Amicus Curiae Brief — Allentown MacK Sales & Service, Inc. v. NLRB

Supreme Court brief1998

Ask Donna

What actually matters in this document.

Text

Supreme Court, U.S.

FILED

(4) APR 17 97

CLERK

No. 96-795

IN THE

Supreme Court of the United States

OCTOBER TERM, 1996

ALLENTOWN MACK SALES AND SERVICE, INC.,

Petitioner,

Vv.

NATIONAL LABOR RELATIONS BOARD,

Respondent.

On Writ of Certiorari to the United States Court of

Appeals for the District of Columbia Circuit

BRIEF OF AMICUS CURIAE THE CHAMBER OF

COMMERCE OF THE UNITED STATES OF

AMERICA IN SUPPORT OF PETITIONER

STEPHEN A. BOKAT JOHN S. IRVING

ROBIN S. CONRAD Counsel of Record

MONA C. ZEIBERG CHRISTOPHER LANDAU

NATIONAL CHAMBER JEFFREY BOSSERT CLARK

LITIGATION CENTER, INC. KIRKLAND & ELLIS

1615 H Street, N.W. 655 Fifteenth Street, N.W.

Washington, D.C. 20062 Washington, D.C. 20005

(202) 463-5337 (202) 879-5000

Counsel for Amicus Curiae

April 17, 1997

TABLE OF CONTENTS

INTEREST OF AMICUS CURIAE

STATEMENT OF THE CASE

SUMMARY OF ARGUMENT

THE BOARD’S POLLING STANDARD

IS ARBITRARY AND CAPRICIOUS

A. The Board’s Standard Renders Polling Useless,

and Thereby Makes It Effectively Impossible

for an Employer to Establish “Good-Faith

Doubt” Regarding Employee Union Sympathies . . 6

. The Board’s Polling Standard Makes it Harder

for an Employer Lawfully to Conduct a Poll

than to Question Individual Employees

. The Board’s Justifications for its Polling

1. The Board’s Polling Standard Does Not

Establish a “Safe Harbor” for Employers ... 15

The Board’s Polling Standard Is Not

Necessary to Protect the Utility of the

Formal Election Mechanism

(ii)

TABLE OF AUTHORITIES

_ CASES:

3. The Board’s Polling Standard Does Not Page

Promote “Workplace Stability” in a 44 Liquormart, Inc. v. Rhode Island,

Permissible Manner ................... 18 5 25

Il. THE BOARD’S POLLING STANDARD LACKS A.L. Gilbert Ce.,

ANY STATUTORY BASIS .................... 20 110 N.L.R.B. 2067 (1954) .................... x

A. The Act Does Not Authorize the Board to Allegheny Ludlum Corp. v. NLRB,

Proscribe Non-Coercive Polling ............. 20 104 F.3d 1354 (D.C. Cir. 1997) ............ 14, 24

B. Any Interpretation of the Act that Would American Ship Bidg. Co. v. NLRB,

Authorize the Board to Proscribe Non- 380 US. 300 (1965) ............0........... 22

Coercive Polling Would, at the Very Least,

Raise Serious First Amendment Questions ecese 23 Apparatus Serv., Inc.,

296 NLRB. 581 (1989) .....................

GED 8 adorn cds ca pee mee rari + becheaise 25 Pe .

Architectural Glass & Metal Co., Inc.,

107 F.3d 426 (6th Cir. 1997) ................. 13

Auciello Iron Works, Inc. v. NLRB,

Po e 6,7

Blue Flash Express, Inc.,

109N.L.RB. 591 (1954) ................., 13

Bourne v. NLRB,

332 F.2d 47 (2d Cir. 1964) (per curiam) ........ 13

Bryan Mem 'l Hosp.,

279 N.L.R.B. 222 (1986),

enf'd, 814 F.2d 1259 (8th Cir.),

cert. denied, 484 U.S. 849 (1987) ...........__. 8

(iii)

(iv)

Cannon Elec. Co.,

151 N.L.R.B. 1465 (1965),

overruled on other grounds,

Resistance Tech., 280 N.L.R.B. 1004 (1986) ..... 24

Celanese Corp.,

tt) ) Perr eee 7

Chevron, U.S.A., Inc. v. Natural Resources

Defense Council, Inc.,

SEP Se I ah bok celcns ba cd iGec cee 21

Curtiss-Wright Corp. v. Schoonejongen,

eee eee 2

Destileria Serralles, Inc.,

289 N.L.R.B. 51 (1988),

enf d, 882 F.2d 19 (ist Cir. 1989) ............ 7-8

EA. Labs., Inc.,

80 N.L.R._B. 625 (1948),

enf‘d as modified, 188 F 2d 885 (2d Cir.),

cert. denied, 342 U.S. 871 (1951) .............. 7

Edenfield v. Fane,

507 U.S. 761 (1993)... 0... ec e cece ccceee. 24

Edward J. DeBartolo Corp. v. Florida Gulf Coast

Bldg. and Constr. Trades Council,

Garment Workers v. NLRB,

(v)

General Mercantile & Hardware Co. v. NLRB,

461 F.2d 952 (8th Cir. 1972) ................. 22

Hajoca Corp.,

291 N.L.R.B. 104 (1988),

enf'd, 872 F.2d 1169 (3d Cir. 1989) ............. 9

Hohn Indus.,

283 NLRB. 71 (1987) ...................... 9

Joy Silk Mills, Inc.,

85 N.L.R.B. 1263 (1949),

enf'd as modified, 185 F.2d 732 (D.C. Cir. 1950),

cert. denied, 341 U.S.914(1951) .............. 7

Lechmere, Inc. v. NLRB,

502 U.S. $27 (1992) ................... 2, 19, 21

Litton Fin. Printing Div. v. NLRB,

CC 2

Lou's Produce, Inc.,

308 N.L.R.B. 1194 (1992),

enfd, 21 F.2d 1114 (9th Cir. 1994) ............. 9

Mingtree Restaurant, Inc. v. NLRB,

736 F.2d 1295 (9th Cir. 1984) ........ 9, 10, 13, 15

Montgomery Ward & Co.,

SIO NLRB. 717 (1974) ..... 2... o eee 9

(vi)

Motor Vehicle Mfrs. Ass'n v. State Farm

Mut. Auto. Ins. Co.,

463 U.S. 29 (1983) ...........--0005- 5, 6, 10, 19

National Ass'n of Regulatory Util. Comm'rs v. ICC,

41 F.3d 721 (D.C. Cir. 1994) ..........-06-66- 23

New Associates,

307 N.L.R_B. 1131 (1992),

enf. denied, 35 F.3d 828 (3d Cir. 1994) .......... K

NLRB v. A.J. Tower Co.,

329 U.S. 324 (1946) ... 6. eee cee eee eee 19-20

NLRB v. Albany Steel, Inc.,

17 F.3d 564 (2d Cir. 1994) ..... 2... - eee ee eens 18

NLRB v. A.W. Thompson, Inc.

