonies. These factors have been repeatedly summarized as follows:

(1) The impeachment value of the prior crime.
(2) The point in time of the conviction and the witness’ subsequent history.
(3) The similarity between the past crime and the charged crime.
(4) The importance of the defendant’s testimony.
(5) The centrality of the credibility issue.

United States v Jackson, 696 F2d 578 (CA 8, 1982); United States v Preston, 608 F2d 626 (CA 5, 1979); United States v Mahone, 537 F2d 922, 929 (CA 7, 1976); United States v Paige, 464 F Supp 99 (ED *587Pa, 1978); 3 Weinstein, Evidence, ¶ 609[04] at 609-77 to 609-78.

B

It was not until the passage of 1861 PA 125 that a criminal defendant in Michigan could make a statement at his own trial. Even after the passage of that act, criminal defendants could not testify under oath. People v Thomas, 9 Mich 314 (1861). It took an additional twenty years before the Legislature, with the passage of 1881 PA 245, removed all obstacles to a criminal defendant testifying on his own behalf. In People v Cummins, 47 Mich 334, 336; 11 NW 184 (1882), decided shortly after the Legislature so acted, this Court held that when criminal defendants testify, they may be impeached by their prior convictions. Impeachment of witnesses, including criminal defendants, by introduction of prior convictions was, for approximately the next ninety years, allowed at the desire of the prosecutor. Finally, in 1974, this Court, in People v Jackson, 391 Mich 323; 217 NW2d 22 (1974), adopted the Luck rule of discretionary exclusion of such evidence.

Justice Swainson concurred in Jackson, but called for a bright line barring all impeachment of criminal defendants by prior conviction:

I think that it is now time for this Court to recognize a fact that has been well illustrated by commentators and confirmed by the facts in the present case. When the prospective witness is a criminal defendant with a prior record, there is often no real distinction between the common-law prohibition from testifying and our present rule. Rather than expose his prior record to the trier of fact the defendant must forego his right to testify in his own behalf. State v Santiago, 53 Hawaii 254, *588258; 492 P2d 657, 660 (1971); H. Kalven and H. Zeisel, The American Jury 146 (1966); McCormick, Law of Evidence (2d ed), § 43, p 84.
If we expect the trier of facts to intelligently and impartially decide if a defendant has committed the crime for which he is standing trial, it makes no sense to permit artificial barriers to remain in the fact-finding process. A total understanding of the events in question at an adversary trial cannot logically be achieved unless the defendant is able to freely testify as to his recollection of them. Conversely, if a defendant with a record does elect to testify, we ask the jury to perform unattainable feats of "mental gymnastics” by exposing it to a defendant’s record for the stated purpose of weighing his credibility and then asking the jury to totally disregard the criminal record when passing on the defendant’s guilt or innocence of the present charge. E. Griswold, The Long View, 51 ABAJ 1017, 1021 (1965); 70 Yale LJ 763, 777 (1961).
I find it undeniable at this time that the prejudice resulting from allowing impeachment of a defendant by prior conviction evidence clearly outweighs whatever small probative value might result from the practice. I believe that under our duty to supervise the administration of justice in the State of Michigan we would promote fairness and uniformity by setting forth a clear rule prohibiting impeachment of a criminal defendant by his prior convictions rather than leaving the decision to be made in each case at the trial court level. [Id. at 343-345.]

The majority, however, rejected the bright-line approach and instead referred to the Gordon factors regarding discretionary admission. However, the Court did not simply state verbatim the factors listed in Gordon. Instead, it stated:

Among the factors to be considered are the nature of the prior offense, whether it is for substantially the same conduct for which the accused *589is on trial, and the effect on the decisional process if the accused does not testify from fear of impeachment by prior convictions. [Id. at 333.]

Four years after Jackson, the Michigan rule, as originally proposed, was developed by a committee established by this Court to draft a code of evidence. By that time the federal rules were largely adopted and Michigan, like many other states at that time, looked to those rules for guidance. In addition, the committee saw Jackson as the basis for any rule in this difficult area.13

Four months after the Michigan Rules of Evidence were promulgated, the Court of Appeals decided People v Crawford, 83 Mich App 35; 268 NW2d 275 (1978), which, along with Jackson, has been consistently cited in MRE 609(a) cases. In that case, defendant was charged with six counts of armed robbery, and the trial court allowed him to be impeached by two prior armed robbery convictions. The Court of Appeals explained the balancing test as follows:

The factors which the judge must weigh in making his determination include: (1) the nature of the prior offense (did it involve an offense which directly bears on credibility, such as perjury?), (2) whether it is for substantially the same conduct for which the defendant is on trial (are the offenses so closely related that the danger that the jury will consider the defendant a "bad man” or infer that because he was previously convicted he likely committed this crime, and therefore create prejudice which outweighs the probative value on the issue of credibility?), and (3) the effect on the *590decisional process if the accused does not testify out of fear of impeachment by prior convictions (are there alternative means of presenting a defense which would not require the defendant’s testimony, i.e., can his side of the story be presented, or are there alternative, less prejudicial means of impeaching the defendant?). [83 Mich App 39.]

The Court held that the trial court erred by not excluding the evidence of prior convictions.14

In People v Kelly, 66 Mich App 634, 637; 239 NW2d 691 (1976), the Court of Appeals enunciated for the first time the Gordon factor not mentioned in either Jackson or MRE 609, but which the dissenters, in their analyses of the cases before us, would elevate to dispositive status:

A consideration not discussed in People v Jackson, supra, but central to Gordon v United States, supra, was the fact that there was a direct conflict between the testimony of the complainant and the defendant. The verdict in Gordon necessarily turned on how the jury resolved the credibility contest between the complainant and the defendant. . . .
This is precisely the situation in the instant case. There was a direct conflict in the testimony between the security guard who testified that he saw the defendant leave with a box of dishes and the defendant who testified that he carried trash out of the store. This conflict in testimony was as compelling as in Gordon for exploring all avenues which would shed light on the credibility of the two witnesses. The trial judge correctly denied the defendant’s motion to suppress his prior felony convictions.

*591This Court has never, until these cases, addressed this particular holding in Kelly.

In People v Hughes, 411 Mich 517, 520-521; 309 NW2d 525 (1981), we described the procedure under 609(a) as follows:

In People v Jackson, 391 Mich 323, 333; 217 NW2d 22 (1974), and People v Baldwin, 405 Mich 550, 552-553; 275 NW2d 253 (1979), we set forth guidelines for the exercise of the trial court’s discretion in balancing the prejudicial effect of evidence of prior convictions against their probative value on the issue of credibility. Among the factors to be considered are the nature of the prior offense, whether the conviction was for substantially the same conduct as that for which the accused is on trial, and the effect on the decisional process if the accused does not testify for fear of impeachment by prior convictions.1
The purpose of this inquiry is two-fold:
1) To put before the jury only those prior convictions indicative of the defendant’s disposition toward truthfulness and veracity; and
2) To keep from the jury those convictions which, although they may be indicative of defendant’s disposition toward truthfulness, may interfere with the jury’s ability to determine the defendant’s guilt or innocence on the basis of the evidence. Such interference is what is meant by "prejudice.” [Emphasis added.]

