The big news last week on the housing price front was the jump in the S&P/Case-Shiller Home Price Indices. Even though the bounce was rather slight – the widely cited index registered a year-over-year increase of just 0.6 percent – the media glommed on to it as evidence that the housing market might be stabilizing.

While great attention gets paid to the slightest gyrations in the oft-cited index, there is another, less-cited component of the Case-Shiller series that’s even more interesting. It’s the index value itself – graphed above. And it essentially tracks housing prices in light of their January 2000 baseline. As this chart shows, the 20-city composite index hit a high of 206 before falling to 145, a decline of roughly one-quarter from its peak.  The 10-city index peaked at 226 before falling to 156, a decline of approximately a third. Still, that’s roughly 50 points above 2000 levels. Nationally, housing prices today are in line with their 2003 levels.

The map above – prepared by the Martin Prosperity Institute’s Zara Matheson – plots the Home Price Index values for the 20 metro regions in their series. The graph below charts the movement in the Index values from 1987 to present. Even though housing prices have declined everywhere across the country, prices have held much better in some places than others. A 100-plus point difference separates the highest city from the lowest one on the index. Greater Washington, D.C. tops the list with an index value of 176.5. Housing prices are above 150 in four other locations – Los Angeles (172), New York (170), San Diego (158), and Boston (151). Sure, housing prices have come down from their peak in these cities, but they remain considerably above 2000 levels. Five regions have index values between 125 and 150: They too remain considerably above 2000 levels – Portland (144), Seattle (144), and San Francisco (135) on the west coast, which sort of make sense; and Miami (147) and Tampa (136) in South Florida which are, frankly, more puzzling, at least to me. Nine more regions have index values between 100 and 125, meaning they are up just a bit or more or less back to where they started a decade ago - Denver, Chicago, Minneapolis, Charlotte, Dallas, Phoenix, Atlanta, Las Vegas, and Cleveland. Only one region, greater Detroit, has seen its housing prices fall below 2000 levels. With an Index value of 70, housing prices in Detroit have fallen to levels last seen in the 1990s. Some experts believe it will take as much as two decades until Detroit area housing prices rebound back to their pre-crisis peak.

Source: From dcist.

I was interested in looking at what might be behind this pattern. With the help of Charlotta Mellander, we compared these Case-Shiller values to a range of regional economic and demographic indicators. We performed a simple correlation analysis and ran a series of scatter-graphs. At 20, the sample size is beneath the threshold required to generate really robust statistical conclusions. And I remind readers that these measures point to association among variables and do not imply causation. Of course, other factors and relationships which we don’t account for may complicate the picture. Still, some findings are interesting enough to warrant discussion.

Source:  Average income is from the American Community Survey.

The golden rule says: Follow the money. A wide range of research, as well as plain old common sense, says housing prices will be higher where people are more affluent. We looked at three basic measures of economic wealth – economic output, income, and wages. All three are associated with higher Case-Shiller values. But just one of these – the correlation for income – was statistically significant (.5).

Source: The creative class measure is based on Rise of the Creative Class, updated with data from the Bureau of Labor Statistics.

Economists have long noted the connection between education levels or human capital, economic success, and housing prices. Not surprisingly, we find that housing prices are associated with human capital, but the correlation is not statistically significant (most likely due to the small number of observations).

Housing prices often reflect local labor markets, as Ryan Avent notes. A study I conducted with Mellander found that housing prices were considerably higher in areas with higher percentages of the creative class as well as higher levels of human capital. We again find a close connection here, on two levels. On the one hand, there is a positive and significant correlation (.51) between housing prices and the creative class – that is, the share of workers in science and technology; arts, media and entertainment; and professional jobs. On the other hand, housing prices are negatively associated (-.65) with blue-collar, working-class jobs.

Source: The tech-pole index is updated based on the Milken Index. This chart uses a logged version of the index.

Housing prices have also held up far better in regions with high-tech economies. When we compare housing prices to the Milken Institute’s “tech-pole index” – a measure of the regional concentration of high-tech firms – we find considerable correlation (.64).

Source: The Bohemian Index measures the percentage of regional employment in artistic and cultural occupations with data from the American Community Survey.

