Journal of Interdisciplinary History, xxxviii:3 (Winter, 2008), 377–399.
BENEFIT OF THE DOUBT
Larry M. Logue and Peter Blanck
“Beneªt of the Doubt”: African-American Civil
War Veterans and Pensions A reading of Civil War pen-
sion statutes and lawmakers’ pronouncements suggests a federal
military-beneªt system that was color-blind. Pension laws rarely
mention race, and a Congressional committee saw “no reason
why the heirs of colored soldiers should not be put on the same
footing as to bounty and pensions as the heirs of white soldiers,
and many reasons why they should.” Nearly 200,000 African
Americans served in the Union armed forces, and statements such
as these would have raised hopes of equal treatment for black sur-
vivors and their heirs.1
Yet race was never far from the minds of the men who ad-
ministered the pension laws. In their certainty about the character
of “ignorant colored people,” ofªcials painted a curious picture of
African-American pension applicants: They were at once devious
(“those of that race who can be counted reliable and absolutely
truthful, are a rarity indeed”) and naïve (a government auditor
Larry M. Logue is Professor of History and Political Science, Mississippi College. He is the
author of “Union Veterans and Their Government: The Effects of Public Policies on Private
Lives,” Journal of Interdisciplinary History, XXII (1992), 411–434; editor, with Michael Barton,
of The Civil War Veteran: A Historical Reader (New York, 2007).
Peter Blanck is University Professor and Chairman, Burton Blatt Institute, Syracuse
University. He is the author, with Eve Hill, Charles D. Siegal, and Michael Waterstone, of
Disability Civil Rights Law and Policy: Cases and Materials (St. Paul, 2005); Employment, Disabil-
ity, and the Americans with Disabilities Act (Evanston, 2000).
This research was funded in part by grants to Blanck from the U.S. Dept. of Education,
National Institute on Disability and Rehabilitation Research (nidrr), for the Rehabilitation
Research and Training Center (rrtc) on Workforce Investment and Employment Policy for
Persons with Disabilities, Grant No. H133B980042-99; Demand Side Employment Place-
ment Models, Grant No. H133A060033; and the Rehabilitation Services Administration
(rsa) for rrtc on Employment Policy for People with Disabilities. The authors also gratefully
acknowledge the support of nih/nia grant number P01 AG10120, Early Indicators of Later
Work Levels, Disease and Death—Robert W. Fogel, principal investigator—and the helpful
comments and insights of an anonymous referee.
© 2007 by the Massachusetts Institute of Technology and The Journal of Interdisciplinary
History, Inc.
1 Report of the Committee on Pensions, Bounty, and Back Pay, House Rpt. 1883, 47th
Cong., 2d sess. 1882/83, serial set 2159, 1. The Pension Bureau also had a number of African-
American employees. See Theda Skocpol, Protecting Soldiers and Mothers: The Political Origins of
Social Policy in the United States (Cambridge, Mass., 1992), 596 n.
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
e
d
u
/
j
i
/
n
h
a
r
t
i
c
e
–
p
d
l
f
/
/
/
/
3
8
3
3
7
7
1
7
0
4
9
0
2
/
j
i
.
.
n
h
2
0
0
8
3
8
3
3
7
7
p
d
.
.
.
f
b
y
g
u
e
s
t
t
o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
378 | LARRY M. LOGUE A ND PETER B LA NCK
wondered at “the general credulity with which the race is apt to
listen to the proposals of a sharper”).2
Government ofªcials were untroubled by inconsistencies in
the intent and administration of pensions for black veterans and
their survivors. Instead, administrators justiªed the attention given
to African Americans’ character as necessary “to afford protection
to the government and honest claimants.” Ofªcials’ motivation,
however, is not the only, or the most signiªcant, window on the
role of race in the Civil War pension system; the experiences of
black veterans, such as Clay Ballard, are equally important.3
Born a slave in Kentucky, Ballard enlisted in the 116th
Colored Infantry in 1864, survived the war without serious disease
or injury, and was discharged in 1867. Ballard returned to Ken-
tucky, where he and his wife were running a boardinghouse when
his health declined in the late 1880s. In 1890, when Congress
changed the law to allow pensions for disabilities that developed
after military service, Ballard submitted his ªrst application, citing
rheumatism and scurvy. Despite the testimony of a private physi-
cian that Ballard was “unªt for hard manual labor,” a board of
physicians found no pensionable disabilities. Ballard asked that a
reevaluation be conducted in Cincinnati, insisting, “I did not re-
ceive a fair and impartial examination” because no “colored ex-
soldier can get justice from that board [in Lexington, Kentucky].”
Ballard received another examination (in Frankfort, Kentucky),
and the physicians found him partially disabled. But, this time, the
Pension Bureau’s reviewers rejected the claim. Ballard ªled a new
application claiming additional disabilities in 1894, but he died the
next year before another examination could be scheduled.4
Ballard’s case would ordinarily have ended at this point,
but his widow Mary, assisted by the pension attorney who had
represented her husband, pursued his claim in order to obtain a
2 E. D. Townsend to Secretary of War, April 1, 1878, Senate Exec. Doc. 57, 45th Cong., 2d
sess., 1877/78, serial set 1781, 5; pension examiner quoted in Donald R. Shaffer, After the
Glory: The Struggles of Black Civil War Veterans (Lawrence, 2004), 130; E. B. French to Secre-
tary of War, April 24, 1872, Senate Exec. Doc. 57, 45th Cong., 2d sess., 1877/78, serial set 1781,
12 (emphasis in original).
3 E. B. French to Secretary of War, Jan. 25, 1878, Senate Exec. Doc. 57, 8.
4 Afªdavit of J. C. Carrick, n.d., afªdavit of Clay Ballard, n.d., Pension File of Clay Ballard
(116th U.S. Colored Infantry), Federal Military Pension Applications (RG 15: Records of the
Veterans Administration), National Archives, Washington, D.C.
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
e
d
u
/
j
i
/
n
h
a
r
t
i
c
e
–
p
d
l
f
/
/
/
/
3
8
3
3
7
7
1
7
0
4
9
0
2
/
j
i
.
.
n
h
2
0
0
8
3
8
3
3
7
7
p
d
.
.
.
f
b
y
g
u
e
s
t
t
o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
BENEFIT OF THE DOUBT | 379
widow’s pension. The Pension Bureau stood its ground, however,
declaring in 1897 that Clay Ballard had not been prevented from
“earning a support,” whereupon the attorney appealed, arguing
that “it certainly looks strange . . . when soldiers die without being
disabled.” The Bureau, suspecting that its rejections “cannot now
be defended,” reopened the case, and after taking depositions, a
special examiner decided that Ballard had indeed been “practically
totally disabled,” clearing the way for Mary Ballard’s pension.5
Other than an endorsement of Clay Ballard as “one of the
better class of negroes,” race is largely absent from his pension
ªle. It would be premature, however, to conclude that Ballard was
the victim of a simple oversight. If other African Americans
disproportionately shared Ballard’s acknowledged mistreatment
at the hands of the Pension Bureau, the pension system’s apparent
color-blindness would have counted for little. If, however, Bal-
lard’s case were atypical, the federal government’s promise to “de-
fend, to the last dollar and the last man,” the rights of black recruits
to be “soldiers of the Union–nothing less and nothing different”
might still retain credence.6
Recent studies of black pension applicants suggest that Bal-
lard’s experience was commonplace. Noting that African Ameri-
cans encountered more outright rejections and smaller pension
awards than did whites, researchers point to biased pension exam-
iners and documentation rules that were more difªcult for black
veterans than white veterans to satisfy. These ªndings, together
with studies that have found longevity beneªts from Civil War
pensions for white veterans, are the starting point for this article.