Not only is the Kelly credibility-contest factor not mentioned here, but the "purpose of the inquiry,” as described in the passage above, indicates that the process is a two-step process of elimination rather than simply a two-factor balancing; i.e., *592first, nonprobative convictions are excluded, and, second, of the probative offenses, those which would cause prejudice should also be excluded. This description does not seem to encompass a consideration of the need to compare the credibility of defendant with that of a prosecution witness.

Similarly, People v Woods, 416 Mich 581; 331 NW2d 707 (1982), does not address Kelly or the credibility-contest factor, citing only Luck and Gordon "as adopted by Jackson ” In addition, People v Wakeford, 418 Mich 95; 341 NW2d 68 (1983), this Court’s most recent statement on 609(a), states that "among” the factors are:

—The nature of the prior offense, whether similar or dissimilar to the offense charged, similarity standing as a factor weighing against admissibility, but not conclusive of the matter;
—Recency of the prior offense as probative of the defendant’s veracity at the time of trial, should he elect to testify;
—The importance to the truth-seeking process, not merely the prospects for acquittal, of obtaining the defendant’s version of the events in question. The notion here is that otherwise presumptively admissible prior convictions should be excluded if the price of a ruling of admissibility is to keep the defendant from the witness stand when the testimony he would offer would be central to the matters seriously controverted. [418 Mich 116-117. Emphasis in original.]

While this Court’s description of the third factor, particularly in light of the emphasized phrase, is opposite to the Kelly view, we did not at that time explicitly reject it, and in some decisions of the Court of Appeals that factor has continued to prove dispositive. In People v Jones, supra, the defendant was charged with larceny from a building and was impeached with two convictions of the *593same crime and one conviction of attempted larceny from a building. The Court of Appeals in that case extended the Kelly factor so that it would apply even where there was not a one-to-one credibility contest. There were two eyewitnesses who testified against defendant, but the panel stated that "[w]hen the defendant testifies, the trial often becomes a contest as to who is telling the truth.” 98 Mich App 428. See People v Monasterski, 105 Mich App 645; 307 NW2d 394 (1981), People v Casey, 120 Mich App 690; 327 NW2d 337 (1982), and People v Holmes, 132 Mich App 730; 349 NW2d 230 (1984), where the Kelly credibility-contest factor was also employed.

iv

As this historical survey indicates, our jurisprudence in this area has long recognized that certain crimes are more strongly related to truthfulness than others. Moreover, as discussed above, the relationship of the commission of a crime to veracity is often present only when seen through the prejudicial conclusion that the witness-accused is of bad general character. The amendment of MRE 609 addresses these realities by dividing the range of prior convictions into three categories.

As previously described, crimes having an element of dishonesty or false statement15 are directly *594probative of a witness’ truthfulness and can be understood as reflecting upon veracity by jurors without the mediation of their deciding that the defendant has a bad general character. Such convictions are of high probative value and possess little likelihood of prejudice. Therefore, the revised MRE 609 does not permit the exclusion of these convictions.16

*595No crimes, other than those including elements of false statement or dishonesty are directly probative of veracity. It may seem, therefore, that evidence of a prior conviction for any other crime should always be excluded. We are, however, mindful that theft offenses have traditionally been viewed as strongly probative of veracity.

In adopting our code of evidence this Court chose to treat theft crimes as it did crimes involving dishonesty or false statement, thus indicating a belief that theft crimes are more probative of veracity than other crimes. Similarly, as noted ante, p 583, Judge Burger, in Gordon v United States, stated:

In common human experience acts of deceit, fraud, cheating, or stealing, for example, are universally regarded as conduct which reflects adversely on a man’s honesty and integrity. Acts of violence on the other hand, which may result from a short temper, a combative nature, extreme provocation, or other causes, generally have little or no direct bearing on honesty and veracity. A "rule of thumb” thus should be that convictions which rest on dishonest conduct relate to credibility whereas those of violent or assaultive crimes generally do not.... [Emphasis added.]

Judge Burger’s analysis and MRE 609, as originally adopted, suggest that crimes having an element of theft should be treated in the same man*596ner as false statement crimes. At the same time, however, the fact that jurors will invariably use evidence of theft convictions to draw conclusions about defendant’s general character suggests that these convictions always be excluded. Rather than adopt one approach or the other, the amendment of MRE 609 leaves the admission of most theft crimes to the trial court’s discretion.17

We do not see countervailing grounds to exclusion where other crimes are involved. Those crimes do not bear an equally strong relationship to credibility. Moreover, their relationship to this trait is mediated through the determination that defendant has a bad general character and thus possess a strong potential for prejudice. Therefore, prior convictions for non-theft crimes which do not contain elements of dishonesty or false statement should never be admitted into evidence.

It could be argued that this three-part division is purely arbitrary. We do not agree. The division is based upon our understanding of which prior convictions are most probative of credibility and which are most prejudicial. It is, however, interesting to note in examining MRE 609 that most of the provisions are arbitrary. Certainly the dividing line between those crimes which are punishable by more than a year’s imprisonment and those which are not will not always divide those prior convictions which should or should not be admitted. Similarly, ten years does not finely divide those prior convictions that are probative from those that are not. Therefore, if there is an element of arbitrariness in the scheme we have defined, it is of a degree inherent in defining practicable evidentiary rules. While the dissent criticizes us for *597categorizing "classes of evidence” (post, p 671), excluding classes of relevant evidence is the very essence of defining evidentiary rules.18 We are also mindful of the appellate burden created by the present rule which the dissent points out "results in the greatest percentage of appeals . . . .” (Post, p 701.) The bright-line rules will facilitate trial court determinations and greatly reduce the number of appeals. In addition, an examination of other jurisdictions demonstrates that in this area the use of trial court discretion is generally circumscribed and that bright-line rules are used much more extensively than has been the case in Michigan.

The federal rule of evidence imposes a bright-line rule on the clearly probative end of the spectrum and a probative versus prejudice balancing for remaining crimes. Approximately thirteen states have adopted the federal approach.19 Five other states have drawn bright lines which permit automatic admission of evidence of all or nearly all prior convictions.20 In addition, eight jurisdic*598tions have bright lines permitting automatic impeachment by certain prior convictions,21 and three other states allow impeachment as long as the prior convictions bear some relevancy to credibility.22 At the same time, four states altogether bar the impeachment of criminal defendants by prior conviction,23 and one has barred such impeachment unless the prior conviction was for perjury or false swearing.24

There are, therefore, thirty-five jurisdictions (including the federal and the District of Columbia) which employ bright lines in addressing the impeachment problem. Of those thirty-five, twenty do not permit the probative/prejudice balance to ever be considered by the court.25 Only sixteen states *599grant their trial courts discretion over the admission of those prior convictions which may be admitted,26 and of those sixteen, five limit, to a greater degree than Michigan, those prior convictions which may even be considered for admission.27

While this survey demonstrates that there is a wide range of evidentiary rules governing the question of impeachment of a witness-accused, it also indicates that bright-line rules, to one extent or another, are employed in the majority of jurisdictions and that the extent of probative/prejudice discretion exercised by trial judges in our state stands at one extreme of the range of approaches now in practice.

