United States v. Romano

14-858 (L)
USA v. Romano

1 UNITED STATES COURT OF APPEALS

2 FOR THE SECOND CIRCUIT

3 ------

4 August Term, 2014

5 (Argued: February 9, 2015 Decided: July 27, 2015)

6 Docket Nos. 14-858, 14-1298

7 _________________________________________________________

8 UNITED STATES OF AMERICA,

9 Appellee,

10 - v. -

11 MICHAEL ROMANO, WILLIAM KEARNEY, a/k/a "Ed Thompson," a/k/a
12 "George,"

13 Defendants-Appellants.
14 _________________________________________________________

15 Before: KEARSE, LIVINGSTON, and CARNEY, Circuit Judges.

16 Appeals from judgments of the United States District Court for the Eastern District of

17 New York convicting defendants of conspiring to commit mail and wire fraud in the sale of coins, in

18 violation of 18 U.S.C. § 1349, and money laundering conspiracy in violation of 18 U.S.C.

19 §§ 1956(a)(1)(A)(i) and 1956(h); sentencing defendant Romano principally to 240 months'

20 imprisonment and defendant Kearney principally to 156 months' imprisonment; and ordering

21 defendants to pay $9,139,727.10 in restitution and to forfeit $32,220,617. Defendants contend

22 principally that the court erred in admitting expert testimony as to the grading and valuation of coins.
 1 See 859 F.Supp.2d 445 (2012). Defendants also contend that their sentences are unreasonable and

2 that the court erred in failing to review de novo the magistrate judge's recommendations for restitution

3 and forfeiture. The government concedes that there should be a remand with respect to restitution and

4 forfeiture.

5 Remanded for de novo review of the recommendations for restitution and forfeiture;

6 in other respects, the judgments are affirmed.

7 LARA TREINIS GATZ, Assistant United States Attorney, Brooklyn,
8 New York (Loretta E. Lynch, United States Attorney for the
9 Eastern District of New York, Susan Corkery, Emily Berger,
10 and Diane Beckmann, Assistant United States Attorneys,
11 Brooklyn, New York, on the brief), for Appellee.

12 RICHARD LEVITT, Levitt & Kaizer, New York, New York (Yvonne
13 Shivers, Emily Golub, on the brief), for Defendant-Appellant
14 Romano.

15 MATTHEW W. BRISSENDEN, Garden City, New York, for
16 Defendant-Appellant Kearney.

17 KEARSE, Circuit Judge:

18 Defendants Michael Romano and William Kearney appeal from judgments of the

19 United States District Court for the Eastern District of New York entered by Sterling Johnson, Jr.,

20 Judge, following a jury trial before Joseph F. Bianco, Judge, convicting defendants of conspiracy to

21 commit mail and wire fraud in connection with the sale of coins, in violation of 18 U.S.C. § 1349, and

22 conspiracy to engage in money laundering, in violation of 18 U.S.C. §§ 1956(a)(1)(A)(i) and 1956(h).

23 Romano was sentenced principally to 240 months' imprisonment, to be followed by a five-year term

24 of supervised release; Kearney was sentenced principally to 156 months' imprisonment, to be followed

25 by a three-year term of supervised release; defendants were ordered, jointly and severally, to pay

2
 1 restitution in the amount of $9,139,727.10 and to forfeit $32,220,617. On appeal, defendants

2 challenge their convictions principally on the grounds that the court failed to perform the necessary

3 gatekeeping function before admitting the testimony of expert witnesses as to the grading and

4 valuation of coins and that admission of one of the government's exhibits violated their rights under

5 the Confrontation Clause. Kearney also argues that the evidence was insufficient to show that he

6 knowingly and willingly participated in a conspiracy. Defendants challenge their sentences as

7 procedurally and substantively unreasonable and contend that the district court erred in failing to

8 conduct de novo review of the magistrate judge's recommendations for restitution and forfeiture. The

9 government concedes that there should be a remand for consideration of defendants' objections to the

10 magistrate judge's recommendations and the extent to which a certain property is forfeitable.

11 For the reasons that follow, we remand for de novo review of the magistrate judge's

12 recommendations for restitution and forfeiture, including forfeitability of the property in question; in

13 all other respects we affirm.

14 I. BACKGROUND

15 The present prosecution focused on the operations of a succession of three companies

16 owned by Romano: Wall Street Rare Coins (or "WSRC"), which operated from approximately

17 December 1990 until October 2004, Atlantic Coin Galleries ("ACG"), which operated from

18 approximately October 2004 until August 2008, and Northeast Gold and Silver ("NEGS"), which

19 operated from approximately September 2008 until November 2008 (collectively, "the Romano

20 Companies"). Kearney was a salesman and sales manager at all three companies and was also a

3
 1 50-percent owner of NEGS. Romano and Kearney were alleged principally to have engaged, through

2 the Romano Companies, in a fraudulent scheme for sales of coins by telephone, making and causing

3 the making of false representations to customers as to the grade and value of the coins they sold, the

4 general profitability of coin sets, and the assistance the Romano Companies would provide to

5 customers who later wished to sell coins bought from the Romano Companies.

6 A. The Government's Evidence at Trial

7 The government's evidence at trial consisted principally of testimony from several of

8 the Romano Companies' former salesmen ("salesmen"), former suppliers ("suppliers"), and customers;

9 government investigators; and three expert witnesses who testified about the grading and valuation

10 of coins.

11 1. General Industry Practice and Procedures

12 The government's first expert witness, Scott Schechter, whose testimony is not

13 challenged on these appeals, provided general information with respect to the rare coin market.

14 Schechter was a vice president of Numismatic Guarantee Corporation ("NGC"), an independent (or

15 third-party) coin grading service, i.e., an entity not in the business of buying or selling coins. He

16 testified that NGC and its competitor Professional Coin Grading Service ("PCGS") each have about

17 45 to 47 percent of the coin-grading market. (See Trial Transcript ("Tr.") 1038.) He described

18 general industry practices, and NGC's procedures, in the grading and certification of rare coins as

19 follows.

4
 1 Coins are generally graded on a 70-point scale, 1-70 (though not all numbers are used),

2 called the Sheldon scale. The Sheldon scale has been in fairly common use since the middle of the

3 20th century. At NGC, each coin is graded seriatim by three experts who ultimately arrive at a

4 consensus on the coin's grade. In evaluating some coins, two graders can reasonably differ in their

5 opinions as to grade; at NGC a third grader examines all coins. The graders consider four attributes:

6 the coin's luster, i.e., the sheen or vibrancy of its surface; its strike quality, i.e., how well the coin has

7 been struck between the two dies that give the coin its details; the existence and number of contact

8 marks, i.e., surface wear; and its eye appeal, i.e., how attractive the coin is. Eye appeal is the most

9 subjective of these factors.

10 At the high end of the Sheldon scale are "mint state" ("MS") coins that can be assigned

11 a number from 60 through 70; they are considered "uncirculated" because they show no evidence of

12 metal loss from having been circulated. The next lower range is for coins near uncirculated condition;

13 they are called "about uncirculated" ("AU") coins and can be graded 50, 53, 55, or 58. In addition,

14 a "+" can be added to a coin's grade to indicate that the coin is in the top 15 percent of the pieces

15 within that grade. Small differences in a coin's grade can result in substantial differences in the coin's

16 market value.

17 Some collectors accumulate complete sets of coins that are of interest to them; that is

18 the most common way to collect coins. However, a set has no greater monetary value than the total

19 values of its individual coins.

20 "Certified" coins are coins that have been graded by a third-party grading service such

21 as NGC or PCGS. After grading a coin, the grading service normally places it in a protective capsule,

22 or "slab," that is then sealed and difficult to open without a tool. An "unslabbed," or "raw," coin is

5
 1 one that is not certified by a grading service. NGC gives a guarantee for any coin it has certified (for

2 so long as the coin remains in its slab), meaning that if a coin fails to trade in the marketplace for as

3 much as the price associated with the grade given it by NGC, NGC either will buy the coin from the

4 owner at the price associated with the certified grade or will regrade the coin, return it to the owner,

5 and pay the owner the difference in value. NGC's grading is relied on in transactions between and

6 among dealers and collectors; indeed, some buyers purchase coins sight unseen based solely on the

7 grades given by NGC. PCGS gives similar guarantees.