The ultimate questions are whether African Americans were dis-
couraged from applying for pensions in the ªrst place; whether,
considering the importance of other characteristics, race swayed
5
J. F. Kinney to Secretary of the Interior, Oct. 9, 1897; J. F. Raub to F. D. Stephenson,
May 21, 1898; Percy S. Crowe to Commissioner of Pensions, Feb. 11, 1899, Senate Exec. Doc.
57.
6 Deposition of F. O. Young, Feb. 9, 1899, Senate Exec. Doc. 57; John A. Andrew, para-
phrasing Edwin Stanton to George T. Downing, March 23, 1863, in Ira Berlin, Joseph P.
Reidy, and Leslie S. Rowland (eds.), Freedom: A Documentary History of Emancipation (New
York, 1982), 88–89. Stanton was referring to free black recruits, but ex-slaves were also con-
vinced that “tings can never go back, because we have showed our energy and our courage
and our naturally manhood” (quoted in Gary Kynoch, “Terrible Dilemmas: Black Enlistment
in the Union Army during the American Civil War,” Slavery and Abolition, XVIII [1997],
123).
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
e
d
u
/
j
i
/
n
h
a
r
t
i
c
e
–
p
d
l
f
/
/
/
/
3
8
3
3
7
7
1
7
0
4
9
0
2
/
j
i
.
.
n
h
2
0
0
8
3
8
3
3
7
7
p
d
.
.
.
f
b
y
g
u
e
s
t
t
o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
380 | LARRY M. LOGUE A ND PETER B LA NCK
pension awards; and whether black veterans shared in any health
beneªts from Civil War pensions.7
The data for this article come from two samples of Civil War
soldiers. The ªrst, containing information about African Ameri-
cans, is contained in a sample created by Fogel and colleagues at
the Center for Population Economics at the University of Chi-
cago. This sample consists of service and pension records of the
men who joined ªfty-two randomly selected companies of the
U.S. Colored Troops. The second is an earlier sample created by
the same researchers, containing the records of men who served in
303 white infantry companies. This essay will refer to these data as
the cpe samples.8
race and pension applications Applying for a pension was a
complex process. A veteran began by submitting the particulars of
his military service and disability to the Pension Bureau, which
requested service conªrmation from the War Department and
occasionally asked applicants (or special examiners in unusually
challenging cases) to obtain corroboration from comrades or
commanders. If his application was thus far acceptable, the veteran
was instructed to obtain an examination from a government-
appointed physician (or more commonly a board of physicians),
who evaluated his medical claim, expressing the ªndings in a rec-
ommended “rating” of the applicant’s degree of disability. Legal
and medical reviewers at the Pension Bureau made the ªnal deci-
sion on what monthly payment, if any, the applicant would re-
ceive. Rejected applicants could ªle new claims when, as in Bal-
7
Shaffer, After the Glory, 119–137; Blanck and Chen Song, “Civil War Pensions for Union
Army Veterans: Race and Disability,” paper presented at the National Bureau of Economic
Research Cohort Studies Meeting, Chicago, 2004; Logue and Blanck, “‘There is Nothing
That Promotes Longevity Like a Pension’: Disability Policy and Mortality of Civil War
Union Army Veterans,” Wake Forest Law Review, XXXIX (2004), 49–67; Martin Salm, “The
Effect of Pensions on Longevity: Evidence from Union Army Veterans,” Discussion Paper
No. 2668, Institute for the Study of Labor (Bonn, Germany, 2007).
8 Both the African-American and white samples are included in Robert W. Fogel et al.,
Aging of Veterans of the Union Army: Military, Pension, and Medical Records, 1820–1940, University
of Chicago, Center for Population Economics, http://www.cpe.uchicago.edu. On the
evolving inclusion of African Americans in the Union army and the formation of the U.S.
Colored Troops, see John David Smith, “Let Us All Be Grateful That We Have Colored
Troops That Will Fight,” in idem (ed.), Black Soldiers in Blue: African American Troops in the
Civil War Era (Chapel Hill, 2002), 1–77; Joseph T. Glatthaar, Forged in Battle: The Civil War
Alliance of Black Soldiers and White Ofªcers (New York, 1990); Noah Andre Trudeau, Like Men
of War: Black Troops in the Civil War, 1862–1865 (Boston, 1998)
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
e
d
u
/
j
i
/
n
h
a
r
t
i
c
e
–
p
d
l
f
/
/
/
/
3
8
3
3
7
7
1
7
0
4
9
0
2
/
j
i
.
.
n
h
2
0
0
8
3
8
3
3
7
7
p
d
.
.
.
f
b
y
g
u
e
s
t
t
o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
BENEFIT OF THE DOUBT | 381
lard’s case, new disabilities arose, and pensioners could request
payment increases when the law changed.9
Ex-soldiers who had changed a slave name undoubtedly had a
tough time documenting their service, as did impoverished Afri-
can Americans who had to travel to examining boards and illiter-
ate black veterans who had to obtain written testimony. Although
“claim houses”—ªrms headed by attorneys specializing in pension
claims—promised to assist veterans with their applications in
exchange for a government-regulated fee, studies that compare
African-American and white pensioners ªnd a substantially lower
proportion of applications among black veterans. Did the applica-
tion process become the ªrst level of discrimination against black
veterans?10
At ªrst glance, it appears that African Americans were indeed
discouraged from seeking a pension. Slightly less than one-third
(31.9 percent) of all black enlisted men in the cpe samples ever ap-
plied for a pension, whereas more than half (52.5 percent) of white
veterans applied. Yet these aggregates mask complexities in gain-
ing access to the pension system: Proportionately more white sol-
diers survived the war (85 versus 79 percent of black recruits), for
example, and any racial difference in application rates might not
have remained uniform throughout changes in pension law.
Table 1 allows for these two inºuences on pension applica-
tions. Examining survivors of the war, the table shows the per-
centage of veterans who applied for a new pension in the ªrst two
major periods of pension law. During the “general-law” period,
from the beginning of pensions in 1862 until mid-1890, applicants
9 The medical-examination system evolved from a roster of individual contract physicians
and boards in major cities through the early 1880s to a network of more than 1,200 boards
(most of which had three members) throughout the country by 1895. See Annual Report of
Commissioner of Pensions, House Doc. 5/5, 54th Cong., 1st sess., 1895/96, serial set 3383, 32.
Useful summaries of the application and evaluation process in the 1890s are in Annual Report
of the Secretary of the Interior, House Exec. Doc. 1/15, 52d Cong., 1st sess., 1891/92, serial set
2933, 70–73; U.S. Government Printing Ofªce, A Treatise on the Practice of the Pension Bureau
(Washington, D.C., 1898).
10
See Shaffer, After the Glory, 123–128; Edward A. Miller, Jr., The Black Civil War Soldiers of
Illinois: The Story of the Twenty-Ninth U.S. Colored Infantry (Columbia, S.C., 1998), 179–183.
On pension attorneys, see Blanck and Song, “Civil War Pension Attorneys and Disability Pol-
itics,” University of Michigan Journal of Law Reform, XXXV (2001/02), 137–217. Shaffer, After
the Glory, 209, ªnds that 64% of black veterans sampled applied for a pension at least once,
compared to 77% of whites. Blanck and Song, “Civil War Pensions,” report that applications
from black veterans amounted to 9.2% of all pension applications, though African Americans
comprised 14.5% of the sample.
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
e
d
u
/
j
i
/
n
h
a
r
t
i
c
e
–
p
d
l
f
/
/
/
/
3
8
3
3
7
7
1
7
0
4
9
0
2
/
j
i
.
.
n
h
2
0
0
8
3
8
3
3
7
7
p
d
.
.