The bright-line approach was also favored by the drafters of the 1953 Uniform Rules of Evidence and the Model Code of Evidence. The 1953 Uniform Rule 21 would have permitted impeachment of witnesses who were not criminal defendants only where the prior conviction involved dishonesty or false statement. Where the witness was a *600criminal defendant, it would have, with one exception, completely barred any such impeachment:

If the witness be the accused in a criminal proceeding, no evidence of his conviction of a crime shall be admissible for the sole purpose of impairing his credibility unless he has first introduced evidence admissible solely for the purpose of supporting his credibility.

The Model Code Rule 106 states in pertinent part:

If an accused who testifies at the trial introduces no evidence for the sole purpose of supporting his credibility, no evidence concerning his commission or conviction of crime shall, for the sole purpose of impairing his credibility, be elicited on his cross-examination or be otherwise introduced against him ....

V

Having left the admission of theft convictions to the judge’s discretion, it is necessary to clarify the manner in which this discretion should be exercised. At the present time there is a split in the Court of Appeals as to what factors are properly considered in the balancing test conducted under MRE 609 prior to the amendment adopted today.28

In light of this split in the Court of Appeals and the contradictory nature of some of the factors employed in the past, we define a more specific probative versus prejudice balancing test consistent with what we feel was the original intent of the existing MRE 609, and our prior decisions on this subject. It will be used to resolve the cases at *601hand, as well as to serve as the balancing test pursuant to the amended MRE 609.

In enumerating his criteria in Gordon, Judge Burger observed that "[o]ne important consideration is what the effect will be if the defendant does not testify out of fear of being prejudiced because of impeachment by prior convictions.” Id. at 347. The subsequent application of the Gordon guidelines makes clear that the credibility-contest factor contradicts this consideration since the defendant’s testimony is never more crucial, a failure to testify more damaging, and the decision to admit prior convictions more prejudicial than when "the case . . . narrow[s] to the credibility of two persons— the accused and his accuser . . . .” Id. at 348. This contradiction lends great credence to Judge Burger’s prophecy that employing the Luck-Gordon rule was "extremely difficult.” Id. Indeed, as one writer has observed, "When one [of these two factors] increases in importance, the other does also, and there appears to be no principled way to determine which factor should prevail.” Surratt, Prior-conviction impeachment under the federal rules of evidence: A suggested approach to applying the "balancing” provision of Rule 609(a), 31 Syracuse L R 907, 945 (1980).

The dissenters’ focus on the credibility-contest factor in two of the cases before us, Allen and Gray, without any balancing of the "important” and countervailing standard of the "effect on the decisional process if the accused does not testify from fear of impeachment by prior convictions,” Jackson, supra at 333, is the latest example of the fluidity and unpredictability of attempts to apply our current rule to these questions. The dissenters would not only employ a factor that has not been previously utilized by this Court or in the oft-cited Court of Appeals Crawford case, but would do so in *602such a manner as to strip the probative/prejudice balance of any real meaning. The dissenters discard the most-cited balancing test and the only test for judicial discretion actually set forth in MRE 609 — probativeness versus prejudice "on the issue of credibility.” It has not heretofore been suggested that the "probative value of admitting this evidence” escalated with its need or that "its prejudicial effect” decreased with its need.

In the remaining three cases before us, the dissenters would find error, again, not based principally on an analysis of the "probative value of admitting this evidence on the issue of credibility” and "its prejudicial effect,” but rather on the need or lack thereof for evaluating the defendants’ credibility.29

It is our view that it is the effect on the decisional process if the defendant does not testify which must predominate and so the contradicting "credibility contest” factor must therefore be eliminated. We note, initially, that both the first federal case and the first Michigan case dealing with the prior impeachment question failed to cite the credibility-contest factor. Luck, supra at 157, instead cited "above all, the extent to which it is more important to the search for truth in a partic*603ular case for the jury to hear the defendant’s story than to know of a prior conviction.” (Emphasis added.) Similarly, Jackson, supra at 333, cited "the eifect on the decisional process if the accused does not testify from fear of impeachment by prior convictions.”

We also acknowledge that when a criminal defendant testifies jurors are quite aware that he has a unique concern with the outcome of the trial and is more likely to have fabricated his testimony than any other witness. His testimony is therefore likely to be given diminished weight irrespective of impeachment. Note, Procedural protections of the criminal defendant — A reevaluation of the privilege against self-incrimination and the rule excluding evidence of propensity to commit crime, 78 Harv LR 426, 440, 450 (1964); 9 Law & Human Behavior, supra at 43.

In addition, as pointed out in Crawford, excluding evidence of prior convictions does not mean that a defendant’s testimony will reach the jury unassailed. The prosecutor on cross-examination may draw out contradictions or indications of unreliability in defendant’s testimony. The prosecutor may also bring forward prior inconsistent statements or challenge the defendant’s sensory capabilities. Assuming, as the dissent seems to do, that the defendant is almost certainly guilty, and thus fabricating his story, it is quite likely that defendant’s story will not hold up under these less prejudicial forms of impeachment.

Finally, unless one were to take the position that a defendant’s testimony is not relevant to truth finding, in which case there is little need for criminal trials at all, we cannot ignore the burden placed upon truth finding when a defendant chooses not to testify out of fear of prior conviction impeachment. Only the dissenters’ assumption *604that "ninety percent of those charged are guilty”30 *605(post, p 672, n 6) explains their failure to apply the same rigorous analysis to this burden that they apply to the burden of barring impeachment by prior conviction.

In sum, the trial judge’s first task, under the amended MRE 609, will be to determine whether the crime contains elements of dishonesty or false statement. If so, it would be admitted without further consideration. If not, then the judge must determine whether the crime contains an element of theft. If it is not a theft crime, then it is to be excluded from evidence without further consideration. If it is a theft crime and it is punishable by more than one year’s imprisonment,31 the trial *606judge would exercise his discretion in determining the admissibility of the evidence by examining the degree of probativeness and prejudice inherent in the admission of the prior conviction. For purposes of the probativeness side of the equation, only an objective analysis of the degree to which the crime is indicative of veracity and the vintage32 of the conviction would be considered,33 not either party’s need for the evidence. For purposes of the prejudice factor, only the similarity to the charged offense and the importance of the defendant’s testimony to the decisional process would be considered. The prejudice factor would, of course, escalate with increased similarity and increased importance of the testimony to the decisional process. Finally, unless the probativeness outweighs the prejudice, the prior conviction would be inadmissible.

vi

The bright-line aspects of the amendment of MRE 609 will, with the exception of the execution of the balancing test, apply to witnesses called by the prosecution and nondefendant witnesses called by the defense in the same way they apply to *607defendants. We recognize that symmetry is "neither an object of criminal procedure nor a proper criterion of fairness.” People v Hayes, 410 Mich 422, 425; 301 NW2d 828 (1981). However, since we have eliminated the credibility contest factor in the balancing test and barred the introduction of most prior convictions against the defendant, it would skew the decisional process if prosecution witnesses, including the complaining witness, were subject to the level of impeachment permitted prior to today’s amendment. Similarly, since the new rule will apply to all prosecution witnesses, it should, in fairness, also apply to all defense witnesses.

