This report was prepared by the NAHB Research Center, Inc., for the U.S. Department of Housing and Urban Development, Office of Policy Development and Research. The contents of this report are the views of the contractor and do not necessarily reflect the views or policies of the U.S. Department of Housing and Urban Development, the U.S. Government, or any other person or organization. Trade or manufacturers' names appear herein solely because they are considered essential to the object of this report.

New homes in the United States are produced in a variety of different ways. Conventional site-built construction has historically predominated, but factory built homes, especially "manufactured homes" built under the preemptive Federal "HUD code," also play a very important role. This report, for the first time, provides a comprehensive comparison of HUD-code manufactured housing, conventional site-built homes and factory-built modular homes along several important dimensions. The comparisons address industry structure, production cost, characteristics of occupants and purchasers, unit designs and construction materials, regulatory processes, code requirements, and buyer costs.

Many of the historical distinctions between manufactured homes and conventional homes have been disappearing. During the 1990's, as HUD-code homes have become larger, multi-section units have become more common than single-section units, and placements on private land have outpaced placements on rented land. At the same time, site builders have slowly been shifting away from construction of compact, relatively inexpensive entry-level homes in favor of larger homes aimed at move-up buyers. However, very important differences still remain. For example, conventional site builders continue to play a much greater role in land and site development than HUD-code producers, and the two groups market their homes to purchasers in entirely different ways.

Based on the success and significant recent growth in the HUD-code sector, the report also recommends strategies by which home builders can improve efficiency, reduce production costs and play a larger role in delivering affordable homes to buyers of modest means. The future may ultimately see a more creative blending of factory production technology with conventional home building activities. Drawing on strengths and talents of both sectors offers a very potent approach to improving affordability while meeting the needs of home buyers and the communities where they live.

Xavier de Souza Briggs

Deputy Assistant Secretary for

Research, Evaluation, and

Monitoring

Executive Summary i

List of Tables. vii

List of Figures viii

Chapter 1. Introduction.

Chapter 2. Overview of the Housing Industry and Recent Trends

2.1 INDUSTRY STRUCTURE

2.2 TRENDS IN HOUSING PRODUCTION

2.3 TRENDS IN SALES PRICES OF NEW HOMES

2.4 GENERAL ANALYSIS

Chapter 3. Characteristics of Conventional and Manufactured Homes

3.1 INTRODUCTION

3.2 COMPOSITION AND LOCATION OF THE U.S. HOUSING STOCK

3.2.1 Age of Housing

3.2.2 Regional Distribution of Housing and Market Shares

3.2.3 Community Characteristics and Land Tenure

3.3 HOUSEHOLD CHARACTERISTICS FOR CONVENTIONAL AND MANUFACTURED HOUSING

3.3.1 Age Composition

3.3.2 Level of Education

3.3.3 Household Income and Housing Expenditures

3.3.4 Recent Movers, Choice of New Housing and Reasons for Moving

3.4 DESIGN AND MATERIAL CHARACTERISTICS OF NEW CONVENTIONAL AND HUD-CODE HOMES

3.4.1 Sizes of Housing Units and Price per Square Foot

3.4.2 Design Features in New Manufactured and Conventional Homes

3.4.3 Design Features in Existing Manufactured and Conventional Homes

3.4.4 Construction Materials in New Manufactured and Conventional Homes

Chapter 4. Comparison of the Regulatory Processes for Industrialized and Site-Built Housing.

