Since I am the final speaker in this conference, Bill Wheaton asked me to focus on what to do in order to get larger quantities of affordable housing. So I will talk about policy responses to the shortage of affordable housing. I will discuss the causes of that shortage only insofar as doing so is necessary to illuminate what the policy responses ought to be.

There are three basic causes of affordability problems: (1) Many households have incomes too low to pay for "decent" quality units. In no U.S. housing market can a person earning the minimum wage afford a modest two-bedroom apartment. (2) Housing prices are too high for many reasons,and (3) The rental stock is very old and many units are being retired.

Raising household incomes is a key way to attack this problem. That might occur through another period of general prosperity like the late 1990s. But housing rents and prices rose faster than incomes, keeping them unaffordable to many low-income households. So the income approach requires directly aiding low-income households.

That can be done through aids like the earned income tax credit, or aid linked to housing through rental housing vouchers or other direct subsidies. Given the fiscal stress of state governments, such aid must come from the federal government. Expanding rental vouchers would be a very effective approach.

The biggest problem with this approach is that federal housing policy is hugely biased towards promoting homeownership, even though poor renters have the greatest housing affordability problems. The biggest budgetary aids are tax benefits that aid high-income homeowners most. Transferring some of those benefits to poor renters would be a great way to improve general affordability. But it is a political non-starter in a society where two-thirds are homeowners.

Even so, increasing federal aid to renter households is the most direct means of improving housing affordability for such households.

Working on the other side of the problem – high housing prices – can be done in several different ways because several different causes are involved. But first let us examine the fundamental difficulty in reducing housing prices generally.

In order to get a general increase in housing affordability, the overall level of housing prices would have to fall – at least in relation to prevailing incomes. That is only possible if the overall supply of housing expands faster than the number of households, and the resulting surplus of units is not offset by high levels of demolitions or other removals. That in turns mean we must build many more new housing units than we create new households or remove older units.

But in the 45 years from 1955 to 2000, total new housing starts – including mobile homes – have outnumbered new households all but four times. The overall average is 1.36 new housing units for every 1 added household. Yet housing prices in nominal terms have risen steadily over that period.

There must be a lot of housing units removed each year, or much migration from some metropolitan regions to others, to explain why housing prices have continued to rise in spite of this apparent "surplus" of new units.

In addition, the homebuilding industry is increasingly dominated by large builders, who operate with very short production lead-times. Appearance of any market surplus immediately shuts down construction – something very different from the office space or shopping center industries.

Moreover, there are enormous economic and political forces arrayed against increasing the supply of new units enough to reduce housing prices generally. These include all the financial institutions that have lent trillions of dollars in mortgages, the secondary agencies like Fannie Mae and Freddy Mac, all the investors who own their bonds, the homebuilding industry, and the 67 percent of all U.S. households who are homeowners. These powers all have vested interests in keeping home prices rising – which means making them less affordable to all potential buyers and all existing renters, including the poor.

At present, these forces have collaborated to keep the housing production boom going without reducing prices by easing the qualifications for buying new homes, thereby expanding market demand – and risk – substantially. This means the chances of getting a general increase in housing affordability through creating an overall market surplus are slim to none.

Consequently, tactics for increasing affordability have to focus not on expanding the general supply of affordable units, but expanding supplies in specific situations in limited areas. This shifts emphasis from federal policies to state and local policies.

Another aspect of this problem is that one in ten apartments affordable to low-income households is federally subsidized. Thousands of those apartments will be removed from subsidy status within the next few years. Over one million units are involved.

Hence a key policy to provide affordable housing is somehow to preserve many of these units in subsidized status, mainly by shifting ownership to non-profits. This cannot be done without sizable federal capital contributions to both equity and rehabilitation, plus the development of appropriate non-profit organizations. Given current budgetary pressures connected with security, this seems unlikely.

A major aspect of the affordability problem is: what specific goals or criteria should any public policies aimed at improving affordability problem pursue?

First, should such efforts focus on improving affordability for the very poorest people with the worst problems, or upon expanding the total supply of units affordable to households with considerably higher incomes? Expanding the housing supply greatly at any income level tends to improve it for all income levels, since there is a "trickle-down" effect through chains of moves that affect many income groups.