Unfortunately, perfect symmetry cannot be attained. While the bright lines can be easily applied to nondefendant witnesses,34 the application of the balancing test for theft crimes poses a problem as the factors relevant to prejudice do not arise where the witness is not the defendant. No one is prejudiced where the offense used to impeach a nondefendant witness is similar to the offense with which defendant is charged. Similarly, since, in general, the only individual who can refuse to testify is the defendant, the specter of impeachment cannot result in the loss of a nondefendant’s testimony. The "effect on the decisional process” factor is, therefore, irrelevant. The dissent would argue that prejudice also lies in the possibility that a prosecution witness will be impeached for a theft crime while the defendant is not.

We agree that this represents an imbalance, but do not consider it to be prejudice as we have described it. Assymetry has long been present in *608this area of the law. For example, under the federal rule a prosecution witness can never escape impeachment with a timely felony or false statement conviction.35 Moreover, today’s amendment does not permit consideration of possible prejudice to defendant if a nondefendant defense witness is to be impeached. As with impeachment of prosecution witnesses, the only "prejudice” arises out of the fact that the witness may not be believed, as the nondefendant witness’ general character will not be considered by the jury.

Therefore, where a party seeks to impeach a nonaccused witness, the bright lines will apply. Where the relevant offense is a theft crime, the judge need only determine that the prior offense, in light of its nature and vintage, is significantly probative36 of veracity.37 If so, the impeachment evidence should be admitted. If not, it should be excluded.

VII

The bright-line test substantially revises our approach to impeachment by prior conviction. Trial judges have, in good faith, attempted to follow the framework described in MRE 609 prior to the amendment we now announce. Thus, parts *609of the amendment of MRE 609 are "clear breaks” in our jurisprudence, and we will apply them only prospectively. See, e.g., Tebo v Havlik, 418 Mich 350; 343 NW2d 181 (1984). Accordingly, the amendment of MRE 609 will take effect March 1, 1988. Trials begun before that date will be governed by the existing version of MRE 609 as interpreted by this opinion.

We do not believe, however, that our clarification of the balancing test represents a "clear break” in our jurisprudence. The balancing test, as such, has long been in place, and, as described above, there has been a split in the Court of Appeals regarding its proper exercise. Resolution of the many cases now on appeal requires clarification of that test. Therefore, the clarified balancing test, as described in this opinion and in the amendment of MRE 609, will be given limited retroactivity as follows: 1) to the instant cases; 2) all cases pending on initial and direct appeal in which the issue of impeachment by prior conviction under the then-existing MRE 609(a) has been raised and preserved;38 3) currently pending appeals in which the appellant’s initial brief has not yet been filed, but the issue was raised and preserved in the trial court; and 4) cases in which the issue has been raised and preserved in the trial court, but no first appeal from the original judgment has yet been filed.

In conclusion, the bright-line tests will become effective March 1, 1988 and will apply to cases begun on that date and thereafter. The balancing test, although part of the amendment of MRE 609, should also be understood as a clarification of the proper procedure pursuant to MRE 609 prior to its *610amendment.39 It will, therefore, be given the limited retroactive application just described.40

We therefore review the instant cases in the context of the clarified balancing test, but not in the context of the bright-line rules.

In Gray, defendant was convicted of first-degree felony murder and felony-firearm. The trial court ruled that he could be impeached with three prior concealed weapons convictions, as well as with a conviction for possession of heroin. The convictions occurred from five to nine years prior to the instant trial and all were for possessory crimes which have a low degree of probativeness. At the same time, the concealed weapon convictions were similar to the felony-firearm charge, and Gray’s own testimony was the only evidence presented in the defense’s case. His testimony was therefore very important to the decisional process. In light of the low probative value of the evidence and its very high level of prejudice, we find that the prejudice outweighed the probativeness of the evidence of prior convictions. Its admission was, therefore, error, and we reverse the conviction and remand for a new trial.

In Pedrin, defendant was, in 1982, tried and convicted of breaking and entering an unoccupied building with the intent to unlawfully drive away an automobile. The trial court ruled that defendant could be impeached with a prior 1981 conviction for breaking and entering with intent to commit larceny. Such a theft crime is moderately *611probative of veracity, and the recentness of the crime accents that probative value. The level of prejudice was, however, significantly greater. The charged offense was very similar to the prior conviction. In addition, defendant’s testimony was very important to the decisional process, as he had no other means of presenting his version of events. Accordingly, we reverse and remand for a new trial.

In Allen, defendant was tried and convicted of first-degree criminal sexual conduct. He was impeached with a conviction for second-degree criminal sexual conduct which occurred less than three years before the case in question. While the conviction was of recent vintage, the probative value of assaultive crimes is low. As to the level of prejudice, the prior crime and the charged offense were extremely similar. In addition, although the defense presented a witness who testified as to the defense claim of prior consensual sexual activity, the only evidence as to defendant’s version of events regarding the charged incident was defendant’s own testimony. We therefore reverse and remand for a new trial.

In Brooks, defendant was, in 1981, convicted of armed robbery and felony-firearm. He was impeached with two 1975 armed robbery convictions and a 1981 unarmed robbery conviction. Since robbery, although it contains an element of theft, is primarily an assaultive crime (People v Wakeford, 418 Mich 95; 341 NW2d 68 [1983]), and does not involve stealth, it has a lower probative value on the issue of credibility than would other theft crimes. At the same time, the recentness of the last conviction heightens its probative value. However, the similarity of the prior convictions to the charged crime was highly prejudicial, and defendant was the only witness presented by the de*612fense. The evidence of prior convictions should therefore have been excluded. Nevertheless, for the reasons described in Chief Justice Riley’s opinion, we find that the error was harmless. We also agree with the Chief Justice that it was not error for the trial court to fail to give a limiting instruction sua sponte on the use of prior convictions. We therefore affirm the conviction.

In Smith, defendant was convicted of two counts of first-degree criminal sexual conduct and one count of assault with intent to murder. The trial court ruled that defendant could be impeached with an eight-year-old manslaughter conviction. Manslaughter, an assaultive offense, is of low probative value, and the conviction was of relatively late vintage. The level of prejudice was high, since manslaughter is similar to the charge of assault with intent to murder. In addition, although the defense presented three witnesses whose testimony lent some credence to defendant’s theory of the case, only defendant’s own testimony could have put his version of events into evidence, since his codefendants testified against him. The high level of prejudice and low level of probativeness lead us to conclude that the trial court’s ruling that the evidence of prior convictions could be admitted was erroneous. Nevertheless, the prosecutor’s case was so strong that we do not believe a reasonable juror could have voted to acquit defendant, even if hé had testified and not been impeached. We therefore find that the error was harmless. Defendant also raises four other issues which are described and carefully examined in Chief Justice Riley’s opinion. We agree with the Chief Justice’s analyses of these issues, and thus we affirm the conviction.41

*613VIII

Because the Court is unable to achieve a majority position on the merits of the rule announced in Luce v United States, 469 US 38; 105 S Ct 460; 83 L Ed 2d 443 (1984), we decline to address its proposed applicability in Michigan, reserving that question for another day.

Levin, Cavanagh, and Archer, JJ., concurred with Brickley, J.

*614APPENDIX A

REVISED MRE 609

IMPEACHMENT BY EVIDENCE OF CONVICTION OF CRIME

(a) General rule. For the purpose of attacking the credibility of a witness, evidence that the witness has been convicted of a crime shall not be admitted unless the evidence has been elicited from the witness or established by public record during cross-examination, and

(1) the crime contained an element of dishonesty or false statement, or

(2) the crime contained an element of theft, and

(A) the crime was punishable by imprisonment in excess of one year or death under the law under which the witness was convicted, and

(B) the court determines that the evidence has significant probative value on the issue of credibility and, if the witness is the defendant in a criminal trial, the court further determines that the probative value of the evidence outweighs its prejudicial effect.

