172 F.3d 1104
99 Cal. Daily Op. Serv. 2205, 1999 Daily
Journal D.A.R. 2892
UNITED STATES of America, Plaintiff-Appellee,
Tony Julius MITCHELL, Defendant-Appellant.
United States of America, Plaintiff-Appellee,
Tony Julius Mitchell, Defendant-Appellant.
Nos. 96-10411, 97-10427.
United States Court of Appeals,
Argued and Submitted June 12, 1998.
Decided March 26, 1999.
Terry Gross, Friedman, Sloan & Ross, San Francisco, California, for defendant-appellant.
Marcia Jensen, Assistant United States Attorney, San Jose, California, for plaintiff-appellee.
Appeals from the United States District Court for the Northern District of California; Ronald M. Whyte, District Judge, Presiding. D.C. No. CR-96-20020-RMW.
Before: WALLACE, T. G. NELSON and KLEINFELD, Circuit Judges.
KLEINFELD, Circuit Judge:
We reverse because inadmissible evidence regarding the defendant's poverty was admitted and relied upon heavily to secure a conviction.
A bank was robbed at gunpoint, and Mitchell was convicted of the robbery. The only issue at trial was whether Mitchell was the man who did it, or was mistakenly identified. There were weaknesses in the prosecution evidence. Mitchell was easily located from the license plate number and his father, who readily told the police how to find him when asked, yet the police did not contact Mitchell for three months after the robbery. The usual strong evidence of bank robbery--dye pack marks on the robber's body and clothing, marked money found on the robber or people he bought things from, bank videotape or film of the robber, were all lacking. The tellers' descriptions of the robber varied, and several did not fit Mitchell. The witnesses said the robber had no facial hair, but those who knew Mitchell said he always had a beard, mustache or both.
There were also some strengths to the prosecution case. The police found clothes, backpack and a fanny pack in Mitchell's closet that matched the descriptions of what the robber had. One of the eyewitnesses was confident about her identification and had properly picked Mitchell from a police photographic array. The most persuasive evidence was that this witness, who had been outside the bank when it was robbed, had noted the license plate number on the robber's car, and it matched Mitchell's car.I. ANALYSIS
Mitchell argues that the district judge erred in denying a new trial based on newly discovered evidence of an alibi. We do not reach this issue because of our resolution of the evidence issue. Likewise we do not reach issues he raises concerning ineffective assistance of counsel, and unduly suggestive in-court identification. Mitchell is entitled to a new trial because the district court erroneously admitted evidence of his poverty.
A. The evidence at issue.
As indicated above, the prosecutor faced some difficulties in proving that Mitchell was the robber. One of the devices she used was to show that Mitchell had motive: he needed the money because he was poor. This evidence was a major part of the trial.
Before trial started, the prosecutor said she intended to prove that Mitchell was poor in order to show that he had a motive to rob the bank:
This [evidence] would be offered in order to show that Mr. Mitchell was in very poor financial condition, was unemployed at the time, and as [Mitchell's counsel] indicated, it would be sought to be argued by the United States that that was providing the motive for him to commit the bank robbery.
At a subsequent pretrial conference, after the judge said that "just general testimony that someone is poor would be inappropriate" and asked for clarification of the prosecutor's proffer, the prosecutor said "it would not be our intent [to] introduce evidence that because the defendant is poor, he's likely to commit a crime." But she did.
The first day and a half of trial consisted of ordinary bank robbery evidence--witnesses to the bank robbery and a brief witness to establish the jurisdictional fact that the bank was federally insured. Then the case veered into poverty evidence. The prosecutor called Mitchell's landlord, to testify that Mitchell paid the rent in cash the day after the robbery, which was also the normal day the rent fell due. The landlord testified that she knew Mitchell paid his August rent in cash, not because she specifically remembered him doing so, but because "[Mitchell] always paid in cash," both the rent and the security deposit. Mitchell's testimony, unrebutted in this respect, was that the landlord had specified that payments must always be in cash because she had been burned in the past by bad checks. Mitchell never paid the September rent (the month following the robbery), and was evicted by the end of October.