Moreover, focusing subsidies on groups above the poorest permits smaller amounts of subsidy per household, and therefore allows creating far more added affordable units per million dollars spent than focusing on very poor households. Also, focusing aid directly on the poorest households reduces the acceptability of adding affordable units to residents of most non-poor suburban localities.

A second criterion is: where should more affordable units be located? My own view is that we most need units in suburban communities near where the most new jobs are being created. This is true both of newly-built units and of units made affordable through financial assistance like rental vouchers.

Therefore, a third criterion is: how politically acceptable is a specific approach to the residents of suburban communities? One benefit of approaches that rely on developers taking the initiative is that they do not have to be accepted by residents of the localities where they are used.

I believe the biggest obstacle to affordability in suburban locations is the resistance of many suburban local governments to permitting multi-family or low-cost housing within their boundaries. This is rooted in the political dominance of suburbs by homeowners. This view has been well stated by Bill Fischel in his book The Homevoter Hypothesis.

Those homeowners believe any less costly housing in their neighborhoods might threaten their ability to achieve such maximization. Since their homes are their major financial assets, they pressure their governments to oppose cost-reducing changes in regulation – such as permitting apartments or other lower-cost housing nearby. Therefore, as long as we leave full regulatory power over housing planning and construction in the hands of local governments, there is no realistic chance that housing costs can be reduced by changing regulations that increase those costs.

This economic motivation favoring high housing costs is reinforced by two widespread social desires. One is to live in neighborhoods occupied by others at least as well-off economically, and surely not worse off. The other is that most whites do not want to live in neighborhoods where African-Americans comprise more than about 25 to 33 percent of the residents. Both these social goals are served by keeping local housing prices high.

If suburban homeowners could be persuaded that allowing lower-cost housing into their communities would not harm them, they might be willing to accept it. Such persuasion would require huge publicity on studies that show no adverse effects have occurred where affordable housing has been built. Or we might even experiment with homeowners' insurance that would guarantee owners living near new affordable units that they would not lose value when the resold their homes. Unfortunately, I do not think either of these tactics is likely to work well.

So merely urging local governments to change their regulations in recognition that society needs more affordable suburban housing is a waste of time, since those governments have no incentives to change policies their voters want. Yet that is what every federal housing commission has done – with no results at all.

If this view is correct, effective tactics for increasing the amount of affordable units involve different methods of limiting the powers of local governments to block them. There are at least six basic ways to do this.

One is creating mechanisms that permit housing developers to override local zoning regulations when they prevent construction of affordable units in order to maintain the exclusivity of the locality. The anti-snob zoning law in Mass. takes this approach. This leaves the initiative for creating affordable units entirely in the hands of private housing developers. It also minimizes public sector costs, but cannot create units for really poor families.

This approach requires setting some type of "target" for affordable housing for every locality, either by state law (10% in Massachusetts) or through the actions of a state-appointed agency (as in New Jersey).

A second approach is for the state or the county to require all private housing developers to include affordable units in any projects they build, or inclusionary zoning. It compels any locality that wants more housing to accept more affordable units. Moreover, this approach does not require public subsidies.

But developers cannot get the rent or sales prices of affordable units very low because those units must be subsidized by transfers from their market-rate units. Hence this tactic does not help the lowest income households.

A second limitation is that the affordable units can pass out of that category if the occupants are allowed to sell them at market prices after holding them for a fixed period, say ten years. If the units are rental, they can be owned by a public housing authority or non-profit agency.

The third approach is to pass state or county regulations that prevent localities from adopting specific rules that inhibit affordable housing. Examples are rules against manufactured housing, or against the use of plastic pipe, or against multi-family housing. Another would be a state law empowering any owner of a single-family unit of a certain size to create an accessory rental apartment, even if local laws do not permit it, as long as the apartment met certain criteria. This basic approach is inexpensive in public funds and would aid many homeowners.

The fourth approach is to create financial and other incentives for local governments to permit more affordable housing. This could involve requiring localities to remove regulatory obstacles before qualifying for state aids for roads, sewer and water systems, and other benefits. Maryland has something like this. This approach has the disadvantage of costing the public sector a lot of money.

Congress has specifically prohibited HUD from making any federal assistance to local governments contingent on the local governments reducing their barriers to affordable housing. If this law were changed, federal aids could be made contingent on such local behavior.