.
f
b
y
g
u
e
s
t
t
o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
382 | LARRY M. LOGUE A ND PETER B LA NCK
Table 1 Percentage of Civil War Survivors Who Applied for a New Pension
by Race, cpe Samples
general law
1862–1890
disability law
1890–1907
Whites
African Americans
Number of cases
43.9
19.7
29,527
91.6
93.6
8,004
had to demonstrate that military service had caused a disability that
limited their capacity “to procure a subsistence by manual labor.”
After passage of the Disability Pension Act in mid-1890, pensions
became a general disability beneªt: Almost any serious impair-
ment, whether it was service-related or not, would qualify a vet-
eran for a pension. The analysis ends in February 1907, when
Congress made old age itself (deªned as sixty-two years old, which
included almost all Civil War veterans) the legal basis for a
pension.11
The table shows two striking patterns in veterans’ behavior.
First, the reactions of white and African-American veterans to the
general law were markedly different. More than 40 percent of
white veterans sought a pension before mid-1890, but fewer than
20 percent of black veterans applied. Under the disability law,
however, veterans of both races overwhelmingly sought access to
the pension system; indeed, the proportion of African-American
veterans who applied exceeded that of whites. This new participa-
tion throws the general-law results into sharp relief. Why did
black veterans disproportionately shy away from the ªrst pension
system?12
Shaffer’s pioneering study of African-American veterans at-
tributes much of the racial difference in pension seeking to the dis-
advantages that black veterans faced in navigating the application
process. Since most African-American soldiers had been slaves,
11 For a summary of pension-law changes, see Blanck, “Civil War Pensions and Disability,”
Ohio State Law Journal, XLII (2001), 117–127.
12 The population “at risk” to apply for a pension after 1890, shown in the second column
of Table 1, includes only those who were known (or estimated, according to the method de-
scribed in the Appendix) to be alive and who had not yet applied.
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
e
d
u
/
j
i
/
n
h
a
r
t
i
c
e
–
p
d
l
f
/
/
/
/
3
8
3
3
7
7
1
7
0
4
9
0
2
/
j
i
.
.
n
h
2
0
0
8
3
8
3
3
7
7
p
d
.
.
.
f
b
y
g
u
e
s
t
t
o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
BENEFIT OF THE DOUBT | 383
and since many ex-slaves were impoverished and illiterate, Shaffer
argues that black veterans were especially unlikely to hear about
pensions and unable to document their claims.13
The effects of deprivation cannot be directly assessed from the
cpe data. Because the literacy of Civil War recruits was not re-
corded, applicants’ signatures (or marks) on pension records are
the principal indicators of sample members’ literacy. Those who
never applied are an even greater mystery. Exploring reasons for
pension seeking thus requires information collected during the
war, one item of which was soldiers’ place of birth.
Indeed, birthplace is the basis of an excellent proxy for gaug-
ing postwar African-American deprivation. Although former ser-
vitude is not listed in military records, ex-slaves can reliably be
identiªed from their state and county of birth. In the 1860 census,
99.6 percent of the black population in thirteen southern states,
plus the majority-slave counties of Maryland, were slaves, whereas
93.9 percent of African Americans living elsewhere were free.
This slave–free distinction was reºected in literacy and wealth af-
ter the Civil War. In 1870, 64 percent of black men aged thirty to
ªfty who had been born in free states could write, as opposed to
13 percent of those born in slave states; freeborn black men held
an average of $436 in property versus $116 for ex-slaves.14
Table 2 weighs the effect of this distinction against two char-
acteristics that were directly related to eligibility for pensions prior
to 1890. A military hospital record of conªnement for a wound or
for a disease characterized as “severe,” “acute,” or “chronic” was a
crucial asset to a claim for a general-law pension, and Table 2 in-
cludes dummy variables for wounds and illnesses. The table re-
ports a proportional-hazards analysis, which models the risk of an
event (in this case, a pension application) taking place; separate
analyses are shown for white and African-American veterans.
Since proportional-hazards analysis takes account of each person’s
Shaffer, After the Glory, 123–126.
13
14 Census comparisons are derived from the ipums samples. See Steven Ruggles et al., Inte-
grated Public Use Microdata Series: Version 3.0 (Machine-readable database), Minneapolis, Min-
nesota Population Center (producer and distributor), 2004. Since county of birth was not
recorded in the census and since Maryland was nearly evenly divided between slaves and free
African Americans, Maryland is excluded from the ªgures. The comparable literacy ªgure for
white men in 1870 is 91%. Property averages are based on census-listed real and personal
property combined.
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
e
d
u
/
j
i
/
n
h
a
r
t
i
c
e
–
p
d
l
f
/
/
/
/
3
8
3
3
7
7
1
7
0
4
9
0
2
/
j
i
.
.
n
h
2
0
0
8
3
8
3
3
7
7
p
d
.
.
.
f
b
y
g
u
e
s
t
t
o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
r
a
W
l
i
v
i
C
f
o
s
r
o
v
i
v
r
u
S
l
l
A
—
e
c
a
R
y
b
0
9
8
1
–
d
i
m
e
r
o
f
e
b
n
o
i
s
n
e
P
w
e
N
a
r
o
f
g
n
i
y
l
p
p
A
f
o
”
k
s
i
R
“
e
h
t
n
o
s
e
c
n
e
u
º
n
I
e
v
i
t
a
l
e
R
)
s
e
l
p
m
a
S
e
p
c
(
p
1
0
0
.
1
0
0
.
2
0
0
.
s
n
a
c
i
r
e
m
a
n
a
c
i
r
f
a
o
i
t
a
r
d
r
a
z
a
h
n
a
e
m
3
2
.
2
2
4
.
1
6
7
.
4
2
1
.
1
2
1
.
8
2
8
.
p
1
0
0
.
1
0
0
.
7
1
7
,
3
4
.
7
7
8
,
0
1
5
.
5
0
1
o
i
t
a
r
d
r
a
z
a
h
s
e
t
i
h
w
4
3
.
2
7
6
.
1
n
a
e
m
7
8
2
.
1
4
1
.
2
1
8
,
5
2
9
.
0
7
0
,
5
1
2
0
.
3
9
2
,
2
s
s
e
n
l
l
i
c
i
n
o
r
h
c
/
e
t
u
c
a
/
e
r
e
v
e
S
d
n
u
o
w
r
o
f
d
e
z
i
l
a
t
i
p
s
o
H
e
v
a
l
s
r
e
m
r
o
F
s
e
t
a
i
r
a
v
o
c
h
t
i
w
t
n
e
m
e
v
o
r
p
m
I
d
o
o
h
i
l
e
k
i
l
g
o
l
2
(cid:2)
l
a
i
t
i
n
I
s
e
s
a
c
f
o
r
e
b
m
u
N
a
2
e
l
b
a
T
e
l
b
a
i
r
a
v
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
e
d
u
/
j
i
/
n
h
a
r
t
i
c
e
–
p
d
l
f
/
/
/
/
3
8
3
3
7
7
1
7
0
4
9
0
2
/
j
i
.
.
n
h
2
0
0
8
3
8
3
3
7
7
p
d
.
.
.
f
b
y
g
u
e
s
t
t
o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
y
l
n
O
s
e
v
a
l
S
r
e
m
r
o
F
—
e
c
a
R
y
b
0
9
8
1
–
d
i
m
e
r
o
f
e
b
n
o
i
s
n
e
P
w
e
N
a
r
o
f
g
n
i
y
l
p
p
A
f
o
”
k
s
i
R
“
e
h
t
n
o
s
e
c
n
e
u
º
n
I
e
v
i
t
a
l
e
R
)
s
e
l
p
m
a
S
e
p
c
(
p
1
0
0
.
2
1
0
.
0
9
8
.
2
0
0
.
o
i
t
a
r
d
r
a
z
a
h
6
9
9
.