4.1 INTRODUCTION

4.2 REGULATION OF UNIT CONSTRUCTION

4.2.1 Site-Built and Modular Housing

4.2.2 Manufactured Housing

4.2.3 Findings and Implications

4.3 APPROVAL, DESIGN REVIEW AND INSPECTION

4.3.1 Site-Built Housing

4.3.2 Manufactured Housing

4.3.3 Modular Housing

4.3.4 Findings and Implications

4.4 LAND DEVELOPMENT, SITE-WORK AND INSTALLATION

4.4.1 Site-Built Housing

4.4.2 Manufactured Housing

4.4.3 Modular Housing

4.4.4 Findings and Implications

Chapter 5. Comparison of Code Requirements

5.1 INTRODUCTION

5.2 BUILDING REQUIREMENTS

5.2.1 General Findings

5.2.2 Significant Differences in Building Requirements

5.2.3 Differences in Coverage of Building Requirements

5.3 ELECTRICAL REQUIREMENTS

5.3.1 General Findings

5.3.2 Significant Differences in Electrical Requirements

5.4 PLUMBING REQUIREMENTS

5.4.1 General Findings

5.4.2 Significant Differences in Plumbing Requirements

5.5 THERMAL REQUIREMENTS

5.5.1 Comparison Methodology

5.5.2 General Findings

5.5.3 Other Differences between the HUD-Code and CABO MEC

Chapter 6. Comparative Cost Analysis

6.1 INTRODUCTION

6.2 HOUSING COST ANALYSIS

6.3 GENERAL RESULTS AND COMPARISON BY COST CATEGORY

6.3.1 Construction Costs

6.3.2 Land Costs

6.3.3 Overhead, Administration and Financing Costs

6.4 CONSUMER FINANCING

6.4.1 Financing Options and Analysis

6.4.2 Results of Financing Comparison

6.5 CONCLUSIONS

Chapter 7. Findings and Recommendations

7.1 GENERAL FINDINGS

7.2 CURRENT TRENDS AND A LOOK TO THE FUTURE

7.3 REGULATORY AND TECHNOLOGY RECOMMENDATIONS

7.4 RECOMMENDATIONS FOR SITE BUILDERS

7.5 RECOMMENDATIONS FOR PRODUCTION BUILDERS

7.6 CONCLUDING REMARKS

Appendix A: Comparison of Specific Code Requirements

A.1 BUILDING REQUIREMENTS

A.1.1 Areas Where the HUD-Code is More Stringent than CABO

A.1.2 Areas Where CABO is More Stringent than the HUD-Code

A.2 ELECTRICAL REQUIREMENTS

A.2.1 Areas Where the HUD-Code is More Stringent than CABO

A.2.2 Areas Where CABO is More Stringent than the HUD-Code

A.3 PLUMBING REQUIREMENTS

A.3.1 Areas Where the HUD-Code is More Stringent than the IPC

A.3.2 Areas Where the IPC is More Stringent than the HUD-Code

Appendix B: Cost Comparisons in Chapter 6

B.1 SITE-BUILT SINGLE-FAMILY HOUSES

B.2 MODULAR SINGLE-FAMILY HOUSES

B.3 MANUFACTURED HOUSES

B.4 Table 20: THE NORMALIZED COMPARISON

B.5 Table 23: THE FINANCING COMPARISON

Tables

Table 1: Manufactured Housing Sector Statistics for 1977, 1982, 1987 and 1992

Table 2: Year of Construction of Year-Round Occupied Housing Units by Type, 1995

Table 3: Regional Distribution of Total Housing Stock and Newly Constructed Housing Units by Housing Type, 1995

Table 4: Ages of Heads of Households by Type of Housing, 1987 and 1995

Table 5: Average Numbers of Rooms in New Conventional and Manufactured Homes, 1996

Table 6: Exterior Finish Materials on New Conventional and Manufactured Homes, Percent of Total, 1996

Table 7: Wall Height in New Conventional and Manufactured Homes, 1996

Table 8: Average Number of Window and Door Openings per Dwelling in New Conventional and Manufactured Homes, 1996

Table 9: Type of Glass in Windows and Sliding Glass Doors of Existing Conventional and Manufactured Homes, Percent of Total, 1993

Table 10: Appliances and Equipment in Existing Single-Family and Manufactured Homes, Percent of Homes, 1993

Table 11: Selected Design Amenities, Existing Conventional and Manufactured Homes, Percent of Total

Table 12: Use of Wall Sheathing Materials in New Conventional Single-Family Housing and Manufactured Housing, 1996

Table 13: Use of Floor and Roof Sheathing Materials in New Conventional Single-Family Housing and Manufactured Housing, 1996

Table 14: State Requirements for Construction of Site-Built Homes, Construction of Modular Homes, and Installation of Manufactured Homes

Table 15: Differences in Stringency of Selected Building Requirements

Table 16: Differences in Coverage between HUD-Code and CABO Building Requirements

Table 17: Differences in Stringency of Selected Electrical Requirements

Table 18: Differences in Stringency of Selected Plumbing Requirements

Table 19: Comparison of "Average" Homes

Table 20: Comparison of "Identical" Homes (same square footage and foundation cost)

Table 21: Foundation Cost for a 2,000 Square Foot Home

Table 22: Overhead, Administration, Financing and Related Costs as a Percent of Sales Price, by Type of Home

Table 23: Comparison of Financing of "Identical" Homes (2,000 square feet)

Table 24: HUD-Code Minimum Size Tubing and Pipe for Water Distribution Systems

Table 25: IPC Water Distribution System Design Criteria

Table 26: Minimum Sizes of Fixture Water Supply Pipes in the IPC

Table 27: IPC Drainage Fixture Units for Selected Fixtures and Groups

Table 28: IPC Drainage Fixture Units Based on Fixture Drain or Trap Size

Table 29: IPC Sizing of Horizontal Branches and Stacks

Table 30: Maximum Distance of Fixture Trap from Vent

Table 31: Characteristics of a Standard House

Table 32: Cost Percentages and Dollar Values for Site-Built Houses Used in the Chapter 6 Tables

Table 33: Assumptions for Manufactured House-Land Examples in Chapter 6 Tables

Figures

Figure 1: Conventional Single-Family Housing Starts, Housing Sales, Manufactured Home Shipments and Modular Home Production, 1980-1997

Figure 2: Manufactured Home Shipments, Placements and Year-End Dealer Inventory, 1980-1997

Figure 3: Manufactured Housing Market Share Based on Total New Housing Sales and on Total Housing Starts, 1980-1996

Figure 4: Average Selling Prices of New Homes by Type of Home in Nominal Dollars and 1996 Constant Dollars, 1980-1996

Figure 5: Average Selling Prices of New HUD-Code Homes in Nominal Dollars and 1996 Constant Dollars, 1980-1996

Figure 6: Placements of Manufactured Homes by Region as a Percent of Regional Single-Family Home Sales, 1980-1996

Figure 7: Location of Manufactured Housing, Owner-Occupied Housing and All Occupied Housing Inside and Outside Urbanized Areas, 1995

Figure 8: Park Placement for New Manufactured Homes, 1980-1996

Figure 9: Distribution of Housing Expenditure by Type of Housing, 1996

Figure 10: Reasons Given by Recent Movers to Owner-Occupied Units for Choice of Present Home, by Type of Unit, 1995

Figure 11: Reasons Given by Recent Movers to Owner-Occupied Units for Leaving Previous Unit, by Type of Unit, 1995

Figure 12: Median Square Footage of New Conventional and Manufactured Homes, 1980-1996

Figure 13: Average Sales Price per Square Foot for New Conventional Homes and Manufactured Homes in 1996 Constant Dollars, 1980-1996

Figure 14: Foundation Types for New Manufactured and Conventional Homes, 1996

Figure 15: Roof Shape and Roof Pitch for New Manufactured and Conventional Homes, 1996

Figure 16: Comparison of HUD-Code and 1995 CABO MEC Thermal Requirements

References

Introduction

During the decade of the 1990's, the United States has seen dramatic changes in the production of single family homes. The decade began with the housing industry approaching a cyclical trough that was reached in 1991 when single-family starts fell to a low of 840,000. As of 1996 starts had risen to 1,160,000 in a sustained period of recovery for the industry and strong growth throughout the economy.