651 F.2d 1141 (Sth Cir. 1981) ........ 9, 10, 13, 15

NLRB v. Associated Dry Goods Corp..,

209 F.2d 593 (2d Cir. 1954) ............--545- 22

NLRB v. Brookshire Grocery Co..,

919 F.2d 359 (Sth Cir. 1990) .............-5.. 13

NLRB v. Dan River Mills, Inc.,

274 F.2d 381 (Sth Cir. 1960) ........--.--0-5- 15

NLRB v. Drivers,

362 U.S. 274 (1960) ........ , covesessosedess 25

(vii)

NLRB v. Financial Inst. Emp’'ees,

Sob. Side d5 Sh dodddidleccce 22

NLRB v. Gissel Packing Co.,

Ee 7,23

NLRB v. Golub Corp.,

See £20 928 C6 Cir, 1967) .. 0. cc cceccccees 25

NLRB v. Health Care & Retirement Corp.,

a 21

NLRB v. Intertherm, Inc.,

596 F.2d 267 (8th Cir. 1979) ........... 13-14, 25

NLRB v. Miami Coca-Cola Bottling Co.,

382 F.2d 921 (Sth Cir. 1967) ................. 22

NLRB v. National Licorice Co.,

104 F.2d 655 (2d Cir. 1939),

aff'd, 309 U.S. 350 (1940) ................... 18

NLRB v. Protein Blenders, Inc.,

215 F.2d 749 (8th Cir. 1954) ................. 21

NLRB v. Remington Rand, Inc.,

94 F.2d 862 (2d Cir.),

cert. denied, 304 U.S. 576 (1938) .............. 7

NLRB v. Superior Fireproof Door & Sash Co.,

289 F.2d 713 (2d Cir. 1961) .................. 18

NLRB v. Tennessee Coach Co.,

191 F.2d 546 (6th Cir. 1951) ................. 22

NLRB v. Virginia Elec. & Power Co.,

ET re 23, 25

(viii)

Phoenix Pipe & Tube, L.P.,

302 N.L.RB. 122,

enf'd, 955 F.2d 852 (3d Cir. 1991) .............. 7

Riley v. National Fed’n of the Blind of N.C, Inc.,

SS 24

Robeson Cutlery Co.,

GP Pee SON CPOE occ ccccccccccccccecs 7

Rossmore House,

269 N.L.R.B. 1176 (1984),

enf'd, 760 F.2d 1006 (9th Cir.1985)............ 14

Sanco Piece Dye Works, Inc.,

SE NLRB. GOS) 20s ccc ccc cece ccctcee 7

Sax v. NLRB, -

171 F.2d 769 (7th Cir. 1948) ................. 22

SEC v. Chenery Corp.,

PEED cbse Sicsoccsecccecsoos 12, 16

Standard-Coosa-Thatcher Co. ,

EEL. cs cccscovesctoedses 24

Stormor, Inc.,

268 N.L.R.B. 860 (1984) .................. 6, 12

Struksnes Constr. Co.,

165 NLRB. 1062 (1967) ............... 8,9, 24

Taft Broadcasting,

201 NLRB. 801 (1973) ...... 000... e eee. »

Teamsters Local 633 v. NLRB,

509 F.2d 490 (D.C. Cir. 1974) ................ 14

(ix)

Texas Petrochemicals Corp.,

296 N.L.R.B. 1057 (1989),

enfd in part on other grounds,

923 F.2d 398 (Sth Cir. 1991) ............. passim

Thomas Indus., Inc. v. NLRB,

687 F.2d 863 (6th Cir. 1982) ......... 9, 10, 13, 15

Thomas v. Collins,

po Ee eee eee 23, 25

United States v. Long,

905 F.2d 1572 (D.C. Cir.),

cert. denied, 498 U.S. 948 (1990) ............. 24

Upper Miss. Towing Corp.,

246 NLRB. 262 (1979) ............000.0005. 8

Varity Corp. v. Howe,

116 S. Ce. 1065 (1996) .....................-- 2

Vic Koenig Chevrolet, Inc.,

321 N.L.R.B. No. 168 (Aug. 27, 1996) .......... %

Village of Schaumburg v. Citizens for a Better

Environment,

Wayside Press v. NLRB,

206 F.2d 862 (9th Cir. 1953) ................. 22

Westbrook Bowl,

293 N.LRB. 1000 (1989) .................... 8

Windham Community Mem’! Hosp.,

230 N.L.R.B. 1070 (1977),

enf d, 577 F.2d 805 (2d Cir. 1978) .............. 7

(x)

STATUTES:

SUSE. BD 6 ooh ores V5 RIFKIN Wie cie 5

SPD bv-00 vnc sincuetedcs dene 3, 5, 20, 21

SP GB GEE ccc cceviece ded Ste iv beteewes 23

BPE BD bedside cccccccctdetei AWW 17

ened vovedncedeGayeceusunens 18

SP OEE ¢ 08 cicwhtehessvsetel civ edede cn}: 6

ity aemys oeeus sebesenedecctedens 6

REGULATION:

FP CEs BURIED ov cccciccccedectvewswerees 12

MISCELLANEOUS:

Flynn, The Costs and Benefits of “Hiding the Ball”:

NLRB Policymaking and the Failure of

Judicial Review, 75 B.U. L. Rev. 387 (1995) .. 10-11

1 P. Hardin, The Developing Labor Law

EE SE Oe ee 17

Note, Employer Postcertification Polls to Determine

Union Support, 84 Mich. L. Rev. 1770 (1986) . 12, 14

IN THE

Supreme Court of the United States

OCTOBER TERM, 1996

No. 96-795

ALLENTOWN MACK SALES AND SERVICE, INC.,

Petitioner,

Vv.

NATIONAL LABOR RELATIONS BOARD,

Respondent.

On Writ of Certiorari to the United States Court of

Appeals for the District of Columbia Circuit

BRIEF OF AMICUS CURIAE THE CHAMBER OF

COMMERCE OF THE UNITED STATES OF

AMERICA IN SUPPORT OF PETITIONER

INTEREST OF AMICUS CURIAE

The Chamber of Commerce of the United States of

America is the largest federation of business companies and

associations in the world. With substantial membership in each

of the fifty States, the Chamber represents an underlying

membership of more than three million businesses and

professional i organizations of every size, in every sector of

business, and in every region of the country. The Chamber

thus serves as the principal voice of the American business

community.

One of the Chamber’s principal functions is to represent

the interests of its members in important matters before the

courts of the United States, and especially this Court. The

Chamber has sought to fulfill that function by filing briefs

amicus curiae in cases of importance to the business

community, including cases (like this one) involving labor

relations.

This case presents an issue of particular legal and practical

importance: when may an employer lawfully conduct a poll of

its employees to ascertain whether an incumbent union

continues to enjoy majority support? The National Labor

Relations Board has announced and applied a standard

whereby an employer cannot conduct such a poll unless it first

knows the result. That standard essentially bans employer-

sponsored polls and makes it practically impossible for an

employer to challenge an incumbent union’s presumption of

majority support. The Chamber’s members have a vital

interest in defending their right to rebut this presumption and

in clarifying this now-confused area of the law.

This brief is filed with the consent of both parties, and

letters reflecting such consent have been filed with the Clerk of

this Court.

STATEMENT OF THE CASE

Petitioner Allentown Mack Sales and Service, Inc., is a

company established in December 1990 by three former

managers of the Mack Truck dealership in Allentown,

Pennsylvania. See Pet. App. at 2a. The new company hired 32

of its predecessor’s 45 employees. See id ll of these

employees had previously been represented by a union, Local

Lodge #724, Intersational Association of Machinists and

Aerospace Workers, AFL-CIO. See id. at 2a, 40a.