1
1
3
3
.
1
7
4
0
.
1
4
0
5
.
1
4
6
7
,
2
5
.
8
5
8
,
7
6
.
8
7
n
a
e
m
3
2
1
.
7
3
1
.
2
2
1
.
8
8
1
.
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
e
d
u
/
j
i
/
n
h
a
r
t
i
c
e
–
p
d
l
f
/
/
/
/
3
8
3
3
7
7
1
7
0
4
9
0
2
/
j
i
.
.
n
h
2
0
0
8
3
8
3
3
7
7
p
d
.
.
.
f
b
y
g
u
e
s
t
t
o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
s
e
t
a
i
r
a
v
o
c
h
t
i
w
t
n
e
m
e
v
o
r
p
m
I
n
r
o
b
e
e
r
f
y
n
a
p
m
o
c
f
o
n
o
i
t
r
o
p
o
r
P
C
D
/
A
V
/
D
M
/
E
D
n
i
d
e
t
s
i
l
n
E
s
s
e
n
l
l
i
c
i
n
o
r
h
c
/
e
t
u
c
a
/
e
r
e
v
e
S
d
n
u
o
w
r
o
f
d
e
z
i
l
a
t
i
p
s
o
H
d
o
o
h
i
l
e
k
i
l
g
o
l
2
(cid:2)
l
a
i
t
i
n
I
s
e
s
a
c
f
o
r
e
b
m
u
N
b
2
e
l
b
a
T
e
l
b
a
i
r
a
v
386 | LARRY M. LOGUE A ND PETER B LA NCK
time spent at risk of an event, veterans’ longevity has been esti-
mated where necessary (see Appendix).15
Table 2 presents hazard ratios, that is, the proportional effect
on the “risk” of submitting a pension application associated with
the reported value of each independent variable. Part A of the ta-
ble shows that former slaves were especially reluctant to apply for a
pension–wartime health aside—and that freedmen were nearly 25
percent less likely to apply than were freeborn black veterans. Part
B explores this difference by drawing on a study of the cpe samples
in which the researchers suggest that interaction between former
slaves and free blacks in the army may have contributed to higher
literacy among ex-slaves. It is conceivable that such interaction
also inspired former slaves to apply for pensions. Hence, Part B of
Table 2 focuses on ex-slaves in the cpe samples. A variable is added
for the percentage of company comrades who were freemen (us-
ing the slave–free classiªcation procedure described above), and a
dummy variable identiªes recruits who enlisted in eastern border
areas, which produced the companies with the most freeborn Af-
rican Americans. The latter variable is meant to account for inter-
action between slaves and free blacks before the war.16
The proportion of freeborn comrades produces a positive sign
on the tendency to apply for a general-law pension, but its effect is
15 For good explanations of proportional hazards and its application to historical data, see
John G. Treble, “On Marrows: Evidence from the Victorian Household Panel Study,” Histor-
ical Methods, XXVIII (1995), 183–193; Dora L. Costa, “Height, Weight, Wartime Stress, and
Older Age Mortality: Evidence from the Union Army Records,” Explorations in Economic His-
tory, XXX (1993), 424–449; J. Morgan Kousser, “‘The Onward March of Right Principles’:
State Legislative Actions on Racial Discrimination in Schools in Nineteenth-Century Amer-
ica,” Historical Methods, XXXV (2002), 177–204. Age, which is another potential inºuence on
the likelihood of applying for a pension, is not included in Table 2 since it was used in esti-
mating the survival of non-applicants (see Appendix). The validity of proportional-hazards
analysis rests on the assumption that the effect of the predictor variables is the same through-
out the time of observation. This assumption is potentially jeopardized by the Arrears Act of
1879, which, though it did not alter the basic pension requirements, stimulated a resurgence
of applications via its authorization of a retroactive payment dating to an applicant’s discharge
from the army (see Skocpol, Protecting Soldiers, 115–118). A useful test for violations of the as-
sumption is to divide the observation time into sub-periods and compare results. In such an
analysis (not given herein), the coefªcients are reasonably stable, especially the chief variable
of interest, former-slave status, which shows little change after the Arrears Act. For economy
of presentation, Table 2 covers the entire general-law period to 1890.
16
See Costa and Matthew E. Kahn, “Forging a New Identity: The Costs and Beneªts of
Diversity in Civil War Combat Units for Black Slaves and Freemen,” Journal of Economic His-
tory, LXVI (2006), 936–962. The places that were the source of units with the most freemen
were Maryland, Delaware, Virginia, and the District of Columbia.
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
e
d
u
/
j
i
/
n
h
a
r
t
i
c
e
–
p
d
l
f
/
/
/
/
3
8
3
3
7
7
1
7
0
4
9
0
2
/
j
i
.
.
n
h
2
0
0
8
3
8
3
3
7
7
p
d
.
.
.
f
b
y
g
u
e
s
t
t
o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
BENEFIT OF THE DOUBT | 387
displaced by the stronger inºuence of enlisting (and probably hav-
ing lived in) a border state, which raised the likelihood of applying
by 50 percent. This ªnding supports the argument that slavery’s
deprivations could discourage pension-seeking and that ex-slaves
who had grown up in the vicinity of free blacks, and so likely to
join companies ªlled with freeborn comrades, were more apt to
seek general-law pensions than were freedmen who had been de-
prived of this contact.17
Both parts of Table 2, however, show that all black veterans
joined their white peers in calculating their chances for a pension.
Having a wound, the surest sign of military service’s toll on the
body, doubled the likelihood that veterans of both races would
seek a pension; a record of a signiªcant service-related illness also
considerably raised the propensity to apply. Part of black veterans’
reluctance to seek general-law pensions undoubtedly came from
lack of experience and the resources necessary to negotiate a bu-
reaucratic maze, but their disproportionate lack of wartime trauma
cannot be denied. Because full-scale recruiting of African Ameri-
cans began nearly two years into the war and because some com-
manders resisted using black troops in combat, only 12 percent of
men in the black sample were hospitalized for wounds, versus 29
percent among the white sample.18
Table 2 cannot capture an additional difference within and
between the races. The health problems of non-applicants after
their mustering-out are largely unknown, but death rates suggest
that African-American veterans, especially former slaves, dispro-
portionately suffered from poor health. The crude death rate from
1865 to 1890 for members of the cpe samples with known death
dates was 15 per 1,000 person-years for ex-slaves, 10.7 for freeborn
black veterans, and 5.6 for whites.
The morbidity that undoubtedly underlay this mortality dif-
ference may have inºuenced pension seeking. Though Table 2
shows that service-related health problems encouraged black vet-
erans to apply for a pension, Table 1 makes it clear that any such
17
If used alone, rather than in conjunction with border-area enlistment, the percentage of
freeborn company comrades produces a statistically signiªcant effect (odds ratio 1.007, mean-
ing that a 50-percentage-point increase in freeborn comrades raises the likelihood of applying
by 35%), but place of enlistment reduces the comrades effect to the net shown in Table 2b.
18 On combat use of black troops, see Glatthaar, Forged in Battle, 164; Miller, Black Civil
War Soldiers, 114. This resistance could be overcome, however (ibid., 59–60). The term
wounds in this article includes injuries as well as gunshot and artillery wounds.
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
e
d
u
/
j
i
/
n
h
a
r
t
i
c
e
–
p
d
l
f
/
/
/
/
3
8
3
3
7
7
1
7
0
4
9
0
2
/
j
i
.
.
n
h
2
0
0
8
3
8
3
3
7
7
p
d
.
.