But conventional site-built housing is only part of the story. An even more dramatic development over the same period of time has been the growing production of industrialized housing, most notably factory-built "manufactured homes" that are produced under a federal regulatory system and shipped throughout the U.S. Evolution in the manufactured housing or "HUD-Code" sector has been particularly rapid. There are many signs of this:

• Shipments of HUD-Code homes more than doubled from 171,000 units in 1991 to over 363,000 units in 1996. Output per firm and per plant are at historical highs.
• When HUD-Code and conventional homes are considered together, HUD-Code homes constituted over 24 percent of U.S. total housing starts and almost 32 percent of all new homes sold in the U.S. in 1996.
• Prices of HUD-Code homes have risen but remain well below prices of new site-built homes even after adjusting for house size, foundation and lot costs.
• HUD-Code homes are growing in floor area, double-section units are now more common than single-section units, and the share of new units placed in rental communities is declining.
• HUD-Code homes are increasingly being placed on permanent foundations and financed with 30-year mortgages rather than personal property loans.
• Technological innovations have made it possible to integrate the chassis with the floor system, and 2-story HUD-Code homes are now being built.
• Large conventional home building firms are becoming active in the HUD-Code sector through acquisitions or joint ventures.

These developments naturally raise questions about the underlying reasons for such strong performance in the manufactured home sector. Those questions lead to others, such as the potential for continuation of this trend, the longer-term significance of industrialization in new home production and its relationship to the "affordable housing" market, and the future role of conventional site-built construction and other types of factory-built housing within the overall new home market. This comparative study of industrialized housing and conventional home building was undertaken to improve understanding of recent developments in the manufactured home market and to identify efficiencies in that sector that may find application to more conventional forms of new home construction. Specifically, the study seeks to:

• Document and analyze the recent growth in industrialized housing,
• Assess technical, market and institutional factors contributing to the growth of industrialized housing, and
• Identify efficiencies that may be applicable to conventional site-built or modular housing.

This report draws on information gathered in a series of site visits and interviews with producers, regulators and others involved in production of manufactured and modular housing; information from a variety of site building firms; and review and analysis of existing published studies from numerous sources. It incorporates the most recent data on industrialized housing from the U.S. Bureau of the Census and the Manufactured Housing Institute (MHI) and presents extensive analysis of results from the 1995 American Housing Survey. Statistical data on housing characteristics and building product usage from the NAHB Research Center's annual survey of new home construction practices and data from a similar Research Center survey of manufactured housing producers are also used to document characteristics of site-built and HUD-Code homes.

The study is organized into seven chapters. Chapter 1 explains the purpose of the study. Chapters 2 and 3 provide detailed information about the products, producers and purchasers of each type of housing. Chapter 2 describes the overall structure of the site-built, HUD-Code and modular sectors of the industry and documents recent production and price trends in Chapter 2. Chapter 3 presents basic information about the stock of conventional and HUD-Code homes, the owners, occupants and purchasers of each type of home, and the design features, amenities, building products and materials that affect the cost and marketability of each type of home.

Manufactured housing is also regulated in a completely different way than site-built homes or modular construction, and this can affect design, construction and cost in each sector. Chapters 4 and 5 deal with these issues. Chapter 4 describes the regulatory processes governing unit construction, approvals and inspections, and land development, and Chapter 5 summarizes important substantive differences between the technical requirements of the HUD-Code and the prevailing model codes that apply to site-built and modular housing. Chapter 6 develops a comparative analysis of housing costs for different configurations of site-built, modular and HUD-Code homes, as well as comparisons of normalized costs of purchasing and monthly costs of home ownership under a variety of assumptions about land tenure, financing and other factors. Finally, Chapter 7 presents recommendations for all sectors of the industry intended to help improve efficiency and take advantage in other ways of experience in the manufactured housing sector. Appendix A and Appendix B present additional documentation to supplement the code comparison of Chapter 5 and the cost calculations of Chapter 6.

Overview of the Housing Industry and Recent Trends

This Chapter begins with a description of the principal sectors of the U.S. housing industry as it has developed in recent years up to the present time, and presents basic information about trends in production, market share and price. Subsequent chapters of the report build on this information.

2.1 INDUSTRY STRUCTURE

For purposes of this report the home building industry is treated as multiple overlapping sectors of production, each with its own approach to building and selling new homes. The principal sectors include site-built housing, manufactured (HUD-Code) housing and modular housing. Statistical information is generally presented from a national perspective even though the competitive overlaps within and between these sectors occur in regional and local markets as well, and all the sectors face some degree of competition from sales of existing housing units. Each sector is discussed below.

Site-Built Housing Sector. The home building business has historically been dominated by the construction of new homes on site through sequential fabrication and assembly of products, materials and systems into finished homes by skilled tradesmen and general laborers. Activities are planned and coordinated by experts with regulatory oversight at the local or state level of government. The resulting "site-built" sector of the home building industry is large and very diffuse. It encompasses not only the construction of houses but ancillary activities including land development, infrastructure planning and sale of the finished product as a complete package. In 1996 the two largest conventional home builders, Pulte Home Corporation and Centex Corporation, each constructed more than 10,000 detached homes. For the same year the top 10 companies built almost 75,000 detached homes, which represented about 6.5 percent of national housing starts. Firm sizes drop rapidly from there; for example, the 100 largest companies built an estimated 162,000 single-family homes in 1994.

Tremendous diversity and an unconcentrated, highly competitive economic structure are apparent when the site-built sector is viewed as a whole. Capital requirements are low and there are few barriers to entry or exit. For example, recent National Association of Home Builders membership information indicates that the site-building segment consists of some 50,000 active home building firms with average production of around 20 housing units a year. Typical firms are very small, with the majority building less than 10 units per year, and about 80 percent building less than 25 units per year. The broadest picture of all appears in the 1992 Census of Construction, which reported over 130,000 residential construction "establishments" with one or more employees, and another 210,000 residential construction establishments without employees. While these residential construction establishments also include firms exclusively involved in remodeling, they do not include the hundreds of thousands of special trade contractors used extensively by home builders as subcontractors performing carpentry, plumbing, electrical, mechanical and other work. The number of residential construction establishments in 1992 was not much changed from the number reported in the 1977 Census of Construction.

