Housing Finance Fact or Fiction? FHA Pioneered the 30-year Fixed Rate Mortgage During the Great Depression? (2024)

Read this post on the International Center on Housing Risk website.

FHA introduced the 30-year, self-amortizing mortgage during the 1930s, which along with low-downpayments, helped raise the homeownership rate from 43.6% in 1940 to 61.9% in 1960.

Summary:

The standard description of the FHA’s contribution to boosting the homeownership rate from the 1930s to 1960 rate cites its use of 30 year mortgages with low down payment. This is largely fiction. Rather, historical documents reveal:

• The 30-year term loan was not authorized by Congress until 1948 for new construction and 1954 for existing homes;
• From FHA’s inception until nearly the end of the period, its average loan term was well under 30 years and averaged about 21 years from 1946-1954;
• From FHA’s inception until nearly the end of the period, its average LTV was not low, averaging 20 percent through 1954 and only dropping to about 10% by 1959;
• While downpayments of 20% and self-amortizing mortgages existed before FHA was established, it did contribute to the development of the modern mortgage finance system;
• An intense period of liberalization in FHA’s loan terms began in 1954 and continued until the end of the period and beyond;
• This liberalization of loan terms during the last quarter of the period:
  • Had little, if any, impact on the unprecedented gains in homeownership, as 53% of the 18.3 percentage point increase had occurred by November 1945 and only 8.2% occurred from
  • December 1956 until April 1960, the period during which most of FHA’s liberalization of credit terms took place;
  • Did result, by the late-1950s and early 1960s, in a tremendous increase in FHA’s claim and foreclosure rates; and
  • Worked to stimulate demand which has a perverse effect during periods of high demand. This liberalization was likely to be absorbed by higher prices thereby diminishing or eliminating any improvement in housing standards or affordability.

This analysis will consist of:

• Background;
• Facts;
• Conclusion regarding FHA’s statement;
• Conclusion regarding UI’s statement;
• Answers to additional relevant questions:
  • What was FHA’s share of the market during this period?
  • Who was FHA was serving?
  • What were FHA’s contributions?
  • What is the relevance for today?
• Conclusion

Background:

Below are two representative quotes about FHA’s financing terms from its inception in 1934 through 1960:

Housing finance polices (especially the low-down payment, 30-year, self-amortizing mortgage created by the Federal Housing Administration and support for veterans under the G. I. Bill) and increasing incomes helped raise the total homeownership rate from 44% at the end of the Depression to 62% in 1960 (The Urban Institute, Headship and Homeownership, June 2015). [1]

The agency pioneered the 30-year mortgage, which was very innovative during the 1930s (FHA Annual Management Report to Congress – FY2014). [2]

There is no disputing that 1940-1960 was a period of unprecedented gains in homeownership for all income groups and minorities.[3] Existing research has identified four primary reasons with no single factor having a sufficiently large direct effect to explain the aggregate increase.[4]

• Fifty-three percent of the increase from 43.6% to 61.9% occurred during from April 1940-November 1945. This outcome is somewhat surprising given World War II, however wartime rent controls made it advantageous for property owners to sell to owner-occupants.[5]
• As inflation pushed taxpayers into higher tax brackets, personal income tax benefits made owning a home preferable to renting. It is estimated this accounted for 25% of the change.[6]
• There was substantial growth in real income. This led to substantial changes in age composition, resulting in a reduction in the median age of homeownership. It has been estimated that changes in income and age composition accounted for more than half of change.[7]
• The development of the modern mortgage finance system, with one researcher attributing 12% of the overall 1940-1960 increase to federal agencies such as FHA.[8]

Thus, it is clear FHA “helped raise the total homeownership rate” from 1940 to 1960. But the Urban Institute (UI) and the FHA go further and attribute this result to highly leveraged loans, specifically ones with low down payments and 30-year terms.

It matters greatly whether FHA’s loans were highly leveraged (low down payments combined with a 30-year term) or moderately leveraged (larger down payments combined with substantially shorter loan terms). It also matters what FHA’s share of the market was during this period and who it was serving.

