Special Studies

Special Study #8:
Foreign Social Security Developments Prior to the Social Security Act


by Lillian Liu*
Research Associate

SSA Historian's Office

December 2001

Social insurance was not an American invention. For the most part, it was a Continental innovation, appearing first in Europe in the late 19th century. By the time the United States adopted its first compulsory, contributory social insurance applicable to a large segment of the working population in the form of Federal old-age benefits (known today as Social Security) in 1935, some 20 nations around the world already had such a program in place, and another 30 or more had introduced at least one other social insurance program (for example, workers' compensation see Tables 1 and 2 below).

This paper documents the knowledge base of the policymakers responsible for introducing Social Security to this country and reviews the extent to which foreign experience in social insurance had an impact on their proposals.

In 1934, the Committee on Economic Security (CES) established by President Franklin D. Roosevelt naturally looked to this foreign experience in income security programs as background for their proposals. Their knowledge of the foreign developments and how they informed the Committee's design of its comprehensive economic security proposal for the country were referred to in the Committee's January 1935 "Report to the President." [1] The CES initially envisioned a wide-range of programs to provide Federally subsidized, regulated, or administered programs for workers, the elderly, for children, and for those facing unemployment, and the related economic and/or health hazards.

Two studies provide us a window to the Committee's knowledge base of foreign income security developments. A 1932 study by Barbara N. Armstrong includes exhaustive legislative histories and experiences of various income security programs abroad up to January 1931 (Armstrong 1932). Armstrong was a law professor at the University of California, and she was subsequently recruited by the CES as its director of old age security studies. Armstrong's book was well-known to the staff of the CES and was an influential source for the staff's own research. [2]

In 1937, following the passage of the Social Security Act in the summer of 1935, the CES staff summarized, in a single volume, "some of the most important information in the staff studies" prepared for the Committee, but limited to that specifically relevant to the Social Security Act (Social Security Board 1937b). The "factual background" of foreign systems presented therein, updated the Armstrong study through 1935 in some cases, but was focused primarily on unemployment and old age security.

This review combines both the comprehensive histories of foreign economic security programs reported in the earlier Armstrong study and the CES staff's updated report on unemployment and old age. [3] Section I presents two tables tallying a wide-range of income security programs in their latest configurations as known to the CES at the time–by program type and by country, respectively. The second and final section (II) highlights overall trends of foreign income security programs reported in both the Armstrong study and the CES staff report. It also reviews the extent to which their knowledge of programs abroad had an impact on policymakers of the day.

I. Overview of Foreign Income Security Programs, Circa 1935

Scope of Income Security Programs

Two tables below take stock of the comprehensive income security programs abroad as of early 1930s under varying legislative formulations. The scope of income security programs abroad extended beyond what is known as Social Security in this country–namely income security provisions for old-age, survivors, and disability (OASDI)–to include four other programs. All five together–OASDI, plus health/medical insurance (cash benefits for sickness and maternity, and medical care), unemployment insurance, workers' compensation, and family allowances–represent the five branches of "social security" as defined by the International Labor Organization (ILO).

Those five programs appear in two groupings in both tables. The first group, under the heading of "Social Security Programs," refers to counterparts of present-day U.S. Federal Government administered programs: old age, survivors, disability, and health/medical insurance. The second group, "Other Income Security Programs," includes unemployment insurance, workers' compensation, and family (or children) allowances (or subsidies to families for childbirths and/or care of dependent children). Both unemployment insurance and workers' compensation have, to this day, remained federally regulated and state administered programs in the United States. Family allowances were relatively rare in the early 1930s–included here to capture the very beginning of such programs anywhere.

Compulsory, Contributory Social Insurance vs."Non-Contributory," Means-Tested Assistance–The Two Main Types of Income Security

Almost all foreign income security programs in the early 1930s were in the form of either compulsory, contributory social insurance or "non-contributory," means-tested assistance. [4] Terms for those two types of income security programs are replicated here from the two studies cited (Armstrong 1932 and Social Security Board 1937b). Both Armstrong and the CES staff made clear distinctions between these two formulae a program that was in the form of compulsory, contributory insurance or of a non-contributory and means-tested variety. Their categorization of those two types of programs is widely adopted to the present day.