The level of site construction activity is reported by the Bureau of the Census as housing "starts" (when ground is broken for construction) and housing "sales" (homes for which a sales contract has been signed). Both statistics customarily exclude HUD-Code housing units, which are reported separately. There are usually many more housing starts than housing sales because about one-quarter to one-third of new site-built homes are started but not "sold." Rather, they are built under contract between an owner and a builder serving as general contractor. This difference between "starts" and sales" can be seen in Figure 1 on page below.

While there is always some level of demand for new homes as population grows, new households are formed and economic activity shifts from one area to another, the housing business has been characterized by powerful cyclical trends as well. As the economy moves into recession housing starts can drop abruptly, and as the economy recovers housing starts often rise very quickly. Since most houses are purchased with long-term loans, the demand for new homes is also very sensitive to interest rates and monetary policy. Site builders operating in this volatile environment have tended to protect themselves by minimizing fixed capital investment and making extensive use of subcontracting arrangements.

Manufactured Housing Sector. New homes can be and often are built partly or almost entirely in factories rather than on site. Factory construction offers many opportunities for economizing and increasing efficiency in the production process modeled after experience gained in other industrialized sectors of the economy. For many years the most common type of factory-built housing was the "mobile home," a narrow, lightweight technological descendant of the self-contained travel trailer that was designed to be towed from one location to another along public roads and hooked up for temporary use. This sector first achieved prominence in the 1960s and early 1970s. By 1976 mobile homes had come under regulation in the form of the pre-emptive federal "Manufactured Home Construction and Safety Standards" or "HUD-Code," and the era of modern "manufactured homes" began. Manufactured homes are required to be produced with a permanent chassis designed for over-the-road transportation. They are usually placed at the site on non-permanent foundations (e.g., block piers) and are almost invariably one-story units.

The manufactured housing sector has a profoundly different economic structure and way of doing business than the site-built sector of the industry. Production is much more concentrated in fewer firms than site-built home construction, and is exhibiting a trend towards consolidation that has not been observed elsewhere in the industry. Producers of manufactured homes have historically been focused on the production process itself and left land development and retailing activities to others, but vertical integration into retailing and operation of manufactured home parks or rental communities is taking place.

The level of economic activity in this sector is generally reported as units shipped from the factory (based on comprehensive production monitoring performed on behalf of HUD), or units placed for residential use (based on survey data). Production of manufactured homes, like site-built homes, is subject to cyclical trends and a sensitivity to interest rates. Compared with the site-built sector, however, larger capital investment and the more concentrated industry structure of HUD-Code producers leads to less flexibility in responding to changes in the level of demand and more incentive to maintain production in slow markets.

The two largest HUD-Code producers in 1996, Fleetwood Enterprises and Champion Enterprises, each built about 60,000 homes and together accounted for about 35 percent of total HUD-Code shipments for the year. The top four firms accounted for over 50 percent of 1996 shipments, and the top ten firms accounted for over 70 percent. On a broader scale, numbers of firms and plants dropped steadily from 1977 to 1992, while output per plant and output per firm both rose by large amounts over the period. Some key statistics are summarized in Table 1.

Table 1: Manufactured Housing Sector Statistics for 1977, 1982, 1987 and 1992

Manufactured Housing	1977	1982	1987	1992
Firms	306	261	207	155
Plants	597	516	395	286
Units Produced (Shipped)	267,289	238,820	232,823	210,453
Average Units per Firm	873	915	1,125	1,357
Average Plants per Firm	1.95	1.97	1.91	1.84
Average Units per Plant	447	462	590	736

Sources: Firm and Plant data from U.S. Department of Commerce, Bureau of the Census, 1992 Census of Manufactures, MC92-I-24D, "Wood Buildings and Mobile Homes - Industries 2451 and 2452." Shipments as reported by NCSBCS or MHI. Counts of firms and plants published by MHI differ somewhat from those based on the Census of Manufactures, but display similar trends.

Modular Housing Sector. Modular housing is the largest of the other segments of the housing industry, each of which are for the most part very small compared to the site-built and HUD-Code sectors. Modular housing includes factory-built homes that are delivered to the building site in largely complete form as multiple modules and placed by crane on conventional basement or crawl space foundations. Unlike HUD-Code homes, however, the design and construction of modular homes is regulated entirely by state and local building codes similar or identical to those that apply to site-built homes. Many modulars are two-story houses, and modular producers often report that they compete directly with site-built homes in terms of design and amenities. Modular homes are usually sold through small builders responsible for preparing the site and foundation as well as required finish work. These builders often construct modulars on land owned by the purchaser. The modular sector represents an intermediate form of new home production and distribution that is of significant interest for the present study.

Definitive information about the structure of production in the modular sector is lacking, but it is very clear that modular houses have never achieved the popularity of HUD-Code homes. A 1987 report estimated that about 152 firms produced modular homes, some operating multiple plants. Average production was estimated at between 300 and 400 units per firm, with the largest 25 percent of modular producers accounting for two-thirds of output. Most firms shipped to five or more states. To some degree modular production was found to be a regional phenomenon concentrated at that time in the northeastern and mid-Atlantic states, and to a lesser degree in the midwestern and southeastern states. More recent estimates of annual modular production vary widely, from around 25,000 to 100,000 homes depending on the source. The Bureau of the Census has only published estimates of modular production since 1992, and reports that from 1992 through 1996 modular production has ranged from 32,000 to 37,000 homes per year. The largest modular producer in 1997 was All American Homes of Elkhart, Indiana (2,300 homes), and the second largest producer was Champion Enterprises (1,631 homes), a firm that is much better known as a producer of HUD-Code homes.

A recent analysis of 1995-96 Census data on modular homes provides more information about how modulars compare to stick-built homes. For example, 26 percent of modulars were 2-story, compared to 48 percent of conventional stick-built homes; the median modular square footage was 1,560 compared to 1,950 for conventional homes. The modulars were more likely to have vinyl siding and less likely to have a fireplace or a garage than the conventional homes. Modular houses are disproportionately sited in the Midwest (45 percent of all modulars vs. 21 percent of stick-built) and uncommon in the West (6 percent of modulars vs. 25 percent of stick-built). Modulars were also more likely to be located outside of metropolitan areas (51 percent vs. 18 percent for stick-built homes). The report concluded that modular homes are geared more towards first-time and non-metropolitan purchasers than conventional homes.

