COVID-19 RESOURCES AND INFORMATION: See the Atlanta Fed's list of publications, information, and resources; listen to our Pandemic Response webinar series.

About


Real Estate Research provides analysis of topical research and current issues in the fields of housing and real estate economics. Authors for the blog include the Atlanta Fed's Jessica Dill, Kristopher Gerardi, Carl Hudson, and analysts, as well as the Boston Fed's Christopher Foote and Paul Willen.

Comment Standards:
Comments are moderated and will not appear until the moderator has approved them.

Please submit appropriate comments. Inappropriate comments include content that is abusive, harassing, or threatening; obscene, vulgar, or profane; an attack of a personal nature; or overtly political.

In addition, no off-topic remarks or spam is permitted.

November 3, 2015

Keeping an Eye on the Housing Market

In a recent speech, Federal Reserve Bank of San Francisco President John Williams suggested that signs of imbalances were starting to emerge in the form of high asset prices, particularly in real estate. He pointed out that the house price-to-rent ratio had returned to its 2003 level and that, while it may not be at a tipping point yet, it would be important to keep an eye on the situation and act before the imbalance grows too large. President Williams is not the only one monitoring this situation. Many across the industry are keeping a watchful eye on the rapid price appreciation (see here, here, and here), including my colleagues and me at the Atlanta Fed.

While it is too soon to definitively know if a bubble is forming, the house price-to-rent ratio seems like a relevant measure to track. Why? Basically, because households have the option to rent or own their home, equilibrium in the housing market is characterized by a strong link between prices and rents. When prices deviate substantially from rents (or vice versa), the cost-benefit calculus in the rent-versus-own equation changes, inducing some households to make a transition. In effect, these transitions stabilize the ratio.

In an effort to better understand house price trends, we chart the house price-to-rent ratio at an annual frequency on top of a stacked bar chart depicting year-over-year house price growth (see chart below). Each stacked bar reflects the share of ZIP codes in each range of house price change. Shades of green indicate house price appreciation from the year-earlier level, and shades of red indicate house price decline. The benefit of considering house price trends through the lens of this stacked bar chart is, of course, that it provides a better sense for the distribution of house price change that is often masked by the headline statistic.

Looking at these two measures in concert paints an interesting picture, one that doesn't appear to be a repeat of the early 2000s. While the house price-to-rent ratio indicates that house prices on a national basis have been increasing relative to rents, the distribution of house price change looks a bit different. In 2003, roughly 20 percent of ZIP codes across the nation were experiencing house price appreciation of 15 percent or more on a year-over-year basis. In 2014 and 2015, less than 5 percent of ZIP codes experienced this degree of appreciation.

chart-1

To better understand the regional variation, we repeated this exercise at a metro level using the Case-Shiller 20 MSAs (see charts below). (House price-to-rent ratios for Las Vegas and Charlotte were not calculated because the Bureau of Labor Statistics does not provide an owners' equivalent rent for primary residence series for these markets.) This more detailed approach reveals that elevated price-rent ratio readings were only present in a few, perhaps supply-constrained, metropolitan areas (see top right corner of each chart for the Saiz supply elasticity measure). Moreover, current home price appreciation across ZIP codes does not have the breadth that was present during the early 2000s.

Notes: (1) All price-to-rent ratios are indexed to 1998, except Dallas and Phoenix, which are indexed to 2002. (2) SE = Saiz's Supply Elasticities. Pertains to city boundaries, not metropolitan areas. For more information, see Albert Saiz, "The Geographic Determinants of Housing Supply," The Quarterly Journal of Economics (August 2010) 125

As John Krainer, an economist at the San Francisco Fed, pointed out in a 2004 Economic Letter, "it is tempting to identify a bubble as a long-lasting deviation in the price-rent ratio from its average value. But knowing how large and long-lasting a deviation must be to resemble a bubble is not obvious." We will continue digging and report back when we think we know something more.

Photo of Jessica DillJessica Dill, economic policy analysis specialist in the Atlanta Fed's research department

September 17, 2015

Do Millennials Prefer to Live Closer to the City Center?

In past posts (Part 1 and Part 2), we examined whether millennials were driving the decline in first-time homebuyers. We concluded that, if anything, first-time homebuyers were becoming younger over time and location and economic conditions appeared to be a much stronger predictor of declines than a generational divide. In this post, we look into whether millennials prefer to live close to the city core or in the suburbs. Where millennials settle could determine whether our cities continue to grow, what our transportation infrastructure expenditures should be, and whether homebuilders should focus their efforts on multifamily housing in urban locations or traditional single-family homes in the suburbs.

