MARKETING 2 MILLION HOMES A YEAR

The marketing of 2 million homes a year will present real challenges to the housing industry and to Government. Data are lacking on many of the most important variables which will affect the willingness of people to buy or rent new homes. These include changes in the distribution of incomes, mobility, the income and price elasticity of demand for housing, and the extent of the required or desired movements to suburban areas. Intensive research on housing markets is necessary if we are to achieve our housing goals in an orderly way. Despite these limitations in our knowledge concerning the housing market, some broad relationships may be outlined with respect to the immediate future.

There appears to be a wide measure of agreement that approximately 1 to 1.2 million homes can be sold or rented under economic conditions similar to those of 1953 and with the then-available FHA and VA credit aids. Leaders of the National Association of Home Builders, the National Association of Real Estate Boards and other trade organizations have expressed such judgments repeatedly during the past year." These estimates accord with detailed projections made in the forthcoming 20th Century Fund study. Table 10 shows the income distribution of families of 2 or more persons in 1949-51, and also shows 42 million families distributed in proportion to 1951. The startling increase in incomes between 1949 and 1951 with its consequences for future housing markets are apparent. The number of families with incomes of over $4,000 increased from 10.7 million in 1949 to 16.8 million in 1951. This should indicate at least 20.2 million such families before 1960. This number, at turnover rates for families purchasing homes might sustain a market of 1 million homes a year. In addition there are demands created by single persons, undoubling and mobility.

This table also emphasizes the direct relationship between full employment and housing markets. Any slackening in employment or economic growth tending to recreate the income distribution of 1949 would sharply reduce the possibilities for sustaining large volumes of new-home sales and rentals. It would normally result in price declines, decreases in construction and increases in vacancies which would defer or prevent the development of an adequate housing supply and the replacement of substandard housing.

Current marketing practices are revealed by table 11 which shows the distribution of incomes of families of two or more persons in 1951 in contrast to the incomes of buyers of new FHA insured homes.19 The table indicates that substantially no homes were sold to families in income groups comprising 30 percent of all families and that only a third of FHA buyers were in income groups representiing over half of the market. This is not fully representative of all new residential construction since it appears that non-FHA homes include more higher-priced homes, and more lower-priced homes, although most of the latter may not meet FHA standards and many of them may be substandard when built. In addition, a very large proportion of FHA homes in the lowerprice brackets are purchased by families who will spend more than 25 percent of their income for housing. Indeed some of these families are spending nearly 50 percent of their incomes for housing. Of the FHA buyers with incomes between $200 and $250 per month, 97 percent were spending more than 25 percent of their income for housing. A safe rule would be that families should not spend more than 20 percent of their incomes for housing and most families above the lowest income groups spend substantially less than this ratio. The average FHA sale requires 19.7 percent or less of the purchaser's income (for all housing expenses) in all income brackets."

These FHA-insured sales are not necessarily unsound since they apply to few families numerically (less than 30,000 units below $3,600 income levels in 1952) and many of these families are clearly living from accumulated funds or are families whose homes are being purchased for them by others. It would be so

18 Cf. National Association of Home Builders, NAHB Correlator, February 1954, pp. 4–6, and New York Times, January 22, 1954, p. 44; National Association of Real Estate Boards, speech of Charles B. Shattuck, president, New York Times, November 11, 1953, p. 48; Miles L. Colean, New York Times, November 13, 1953, p. 42.

19 Data are sec. 203 homes. Similar income data are not available for other titles. The price distribution of all FHA new units is wider than that here used.

20 FHA data on monthly housing expenses and income-expense ratios have been used throughout this report. These data have been criticized as understating housing expense. Despite this weakness, these data are among the best and most consistent in the field, and the only adequate source on this subject. Wider publication of local data by FHA would be a distinct service.

cially unwise and economically disastrous, however, if any substantial proportion of our families began to purchase or rent homes which required such high expenditures for housing as to prevent normal expenditures for food, clothing, medical expenses and other necessities. If purchases involving more than 25 percent of income for housing are excluded from the FHA experience table, the results would be more representative of the sales and rental possibiliities of enlarged private housing production under present Government-aid systems. This distribution is shown in the third column of table 11. Limitations of the data are described above. This column suggests that 80 percent of new private construction is produced for income groups representing half of the market.