Economists have found that housing prices are higher in locations with greater amenities. University of Michigan economist David Albouy has found that housing prices are highest in places which combine lots of  amenities with high productivity. My own research– research that landed me on The Colbert Report – found that openness also matters, identifying a significant correspondence between housing prices and the percentage of gays and bohemians. We examined the relationship between housing prices and three measures: the gay index (a measure of the concentration of gay people), the bohemian index (which measures the concentrations of artistic and culturally creative workers), and the immigration index (the percentage of foreign-born population). Housing prices are positively and significantly associated with all three – two of which are statistically significant: the immigration index (.63) and the bohemian index (.65).

A number of economic commentators, including Robert Shiller, argue that U.S. housing prices still have a ways to fall, especially as the government winds down the various supports that have been propping it up. I concur with this assessment. But I also believe future declines will continue to be uneven and vary substantially across cities and regions.

Housing prices are likely to begin to rebound in greater Washington, D.C., greater NY, San Francisco, L.A., Seattle, and Boston. These places have higher income levels, more skilled populations, greater concentrations of knowledge and creative work, more high-tech oriented economies, and higher relative levels of amenity and openness. I expect housing prices to stay relatively stable – neither rising or falling much – in Denver, Chicago, Minneapolis, Charlotte, Dallas, and Atlanta. Even though they have already declined substantially, housing prices could fall even further in Phoenix and Las Vegas, two Sunbelt cities of sand, and the Rustbelt cities of Cleveland and Detroit.

But there are two places where I believe prices are likely to fall considerably further. One is Miami where, despite falling prices, the Case Shiller Index stands at 147 – about the same, relatively speaking, as Seattle or Boston and ahead of San Francisco. The other is Tampa with an Index value of 136, also ahead of  San Francisco in relative terms. Sure, both places offer warm climates, plenty of sun, and great beaches. And it’s certainly true that their housing markets – especially their condo markets – turn on second home buyers who make their money elsewhere. But even then, we’d need to live in a world where demand for sun and beaches has surged considerably, say since the late 1990s. And, it’s rather  hard to see how that could be the case in light of national and global economic conditions. My simple view is that housing prices in these two cities are plainly out of whack with both their historic trends and their economic fundamentals.

Housing prices across the United States have fallen considerably since the bubble burst, but the pattern has been far from uniform. Housing prices have held up better in wealthier and more productive regions, with higher concentrations of knowledge, professional, and creative work and high-tech industry as well as higher levels of amenity (measured as working artists and cultural creatives) and openness (measured as greater percentages of immigrants). Housing prices have fallen further in locations with lower incomes and wages to begin with, with blue-collar manufacturing economies, lower levels of skill, and lower levels of amenity and openness. Expect that pattern to continue.

On Morning Edition yesterday, I discussed the merits of owning versus renting your home with host Steve Inskeep. The ideal of homeownership is deeply ingrained in the American psyche. For the past half century, owning a home of your own has been the veritable cornerstone of the American Dream. We more or less take it for granted that homeownership is a good thing. Homeownership, it is commonly thought,  goes along with higher incomes. It causes people to be more diligent, hard-working, and productive. It leads to stable families, stable communities, and higher levels of happiness and well-being.

But, a whole slew of recent research suggests that there are considerable costs as well as benefits to owning your home. A 1998 Federal Reserve Bank of Dallas study, undertaken well before the boom and bubble, provided detailed empirical evidence of America’s over-investment in housing. Yale University’s Robert Shiller, the world’s leading student of bubbles, housing, and otherwise, found that from “1890 to 1990, the rate of return on residential real estate was just about zero after inflation.” Or as Nobel prize-winning Columbia University economist Edmund Phelps puts it: “It used to be that the business of America was business. Now the business of America is homeownership.” He adds, “To recover and grow again, America needs to get over its ‘house passion.”‘ I delve into these issues in greater depth in The Great Reset.

My conversation with Inskeep got me interested in trying to better understand and explain the variation in homeownership across America and also to see to what degree these differences are tied to social and economic factors that are reshaping its economic geography. We use a a simple gauge of homeownership – the percentage of owner-occupants in a region. Our data are from the 2008 American Community Survey and cover 339 U.S. metropolitan regions.

