THE IMPACT OF MONETARY POLICY ON RESIDENTIAL

CONSTRUCTION, 1948-58

WARREN L. SMITH, University of Michigan and Harvard University This paper was prepared in response to the questions:

How have the volume and price of nonfarm residential credit been affected by policies of the Federal Reserve Board during the period 1948-57? What principles can be drawn from this experience, which principles may be used to formula policies of the Federal Reserve Board during the period 1961-70? How should Federal programs be supplemented or modified to achieve a more balanced impact of Federal Reserve Board policies upon all major sectors of the economy?

I. INTRODUCTION

Residential construction has been sustained at very high levels since the end of World War II. In the decade 1948-57, 11 million private nonfarm dwelling units were started, an average of 1.1 million per year. This is substantially higher than the levels of construction reached during the housing boom of the 1920's, although, of course, the level of total national output has also been substantially higher than in the earlier period.

While the level of housing construction has, in general, been high, by past standards, it has also exhibited very substantial fluctuations from year to year, as can be seen in table 103. For example, housing starts increased by 37 percent from 1949 to 1950, dropped by 25 percent from 1950 to 1951, rose by 13 percent from 1953 to 1954 and by a further 9 percent from 1954 to 1955, and fell by 17 percent from 1955 to 1956. The volume of residential construction valued at constant prices has behaved in much the same way as housing starts, although the fluctuations have, for the most part, been less severe. is interesting to note that, in 1957, housing starts were only 9 percent higher than in 1948, whereas total outlays at constant prices were 36 percent higher, an indication of a substantial upgrading in the size and quality of houses built. The prices of new houses, as measured by the Department of Commerce implicit price deflator for residential nonfarm construction, rose by 24 percent, or only slightly more than the rise (22 percent) in gross national product prices generally. In real terms, residential non farm construction varied between approximately 312 and 5 percent of gross national product, a further indication of substantial fluctuation.

1 Valued at 1954 prices, outlays per start increased from $12,500 in 1948 to $15,600 in 1957. a rise of about 25 percent. The data of the HHFA on number of rooms, floorspace, and proportion of houses with garages for new houses covered by mortgages insured under sec. 203 of the National Housing Act corroborate this upgrading trend. See HHFA annual reports, 1948-56.

1948

1949

1950..

1951.

1952.

1953.

1954

1955.

1956.

1957. 1958

TABLE 103.-Data relating to residential construction, 1948-58

1 Total value of new construction, private and public, deflated by implicit price deflator for new construction component of GNP.

2 Seasonally adjusted annual rate for first 6 months.

3 Seasonally adjusted annual rate for first 2 quarters, deflated by average of consumers' and wholesale priceindexes.

Seasonally adjusted annual rate for first 9 months deflated by average of consumers' and wholesale price

indexes.

Sources: HHFA, Annual Reports; Housing Statistics; Survey of Current Business; Federal Reserve Bulletin.

Total outlays for new construction do not show the same instability as the residential component. In fact, total construction rose each year except for slight decline from 1955 to 1956. Residential construction seems to have followed, in a general way at least, a somewhat contracyclical pattern. In the downturns of 1949 and 1953-54, housing starts picked up and the momentum thus generated carried over in the form of very high levels of construction in 1950 and 1955. As inflationary pressures developed, the measures taken to counteract these pressures produced cutbacks in residential construction in 1951-52 and in 1956-57. However, both the forces producing the initial expansion and the means by which restrictive policies produced a cutback were different in the 1949-52 period and in the 1954-57 period, as we shall see.

Since our main interest is in the effects of monetary and financial forces on residential construction, we shall divide the period from 1948 to 1958 into 3 subperiods in which these forces worked in different ways and through different channels. First, we shall consider the period 1948-50, before the Treasury and Federal Reserve arrived at the accord of March 1951. Next, we shall take up the period 1951-52, when a moderately active monetary policy was employed while at the same time the mortgage market was subjected to the selective controls imposed after the outbreak of the Korean war. Finally, we shall turn to the period since 1953, during which the Federal Reserve has used its general monetary controls rather vigorously and these controls have exerted considerable effects on residential construction.

41549-59-46

II. THE PREACCORD PERIOD: 1948-50

Expenditures on residential construction are influenced by many factors in addition to the cost and availability of credit. Table 104 summarizes what seem to be the more important of these factors for the years 1948-57. It seems apparent that they had much to do with the rapid expansion of residential construction between 1948 and 1950, culminating in the record total of 1.35 million housing starts in the latter year. Household formation was proceeding at a very rapid pace during these 3 years, reaching a peak in 1950, after which it declined rather sharply. Although the number of doubled-up families declined between 1948 and 1950, it was still in excess of 2 million in the latter year. Real per capita disposable income at 1954 prices, increased by nearly $100 or about 6 percent in 1950, the largest year-to-year increase of the postwar period to date. And finally liquid asset holdings were large and widely distributed during the period 1948-50.

TABLE 104.-Factors affecting demand for housing accommodation, 1948–57

1 Median holding of liquid assets of families having some liquid assets, as shown by Survey of Consumer Finances, deflated by the implicit price deflator for residential nonfarm construction.

Sources: HHFA Annual Report, 1956; Survey of Current Business; Federal Reserve Bulletin; U. S. Bureau of the Census.