.
f
b
y
g
u
e
s
t
t
o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
388 | LARRY M. LOGUE A ND PETER B LA NCK
encouragement took place far less often than among whites. Cur-
rent illnesses and injuries, which probably plagued black veterans
more than their white peers, could easily have overwhelmed
health problems that had arisen in the army. If a veteran had been
hospitalized for diarrhea and then developed tuberculosis or heart
disease after the war, the general law offered him no assistance. As
their health continued to worsen, black veterans who survived
into the 1890s would have been especially enthusiastic about new
pensions for recent health problems.19
Multivariate analysis of pension seeking under the disability
law is unnecessary, because the nearly universal participation
shown in Table 1 leaves little variation to explain. Application
procedures under the new law became only marginally easier.
Military service still had to be proven, and a disability not due to
“vicious habits” had to be veriªed, though the burden of connect-
ing it to the war was lifted. Yet, despite the still-daunting bureau-
cratic rules, former slaves now pursued pensions as eagerly as did
white veterans.
The new law’s acceptance of current disabilities was poten-
tially enough to trigger a ºood of new applications, but not among
ex-slaves unable to learn about its provisions or intimidated by the
pension system’s continuing intricacies. Contemporary develop-
ments, however, may have overcome these problems. Chief
among such developments was a campaign in the 1890s aimed at
persuading Congress to create a system of bonuses and pensions
for all former slaves. The plan was never enacted, but Pension Bu-
reau ofªcials were sufªciently worried about confusion between
existing and proposed pensions to order repeated investigations
into the ex-slave plan’s promoters. The promoters were “setting
the negroes wild,” wrote one investigator, “robbing them of their
money and making anarchists of them.” As an ex-slave pension as-
19 On a late nineteenth-century “health crisis” among African Americans, see Kenneth F.
Kiple and Virginia H. King, Another Dimension to the Black Diaspora: Diet, Disease, and Racism
(New York, 1981), 187–190. For a more polemical view, see W. Michael Byrd and Linda A.
Clayton, An American Health Dilemma: A Medical History of African Americans and the Problem of
Race (New York, 2000), I, 325–357. See also Costa, “Race and Older Age Mortality: Evi-
dence from Union Army Veterans,” National Bureau of Economic Research Working Paper
10902, 2004 (http//www.nber.org/papers/w10902); Douglas Ewbank, “History of Black
Mortality and Health before 1940,” Milbank Memorial Fund Quarterly, LXV (supp. 1) (1987),
100–128; Samuel H. Preston and Michael R. Haines, Fatal Years: Child Mortality in Late Nine-
teenth-Century America (Princeton, 1991), 81–85.
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
e
d
u
/
j
i
/
n
h
a
r
t
i
c
e
–
p
d
l
f
/
/
/
/
3
8
3
3
7
7
1
7
0
4
9
0
2
/
j
i
.
.
n
h
2
0
0
8
3
8
3
3
7
7
p
d
.
.
.
f
b
y
g
u
e
s
t
t
o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
BENEFIT OF THE DOUBT | 389
sociation formed new chapters and pension advocacy spread by
word of mouth, black veterans became emboldened to apply for
Civil War pensions.20
Claim houses and their agents also redoubled their efforts to
stimulate new pension applications in the wake of the disability
law. A member of Congress declared in 1900 that claim agents
were “writing all over the country, . . . hunting up everybody to
whom we owe anything.” Indeed, the already heavy reliance on
claim houses, which was greater among African Americans than
among white veterans, became nearly universal after 1890.21
It is improbable that any post-1890 change by itself produced
the extraordinary disappearance of a racial difference in pension
seeking. Acting together, however, the provisions of the new law
and the pension-seeking crusades contributed to a revolution in
behavior. In an era usually known for proscription and violence
directed at African Americans, black veterans began to assert their
right to participate as equals in a government program.
application success and rejection The case of Willis Pleasant
points to the complexity of black veterans’ behavior and raises
questions about their treatment in the pension system. Pleasant,
like Ballard, served in a Kentucky regiment of the U.S. Colored
Troops, but Pleasant had extensive health problems while in uni-
form. Serving in Kentucky after the end of hostilities, Pleasant
sought medical attention on fourteen occasions for such maladies
as diarrhea, rheumatism, and a foot injury. Pleasant submitted his
ªrst pension application in 1890 after passage of the Disability Pen-
sion Act, and his actions suggest that he had just discovered the
system. He ªrst applied under the disability law, later supplement-
ing the claim with another in 1891 in which he cited the military
origins of his present neuralgia and intestinal disorder. Examining
surgeons rated Pleasant as one-third disabled under the later law,
20 On the Pension Bureau’s alarm, see Annual Report of Commissioner of Pensions, House
Doc. 5/6, 56th Cong., 1st sess., 1899/1900, serial set 3917, 47–48. Mary Frances Berry, My
Face Is Black Is True: Callie House and the Struggle for Ex-Slave Reparations (New York, 2005),
83, 33–74, 93–121.
Joseph Cannon, quoted in Berry, My Face, 74. Claim-house assistance to applicants, not
21
differentiated between that offered by attorneys or their brokers, is recorded for members of
the cpe samples: 87% of black applicants and 85% of white applicants for new pensions enlisted
claim houses under the general law, and 96% of black applicants and 92% of white applicants
used claim houses under the disability law.
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
e
d
u
/
j
i
/
n
h
a
r
t
i
c
e
–
p
d
l
f
/
/
/
/
3
8
3
3
7
7
1
7
0
4
9
0
2
/
j
i
.
.
n
h
2
0
0
8
3
8
3
3
7
7
p
d
.
.
.