The map (immediately above) – prepared by the Martin Prosperity Institute’s Zara Matheson – shows the striking variation in the rate of owner-occupancy across U.S. metros. On average, roughly 68 percent of Americans who live in metro areas own their own homes. But the variation ranges from more than 80 percent (81.7 percent) on the high-end to less than half (49.3 percent) on the low end. Among large metros – those with more than one million people – more than seven in 10 people own their homes in Minneapolis-St. Paul (74.2 percent), Detroit (74.1 percent), St. Louis (72.9 percent), Birmingham (72.8 percent), Pittsburgh (71.7 percent), Philadelphia (70.5 percent), Louisville (70.4 percent), and Tampa (70.3 percent). On the other hand, the lowest rates of homeownership are found in Los Angeles (51.9 percent), greater New York (53.7 percent), San Diego (56.7 percent), San Francisco (57.6 percent), Las Vegas (58.9 percent), Austin (60 percent), and San Jose (60.7 percent). College towns also have relatively low rates of homeownership - like, for example, College Station  (52.1 percent), Ithaca (52.9 percent), Ames (55.9 percent), and Corvalis (56.1 percent).

But what might underlie these patterns? What factors might be associated with regional differences in homeownership? And, more importantly, in what ways does homeownership reflect the broader forces that are reshaping America’s regions?

With help from Charlotta Mellander, I was able to take a reasonably close look. We performed a simple correlation analysis and ran a series of scatter-graphs between homeownership and key regional demographic and economic factors. We did this for all 300-plus  U.S. metros and for the 46 large metros (again, those with more than one million people). As usual, I note that our analysis points to association between variables only. It does not imply causation, and other factors may complicate the picture. Our findings are interesting and challenge the conventional wisdom on several fronts.

First of all, homeownership is much more common where houses are cheaper. We found a negative correlation between owner-occupancy and housing prices. The correlation was -.37 for all metros and a whopping -.67 for large metros.

Even more striking perhaps is the relationship between homeownership and the level of economic development. The association here is negative. That is, more people own their own homes in places where incomes, wages, and economic output are lower. Homeownership is negatively related to economic output (which we measure as gross metropolitan product per person). The correlation is -.16 for all metros and -.28 for large metros. It is also negatively related to incomes with a correlation of -.36 for large metros. And, homeownership is negatively related to wages: The correlation is -.18 for all metros and even higher for large metros (-.35).

Smart regions have higher levels of income, economic output, and overall well-being. But homeownership is substantially lower in smart regions, whether measured by the level of human capital; the share of creative, knowledge-based, and professional occupations; or the level of technology-based industry. Homeownership is negatively related to human capital (measured as the share of metro population with a bachelor’s degree or higher): The correlation being -.26 for all metros. It is also negatively related to the creative class, with a correlation of -.19 across all metros.

Homeownership rates are higher, however,  in communities with a higher percentage of traditional blue-collar working-class jobs. The correlation between owner-occupancy and working class jobs is .36 for all metros.

Source: Techpole is for 2006, based on Census Industry Data. Definition based on Tech Pole Index from 2000, published by the Milken Institute.

The relationship between homeownership and high-tech industry is even more striking. Our measure of high-tech industry is the tech-pole index originally developed by Ross DeVol of the Milken Institute. Homeownership is significantly lower in regions with larger concentrations of high-tech industry. The correlation is -.25 for all metros and considerably higher (-.61) for large regions.

Homeownership is also significantly lower in more open and diverse regions. One measure of openness is the share of immigrants or foreign-born people. Homeownership is negatively associated with the percentage of foreign-born people: The correlation is -.45 for all metros and a whopping -.69 for large regions.

Another gauge of openness is the percentage of gay and lesbian people living in a metro. Our measure of gay populations is the gay index -  initially developed by Dan Black, Gary Gates, Seth Sanders, and Lowell Taylor – which we updated based on more recent data. Homeownership is negatively associated with this gay index measure. The correlation is -.48 for all metros and -.56 for large metros.

Research by economist Andrew Oswald has found that homeownership is associated with higher levels of unemployment. In his research on European cities, he found that a 10 percent increase in homeownership correlated with a two percent increase in unemployment. Homeownership was a better predictor of unemployment, Oswald found, than rate of unionization or welfare benefits. Our findings are in line with his. Regions with higher levels of homeownership have also experienced worsening unemployment rates over the past year. The correlation between homeownership and the change in the unemployment rate from 2008 to 2009 is .23 for all metros.