8 2. Testimony by Romano Companies Suppliers or Salesmen

9 Suppliers to the Romano Companies testified that before purchasing coins from them,

10 Romano examined the coins only briefly, if at all. (See Tr. 1583-86, 1878, 2392-93.) For example,

11 one supplier, who over the years sold some $2,000,000 of coins to Romano, testified that Romano

12 typically, without gloves or magnifier or other aid, would look at and purchase some 300-400 coins

13 in 20-30 minutes. (See id. at 1583-86.) Romano would sometimes remove certified coins from their

14 slabs and hand the coins to his secretary. When Romano had asked the supplier for rolls of coins, the

15 supplier brought rolls and Romano bought them without looking at the coins; the grades of the coins

16 were not discussed. (See, e.g., id. at 1584-86.)

17 Romano periodically created a coin inventory for the use of his salesmen, and during

18 that process he assigned grades to the coins he had purchased. Former employees testified that when

19 Romano took inventory he would call a particular salesman, usually Robert Lepolszki, to learn the

20 current prices for a given coin at different grades. Lepolszki testified that once a week or twice a

21 month, Romano would call, tell him to grab his coin-pricing book and turn to a particular page, and

6
 1 would ask, e.g., "what is the 1948 mint state 64 grade condition, what is a 65 grade condition. I would

2 give him the prices" (Tr. 2291). Lepolszki would then hear Romano telling the secretary to "put it

3 down mint state 64, $10,000, or whatever the amount [was]" (id. at 2291-92), but it was always at

4 "[t]he high end" (id. at 2293). When Romano had all the prices, an inventory sheet would be printed

5 and distributed to all the salesmen. All of the coins Romano and Lepolszki discussed would be on

6 the inventory sheet; and they had been graded at the higher end of the prices Lepolszki had relayed.

7 (See id. at 2294; see also id. at 1742-43 (Romano told one of the salesmen he "marked up the coins

8 to the max").)

9 A former secretary testified that in five years of working for Romano, she never saw

10 him examine coins, either while he was taking inventory or at any other time. (See id. at 486.)

11 Romano's salesmen used to joke "about Mike's grading practice" (Tr. 1157; see, e.g., id. at 1148; id.

12 at 851 (it was "a running joke")), and "even joke about customers, how gullible they [we]re" (id.

13 at 1157).

14 Romano tended to assign particular grades for certain types of coins; these included

15 Benjamin Franklin half dollars, of which the Romano Companies sold many rolls. All or nearly all

16 of the rolls of Franklin half dollars sold by the Romano Companies were graded by Romano as

17 MS 64+. However, suppliers testified that they never sold entire rolls of Franklin half dollars to

18 Romano at grade 64 or above, because such rolls are simply not available in the marketplace. (See,

19 e.g., id. at 1503-04 ("You won't buy a complete 64. The problem is it doesn't exist."); id. at 1587

20 ("They just don't come that way."); id. at 2393-94 ("Generally on most of the dates they don't exist

21 on those grades.").) Suppliers testified that the Franklin half dollars they sold Romano in rolls were

22 typically graded lower, i.e., MS 62 to 64 (see id. at 1469), or MS 60 to MS 64 (see Tr. 1553,

7
 1 1586-87), or MS 62 to 63 (see id. at 2390-91), or MS 60 to 62 (see id. at 1877-81). Schechter testified

2 that the proper grade for a roll of coins is the grade of the lowest included coin. (See Tr. 1031.)

3 In addition, one supplier testified that Romano evinced a preference for $20 gold pieces

4 in the AU 58 to MS 61 grade range. The supplier testified that "AU-58[ coins] are called sliders

5 because they could be graded as MS-61 or 60 or 58, because to one untrained eye those coins may

6 look uncirculated . . . ." (Tr. 1551.) Generally, the market prices for $20 gold pieces graded MS 62

7 or MS 63 were much higher than the prices for those in the AU 58 to MS 61 range. Customers

8 submitted for analysis in this case $20 gold pieces they had bought from the Romano Companies;

9 each of the 17 for which the Romano Companies' grading could be determined was sold as MS 63

10 grade or higher.

11 Romano and Kearney closely supervised the companies' sales force, and Kearney,

12 using an alias--as did all of the salesmen--also made sales calls himself. Both men worked mostly

13 from desks on the small sales floor (see Tr. 519, 839-40, 1642); one salesman testified that Kearney

14 was on the sales floor "[p]robably 90 percent of the time" (id. at 1109). Kearney, as the sales

15 manager, trained the salesmen and provided them with sales pitches. (See, e.g., id. at 838-39, 2182.)

16 He also directed salesmen to submit their own handwritten pitches for his approval; he returned typed

17 versions of the pitches he approved and gave feedback on those he did not. According to deposition

18 testimony from Romano that was read into the trial record, "everything" the salesmen said "ha[d] to

19 be sanctioned" by management, and "[w]e monitor[ed] them" to ensure compliance. (Id. at 815-17.)

20 Salesmen also testified that Romano and Kearney regularly listened in to sales calls and critiqued

21 them, sometimes feeding them lines during calls or even taking over calls and pitching to customers

22 directly.

8
 1 The company-approved sales pitches contained numerous false representations, with

2 salesmen stating that they were offering bargains, i.e., coins priced below the normal retail price for

3 coins of the same type and grade, when in fact the quality of the offered coins was simply lower than

4 represented. The salesmen represented that they could offer such bargains because the company was

5 operating as a wholesaler, which it was not (see id. at 550-51, 1146); or because they were liquidating

6 another dealer's inventory, which they were not (see id. at 845); or because they had just returned from

7 a coin show, though the salesmen never actually attended coin shows (see id. at 1146, 2210).

8 Salesmen were also instructed by Kearney to tell customers that by collecting a complete set of certain

9 coins, they could reap a "set premium," i.e., that "[t]he set is worth more than the sum of its parts" (id.

10 at 856)--anywhere from five to 50 percent above the total value of the coins individually (see id.

11 at 854-55). Salesmen testified that they used to joke about this pitch and knew it was false because

12 the Romano Companies had in inventory entire sets listed at prices below the aggregate prices of the

13 individual coins in the set. (See, e.g., id. at 856.) Nonetheless, the salesmen, and Kearney himself,

14 used the "set premium" pitch to persuade customers to buy coins toward completion of sets; and

15 multiple customers testified that they heard and relied on it.

16 Salesmen also used to joke about the fact that as far as the Romano Companies were

17 concerned, "[t]here really was no complete set" (Tr. 1126). They were instructed by Romano and

18 Kearney to keep changing the parameters for what would constitute a set: "It depend[ed] on what the[

19 customer] bought originally, but it would always change. . . . [Y]ou could make up whatever you

20 wanted as a set. And you would never let them complete the set," so the customer would always have

21 to buy more. (Id.; see id. at 1129.) The salesmen would sing a little ditty that went, "This is the set

9
 1 that never ends, yes, it goes on and on my friends," and Romano and Kearney would laugh or join in.

2 (See id. at 1130-31.)

3 Kearney also instructed salesmen to tell customers that the Romano Companies

4 employed three in-house graders (see Tr. 853, 1120); that the coins were graded in accordance with

5 the highest standards of the American Numismatic Association ("ANA") (see id. at 853, 2265); and

6 that their grading process was similar to those used by NGC and PCGS (see id. at 853). Kearney

7 himself, and all of the salesmen, made such representations in sales calls. (See, e.g., id. at 866, 1518.)

8 Customers testified that they relied on these representations about the grading process. But in fact

9 the only "grader" was Romano, and he did not examine the coins.

10 Several salesmen testified that they not only knew that their descriptions of the

11 companies' grading process were false, but also knew that the grades themselves were inflated. One

12 testified that he repeatedly overheard Romano ordering uncirculated Franklin half dollars with no

13 specification as to their grade; nonetheless, the Romano companies represented nearly all of the

14 Franklin half dollars they sold to be MS 64+. That salesman also sold Morgan silver dollars that the

15 Romano Companies graded MS 64; he did not believe that grading was accurate, given the high

16 volume and large profit margins on these coins, as well as the "constant customer complaints about

17 quality." (Tr. 851-52.) On one occasion when he failed to sell a coin that Romano had graded MS 62,

18 because the customer wanted a coin graded MS 63, Romano instructed him to call back and sell that

19 coin by simply telling the customer he had an MS 63. (See Tr. 876.)

20 In addition, although Romano and Kearney required their salesmen to sign a form

21 promising that they would not hype the coins as an "investment," and did not allow them to use that

22 specific word (see, e.g., Tr. 874-75, 1152, 1703), the salesmen viewed the form as "a big joke" (id.