Knowing the facts relating to FHA’s financing terms during the 25 year period subsequent to its inception in 1934 has important policy implications for today. First, the homeownership rate increased from 63 percent in 1994 to 69 percent in 2004, a period of high and increasing leverage.[9] Second, with the today’s homeownership rate approaching the level attained in 1960,[10] many see increased leverage as the remedy (WSJ: Low down mortgages picking up; some see opportunity for the market to regain vigor).

Thus getting this history right is central to evaluating whether policies suggestions will in fact achieve the desired goal of encouraging sustainable homeownership and building wealth for low- and moderate income and minority households.

Facts:

Undertaking research of this type is best done using reliable primary and secondary documents from the period (many of the documents cited may be found in: AEI bibliography of historical housing finance documents). It is also important to use data sources appropriate to the context. For example, the U.S. Census Bureau data on the nonfarm ownership rate better reflects trends for the period before 1960 and particularly before 1930, than the overall homeownership rate cited by many including UI.[11] As shown by the chart below, prior to 1960, the overall rate was higher than the nonfarm rate. This was due to the farm rate being substantially higher than the nonfarm rate. By 1960 the migration from the farm sector had largely run its course and the two homeownership trends substantially merged. Both series demonstrate that 1940-1960 was a period of unprecedented increase in the homeownership rate—three times the size of the more recent 6 percentage point increase noted above and, in the case of the nonfarm series, double the size of the increase from 1900 to 1930. This has led at least one observer to note that the 1940-1960 nonfarm trend was in part a continuation of the pre-1930 nonfarm trend (blue line).[12]

It is also important to track down any data points between decennial censuses. For the period 1940-1960 it is fortunate that the Census Bureau published four intra-census home ownership data points (only 3 are shown on the chart below). This allows this 20-year period of unprecedented homeownership growth to be analyzed using smaller time periods.

Looking at the nonfarm ownership rate for the 1940-1960 period more closely:

• The nonfarm ownership rate for all races rose to 61.0% in 1960 from 53.4% in 1950 and 41.1% in 1940, up 19.9 ppts. (48%) from 1940 to 1960.[13]
• The nonfarm ownership rate for blacks rose to 38.4% in 1960 from 35.2% in 1950 and 23.9% in 1940, up 14.5 ppts. (61%) from 1940 to 1960.[14]

Thus 62% of the total 19.9 percentage point gain for all races occurred during the first half of the period and 78% of the 14.5 percentage point gain for blacks occurred during the first half of the period. It is noteworthy 53% of the overall increase in total homeownership rate had occurred by November 1945.[15] As noted earlier, this has been attributed to wartime rent controls. Clearly an examination of FHA’s LTV and loan term policies and practices during segments of this 20-year period is in order.

Conclusion regarding FHA’s statement:

FHA’s statement to Congress that it pioneered the 30-year mortgage during the 1930s is easily determined to be housing finance fiction. Congress did not authorize FHA to make 30-year term mortgages on newly constructed homes until 1948[16] and on existing homes until 1954.[17]

Conclusion regarding UI’s statement:

UI’s claim is “the low-down payment, 30-year, self-amortizing mortgage created by the Federal Housing Administration helped raise the total homeownership rate from 44% in at the end of the Depression to 62% in 1960.”

It has already been established that the 30-year loan was not even authorized by Congress until 1948 for new construction and 1954 for existing homes. An examination of published average loan level data finds the impact of the 1948 change on new home loan terms was muted—increasing from 19.5 years in 1948 to an average of 22.6 years over 1949-1954. Over the period 1946[18] to 1959 FHA’s unweighted average loan term was 22.3 years, reaching an average of 27 years in 1959 (see appendix 1).[19] It has also been established that 62 percent of the nonfarm homeownership increase had occurred by 1950 (the same percentage pertains using the total homeownership rate).

While UI does not define “low-downpayment”, its overall policy argument pre-supposes a definition at or near today’s FHA average level of 4%.[20] Published average loan level data dispels UI’s low downpayment assertion. Over the period 1946[21] to 1959 and 1946 to 1954 FHA’s unweighted average LTV was 81.6% and 79.7% respectively, reaching an average of 90 percent in 1959 (see appendix 1).[22] FHA’s average today is 96%.