(b) Determining probative value and prejudicial effect. For purposes of the probative value determination required by subrule (a)(2)(B), the court shall consider only the age of the conviction and the degree to which a conviction of the crime is indicative of veracity. If a determination of prejudicial effect is required, the court shall consider only the conviction’s similarity to the charged offense and the possible effects on the decisional process if admitting the evidence causes the defendant to elect not to testify. The court must articulate, on the record, the analysis of each factor.

(c) -(f) [(b)-(e) redesignated but otherwise unchanged.]

*615APPENDIX B

SUMMARY OF ABEYANCES FOR PEOPLE VALLEN

Abeyance Summary Abbreviation Key

A Assault

A&B Assault and battery

AR Armed robbery

Att Attempted

AWI Assault with intent, plus suffixes for:

GBH Great bodily harm

M Murder

RA Robbery while armed

RU Robbery while not armed

B&E Breaking and entering

CCW Carrying a concealed weapon

CSC Criminal sexual conduct, plus

suffixes I-III

Entry w/o B Entry without breaking

FA Felonious assault

FF Felony firearm

FPF Felon in possession of firearm

GL Grand larceny

HO Habitual offender, plus suffixes I-IV

K Kidnapping

L Larceny

LIB Larceny in a building

LIMV Larceny in a motor vehicle

LP Larceny from a person

$100L Larceny over $100

$100MD Malicious destruction over $100

Ml First-degree murder

M1[F] Felony first-degree murder

M2 Second-degree murder

NSF Non-sufficient funds

Poss cocaine Possession of cocaine

R&C Receiving and concealing

SA Simple assault

U&P Uttering and publishing

UR Unarmed robbery

Example: "AR (x3)” — defendant had three armed robbery convictions.

*616Abbreviations pertinent to defendants’ testimony:

T Testified

DNT Did not testify

*617[[Image here]]

*618[[Image here]]

*619[[Image here]]

*620[[Image here]]

*621[[Image here]]

MRE 609 reads in pertinent part:

(a) General rule. For the purpose of attacking the credibility of a witness, evidence that he has been convicted of a crime *565shall be admitted if elicited from him or established by public record during cross-examination but only if
(1) the crime was punishable by death or imprisonment in excess of one year under the law under which he was convicted, or the crime involved theft, dishonesty or false statement, regardless of the punishment, and
(2) the court determines that the probative value of admitting this evidence on the issue of credibility outweighs its prejudicial effect and articulates on the record the factors considered in making the determination.

According to the recent report of the Michigan Rules of Evidence Committee, "[w]hen the federal rules were being considered by Congress, Rule 609 drew the most heated comments; and the Michigan Committee, meeting in 1976, spent more time on this rule than on any other.” Report of the Committee on the Rules of Evidence, December 23,1983, p 17.

See, e.g., note, Other crimes evidence at trial: Of balancing and other matters, 70 Yale L J 763 (1961); note, Procedural protections of the criminal defendant — A reevaluation of the privilege against self-incrimination and the rule excluding evidence of propensity to commit crime, 78 Harv L R 426 (1964); comment, Use of bad character and prior convictions to impeach a defendant-witness, 34 Fordham L R 107 (1965-66); note, To take the stand or not to take the stand: The dilemma of the defendant with a criminal record, 4 Colum J of L & Social Problems 215 (1968); Glick, Impeachment by prior convictions: A critique of Rule 6-09 of the proposed rules of evidence for US district courts, 6 Crim L Bull 330 (1970); Spector, Impeaching the defendant by his prior convictions and the proposed federal rules of evidence: A half step forward and three steps backward, 1 Loyola Univ L J 247 (1970); Krauser, The use of prior convictions as credibility evidence: A proposal for Pennsylvania, 46 Temple L Q 291 (1972); note, An eclectic approach to impeachment by prior convictions, 5 J of L Reform 522 (1972); Spector, Rule 609: A last plea for its withdrawal, 32 Okla L R 334 (1979); Nichol, Prior crime impeachment of criminal defendants: A constitutional analysis of Rule 609, 82 W Va LR 391 (1980); Surratt, Prior-conviction impeachment under the federal rules of evidence: A suggested approach to applying the "balancing” provision of Rule 609(a), 31 Syracuse LR 907 (1980); Beaver & Marques, A proposal to modify the rule on criminal conviction impeachment, 58 Temple L Q 585 (1985).

Much of the dissent’s argument appears to rest on the erroneous assumption that the relevancy of MRE 404 protections disappears at the close of the prosecution’s case in chief and that therefore there is no practical contradiction between the purposes of MRE 404 and MRE 609. We find this approach flawed since the language of MRE 404 explicitly lists MRE 609 as an exception.

This fundamental principle is the basis for MRE 404 which provides in part:

(a) Character evidence generally. Evidence of a person’s character or a trait of his character is not admissible for the purpose of proving that he acted in conformity therewith on a particular occasion, except:
(4) Character of witness. Evidence of the character of a witness, as provided in Rules 607, 608, and 609.
(b) Other crimes, wrongs, or acts. Evidence of other crimes, wrongs, or acts is not admissible to prove the character of a person in order to show that he acted in conformity therewith.
It may, however, be admissible for other purposes, such as proof of motive, opportunity, intent, preparation, scheme, plan, or system in doing an act, knowledge, identity,' or absence of mistake or accident when the same is material, whether such other crime, wrongs, or acts are contemporaneous with, or prior or subsequent to the crime charged.

This rule bars a prosecutor, with specific and narrow exceptions, from introducing evidence of an accused’s prior convictions in order to show that he acted in conformity with such prior activities.

This is in marked contrast to civil law jurisdictions where the prosecutor, as part of his case in chief may introduce evidence of the accused’s character, past activities, and criminal convictions. Comment, Use of bad character and prior convictions to impeach a defendant-witness, 34 Fordham L R 107, 108-109 (1965-66).

The dissent seems to suggest that a defendant who does not have a criminal record will be prejudiced by today’s decision and the adoption of the amendment of MRE 609. We do not agree. It is unreasonable to suggest that typical jurors will be sufficiently versed in the law to know that a witness-accused may have a criminal past which is not raised due to a rule of evidence. Even if this were to occur in an isolated case, it would not raise new problems. Any juror who is so informed as to the revised rule that he may suspect the bona lides of an apparently crime-free history would certainly do the same under the present rule, as it too results in the exclusion of *568evidence of some prior convictions. In addition, the suggestion that we must evaluate rules of exclusion on the grounds that jurors might suspect exclusion where none has occurred is, in fact, an argument for never allowing exclusion of evidence.

The American Jury by Kalven and Zeisel (Boston: Little, Brown & Co, 1966), involved a study of jury behavior and indicates that the introduction of a defendant’s prior convictions substantially increases the conviction rate. At one point, the authors divided defendants into two groups. The first group was composed of those defendants who took the stand and either had no record or were able to keep it from the jury as well as those defendants who did not take the stand but of whom the jury learned that they had no record. The second group was comprised of all other defendants, i.e., those who the jury learned had a criminal record and those who did not take the stand and of whom the jury received no information as to their record.