10
 1 at 1157) that had no real effect on their sales pitches, even when Romano and Kearney were listening

2 in (see id. at 1642, 2301-02; see also id. at 875, 1152). Indeed, sales pitches distributed by Kearney

3 instructed salesmen to tell customers that buying Franklin half dollars from the Romano Companies

4 was "a can't miss situation" (id. at 2247-48) and to say, "'If these coins do half as well as they've done

5 in the past, believe me, you're going to wish you have 10,000 of them'" (id. at 2238-39). Thus, despite

6 knowing that the Romano Companies falsely inflated the coins' value, Kearney and the other salesmen

7 regularly told customers that owning the coins would be profitable.

8 Kearney and the salesmen also told customers that if the customers decided to sell the

9 coins in the future, the Romano Companies would buy the coins back or help customers with resale;

10 several customers testified they relied on this representation. (See Tr. 552-53, 1518-19, 1838.)

11 However, Kearney trained the salesmen on how to "push[ customers] off until a later time" when they

12 wanted to sell, and one salesman testified that he overheard Kearney telling customers who made such

13 requests that "it wasn't a good time right now" and "we should wait." (Id. at 869.) Thus, one

14 customer testified that when she contacted a salesman about selling some coins, the salesman became

15 "evasive," told her "[i]t was not a good time to sell," and said he would get back to her, which he

16 never did. (Id. at 1326.) And when another customer tried to contact ACG "to try to get them to buy

17 the coins back," he discovered that ACG "had closed shop and disappeared." (Id. at 1906.)

18 Salesmen and secretaries testified that they forwarded customer complaints to Romano

19 and Kearney, including complaints that coins were overgraded and not worth what the customers had

20 paid. (See, e.g., Tr. 384, 388-89, 472-73, 876-77, 1673, 2280-81.) One customer, Ms. Williams,

21 testified that she complained to Kearney that he had misrepresented the value of the coins he had sold

22 her, and that Kearney responded by saying that grading companies were not necessarily reliable and

11
 1 reiterating that the Romano Companies had three in-house graders. Williams testified that Kearney

2 (who used the alias "Ed Thompson") had sold her a 1936 Walking Liberty half dollar proof coin as

3 an MS 67; but when she had it appraised by NGC, NGC graded it "a proof 64. So instead of being

4 worth almost $10,000, it was worth $2,000." (Id. at 1359.) Williams also testified that Kearney

5 consistently steered her away from buying certified coins when she requested them.

6 Romano and Kearney made extensive efforts to conceal their business practices from

7 potential law enforcement agents and from customers. One salesman testified that Kearney repeatedly

8 worried aloud about the possibility of undercover police officers infiltrating the sales staff (see

9 Tr. 872), as did Romano, and that in fact "there was a general concern in the office about that" (id.

10 at 873). Romano twice renamed his company and relocated it within the same geographic area, for

11 no apparent reason other than to conceal its business practices. The entity, its personnel, and its

12 practices remained the same. Romano and Kearney instructed salesmen and secretaries to shred

13 invoices or letterheads bearing the name of the prior company and to deny any affiliation with

14 it--unless they thought the affiliation would help make a sale. (See, e.g., Tr. 405, 879-81, 1174-75.)

15 In addition, one supplier testified that Romano asked him to change items listed on

16 some invoices to reflect coins different from the ones Romano had actually purchased. (See

17 Tr. 1595-96.) And Kearney instructed the secretaries not to put any coin grades on invoices for

18 particularly expensive orders (see id. at 383-84); he argued to Romano that the company should not

19 put grades on invoices (see id. at 1171), and came up with the idea to stamp some orders as

20 nonreturnable "special orders" (id. at 540-42), which the companies used for customers "who were

21 considered problematic or [who] had returned [coins] in the past" (id. at 871).

12
 1 William Hessle, who as a postal inspector had participated in the investigation into the

2 Romano Companies, testified that bank records showed that, between 2000 and 2008, the companies

3 sold coins to approximately 1,450 customers and generated revenue of some $32 million. Of that

4 sum, Romano personally received approximately $6.7 million, and Kearney personally received

5 approximately $3.9 million.

6 3. Expert Testimony as to Romano's Grading

7 The government introduced the testimony of two numismatic experts, Richard

8 Montgomery and Anthony Swiatek, who respectively testified with regard to, inter alia, the grading

9 and the valuation of coins sold by the Romano Companies. Defendants had moved for an in limine

10 ruling under Fed. R. Evid. 702 precluding such testimony and precluding evidence as to the Sheldon

11 scale, arguing that the grading of coins is inherently subjective and thus an inappropriate subject for

12 expert testimony. That motion was denied. (See Part II.A. below.)

13 Richard Montgomery was president of NGC and had been president of PCGS. He had

14 been trained as a coin grader and authenticator in the early 1980s; he was employed as a grader by

15 PCGS from 1987 to 1997 and by NGC from 2002 to 2006. As a grader, he graded at least 1,000 coins

16 a day; he estimated that he had personally graded several million coins. Montgomery also taught

17 seminars on coin grading, and he continued to grade coins while serving as the president of each

18 organization. He described the industry grading scale and the grading practices and procedures at

19 NGC in the same way Schechter had.

20 In connection with the present case, NGC had examined numerous coins sold by the

21 Romano Companies; the government introduced a spreadsheet, Government Exhibit 109 ("GX 109"),

13
 1 which listed a total of 1,823 such coins. Among the coins examined by NGC were 66 20-coin rolls

2 of Franklin half dollars. As indicated in Part I.A.2. above, nearly all of the Franklin half dollars that

3 the Romano Companies sold were represented to be MS 64+ or above. However, Montgomery

4 testified that not one of the 66 Romano Companies rolls was graded by NGC as high as MS 64+.

5 Indeed, NGC graded all but one roll below 64. In addition to the rolls of Franklin half dollars, NGC

6 had examined several hundred other individual coins sold by the Romano Companies. With respect

7 to 420 such coins, only 27 were graded the same by both Romano and NGC. And only 10 of the 420

8 were graded higher by NGC than by Romano; 383 were graded lower by NGC than by Romano.

9 Montgomery cited 23 examples of coins that the Romano Companies sold as high-quality mint-state

10 coins, but which NGC graded in the "about uncirculated" range, including one gold coin that the

11 Romano Companies sold as an MS 65, which NGC graded AU 55, and one Walking Liberty half

12 dollar sold by the Romano Companies as an MS 65, that NGC graded AU 50.

13 Montgomery acknowledged that coin grading is an art rather than a science and that

14 it is "a subjective measure" but stated that it is "built around a core set of principles." (Tr.

15 at 1971-72.) He testified that attributes such as luster and strike quality, though difficult to articulate,

16 can be illustrated through "visual examples" (id. at 1975-76), and that NGC maintains coin samples

17 of various grades that can be consulted for reference. Montgomery testified that grading discrepancies

18 of one or two points could occur occasionally but that three- or four-point discrepancies would be

19 quite unusual. (See id. at 1993-94.)

20 Anthony Swiatek testified that he was a professional numismatist who had studied with

21 leading numismatists to learn the trade and had been engaged in buying and selling coins full time

22 since 1979. He had written books about coins, lectured at major coin shows, written for the industry

14
1 magazine Coin World, and contributed to price guides or coin valuation books; he was a member of

2 the Professional Numismatist Guild--an organization of coin dealers--and a coin grading consultant

3 for NGC, PCGS, and the ANA's grading service. He was a former president of the ANA and member

4 of its Board of Governors; in the latter capacity he had served on a committee that handled frauds and

5 coin improprieties. He had graded and valued at least a million coins.

6 Swiatek testified that in order to determine the value of a graded coin, he typically

7 consults at least one of two periodicals. One is a weekly publication called the Coin Dealer

8 Newsletter and known as the "Greysheet," which lists bid and ask prices for various types of coins

9 at various grades. The other is the Certified Coin Dealer Newsletter, known as the "Bluesheet," which

10 lists bid and ask prices for coins certified by NGC or PCGS.