f
b
y
g
u
e
s
t
t
o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
390 | LARRY M. LOGUE A ND PETER B LA NCK
but the Pension Bureau rejected the recommendation. Both sur-
geons and Bureau rejected Pleasant’s general-law claim. In 1900,
since Pleasant was well past the recently deªned age of “senility,”
he was granted a pension of $8 a month. Were white applicants treated in the same manner as was Pleasant? How much did his race, and that of other applicants, inºuence pension decisions?22 To explore this question, the focus must shift from veterans’ actions to their applications and their manner of evaluation. The centerpiece of the judgment process was supposed to be the medi- cal examination. As one pension commissioner asserted, the ex- amination was nothing less than “the basis of pension–it deter- mines the disability.” Yet the Pension Bureau had a contentious relationship with its appointed physicians. The same commis- sioner complained that the examining physicians “are appointed and assigned to duty without any knowledge of the law, and prac- tically without any experience or instruction,” citing test exami- nations that revealed considerable inaccuracy and inconsistency in the ratings of applicants. On the basis of the Bureau’s rejection of examining physicians’ ªndings for Ballard and Pleasant, apparently two judgment processes affected pension applicants.23 Table 3 shows the results of these processes. Decisions by both physicians and the Pension Bureau are analyzed as dichoto- mies. Additional gradations of “success” were possible because ap- plicants could be awarded varying payments under the law, but the starkest difference, and the one most likely to elicit the resent- ment expressed by Ballard, was the distinction between a pension and no pension.24 The table reveals differences in physicians’ judgment of the races. Although most African-American applicants were recom- mended for a pension, the physicians were even more likely to 22 Pension File of Willis Pleasant (119th U.S. Colored Infantry), Federal Military Pension Applications (RG 15: Records of the Veterans Administration), National Archives,Washing- ton, D.C. In the late 1890s, the Pension Bureau ruled that applicants sixty-ªve and older would be allowed the minimum pension, “unless the evidence disclose[d] an unusual vigor and ability for the performance of manual labor in one of that age”; those seventy-ªve and older were to be assigned “senility” as a disability. See Annual Report of Commissioner of Pensions, House Doc. 5/5, 55th Cong., 2d sess., 1897/98, serial set 3642, 52. A presidential or- der in 1904 authorized the Bureau to classify applicants sixty-two and older as one-half dis- abled and seventy and older as fully disabled. Pleasant was born in approximately 1830. 23 Annual Report of Commissioner of Pensions, House Doc. 5/6, 57th Cong., 1st sess.,1901/02, serial set 4292, 73, 75. 24 Physicians’ reports in Fogel et al., Aging of Veterans of the Union Army. l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / j i / n h a r t i c e – p d l f / / / / 3 8 3 3 7 7 1 7 0 4 9 0 2 / j i . . n h 2 0 0 8 3 8 3 3 7 7 p d . . . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 BENEFIT OF THE DOUBT | 391 Table 3a Examining Physicians’ Pension Recommendations by Race, Appli- cants for New Pensions, cpe Samples before 1890 1890–1907 % of white applicants recommended for a pension % of African-American applicants recommended for a pension Number of examinations 75.1 67.9 8,407 90.4 78.4 6,803 Table 3b Pension Bureau Awards by Race, Applicants for New Pensions, cpe Samples % of white applicants awarded a pension % of African-American applicants awarded a pension Number of rulings 77.9 39.4 8,224 72.5 45.7 6,621 endorse white applicants’ claims, especially under the disability law. Yet the Pension Bureau’s decision making shows a far larger racial disparity. Under the general law, a racial difference of less than 10 percentage points in surgeons’ approvals became a gap of thirty-eight percentage points in the hands of the Bureau, and a twelve-percentage-point difference grew to nearly twenty- seven points after 1890. The physicians’ behavior raises suspicions of prejudice, but the decisions of the Pension Bureau are the cen- tral focus in this context, because the racial difference in its rulings was much larger and because the Bureau’s judgment decided the actual awarding of a pension.25 Black veterans’ lower approval rates might have stemmed from eligibility differences originating in their wartime service, or from the caprices of partisan politics. Two-thirds of black appli- cants lived in the South, without beneªt of the alleged “order 25 The starting points for Table 3 are medical examinations that followed ªrst applications, and the Pension Bureau’s rulings on ªrst applications. The Bureau occasionally overruled ex- amining physicians’ rejections, accounting for the higher approval rate for whites under the general law. Cases would sometimes lapse between examination and ruling, usually because of the applicant’s death, thus explaining the difference between the examinations and rulings. l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / j i / n h a r t i c e – p d l f / / / / 3 8 3 3 7 7 1 7 0 4 9 0 2 / j i . . n h 2 0 0 8 3 8 3 3 7 7 p d . . . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 392 | LARRY M. LOGUE A ND PETER B LA NCK [that] was issued [in 1880] to reject no cases [in ‘swing’ states] pending the election,” and African Americans’ rush to apply under the disability law may have caught them in the “orgy of punishing and humiliating the pensioners of the Civil War” that supposedly occurred during President Cleveland’s second term.26 Table 4 includes dummy variables to control for residence in the South and swing states and application during the sec- ond Cleveland term, plus variables for the applicant’s age, war- time health, noncommissioned rank, and use of a claim house. To gauge the importance of race against these control vari- ables, the white and black cpe samples have been combined and weighted.27 The table shows odds ratios (analogous to the hazard ratios presented above) from a logistic regression for each of the ªrst two pension-law periods; application approval or rejection is the de- pendent variable. Instead of explaining away the racial difference in Pension Bureau approval rates, multivariate analysis underscores the importance of race in the judgment of applications. No matter their place of residence, their health while in the army, or their ac- 26 Testimony of Thomas P. Kane, Report of Select Committee on the Payment of Pen- sions, Bounty, and Back Pay, House Rpt. 387, 46th Cong, 3rd sess., 1880/81, serial set 1983, 389; Wiley Britton, A Traveling Court: Based on the Investigation of War Claims (Kansas City, 1926), 311, 308–309. See also Heywood Sanders, “Paying for the ‘Bloody Shirt’: The Politics of Civil War Pensions,” in Barry Rundquist (ed.), Political Beneªts (Lexington, 1980), 154– 156; Mary R. Dearing, Veterans in Politics: The Story of the G.A.R. (Baton Rouge, 1952), 451– 453; William H. Glasson, Federal Military Pensions in the United States (New York, 1918), 224; Donald L. McMurry, “The Political Signiªcance of the Pension Question, 1885–1897,” Mis- sissippi Valley Historical Review, IX (1922), 19–36; Blanck and Song, “Civil War Pension Attor- neys.” 27 The South is deªned as the former Confederate states, and swing states those that experi- enced two or more party switches in presidential elections from 1872 to 1892, plus Ohio, which favored Republican candidates throughout but is speciªcally cited in the testimony and in Sanders, “Paying for the Bloody Shirt.” The cpe African-American sample constitutes an oversample: Black recruits comprised 14.7% of the combined white and black samples, whereas African Americans constituted approximately 8.4% of all Union recruits. On the Union total of approximately 2.1 million soldiers (some estimates include sailors as well), see Annual Report of the Commissioner of Pensions, House Exec. Doc. 1/15, 47th Cong., 2d sess., 1882/83, serial set 2100, 723–725; Historical Statistics of the United States, Colonial Times to 1970 (Washington, D.C., 1975), II, 1140; James M. McPherson, Battle Cry of Freedom: The Civil War Era (New York, 1988), 306–307 n.; Benjamin A. Gould, Investigations in the Military and Anthropological Statistics of American Soldiers (New York, 1869), 27. Applying the ratio of enlist- ments to individuals in this article’s samples to the total enlistments given in U.S. Government Printing Ofªce, War of the Rebellion: A Compilation of the Ofªcial Records of the Union and Con- federate Armies, 1861–1865 (Washington, D.C.,1880–1901), IV, 1269–1270, produces a similar total of just under 2.1 million soldiers. For the number of black soldiers, see ibid. Black veter- ans are weighted downward statistically to comprise 8.4% of the total in Table 4. l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / j i / n h a r t i c e – p d l f / / / / 3 8 3 3 7 7 1 7 0 4 9 0 2 / j i . . n h 2 0 0 8 3 8 3 3 7 7 p d . . . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 e p c e t i h W d n a n a c i r e m A – n a c i r f A d e n i b m o C , s n o i t a c i l p p A n o i s n e P w e N r o f s s e c c u S f o s d d O f o s n o i s s e r g e R c i t s i g o L d e t h g i e W s e l p m a S 4 e l b a T p 2 3 0 . 9 3 8 . 1 0 0 . 2 0 9 . 5 0 7 . 