' See, e.g., Varity Corp. v. Howe, 116 S. Ct. 1065 (1996); Curtiss-Wright

Corp. v. Schoonejongen, 115 S. Ct. 1223 (1995); Lechmere, Inc. v. NLRB,

502 U.S. 527 (1992); Litton Fin. Printing Div. v. NLRB, 501 U.S. 190

(1991); NLRB v. Curtin Matheson Scientific, Inc., 494 U.S. 775 (1990); Fall

River Dyeing & Finishing Corp. v. NLRB, 482 U.S. 27 (1987).

seven of petitioner’s employees indicated to management that

they no longer supported the union. One of these employees

stated that the “entire night shift” (consisting of five or six

employees) did not want the union. See id at 14a, 5la

Another employee (a union shop steward) stated that the

employees did not need a union. See id at 14a. Yet another

employee, who was a member of the union bargaining

committee and the shop steward for the 23-member service

department, stated that “‘with a new company, if a vote was

taken, the union would lose and that it was his feeling that the

employees did not want a union.” Jd.

Based on these statements, petitioner postponed

recognition of the union “until further investigation.” Jd. at

30a. It then polled its employees to determine whether a

majority supported the union. See id. at 44a, 58a. The poll was

supervised by a Roman Catholic priest, and complied fully

with the procedural safeguards established by the National

Labor Relations Board. See id. at 2a, 58a-60a.

The employees rejected the union by a 19-to-13 margin.

See id at 2a, 14a, 44a. Accordingly, petitioner withdrew

recognition from the union. See id. at 2a. The union responded

by filing unfair labor practices charges with the Board. See id.

The Board held that the poll violated § 8(a\(1) of the

National Labor Relations Act, 29 U.S.C. § 158(a)(1). See id

at 26a n.9. The Board did not contest the poll’s fairness or

accuracy, see id. at 25a-26a, 58a-60a, but held instead that it

violated the Act because petitioner did not have sufficient

evidence of loss of union support to justify a poll in the first

place, see id. at 25a-26a. The Board based this conclusion on

its precedents holding that the same standard governs an

employer’s ability to poll its employees as to withdraw

recognition from an incumbent union. To satisfy that standard

in either context, according to the Board, an employer must

have a good-faith doubt regarding the union’s majority status.

See id. at 25a-26a & n.9 (citing Texas Petrochemicals Corp.,

296 N.L.R.B. 1057, 1058-63 (1989)). Because a majority of

petitioner’s employees had not specifically informed

management that they opposed the union, the Board held that

petitioner lacked such doubt. See Pet. App. at 21a-26a. The

Board further asserted that “because the poll itself was an

unfair labor practice, [the employer] could not lawfully rely on

the results of the poll in declining to recognize the [u]nion.”

Id. at 27a. Accordingly, the Board held that petitioner’s

withdrawal of recognition was also unlawful and ordered the

company to bargain with the very union rejected by a majority

of its employees. See id. at 27a, 62a-63a.

A divided panel of the U.S. Court of Appeals for the D.C.

Circuit enforced the Board’s order. See id at la-18a. The

panel majority expressly rejected the decisions of three other

Courts of Appeals, all of which had refused to enforce the

Board’s polling standard. See id at 3a-8a. The dissenting

judge noted that the Board’s standard led to the “bizarre result”

that “an employer cannot conduct a poll to determine majority

support unless it already has so much evidence of no majority

support as to render the poll meaningless.” Jd. at 15a.

SUMMARY OF ARGUMENT

This case highlights an inherent conflict among several

strands of federal labor law developed by the National Labor

Relations Board. First, the Board has long acknowiedged an

employer’s right to poll its employees in a non-coercive

manner to determine their level of union support. Second, for

more than half a century, the Board has recognized an

employer’s right to withdraw recognition from an incumbent

union on the basis of a “good-faith doubt” regarding the

union’s majority status. But third, and at issue here, the Board

in recent years has held that an employer cannot lawfully

conduct a poll unless a majority of its employees have first

expressed a specific desire to repudiate an incumbent union.

An employer, in other words, is free to poll its employees, but

only when it already knows the results of that poll. As the

dissenting judge noted below, that result gives new meaning to

Dickens’ phrase “the law is a ass.” The Board’s contradictory

matrix of rules governing workplace polling is arbitrary and

capricious, and hence invalid under basic principles of

administrative law.

Even if the Board’s polling standard could survive routine

“arbitrary and capricious” review, moreover, the Board lacks

the statutory authority to restrict non-coercive polling. The

Board in this case invoked its power under § 8(a)(1) of the

National Labor Relations Act, which makes it an unfair labor

practice for an employer to “interfere with, restrain, or coerce

employees” in the exercise of their statutory right to self-

organization. 29 U.S.C. § 158(a)(1). But a poll (like the one

involved here) that is concededly non-coercive under the

Board’s own rigorous standards cannot be proscribed as an

unfair labor practice for the simple reason that it does not

“interfere with, restrain, or coerce” the exercise of any

employee rights. Indeed, any construction of the Act that

would restrict employers from polling their employees in a

non-coercive manner would, at the very least, raise First

Amendment problems of the first order.

ARGUMENT

I. THE BOARD’S POLLING STANDARD IS

ARBITRARY AND CAPRICIOUS.

The Board order under review in this case must be set

aside, first and foremost, because it violates the bedrock

capricious” agency action. See 5 U.S.C. § 706(2)(A). To be

sure, “[t]he scope of review under the ‘arbitrary and capricious’

standard is narrow and a court is not to substitute its judgment

for that of the agency.” Motor Vehicle Mfrs. Ass'n v. State

Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983). But a

reviewing court need not, and must not, rubber-stamp agency

action. See id That is especially true where, as here, Congress

has required the courts to ascertain that challenged agency

action is “supported by substantial evidence on the record

considered as a whole.” 29 U.S.C. § 160(e), (f). The order iin

this case must be set aside because the Board’s confusing amd

contradictory matrix of rules governing workplace polling

simply defies reasoned analysis. See, e.g., State Farm, 463

US. at 42-44.

A. The Board’s Standard Renders Polling Useless,

and Thereby Makes It Effectively Impossible for

an Employer to Establish “Good-Faith Doubt”

Regarding Employee Union Sympathies.

The Board’s order concluding that petitioner committed an

unfair labor practice by polling its employees is fundamentally

inconsistent with longstanding precedent holding that an

employer enjoys the right (1) to poll its employees in a nom-

coercive manner to determine their level of union support, amd

(2) to withdraw recognition from an incumbent union on the

basis of a “good-faith doubt” regarding the union’s majority

status. To appreciate this inconsistency, it is necessary to

understand the complicated matrix of legal rules governing this

area.

For one year after certification or voluntary recognition

and up to three years after entering into a collective bargaining

agreement, a union typically enjoys an _ irrebuttable

presumption of majority support. See, e.g. Auciello Iron

Works, Inc. v. NLRB, 116 S. Ct. 1754, 1758 (1996); NLRB v.

Curtin Matheson Scientific, Inc., 494 U.S. 775, 777-78 (1990).