Then the prosecutor called Mitchell's last boss, who said they had fired Mitchell because in his sales job he had only sold two orders, receiving $153.17 in commission. But this job was not Mitchell's only source of income. He was bringing in money through another job and self-employment. His last boss conceded that Mitchell had self-employment and another employer at the same time. Mitchell later testified that his self-employment selling football statuettes produced about $350 per month.
The prosecutor called Mitchell's father as a witness, and used him mainly to show that Mitchell could not adequately support himself and his wife and children. The father paid for his car, even though Mitchell was the only driver, and his parents helped him financially many times. The prosecutor established that Mitchell's father made Mitchell's July and August car payments. But his father always made the car payments on Mitchell's car and continued to do so after the robbery. The father said Mitchell was encouraged by his parents to come to them when he needed money, he came regularly, and they wanted him to. Although the father's testimony was also useful in tying the car to Mitchell, he was harmful to the prosecution in saying that Mitchell always had facial hair--beard, mustache or both (the witnesses testified that the robber was clean-shaven).
The government wound up its case with the FBI case agent, who tied the car to Mitchell, described the photo identifications by the eye witnesses and the clothing found in Mitchell's closet, and said Mitchell was overdrawn at his bank $274.45 on June 27 (a month before the robbery), and $51.26 on July 26 (a week before the robbery). But the bank statements introduced as evidence showed that Mitchell was overdrawn after the robbery as well as before it.
In the prosecutor's closing argument in chief, she spent most of her time on the robbery and identification evidence, but wound up with a reference to Mitchell's inability to pay his bills, reminded the jury that he paid his August rent in cash the day after the robbery, and neither Mitchell, nor his wife, nor any other defense evidence had explained where the money came from. Defense counsel focused on the weaknesses and inconsistencies in the identification evidence, and the degree to which it suggested that Mitchell was not the man who robbed the bank. He pointed out that the government case agent had run six possible combinations of numbers on the license plate, suggesting some uncertainty. Then in her rebuttal argument, when the judge's instructions were long past and defense counsel could not respond, the prosecutor hit the poverty evidence hard:
He didn't have any money. He had minus $57 in his bank account. Look at his bank records. It's about consistent for the entire time that you've got bank records for. [Mitchell's employer] came in and said, "During the course of May to September when Mr. Mitchell was an independent contractor, I paid him a grand total of $157.13," I think it was. Mr. Mitchell is staring down the face of the getting kicked out of his condominium. He doesn't have any money. His wife is not getting the usual [A.F.D.C.] money she's used to getting that month. Lo and behold, the day after the robbery, unexplained he can go in and pay the full amount of the rent in cash. It's not a coincidence. He got the money from the bank. He was unemployed. His account was overdrawn. He paid in cash.
B. The arguments.
Mitchell argues that evidence of a defendant's poverty is not admissible to show motive, citing an evidence treatise by Wigmore and several cases from other circuits such as Davis v. United States, 409 F.2d 453, 457-58 (D.C.Cir.1969), United States v. Reed, 700 F.2d 638, 642-43 (11th Cir.1983), and United States v. Zipkin, 729 F.2d 384, 390 (6th Cir.1984). The government argues that evidence of financial condition is admissible to show motive under United States v. Feldman, 788 F.2d 544, 556-57 (9th Cir.1986), United States v. Jackson, 882 F.2d 1444, 1449-50 (9th Cir.1989), and United States v. Miranda, 986 F.2d 1283, 1285 (9th Cir.1993). In her argument heading, the prosecutor also says the evidence was properly admitted under Federal Rule of Evidence 404(b). She notes that the jury was given a limiting instruction regarding use of the evidence.