1 2 0 . 1 0 0 . 5 4 8 . 1 0 0 . w a l y t i l i b a s i d 7 0 9 1 – 0 9 8 1 o i t a r s d d o n a e m 6 3 8 . 3 2 0 . 1 6 4 0 . 1 1 1 0 . 1 9 6 9 . 9 9 7 . 4 4 3 . 5 7 9 . 0 3 4 . 7 9 1 . 2 0 1 . 0 . 3 5 7 0 2 . 6 4 2 . 7 2 2 . 4 6 0 . 2 2 9 . 4 5 1 . p 1 0 0 . 1 0 0 . 1 0 0 . 1 0 0 . 1 0 0 . 2 0 0 . 1 0 0 . 1 0 0 . 6 4 1 , 5 4 . 2 3 4 , 5 2 . 4 4 2 l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / j i / n h a r t i c e – p d l f / / / / 3 8 3 3 7 7 1 7 0 4 9 0 2 / j i . . n h 2 0 0 8 3 8 3 3 7 7 p d . . . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 w a l l a r e n e g 0 9 8 1 – 2 6 8 1 o i t a r s d d o n a e m e l b a i r a v 8 9 2 . 1 9 5 4 . 1 2 7 9 . 8 0 4 . 1 6 5 2 . 1 8 6 7 . 5 2 3 . 5 4 6 . 1 7 9 3 . 4 1 2 . 5 . 3 4 8 1 2 . 3 3 3 . 6 3 1 . 2 6 8 . 1 7 0 . 8 2 5 , 7 7 . 4 1 7 , 8 7 . 4 0 4 s s e n l l i c i n o r h c / e t u c a / e r e v e S d n u o w r o f d e z i l a t i p s o H r e c ª f o d e n o i s s i m m o c n o N e t a t s g n w i s h t u o S n i n i g n i v i L g n i v i L e g A m r e t d n 2 s ’ d n a l e v e l C n i e s u o h m i a l c y b d e i l p p A d e t s i s s A n a c i r e m A n a c i r f A s n o i t a c i l p p a f o r e b m u n d e t h g i e w n U d o o h i l e k i l g o l 2 (cid:2) l a i t i n I s e t a i r a v o c h t i w t n e m e v o r p m I 394 | LARRY M. LOGUE A ND PETER B LA NCK cess to a claim house, black applicants were less than one-third as likely as white veterans to be approved before 1890, and less than one-half as likely under the disability law. This discrepancy was conceivably due to deªciencies in black veterans’ applications. Scrutinizing the amount and the cogency of the corroborating evidence for black applicants, or following the advice of special examiners who had sought such evidence, Pen- sion Bureau reviewers probably found black claims less compelling than those of whites who could more readily obtain conªrmation of service and disability. Yet Table 4’s results are a warning against overemphasizing application skills: Not even black veterans with military records of service-related trauma to support their cases or with the rank of corporal or sergeant, which presumably would have accustomed them to dealing with ofªcial procedures, were as likely as their white peers to receive a pension.28 A key distinction that underlies Table 3 helps to explain the discrimination suggested by Table 4. Examining physicians, who encountered applicants face to face, were hardly unaffected by what they saw. Examining boards noted that both Ballard and Pleasant had lost all of their teeth, and Pleasant’s examiners ob- served that he “walks with the unsteady gait of an old man.” One examining physician testiªed that “it is very natural to form an opinion of a man’s character when you examine him. . . . If he ap- pears like an honest man and not disposed to exaggerate, we are very likely to give him the beneªt of the doubt.” Although Table 3 shows that physicians were more skeptical of African Americans than of white applicants, the physicians were far more likely to give black veterans the beneªt of the doubt than were Pension Bureau ofªcials. Table 4 strongly suggests that seeing a notation of service in the U.S. Colored Troops on an application was enough to set Pension Bureau reviewers against the appli- cant.29 pensions and mortality Prejudiced actions have conse- quences, but they are not always readily identiªable. In the case of See Shaffer, After the Glory, 128–132, for a discussion of special pension examiners and 28 African-American applicants. 29 Surgeon’s Certiªcate, June 8, 1891, Pension File of Clay Ballard; Surgeon’s Certiªcate, Oct. 3, 1900, Pension File of Willis Pleasant; Testimony of Joseph F. Atwood, Report of Se- lect Committee on the Payment of Pensions, Bounty, and Back Pay, House Rpt. 387, 46th Cong, 3d sess., 1880/81, serial set 1983, 133. l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / j i / n h a r t i c e – p d l f / / / / 3 8 3 3 7 7 1 7 0 4 9 0 2 / j i . . n h 2 0 0 8 3 8 3 3 7 7 p d . . . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 BENEFIT OF THE DOUBT | 395 government payments to aging, disabled men, however, the con- sequences of prejudice would seem to be evident in veterans’ health after their pension decisions. The tribulations of Ballard and Pleasant provide two cases in point. Ballard was convinced that his race had cost him a pension in 1890; Pleasant and other rejected black applicants may well have reached the same conclusion. Ballard and Pleasant lived in the South at a time when whites were employing violence, disfran- chisement, and segregation to secure their supremacy, making criticism of racist behavior increasingly risky. A less hazardous ap- proach to justice was perseverance. Both men continued to apply, and Pleasant followed his successful application with requests for increased payments. He collected more than $960 in pension
income until his death in 1908 when he had reached his late
seventies. Little else about these two veterans’ last years is known,
but the extent to which their experiences, together with those of
their peers, are on record serves to highlight pension ofªcials’
actions.30
Not surprisingly, pensions allowed veterans to acquire posses-
sions. Almost half of the pensioners in the African-American sam-
ple who appear in the 1900 census owned their homes, whereas
just under one-third of men like Pleasant, who had received no
pension at the time of the U.S. census, were homeowners. Yet the
approach of death for the Civil War’s survivors became the central
concern of policymakers as the twentieth century began. Calling
attention to veterans’ advancing age and declining numbers, one
senator insisted that Congress’ duty was “assisting [veterans] while
they are on this side of the grave.” Did pensions play some part in
keeping recipients out of the grave?31
There was no clear and simple connection between income
and mortality in the late nineteenth century. The superior diet that
money could buy mitigated certain diseases, but variations in mor-
tality also depended on the state of public health and sanitation as
well as the halting progress of medical care. But one key reason for
seeking a link between pensions and mortality for African Ameri-
cans lies in the cpe samples: Previous articles have discovered that,
30 On conditions in the South, see especially Edward L. Ayers, The Promise of the New
South: Life after Reconstruction (New York, 1992), 132–159.
31
806.
Speech by Porter McCumber, Congressional Record, 59th Cong., 2d sess., Jan. 9, 1907,
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
e
d
u
/
j
i
/
n
h
a
r
t
i
c
e
–
p
d
l
f
/
/
/
/
3
8
3
3
7
7
1
7
0
4
9
0
2
/
j
i
.
.
n
h
2
0
0
8
3
8
3
3
7
7
p
d
.
.
.
f
b
y
g
u
e
s
t
t
o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
396 | LARRY M. LOGUE A ND PETER B LA NCK
despite elderly white pensioners’ disabilities, pensions were associ-
ated with longevity.32
Linking pensions to mortality, however, raises the question of
endogeneity. The longer veterans lived, the more opportunity
they had to apply for, and receive, a pension—as in Pleasant’s case.
Longevity might have contributed to pension income as well as
the other way around. Table 5, which uses proportional-hazards
analysis to explore the risk of dying that the African Americans in
the cpe samples faced after 1900, is designed to control for this is-
sue. Observation of each veteran’s risk of dying begins in mid-
1900, allowing inclusion of home ownership as a measure of eco-
nomic well-being from that year’s census, and ends with the next
major expansion of pensions in early 1907. To control for a recip-
rocal effect of pensions and longevity, pension income is “lagged”
by terminating it at 1900. That is, each veteran’s pension income
to that year is summed and converted to a logarithm. Hence, all
veterans included in the table had the same opportunity to apply
for a pension prior to observation.
Other independent variables include wartime health and
noncommissioned rank as they were employed in earlier tables.
Since turn-of-the-century cities were particularly deadly, a sepa-
rate variable identiªes black veterans who lived in cities with more
than 30,000 in population. Finally, age is included as a control
variable.33
With these characteristics included, Table 5 shows that pen-
sion income nonetheless had a clear inºuence on mortality: The
more pension income a black veteran had received prior to 1900,
the less likely he was to die before 1907. Note, however, that any
pension beneªt operated against the background of high overall
African-American mortality. The information available for these
veterans will not let us point to better diets, living conditions,
32
See S. Ryan Johansson, “Food for Thought: Rhetoric and Reality in Modern Mortality
History,” Historical Methods, XXVII (1994), 101–126; Gerald N. Grob, The Deadly Truth: A
History of Disease in America (Cambridge, Mass., 2002), 209–215; Costa and Kahn, “Public
Health and Mortality: What Can We Learn from the Past?” in Alan J. Auerbach, David Card,
and John M. Quigley (eds.), Public Policy and the Income Distribution (New York, 2006), 359–
399. On developments in African Americans’ health and mortality, see Kiple and King, Black
Diaspora, 187–207; Byrd and Clayton, American Health Dilemma, 325–357. On pensions’ effect
on mortality among white veterans, see Logue and Blanck, “Disability Policy and Mortality.”