After that point, the presumption is subject to rebuttal; an

employer can overcome it by proving that either “(1) the union

did not in fact enjoy majority support, or (2) the employer had

a ‘good-faith’ doubt, founded on a sufficient objective basis, of

the union’s majority support.” Curtin Matheson, 494 U.S. at

778 (emphasis added). See also Stormor, Inc., 268 N.L.R.B.

860, 866-67 (1984). The employer’s good-faith doubt, in other

words, has long been recognized as a complete defense to an

unfair labor practice charge. See, e.g, Windham Community

Mem 'l Hosp., 230 N.L.R.B. 1070, 1073 (1977), enf'd, 577 F.2d

805 (2d Cir. 1978); Celanese Corp., 95 N.L.R.B. 664, 675

(1951); EA. Labs., Inc., 80 N.L.R.B. 625, 683 (1948), enf'd as

modified, 188 F.2d 885 (2d Cir.), cert. denied, 342 U.S. 871

(1951). “The right of an employer lawfully to refuse to bargain

if he had a good faith doubt as to the Union’s majority status,

even if in fact the Union did represent a majority, was

recognized early in the edministration of the Act.” NLRB v.

Gissel Packing Co., 395 U.S. 575, 597 n.11 (1969) (emphasis

added) (citing NLRB v. Remington Rand, Inc., 94 F.2d 862, 868

(2d Cir.) (L. Hand, J.), cert. denied, 304 U.S. 576 (1938)). See

also Joy Silk Mills, Inc., 85 N.L.R.B. 1263 (1949), enfd as

modified, 185 F.2d 732 (D.C. Cir. 1950), cert. denied, 341 U.S.

914 (1951); Robeson Cutlery Co., 67 N.L.R.B. 481, 481, 499-

500 (1946); Sanco Piece Dye Works, Inc., 38 N.L.R.B. 690,

710-11 (1942).

“Taan otek tee tite See, e.g., Garment Workers

union without delay as soon as it has a good-faith doubt

regarding majority status. See, e.g, Auciello, 116 S. Ct. at

1760-61.

It is far from easy for an employer to rebut the

to repudiate the union. See, e.g., Phoenix Pipe & Tube, L.P.,

302 N.L.R.B. 122, 122, enf'd, 955 F.2d 852 (3d Cir. 1991);

Destileria Serralles, Inc., 289 N.L.R.B. 51, 51-52 (1988),

enf'd, 882 F.2d 19 (ist Cir. 1989); Bryan Mem'l Hosp., 279

N.L.R.B. 222, 225 (1986), enfd, $14 F.2d 1259 (8th Cir.),

cert. denied, 484 U.S. 849 (1987). The test is exceedingly

demanding: indeed, the Board has held that the fact that a

majority of employees responded negatively to the question of

whether they “wish[ed] to remain in the Union” did not

“express the unequivocal repudiation of 2 union which the

Board requires to serve as the basis for a good-faith doubt of

majority status.” Vic Koenig Chevrolet, Inc., 321 N.L.R.B. No.

168, at 6 (Aug. 27, 1996). Anti-union sentiments expressed

during a job interview, general employee statements of

dissatisfaction with the quality of union representation, or one

employee’s report of other employees’ antipathy toward a

union typically are insufficient. See, e.g, Apparatus Serv.,

Inc., 296 N.L.R.B. 581, 583 (1989) (job interview); Upper

Miss. Towing Corp., 246 N.L.R.B. 262, 273-74 (1979) (general

expressions); Westbrook Bowl, 293 N.L.R.B. 1000, 1001 n.11

(1989) (employees speaking for other employees). See also

New Associates, 307 N.L.R.B. 1131, 1136 (1992) (employee-

initiated decertification petition not sufficient to raise good-

faith doubt regarding union’s majority status), enf. denied, 35

F.3d 828 (3d Cir. 1994).

Because it is all but impossible as a practical matter

(especially in large bargaining units) for a majority of

employees to approach an employer to express a “specific

desire” to repudiate an incumbent union, the Board has long

recognized the validity of polling as an alternate means for an

employer to ascertain the level of union support among its

employees. See, e.g., Taft Broadcasting, 201 N.L.R.B. 801,

803 (1973), AL. Gilbert Co., 110 N.L.R.B. 2067, 2072 (1954).

Over time, the Board has developed an array of procedural

safeguards designed to ensure that any such poll is conducted

in a fair and non-coercive manner. See Struksnes Constr. Co.,

165 N.L.R.B. 1062 (1967). A poll is lawful, the Board has

held, if (1) its purpose is to determine the validity of a union’s

claim of majority status, (2) that purpose is communicated to

the employees, (3) assurances against reprisals are given, (4)

the employees are polled by secret ballot, and (5) the employer

has not engaged in unfair labor practices or otherwise created

a coercive atmosphere. See id at 1063. Although Siruksnes

involved a pre-certification poll, the Board has applied its

procedural safeguards to post-certification polls as well. See,

e.g., Lou's Produce, Inc., 308 N.L.R.B. 1194, 1204 (1992),

enf'd, 21 F.2d 1114 (9th Cir. 1994); Hajoca Corp., 291

N.L.R.B. 104, 1121-3 (1988), enf'd, 872 F.2d 1169 (3d Cir.

1989); Hohn Indus., 283 N.L.R.B. 71, 71 n.2 (1987).

The dispute in this case arises from the Board’s further

holding that an employer seeking to poll its employees after

not only comply with the Struksnes safeguards but also must

show “a reasonable basis for believing that the union has lost

its majority [status] since its certification.” Montgomery Ward

& Co., 210 N.L.RB. 717, 717 (1974). Because that standard

is, by the Board’s admission, precisely the same standard that

governs an employer's ability to withdraw recognition from an

incumbent union, the Montgomery Ward polling standard

created. the anomaly at issue here: an employer cannot poll its

employees unless it already knows the results of that poll.

Over the years following Montgomery Ward, three federal

courts of appeals refused to enforce orders applying the

Board’s stringent polling standard. See, eg, Mingtree

Restaurant, Inc. v. NLRB, 736 F.2d 1295, 1297 (9th Cir. 1984);

Thomas Indus., Inc. v. NLRB, 687 F.2d 863, 867 (6th Cir.

1982); NLRB v. A.W. Thompson, Inc., 651 F.2d 1141, 1144

(Sth Cir. 1981). In light of these judicial rebukes, the Board in

1989 reconsidered Montgomery Ward. See Texas

Petrochemicals Corp., 296 N.L.R.B. 1057 (1989), enfd in part

on other grounds, 923 F.2d 398 (Sth Cir. 1991). By a divided

vote, however, the Board decided to reaffirm Montgomery

Ward, see 296 N.L.R.B. at 1059-63; id. at 1065-66 (Chairman

Stephens, concurring), while at the same time reaffirming an

employer’s “right” to poll its employees and disclaiming any

desire to “do away with such polls,” id at 1061. The Board

proceeded to apply the Montgomery Ward/Texas

Petrochemicals standard in this case to hold petitioner guilty of

an unfair labor practice, see Pet. App. at 20a-27a, 42a-60a, and

a divided panel of the D.C. Circuit enforced that order, see id

at 3a-12a, 13a-18a (Sentelle, J., dissenting).