The government's characterization of the evidence as admissible under Federal Rule of Evidence 404(b) to show motive is meritless. Rule 404(b) is an exception to the general rule against evidence of character to prove conduct, where evidence of "other crimes, wrongs, or acts" may be proof of motive etc. Fed.R.Evid. 404. The rule speaks only to "[e]vidence of other crimes, wrongs, or acts." Being poor is not a crime, wrong, or act. Rule 404(b) therefore has no application.
C. Relevance and prejudice.
The issue on evidence of poverty is whether it has any relevance, that is, "any tendency to make the existence of any fact that is of consequence to the determination of the action more probable or less probable than it would be without the evidence." Fed.R.Evid. 401. Of course, relevant evidence may be excluded in the discretion of the trial judge "if its probative value is substantially outweighed by the danger of unfair prejudice." Fed.R.Evid. 403. The issue is whether the trial judge was within his discretion in admitting the evidence in this case under these two rules, not Rule 404(b).
The issue of evidence of poverty is something of an old chestnut in the law of evidence, so we do not write on a blank slate. Wigmore expresses the traditional suspicion of admitting evidence of impecuniousness:
The lack of money by A might be relevant enough to show the probability of A's desiring to commit a crime in order to obtain money. But the practical result of such a doctrine would be to put a poor person under so much unfair suspicion and at such a relative disadvantage that for reasons of fairness this argument has seldom been countenanced as evidence of the graver crimes, particularly those of violence.
II Wigmore, Evidence § 392 (Chadbourne rev.1979). The cases Mitchell cites from other circuits do, as he says, stand for the traditional view that evidence of poverty is not admissible to show motive, because it is of slight probative value and would be unfairly prejudicial to poor people charged with crimes.
Of our cases, two are relevant (the third cited by the government, United States v. Miranda, 986 F.2d 1283, 1285 (9th Cir.1993), is not, because we discussed the issue as "evidence of a drug habit," not evidence of financial need). In United States v. Feldman, 788 F.2d 544 (9th Cir.1986), we found no abuse of discretion in a bank robbery trial, where the court had admitted evidence that the defendant's joint account with his father was overdrawn by more than $8,000, and the father had told the bank that his signature was forged on the checks creating the overdraft. We said that the evidence that Feldman owed "substantial sums" was relevant to show motive, quoting another case that said "[e]vidence that tends to show that a defendant is living beyond his means is of probative value in a case involving a crime resulting in financial gain." Id. at 557 (quoting United States v. Saniti, 604 F.2d 603, 604 (9th Cir.1979)).
Likewise, in United States v. Jackson, 882 F.2d 1444, 1449 (9th Cir.1989), we held that evidence was properly admitted that the defendant was "short on funds," "having financial difficulty," and borrowed money because he "couldn't pay for things he needed to have done." Id. We reviewed the traditional "[r]esistance to linking poverty with motivations to commit crimes." Id. We noted that where evidence of impecuniosity properly came into evidence, there was "more than the mere fact that the defendant is poor." Id. We also noted that in the case quoted by Feldman for the "living beyond his means" language, the defendant had a $250 per day heroin habit. The reason we resolved Jackson in favor of admissibility was "an unexplained and abrupt change in that status for the better." Id. at 1450. A witness had testified that he was surprised when the defendant paid $100 for a post office box "because he never had any money." Id.
To determine whether evidence of impecuniousness has relevance, and that its probative value is not outweighed by the risk of unfair prejudice, it is necessary to consider the facts of the particular case. No general proposition can properly resolve all cases, because the multiplicity of circumstances in human conduct is too great. If a man is notoriously broke and cannot buy a pack of cigarettes Tuesday, that night a laundromat is burglarized, and on Wednesday the man buys a carton of cigarettes and a $40 bottle of scotch, all with quarters, the man's financial circumstances have obvious and significant probative value.
Poverty as proof of motive has in many cases little tendency to make theft more probable. Lack of money gives a person an interest in having more. But so does desire for money, without poverty. A rich man's greed is as much a motive to steal as a poor man's poverty. Proof of either, without more, is likely to amount to a great deal of unfair prejudice with little probative value.