Salm, “Effect of Pensions.”
33
1940,” Annales de Démographie Historique, 101 (2001), 33–64.
See Michael R. Haines, “The Urban Mortality Transition in the United States, 1800–
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
e
d
u
/
j
i
/
n
h
a
r
t
i
c
e
–
p
d
l
f
/
/
/
/
3
8
3
3
7
7
1
7
0
4
9
0
2
/
j
i
.
.
n
h
2
0
0
8
3
8
3
3
7
7
p
d
.
.
.
f
b
y
g
u
e
s
t
t
o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
BENEFIT OF THE DOUBT | 397
Table 5 Relative Inºuences on the Risk of Dying from 1900 to 1907, African
Americans in cpe Samples
variable
mean
hazard ratio
p
Hospitalized for wound
Severe/acute/chronic illness
Noncommissioned ofªcer
Age in 1900
Owned own home in 1900
Living in city > 30,000
Total pension income to 1900 (log)
Number of cases
Initial (cid:2)2 log likelihood
Improvement with covariates
.161
.115
.202
59.0
.426
.261
4.56
884
2,808.4
59.4
1.081
.931
.903
1.069
.876
1.379
.954
.678
.742
.566
.001
.367
.034
.033
medical care, or any other amelioration as an explanation for the
apparent pension beneªt. The evidence points solely to a human
toll of discrimination. If pension ofªcials’ evaluations had not been
susceptible to the race of applicants, some veterans’ time “on this
side of the grave” would have been prolonged.
The egalitarian principles of the Union army’s pension system
were ultimately compromised in practice. To be sure, the initial
failure of African-American veterans to apply for pensions was
temporary. When Congress made the law more inclusive of dis-
abilities, and when nonmilitary pensions for freedpeople became
an issue of widespread appeal, black veterans claimed their right to
a pension as insistently as did their white comrades-in-arms.
When they entered the process, however, black veterans
became enmeshed in the pension system’s devotion to what
might be termed moral cost containment. As discussed above,
partisan maneuvering affected pension awards, generating favorit-
ism to applicants living in some places and prejudice against those
who applied at the wrong time. But a broader political commit-
ment overshadowed these shifting tides of partisan politics:
Pension ofªcials were unalterably dedicated to upholding moral
dichotomy—worthiness versus unworthiness—as public policy.
The applicants who were trying to cheat the pension system were
not simply unworthy now; they had never been worthy. Pension
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
e
d
u
/
j
i
/
n
h
a
r
t
i
c
e
–
p
d
l
f
/
/
/
/
3
8
3
3
7
7
1
7
0
4
9
0
2
/
j
i
.
.
n
h
2
0
0
8
3
8
3
3
7
7
p
d
.
.
.
f
b
y
g
u
e
s
t
t
o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
398 | LARRY M. LOGUE A ND PETER B LA NCK
ofªcials were unceasingly on guard against the “cowards” and
“fraudulent malingerers” who “enlisted near the close of the war
for large bounties and did little actual service”; those who actually
deserved pensions were “the real soldiers of the war” who had
done “patriotically what [their] duty require[d].” Moreover, ad-
ministrators complained that their vigilance was handicapped by
the “army of attorneys . . . subagents, solicitors, and ‘grafters’”
who encouraged fraudulent claims, as well as by examining physi-
cians who were incompetent and unreliable.34
Thus besieged and betrayed, Pension Bureau ofªcials saw
themselves as the last line of defense for the public trust. Over-
whelmed by the number of veterans seeking a pension, Bureau re-
viewers apparently fell back on preconceived notions of what sorts
of people were most likely to be malingerers, making race once
again a factor in the reviewers’ rulings. A seemingly unassailable
policy of protecting the public and worthy pensioners became a
systematic denial of justice, and possibly shortened some veterans’
lives.35
APPENDIX: ESTIMATING VETERANS’ LONGEVITY
Investigation of pension-application rates calls for modiªcation of
the cpe samples. Any analysis of pension seeking must take into account
each veteran’s opportunity to apply, that is, how long he lived. Informa-
tion on veterans’ death dates is biased toward pension applicants. Be-
cause the primary means of ascertaining the date of death for survivors of
the war is from documents in a pension applicant’s ªle, examining the
behavior of veterans with known death dates would disproportionately
exclude non-applicants.
Enough death dates appear on widows’ pension applications and
other documents, however, to supply survival information for about 15
percent of veterans who never applied for a pension, and the informa-
tion points to a stark difference between applicants and non-applicants.
More than 80 percent of non-applicants died before 1890, whereas
fewer than 5 percent of applicants were dead by that year. To reºect
this difference and to produce a more appropriate population at risk
for
of applying, we have estimated years-to-live at mustering-out
34 Annual Report of Commissioner of Pensions, House Doc. 5/5, 54th Cong., 1st sess.,
1895/96, serial set 3383, 13–14; Annual Report of Commissioner of Pensions, House Exec.
Doc. 1/19, 53d Cong., 3rd sess., 1894/95, serial set 3307, 7.
35 For additional information about disability law and policy issues and this program of
study, see the Burton Blatt Institute: http://bbi.syr.edu.
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
e
d
u
/
j
i
/
n
h
a
r
t
i
c
e
–
p
d
l
f
/
/
/
/
3
8
3
3
7
7
1
7
0
4
9
0
2
/
j
i
.
.
n
h
2
0
0
8
3
8
3
3
7
7
p
d
.
.
.
f
b
y
g
u
e
s
t
t
o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
BENEFIT OF THE DOUBT | 399
Table A Percentage of Civil War Survivors Who Applied for a New Pension
by Race, cpe Samples (Cases with Known Death Dates)
general law
1862–1890
disability law
1890–1907
Whites
African Americans
Number of cases
54.7
28.6
21,741
91.5
93.1
7,518
“unknown” non-applicants based on the known years-to-live of non-
applicants of the same age and race. For example, since black non-
applicants mustered out at ages 20 to 25 lived an average of 15.3 years, a
black non-applicant with an unknown death date who was discharged in
1865 in his early twenties would be given an estimated death year of
1880.
Comparison of Table A and Table 1 reveals the effects of this adjust-
ment. Since the adjustment’s central purpose is to incorporate the be-
havior of veterans who died without pension records before 1890, its
main effect is to raise the number of non-applicants in the general-law
period. The adjustment adds 7,786 cases before 1890 and 486 afterward,
and lowers the general-law application rate by 10.8 percentage points for
whites and 8.9 points for African Americans, with smaller changes under
the disability law. The key hazard ratio in Table 2—that is, the effect on
the risk of applying under the general law associated with having been
born a slave—would be .741, using only veterans with known death
dates rather than the adjusted ratio of .755.
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
e
d
u
/
j
i
/
n
h
a
r
t
i
c
e
–
p
d
l
f
/
/
/
/
3
8
3
3
7
7
1
7
0
4
9
0
2
/
j
i
.
.
n
h
2
0
0
8
3
8
3
3
7
7
p
d
.
.
.
f
b
y
g
u
e
s
t
t
o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
e
d
u
/
j
i
/
n
h
a
r
t
i
c
e
–
p
d
l
f
/
/
/
/
3
8
3
3
7
7
1
7
0
4
9
0
2
/
j
i
.
.
n
h
2
0
0
8
3
8
3
3
7
7
p
d
.
.
.
f
b
y
g
u
e
s
t
t
o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
Download pdf