As the Fifth, Sixth, and Ninth Circuits have recognized,

the Board’s polling standard gives rise to two basic anomalies.

First, the standard effectively outlaws polling sub silentio by

allowing an employer to conduct a poll only when it already

has enough information to render any such poll superfluous.

See, e.g., Mingtree Restaurant, 736 F.2d at 1297 (“[T]he

Montgomery Ward rule is tantamount to an outright prohibition

of employer-sponsored polls.”). The point of a poll, after all,

is to allow an employer to ascertain whether it might be

necessary or appropriate to withdraw recognition from an

incumbent union. It makes no sense, thus, to allow polling in

circumstances where an employer would already be justified in

withdrawing recognition. “[Tjhe Board’s position [is]

untenable [because] an employer would only be allowed to take

a poll under circumstances where no poll was necessary.”

Thomas Indus., 687 F 2d at 867. See also Mingtree Restaurant,

736 F.2d at 1297; A.W. Thompson, 651 F.2d at 1144.

Even assuming that the Board had the statutory and

constitutional authority to prevent an employer from

conducting a non-coercive poll of its employees—an

assumption that amicus challenges in Part Il, infra—the Board

may not do so by subterfuge. An administrative agency

cannot, by semantic sleight-of-hand, outlaw a practice that it

purports to permit. If the Board wishes to change course, it

must at the very least do so forthrightly—and thrs

acknowledge and accept the legal and political consequences

of its actions. See, e.g. State Farm, 463 U.S. at 41-42; cf

Flynn, The Costs and Benefits of “Hiding the Ball”: NLRB

Policymaking and the Failure of Judicial Review, 75 B.U.

L. Rev. 387, 393-404 (1995).

Second, the Board’s polling standard effectively abolishes

the venerable “good-faith doubt” standard sub silentio by

making it all but impossible for an employer to satisfy that

standard. If an employer cannot poll its employees, the only

way it can lawfully ascertain the level of union support is to

wait for a majority of the employees to express a specific and

unequivocal desire to repudiate an incumbent union. But that

requires the employer to wait until it knows that a union does

not in fact command majority support; it leaves no room

whatsoever for a “good-faith doubt.”

This anomaly did not escape notice during this Court’s

most recent foray into this area in Curtin Matheson. The Chief

Justice bighlighted the inherent tension between Board

decisions “requir{ing] an employer to show that individual

employees have ‘expressed desires’ to repudiate the incumbent

union in order to establish a reasonable doubt of the union’s

majority status,” and the Board’s Texas Petrochemicals rule

“prevent[ing] the employer from polling its employees unless

it first establishes a good-faith doubt of majority status.” 494

U.S. at 797 (concurring opinion). “I have considerable doubt

whether the Board may insist that good-faith doubt be

determined only on the basis of sentiments of individual

employees, and at the same time bar the employer from using

what might be the only effective means of determining those

sentiments.” Jd Justice Blackmun emphasized the same

anomaly, observing that it places an employer in “a difficult

bind.” Jd at 799 & n.3 (dissenting opinion). That is something

of an understatement: in fact, the Board’s polling standard

effectively destroys an employer’s right to withdraw

recognition from a minority union. Indeed, the standard places

employers squarely in a “Catch-22”: they are guilty of an

unfair labor practice if they bargain with a minority union, see,

e.g., Garment Workers, 366 U.S. at 736, but they also are

guilty of an unfair labor practice if they attempt to ascertain

whether an incumbent union has lost majority support.’

Whatever the Board’s power to eliminate its longstanding

“good-faith doubt” rule (a question that remains open, see

Curtin Matheson, 494 U.S. at 788 n8; id at 800-01

(Blackmun, J., dissenting)), it plainly cannot do so by stealth.’

The upshot of these anomalies, of course, is not that the

Board must allow unrestricted polling to rebut the presumption

of majority support. Rather, the point is only that the Board

? Although the Board and the court below attempted to deny this

contradiction by asserting that an employer may safely rely on the

presumption of continued majority support, see Texas Petrochemicals, 296

N.L.R.B. at 1062; Pet. App. at 7a, their assertion is hollow. An employer

is statutorily forbidden from recognizing a minority union regardless of a

good-faith basis for doing so. See Garment Workers, 366 U.S. at 736. The

Board’s assertion, moreover, ignores the destruction of the employer’s

statutory right (the corollary of its statutory duty) to withdraw recognition

from a minority union.

For the same reasons, the Board cannot defend its polling standard on

the ground that employees themselves have the right to petition for a Board-

conducted election. See Texas Petrochemicals, 296 N.L.R.B. at 1062. Such

petitions are not only cumbersome in practice—they must be signed by at

least 30% of the employees, see 29 CF.R. § 101.18(a)-(b)—but also have

nothing to do with an employer’s right and duty to withdraw recognition

from a minority union. See general’y Note, Employer Postcertification

Polls to Determine Union Support, 84 Mich. L. Rev. 1770, 1781-82 (1986).

> The AFL-CIO has previously urged this Court to reject the “good-faith

doubt” standard altogether and to hold that an employer can withdraw

recognition from an incumbent union only by showing an actual loss of

majority status following a Board-conducted election. See Curtin

Matheson, 494 U.S. at 788 n.8. That aggressive argument is not only

misguided on the merits, but—most pertinent here—misdirected to this

Court. The Board has not remotely purported to abolish the longstanding

“good-faith doubt” rule. See id (citing Stormor, Inc., 268 N.L.R.B. 860,

866-67 (1984)). The validity of that rule is not before the Court in this case.

Under fundamental principles of administrative law, agency action may be

upheld only on the basis relied upon by the agency. See, e.g, SEC v.

Chenery Corp., 318 U.S. 80, 94-95 (1943).

cannot—without even purporting to ban polling or eliminate

the good-faith doubt rule—set the same standard for polling as

for withdrawing recognition from an incumbent union. As the

Fifth, Sixth, and Ninth Circuits have sensibly recognized, the

most stringent feasible polling standard is one that allows an

employer to conduct a poll on the basis of “substantial,

objective evidence [that] the union has lost support,” even if

such evidence would be insufficient by itself to justify

withdrawal of recognition. A.W. Thompson, 651 F.2d at 1145.

See also Mingtree Restaurant, 736 F.2d at 1299; Thomas

Indus., 687 F.2d at 867. The Board’s current polling standard

cannot survive even deferential “arbitrary and capricious”

review.

B. The Board’s Polling Standard Makes it Harder for

an Employer Lawfully to Conduct a Poll than to

Question Individual Employees.

In addition to the anomalies noted above, the Board’s

polling standard also leads to the curious result that it is harder

for an employer lawfully to conduct a non-coercive poll of ail

employees than to question individual employees about their

union sympathies. Under longstanding Board precedent, an

employer has the right to “interrogate” individual employees as

long as the questioning is not coercive. See, e.g., Blue Flash

Express, Inc., 109 N.L.R.B. 591 (1954). The coercion inquiry

is governed by a multi-factor totality-of-the circumstances

analysis. See id at 594 (relevant factors include “[t}he time,

the place, the personnel involved, the information sought and

the employer’s conceded preference”). Under that standard (or

analogous standards applied by the courts of appeals), the

questioning of individual employees “is not held to be an unfair

labor practice unless it meets certain fairly severe”

requirements. Bourne v. NLRB, 332 F.2d 47, 48 (2d Cir. 1964)

(per curiam). See also Architectural Glass & Metal Co., Inc.,

107 F.3d 426, 434 (6th Cir. 1997); NLRB v. Brookshire

Grocery Co., 919 F.2d 359, 366-68 (Sth Cir. 1990); NLRB v.