Ninety-two percent of the increase in the total homeownership rate from 1940 to 1960 had already occurred by January 1957.[23] It is an interesting to note that a series of amendments to the National Housing Act liberalizing FHA lending standard began about the same time as the 1940-1960 homeownership growth spurt slowed dramatically (see chart below). Amendments were made in 1954, 1956, 1957, 1958, 1959, and 1961.[24]

Clearly, the shift to low downpayments and 30-year lending late in the period was not a significant factor in the overall homeownership increase for the total period. As we shall see, the substantial liberalization in loan terms that kicked in beginning in 1955 as a result of the series of amendments to the National Housing Act starting in 1954, would soon lead to dramatic increases in loan foreclosures.

A reasonable reader of UI’s statement would have been left with the mistaken impression that FHA lending during 1940-1960 included low-downpayment loans with 30-years terms for much if not the entire period. Leaving the reader with this impression is misleading since this premise is a fundamental foundation to UI’s later statement that “very tight credit availability [today and in the future] will retard homeownership.” It is this combination that qualifies this statement as housing finance fiction.

Additional relevant questions:

1. What was FHA’s share of the market during this period?

In the 1950s, the mortgage market was dominated by conventional lending.

• In 1950, 73% of the stock of first mortgages were conventional ones, with FHA accounting for 17 %.[25]
• In 1959, 72% of new loan originations with a dollar amount of $20,000 or less were conventional, with FHA accounting for 19%.[26]

2. Who was FHA was serving?

During the period in question, there is strong evidence that FHA borrower income exceeded that for conventional borrowers.

• As of 1950 conventional lenders had a greater propensity to make loans to households with an income below the median U.S. income of $3,319[27]
  • Stock of conventional loans: 42% below median income
  • Stock of FHA loans: 34% below median income
  • Stock of VA loans: 37% below median income
  • In 1960, FHA insured borrowers had a median income of $6,000, higher than the US median of $5,620.[28]

At the same time, average LTV and loan term respectively for conventional loans made by savings and loans (the predominant conventional lender) were substantially lower than for FHA loans:[29]

• Over the period 1950[30] to 1959 the unweighted average loan term for conventional loans was 16 years compared to 23.2 years for FHA (see appendix 2).[31]
• Over the period 1946[32] to 1959 the unweighted average LTV for conventional loans was 68.4% compared to FHA’s unweighted average LTV of 82.4% (see appendix 2).[33]

It bears emphasis that the predominant form of lending during the period was conventional, which relied on even less leverage and appears to have served a lower income group than FHA.

What were FHA’s contributions?

FHA made three major contributions during its first 20 years:

• While downpayments of 20%and self-amortizing mortgages existed before FHA was established, it did contribute to the development of the modern mortgage finance system, including risk rating every loan, applying rigorous compensating factors, a robust appraisal methodology, strict limitations on purchase money second mortgages, and application of ability-to-pay metrics, including the residual income method (Federal Housing Administration from 1934-1938: lessons for wealth building).[34]
• FHA’s default claims literally rounded to zero during its first 20 years, experiencing a cumulative claim rate of 0.2%.[35] A key contributor was it use of shorter term mortgages, which it described as “very much like buying a house and then renting it to yourself.”[36]
• Its “new mortgage system [was designed to provide]] a straight, broad highway to debt-free ownership,…[The goal being] “the possession of a home, free and clear of all debt at the earliest possible date, should be the goal of every American family.”[37] It relied heavily on an evaluation the prospective homebuyer’s current rent in relationship to the new monthly housing expense and the relationship between the earnings capacity (rent) on the home being purchased and the new month housing expense.

What is the relevance for today?

By the 1950s and 1960s the adverse impact of higher leverage on default rates was already being recognized both inside and outside FHA. The adverse impact of this change in housing policy has been largely ignored by UI and today’s FHA.