11 Swiatek gave his opinions as to the proper valuations of approximately 1,800 coins

12 sold by the Romano Companies. In determining such valuations, he did not examine the actual coins

13 but relied on the grades given to them by NGC. To determine the market value of the coins, Swiatek

14 relied in part on the bid and ask prices listed in the Greysheet and/or Bluesheet as of the time the

15 Romano Companies sold each coin; where the difference between NGC's grade and Romano's grade

16 was particularly substantial, Swiatek used the bid price. Swiatek testified that he generally relied on

17 the Greysheet rather than the Bluesheet except in instances where he knew that, at the time of the

18 Romano Companies' sale, buyers and sellers in the industry were using the Bluesheet to determine

19 prices for that type of coin. Swiatek also relied in part on auction prices, especially for coins that were

20 particularly rare, and on his own experiences as a dealer. Using these sources, Swiatek determined

21 (a) the market prices for the 1,800 coins using the grades assigned by Romano and (b) the market

22 prices for those coins as graded by NGC; Swiatek compared those market prices against the prices

15
 1 at which the Romano Companies actually sold the coins. His resulting spreadsheet showed that those

2 coins, sold by the Romano Companies for a total of $959,494, were, as graded by NGC, worth a total

3 of $136,859.65, approximately 14 percent of their aggregate sales price. Thus, with respect to those

4 coins, the Romano Companies had overcharged their customers by $822,634.35.

5 For example, the Romano Companies had sold a gold coin that Romano graded an

6 MS 63 for $8,800. Although the market price for a coin of that type and grade was $11,000 according

7 to the Greysheet or auction records, the coin sold by the Romano Companies was graded only an

8 AU 55 by NGC. At the latter grade, the coin's value was $400. Another coin sold by the Romano

9 Companies for $12,500 as an MS 65 would, at that grade, have had a market price of $10,500

10 according to Swiatek's analysis; but it was graded an AU 58 by NGC and as such was worth just $600.

11 Yet another coin sold by the Romano Companies as an MS 65 for $21,000 would, at that grade, have

12 had a market price of $4,050; it was graded MS 62 by NGC and as such was worth $1,150.

13 Swiatek conceded that he had made errors in entering some of the numbers on his

14 spreadsheet, but he stated that he had thoroughly checked the numbers for the rarer coins and that his

15 methodology was sound. Swiatek stated that of the 1,800 coins he had considered, he had made errors

16 as to 45, all of which were Franklin half dollars; and having overvalued some of those coins and

17 undervalued others, the cumulative effect of his errors was that the amount by which his spreadsheet

18 showed the Romano Companies had overcharged their customers ($822,634.35) was overstated by

19 just $176.75.

16
1 B. The Defense Case

2 Defendants presented the testimony of an investigator, Josh Leicht, who was not a coin

3 expert or collector or buyer or seller. (See, e.g., Tr. 2588.) Based on the spreadsheets introduced at

4 trial by the government, Leicht had performed a comparative analysis of the grades assigned by

5 Romano and those assigned by NGC. According to his analysis, NGC had graded 1,671 of 1,823

6 coins (determining that the remainder could not be graded). Leicht testified that the spreadsheets

7 showed disagreements between the first and second NGC grader as to 167 coins, including 11

8 involving a discrepancy of two or more points and/or 12 as to whether a coin fell in the MS or AU

9 range.

10 As to some of the 1,671 coins graded by NGC, Romano's grade was unknown; there

11 were thus approximately 1,600 coins for which NGC's grade and Romano's grade could be compared.

12 Leicht testified that as to 570 coins, Romano had given a grade one point higher than NGC; and 291

13 coins had been graded by Romano two points higher than NGC. Leicht testified that Romano had

14 given the same grade as NGC to 399 coins; however, Leicht had rated grades of 64+ by Romano as

15 equal to grades of 64 by NGC, because defense counsel had instructed him to disregard pluses. Leicht

16 also testified that Romano's grade was lower than NGC's grade on 66 coins. However, as to 24 of

17 those coins (depicting the Liberty Bell, with lines along the bottom of the bell), Leicht had found

18 NGC's grade higher than Romano's only because it bore the notation "FBL," knowing that that stood

19 for "full bell line[s]" (Tr. 2607), but not being aware that an FBL notation is merely descriptive of the

20 continuity of lines on the coin and did not enhance the grade given to it (see id. at 2008-09, 2607-08).

21 Leicht's calculations apparently left 275 coins that Romano had graded three or more points higher

22 than NGC had.

17
1 C. The Verdicts and the Motions for a New Trial

2 The jury found Romano and Kearney guilty on each of the counts against them, i.e.,

3 conspiracy to commit mail and wire fraud in the sale of coins, and conspiracy to engage in money

4 laundering through financial transactions involving the proceeds of their mail and wire fraud

5 conspiracy. Defendants moved pursuant to, inter alia, Fed. R. Crim. P. 33 for a new trial on the

6 ground that the testimony of Swiatek did not meet the Fed. R. Evid. 702 standard for the admission

7 of expert testimony. The motions were denied, see United States v. Romano, 859 F.Supp.2d 445, 450

8 (E.D.N.Y. 2012); and after a Fatico hearing, defendants were sentenced as indicated above.

9 II. CHALLENGES TO THE CONVICTIONS

10 On appeal, defendants contend principally that they are entitled to a new trial on the

11 ground that the district court should have excluded the testimonies of Montgomery and Swiatek

12 pursuant to Rule 702 as unsuitable for expert testimony, and should have excluded GX 109 on

13 Confrontation Clause grounds. Romano also contends that he was denied due process because he

14 "received no notice that the lawfulness of his conduct depended on whether his grades were similar

15 to those assigned by NGC" (Romano brief on appeal at 76); and Kearney contends that the evidence

16 was insufficient to show that he knowingly and willingly participated in a conspiracy. For the reasons

17 that follow, we find no merit in defendants' challenges to their convictions.

18
 1 A. Admission of the Testimony of Montgomery and Swiatek

2 Prior to trial, defendants moved for an in limine ruling precluding the government from

3 calling Montgomery, Swiatek, and Schechter as witnesses and from giving expert testimony as to the

4 grading and valuation of coins and the Sheldon scale. Defendants contended that the coin grading and

5 valuation processes are subjective and unreliable and that the proposed testimony was inadmissible

6 under Rules 403 and 702 of the Federal Rules of Evidence. On appeal, they seek a new trial, arguing

7 that the district court wrongly denied the Rule 702 motion as to Montgomery and Swiatek and that

8 their testimony was based on insufficiently reliable methods that were unreliably applied to the facts

9 of this case. We disagree.

10 Rule 702, as it read during the pretrial and trial proceedings in this case, provided that

11 [i]f scientific, technical, or other specialized knowledge will assist the trier of
12 fact to understand the evidence or to determine a fact in issue, a witness
13 qualified as an expert by knowledge, skill, experience, training, or education,
14 may testify thereto in the form of an opinion or otherwise.

15 Fed. R. Evid. 702. "The fact that the subject matter is not 'scientific' is no bar to admissibility of

16 expert testimony." United States v. Kayne, 90 F.3d 7, 11 (1st Cir. 1996), cert. denied, 519 U.S. 1055

17 (1997). Rule 702 itself refers to "other specialized knowledge," Fed. R. Evid. 702, and "there are

18 many different kinds of experts, and many different kinds of expertise"; and in some "cases, the

19 relevant reliability concerns may focus upon personal knowledge or experience," Kumho Tire Co. v.

20 Carmichael, 526 U.S. 137, 150 (1999).

21 Rule 702 imposes on the trial judge an obligation to determine whether the expert's

22 specialized knowledge will assist the trier of fact, i.e., will be not only relevant, but reliable. See, e.g.,

23 id. at 147, 152; Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 589 (1993). Daubert,

24 which dealt with a scientific theory, stated that factors the district court could consider included the

19
 1 theory's testability, the extent to which it "has been subjected to peer review and publication," the

2 extent to which a technique is subject to "standards controlling the technique's operation," the "known

3 or potential rate of error," and the "degree of acceptance" within the "relevant scientific community."

4 509 U.S. at 593-94 (internal quotation marks omitted). But "the Rule 702 inquiry [i]s 'a flexible one,'"

5 and "Daubert makes clear that the factors it mentions do not constitute a 'definitive checklist or test.'"

6 Kumho, 526 U.S. at 150 (quoting Daubert, 509 U.S. at 594, 593 (emphasis in Kumho)). "And

7 Daubert adds that the gatekeeping inquiry must be 'tied to the facts' of a particular 'case.'" Kumho,

8 526 U.S. at 150 (quoting Daubert, 509 U.S. at 591).