The acquittal rate of the first group was forty-two percent while that of the second group was twenty-five percent. In trial settings where the evidence suggested the highest probability of acquittal, the acquittal rate of the first group was sixty-five percent and that of the second group, thirty-eight percent.

While the study did not determine what percentage of those choosing not to take the stand did so out of fear of impeachment by prior conviction, it did conclude that defendants without records testify in ninety-one percent of the cases, those with records testify in only seventy-four percent. Kalven & Zeisel, supra at 146. Where the case is clear for acquittal, only fifty-three percent of defendants with records testify, apparently not wishing to risk their high chance of acquittal on impeachment. Id. If it were true that impeachment went only to defendants’ credibility, there would be little reason for defendants to shy away from testifying. At worst, their testimony would be disbelieved. It would appear, however, that the fact that the risks created by impeachment go far beyond its legitimate purpose is not lost on defendants.

The dissent ascribes the higher conviction rate to the use of prior conviction evidence in determining credibility. A recent study addressed that question and found that "prior conviction evidence does not have its impact on verdicts by way of an intervening impact on perceptions of credibility” and that mock jurors "were willing to state *569that the prior conviction evidence increased the likelihood of the defendants’ guilt and was the reason they found him guilty, even though they had been instructed not to use the information for that purpose.” Wissler & Saks, On the inefficacy of limiting instructions: When jurors use prior conviction evidence to decide on guilt, 9 Law & Human Behavior 34, 44 (1985). Thus, juries exposed to prior conviction evidence may decide on a defendant’s guilt upon the basis of the inference that prior criminal activity indicates guilt of a charged crime. What the dissent views as contradictory results in this study are in fact consistent. See, id.

We also note that at least one commentator takes issue with the psychological assumptions upon which even this notion is based. Spector, Rule 609: A last plea for its withdrawal, 32 Okla L R 334 (1979).

The dissenters cite Bruton, supra, Tennessee v Street, 471 US 409; 105 S Ct 2078; 85 L Ed 2d 425 (1985), and Richardson v Marsh, 481 US —; 107 S Ct 1702; 95 L Ed 2d 176 (1987), as the basis for the analysis described in section n<2), post, p 667 of the dissent. Initially, we note that the three-step "proper approach” (post, p 667) has never been so defined by the United States Supreme Court, nor was it employed in the determination of the federal or Michigan Buies of Evidence. Moreover, in Street, the defendant claimed that his confession was coerced and that it was based upon his accomplice’s statement. The fact that his confession contained details of the crime not described in the codefendant’s statement was extremely probative of the inaccuracy of defendant’s explanation of his confession. Indeed, it is difficult to imagine more probative evidence of such a claim.

We do not view prior conviction evidence in the same light. It does not, as did the evidence in Street, go directly to the heart of defendant’s explanation. Prior conviction evidence, on the other hand, is, in general, probative of the truthfulness of a defendant’s claim only when mediated through the notion that defendant is of general bad character. (See above.) Consistent with this is the fact that prior conviction evidence, unlike the evidence in Street, may provide a jury with an independent and improper basis for a guilty verdict. The admission of Street’s codefendant’s statement did not serve to convince the jury that defendant was a bad man deserving of imprisonment regardless of his guilt in the tried case, or that the burden of proof would be lowered, or that defendant was the type of man who would have likely committed the charged offense. The evidence in *573Street provided little, if any, inculpatory information to the jury that was not already before- them in defendant’s own confession. The degree of prejudice was, as was the degree of probativeness, of a different magnitude than in the cases before us today which involve a very different type of evidence.

The dissenters also cite Spencer v Texas, 385 US 554; 87 S Ct 648; 17 L Ed 2d 606 (1967). In that case, the United States Supreme Court held that introducing a defendant’s prior convictions did not violate the Due Process Clause of the United States Constitution. However, our decision today is not based on constitutional grounds, but rather on policy. Policy, not constitutional limitations, underlies most of Michigan’s evidentiary code, and the determination of that policy has been explicitly left to this Court by art 6, § 5, of the 1963 Constitution. Moreover, the Spencer Court based its decision in part upon its belief that

[i]t would be a wholly unjustifiable encroachment by this Court upon the constitutional power of States to promulgate their own rules of evidence to try their own state-created crimes in their own state courts, so long as their rules are not prohibited by any provision of the United States Constitution .... [385 US 568-569.]

The rule which we today adopt is not, in our view, prohibited by any provision of the United States Constitution. It is also worth noting that four justices in Spencer believed that the procedure in that case was unconstitutional, and a fifth justice, while agreeing with the majority that there was no constitutional violation, viewed the procedure as bad policy.

120 Cong Rec, H 1414 (January 30, 1974).

The dissent incorrectly states (post, p 685) that this study concluded that conviction was more likely only when the prior conviction was for a crime similar to the one charged. To the contrary, while similar crimes were the most prejudicial, the study resulted in a finding of jury bias where the prior conviction was for a dissimilar offense as well.

Professor John Reed’s memorandum of November 11, 1975, to the evidence committee summarizing its initial rule 609(a) activities stated that the proposed rule as to a witness-accused "was in accord with People v Jackson, supra, in directing the trial court to make a determination that probative value outweighs prejudicial effect . . . .”

While Crawford was initiated prior to the enactment of the Rules of Evidence, the Court of Appeals noted that "[a] similar analysis is appropriate under new MRE 609 and the result in a case such as this would be identical.” 83 Mich App 40, n 1.

The trial in this case preceded the March 1, 1978 effective date of the Michigan Rules of Evidence. The listed factors continue to be relevant in the balancing required by MRE 609(a)(2).

While a small number of federal courts have interpreted the phrase "involving dishonesty or false statement” broadly, e.g., United States v Del Toro Soto, 676 F2d 13 (CA 1, 1982), most have adhered to the House Conference Report’s explanation that

by the phrase "dishonesty and false statement” the Conference means crimes such as perjury or subornation of perjury, false statement, criminal fraud, embezzlement, or false pretense, or any other offense in the nature of crimen falsi, the commission of which involves some element of deceit, untruthfulness, or falsification bearing on the accused’s propensity to testify truth*594fully. [Conference Report, HR No 93-1597, reprinted at 120 Cong Rec, H 39939, 39941 (December 14, 1974). Emphasis added.]

See, e.g., United States v Lewis, 200 US App DC 76, 82; 626 F2d 942 (1980); United States v Donoho, 575 F2d 718 (CA 9, 1978).

In order to avoid any confusion in our courts, we have drafted the new rule so as to make clear that prior convictions are only to be admitted where dishonesty or false statement is an actual element of the offense in question. This is consistent with the fact that juries do not reach decisions as to how a crime was committed unless such is an element of the offense, e.g., larceny by false pretenses. In this regard, it is also worth noting that the committee which drafted the proposed Michigan Rules of Evidence "agreed that the Comments [to Rule 609] will urge strict construction of the phrase 'dishonesty or false statement ....’” Memorandum of John Reed to the Michigan Rules of Evidence Committee (November 11, 1975).