• “The major liberalization in terms for FHA insured mortgages which became effective between 1954 and 1961…played a role in increased rates of mortgage mortalities.[38]
• “Mortgages with low downpayments had much higher acquisition ratios than mortgages with higher downpayments.”[39]
• “The acquisition ratios for longer term mortgages exceed those for shorter terms.”[40]
• “During the 1950’s, foreclosure rates on VA, FHA and conventional mortgages did not diverge greatly. In the early 1960’s, however, rates on VA loans rose appreciably faster than those on conventionals, and rates on FHA’s rose especially rapidly. By 1963, foreclosure rates on VA loans were more than twice as high as estimated rates on conventionals, and rates on FHA loans were roughly four times as high.”[41]

Ignoring these early warning signs took an increasing toll on FHA borrowers—from 1977 to 2013, one in eight FHA borrowers lost their homes to foreclosure—well over 3 million families. This was 60 times the FHA claim rate from 1934 to 1954[42], as result of the major liberalization in terms for FHA insured mortgages which were enacted by a series of amendments to the National Housing Act from 1954 onward.

History provides a further lesson regarding how liberalization of credit terms effects demand. How this phenomena operates under liberalizing lending standards is as follows:[43]

As the market swings in favor of the seller, there is a tendency for more liberal credit terms to be absorbed in price advances rather than result in improved standards of housing….

Thus, the liberalization of terms easily becomes capitalized in higher prices….[t]he data indicate that from 1938 through 1941 borrowers in the same income groups paid higher prices when more liberal credit was available, borrowed larger amounts in proportion to their incomes, and incurred debt service burdens that absorbed more of their expected incomes.

In a buyer’s market, it seems that when there is an opportunity to select from a number of homes having about the same price and quality, more liberal credit probably raises housing standards; but in a seller’s market, when choice is restricted and the seller virtually dictates sales terms, more liberal credit is likely to be absorbed in price with probably a reduction in housing standards (emphasis added).

There is a recent, real-life example of how liberalized credit gets absorbed in price. In January 2015 FHA reduced its annual mortgage insurance premium by 0.50%. This had the effect of increasing buying power by 6.9 percent. The following reaction was typical:

‘Lots of people have been locked out of the market, particularly lower-wealth borrowers and borrowers of color, by the high prices at FHA,’ said Julia Gordon, director of housing finance and policy at the Center for American Progress, a group affiliated with Democrats. The premium cut ‘does put homeownership within the reach of more people.'[44]

However, the result was not as Ms. Gordon predicted.

New groundbreaking research, involving a review of over 2½ million first-time homebuyer loans,[45] demonstrates that FHA’s recent action to reduce its mortgage insurance premium did little to expand access to middle-and lower-wealth borrowers. Instead the benefits were largely captured by the National Association of Realtors and other housing interest groups, as the premium cut was largely capitalized into the purchase of higher priced homes.

In April and May, 2015 the median price of FHA-insured homes to first-time buyers paying the lower premium went up by about 5% more than GSE- and VA-insured homes. This suggests that half or more of the extra 7% in buying power was used to purchase either larger, more expensive homes or that prices rose in response to increased demand pressure. Further the resulting increase in FHA‘s volume appears to be largely a zero-sum game with 93 percent of the share pickup coming at the expense of Fannie Mae and Rural Housing Services, its closest competitors.

Link to full research results (link live on 6.25.15): FHA’s MIP cut: who was helped more? First-time homebuyers or the NAR.

Conclusion:

UI has pointed out that minorities will constitute approximately 80 percent of future housing demand.[46] However, if UI once again prevails in its calls for liberalized underwriting,[47] millions of new homebuyers will find that “‘Mortgage’[will] became just another word for trouble—an epitaph on the tombstone of their aspirations for home ownership.”[48]

This may be demonstrated with a thought experiment. To raise Black and Hispanic homeownership rates to the non-Hispanic white rate of 72 percent would require the origination of 10 million FHA-style 30-year loans totaling $1.7 trillion. These loans would have:

• An average credit score of 630[49]
• A downpayment of 3.5%, and
• A median debt-to-income ratio of 42%

AEI’s National Mortgage Risk Index indicates that loans with these characteristics would have a default rate of nearly 40% under the severely stressed conditions, more than three times the stressed default rate for the full cohort of recently originated home purchase loans.

These loans would present unacceptably high risks to homebuyers, neighborhoods, lenders, mortgage insurers, and taxpayers. Tragically, they will not reliably build wealth for low- and moderate-income and minority homebuyers.