This question has received a fair amount of attention—see here, here, here, and here. A number of observers have speculated whether the recent surge in millennials living in cities represents a change in preferences or whether it's simply an artifact of financial constraints—tighter underwriting standards, weak income growth, or larger student debt. Nielsen's survey of young adults finds that millennials prefer the lifestyle afforded by dense urban environments, but the National Association of Homebuilder's survey of young homebuyers finds that just 10 percent would prefer to live in the city while a whopping 66 percent want to live in the suburbs. Setting preferences aside, others debate whether millennials really are moving to the city. While recent data confirm that young people are moving to the cities at much higher rates than in the 1990s, it's also true that the raw majority of young people choose the suburbs over the city.

This research on young adults tends to combine renters and homeowners in one category. Renters tend to experience credit and financial barriers to location, and are limited in their location choice by the distribution of rental housing stock. That can make it difficult to distinguish whether young people who move to the city do so because they prefer urban life or because there is more rental housing stock in the city than in the suburbs. To shed light on the question of where millennials prefer to live, we segment out a group of young adults who experience relatively fewer restrictions on where to live: first-time homebuyers. Our data set allows us to identify first-time millennial homebuyers and the census tracts where they bought their first homes (a previous post describes the data).

Using this data, we ask if first-time millennial homebuyers are more likely to live near the city center than either existing homeowners or older first-time homebuyers. Finally, we look at how other factors like creditworthiness and student debt levels appear to influence this decision.

Below, we chart the median distance from the central business district (CBD) of first-time and existing homeowners by age bracket from the years 2001 to 2014. We find that existing homeowners tend to live, on average, 6.3 to 6.5 miles from the city center. First-time homebuyers tend to live closer in regardless of age, on average 5.8 to 5.9 miles away from the city center. Beginning in 2003, younger first-time homebuyers trended towards more central locations. During the 2007–09 recession, the spread between older and younger first-time homebuyers collapsed. After the recession, the spread widened again. It's difficult to say whether the shift in purchase patterns is the result of financial constraints or changing preferences, but the tendency appears to be for newer and younger homeowners to purchase homes closer to the city center.

Median-distance-of-purchase-from-cbd-core

What this chart cannot tell us is whether the trend that has younger people living closer to the city center reflects uniform preferences or whether this is an artifact of stronger economic growth in denser cities. In other words, is this trend the result of strong home buying in compact cities and weak sales in sprawling metropolises (that is, between cities), or is it the result of all buyers nationwide choosing to move closer to the city center (that is, within cities)?

To further investigate whether millennials prefer to live close to the city center, we perform several regressions to see how age relates to first-time homebuyer location decisions before and after the crisis. We control for a few key variables—namely, credit score, mortgage size, and student debt levels. The sample includes first-time homebuyers aged 18–60 who chose to purchase homes in the 50 largest metropolitan statistical areas (MSA) in the United States. We calculate distance by matching the census tract variable in the Federal Reserve Bank of New York Consumer Credit Panel with census tract data on distance from city center provided by CityObservatory.

Because some cities are more compact than others, we add MSA-level fixed effects. To control for the influence of nationwide effects such as the introduction of quantitative easing and the first-time homebuyer tax credit, we control for year-fixed effects as well in each regression. These controls should adjust for all region and time invariant factors that might affect both the age and location choice of home purchases.

Since creditworthiness typically increases with age and households with higher credit scores tend to be less constrained in their location choice, we also add a risk score variable to see whether age is simply a proxy for the ability to borrow. Similarly, we include mortgage balances. Finally, we add student debt balances to see whether the higher student debt burdens of young people can explain the discrepancy between the location choices of older and younger buyers.

The results are featured in the table below. On the right side of the table—from 2001–05 (that is, before the housing market crisis)—age appeared to have had a small impact on location and was not significant. Other factors such as size of mortgage and amount of student debt seemed to be larger determinants of location. Homebuyers with larger mortgages and with more student debt were more likely to live farther from the city center.

On the left side of the table—from 2006 to 2014 (that is, during and after the housing crisis)—age appeared to have had a small but significant relationship with location. Buyers who were one standard deviation younger located 0.03 standard deviations closer to the city center. With more controls included in the regression, this relationship declined to 0.02 standard deviations closer to the city center.