Thus, forces other than the cost and availability of mortgage credit seem capable of accounting for the rapid expansion of residential construction during this period. However, changes in the mortgage market were a contributory factor also. In part, these changes were the result of amendments to Federal housing legislation. The Housing Act of 1948 approved in August of that year, contributed significantly to the expansion of the housing market by raising the limits on the size, and increasing permissible maturities of mortgages insured by the FHA. The Housing Act of 1950, adopted in April, eased

2 Systematic discussions of the factors influencing the housing and mortgage markets are to be found in J. S. Duesenberry, Business Cycles and Economic Growth (New York, 1958), ch. 7 D. M. Blank and L. Winnick. The Structure of the Housing Market, Quarterly Jourval of Economics, LXVII. May 1953, pp. 181-208: L. J. Atkinson, The Housing Market, Survey of Current Business, May 1955, pp. 13-19: L. R. Klein, Economic Fluctuations in the United States. 1921-41 (New York, 1950), pp. 89-95. 130-132; and J. E. Morton, Urban Mortgage Lending: Comparative Markets and Experience (Princeton, 1956).

3 The 1948 act increased the maximum amount of a 90 percent FHA-insured mortgage from $5.400 to $6.300 and the maximum amount of a 90-80 percent mortgage from $8.600 to $9.500 (90 percent of the first $7.000 of appraised value and 80 percent of the remainder up to $11.000) and increased the maximum mortgage eligible for a 25-year maturity from $5.400 to $16.000. A new subsection was added authorizing mortgages up to 95 percent of appraised value and 30 years' maturity on smaller houses. Another significant provision authorized an increase from 4 percent to 4 percent in the interest rate on VA-guaranteed home mortenges if conditions in the mortgage market should require it. HHFA, Annual Report, 1948, pp. 74-76, 159-160.

credit terms still further on both FHA-insured and VA-guaranteed mortgages. Legislation passed in 1948 and 1949 authorized the Federal National Mortgage Association (FNMA) to purchase VA-guaranteed mortgages and vastly expanded its authorization for the conduct of secondary market operations in both FHA-insured and VA-guaranteed mortgages.

TABLE 105.—Characteristics of FHA-insured and VA-guaranteed mortgage loans on new houses, 1948-56

1 Median effective income, as tabulated by FHA, deflated with the implicit-price deflator for personal consumption expenditures.

Not available.

First 9 months.

Sources: HHFA Annual Reports; Survey of Current Business; Veterans' Administration.

Some evidence of the effects of the liberalization of FHA credit terms is given in table 105 which presents some of the characteristics of FHA-insured and VA-guaranteed mortgages in the years 1948-58. The table indicates that the average term of FHA-insured mortgages increased by 4 years between 1948 and 1950, and at the same time the loan/value ratio increased from 81 to 88 percent. The percent of VA-guaranteed mortgages involving no downpayment increased from 19 in 1948 to 54 in 1950. These changes represent a substantial easing of mortgage credit terms as to both maturities and downpayments. Table 105 shows that there was a slight drop in the ratio of monthly payments to income for FHA mortgage borrowers. This occurred as a result of the extention of maturities which reduced the size of monthly payments. It is especially interesting that there was a drop of nearly $200 in the median real income (in 1954 dollars) of FHA-insured borrowers between 1948 and 1950. This occurred at the same time that average per capita real disposable income (also in 1954 dollars) increased by $90 as indicated in table 104. Thus, it appears that, as a result of the easing of credit terms, the demand for residential housing reached down into lower income brackets between the 2 years.

The 1950 act raised the maximum loan guaranty under the VA program from $4.000 or 50 percent of the loan to $7.500 or 60 percent of the loan and raised the maximum maturity of VA-guaranteed mortgages from 25 to 30 years. HHFA, Annual Report. 1950, pp. 6-7, 138-140, 213-215.

TABLE 106.-Privately owned permanent nonfarm dwelling units started, classified by type of financing, 1948–58

NOTE.-Details may not add to totals because of rounding.

Sources: HHFA, Annual Report, 1956; Housing Statistics; Federal Reserve Bulletin.

Table 106 shows privately owned nonfarm housing starts classified according to type of financing. FHA-insured starts increased by nearly 200,000 units or 66 percent between 1948 and 1950. Unfortunately, separate data are not available for VA-guaranteed and conventional mortgages covering these years, but the table shows that the 2 together increased more than 250,000 units or about 40 percent. Data on mortgage recordings, which cover borrowings to purchase both new and existing houses show increases of 13 percent for FHAinsured mortgages, 64 percent for VA-guaranteed mortgages, and 35 percent for conventional borrowing. Thus, VA financing apparently experienced a large increase. Actually VA loan applications declined steadily from late 1947 to mid-1949 and then expanded rapidly. The main reason for this pickup was probably the large volume of secondary market support provided by the FNMA after that agency's operations were expanded, beginning in mid-1948. These operations picked up momentum in 1949 and 1950, as indicated in table 107, which summarizes FNMA operations for the period 1948-57. Net purchases of mortgages rose from virtually nothing in 1947 to $198 million in 1948, $652 million in 1949, and $575 million in 1950. It is apparent from the changes in the FNMA portfolio that purchases were heavily concentrated in VA-guaranteed mortgages.

Another factor which helped to account for the rapid expansion of mortgage credit in 1949 and 1950 was undoubtedly the change that occurred in credit conditions and interest rates generally. As shown in figure 24, the maximum interest rate on FHA-insured mortgages was 412 percent during 1948 and 1949 and until April 1950, when it was lowered to 414 percent by the authority of the FHA. The maximum rate for VA-guaranteed mortgages remained at 4 percent throughout the 1948-50 period. With these rigid ceiling rates in effect, the supply of funds available for mortgages depended somewhat on the level of yields prevailing on competitive investments, such as corporate and Government securities. As figure 24 indicates, during 1948, the yield on high-grade corporate bonds remained in the neighborhood of 2.80 percent, being maintained there by the Federal

See Annual Report of the Administrator of Veterans' Affairs, 1950, pp. 92–97.