If a 20 percent ratio of income to housing expense were to be utilized, these conclusions would be even more apparent and might well extend to larger proportions of the market. In addition it should be noted that national housing market data tend to set low-priced homes of the South and West against relatively higher incomes of the North and East. When new house prices are compared with family incomes on a city-by-city basis, we rarely find situations in which homes are offered at prices within the means of more than 40 percent of the families; and these are in the main already well housed and not in the market in any urgent sense for a new home. Finally, many families in these middle- and lower-income groups cannot use the lowest-priced houses produced because such homes are universally 2-bedroom homes and most of our families have more than 1 child. The disparity between incomes and housing prices is thus widened further by the factors of location and family size.

1 Unpublished HB tabulations of owner and renter occupied units, United States Census 1950. Families not reporting certain housing items are excluded from the tabulations.

United States Census, Series P-60, No. 12, 1953. These are families of 2 or more persons, and exclude single person families. The former are believed to be more representative of heads of households shown in the 1st column.

TABLE 11.-Family incomes, 1951, and incomes of buyers of new FHA-insured homes, 1952

A second approach to the problem of marketing a high volume of new homes is summarized in table 12. The method here used is a modification of the method developed by the National Housing Agency in its 1944 projection of housing needs. Basically, it involves a filtration of the 1950 supply, elimination of losses due to clearance and demolition, and the addition of new units during the 1950-54 period. The result is subtracted from a future distribution based upon current or past experience. Limitations in available data and in the number of market factors which can be treated in the model suggest that this method can be used only to identify broad magnitudes and relationships and should not be used to forecast actual market trends. Detailed calculations are shown in the appendix.

TABLE 12.-Projections of nonfarm housing needed by price class under different economic assumptions and with low filtration and clearance rates and high construction volume, 1955-60

1 Detail of method in appendix E. See text note for results of alternative assumptions regarding filtration. Other assumptions, demolition of 2.5 million substandard units, 5 percent filtration, moderate current losses, 6 million units built 1950-54 and 8.8 million units required 1955-59.

2 The 1960 families distributed by classes of rents and prices paid in 1950.

The 1960 families distributed by classes of rents and prices paid in 1950 and adjusted upward for changes in income between 1949 and 1951.

Negative quantities imply surpluses in the classes indicated. This presumes rapid filtration, price declines, and reduction of building in these price classes and the price classes immediately above. Under alternative assumptions regarding filtration. See text note and appendix.

The method does suggest these important conclusions:

1. The current excessive produ tion in higher priced houses, i. e., those priced to sell or rent at monthly costs of $75 or more, will materially narrow the market for such homes in future years. To the extent that these homes alter down to the prices between $60 and $100 per month, they may create a surplus of such homes under unfavorable conditions. This could result in excessive filtration of higher priced units and marketing problems for new homes.

2. There is a very large market for homes at prices and rents ranging from $60 per month downward, a market which could account for over 3 million units in the next 5 years. This is the market now not served by either private or public housing. In northern cities, it consists roughly of homes in the range of $40-$60 monthly cost, and in southern areas, of $25-$50 monthly cost.

3. Unless full employment and a steady rate of economic growth are maintained, there will appear positive surpluses in some price ranges, accompanied by softening of markets which will jeopardize all private housing production. On the other hand, a maintenance of full employment conditions could produce substantially greater markets than those shown in the table. Housing markets are highly sensitive to changes in family incomes.

4. There remains a large need for housing for low-income families, a need which is increased by 50 percent under less than full-employment income distributions. This need is at price and rent levels far below any which private construction can approach.

5. Other models of future income distributions suggest that very high levels of housing production could be maintained in all price classes if rates of redistribution of income comparable to those which occurred in 1949-51 could be repeated.

6. A stable demadn for about 400,000 units of higher priced homes appears in all estimates based upon this method. If the method overstates needs in the higher price groups, this demand will appear in the price groups of $70-$100 per month, just below the levels shown in the table."

These conclusions support the views widely recognized by industry leaders that future construction must serve the broad middle and lower income groups if it is to be maintained at high levels. Under favorable economic conditions, there appears to be a sustained demand for approximately 1 million dwellings a year in the price classes now produced by industry with current Federal aids. In addition there appears to be a need for approximately 500,000 units a year in price classes below the levels now served by the industry with present credit aids, and for 200,000 to 300,000 units per year for families requiring public housing. These nonfarm needs are supplemented by farm needs of unknown price distribution of 200,000 to 300,000 units per year.