While homeownership rate differences among income levels and minorities and whites are worthy of national discussion, any discussion regarding potential sustainable home lending solutions must be fact-based. Lower income and minority borrowers’ homeownership opportunities are ill-served by the Urban Institute and FHA clinging to a mythical link between high leverage and the growth in homeownership during the mid-20th century. The data demonstrate that continuing to espouse the use of more leverage just repeats a failed housing policy in place since the late-1950s.

Policy lessons from 1940-1960 to sustainably grow the homeownership rate and reliably build lower-income and minority wealth:

• Use moderate leverage combined with robust income growth
• By relying on moderate leverage solutions, such as the Wealth Building Home Loan developed at AEI, provide such a path safely and reliably
• Provide tax incentives that reduce leverage rather than promote higher debt

Appendix 1:

Average loan-to-value ratios and loan terms on new FHA loans remained quite moderate through the mid-1950s, but thereafter began to rise rapidly.

Average LTV and loan term (1946-1959 only) respectively for FHA loans:[50]

1937: 74 percent
1939: 79 percent
1940: 80 percent
1941: 81 percent
1942: 82 percent
1946: 79.5 percent and 19.2 years
1947: 77.8 percent and 19.2 years
1948: 77.3 percent and 19.5 years
1949: 80.1 percent and 21.3 years
1950: 81.9 percent and 22.7 years
1951: 79.1 percent and 22.5 years
1952: 78.1 percent and 20.6 years
1953: 80.3 percent and 21.2 years
1954: 79.9 percent and 21.4 years
1955: 84 percent and 24 years
Increases due to changes made by 1954 amendments to National Housing Act
1956: 82 percent and 24 years
1957: 82 percent and 24 years
1958: 88 percent and 26 years
Increases due to changes made by 1957 amendments to National Housing Act
1959: 90 percent and 27 years
Increases due to changes made by 1958 amendments to National Housing Act

Appendix 2:

Average loan-to-value ratios on new conventional loans remained quite low through the mid-1950s. As changes to the National Housing Act took effect, FHA’s LTVs and terms increased and this put competitive pressure on conventional lenders.

Average LTV and loan term respectively for conventional loans:[51]

1950: 67 percent and 13 years
1951: 65 percent and 15 years
1952: 66 percent and 16 years
1953: 66 percent and 15 years
1954: 67 percent and 16 years
1955: 70 percent and 16 years
1956: 70 percent and 16 years
1957: 69 percent and 17 years
1958: 71 percent and 18 years
1959: 73 percent and 19 years

Editor’s note: The original article was revised on September 14, 2015.

NOTES

[1] Goodman, et. al., Urban Institute, Headship and Homeownership, June, 2015, http://www.urban.org/research/publication/headship-and-homeownership-what-does-future-hold.

[2] http://portal.hud.gov/hudportal/HUD?src=/program_offices/housing/hsgrroom/fhaamr.

[3] US Census Bureau, Census of Housing: 1960, vol. 1, part 1, tables G, W, and X.

[4] White, Snowden, Fishback (editors), Housing and Mortgage Markets in Historical Perspective, 2014.

[5] Fetter, The Home Front: Rent Control and the Rapid Wartime Increase in Homeownership, NBER, 2013.

[6] Chevan, The Growth of Homeownership: 1940-1980, Demography, 1989.

[7] Ibid.

[8] Supra, White.

[9] US. Census Bureau.

[10] Ibid.

[11] The nonfarm homeownership rate increased by 9.5 percentage points from 1900 to 1930, with more than half occurring from 1920 to 1930. While beyond the scope of this paper, the US government was involved in promoting homeownership through the Better Homes Movement that began in 1923. The increase from 1900 to 1930 was followed by a decline of 4.9 percentage points from 1930 to 1940.

[12] Supra, White.

[13] Supra, US Census Bureau, Census of Housing: 1960.

[14] Ibid.

[15] Fisher, Market for Homes in Fee, NBER, 1951. Ernest Fisher was the chief economist at the FHA during the 1930s.

[16] Congressional Research Service, A Chronology of Housing Legislation, 1892-2003, 2004.