9 "[A] court of appeals is to apply an abuse-of-discretion standard when it 'review[s] a

10 trial court's decision to admit or exclude expert testimony.'" Kumho, 526 U.S. at 152 (quoting

11 General Electric Co. v. Joiner, 522 U.S. 136, 138-39 (1997)). "That standard applies as much to the

12 trial court's decisions about how to determine reliability as to its ultimate conclusion." Kumho, 526

13 U.S. at 152. "[W]hether Daubert's specific factors are, or are not, reasonable measures of reliability

14 in a particular case is a matter that the law grants the trial judge broad latitude to determine." Kumho,

15 526 U.S. at 153; see, e.g., General Electric Co. v. Joiner, 522 U.S. at 143.

16 In the present case the district court held a hearing on defendants' in limine motions

17 ("Daubert hearing") and heard testimony from Montgomery and Swiatek. Their testimony at that

18 hearing as to, inter alia, their credentials in the rare coin industry, the industry's use of the Sheldon

19 coin grading scale, NGC's process for grading coins, and Swiatek's process in valuing the coins sold

20 by the Romano Companies, was essentially the same as their trial testimony, described in part I.A.3.

21 above. (See, e.g., Hearing Transcript, April 11, 2011 ("Daubert Hr'g"), at 20-27, 29-49, 116-30.)

20
 1 In cross-examining Montgomery at the Daubert hearing, defendants brought out, inter

2 alia, that NGC's website stated that discrepancies in grading of one to two points in coins at MS 60

3 and higher are not uncommon. Montgomery responded that he viewed two-point discrepancies as

4 "scarce," occurring only "occasionally." (Daubert Hr'g 105.) He testified that a four-point difference

5 between graders with respect to a particular coin is "quite extreme," "[t]hree points is also extreme,"

6 a two-point difference happens "once in a blue moon," and one-point differences are "not . . .

7 improbable at all," which is why services use consensus grading. (Id. at 49.)

8 Defendants pointed to descriptions by ANA and PCGS that differed as to the

9 characteristics of an MS 64 grade coin. Montgomery responded that although the ANA and PCGS

10 descriptions were not identical they were not inconsistent. (See id. at 70-71, 111.) Defendants also

11 pointed out that an article published by Coin World in 2003 reported disparities in the grades assigned

12 by different grading services to 15 coins the magazine had submitted for grading. Montgomery

13 responded that coin grading has become increasingly more refined since 2003, and that even the

14 author of that 2003 article viewed a sample of 15 coins as too small to be meaningful. (See id.

15 at 106-07.)

16 In cross-examining Swiatek, defendants brought out, inter alia, that he was generally

17 unaware of Romano's business practices and did not know whether the Romano Companies sold coins

18 wholesale or retail, and that he did not use full retail prices in estimating the value of the coins

19 Romano sold. (See Daubert Hr'g 131-32, 144-45, 147-48.) Defendants also sought to bring out that

20 the methods used by Swiatek, which included his reliance on his own personal experiences, could not

21 be scientifically replicated.

21
 1 After hearing the testimony and receiving briefing, the district court denied defendants'

2 motion to preclude the testimony of Montgomery, Swiatek, and Schechter, stating principally as

3 follows:

4 I find it certainly relevant under the rules of evidence. And I have conducted
5 a 403 balancing, as [Romano's trial counsel] urged the court to do.

6 And, in addition, I find, in terms of the Daubert challenge, that the
7 government has demonstrated by a preponderance of the evidence that the
8 facts and data and the principles and methods relied upon by both Mr.
9 Montgomery and Mr. Swiatek were sufficient under Daubert to allow the
10 testimony to be admitted.

11 . . . . The issue that [Romano's counsel] has raised both in the hearing
12 and in his submissions, these disparities as to grading, go to the weight of the
13 evidence and not to its admissibility.

14 And given the evidence that I heard at the hearing . . . with respect to
15 [the] opinion of Mr. Swiatek, again, the failure to assign a range of values, the
16 use of the blue sheet versus the gray sheet, wholesale versus retail, I believe
17 all of those issues go to the weight to be given his testimony, not to its
18 admissibility.

19 ....

20 . . . [A]ll the issues that have been raised can be the proper subject of
21 cross-examination, but I think to just submit gray sheets and blue sheets to a
22 jury as business records would not be sufficient to have them understand how
23 the industry works and how the grading plays into the ultimate evaluation of
24 the coins within the industry. So I'm going to allow them to testify.

25 (Hearing Transcript, May 4, 2011 ("2011 Ruling"), at 4-6 (emphases added).)

26 In its posttrial ruling denying defendants' motion for a new trial based on the admission

27 of the Swiatek testimony, the district court adhered to the views stated in the 2011 Ruling. The court

28 stated that despite the undisputed fact that both grading and valuation have a "subjective component,"

29 both Montgomery and Swiatek at the Daubert hearing had "demonstrated that the system for coin

22
 1 grading and valuation rests on well-established industry standards, publications and market trends,

2 which are sufficiently reliable to form the basis of expert testimony." 859 F.Supp.2d at 458.

3 Having noted that a new trial may be granted pursuant to Fed. R. Crim. P. 33 if the

4 interest of justice so requires, see 859 F.Supp.2d at 455-56, and that "'[t]he ultimate test on a Rule 33

5 motion is whether letting a guilty verdict stand would be a manifest injustice,'" id. at 456 (quoting

6 United States v. Ferguson, 246 F.3d 129, 134 (2d Cir. 2001)), the court found no basis for a new trial,

7 see 859 F.Supp.2d at 467.

8 We see no abuse of discretion in either the district court's decision to admit the

9 testimony of the experts or its decision that the admission of their testimony does not warrant a new

10 trial. Although we have not directly addressed the need for expert testimony with respect to alleged

11 fraudulent conduct in the rare coin industry, we think it clearly was appropriate for the court to find

12 that such testimony was relevant and would be of assistance to the jury, for we agree with the First

13 Circuit that "[o]ne could hardly expect a lay jury to form conclusions about such an esoteric subject

14 as the value of rare coins without the help of experts," United States v. Kayne, 90 F.3d at 12; see

15 generally United States v. Kail, 804 F.2d 441, 448 (8th Cir. 1986) (approving admission of expert

16 testimony that "ANA grading standards enjoyed almost industry-wide acceptance, and that the gray

17 sheet is recognized as the chief guide to the current price of rare coins"). See also United States v.

18 Numisgroup Int'l Corp., 170 F.Supp.2d 340, 350 (E.D.N.Y. 2001) ("[t]he grade of a coin is the

19 predominant factor in a consumer's decision to purchase a coin; it is what the customer is bargaining

20 for," and "coin grading involves various factors, including some that are subjective, that are beyond

21 the knowledge and experience of the unsophisticated purchaser" (emphasis added)), aff'd, 368 F.3d

22 880, 880 (2d Cir. 2004) (affirming the denial of motions for acquittal "for substantially the reasons

23
 1 stated by the District Court in its thorough decision"), rev'd on other grounds sub nom. Dupurton v.

2 United States, 543 U.S. 1098 (2005).

3 In the context of this industry, it was also well within the bounds of the district court's

4 discretion to conclude that the experts' testimony was sufficient to show that the grading of coins

5 provides a reliable indicator of their worth. The testimony in the record--including that of Schechter,

6 which defendants do not challenge on appeal--showed that the 70-point Sheldon scale has been in

7 fairly common use since the middle of the 20th century and that grading using that scale is relied on

8 in the industry. Both NGC and PCGS, which together account for 90-94 percent of the independent

9 grading service industry, use the Sheldon scale, and they give guarantees for the grades they assign

10 to coins. And many purchases of coins are made on the basis of the certifications by NGC and PCGS

11 without the buyers' viewing the coins themselves. Further, there was evidence that the Romano

12 Companies' salesmen were instructed by Kearney to tell customers who expressed concern about

13 quality and value that the company had in-house graders who used a grading process similar to that

14 used by NGC and PCGS--thereby confirming the general acceptance of those processes as reliable

15 indicators of value.

16 While defendants argue that expert testimony is inherently unreliable because of the

17 subjective aspects of coin grading, the lack of unanimity as to the grade a particular coin should be

18 given is not a basis for a conclusion that use of the Sheldon scale is unreliable or leads to wild

19 disparities. Defendants' own trial witness's study of the government's spreadsheets revealed that when

20 NGC graded 1,671 coins sold by the Romano Companies, NGC's first two graders agreed on the

21 proper grade for 90 percent of the coins.