That brings us to happiness and well-being. A study by University of Pennsylvania economist Grace Wong Bucchianeri found that after controlling for income and demographics homeowners are no happier than renters, and that homeowners reported considerably higher levels of stress. Our findings also align with  hers: We find that homeownership is associated with lower levels of  happiness and well-being, with a correlation of -.19 for all metros.

The varied picture of homeownership across metropolitan America reflects the deep and fundamental forces which are and have been reshaping its underlying economic geography. In a previous Atlantic article, I summarized the powerful undercurrents which have been sorting people and economic activity across U.S. regions. On the one hand, the variation in homeownership across regions reflects these forces.  Homeownership is more prevalent in smaller regions with more traditional economic structures, lower levels of human capital, and lower rates of economic growth, where houses are more affordable. But it is significantly lower in more affluent knowledge- and technology-based regions where housing is considerably more pricey. These high prices, as my own research and that of University of Michigan economist David Albouy show, reflect their higher rates of productivity and higher levels of amenity.

This also has implications for where and how we want to live. If owning a home is part of your dream, it’s far more likely in smaller regions with less highly charged economies. But if you wish to live in a superstar city or in a high-tech center, get ready to pay a lot more or to settle into long-term renting.

America’s housing patterns not only reflect these underlying forces, they may well reinforce them. Places like Los Angeles and New York, Seattle, Silicon Valley, and Austin not only benefit from diverse, knowledge-based economies, they have housing systems which contribute to their flexibility. Lower levels of homeownership mean that fewer of their residents are locked into mortgage payments and thus can much more easily adjust to economic downturns. They can more easily downshift their housing or move to new opportunity. And with more options for rental housing, these regions can more easily attract new people – including young people, those just starting out or starting over. Regions with higher levels of homeownership like Detroit or St. Louis have far less flexibility in adjusting to economic downturns. More of their residents are locked into their homes and are unable to relocate, whether to less expensive housing nearby or better opportunities elsewhere. These regions also have smaller rental stocks to accommodate newcomers. A region’s housing market entails more than providing shelter for residents; it is a key contributing factor to its flexibility and adaptability to changing economic conditions.

Owing your own home may be a cornerstone of the American Dream, but it is no longer associated with either greater levels of regional development or higher levels of regional happiness. Too much of it may make regions less flexible and resilient in dealing with economic change.

Next up: Later this week, I perform a similar analysis of renting versus owning, looking in detail at the factors associated with regional differences in the so-called housing price-to-rent ratio across the United States.

Housing prices continue to fall nationally but the economic impacts of the crisis are being felt unevenly across the country. Housing values are off roughly a third from their peak in mid-2006, according to the Case-Shiller Home Price Index. Phoenix and Las Vegas have taken the biggest hits, suffering declines of more than 50 percent in the past year. Miami, San Diego, L.A., and Tampa have also been hard hit. Detroit has seen housing prices sink to mid-90s levels. Housing prices have declined less significantly in greater D.C., Chicago, Seattle, Atlanta, New York, Portland, Boston, Denver, Dallas, and Charlotte. But the Case-Shiller data only covers 20 large metro regions.

This week, I take a look at how housing prices have fared across the full set of more than 300 American metropolitan areas. The posts are based on statistical analysis by my colleague Charlotta Mellander. Today and tomorrow, I’ll look at how housing prices have fared since their 2006 peak. Later in the week, I’ll look at the relationship between housing prices and incomes and wages.

The graph below compares housing prices in 2009 to their 2006 baseline price. It’s based on “residual analysis,” comparing the change in housing prices between 2006 and 2009.

Clearly, the two are related – the correlation is 0.903 and the R2 is 0.815. But the slope of the fitted line suggests that, on average, housing values in these regions have dropped by approximately 15 percent. Metros above the line have lost less value than their 2006 worth would predict, while those below the line have lost more.