During and after the crisis, risk score became a stronger determinant of location as well. As risk scores increased by one standard deviation, buyers moved closer to the city center by 0.04 to 0.03 standard deviations, depending on the specification. This suggests that credit-constrained homebuyers are more likely to live father away from the city center and that, all else equal, younger homebuyers prefer to live closer to the city center.
Percentage_distance_from_city_core

While it appears that, on average, younger homebuyers prefer to live closer to the city center, can we say this reflects a preference for urban life? The average distance from city center—five-and-a-half miles—could very well describe areas with moderate density and single-family housing stock in moderate-sized cities. To focus on whether younger homebuyers are interested in living in the central city, we repeat these results using a logistic regression predicting the likelihood a first-time homebuyer will purchase within one mile of the city center. Controlling for all available factors, we find that younger buyers are significantly more likely to live in the heart of downtown. For each additional year, the odds that a buyer will decide to live within one mile of the city center drop by 6 percent.

By using this unique data set, we hope that we have shed some additional light on the age and location decisions of first-time homebuyers. Our interpretation of the data suggests that first-time homebuyers became more likely to buy closer to the city center during and after the housing market crisis and that young homeowners (first-time and existing) are more likely to live closer to the city center than older homeowners. Moreover, creditworthiness, total mortgage balance, and student debt loads appear to matter when the time comes to decide where to buy. In short, although age may not affect whether someone buys a house, our analysis suggests it may influence where they buy.

Photo of Elora RaymondBy Elora Raymond, graduate research assistant, Center for Real Estate Analytics in the Atlanta Fed's research department, and doctoral student, School of City and Regional Planning at the Georgia Institute of Technology, and

Photo of Jessica DillJessica Dill, economic policy analysis specialist in the Atlanta Fed's research department

May 20, 2015

Are Millennials Responsible for the Decline in First-Time Home Purchases?

First-time homebuyers play a critical role in the housing market because they allow existing homebuyers to sell their homes and trade up, triggering a cascade of home sales. While their share of all purchases has remained fairly flat over time (see our previous post on this topic), the number of first-time homebuyers has declined precipitously since the real estate crash. Many think of first-time homebuyers as younger households, and believe millennials are largely behind the decline in first-time homebuying. There are a variety of theories about why millennials have been slow to enter homeownership. One theory says that millennials would rather rent in dense urban areas where land is scarce than buy homes in the suburbs. Another theory blames steep increases in student debt for crowding out mortgage debt and reducing the homeownership opportunities of younger generations. Yet another theory argues that because the recession lowered incomes, younger people can't afford to buy. Finally, underwriting standards tightened after the recession, causing mortgage lenders to require larger down payments and higher credit scores in order to buy a home. Some worry that this more stringent lending environment has raised the bar too high for millennial homebuyers in particular.

We can't examine all these theories in a blog post, but we can examine the validity of the assumption that millennials are driving the decline in first-time homebuyers. We approached this from two angles. We first looked at whether the age distribution of first-time homebuyers has changed, and then we tried to discern patterns in first-time home buying across states. In general, we find that the age distribution of first-time homebuyers has become younger, not older, since the crisis. We also found that the dramatic fall in purchases varies much more strongly across states than by age. The preliminary figures suggest that housing market and local economic conditions may explain as much or more of the decline in first-time homebuyers than a generational divide.

Searching the data for first mortgages
Our analysis is based on the Federal Reserve Bank of New York Consumer Credit Panel/Equifax data. This data set provides longitudinal, individual data, using a 5 percent sample of all persons with a credit record and social security number in the United States.i We examined the age, location, and credit scores of people who bought homes for the first time and looked at how these characteristics changed after the crisis.ii

To identify first-time homebuyers, we flagged the first year of the oldest mortgage for each individual in the credit panel. This reveals the first instance of someone obtaining a mortgage, even if they subsequently buy another home or even transition back to renting. The trade-off is we were able to observe only those who use debt finance, and thereby excluded all cash purchases. While many homeowners do own their homes outright, we expect most first-time buyers and certainly most young buyers to have a mortgage.iii

Having isolated first-time homebuyers in this data set, we looked at their purchasing trends and demographic attributes from 2000 to 2014. In this data set, we found that roughly 1 percent to 2 percent of the population purchased a mortgage-financed home for the first time in a given year. Forty-nine percent to 53 percent had no mortgage (this category combines renters and those who own their homes outright), and 45 percent to 50 percent were paying down an existing mortgage.

Image

Buyers aren't getting older
We found that the number of first-time home buyers fell precipitously after the crash, from 3.3 million a year to around 1.5 million to 1.8 million. However, the age distribution of these first-time homebuyers does not change dramatically, though the median age of actually went down slightly since the peak. If we were to believe that the decline in first-time buyers was driven primarily by younger workers requiring more time to amass a down payment or pay off student debts, then we would expect to see first-time buyers getting older.