Of the first group–"Social Security Programs"–presented in the two tables in this overview, the pre-1935 programs generally took the form of either compulsory, contributory insurance or non-contributory and means-tested assistance. The second group, "Other Income Security Programs," consisted mostly of contributory insurance plans, though not always applied compulsorily or nationally. Table 1 below summarizes income security programs by either type.

Generally, contributory, social insurance imposes payroll taxes upon the covered working population in order to fund benefit payments to eligible contributors or insured persons. The so-called "non-contributory," means-tested programs typically provide benefits out of general tax revenues. Their sources of funding come from compulsory collections of taxes contributed by all taxpayers to the central government–albeit not earmarked for specified social programs. This unambiguous element of compulsion by the central government requiring citizens, residents, and/or insured working population to contribute to funding either form of social security programs reflects that both types of social security programs share a common set of principles grounded in the efficacy of collective responsibility, resource pooling, and risk sharing.

There are major differences between these two types of programs in terms of their respective impact on individuals. The most obvious distinction between the two forms of income security programs is whether such programs are limited to public assistance for the indigent. The tax-financed, "non-contributory," program is generally dedicated to only the needy and requires the potential recipient to undergo a means-test (namely, an income and/or asset test). The government presumably can better control the costs by setting the threshold of the means-test and the benefit amount at specified levels. There are several disadvantages to this type of benefit. First, due to the stigma for recipients of being on public assistance and the requirement of undergoing means-tests, such programs often have a low "take-up" rate and are not fully effective in reaching all the deserving poor. Means-tested programs have also been associated with what is generally known as "moral hazard"–namely, they may discourage savings and/or productive work for some individuals. Finally, funding can be a problem because many taxpayers are unwilling to support programs from which they do not expect to receive benefits.

Contributory social insurance is believed to remedy the disadvantages of tax-financed means-tested programs. There is no stigma for participants under social insurance because beneficiaries are themselves previously (or simultaneously) contributors to the program, they have therefore earned their "entitled right" to the fruits of their own contributions upon reaching old age, or becoming disabled, unemployed or injured. Making such programs available to the working population would therefore encourage productive work and not induce "moral hazard." The CES staff observed, in reference to old-age income security program, that there was a noticeable shift abroad from tax-financed old-age "assistance" to contributory insurance, because of "the widespread objection to the ‘means-test' basis of noncontributory old-age assistance and the desire to make grants available as of right (emphasis added) on arrival at old age." [5] Further, the CES staff found the government's presumed power to keep down the costs of means-tested programs to be limited. They reported that both domestic and foreign experience showed that there seemed to be an ever-growing proportion of aged becoming in need of relief at home and abroad. [6]

The CES staff concluded (while espousing the introduction of old-age insurance in the United States): "a system of contributory old-age insurance should be established at the earliest possible time to control the upward trend in expenditures for old-age assistance. Only through the method of preventing dependency through some form of cooperative thrift can the cost of relief be kept down. Old-age insurance financed in large measure from the contributions of workers and their employers would serve to protect an increasing proportion of our citizens from the hazard of old-age dependency and at the same time retard the mounting trend of assistance expenditures.... Since insurance benefits would be received as a matter of right, based on contributions related to wages, workers would be encouraged to maintain the best possible record of employment and wages in order to earn the right to a high rate of benefits. Savings or other assets would in no way reduce the amount of benefits received but would provide a means of augmenting income in the later years of life." [7]

Indeed the contributory social insurance programs would serve to stabilize the economy, so the CES staff believed, because "The certainty and regularity of benefit payments under an old-age insurance program would serve to stabilize in some degree the flow of consumption expenditures on the part of our superannuated population." [8]