Other Industry Sectors. There are several other, smaller sectors that constitute the remainder of the housing industry, though they are not focused on in the present study. These include log homes, pre-cut package homes and various types of panelized construction. Total production of all these types is currently estimated by the Bureau of the Census at less than 30,000 units per year. Producers are small and geographically dispersed.

2.2 TRENDS IN HOUSING PRODUCTION

Figure 1 below shows how the overall production of single-family housing has been divided among the conventional (site-built), HUD-Code and modular segments of the industry for the period 1980 to 1997. The Figure gives a good sense of the volatility in housing starts and housing sales, the relative shares of the new home market occupied by each industry sector, and the recent growth in manufactured housing. The data clearly shows the recession in the housing industry that reached bottom for conventional homes and HUD-Code homes alike in 1991. Since then both sectors have displayed strong recoveries. Yet although conventional housing starts and sales have grown in number, they have also dropped as a share of all new housing units. Manufactured housing shipments reached a peak in 1996 at 363,000 units. Modular production has fluctuated between 30,000 and 40,000 units per year since 1992, and has not evidenced the degree of growth of HUD-Code homes.

Figure 1: Conventional Single-Family Housing Starts, Housing Sales, Manufactured Home Shipments and Modular Home Production, 1980-1997

Production of manufactured homes is of particular interest for this study, and is documented in greater detail in Figure 2. The curves show substantial growth in shipments and placements during the overall period of recovery for the housing industry that started from the low point in 1991. Figure 2 also divides placements into single-section and double-section units, and shows that double-sections grew from less than 25 percent of all placements in the 1980's to more than 50 percent by 1997. Finally, Figure 2 tracks dealer inventories (the cumulative difference between shipments and placements) over the period, showing significant accumulation since 1991. The inventory data is further discussed in a later section.

Figure 2: Manufactured Home Shipments, Placements and Year-End Dealer Inventory, 1980-1997

Source: Bureau of the Census, Current Construction Reports, Series C20, "Housing Starts", various years.

Trends in Market Shares. The market share of manufactured housing relative to conventional housing can be measured in two distinct ways, depending on the source and the purpose of the comparison. One way is to compute manufactured home placements as a fraction of total new housing sales (with sales also including placements), while the other is to compute manufactured home shipments as a fraction of total housing starts (with starts also including shipments). The form of measurement makes a significant difference because even though manufactured housing placements and shipments tend to come into balance over time, new housing sales (as previously noted) are consistently less than housing starts. Thus, for example, manufactured housing had a 30 percent market share in 1996 based on placements and new housing sales, compared to a 24 percent share based on shipments and housing starts. Figure 3 gives data since 1980 expressed in both ways.

Figure 3: Manufactured Housing Market Share Based on Total New Housing Sales and on Total Housing Starts, 1980-1996

Source: Bureau of the Census, Current Construction Reports, Series C20 and C25, various years.

Figure 3 also shows that while manufactured housing has grown considerably in market share since the start of the 1990's, it remains below a peak that was reached in 1982. Of course, the 1982 market was totally unlike the situation in the mid-1990s. The 1982 peak did not reflect high production of manufactured homes (1982 shipments were 240,000) so much as abnormally depressed sales of site-built homes in a very weak economy with high interest rates. Furthermore, placements in 1982 were overwhelmingly single-wide units, so the degree of competitive overlap between manufactured and site-built housing was considerably less than in today's environment where the majority of units are double-wide.

Trends in Manufactured Housing Inventories. The balance over time between manufactured housing shipments to retailers and placements from retailer lots determines the number of homes in dealer inventories. In this market and elsewhere in the economy, inventories are frequently studied as indicators of future economic activity in an industry. The shipments and placements curves in Figure 2 show that every year since 1992 there has been an excess of shipments over placements, with the difference growing every year. By 1997, placements were just 80 percent of shipments. The result has been a large jump in estimated dealer year-end inventories, which rose to more than 50 percent of annual placements by 1997.

Some growth in inventories is to be expected as the HUD-Code sector grows and more retail outlets come on line, which has been the case in most recent years. But some of it may simply represent the accumulation of unsold homes on dealer lots, which would ultimately dampen orders for new homes from the factory. Indeed, in 1997 there were reports of some consolidation in retailers at the same time as the industry experienced its first production decline in six years. It is not at all surprising that shipments would stabilize or even decline once dealer inventories reach unusually high levels. And it is too soon to tell whether this decline is temporary, until the manufactured home industry works off its unsold inventory, or represents the end of the rapid growth period of the 1990s. Finally, the estimates of placements and inventory should be viewed cautiously, since the Census sampling methodology does not survey dealer inventory directly. Rather, a sample of HUD-Code homes shipped to dealers is tracked over time to simultaneously estimate placements and inventory. If a given home in the sample cannot be confirmed to have been placed for use, it is assumed to remain in inventory. This means that any problems in tracking the sample over time may tend to raise the estimated level of inventory and depress reported placements.

2.3 TRENDS IN SALES PRICES OF NEW HOMES

One of the most obvious and potentially important factors distinguishing conventional and manufactured housing is selling price. Average selling prices for new conventional homes and for single-wide and double-wide HUD-Code homes as tabulated by the U.S. Bureau of the Census are shown below in Figure 4.

Figure 4: Average Selling Prices of New Homes by Type of Home in Nominal Dollars and 1996 Constant Dollars, 1980-1996

Source: Bureau of the Census, Current Construction Reports C25, Characteristics of New Housing, various years.