We also approve of the view expressed in the House Conference Report and nearly all federal cases that "dishonesty” is not demonstrated by a mere willingness to engage in criminal conduct. The term refers specifically to lying, deceit, misrepresentation or a lack of veracity. See United States v Ashley, 569 F2d 975 (CA 5, 1978); United States v Fearwell, 193 US App DC 386; 595 F2d 771 (1978); Government of Virgin Islands v Toto, 529 F2d 278 (CA 3, 1976); United States v Smith, 179 US App DC 162; 551 F2d 346 (1976); United States v Barnes, 622 F2d 107 (CA 5, 1980); United States v Entrekin, 624 F2d 597 (CA 5, 1980); United States v Yeo, 739 F2d 385 (CA 8, 1984); United States v Glen, 667 F2d 1269 (CA 9, 1982); United States v Lane, 708 F2d 1394 (CA 9, 1983); United States v Mansaw, 714 F2d 785 (CA 8, 1983); United States v Bay, 762 F2d 1314 (CA 9, 1984); United States v McClintock, 748 F2d 1278 (CA 9, 1984); United States v Hans, 738 F2d 88 (CA 3, 1984).

This bright-line rule is, in essence, based upon our view that impeachment through reference to crimes for which false statement or dishonesty is an element is inherently more probative than prejudicial. Therefore, defendants who are so impeached may not claim that such impeachment violates MRE 403. That rule requires that the probative value of the evidence be "substantially outweighed” by prejudice. Since we find that as a matter of law prior convictions of crimes involving dishonesty or false statement are more probative *595than prejudicial, it obviously cannot be argued that the probative value is "substantially outweighed” by prejudice.

This view has been accepted in every federal Court of Appeals which has addressed the question, and six federal circuits now bar exclusion of prior convictions involving dishonesty or false statement on FRE 403 grounds. United States v Kuecker, 740 F2d 496 (CA 7, 1984); United States v Wong, 703 F2d 65 (CA 3, 1983); United States v Kiendra, 663 F2d 65 (CA 1, 1981); United States v Leyva, 659 F2d 118 (CA 9, 1981); United States v Coats, 209 US App DC 205; 652 F2d 1002 (1981); United States v Toney, 615 F2d 277 (CA 5, 1980).

Where a theft crime includes an element of dishonesty or false statement, e.g., larceny by false pretenses, it will be treated as an automatically admissible prior offense.

The exclusion of classes of relevant evidence is the effect of many evidentiary rules including MRE 404, as well as the rules excluding: evidence of subsequent remedial measures (MRE 407); evidence of withdrawn guilty or nolo contendere pleas (MRE 410); evidence of the existence of liability insurance (MRE 411); opinion testimony by lay witnesses (MRE 701); hearsay (MÍRE 802).

The following states have adopted the federal approach: Arkansas (ARE 609); Delaware (DRE 609); Minnesota (MRE 609); Nebraska (NRE 609); New Mexico (NMRE 609); North Dakota (see State v Eugene, 340 NW2d 18, 30 [ND, 1983]); Ohio (ORE 609); Oklahoma (OEC 2609); Oregon (OEC 40.355); Tennessee (see State v Morgan, 541 SW2d 385 [Tenn, 1976]); Utah (URE 609); Washington (see State v Anderson, 31 Wash App 352; 641 F2d 728 [1982]); Wyoming (WRE 609).

The states permitting automatic admission of evidence of all or nearly all prior convictions are: Louisiana (La Stat Ann — Rev Stat 15:495); Mississippi (see Sanders v State, 352 So 2d 822 [Miss, 1977]); Missouri (Mo Ann Stat, § 491.050); Rhode Island (see State v Lombardi, 113 RI 206; 319 A2d 346 [1974]) (the trial court may exclude a prior conviction only on the basis of remoteness in time); Indiana (see Ashton v Anderson, 258 Ind 51; 279 NE2d 210 [1972]) (prior convic*598tions are automatically admissible if they involved dishonesty or false statement or were for treason, murder, rape, arson, burglary, robbery, kidnapping, forgery, or wilful perjury).

The states allowing for automatic admission of certain types of prior conviction evidence of a witness-accused are: Alabama (Ala Code 12-12-162); Colorado (§ 13-90-101, CRS 1973); District of Columbia (DC Code Ann, § 14-305); Florida (Fla Stat Ann 90.610); Maryland (Md Code 10-905); Nevada (Nev Rev Stat 50.095); North Carolina (NC Gen Stat 609); Virginia (see Hackman v Commonwealth, 220 Va 710; 261 SE2d 555 [1980]; Sadoski v Commonwealth, 219 Va 1069; 254 SE2d 100 [1979]).

The three states allowing impeachment providing the prior convictions bear some relevance to credibility are: California (see People v Miles, 153 Cal App 3d 652; 200 Cal Rptr 553 [1984]); Idaho (Idaho R Civ P 43[b][6]); South Carolina (see Taylor v State, 258 SC 369; 188 SE2d 850 [1972]).

The states barring any impeachment of a criminal defendant by prior convictions are: Georgia (Ga Code Ann, § 38-415) (prior convictions are admissible only if defendant first puts his character in evidence or where prior felonies have been "alleged in the indictment as provided by law”); Hawaii (see State v Santiago, 53 Hawaii 254; 492 P2d 657 [1971]) (prior conviction impeachment of a criminal defendant violates his due process right to testify); Kansas (Kan Stat Ann 60-421); Montana (MRE 609).

In State v McAboy, supra at 508, the West Virginia Supreme Court limited prior conviction impeachment of criminal defendants to convictions for perjury or false swearing as "[c]onviction of these crimes goes directly to the credibility of the defendant . . . .” See also State v Clements, — W Va —; 334 SE2d 600 (1985).

The jurisdictions which completely reject the probative/prejudice *599balancing approach are: California, Colorado, District of Columbia, Florida, Georgia, Hawaii, Idaho, Indiana, Kansas, Louisiana, Maryland, Massachusetts, Mississippi, Missouri, Montana, Nevada, North Carolina, Rhode Island, South Carolina, and Virginia.

The states in which trial judges employ the probative/prejudice balancing whenever prior convictions impeachment of a witness-accused is sought are: Alaska (see Frankson v State, 645 P2d 225, 227 [Alas, 1982]); Arizona (ARE 609); Connecticut (see State v Bitting, 162 Conn 1; 291 A2d 240 [1971]); Illinois (see People v Montgomery, 47 Ill 2d 510; 268 NE2d 695 [1971]); Iowa (IRE 609); Kentucky (see Cotton v Commonwealth, 454 SW2d 698 [Ky, 1970]); Maine (see State v Grover, 518 A2d 1039, 1049 [Me, 1986]); Massachusetts (Mass Ann Laws, ch 233, § 21); New Hampshire (see State v Staples, 120 NH 278; 415 A2d 320 [1980]); New Jersey (NJSA 2A.81-12); New York (see People v Sandoval, 34 NY2d 371; 314 NE2d 413 [1974]); Pennsylvania (see Commonwealth v Bighum, 452 Pa 554; 307 A2d 255 [1973]); South Dakota (SD Codified Laws Ann, § 19-14-12); Texas (TRE 609); Vermont (VRE 609); Wisconsin (WRE 906.09).

These five states are Alaska, Connecticut, Kentucky, Pennsylvania, and Vermont.