Table 1 highlights the income security programs abroad by either type–contributory or "non-contributory"–under the broad grouping of "Social Security Programs" and "Other Income Security Programs." Under both groupings–"Social Security Programs" and "Other Income Security Programs"–those in the form of "non-contributory" are separated from contributory systems for easy comparison. The tally in Table 1 shows that by 1935, the contributory social insurance form of income security programs had predominated every major program category. Workers' compensation was the most prevalent of social insurance programs, adopted by all 54 countries that had introduced income security to protect workers from accidents and occupational diseases. The second most popular program was old-age security totaling 36 countries–with 27 countries already offering contributory, social insurance plans to at least certain segments of the working population. Twenty eight countries offered disability programs including four that opted for the "non-contributory" and means-tested format. Out of 23 countries offering survivor benefits only three chose the "non-contributory" type. As for health/medical insurance, unemployment and workers' compensation, all were in the form of contributory social insurance, albeit in some countries the participation was voluntary, not compulsory. Of the 17 countries that had introduced "family allowances" programs, a great majority of them (14) had only localized application for their contributory social insurance format; two had introduced voluntary coverage, and New Zealand was the only country that had national application but opted for the "non-contributory" and means-tested assistance (see also Table 2 below to identify countries under each program).

SEE Table 1: Summary of Major Program Categories--Number of Nations Having Programs of Each Type Prior to 1935

Summary of Program Profiles by Country

Table 2 identifies the countries that had adopted the many programs by 1935, under the same two groupings-"Social Security Programs" and "Other Income Security Programs." It shows how many programs each country had legislated in what format (contributory or "non-contributory") by 1935 and for how long. To the extent possible, the date of the first legislative act of a program in each country (in its latest configuration) and subsequent amendments are included here to show that at best each case was a "work in progress." For example, in France and the Great Britain, Armstrong documented the legislative act in 1910 as it first introduced contributory social insurance, followed by amendments to augment the program through 1930. [9] A review of Table 2 follows.

Social Security Programs

The heading "Social Security Programs" refers to nationally applicable legislation of income security programs in the form of either compulsory and contributory social insurance or "non-contributory" and means-tested assistance. Where the legislation was yet to take effect, [N/E] is assigned to the specific program for the country involved.

In the early 1930s, old age income security was the most prevalent in this group of programs, provided by 36 countries. Twenty seven of those countries had introduced nationally- legislated compulsory, contributory insurance; nine countries rely solely on "non-contributory," means-tested programs for old-age income security; and Uruguay had a "non-contributory" and means-tested program as its primary source for public relief for seniors, and also begun to introduce contributory social insurance to specified sectors of the working population. (It is entered in both contributory and "non-contributory" columns in both tables).

Contributory Social Insurance

Among the 27 countries (across Europe and extending to Latin America) that required contributory insurance for old-age benefits, a great majority (20) had a national program offering "general" coverage, namely, applicable to employees in industry and commerce or beyond. Six of the 27 are marked as [L] in the table, to indicate limited coverage of such programs to specified occupational groups, for example, railroad, bank, or public utilities, or government employees-reflecting the early stages of contributory insurance in those countries. (The United States, for example, had similarly "limited" coverage prior to the 1935 Social Security Act in the contributory insurance coverage for railroad workers.) One of the six was Uruguay, it is entered in the contributory column with [L] for limited coverage, and also in the "non-contributory" column for its general application. Of the 27 countries, Iceland had adopted a hybrid program that required contributions from all citizens aged 18 to 60 (therefore included in the "contributory" column in both tables), but benefits were provided to only the needy.

Nineteen of the 20 countries that had national, general coverage for old age insurance legislated a contributory program for disability as well-the lone exception being Spain. Again, of the same 20 countries that required contributory social insurance, 14 had introduced contributory insurance for surviving spouse and/or dependent children-Chile, Italy, Portugal, Romania, Spain, and Sweden excepted. Of the six countries with "limited" contributory insurance, four offered benefits of survivors and disability as well (including Uruguay). Only Ecuador and Switzerland extended their program to both the aged and survivors, but not the disabled.

Health insurance abroad generally provided cash benefits for sickness and/or maternity as well as medical care for the working population. Twenty seven countries had adopted such a program by January 1931. Most (22) opted for compulsory insurance, only five countries were marked [V] for voluntary coverage (depending on employers) with national government subsidies. [10] (This pattern of health insurance for the working population differs from the Federally-administered Medicare program in this country in two respects: Medicare insures only seniors aged 65 or over or the disabled at any age [11] with cost sharing for out-patient and in-patient care.) Estonia, Japan, Latvia, and Lithuania offered contributory social insurance for health and medical care but offered no program for old age, survivors, or disability.