Under-performers: These are regions where housing values have slipped even more than predicted. Among large metros, the under-performers include: Los Angeles (where values are off $79,789 more than expected based on the national trend), San Francisco (-$79,029), Las Vegas ($-72,421), Phoenix (-$69,897), and Miami (-$53,021). Cape Coral, FL saw the biggest relative decline (- $111,797), followed by Riverside, CA (-$103,683), Sacramento, CA (-$91,640), and Sarasota, FL (-$82,353). Akron, OH (-$59,635) and Lansing, MI (-$57,574) also saw significant declines. Housing values were down slightly more than would have been expected in Atlanta (-$27,413), Chicago (-$16,580), and greater D.C. (-$14,411). 2009 data for Detroit were not available.

Over-performers: The analysis turned up a number of over-performing regions. By that I mean regions with housing values performed better than expected relative to the national trend. Over-performers include: Honolulu (where housing values remain $160,414 more than expected), Boulder ($72,172), Salt Lake City ($68,935), Seattle ($61,997), New York ($58,407), Raleigh, NC ($57,552), Portland, OR ($42,173), Baltimore ($39,896), Austin ($38,181), Philadelphia ($29,011), Boston ($13,644), Houston ($8,693), and Dallas ($5,661).

Stay tuned for more tomorrow.

Writing in The Atlantic, I argued that the economic crisis was reshaping America’s economic geography, with big city centers and mega-region hubs like New York City, talent-rich regions like greater D.C., and college towns weathering the storm relatively well, while Rustbelt cities and shallow-rooted Sunbelt economies being much harder hit.

Take a look at the graph below from the newly released SP/Case-Shiller Home Price Index for April.

Phoenix and Las Vegas have taken the biggest hits: Housing prices there have declined more than 50 percent in the past year. Miami is next, then Detroit where housing prices have sunk to mid-90s levels. San Francisco is the only significant talent region to be pummeled. Part of this is to be expected given the tremendous run-up in housing prices there, but still prices remain higher than 2000 levels. San Diego, L.A., and Tampa have all seen declines in excess of 40 percent.

Housing prices have declined less significantly in greater D.C., Chicago (hub of the great Chi-Pitts mega-region and a magnet for regional talent), Seattle (a high-tech, high human capital center), Atlanta (a talent hub for the southeast), New York, Portland, Boston, Denver – (talent hub for the Rockies), Dallas (a mega-region hub), and Charlotte (which along with Atlanta hubs the great Char-lanta mega-region). Cleveland breaks the pattern, but like Detroit its absolute housing values have fallen. Prices in greater D.C., along with Denver, Dallas, and Cleveland, were actually up in April.

The Index also tracks prices in terms of their 2000 baseline. Nationally, it’s at 140, meaning housing prices remain 40 percent higher than in 2000, more or less in line with 2003 prices. Looked at it this way, the geographic pattern could not be more striking.

Rustbelt cities have seen, by far, the biggest declines relative to 2000 prices. Detroit and Cleveland are the only two cities where housing values have slipped below 2000 values – Detroit at 69 percent and Cleveland at 98 percent.

Prices have just about fallen back to 2000 levels in Sunbelt cities like Phoenix (105), Atlanta (105), Las Vegas (112), Dallas (115) and Charlotte (118). Miami (145) and Tampa (140) break the pattern; their prices remain significantly above 2000 levels. My guess is that prices will continue to fall and sharply in these two markets in the coming months.

But prices in prices in Boston (146), L.A. (149), greater D.C. (167), and New York (170) remain significantly above – 50 to 70 percent above – 2000 levels. While these prices may dip some, my hunch is these markets will not be devastated and will remain substantially above 2000 levels.

The SP/Case-Shiller Index suggests that housing prices are still falling and have another 30 or more to go before they hit bottom. One thing you can be sure of, it will continue to be felt unequally across regions.

Felix Salmon says there’s no end in sight for the housing bust, pointing to the latest edition of the Case-Shiller Home Price Index. Housing prices are off 36 percent since their 2006 peak.  Housing prices have fallen back to 2002 levels in nominal terms but, as Business Week’s Prashant Gopal notes, they’ve plunged to 2000 levels when adjusted for inflation. Calculated Risk (with great graphics as usual) predicts another 10-20 percent drop,

Regional differences remain pronounced. Phoenix and Las Vegas are down more than 50 percent from their peak values, while Dallas is off only 11 percent. Dallas, Denver, Boston, Charlotte, and New York appear to be holding up best. New York prices remain 73 percent above their 2000 levels, Detroit’s are nearly 30 percent below – in line with their 1995 levels.