[17] John P. Herzog and James S. Earley, Home Mortgage Delinquency and Foreclosure (Cambridge, MA: National Bureau of Economic Research, 1970), www.nber.org/books/herz70-1.

[18] Earliest data found.

[19] Supra. Herzog. Annual data is provided for new and existing homes. The two data points were averaged for each year, as the new-existing volume split was about even.

[20] Supra, Goodman, AEI National Mortgage Risk Index.

[21] Earliest data found.

[22] Supra. Herzog. Annual data is provided for new and existing homes. The two data points were averaged for each year, as the new-existing volume split was about even.

[23] St. Louis Fed, Historical Homeownership Rate in the United States, 1890-Present, https://fraser.stlouisfed.org/scribd/?item_id=5808&filepath=/docs/publications/histstatus/hstat1970_cen_1975_v2.pdf&start_page=39#scribd-open.

[24] McFarland, FHA experience with mortgage foreclosures and property acquisitions, 1963, GPO.

[25] Ratcliff, Residential Finance, 1950, 1957, Social Sciences Research Council in conjunction with U.S Census Bureau. Richard Ratcliff was an economist at the FHA during the 1930s.

[26] Survey of Current Business, US Census Bureau, 1960.

[27] Supra, Ratcliff.

[28] Supra, McFarland and US Census.

[29] Supra. Herzog.

[30] Earliest data found.

[31] Supra. Herzog.

[32] Earliest data found.

[33] Supra. Herzog.

[34] FHA has been rightly criticized for condoning segregation based on race and national origin and enforcing local customs and practices including racial and ethnic covenants through its underwriting practices. That said, a total of 1¼ pages of its five hundred page Underwriting Manual consisted of references to racial and ethnic customs and practices. FHA, FHA Underwriting Manual, February 1938 revision, GPO.

[35] Supra, Herzog, Home Mortgage Delinquency and Foreclosure, 1970.

[36] FHA, How You Can Build, Buy, or Refinance Your Home, 1935, GPO. This was FHA’s first consumer brochure.

[37] FHA, How to Have the Home You Want, 1936, GPO.

[38] Supra., McFarland.

[39] Ibid.

[40] Ibid.

[41] Supra. Herzog.

[42] Ibid. Herzog.

[43] Fisher, Financing Home Ownership, NBER, 1951 http://papers.nber.org/books/fish51-1. Ernest Fisher was the chief economist at the FHA during the 1930s. See also Ratcliff.

[44] http://www.bloomberg.com/news/articles/2015-01-07/obama-said-to-announce-cut-in-fha-mortgage-insurance-premiums.

[45] The study controlled for variation in CLTV, FICO, total DTI, seasonality, and property state. The AEI International Center on Housing Risk maintains a national mortgage database which covers an estimated 98% of gov’t-guaranteed mortgages for home purchases (about 80% and 85% respectively of all purchase loans and all owner-occupied purchase loans by count). With a nearly complete census of gov’t-guaranteed loans, the NMRI data allow for accurate, timely, and in-depth coverage of purchase mortgage trends.

[46] Supra. Goodman.

[47] Temkin, A Study of the GSEs’ Single-Family Underwriting Guidelines, Urban Institute, 1999. http://www.urban.org/publications/1000205.html, 1999. Study documented the easing of standards by the GSEs through 1998 but also noted that “The GSEs’ guidelines, designed to identify creditworthy applicants, are more likely to disqualify borrowers with low incomes, limited wealth, and poor credit histories; applicants with these characteristics are disproportionately minorities.” HUD relied on this study when it greatly expanded the affordable housing goals in 2000.

[48] Supra. FHA, How to Have the Home You Want.

[49] A 2007 Federal Reserve report to Congress on credit scoring found that the median scores for Blacks, Hispanics, and Whites were 618, 670, and 737 respectively (medians interpolated from the Fed data) http://www.federalreserve.gov/boarddocs/rptcongress/creditscore/.

[50] Supra. Herzog (1946-1959), Supra, Fisher, Financing Home Ownership (1937-1942), Fisher, The Mutual Mortgage Insurance Fund, 1956 for new and existing construction weighting for years 1946-1954.

[50] Ibid. Herzog.