The raw Census price data (shown as solid lines in the figure) obviously reflects substantial inflation for all types of housing over the 17-year period, while the 1996 constant-dollar prices (the dashed lines, adjusted for inflation by using the overall Consumer Price Index) show much less overall change. In addition, the conventional new home prices include land and site improvements, while the manufactured home prices do not include land or site improvements. No attempt has been made to adjust for differences or changes over time in new home size, quality or amenities. Nevertheless, the figure makes it clear that site-built homes have been and still remain significantly more expensive to purchase than manufactured homes. Much of this report focuses on identifying and analyzing the reasons for this difference, which clearly has been an important factor underlying growth in the HUD-Code sector.

Figure 5 gives more details about the price trends for HUD-Code homes. It shows the average prices for single-section, double-section and all HUD-Code homes, in both nominal dollars (solid lines) and adjusted for inflation to constant 1996 dollars.

Figure 5: Average Selling Prices of New HUD-Code Homes in Nominal Dollars and 1996 Constant Dollars, 1980-1996

Source: Bureau of the Census, Characteristics of New Housing, Current Construction Reports C25, various years.

Several important trends are apparent from Figure 5. The real prices of single-section and double-section manufactured homes both declined slightly on balance from 1980 to 1996, but at the same time the real price of the average HUD-Code home rose by a small amount. This increase reflects a shift in the output mix towards double-section homes. Furthermore, real prices of both single-section and double-section homes have been increasing significantly since 1992. This is believed to reflect enhancements to new manufactured homes, both single-section and double-section, that have added cost but also expanded market appeal. Analysis of median prices as opposed to average prices points to similar trends and conclusions.

2.4 GENERAL ANALYSIS

The data on industry structure, production trends and sales prices presents a very basic picture of the market for new housing and highlights issues discussed throughout this report. Manufactured housing and conventional homes are produced in fundamentally different ways by very different types of organizations, and sell for considerably different prices. Yet the finished products can be remarkably similar, and there is some degree of overlap in the pool of interested buyers. The situation is also highly dynamic because manufactured homes themselves have evolved in design and production technology to add buyer appeal at the same time as output has grown.

A major purpose of this report is to identify efficiencies in manufactured housing that may be applicable to more conventionally produced homes. Since opportunities to realize cost savings are obviously important to any producer of housing, one of the major questions considered throughout this report is the degree to which manufactured homes actually cost less to produce than comparable site-built homes and the reasons for such a difference. There are clearly differences in production cost, though the actual picture is far more complicated than indicated by the highly simplified price statistics discussed so far. There are also differences in the finished product and other reasons for real or apparent cost differences considered throughout the report.

One very basic reason is quite straightforward. It is logical to expect manufactured housing to enjoy some inherent cost savings over conventional homes due to fundamental differences in the production process. Production of new homes in a factory differs in many important ways from construction of homes on site, and general opportunities for efficiency exist in this centralized, controlled environment compared to construction on scattered sites. For example, factories built in low-cost areas where prevailing wage rates are lower can achieve a competitive advantage by selling finished products into markets where the wage rates for similar work performed on site would be much higher. These savings are attractive to the extent they outweigh the costs of transportation and installation that are unique to factory-built housing. In addition, capital investment for plant and specialized equipment used in assembly-line operations shifts the mix of labor and capital inputs and raises labor productivity. Unskilled, less expensive labor can be used more effectively when production takes place on an assembly line and can be organized into simple, repetitive operations. Workers in the plant are generally employees of the firm, not subcontractors, and as such can be scheduled, managed, trained and deployed by a single authority in the interests of productivity and efficiency. These factors, which characterize both HUD-Code and modular home production, differ greatly from the institutional contracting and subcontracting arrangements that characterize conventional site-built homes and home building firms. Industry sources report that the labor content of HUD-Code homes typically ranges from 8 to 12 percent of total cost, compared to total labor costs for site-built homes which have been estimated to constitute 40 percent or more of total cost. Of course, to some degree a smaller labor share will be offset by higher costs of capital for any firm with investment in fixed production facilities, but the successes of industrialization throughout the economy are powerful evidence of the opportunities to reduce production cost by substituting capital for labor.

The factory setting offers other advantages that can help control cost. It minimizes delays due to poor weather, exposure to theft and vandalism, and damage to building products and materials stored for use in construction. It also affords opportunities to realize economies of scale or lower unit cost of production through expansion of highly centralized facilities, a result that has proven extremely difficult to achieve in the decentralized site-built environment. The dramatic differences in competitive structure between the site-built and manufactured sectors of the industry emphasize this point. As a related point, the large size and purchasing power of firms that produce most manufactured homes undoubtedly gives them the pure economic leverage needed to negotiate the lowest possible prices for all commodity-type building products. HUD-Code producers indicate that they can save up to 30 percent of cost on standard building materials, although a ten percent savings is reportedly more common. Producers achieve savings through large-scale purchases direct from manufacturers instead of distributors or wholesalers, and by taking delivery at centralized production facilities rather than at multiple building sites. Such a difference in the cost of inputs alone goes far to explaining differences in selling prices, even where products or materials are identical in conventional and manufactured homes. It is significant that modular producers have failed to grow to the point where they can realize this benefit of size.

A more detailed look at other factors contributing to differences in production cost will be found in subsequent chapters. Conventional and manufactured homes are similar, but they are hardly identical. Chapter 3 looks at differences in design features, amenities and various building products found in conventional and manufactured homes that undoubtedly contribute to differences in production cost and selling price. Chapters 4 and 5 compare the regulatory systems and code requirements applicable in each sector of the industry. Substantive differences in codes, regulatory procedures and compliance costs also contribute to differences in production cost. Itemized cost comparisons in Chapter 6 control for various factors and give the most complete picture of where costs differ, where they are similar, and how they relate to the overall cost of purchasing a completed home or financing purchase through a mortgage.

Characteristics of Conventional and Manufactured Homes

3.1 INTRODUCTION

This chapter presents statistical data documenting several types of similarities and differences between conventional and manufactured homes, including location and land tenure, occupant characteristics, and design and material characteristics. Some of the comparisons indicate underlying differences in cost of production that contribute to differences in selling prices between the two categories of housing as documented in Chapter 2. Other information provides insights concerning the degree of market overlap, i.e. the similarities and differences between actual or potential buyers of either type of home. Where possible the data is based specifically on newly constructed units but in some cases the statistics reflect all units in the housing stock.