There is a distinction between an interest, in the sense that it is in anyone's interest to be richer rather than poorer, and an inclination. A mere interest, unconnected with inclination, desperation, or other evidence that the person was likely to commit the crime does not add much, in most cases, to the probability that the defendant committed a crime. If people commonly committed crimes whenever they needed money and could get it by crime, no company would sell life insurance. There is usually a moral disinclination and an interest in avoiding punishment that restrains people from committing crimes out of mere financial interest. The problem with poverty evidence without more to show motive is not just that it is unfair to poor people, as Wigmore says, but that it does not prove much, because almost everyone, poor or not, has a motive to get more money. And most people, rich or poor, do not steal to get it.
The reason the financial circumstances evidence could come in in Jackson was abrupt change of circumstances. We required "more than the mere fact that the defendant is poor," such as "an unexplained and abrupt change in that status for the better." Jackson, 882 F.2d at 1449, 1450. An unexplained abrupt change of circumstances is not merely proof of motive, but also amounts to circumstantial evidence of the crime. In the laundromat hypothetical, that the man who was broke yesterday has a couple of hundred quarters today tends to show that the man actually did burglarize someplace that kept its money in quarters, not merely that he had a financial interest in doing so. In Jackson, that the man had no money before, and suddenly had an unexplained $100 to pay for something he could not ordinarily afford, tended to show that he had stolen the $100, not just that he had a financial interest in having $100 more than he did before the theft.
Likewise in Feldman, the size of the $8,000 overdraft and the defendant's father's complaints to the bank that his signature had been forged tended to prove that Feldman had a desperate need to cover the overdraft, not just that he would be better off if he were a few thousand dollars richer. Feldman was squeezed, not just poor. Likewise in the dope addict cases, such as Miranda, the addiction establishes a likelihood of desperate need and lack of self control, not just financial interest in being richer.
In this case, the evidence did not show "more than the mere fact that the defendant is poor." Jackson, 882 F.2d at 1449. The prosecutor implied in her closing argument that paying the rent in cash August 4, the day after the robbery showed an abrupt change of circumstances, but the evidence did not support that. Mitchell was as overdrawn when he paid the rent in cash July 5 or 6, as he was in August. Because Mitchell was chronically overdrawn, before and after the robbery, whether he had cash or not, his overdraft did not show lack of money to pay the rent. Nor did the evidence establish desperation. There is no reason why a man who has maintained an empty, overdrawn checking account would suddenly need to rob one bank to cover his small overdraft at another. Unlike Feldman, neither the size of the overdraft nor the defendant's father's conduct put any pressure on Mitchell to rob a bank to cover it. Nor did the evidence establish desperation to keep a roof over his family's head. Mitchell's family stayed in the apartment to the end of October, even though he did not pay the rent in September or October, and Mitchell's father said and had demonstrated readiness to help Mitchell out financially as necessary. The cash rent payment was not circumstantial evidence of the robbery, because no marked bills were proved, the day he paid was when the rent was due, and Mitchell always paid in cash pursuant to his landlord's requirement.
There is a distinction that cuts in favor of Mitchell between the unexplained ability to pay $100 for a post office box in Jackson and the unexplained ability to pay $885 rent in the case at bar, even though the amount of money cuts against him. In Jackson, a witness testified that "he was surprised" when Jackson paid the $100 "because he never had any money." Jackson, 882 F.2d at 1450. But it was no surprise when Mitchell paid the $885 in cash. That was the day his rent was due. He had paid it in cash before. The landlord required payment in cash. He had previously paid his landlord even though he was overdrawn at the bank and did not have earnings to cover the payment. Perhaps he was in the habit of keeping his cash out of the bank when the bank might use it to cover his overdraft and overdraft charges. We explained why the defendant's finances could come into evidence in Jackson by saying that the evidence showed "an unexplained and abrupt change in that status for the better." Id. It was the "abrupt change" that made Jackson more like the person with all the quarters in the laundromat hypothetical. In this case there was no such abrupt change.