Intertherm, Inc., 596 F.2d 267, 277-78 (8th Cir. 1979);

Teamsters Local 633 v. NLRB, 599 F.2d 490, 493-95 (D.C. Cir.

1974); cf Rossmore House, 269 N.L.R.B. 1176, 1178 n.20

(1984) (applying Bourne coercion standard), enf’d, 760 F.2d

1006 (9th Cir.1985).

Needless to say, it makes no sense for the Board to limit an

employer’s right to poll ai] employees more stringently than its

right to question individual employees. If anything, employees

need more protection when they are questioned one-on-one

than when they vote en masse in a non-coercive poll. As noted

above, see supra at 8-9, employer-sponsored polls are subject

to a host of procedural protections, including secret balloting.

No such categorical protections apply to individual

questioning; there are—as yet—no Miranda warnings in this

context. Indeed, the juxtaposition of the Board’s de facto ban

on polling with its flexible treatment of individualized

interrogation is all the more startling “because the Board itself

has recognized that a properly conducted secret ballot poll of

an entire bargaining unit is inherently less coercive than the

direct questioning of an individual employee.” Note, Employer

Postcertification Polls to Determine Union Support, 84 Mich.

L. Rev. 1770, 1776 (1986) (citing Struksnes Constr. Co., 165

N.L.R.B. 1062, 1063 (1967), and Rossmore House, 269

N.L.R.B. 1176 (1984)).

This confusion is exacerbated by the Board’s failure to

explain “where or why [it] draws the line between ‘polling’ and

other types of ‘interrogation’ in individual fact scenarios.”

Allegheny Ludlum Corp. v. NLRB, 104 F.3d 1354, 1359 (D.C.

Cir. 1997). An employer cannot conduct a de facto poll by

(und 1 D polet, bod rev ~ = ; vid = —

become a poll. “[T]he Board has rarely articulated its rationale

for choosing ‘polling’ analysis over ‘interrogation’ analysis in

particular decisions, and some of its choices between the two

defy classification.” Jd at 1360. Accordingly, it is virtually

impossible for an employer lawfully to ascertain whether a

majority of its employees continue to support an incumbent

union. The Board’s conflicting decisions in this area leave

employers mired in a “Serbonian bog.” NLRB v. Dan River

Mills, Inc., 274 F.2d 381, 388 (Sth Cir. 1960).

C. The Board’s Justifications for its Polling Standard

Are Unpersuasive.

The Board has attempted to justify its stringent polling

standard by asserting that the standard (1) does not render

polling useless, see Texas Petrochemicals, 296 N.L.R.B. at

1063, (2) prevents the anomaly of having a lower standard for

polling than for Board-conducted elections, see id. at 1060-61,

and (3) “promote[s] industrial and workplace stability” by

protecting incumbent unions, id at 1061. None of these

asserted justifications, however, can save the standard.

1. The Board’s Polling Standard Does Not

Establish a “Safe Harbor” for Employers.

As an initial matter, the Board rejects the criticism that its

polling standard effectively bans employer-sponsored polls by

rendering them useless. See id at 1063 (challenging reasoning

in Mingtree Restaurant, 736 F 2d at 1297-98, Thomas Indus.,

687 F.2d at 867, and A.W. Thompson, 651 F.2d at 1144).

Importing the stringent standard for withdrawal of recognition

into the polling context, the Board declares, “provides an

employer with a clear choice”: it may either (1) withdraw

recognition or (2) conduct a poll. Texas Petrochemicals, 296

N.L.RB. at 1063. A poll remains a viable option, according to

the Board, because it provides an employer with greater

certainty regarding the level of union support among

employees. “Rather than simply withdraw recognition from a

union that might still in fact have majority support, an

employer may wish first to poll its employees to obtain more

certain, precise information about the union’s support than is

provided by its own reasonable doubt. The employer can then

act with confidence and certainty in light of the results of the

poll.” Jd The Board thus suggests that, even under its

stringent standard, a poll operates as a sort of “safe harbor” for

an employer uncertain about the propriety of withdrawing

recognition from an incumbent union.

Any such “safe harbor,” however, is wholly illusory. A

poll is useful only as a means for an employer to determine

whether to withdraw recognition. As long as the standard for

polling and withdrawal of recognition is the same, a poll is

meaningless: if the employer would not have been justified in

withdrawing recognition, it would not have been justified in

taking the poll. The certainty added by the poll, in other

words, has no legal or practical effect. If an employer does not

know whether a union has lost majority support, it commits an

unfair labor practice either by conducting a poll or by

charged with unfair labor practices on both grounds.‘

2. The Board’s Polling Standard Is Not

Necessary to Protect the Utility of the Formal

Election Mechanism.

The Board also asserts that its stringent polling

standard—whatever anomalies it might create—is necessary to

prevent the further “anomalous” situation that would allegedly

arise if the standard for polling were lower than the standard

* The majority opinion below suggested that the Board could cure this

fundamental anomaly “by raising [its] withdrawal-of-recognition standard.”

Pet. App. at 6a. That suggestion, however, provides no basis for upholding

the challenged polling standard. As an initial matter, it is elementary that

administrative action can be upheld only on the grounds provided by the

agency. See, e.g., Chenery, 318 U.S. at 94-95. A court, in other words,

cannot substitute its own justifications. See id In any event, the possibility

that the Board might at some future date resolve this anomaly cannot justify

upholding the order issued against petitioner in this case, especially because

it is probably impossible either in theory or in practice for the Board to raise

any higher its standard for withdrawal of recognition.

for petitioning for a Board-conducted election. See id. at 1060.

Unless the standard for polling is “at least as stringent zs that

for [Board-conducted] RM elections,” the Board insists, an

employer would have no incentive to petition for an election

and the formal election mechanism would be undermined. Jd.’

The most obvious answer to that argument is that the

Board cannot justify one anomaly by conjuring up another.

The various standards at issue here are not immutable truths of

nature; they are simply a product of past Board decisions. An

administrative agency, of course, cannot possibly prevail in a

“battle of the anomalies” of its own creation. Indeed, the

Board could readily cure the anomaly that results from

applying the same standard to polling and withdrawal of

recognition without in turn setting different standards for

polling and RM elections: it could simply relax both the

standards for polling and for elections. Perhaps there is a

reason why the Board has refused to take that step, but, if so,

that reason has never been articulated.

In any event, it is not true that it would be anomalous to set

a lower standard for polling than for petitioning for an RM

election. The legal consequences of polls and elections differ

markedly. A poll is simply a mechanism for an employer to

determine whether it is necessary and appropriate to withdraw

recognition from an incumbent union. Following any such

withdrawal, the union would remain free to challenge the poll

and seek an election. An election, in contrast to a poll, is

conclusive. It requires the Board’s participation and carries not

only the imprimatur of Board certification but also (if the union

wins) a one-year irrebuttable presumption of majority status, or

* Employer-initiated elections are authorized by § %c\1)\(B) of the

National Labor Relations Act, 29 U.S.C. § 159(c\1)(B), and are often called

“RM elections” in light of the Board’s internal classification system. See,

eg, | P. Hardin, The Developing Labor Law 378, 383 (3d ed. 1992). They

entail a host of procedural requirements. See id. at 414-47.