In short, a total of 13 countries had compulsory coverage for OASDI and health insurance as well-including Greece and the Netherlands, where the health insurance legislation was yet to take effect as of January 1931.

"Non-Contributory," Means-Tested Assistance

Of the 10 countries that had "non-contributory," means-tested programs for old age, Denmark was the only country that had extended similar coverage to both survivors and the disabled; it also offered voluntary, contributory health insurance. Ireland required contributory insurance for disability and health care instead; Norway chose contributory insurance for health care. Uruguay's "non-contributory," means-tested program was applicable to only seniors and the disabled.

Other Income Security Programs

Under this heading are mostly contributory insurance programs with general application, compulsory or voluntary (labeled as [V]), except for family allowances.

Twelve countries had established national compulsory unemployment insurance by the end of 1935, including Canada, where its income security for old age was in the form of "non-contributory" and means-tested assistance. Ten other European countries provided subsidies to voluntary plans, including Denmark and Spain, where eligibility for receiving benefits were means- or earnings-tested. In Switzerland, however, about half of its cantons adopted compulsory insurance for unemployment, and the other half opted for federally-subsidized voluntary plans. Russia was the only country that had repealed its unemployment insurance by 1935, only because the government there declared that unemployment no longer existed upon its adoption of the Communist model of centrally-planned economy. (Russia was entered in Table 2 but not counted toward the total in Table 1 for this reason).

Workers' compensation was often the first national social insurance program to emerge in most countries due to increasing incidents of injuries and occupational diseases rising out of industrialization. Fifty four of the 55 countries presented here had opted such a program; Greenland being the only exception. By the early 1930s, the program typically offered cash benefits and medical care for short- and long-term disability caused by work-related injuries and/or occupational diseases. In many cases the benefits had already extended from employees in industry and commerce to agricultural and domestic workers. In a large majority of countries (34), employers were required to be insured against liabilities rising out of workers' compensation or to contribute to a Guaranteed Fund for such purposes. [12]

Finally, family allowances (or child endowment) plans grew in the post-WWI era out of the "living wage" movement or of the government policy to increase birth rates. From a total of 18 countries that offered such programs, New Zealand was the only nation that paid these allowances out of general tax funds as a matter of national policy to help low income families with dependent children and, indirectly, to subsidize employers paying low wages. In 15 other countries, this program was available in specified localities (indicated as [L-L]). In Belgium and France, employers voluntarily offered allowances, as an incentive for childbirths. In parts of Australia, for example, in New South Wales, where such a program was adopted to supplement the basic wage with allowances for each dependent child, this move enabled the employers to keep the basic wage down. [13]

SEE Table 2: Social Security Program Abroad by Country, Pre-1935

II. Summary of the Developments and Trends Abroad Prior to the 1935 Social Security Act

The two overview tables merely tally programs abroad in their latest form in the early 1930s. The most valuable contributions of the two studies cited here are their compilation of voluminous factual information and the authors' detailed analysis of data. The authors presented thorough reviews of the evolution of income security programs of all types. Foreign experiences were, in effect, the results of experiments in social laboratories around the world cultivated in incubators of varying combinations of social, economic and political makeup. Results from each experiment at every stage of its development, spanning over half a century in many cases, were subject to careful and critical examination. Repeatedly, those results showed that in most countries, demands for income security under various contingencies initially generated voluntary programs, organized by guilds, mutual aid societies, local communities, or trade unions–at first without government assistance.