The poverty evidence was not only of negligible probative value, but also produced a high "danger of unfair prejudice." Fed.R.Evid. 403. The prosecutor did not merely show that Mitchell would be better off if he had a few thousand dollars more. She effectively portrayed him as a feckless man who did not support his wife and children. She showed with the poverty evidence that Mitchell let his wife draw welfare while he went to the basketball court, lived on his parents' bounty at an age where most people do not, and let his family get evicted from their apartment. Jurors' feelings about a man who lives that way have no legitimate bearing on whether he should be convicted of robbing a bank. That a person is feckless and poor, or greedy and rich, without more, has little tendency to establish that the person committed a crime to get more money, and its probative value is substantially outweighed by the danger of unfair prejudice. The district court's discretion was not broad enough to allow admission of the evidence of Mitchell's impecunious financial circumstances.
The next issue is whether the error was harmless. Error not affecting substantial rights must be disregarded. Fed.R.Crim.P. 52(a).
The government does not, in so many words, argue in its brief that error if any was harmless. All it does is say that the judge's instruction "assured that the jury would consider the evidence ... solely for the proper purpose." The problem with this argument is that there was no proper purpose for which the evidence of Mitchell's poverty before, during and after the bank robbery could be considered. This was not analogous to evidence admitted properly for one purpose which the jury is admonished to use only for that purpose.
The judge instructed the jury, using language from United States v. Jackson, 882 F.2d 1444 (9th Cir., 1989), that while poverty alone was insufficient to prove motive, evidence of an unexplained or abrupt improvement in financial circumstances might indicate commission of a crime:
Evidence was introduced as to defendant's financial situation at and near the time of the charged bank robbery. Evidence of poverty or financial need alone does not indicate a motive to commit a crime. Therefore, you cannot conclude that just because a person is poor or needy he is likely to commit a crime to obtain money. However, evidence that shows a person has an unexplained and abrupt change of financial status for the better may, but of course does not necessarily, indicate a commission of a crime.
The instruction is a fair paraphrase of Jackson. But because there was no evidence at Mitchell's trial of "an unexplained and abrupt change of financial status for the better," the instruction should not have been given. The jury had evidence before it of poverty, which it should not have had, and had no evidence of abrupt change, so telling the jury that it could consider the poverty evidence only insofar as it established abrupt change could not cure the erroneous admission. The instruction might be good for another case, but not this one.
We must reverse for the objected to error in admitting the poverty evidence unless it is more probable than not that the error did not materially affect the verdict. United States v. Rahm, 993 F.2d 1405, 1415 (9th Cir.1993). The government bears the burden of persuasion with respect to proving that the error was harmless. United States v. Annigoni, 96 F.3d 1132, 1145 (9th Cir.1996).
We are just not sure whether the error was harmless. Most of the prosecution case for identifying Mitchell as the robber was equivocal. But one fact, an eyewitness's noting of what turned out to be his license plate number, seems very persuasive to us. On the other hand, it evidently was not overwhelmingly persuasive to the jury, so we may be missing something about the nuances of how the number came into evidence. Defense counsel had demonstrated some ambiguity about what number was noted, and how it was searched, and the jury asked a question during deliberations about whether the numbers would identify the color and kind of vehicle. The jury was out for eight hours over two days, so they evidently found this to be a hard case. One juror at one point sent out a note saying she "could not live with myself" if she decided the case on the evidence they had. Because the judge instructed before argument, as he had discretion to do, the closing arguments came subsequently, so the last thing the jury heard before retiring to deliberate was the prosecutor hammering on the poverty evidence in her rebuttal argument.
Because we are in equipoise about whether the error was harmless, cf. O'Neal v. McAninch, 513 U.S. 432, 437-38, 115 S.Ct. 992, 130 L.Ed.2d 947 (1995), the government has not established harmlessness, so we REVERSE and REMAND for a new trial.