(if the union loses) a one-year statutory bar to another election.

See 29 U.S.C. § 159(c\(3). It only stands to reason, thus, that

a higher standard should apply to trigger an RM election than

a poll.

Furthermore, even if the Board were correct that setting a

lower standard for polls than for RM elections would create a

disincentive for employers to petition for elections, that would

in no way undermine the utility of the formal election

mechanism. Unions and employees would always remain free

to challenge polls by petitioning for an election. Employers

themselves might prefer to request an RM election rather than

face a challenge to a poll. Indeed, even the courts have long

exercised authority to order elections to resolve disputes over

withdrawals of recognition, some of them resulting from

contested polls. See, e.g., NLRB v. Albany Steel, Inc., 17 F.3d

564, 568, 571-72 (2d Cir. 1994); Texas Petrochemicals v.

NLRB, 923 F.2d 398, 405-06 (Sth Cir. 1991); NLRB v. Superior

Fireproof Door & Sash Co., 289 F.2d 713, 723 (2d Cir. 1961);

NLRB v. National Licorice Co., 104 F.2d 655, 658 (2d Cir.

1939) (L. Hand, J.), aff'd, 309 U.S. 350 (1940). In the final

analysis, there is nothing wrong with an employer taking

“reasonable steps to verify union claims” of majority support

in order to “obviate a Board election.” Garment Workers, 366

US. at 739. Polling simply happens to be the most reliable and

efficient of such “reasonable steps.”

3. The Board’s Polling Standard Does Not

Promote “Workplace Stability” in a

Permissible Manner.

Finally, and most generally, the Board asserts that its

stringent polling standard is necessary “to promote industrial

and workplace stability.” Texas Petrochemicals, 296 N.L.R.B.

at 1061. “[PJolling employees about their continued support

for an incumbent union is itself potentially, if not inherently,

both disruptive of the collective-bargaining relationship

between an employer and a union and also unsettling to the

employees involved.” Jd.

Those assertions cannot validate the Board’s illogical

standard. Although the Board surely enjoys a degree of

Stateline tin einienitiin “tin lexities of industrial life,”

Lechmere, Inc. v. NLRB, 502 U.S. 527, 536 (1992) (internal

quotation omitted), such discretion is by no means unbounded.

Whatever the Board’s policy preferences, it may not—under

basic principles of administrative law—pursue them in a

manner that is arbitrary or contrary to the governing statute.

See, e.g., State Farm, 463 U.S. at 43-44.

Thus, even if the Board were correct that its polling

standard is likely to “promote industrial and workplace

stability” by making it difficult for an employer to challenge an

standard is preferable to an outright ban on polling. The issue

in this case, after all, is not the validity of polling—a practice

that the Board has not purported to outlaw—but rather the

validity of the current regime whereby an employer is free to

poll its employees, but only when it already knows the results

of the poll. That regime makes no sense, and cannot be

rationalized by platitudes about industrial peace.

The Board’s “stability” justification also proves too much

from a statutory point of view. Any rule that favors an

incumbent union can be said to “promote industrial and

workplace stability.” Indeed, that goal would best be promoted

by a rule imposing a perpetual irrebuttable presumption of

majority support. But such a rule would undoubtedly be

invalid in light of the emphasis on labor democracy in the

National Labor Relations Act. Employees should not be forced

to be represented by, and employers should not be forced to

bargain with, a minority union. See, e.g, Garment Workers,

366 U.S. at 737-38. The Act is squarely premised on the

principle of majority rule. See, e.g., NLRB v. A.J. Tower Co.,

329 U.S. 324, 331 (1946). A legitimate interest in “stability”

is not a license to ignore democracy.

Il. THE BOARD’S POLLING STANDARD LACKS ANY

STATUTORY BASIS.

A. The Act Does Not Authorize the Board to

Proscribe Non-Coercive Polling.

The Board order under review is not only arbitrary and

capricious; it is also ultra vires. The Board expressly invoked

its authority under § 8(a)(1) of the National Labor Relations

Act, 29 U.S.C. § 158(a)(1), which prohibits employers from

“interfer[ing] with, restrain[ing], or coerc[ing] employees” in

the exercise of their statutory right to self-organization. See

Pet. App. at 26a n.9. The Board did not, however, conclude

that the poll in this case involved any coercion or otherwise

impinged upon the rights of petitioner’s employees. Rather,

the Board simply asserted that any poll conducted before an

employer knows the outcome—regardless of how procedurally

proper and non-coercive—yper se amounted to an unfair labor

practice proscribed by § 8(aX(1). See Pet. App. at 25a-27a;

Texas Petrochemicals, 296 N.L.R.B. at 1057.

The Court of Appeals deferred to the Board’s polling

standard on the theory that “[nJothing in the National Labor

Relations Act specifically governs this practice.” Pet. App. at

7a. While acknowledging that the Board’s standard “has its

faults,” id. at 8a, the panel majority declared that “[t]he Board,

not the courts, has the primary responsibility for developing

and applying national labor policy,” id (internal quotation

omitted).

The Court of Appeals framed the issue precisely

backwards. An agency does not enjoy unfettered authority

absent any specific restriction in its governing statute; rather,

an agency enjoys only such authority as that statute specifically

confers. “[A]n agency literally has no power to act . . . unless

and until Congress confers power upon it.” Louisiana Public

Serv. Comm'n v. FCC, 476 U.S. 355, 374 (1986). Thus, the

Act’s silence on the issue of polling does not authorize the

Board to regulate the practice at will. The Board cannot deem

a poll to be an unfair labor practice without first identifying a

valid statutory source of authority. “[TJhe statute must control

the Board’s decision, not the other way around.” NLRB v.

Health Care & Retirement Corp., 114 S. Ct. 1778, 1784

(1994).

To be sure, the Board is entitled to judicial deference when

interpreting ambiguoc3 provisions of the Act. See, e.g,

Lechmere, 502 U.S. at 536; cf Chevron, U.S.A., Inc. v. Natural

Resources Defense Council, Inc., 467 U.S. 837, 842-43 (1984).

question at hand. If so, “that is the end of the matter,” because

both agencies and courts must implement statutory commands.

Chevron, 467 U.S. at 842. If the statute is ambiguous,

however, a court must defer to a reasonable agency

NLRB v. Protein Blenders, Inc., 215 F.2d 749, 750 (8th Cir.

1954). Before the Board can hold an employer guilty of an

unfair labor practice for conducting a poll (or otherwise

questioning its employees about their union sympathies), in

short, it must first establish that the employer’s actions in some

way infringed upon the employees’ right of self-organization.

See, e.g., General Mercantile & Hardware Co. v. NLRB, 461

F.2d 952, 954 (8th Cir. 1972); NLRB v. Miami Coca-Cola

Bottling Co., 382 F.2d 921, 924 (Sth Cir. 1967); NLRB v.

Associated Dry Goods Corp., 209 F.2d 593, 595 (2d Cir.

1954); Wayside Press v. NLRB, 206 F.2d 862, 864 (9th Cir.