24
 1 Defendants also argue that Swiatek's testimony should have been excluded because

2 his methodology was incapable of replication and hence was untrustworthy. But at the Daubert

3 hearing, Swiatek explained how he used the Greysheet, the Bluesheet, and auction prices, as well as

4 his own experience, to value the coins that were examined in the present action. Although defendants

5 highlight some portions of the trial transcript that indicate less-than-clear descriptions of Swiatek's

6 methods (see Kearney brief on appeal at 35-37), Swiatek's trial testimony as a whole described his

7 methods coherently and was consistent with his testimony at the Daubert hearing. To be sure, it is

8 possible that Swiatek's methods are not entirely replicable because they are based in part on his

9 personal experience as a coin dealer for several decades; but Rule 702 itself provides that the court

10 may admit evidence that will assist the jury based on the witness's specialized knowledge. We see

11 no abuse of discretion in the district court's conclusion that the criticisms raised by defendants went

12 to the weight of the evidence, not to its admissibility, and were matters for cross-examination and

13 argument to the jury.

14 Finally, although defendants argue that Swiatek's valuations should have been excluded

15 on the theory that it was absurd to believe that Romano would pay his suppliers three times as much

16 as Swiatek opined the coins were worth, that theory too was a matter for argument to the jury, not a

17 basis for excluding Swiatek's evidence. As discussed in Part I.A.2. above, suppliers testified that

18 Romano gave little or no attention to the coins themselves before buying them. And the jury was

19 entitled to infer that, as Romano was driven by a profit motive, the price he paid for a coin was of no

20 concern to him as long as he could market it for substantially more.

21 We conclude that the district court did not abuse its discretion in admitting the

22 testimony of Montgomery and Swiatek or in denying a new trial.

25
 1 B. Other Challenges to the Convictions

2 Defendants also raise other challenges to their convictions, including a Confrontation

3 Clause claim, a due process claim, and an insufficiency claim. These challenges do not require

4 extended discussion.

5 1. The Present Confrontation Clause Challenge to GX 109

6 Defendants argue that the admission of GX 109, the spreadsheet setting out the grades

7 that NGC assigned to coins submitted in this case by customers of the Romano Companies, violated

8 their Confrontation Clause rights because the NGC graders "were not themselves witnesses at trial

9 and . . . therefore were not subject to cross-examination" (Romano brief on appeal at 74). "The right

10 to confrontation may, of course, be waived," Melendez-Diaz v. Massachusetts, 557 U.S. 305, 314 n.3

11 (2009), and we conclude that it was waived here.

12 Defendants acknowledge that they did not make a Confrontation Clause objection to

13 the district court (see Romano brief on appeal at 72), but they argue that they are entitled to plain-

14 error review because they only "explicitly disclaimed a hearsay argument," and "did not disclaim the

15 [Confrontation Clause] argument" (id. at 73 n.31). The record does not support their argument, for

16 "a strategic decision[ that is] evidenced not merely by silence but by a negative response on the record

17 to a district court invitation to voice objection, does more than forfeit the unraised objection; it waives

18 it," United States v. Agrawal, 726 F.3d 235, 259 (2d Cir. 2013) (emphasis added).

19 In the district court, defendants objected to the introduction of GX 109 on the basis of

20 their contention that coin grading is not "an appropriate subject for expert testimony," characterizing

21 the exhibit as "a compendium of expert opinion." (Tr. 1949.) Although defendants did expressly

26
 1 agree that they were "not objecting to this as a hearsay document," they also answered "No" to the

2 court's far more elemental question: whether defendants were objecting to GX 109 on the ground

3 "that each individual grader needs to come in." (Id. at 1950.) Their "No" answer to that question

4 waived their right to have the individual graders themselves come in to testify as witnesses.

5 2. The Due Process Challenge

6 Romano's argument that he was denied due process because he had no notice that he

7 could be convicted of conspiracy to commit mail or wire fraud merely for selling coins whose grades

8 as he represented them "were [not] similar to those assigned by NGC" (Romano brief on appeal at 76)

9 is merely a repackaging of his objection to the admission of expert testimony about the grading

10 standards in the rare coin industry, and is meritless. The jury was properly instructed that

11 [t]he government does not allege that the defendants engaged in
12 criminal conduct merely because they allegedly charged more for certain coins
13 than their actual value in the opinion of an expert witness [and that] even if
14 you should determine that in some cases the amount the subject companies
15 charged customers for coins exceeded Mr. Swiatek's value for those coins, you
16 may not convict the defendant you are considering of the charges contained in
17 the indictment unless all of the elements of the crimes charged have been
18 proven beyond a reasonable doubt.

19 (Tr. 2619.)

20 3. Sufficiency of Evidence Supporting Kearney's Convictions

21 Kearney argues that the evidence was insufficient to show that he was a knowing

22 participant in the conspiracy to commit mail and wire fraud because, he asserts, there was no evidence

23 that he knew Romano was grading coins fraudulently. And he argues that since that conspiracy

24 charge was the basis for the charge of money laundering conspiracy, the evidence was insufficient to

27
 1 support his conviction on either count. These arguments are meritless. There was ample evidence

2 that Kearney knew of Romano's fraudulent misgrading; and given the many other respects in which

3 the Romano Companies defrauded their customers, there was abundant evidence that Kearney

4 participated in the mail and wire fraud conspiracy even if he had not known the coins were misgraded.

5 "[S]olicitation of a purchase by means of false representations" constitutes fraud when

6 such representations are "directed to the quality, adequacy or price of goods to be sold . . . ." United

7 States v. Regent Office Supply Co., 421 F.2d 1174, 1179 (2d Cir. 1970). A defendant's fraudulent

8 intent may be proven entirely through circumstantial evidence. See United States v. Guadagna, 183

9 F.3d 122, 129 (2d Cir. 1999). "When the necessary result of the [defendant]'s scheme is to injure

10 others, fraudulent intent may be inferred from the scheme itself." United States v. D'Amato, 39 F.3d

11 1249, 1257 (2d Cir. 1994) (internal quotation marks omitted).

12 In considering a challenge to the sufficiency of the evidence to support a conviction

13 "we view the evidence in the light most favorable to the government, drawing all inferences in the

14 government's favor and deferring to the jury's assessment of the witnesses' credibility." United States

15 v. Sabhnani, 599 F.3d 215, 241 (2d Cir. 2010) (internal quotation marks omitted), cert. denied, 131

16 S. Ct. 1000 (2011). Deference to the jury's findings "is especially important when reviewing a

17 conviction for conspiracy . . . because a conspiracy by its very nature is a secretive operation." United

18 States v. Jackson, 335 F.3d 170, 180 (2d Cir. 2003) (internal quotation marks omitted). A conviction

19 will be upheld so long as, with the evidence at trial viewed in the light most favorable to the

20 government, "any rational trier of fact could have found the essential elements of the crime beyond

21 a reasonable doubt." Jackson v. Virginia, 443 U.S. 307, 319 (1979) (emphasis in original).

28
 1 Here, the record shows that customers of the Romano Companies were defrauded in

2 several ways in addition to the misgrading of the coins they bought. As described in Part I.A.2.

3 above, the salesmen fraudulently told customers and prospective customers that the coins would

4 increase in value; that sets of coins were worth more than the total value of the individual coins in the

5 set; that the Romano Companies employed in-house coin graders; and that the Romano Companies

6 would help the customers sell the coins they bought if they so desired. All of the salesmen were

7 trained by Kearney; he directed them to make all of these fraudulent statements. He also, having

8 trained them to say the Romano Companies would help customers sell their coins if they so requested,

9 trained the salesmen to deflect any such request.

10 Further, Kearney himself made sales calls and included in his sales pitch the same

11 material misrepresentations that he required the salesmen to make--and he even embellished on some

12 of them. Kearney told one customer, for example, that WSRC had in-house graders and "that each

13 grader was independent of the others. They had all worked for grading companies and had come to

14 work at Wall Street Rare Coins now, and that they looked at the coins independently and graded them

15 independently, and that the lowest grade was chosen to represent the coin." (Tr. 1354.)

16 Second, there was ample circumstantial evidence from which the jury could infer that

17 Kearney knew Romano was engaging in fraudulent grading. Salesmen testified that all of the

18 salesmen were aware of Romano's method of selecting grades for the coins in inventory; he would

19 call Lepolszki (or one other salesman) to learn the book prices for various grades of a certain coin,

20 and Romano's ensuing inventory sheet would show that coin at the high end of the prices Lepolszki

21 had reported. The salesmen were aware of this process because they could hear Lepolszki's responses

29
 1 to Romano's grade/price inquiries, as they all worked in a small office at desks merely five or six feet

2 apart. Kearney's desk was one of them, and he was there 90 percent of the time.