3.2 COMPOSITION AND LOCATION OF THE U.S. HOUSING STOCK

The American Housing Survey (AHS) is a comprehensive national compilation of information about all types of housing units, and is performed jointly at regular intervals by the U.S. Department of Commerce, Bureau of the Census, on behalf of the U.S. Department of Housing and Urban Development. The most recent AHS estimated that as of 1995 the U.S. housing stock consisted of more than 109 million housing units. This total included 66 million detached houses and 7.6 million other units classified as "mobile home or trailer." After eliminating seasonal and vacant properties the year-round occupied housing stock was estimated at 98 million units, including almost 61 million detached houses (86 percent owner-occupied) and 6.1 million manufactured houses (78 percent owner-occupied). A comparison to earlier data shows that manufactured housing rose from 3.3 percent of the occupied housing stock in 1970 to 6.3 percent by 1995.

3.2.1 Age of Housing

As might be expected, manufactured homes are considerably newer than other units in the housing stock. In 1995 the median manufactured home was 15 years of age, compared to 30 years for all other housing units. Table 2 shows the distribution of year of production for manufactured homes and all other housing units as of 1995. The Table clearly shows that large numbers of mobile homes were first introduced in the 1960s; only 2 percent were built earlier than 1960. About 35 percent of the existing manufactured housing units were built prior to 1975, the year the HUD-Code went into effect. More than 19 percent of the stock of manufactured homes were built in the 1970-74 period, the highest number of manufactured homes produced in a five-year period since large-scale production began in the early 1960s. Much of the perception of manufactured homes is based on the majority of units that predated the early 1980's.

Table 2: Year of Construction of Year-Round Occupied Housing Units by Type, 1995

Year of Construction	Manufactured Homes		All Other Housing Units
	Number	Percent of Total	Number	Percent ofTotal
1995 (part year)	136,000	2.2%	674,000	0.7%
1990-94	1,183,000	19.2%	5,795,000	6.3%
1985-89	852,000	13.8%	7,266,000	7.9%
1980-84	811,000	13.2%	6,484,000	7.1%
1975-79	1,054,000	17.1%	10,054,000	11.0%
1970-74	1,184,000	19.1%	8,741,000	9.5%
1960-69	809,000	13.1%	13,458,000	14.7%
pre-1960	134,000	2.1%	39,059,000	42.7%
Total Units	6,164,000	100%	91,531,000	100%
Median Year	1980		1965

Source: American Housing Survey, 1995.

3.2.2 Regional Distribution of Housing and Market Shares

The regional distribution of the overall housing stock and the newly constructed stock, as well as the corresponding distributions for manufactured homes, appear in Table 3.

Table 3: Regional Distribution of Total Housing Stock and Newly Constructed Housing Units by Housing Type, 1995

Region	Total Housing Stock, 1995		New Construction, 1995
	All Types	Manufactured	All Types	Manufactured
Northeast	19.6%	8.5%	8.7%	4.7%
Midwest	23.8%	18.0%	21.4%	18.0%
South	35.8%	53.6%	45.4%	63.8%
West	20.8%	20.0%	24.5%	13.5%
Total	100%	100%	100%	100%
Total Housing Units	109,457,000	7,647,000	1,354,100	310,700

Sources: Total housing stock data from American Housing Survey, 1995. New construction data for 1995 from Bureau of the Census, Current Construction Reports, Series C20, "Housing Starts."

The Table shows that while the South has just over one-third of all existing housing units, it contains over half of the nation's manufactured housing. New construction data shows an even stronger trend: in 1995 over 45 percent of all housing starts and almost two-thirds of all HUD-Code placements were in the South. This suggests that some of the growth in manufactured housing production during the 1990s can be explained simply by the general regional shift in housing production towards its historically strongest region.

There is, however, more at work than just an overall regional shift in new home location. This can be seen by examining trends in the composition of housing sales by region, as shown in Figure 6. In 1996, manufactured home placements had a market share of nearly 38 percent of all single-family homes sold or placed in the South, compared with a 30 percent share in the Midwest and 17 percent shares in both the West and Northeast. This represents an increase of 8 percentage points in market share for manufactured homes in the South from its low of 30 percent of the single-family home market in 1989. Shares for the other regions have fluctuated within ranges of about five percentage points over the last 10 years, with a slight increase suggested in the West that has yet to make up for a substantial decline in that region since 1980.

Figure 6: Placements of Manufactured Homes by Region as a Percent of Regional Single-Family Home Sales, 1980-1996

Source: U.S. Department of Commerce, Bureau of the Census, Current Construction Reports, C20 and C25, various years.

The resulting picture is a complex one. While Figure 3 shows manufactured homes with an overall market share that has been rising since the late 1980s, Table 3 documents a general shift in housing production towards the South that would in itself be expected to increase overall market share of manufactured homes based on the historical mixture of conventional and manufactured homes by region. Finally, Figure 6 gives a regionally disaggregated picture of the market share of manufactured homes, showing a growing market share in the South and possibly the West but no clear trend in the Midwest or the Northeast. In conclusion, the statistics indicate that recent growth in the market share of manufactured housing appears to be more a regional than a nationwide phenomenon, concentrated in the Southern and possibly the Western states.

3.2.3 Community Characteristics and Land Tenure

Manufactured homes are disproportionately sited in non-metropolitan areas, and in the less-densely populated suburban fringes of metropolitan areas (see Figure 7). More than eighty percent of manufactured homes are placed outside urbanized areas, with 89 percent of those in rural settings. Of the 20 percent inside urbanized areas, three-fourths are in suburban areas and only 25 percent are in central cities. Conversely, over 55 percent of all owner-occupied homes are located in urbanized areas and more than 40 percent of that group is located in central cities. Thus, only about 5 percent of all manufactured homes are in central cities, compared to about 22 percent of all owner-occupied homes. The overall pattern most likely reflects a combination of market forces and historical zoning restrictions on the placement of manufactured homes.