1953); NLRB v. Tennessee Coach Co., 191 F.2d 546, 555 (6th

Cir. 1951); Sax v. NLRB, 171 F.2d 769, 772-73 (7th Cir. 1948).

The Board in this case failed to establish anything of the

kind. There is no question that petitioner’s poll satisfied the

Board’s rigorous Struksnes standards for procedural fairness.

It was conducted by none other than a priest. The

Administrative Law Judge specifically found that “there is

nothing in the evidence . . . to indicate that there was anything

; or coercive about the poll.” Pet. App. at 60a; see

also id. at 58a-59a. The Board itself in no way purported to

find that the poll “interfere(d] with, restrain(ed), or coerce[d]”

petitioner’s employees in exercising their right to self-

organization. Accordingly, the Board had no power in this

case to hold petitioner guilty of an unfair labor practice. “[T]he

Act’s provisions are not indefinitely elastic, content-free forms

to be shaped in whatever manner the Board might think best

conforms to the proper balance of bargaining power.”

American Ship Bldg. Co. v. NLRB, 380 U.S. 300, 310 (1965).

Because the Board overstepped its statutory bounds, the Court

of Appeals erred by deferring to the Board’s interpretation of

the statute. “Deference to the Board cannot be allowed to slip

into judicial inertia.” NLRB v. Financial Inst. Emp’ees, 475

U.S. 192, 202 (1985) (internal quotation omitted).°

* Even assuming, arguendo, that § 8(a)1) could be read to authorize the

Board to proscribe non-coercive polls, the Board’s incoherent and

(continued...)

B. Any Interpretation of the Act that Would

Authorize the Board to Proscribe Non-Coercive

Polling Would, at the Very Least, Raise Serious

First Amendment Questions.

Indeed, any interpretation of the Act that would authorize

the Board to proscribe non-coercive polling would, at the very

least, raise serious First Amendment questions. Employers, no

less than other citizens, enjoy the right to freedom of

expression. See, e.g, Gissel Packing, 395 U.S. at 617; Thomas

v. Collins, 323 U.S. 516, 532 (1945); NLRB v. Virginia Elec. &

Power Co., 314 U.S. 469, 478 (1941). That right is guaranteed

not only by the First Amendment, but also by § 8(c) of the Act,

29 U.S.C. § 158(c), which “implements the First Amendment”

by establishing “an employer's free speech right to

communicate his views to his employees.” Gissel Packing,

395 U.S. at 617.’

The Board, however, has long held that the protections of

§ 8(c)}—and, by extension, the First Amendment—are not

implicated when an employer seeks to poll or otherwise

question its employees about their union sympathies.

According to the Board, such employer activities are not

expressive in nature. “[A]n employer, in questioning his

employees as to their union sympathies, is not expressing

* (continued)

contradictory rules on polling do not reflect a reasonable interpretation of

the statute for the reasons set forth in Part |, supra. Cf National Ass'n of

Regulatory Util. Comm'rs v. ICC, 41 F.3d 721, 726-27 (D.C. Cir. 1994)

(Silberman, J.) (noting overlap between second step of Chevron inquiry and

conventional “arbitrary and capricious” review).

” Section 8(c), which was added to the Act by the Taft-Hartley

Amendments of 1947, provides in pertinent part as follows: “The expressing

of any views, argument, or opinion, or the dissemination thereof, whether

in written, printed, graphic, or visual form, shall not constitute or be

evidence of an unfair labor practice .. . if such expression contains no threat

of reprisal or force or promise of benefit.” 29 U.S.C. § 158(c).

views, argument, or opinion within the meaning of Section 8(c)

of the Act, as the purpose of an inquiry is not to express views

but to ascertain those of the person questioned.” Siruksnes,

165 N.L.R.B. at 1062 n.8. See also Cannon Elec. Co., 151

N.L.R.B. 1465, 1469 (1965) (“Interrogation, particularly

systematic polling of employees as to their union sympathies,

is not an expression of ‘views, argument, or opinion’ within the

meaning of Section 8(c). The purpose of interrogation is not to

express views, but to ascertain those of the person

interrogated.”), overruled on other grounds, Resistance Tech.,

280 N.L.R.B. 1004 (1986); Standard-Coosa-Thatcher Co., 85

N.LRB. 1358, 1363 (1949) (“[W]e again reject the contention

that interrogation is protected by Section 8(c) of the amended

Act. ... [T]he purpose of [§ 8(c)] is to permit an employer to

express his views, not to license him to extract those of his

employees.”).

The Board has thereby evinced a seriously deficient

understanding of the First Amendment. A question or

solicitation, no less than a declarative statement, is a form of

expression. See, e.g., Edenfield v. Fane, 507 U.S. 761, 765-66

(1993); Village of Schaumburg v. Citizens for a Better

Environment, 444 U.S. 620, 628-32 (1980); cf United States v.

Long, 905 F.2d 1572, 1579-80 (D.C. Cir.) (Thomas, J.), cert.

denied, 498 U.S. 948 (1990). Indeed, the Board’s rationale for

restricting employer polling and other forms of

questioning—that such activities implicitly convey a message

of anti-union animus—flatly contradicts the notion that such

activities are non-expressive. See Allegheny Ludlum, 104 F.3d

at 1363. A free and open interchange of ideas and information

necessarily entails the abjlity to solicit the views of others. Cf

International Soc'y for Krishna Consciousness v. Lee, 505 U.S.

672, 677-78 (1992); Riley v. National Fed'n of the Blind of

NC,, Inc., 487 U.S. 781, 788-89 (1988). The Board cannot, by

simply reciting the mantra “industrial and workplace stability,”

muzzle an employer’s freedom to communicate with its

employees, at least insofar as such communication is truthful

and non-coercive. See, e.g., Thomas, 323 U.S. at 532; Virginia

Elec. & Power Co., 314 U.S. at 478; NLRB v. Intertherm, Inc.,

596 F.2d 267, 277-78 (8th Cir. 1979); NLRB v. Golub Corp.,

388 F.2d 921, 926-29 (2d Cir. 1967) (Friendly, J.); cf 44

Liquormart, Inc. v. Rhode Island, 116 S. Ct. 1495, 1507-08

(1996).

Accordingly, the Board cannot, consistent with the First

Amendment, prevent an employer from conducting a

concededly non-coercive poll (like the one at issue here) to

assess the level of union sympathy (or antipathy) among its

employees. It necessarily follows that the Act must be

construed in a manner that would avoid such an

unconstitutional result. See, e.g, Edward J. DeBartolo Corp.

v. Florida Gulf Coast Bldg. and Constr. Trades Council, 485

US. 568, 575-78 (1988); NLRB v. Catholic Bishop of Chicago,

440 US. 490, 506-07 (1979). Indeed, this Court has repeatedly

rejected the Board’s efforts to characterize communicative

activity protected by the First Amendment as an “unfair labor

practice.” See, e.g, DeBartolo, 485 U.S. at 577-78 (employee

handbilling); NLRB v. Drivers, 362 U.S. 274, 284 (1960)

(employee picketing).

CONCLUSION

For the foregoing reasons the judgment of the Couri of

Appeals for the District of Columbia should be reversed and

enforcement of the Board’s order should be denied.

Respectfully submitted,