Compare People v Crawford with People v Kelly. See ante, pp 585-593.

While the dissenters claim they prefer to leave MRE 609 jurisprudence unchanged, their disposition of the instant cases demonstrates their view that the credibility-contest factor should predominate over the effect-on-the-decisional-process factor.

In Gray, Chief Justice Riley’s opinion states: "Standing alone, evidence of convictions of carrying a concealed weapon and possession of heroin may not be highly probative of an individual’s credibility. However, the value of such evidence to the truth-seeking process was critical because the trial had boiled down to a swearing contest between two opposing witnesses.” (Post, pp 641-642.) And in Allen, Chief Justice Riley’s opinion states: "The similarity of the prior offense and the fact that it is not an offense like perjury which bears directly on credibility weigh against its admission. Nevertheless, the rationale underlying our decision in Gray, supra, is similarly applicable here where the question of guilt or innocence rests on the credibility of opposing witnesses.” (Post, p 648.)

The dissent’s emphasis on the numbers of criminal defendants who are convicted indicates an approach to rulemaking based upon probabilistics methods and mathematical techniques. This approach is utilized by the dissent most pointedly and erroneously in its analysis of the Kalven & Zeisel materials cited ante, p 568, n 8.

The dissent states that the Kalven & Zeisel study "assumed” that the judges’ findings as to guilt or innocence of defendants were correct. However, the study did not make any such assumption. It instead compared judge and jury decisions and attempted to discern the reasons for the differences. The only assumption made was that the judges were able to accurately explain why they and the juries disagreed. Never in the work did the authors make the assumption that the judges’ evaluations of the defendants’ guilt were correct.

Even assuming that the statistical analysis is correct, it represents no more than a conclusion that such a number represents the difference in judge and jury conviction rates where the same prejudicial material is made known to each. The fact that judges may also have difficulty ignoring the prejudicial aspects of this evidence does not run counter to our analysis. In other words, some of what the dissent terms "the most significant observations in The American Jury” (post, p 695 were not, in fact, observed by its authors.

Even more important, however, is that apart from the accuracy of the dissent’s statistics, it proposes that in order to gauge the effectiveness of procedural rules we must first and foremost consider what percentage of defendants are guilty. However, as pointed out by Professor Tribe:

[0]nce one is precise and calculating about rulemaking, one can no longer so easily enjoy the benefits of those profoundly useful notions — like the "presumption of innocence” and "acquittal in all cases of doubt” — that [are] threatened by mathematical proof. After deciding in a deliberate and calculated way that it is willing to convict twelve innocent defendants out of 1000 in order to convict 800 who are guilty — because that is thought to be preferable to convicting just six who are innocent but only 500 who are guilty — a community would be hard pressed to insist in its culture and rhetoric that the rights of innocent persons must not be deliberately sacrificed for social gain. [Tribe, Trial by mathematics: Precision and ritual in the legal process, 84 Harv LR 1329, 1390 (1971).]

The dissent, in essence, argues that rules should be determined in the context of the fact that according to its statistics, at least ninety-four percent of all defendants are guilty (not found guilty, but are guilty). While the dissent does not by any means suggest the altering of the presumption of innocence in any given case, this approach would create a general presumption of guilt where the drafting of procedural and evidentiary rules are concerned. Carried to its logical consequences, this approach would result in less protective rules, more convictions, and therefore a self-fulfilling stronger presumption *605of guilt suggesting even less protective rules. Such a focus on trial outcome entails dangers for innocent defendants and for society at large.

[R]ules of trial procedure in particular have importance largely as [expression of certain ends and values] and only in part as means of influencing independently significant conduct and outcomes. Some of those rules, to be sure, reflect only "an arid ritual of meaningless form,” but others express profoundly significant moral relationships and principles — principles too subtle to be translated into anything less complex than the intricate symbolism of the trial process. Far from being either barren or obsolete, much of what goes on in the trial of a lawsuit — particularly in a criminal case — is partly ceremonial or ritualistic in this deeply positive sense, and partly educational as well; procedure can serve a vital role as conventionalized communication among a trial’s participants, and as something like a reminder to the community of the principles it holds important. The presumption of innocence, the rights to counsel and confrontation, the privilege against self-incrimination, and a variety of other trial rights, matter not only as devices for achieving or avoiding certain kinds of trial outcomes, but also as affirmations of respect for the accused as a human being — affirmations that remind him and the public about the sort of society we want to become and, indeed, about the sort of society we are. [Id., pp 1391-1392.]

We agree with the dissenters that the integrity of the fact-finding process is fundamental to rulemaking. However, we do not agree that the integrity of that process is to be measured by conviction rates.

The amendment we adopt today bars impeachment by a prior conviction not punishable by more than one year’s imprisonment *606unless dishonesty or false statement was an element of that prior offense.

We continue the ten-year cutoff for the use of any prior convictions. In addition, for those theft convictions occurring less than ten years prior to the relevant case, the vintage of the prior conviction and the defendant’s behavior subsequent to that conviction are relevant to probativeness. Contrary to the dissent’s assertion, as recently as People v Wakeford, 418 Mich 95; 341 NW2d 68 (1983), we indicated that the vintage of the prior conviction is a proper consideration.

Even though the only crimes falling within the balancing test under the new rule are theft crimes and we have indicated that theft crimes are more probative of credibility than most other prior convictions, the determination of probative value must still be made in each case governed by the balancing test. We are not prepared to say that all theft crimes are of equal probativeness. In addition, the vintage of convictions will vary from case to case.

We note that application of the bright line mandating admission may be improper where the witness is an accomplice of the defendant and impeachment is sought upon the basis of conviction for crimes growing out of the events for which defendant is on trial. People v Lytal, 415 Mich 603; 329 NW2d 738 (1982).

FRE 609 allows the trial judge to consider only prejudice "to the defendant” when a prior conviction is introduced. We note that while this phrase is not included in the Michigan rule, its absence is not reported in the committee note to MRE 609, as are other differences from the federal rule, indicating that the original intent of the Michigan rule may not have differed in this regard from its federal counterpart.

While crimes having an element of theft are probative of veracity, we recognize that there are variations within this category (see n 33) and that the probative value of a particular prior conviction may be insignificant and so it should not be admitted.

An independent problem may arise where the witness was an accomplice of the defendant. See n 34.

The clarified balancing test shall not apply where an initial and direct appeal has been concluded.

After March 1, 1988, only theft crimes will fall within the balancing test. See n 33.

The dissent points out that retroactive application of this decision cannot, "in good faith, be justified by rule-making authority.” (Post, p 661.) We agree, and we do not rely on such authority. The sole portion of this opinion which has retroactive effect is the clarification of the existing balancing test. This portion resolves a split in the lower courts and does not require amendment of MRE 609.

We note that application of the bright-line tests would have *613resulted in the same dispositions in these cases. In Allen, Smith, and Gray, the admission of the prior convictions would have been deemed error, since criminal sexual conduct, the carrying of a concealed weapon, and possession of heroin are not theft crimes, nor do they involve dishonesty or false statement. They therefore would have been excluded pursuant to the amended MRE 609(a).

In Pedrin and Brooks, since the prior convictions were theft offenses, we would have applied the balancing test just as we have today.

However, in both Brooks and Smith, the error would still have been held to be harmless.