3 Further, in addition to Kearney's making and orchestrating the false representations

4 about having well-qualified in-house graders--which the jury was entitled to infer constituted a

5 knowing effort to conceal false grading--customers' complaints that their coins were not of as high

6 a grade as the Romano Companies represented them to be were referred to Romano and to Kearney.

7 One of Kearney's solutions was to instruct that grades not be shown on invoices.

8 The evidence was more than sufficient to permit the jury to find that Kearney knew

9 of the false grading of the coins.

10 III. SENTENCING CHALLENGES

11 The presentence report ("PSR") prepared on Romano calculated that his total offense

12 level was 39 which, with his criminal history category of I, resulted in an advisory Guidelines

13 recommended range of 262-327 months' imprisonment. Among the components of the PSR

14 calculation were offense-level enhancements of six steps on the ground that the offenses involved

15 1,453 victims, see Guidelines § 2B1.1(b)(2)(C) (the increase for 250 or more victims) (see Part III.A.

16 below), and 20 steps for the losses of more than $7 million but not more than $20 million caused by

17 the offenses, see id. § 2B1.1(b)(1)(K). Noting the difficulty of calculating actual loss, the PSRs for

18 both Romano and Kearney used gain from the offenses in calculating loss, see id. § 2B1.1 Application

19 Note 3(B). The PSR found Romano responsible for the total Romano Companies gains, or

30
 1 $11,711,583. The Probation Department recommended for him a below-Guidelines prison term of

2 180 months.

3 The PSR prepared on Kearney calculated that his total offense level was 36, which,

4 given his criminal history category of I, resulted in an advisory Guidelines recommended sentencing

5 range of 188-235 months. The components of this calculation included the same six-step offense-

6 level enhancement for 250 or more victims ("multiple-victim enhancement"), and it applied an 18-step

7 offense-level enhancement for losses of more than $2.5 million but not more than $7 million, see

8 § 2B1.1(b)(1)(J), on the basis that Kearney was accountable for $4,995,958 in gains from the offenses.

9 The Probation Department recommended for Kearney a below-Guidelines prison term of 120 months.

10 Defendants objected to the PSR findings as to, inter alia, the amount of loss,

11 contending that the court should consider only the losses of the 10 victims who testified at trial. They

12 also argued that there should have been no multiple-victim enhancement because the Guidelines

13 define victims as persons who suffered an actual loss from the offenses, and the court found that it

14 could not determine actual loss.

15 The district court held a Fatico hearing. The sole live witness was Hessle, the former

16 postal inspector, who testified, largely as he had at trial, that he had obtained the Romano Companies'

17 bank records for the period of 1999 to 2008 and that the records for that period showed the companies

18 had generated $32 million in revenue from 1,453 people who purchased coins. Hessle stated that he

19 considered all of the 1,453 coin purchasers to be potential victims and had attempted to contact them,

20 but many had died or had moved. Hessle also testified that it would have been possible to estimate

21 the loss amount using the $32-million revenue figure and the evidence presented at trial as to the

22 actual value of the coins sold by the Romano Companies compared to their sales price. He stated that

31
 1 this calculation was complicated, however, by fluctuations in the prices of gold and silver, and that

2 "[i]t was much fairer" to use the defendants' gain for the loss amount. (Fatico Hearing Transcript,

3 November 22, 2013, at 35-36.) The parties also submitted parts of various trial witnesses' testimony.

4 At defendants' respective sentencing hearings, the district court rejected their

5 challenges to the PSRs' Guidelines calculations. (See Romano Sentencing Transcript, February 27,

6 2014 ("Romano S.Tr."), at 20-23; Kearney Sentencing Transcript, March 10, 2014 ("Kearney S.Tr."),

7 at 23.) The court adopted the findings in each defendant's PSR, including those with respect to the

8 use of gain to measure the amount of loss and with respect to the number of victims of the offenses.

9 (See Romano S.Tr. 20, 21, 23, 25; Kearney S.Tr. 23.)

10 In imposing sentence on Romano, the court stated principally as follows:

11 The guideline range is 262 to 327 months. I have read the submissions
12 and heard oral arguments by counsel, and I think a sentence that is sufficient
13 but not greater than necessary to meet the aims of the statute is the bottom end
14 of the guideline. I'm going to sentence the defendant to the custody of the
15 attorney general . . . for a period of 262 months . . . .

16 (Romano S.Tr. at 23.) The government requested clarification as to whether the court intended to split

17 the 262-month term of imprisonment between Romano's two convictions, indicating that each carried

18 a 20-year statutory maximum sentence. After some discussion with respect to the applicable statutory

19 maxima, the court stated that "to simplify it, I'll make it 240 months to run concurrently" (id. at 24).

20 In imposing sentence on Kearney, the court stated principally that in arriving at a

21 "sentence that is sufficient, but not greater than necessary to meet the aims of the statute," the court

22 determined, in a downward departure from the Guidelines-recommended range, that Kearney's term

23 of imprisonment should be 156 months. (Kearney S.Tr. 26.)

32
 1 Defendants contend that the sentences imposed on them are unreasonable, arguing

2 principally that the court improperly applied the multiple-victim enhancement without making the

3 prerequisite determination of actual loss and that the prison terms are unduly long. Defendants also

4 contend that the district court erred in failing to review de novo the recommendations of the

5 magistrate judge for restitution and forfeiture. We find merit only in this last contention.

6 A. The Multiple-Victim Enhancement

7 Defendants, in arguing that the district court erred in increasing their offense levels on

8 account of the number of victims involved in their offenses, contend that there were no victims within

9 the meaning of the relevant guideline definition, because the district court did not determine actual

10 loss and instead calculated defendants' offense levels with respect to their respective gains. We

11 disagree.

12 For fraud offenses, the Guidelines provide that the defendant's offense level is to be

13 calculated based in part on the amount of loss the offense of conviction either caused or was intended

14 to cause, see Guidelines § 2B1.1, whichever is greater, see id. Application Note 3(A). After providing

15 a base offense level in subsection (a), § 2B1.1 sets out in subsection (b)(1) a loss table that prescribes

16 offense-level increases pegged to the amount of loss. Application Note 3 to § 2B1.1, which begins

17 by stating that it "applies to the determination of loss under subsection (b)(1)," provides that "[t]he

18 court need only make a reasonable estimate of the loss . . . . based on available information," id.

19 Application Note 3(C), and that "[t]he court shall use the gain that resulted from the offense as an

20 alternative measure of loss only if there is a loss but it reasonably cannot be determined," id.

33
 1 Application Note 3(B) (emphasis added). When the court has so used gain, the defendant's offense

2 level is increased for the calculated amount in accordance with the loss table.

3 The Guidelines prescribe further increases in a defendant's offense level that depend

4 on the number of victims his offense involved. See Guidelines § 2B1.1(b)(2). "Victim," in

5 connection with this subsection, is defined, in pertinent part, as "any person who sustained any part

6 of the actual loss determined under subsection (b)(1)." Id. § 2B1.1 Application Note 1.

7 Given this definition of "[v]ictim" in terms of an "actual loss determined," and the

8 statement in Application Note 3(B) that reliance on "gain" is authorized only if "loss . . . reasonably

9 cannot be determined," defendants contend that a district court's use of gain for purposes of

10 § 2B1.1(b)(1) precludes a finding that there were any victims of the defendant's offense. (See, e.g.,

11 Kearney brief on appeal at 51.) In support of this position, defendants principally cite United States

12 v. Skys, 637 F.3d 146 (2d Cir. 2011) ("Skys"), for the proposition that a district court cannot properly

13 "make any determination regarding the number of victims without first making a determination

14 concerning actual loss under § 2B1.1(b)(1)." (Kearney brief on appeal at 51.) Defendants' reliance

15 on Skys is misplaced, and their argument is unpersuasive.

16 First, Skys is inapposite because it did not involve a calculation of gain. It instead

17 involved the difference between actual loss and intended loss. The defendant there had caused actual

18 losses to some individual investors and had unsuccessfully sought to perpetrate more massive frauds

19 on financial institutions. The court did not make any determination as to the amounts actually lost

20 by the defendant's investors. Because the Guidelines define loss as "the greater of actual loss or

21 intended loss," Guidelines § 2B1.1 Application Note 3(A), the district court had simply applied the

22 step on the § 2B1.1(b)(1) loss scale that was applicable to the massive losses that were intended.