The fifth approach is for the state legislature to adopt goals or procedures that all local governments are required to follow. These include having to prepare comprehensive plans with housing elements that estimate needs for affordable units, and development of specific programs for meeting those needs. Only New Jersey does this now, although California is moving closer. Oregon has statewide planning goals, but they don't require more affordable housing units.

For this approach to be meaningful, it requires that the state or county set up some agency to review the housing plans of individual localities, with the power to impose sanctions on localities that do not follow state rules. Hence this approach must usually be tied to state provision of some type of financial incentives, such as infrastructure financing aid.

A sixth approach would be for the state or county to create a region-wide agency that can become directly involved in housing location decisions. Only New Jersey has done this, insofar as I know. However, federal laws permitting city public housing authorities to use their housing vouchers to enable city residents to move into individual apartments in suburban areas perhaps exemplify this approach too. That is behind HUD's "moving to opportunities" program.

It would be possible to create a regional non-profit agency aimed at putting affordable units around the region by controlling low-income tax credits.

Up to now, none of these approaches have been widely enough adopted to put any dent in the national shortage of affordable housing for low-income households. Political opposition at the local level has been too strong. What would have to happen to get one or more of these approaches more widely adopted?

One possibility is recognition within a specific region that there is a housing crisis injuring politically important groups there. Poor households are usually not such a group – otherwise we would already have recognized such a crisis. But two other groups are: (1) employers who cannot find either low wage or medium wage workers because of housing shortage – as in Silicon Valley, and (2) middle-income households who are being squeezed by very high housing prices. These include public sector workers like teachers, police, and fire-fighters, plus many health sector workers, plus high-wage workers being transferred to high-cost areas from other regions.

This situation is most likely to occur in areas with very high housing prices, mainly California, Seattle, Boston, New York, and Washington. However, the current slow economy takes the pressure off of employers there.

Another possibility would be heavy publicity about overcrowding from local media, especially in areas with big immigration from abroad. Thousands of such households are living three or four to a dwelling unit in immigration entry ports, such as East Los Angeles. If there were a fire that killed a lot of people, or high disease rates, or some other notable event, in such housing, the local media might make a crusade out of doing something about it. However, without such publicity, nothing much is likely to happen. The well-housed majority is not conscious of this problem.

We Americans do not like to admit that overcrowded slum housing has always been, and still is, an essential part of our housing strategy. We pretend it has been eliminated by all the high-quality new units we build, but that is a delusion.

A third possibility is that the proponents of Smart Growth will wake up and realize that they cannot even slow down the outward spread of sprawl unless they accept higher density dwellings in existing built-up communities. Up to now, neighborhood resistance to higher densities has prevented this in most regions. So outward sprawl continues in high-growth regions, causing heavy traffic flows through inboard communities too. If substantial in-fill development at higher densities were to occur, even at high-rent levels, that might expand the local supply enough to aid the poor too.

In Brazil and Mexico, U.S. developers are building low-cost housing for $25,000 per unit, for 500 square foot units either detached or in high-density high-rise buildings, with only a small interest subsidy from the government. But we cannot get suburban communities to accept such structures, or anything close to them, even though they would be far superior to the present huge overcrowding in many slum areas, including in the suburbs.

Frankly, I think that all of these possibilities are rather unlikely today. With low interest rates and biased federal policies, our housing policy focus is entirely on expanding homeownership, at the expense of ignoring the worst housing problems.

Our success at providing the vast majority of Americans with good housing has greatly undermined our willingness to pay attention to the significant minority that has heavy housing cost burdens. Moreover, aiding that minority would require many people in the well-housed majority to make some sacrifices of what they believe are current benefits – especially the benefit of excluding poor people from their own communities, and thereby protecting their home values.

In a true democracy like ours, one of the most difficult policy challenges is getting a large numerical majority that benefits from certain current arrangements to change those arrangements in order to benefit a numerical minority that is injured by those same arrangements. Majority rule makes this extremely difficult – but a majority-ruled system is greatly preferable to a minority-ruled system!

Therefore, I believe those of us who want to improve our housing affordability situation will just have to keep pressing for adoption of one or more of the approaches I have described as best we can, without being discouraged by limited success.

After all, it took over 100 years to establish universal public education in America. But persistence in that good cause finally won out – although we still have great iniquities in our school systems.

If we exhibit similar courage and persistence in spite of current majority resistance, perhaps we will someday live up to Congress's 1949 resolution to "provide every American family with a decent dwelling in a sound neighborhood." I hope so.