With meticulous documentation, the authors showed how those earlier forms of income security failed to provide intended protections for workers, the aged, and children for their economic and physical well being. Through trial and error, income security programs in nations of varying sizes and political orientations followed a pattern of voluntary, localized, or limited coverage plans evolving into nationally applied programs, either means-tested or social insurance. [14] The CES staff's fastidious adherence to and unrelenting scrutiny of factual information, could justifiably pass the standards set by Frederick A. Hayek, Nobel Laureate, who held that "Past experience is the foundation on which our beliefs about the desirability of different policies and institutions are mainly based.... Yet we can hardly profit from past experience unless the facts from which we draw our conclusions are correct." [15]

Given this extensive knowledge base, to what extent did the CES recommendations follow the world-wide developments of social insurance? Authors of the CES Report pointed out that their proposals were deemed as most appropriate "under American conditions" and did not follow any single pattern or "copy European methods." They recommended "wide application of the principles of social insurance, but not without deviation from European models." [16]

A review of the Committee proposals as compared with findings in the Armstrong study appears to bear this out. The CES' overall preference for contributory social insurance over "non-contributory," means-tested assistance has been discussed earlier in section I of this paper. Their recognition of the advantages of the social insurance method, however, did not lead to indiscriminate recommendations for this form of income security program to be installed for all types of contingencies. The Armstrong study and the CES staff report were equally diligent, if not more so, in their research and assessment of domestic needs and appropriate forms of income security programs on U.S. soil.

The CES' proposal for a compulsory, contributory "old age annuity"–a social insurance program–in its January 1935 Report (later renamed as "Federal old-age benefits" in the 1935 Social Security Act), was grounded in an analysis of the advantages of the insurance method and the necessity of a single Federal system, preceded by lengthy discussions of contemporary needs of the aged in the United States. [17] The 1935 Committee proposals did not follow models in foreign lands to recommend provisions for survivors or the disabled in addition to old-age insurance, or suggest national government administered unemployment insurance. Further, it did not recommend a Federal contributory social insurance program for workers' compensation, despite the preponderance of this practice abroad. The CES concluded in the same Report, having examined the safeguards for workers' accident compensation laws practiced in all but four states in this country, that "substitution of continental European form of contributory accident insurance for our noncontributory accident compensation laws...or any other fundamental change is unwarranted." [18]

To CES members and staff, the ultimate test for the formulation of any income security program for the United States in 1935 was efficacy, not affinity to any model or ideology.


Armstrong, Barbara Nachtrieb. 1932. Insuring the Essentials: Minimum Wage Plus Social Insurance–A Living Wage Program. New York: Macmillan Company.

Committee on Economic Security. 1935a. "Report to the President of the Committee on Economic Security." 74 pp., reproduced in Reports Book, 1935-69. Volume 1. Baltimore, MD: Social Security Administration [n.d.]. Available online at: http://www.ssa.gov/history/reports/ces/ces.html

Committee on Economic Security. 1935b. "Supplement to ‘Report to the President of the Committee on Economic Security.'" 18 pp., reproduced in Reports Book, 1935-69. Volume 1. Baltimore, MD: Social Security Administration. [n.d.] Available online at: http://www.ssa.gov/history/reports/ces/ces.html

Social Security Board. 1937a. Principal Provisions of Foreign Compulsory Contributory Insurance Laws Covering the Risks of Old-Age, Invalidity, and Death. Washington, D.C.

Social Security Board. 1937b. Social Security in America: The Factual Background of the social Security Act as Summarized from Staff Reports to the Committee on Economic Security. Washington, D.C.: Government Printing Office. Available online at: http://www.ssa.gov/history/reports/ces/cesbook.html

Weaver, Carolyn L. 1982. The Crisis in Social Security: Economic and Political Origins. Durham, N.C.: Duke Press Policy Studies.

Witte, Edwin. E. 1963. The Development of the Social Security Act: A Memorandum on the History of the Committee on Economic Security and Drafting and legislative History of the Social Security Act. Madison: University of Wisconsin Press.


* The author is indebted to the SSA Historian, Larry DeWitt, for suggesting the study, introducing her to the sources cited, and for guiding her throughout the writing of this paper. His comments and suggestions have substantially improved the quality of the product.

1. See, Committee on Economic Security 1935a, p. 50; and Committee on Economic Security 1935b, pp. 7-9 and 13-14.

2. Armstrong's broad expertise in social programs also came to bear on CES' proposals on unemployment compensation and other issues (Witte 1963, pp. 30, 33n, 55-56, 82n, 83n, 116, and 122). She was also one of the contributors to the basic data for old age security studies of the CES staff report cited below (Social Security Board 1937b).