Figure 7: Location of Manufactured Housing, Owner-Occupied Housing and All Occupied Housing Inside and Outside Urbanized Areas, 1995

Source: American Housing Survey, 1995

Land Tenure. While conventional single-family homes are built almost exclusively on private land, manufactured homes can be placed either in rental communities or on private land. Rental communities are referred to by various terms including "mobile home parks," "manufactured home communities," and "land-lease communities." Placing a manufactured home on a rented house site allows the buyer to avoid the cost of land and related infrastructure when purchasing a home. This minimizes the required down payment and closing costs, but adds monthly land rent or "pad rent" fees to housing costs. In many rental communities the leases are relatively short-term, but longer leases also can be found.

About 50,000 to 55,000 manufactured home communities currently exist in the U.S, ranging in size from three to 1,000 homes. About 80 to 85 percent of these communities have 100 or fewer sites. The largest 15 to 20 percent of parks, sometimes referred to as "institutional investment- grade" parks, each have more than 100 sites. One source estimates that 500 major owner/operators, each with a minimum portfolio of five manufactured home communities (500 home sites), control about 15 percent of the inventory of community parks. The annual turnover of manufactured home residents in these communities is estimated to be just five percent, compared with 10 and 60 percent for conventional owner/rental and apartment rentals respectively. The 1995 AHS reported that almost two-thirds of manufactured homes were in groups of one to six units, while nearly 32 percent were in groups of more than 20 units. Most of the units in every size group were owner-occupied. The highest proportion of rental units was 36 percent, found in the 7- to 20-unit group size.

Development of new parks is seen as a significant problem by the manufactured home industry. This may reflect negative perceptions of existing park communities, some of which are poorly maintained and permitted only as pre-existing non-conforming uses. From 1989 to 1996 the number of manufactured homes placed outside parks each year increased more than 80 percent, four times as much as the number of homes placed inside parks. Single-section placements in parks grew more than double-section placements. Trends in the proportions of single-section and double-section placements in parks for 1980 through 1996 are shown in Figure 8. The Figure shows that single-section park placements dropped from over 50 percent to just over 40 percent during this period, and double-section park placements dropped from over 40 percent to slightly over 25 percent. Overall park placements dropped from over 50 percent to about 33 percent. Even though the absolute number of homes placed in parks actually grew by 20 percent from 1989 to 1996, reflecting overall growth in sales and development of new communities, Figure 8 shows that a clear shift away from park placement has been taking place since 1980, and has been pronounced in recent years. The trend towards locating manufactured homes on private lots rather than in rented park spaces is quite important because it bridges some of the differences historically associated with location of manufactured homes and conventional homes.

Figure 8: Park Placement for New Manufactured Homes, 1980-1996

Source: U.S. Department of Commerce, Bureau of the Census, Characteristics of New Housing, various years.

Lot Size. According to the 1995 AHS, the average lot size for manufactured homes is 0.88 acres, compared with a 0.43 acre average for both all detached houses and detached owner-occupied houses, and a 0.48 acre average for new housing. Thus, manufactured homes appear to be concentrated at both ends of the density spectrum. Compared with other housing, not only is the stock of manufactured homes concentrated in high densities of eight units or more per acre (14 percent vs. 6 percent), but it is also concentrated at low densities exceeding one acre per unit (32 percent vs. 15 percent). This may reflect location of many older mobile homes in rental parks at very high densities. In more recent years, changing standards in land development and zoning have decreased density for all housing. The larger lot size for new manufactured homes compared with other homes results from the tendency to place manufactured homes in less populated areas where lots are larger, land is less expensive, and zoning is less restrictive.

Summary. Locations of both new and existing manufactured homes differ considerably from other types of housing. The overall stock of manufactured homes is disproportionately located in rural areas, and the majority of new manufactured homes are being placed on parcels of land owned by individuals outside of urbanized areas, particularly in rural areas. Notwithstanding the trend to place owned manufactured homes on their own parcels of land, over 40 percent of new single-section units and about 25 percent of new double-section units are still being placed on rental sites in manufactured home communities.

3.3 HOUSEHOLD CHARACTERISTICS FOR CONVENTIONAL AND MANUFACTURED HOUSING

This section compares the households that occupy conventional and manufactured homes with respect to age, education, household income and housing expenditures. It also compares reasons given by recent movers to owner-occupied manufactured and conventional homes for leaving their previous home and for choosing their present home.

3.3.1 Age Composition

According to AHS data the three major age groups, defined as the young (under 35), middle-age (35-64) and elderly (65 and above), are relatively evenly distributed in manufactured housing. The highest proportion of residents is in the middle-age group. Manufactured homes have been increasingly attractive to households in the young age group and to a lesser extent to the middle-age group. Counterbalancing trends among age groups in manufactured housing have resulted in little change in the median age of heads of households of manufactured homes from 1987 to 1995. Heads of households living in manufactured homes have a median age of 44 years, compared with 46 years for all households, 38 years for renters, and 51 years for households in owner-occupied units.

The following trends were observed in the age of heads of households residing in manufactured housing compared to all households and all owner-occupied households (see Table 4):

• Under age 35. In 1995, 29 percent of heads of households in manufactured housing were under age 35, a larger share than in all households or owner-occupied households. This represented a decrease from 1987 when this age group had a 35 percent share.
• Age 35 to 54. In 1995, 37 percent of heads of households in manufactured housing were age 35 to 54. This was a lower share than in other types of housing. However, since 1987 this group has been increasing more rapidly in manufactured housing than in owner-occupied homes or all housing.
• Age 55 and above. The 34 percent share of households over age 55 in manufactured homes is nearly the same as the share of such households in all occupied units, but significantly less than the 40 percent share of such households in owner-occupied housing units. The share of this age group in manufactured homes has declined less since 1987 than in owner-occupied households or all households.