34
 1 Because there had been no determination of actual loss, we concluded that the multiple-victim

2 enhancement was inappropriate, and we remanded for, inter alia, a determination of the amounts of

3 actual loss suffered by individuals defrauded by the defendant.

4 In the present case, the issue is not actual loss as contrasted with intended loss.

5 Defendants' mail and wire fraud conspiracy succeeded in costing customers of the Romano

6 Companies millions of dollars. Any intended but unrealized loss was not a factor in the sentencing

7 of Romano or Kearney. And we interpret the authorization for a sentencing court to use gain for "the

8 determination of loss under subsection (b)(1)," Guidelines § 2B1.1 Application Note 3, to be premised

9 solely on actual loss: Gain can be resorted to "only if there is a loss." Id. Application Note 3(B)

10 (emphasis added).

11 Second, we reject defendants' contention that the definition of "[v]ictim" as a "person

12 who sustained any part of the actual loss determined under subsection (b)(1)" means that there cannot

13 be a victim of the offense where the court has applied subsection (b)(1) after calculating the

14 defendant's "gain" because of an inability to "calculate the customers' actual loss" (Kearney brief on

15 appeal at 50). We interpret the Guidelines' stated precondition for consideration of gain--i.e., that

16 "there is a loss but it reasonably cannot be determined," Guidelines § 2B1.1 Application Note 3(B)

17 (emphasis added)--as hypothesizing (a) an actual loss that (b) cannot reasonably be determined as to

18 amount. And by the terms of the Application Note itself, where "there is a loss," the use of gain is

19 an "alternative measure of [the] loss." Id. (emphases added). Thus, in using that alternative measure

20 of the loss that actually occurred, and applying to it the § 2B1.1(b)(1) loss table, the sentencing court

21 has, within the meaning of the Guidelines, made an actual loss determination under subsection (b)(1).

35
 1 B. Other Reasonableness Challenges

2 Defendants' contentions that their sentences are substantively unreasonable do not

3 require extended discussion. Where there is

4 no procedural unreasonableness in sentencing, our review of a claim of
5 substantive unreasonableness is narrow, akin to review for abuse of discretion
6 . . . . In conducting such review, we evaluat[e] the length of the sentence
7 imposed in light of the factors enumerated under 18 U.S.C. § 3553(a).

8 United States v. Norman, 776 F.3d 67, 85-86 (2d Cir.) (internal quotation marks omitted), cert.

9 denied, 135 S. Ct. 2333 (2015). We do not substitute our own judgment for that of the district judge,

10 nor will we "second guess the weight (or lack thereof) that the judge accorded to a given factor or to

11 a specific argument made pursuant to that factor." United States v. Perez-Frias, 636 F.3d 39, 42 (2d

12 Cir. 2011) (internal quotation marks omitted). We will "set aside a district court's substantive

13 determination only in exceptional cases where the trial court's decision cannot be located within the

14 range of permissible decisions." United States v. Cavera, 550 F.3d 180, 189 (2d Cir. 2008) (en banc)

15 (emphasis in Cavera) (internal quotation marks omitted).

16 At their respective sentencing hearings, counsel for Romano or Kearney called the

17 court's attention to, inter alia, their objections to the PSRs' loss calculations and recommendations that

18 the court look to defendants' gain, their objections to the recommended offense-level increases

19 because the offenses targeted vulnerable victims, and their objections that the PSRs' calculations as

20 to the amount of loss and number of victims created an overlap that inflated their offense levels.

21 Counsel also argued for leniency in light of the defendants' respective ages, personal histories, and

22 family circumstances. We see no indication in the record that the district court failed to consider these

23 contentions and representations. The court also gave appropriate consideration to the record of

24 defendants' participation in a massive, long-running fraudulent scheme that specifically targeted a

36
 1 vulnerable class of victims and to the letters submitted by victims (or their relatives or heirs) of

2 defendants' offenses. We cannot say that the prison term imposed on either defendant constituted an

3 abuse of discretion.

4 Defendants also argue that the district court "did not adequately consider the § 3553(a)

5 sentencing factors" (Romano brief on appeal at 85) and did not adequately state its reasons for the

6 selected sentences in open court (see id. at 90). As defendants did not voice these objections in the

7 district court, they are reviewable only for plain error. See, e.g., United States v. Verkhoglyad, 516

8 F.3d 122, 127-28 (2d Cir. 2008); United States v. Villafuerte, 502 F.3d 204, 207-08 (2d Cir. 2007);

9 United States v. Carter, 489 F.3d 528, 540 (2d Cir. 2007), cert. denied, 552 U.S. 1144 (2008). We

10 see no such error here.

11 In the absence of record evidence suggesting otherwise, we presume that a sentencing

12 judge has considered the relevant statutory factors. As discussed above, the parties had submitted

13 objections to the respective PSRs, and the record shows that the district court considered their

14 objections. The court expressly adopted the findings in the respective PSRs. It heard arguments and

15 pleas for leniency from counsel at each sentencing hearing; and, stating that it was imposing the

16 sentence it believed was sufficient, but not greater than necessary, to meet the aims of the sentencing

17 statute, the court proceeded to impose sentences that were below the Guidelines-recommended ranges.

18 Even assuming that the above proceedings contained any error, we cannot see that defendants'

19 substantial rights were affected.

37
1 C. Restitution and Forfeiture

2 Defendants contend that the district court erred in not conducting a de novo review of

3 the recommendations of the magistrate judge that they be ordered to pay $9,139,727.10 in restitution

4 and to forfeit $32,220,617. Kearney also argues that a property in East Islip, New York, "was

5 acquired with a mix of funds, only some of which were derived from the alleged fraud," and should

6 not have been subject to total forfeiture (Kearney brief on appeal at 53). The government concedes

7 that these matters should be the subject of a remand.

8 If a party timely objects to any portion of a magistrate judge's report and

9 recommendation, the district court must "make a de novo determination of those portions of the report

10 or specified proposed findings or recommendations to which objection is made." 28 U.S.C.

11 § 636(b)(1). We normally "presume that the district court has made [such] a de novo review unless

12 affirmative evidence indicates otherwise." Claude v. Peikes, 534 F.3d 801, 801 (2d Cir. 2008)

13 (internal quotation marks omitted).

14 Here, the district court had referred questions of restitution and forfeiture to a

15 magistrate judge for report and recommendation. The magistrate judge had issued a report with

16 respect to both defendants as to restitution and a report with respect to both defendants as to forfeiture.

17 There were timely defense objections to both reports. At the sentencing hearing of Romano, the

18 colloquy on restitution and forfeiture was as follows:

19 [THE COURT:] . . . . I will adopt the report and recommendation of
20 Magistrate Judge Scanlon with respect to restitution.

21 Was there one for forfeiture too?

22 [ASSISTANT U.S. ATTORNEY]: Yes, Your Honor.

23 THE COURT: I'll adopt that too.

38
 1 (Romano S.Tr. 23.) At the sentencing hearing of Kearney, those recommendations were discussed

2 as follows:

3 [ASSISTANT U.S. ATTORNEY]: . . . . I would . . . ask the court to
4 adopt the report and recommendation . . . that there should be a forfeiture
5 amount of $32,220,617 on --

6 THE COURT: I have signed the forfeiture order."

7 (Kearney S.Tr. 25.)

8 This record, with the court's question at Romano's hearing as to the existence of a

9 recommendation for forfeiture, and the court's immediate adoption of the recommendation, effectively

10 eliminates the presumption that the court gave the recommendation proper consideration. And in the

11 absence of any discussion of defendants' objections to either report, it also raises a question as to

12 whether the court gave de novo consideration to the magistrate judge's other recommendation.

13 Accordingly, we remand for the district court to give de novo consideration to the

14 magistrate judge's recommendations for restitution and forfeiture, including Kearney's challenge with

15 respect to the East Islip property. Defendants ask that the matter be remanded to a different district

16 judge. We are not persuaded that the record warrants such action.

17 CONCLUSION

18 We have considered all of defendants' contentions on these appeals and, except to the

19 extent indicated above, have found them to be without merit. The judgments of the district court are

20 affirmed in all respects except that the matter is remanded as indicated above for de novo review of

21 the magistrate judge's reports and recommendations concerning restitution and forfeiture.

39