3. The two books cited here (Armstrong 1932 and Social Security Board 1937b) form the basis of this paper for purposes of documenting the knowledge base of the CES as its members and staff deliberated proposals for and provisions of the 1935 Social Security Act. (No attempt is made to review the unpublished individual reports on foreign programs prepared by the staff.) Those two studies represent perhaps the most exhaustive and valuable documentation of foreign programs in its day by U.S. authors. They also inform later-day scholars interested in the early stages of social security developments. (See, for example, Weaver 1982, pp. 33-35 and, especially, Table 3.3 "Compulsory old-age insurance and pension programs abroad, 1933" on p. 34).

The widely referenced book Social Security Programs Throughout the World (SSPTW) published by the Social Security Administration made its first appearance in 1937 under the auspices of its predecessor (Social Security Board 1937a). It cited a single 1933 source as reference–namely, the International Labour Office's Compulsory Pension Insurance (Studies and Reports Series M (Social Insurance) No. 10). This first issue included only provisions for old age, survivors, and disability in countries that opted for compulsory, contributory social insurance. Later editions of the same series (1999 being the latest available) have since included countries where "non-contributory," means-tested program is the preferred format, and also expanded the scope to include other branches of social security–health/medical insurance, unemployment insurance, workers' compensation, and family allowances. The biennial revisions of this publication are designed to update the programs in their current form and, as such, they typically date the first legislation of programs in each country, but provide no further details of those programs in their beginning years. Where inconsistencies exist between this first edition of foreign social security programs (Social Security Board 1937a) and the Armstrong (Armstrong 1932) and the CES staff study (Social Security Board 1937b), the latter two studies are the sources of information presented here without independent verification of the information cited, because the primary purpose of this paper is to document the knowledge base of CES members and staff.

4. There were minor exceptions to this pattern. For example, all citizens in Iceland aged between 18 and 60 were required to contribute to a poll tax to finance its means-tested old-age assistance; Denmark's and Spain's contributory unemployment insurance was established with voluntary participation but beneficiaries had to meet specified means or earnings tests for eligibility (see also Table 2 below) In addition, there were countries where both forms co-existed so that contributory old-age social insurance would cover the working population and tax-financed means-tested programs would target those who for one reason or another were not able to participate in the labor force and became indigent in their later years.

5. Social Security Board 1937b, p. 184.

6. Social Security Board 1937b, pp. 137-167; 181-188.

7. Social Security Board 1937b, pp. 197-198.

8. Social Security Board 1937b, p. 198.

9. It should be noted that dates of legislative acts in Table 2 refer to only income security programs in their respective latest configuration. In other words, as each country's programs evolved from voluntary, localized, or ‘non-contributory" means-tested" assistance, the legislative history or dates for the earlier forms are not recorded here. Thus, France's earlier experiment with "non-contributory" and means-tested program in 1905 is not reported in Table 2.

10. Compulsory, contributory health insurance for the working population is a common practice today among industrialized countries, excluding South Africa and the United States.

11. Medicare is available to the disabled at any age under two specified conditions: if the disabled (1) have been receiving disability benefits for more than 24 months; or (2) are receiving disability benefits for treatment of chronic renal disease (CRD)

12. Armstrong 1932, pp. 240, 249, and 571-572.

13. Armstrong, 1932, pp. 134-137; 142-148.

14. For old age security, among several principal European nations, France (1905) and Great Britain (1908) first introduced "non-contributory" means-tested programs, while Belgium and Spain experimented with voluntary plans with government subsidies. All abandoned their earlier programs and opted for contributory, social insurance instead within two decades (Armstrong 1932, pp. 407-432; Social Security Board 1937b, pp. 181-188). By the 1930s, several countries had gradually expanded social insurance coverage from one of limited sectors of the working population to include most of their labor force, even domestic and casual workers in some cases.

15. Cited in Weaver 1982, p. 4.

16. Committee on Economic Security 1935a, p. 50.

17. Armstrong 1932, pp. 376-438; and Social Security Board 1937b, pp. 139-180; 197-202.

18. Committee on Economic Security 1935a, p. 46.

SSA Historian's Office
December 2001