First-time-homebuyers-2001-14

We did not see a strong explanation for dramatic declines in first-time homebuyers when we compared younger and older adults. It doesn't appear that millennials are driving the decline. By comparison, when we reviewed the number of first-time home purchases by state, we found very dramatic differences that population alone cannot explain. Unsurprisingly, first-time homebuying fell further in places where the housing crisis hit the hardest.

The chart shows the number and percent change in first-time homebuyers from 2001 to 2011 by state. There is a wide variety in the percent change in first-time homebuyers, with declines as strong as 65 percent in some states and as low as 10 percent in others. North Dakota was the only state to have increases in first-time homebuyers, likely due to the oil industry growth there.

Changes-in-first-time-home-buyers-by-state
(enlarge)

This analysis does have some weaknesses. For one, as we mentioned, it omits cash buyers, who are an increasingly important segment of the housing market, especially in hard-hit states like Georgia and Florida. Also, other research has shown that the transition from renter to owner and back can happen many times in a person's lifetime, and this data set does not control for homeownership "spells" older than one year (see Boehn and Schlottman 2008). Notwithstanding, this analysis suggests that the decline in first-time homebuying is driven not by swiftly changing preferences nor the economic constraints of the younger generation but by regional and local economic conditions. Stay tuned for more, as we plan to look further into how the real estate crisis altered the home purchase decisions of young first-time homebuyers relative to older generations.

Photo of Carl Hudson By Elora Raymond, graduate research assistant, Center for Real Estate Analytics in the Atlanta Fed's research department, and doctoral student, School of City and Regional Planning at the Georgia Institute of Technology, and

Photo of Jessica DillJessica Dill, economic policy analysis specialist in the Atlanta Fed's research department

_______________________________________

i The data is a 2.5 percent sample of all individuals with a credit history in the United States. So, for example, this sample resulted in 636,638 records in 2014, which would correspond to an estimated 254,655,200 individuals with credit records and social security numbers in 2014.

ii We excluded anyone who had an older mortgage in a prior year. Doing so resulted in only a very small percentage of records being excluded.

iii Our approach and results are similar to those cited in Agarwal, Hu, and Huang (2014), who find that the homeownership rate between 1999 and 2012 varies between 44 percent and 47 percent for individuals aged 25—60 using a different time frame and age distribution of the same data set. Because our definition—and that of Agarwal, Hu, and Huang—is unique, it differs from the widely cited homeownership rate published by the U.S. Census Bureau. The rate published by the Census Bureau ranges between 65 and 68 percent for individuals over 25 years old and is calculated by dividing the number of owner-occupied households by the total number of occupied households. Homeownership rates have also been derived using other data. Gicheva and Thompson (2014) derive a homeownership rate using the Survey of Consumer Finance and find the mean homeownership rate to be 61 percent between 1995 and 2010. Gerardi, Rosen, and Willen (2007) used the Panel Survey of Income Dynamics (PSID), which tracks households over time and captures changes in tenure status, to identify home purchasers. They reported a range of 5.6 percent (in 1983) to 9.6 percent (in 1978) of households buying homes in the 1969—99 timeframe.

April 20, 2015

Income Growth, Credit Growth, and Lending Standards: Revisiting the Evidence

Almost a decade has passed since the peak of the housing boom, and a handful of economics papers have emerged as fundamental influences on the way that economists think about the boom—and the ensuing bust. One example is a paper by Atif Mian and Amir Sufi that appeared in the Quarterly Journal of Economics in 2009 (MS2009 hereafter). A key part of this paper is an analysis of income growth and mortgage-credit growth in individual U.S. ZIP codes. The authors find that from 2002 to 2005, ZIP codes with relatively low growth in incomes experienced high growth in mortgage credit; that is, income growth and credit growth were negatively correlated during this period.

Economists often cite this negative correlation as evidence of improper lending practices during the housing boom. The thinking is that prudent lenders would have generated a positive correlation between area-level growth in income and mortgage credit, because borrowers in ZIP codes with high income growth would be in the best position to repay their loans. A negative correlation suggests that lenders instead channeled credit to borrowers who couldn't repay.

Some of the MS2009 results are now being reexamined in a new paper by Manuel Adelino, Antoinette Schoar, and Felipe Severino (A2S hereafter). The A2S paper argues that the statistical evidence in MS2009 is not robust and that using borrower-level data, rather than data aggregated up to the ZIP-code level, is the best way to investigate lending patterns. The A2S paper has already received a lot of attention, which has centered primarily on the quality of the alternative individual-level data that A2S sometimes employ.1 To understand the relevant issues in this debate, it's helpful to go back to MS2009's original statistical work that uses data aggregated to the ZIP-code level to get a sense of what it does and doesn't show.