FINANCING 2 MILLION HOMES PER YEAR

Both the construction and the marketing of new residential construction will depend upon the availability of an adequate supply of construction and mortgage funds on terms which will meet the needs of the industry and of consumers. The Federal Housing Administration prepared an estimate for the President's Advisory Committee on Government Housing Policies and Programs of the probable mortgage fund requirements for the period 1954-58, and probable volume of savings available for mortgage lending." These estimates are extremely conservative. They assume construction of only a million new units per year," and a decline in the total volume of mortgage lending," even under assumed conditions of prosperity in the rest of the economy. Selected measures from this report are presented in table 13. The average mortgage requirement is not to be interpreted as a function of prices alone. It is a product of the aggregate volume of new savings, repayments on existing mortgages, the volume and amount of mortgage loans on new and existing housing and prices on new and existing homes.

23

The same study estimates that $9 billion annually will become available for new mortgage loans of which $5.5 billion will be utilized by the low level of construction assumed. This would leave unutilized $3.5 billion. In the last column of table 13 these funds have been applied to FHA's estimate of the average mortgage amount for new homes during the period. At this amount, funds would be available to finance an additional 450,000 units per year, or a total of 1,450,000 units per year. The FHA estimate of new savings is probably low and the implication of its assumed low volume of construction is that mortgage interest rates will decline sharply and that funds will be readily available. If these estimates of available funds prove correct, full utiliaztion of such funds would tend to hold interest rates at higher levels. Thus even if savings available for mortgage loans (but a small part of total savings) were assumed to average somewhat higher than the FHA estimate, such funds would probably not be available freely and on favorable terms at construction volumes exceeding 1.5 million new units per year under current institutional arrangements for savings and mortgage lending.

The method assumes that all units in the existing supply will filter by 5 percent. In fact filtration rates will vary widely by location and price class, and recently built homes may not filter (i. e., suffer price or rent reductions) at all during the first 5 years of occupancy. If this is assumed to be true, construction requirements in the highest class interval above are sharply reduced and those in the next two classes are correspondingly increased. This would indicate a low demand (0.6) in the class over $100 under the first assumption, and a substantial demand (1.2) in the $75-$100 class. Under the second assumption, demand in the over $100 category would be the same (0.6) and in the $75-$100 class would jump to 3.4 million, or nearly 700,000 units a year. These shifts are shown in the appendix, table E.

Prospective Level of Residential Construction and Availability of Mortgage Money, 1954-58, Federal Housing Administration, 1953. These estimates and the following text deal with nonfarm financing.

Ibid., pp. 4, 12, 13.

Ibid., table 5.

44022-54-37

TABLE 13.-Increases in mortgage holdings, 1950–53; new residential construction and projections, 1954-58

1 HHFA, Housing Statistics, January 1953, p. 23.

FHA, Prospective Level of Residential Construction and Availability of Mortgage Money, 1954-58, table 6.

3 Ibid., table 5. Estimate is for average year in this period.

4 Computed from ibid., tables 5 and 6. Assumed average mortgage amount is the marginal rate assumed by FHA table.

These considerations suggest that under conditions of prosperity, ample mortgage funds may be available on favorable terms through existing channels for 1.2 to 1.4 million units a year. Additional flows of money will probably be required to meet housing goals of 1.8 million nonfarm units a year in the next few years, and in part to meet farm goals. The prospective large volume of savings, increases in corporate savings to meet corporate investment requirements, and reductions in defense financing requirements suggest that ample funds will be available in fields which have not financed housing construction in recent years. These include particularly institutional savings which have been utilized for corporate investments and institutional and personal savings which have been going into Government bonds. The fuller utilization of these savings channels may be expected to provide funds on terms acceptable to the housing market. If funds are to be diverted directly from other sources, the implied increases in interest rate would make impossible the marketing of larger volumes of new construction.

These conclusions are consistent with the marketing requirements suggested in the preceding section. A substantial market appears to exist in the higher income levels for new homes at rents and monthly prices corresponding to the present output of the industry. If larger volumes are to be produced, they must be produced at prices and on terms which will be available to income groups substantially below those now in the market for new homes. If such terms are available, and only if such terms are available, it will be possible to obtain the goal of a decent home in a suitable living environment for every American family.

APPENDIX

TABLE A.-Projections, substandard units remaining in use with housing construction at volumes of 1,200,000 to 1,400,000 units and rehabilitation of 400,000 units per year, 1955-70