Chart 1: Measures of the Relationship between within-County Income Growth and Credit Growth for U.S. ZIP Codes

Chart 1 summarizes the central MS2009 result. We generated this chart from information we found in either MS2009 or its supplementary online appendix. The dark blue bars depict the coefficients from separate regressions of ZIP-code level growth in new purchase mortgages on growth in ZIP-code level incomes.2 (These regressions also include county fixed effects, which we discuss further below.) Each regression corresponds to a different sample period. The first regression projects ZIP-level changes in credit between 1991 and 1998 on ZIP-level changes in income between these two years. The second uses growth between 1998 and 2001, and so on.3 During the three earliest periods, ZIP-level income growth enters positively in the regressions, but in 2002–04 and 2004–05, the coefficients become negative. A key claim of MS2009 is that this flip signals an important and unwelcome change in the behavior of lenders. Moreover, the abstract points out that the negative coefficients are anomalous: "2002 to 2005 is the only period in the past eighteen years in which income and mortgage credit growth are negatively correlated."

Chart 2: Serious Delinquency Rates by Loan Vintage

There are, however, at least three reasons to doubt that the MS2009 coefficients tell us anything about lending standards. First of all, the coefficients for the 2005–06 and 2006–07 regressions are positive—for the latter period, strongly so. By MS2009's logic, these positive coefficients indicate that lending standards improved after 2005, but in fact loans made in 2006 and 2007 were among the worst-performing loans in modern U.S. history. Chart 2 depicts the share of active loans that are 90-plus days delinquent or in foreclosure as a share of currently active loans, using data from Black Knight Financial Services. To be sure, loans made in 2005 did not perform well during the housing crisis, but the performance of loans made in 2006 and 2007 was even worse.4 This poor performance is not consistent with the improvement in lending standards implied by MS2009's methodology.

A second reason that sign changes among the MS2009 coefficients may not be informative is that these coefficients are not really comparable. The 1991–98 regression is based on growth in income and credit across seven years, while later regressions are based on growth over shorter intervals. This difference in time horizon matters, because area-level income and credit no doubt fluctuate from year to year while they also trend over longer periods. A "high-frequency" correlation calculated from year-to-year growth rates may therefore turn out to be very different from a "low-frequency" correlation calculated by comparing growth rates across more-distant years. One thing we can't do is think of a low-frequency correlation as an "average" of high-frequency correlations. Note that MS2009 also run a regression with growth rates calculated over the entire 2002–05 period, obtaining a coefficient of -0.662. This estimate, not pictured in our graph, is much larger in absolute value than either of the coefficients generated in the subperiods 2002–04 and 2004–05, which are pictured.

A third and perhaps more fundamental problem with the MS2009 exercise is that the authors do not report correlations between income growth and credit growth but rather regression coefficients.5 And while a correlation coefficient of 0.5 indicates that income growth and credit growth move closely together, a regression coefficient of the same magnitude could be generated with much less comovement. MS2009 supply the data needed to convert their regression coefficients into correlation coefficients, and we depict those correlations as green bars in chart 1.6 Most of the correlations are near 0.1 in absolute value or smaller. To calculate how much comovement these correlations imply, recall that the R-squared of a regression of one variable on another is equal to the square of their correlation coefficient. A correlation coefficient of 0.1 therefore indicates that a regression of credit growth deviated from county-level means on similarly transformed income growth would have an R-squared in the neighborhood of 1 percent. The reported R-squareds from the MS2009 regressions are much larger, but that is because the authors ran their regressions without demeaning the data first, letting the county fixed effects do the demeaning automatically. While this is standard practice, this specification forces the reported R-squared to encompass the explanatory power of the fixed effects. The correlation coefficients that we have calculated indicate that the explanatory power of within-county income growth for within-county credit growth is extremely low.7 Consequently, changes in the sign of this correlation are not very informative.

How do these arguments relate to A2S's paper? Part of that paper provides further evidence that the negative coefficients in the MS2009 regressions do not tell us much about lending standards. For example, A2S extend a point acknowledged in MS2009: expanding the sample of ZIP codes used for the regressions weakens the evidence of a negative correlation. The baseline income-credit regressions in MS2009 use less than 10 percent of the ZIP codes in the United States (approximately 3,000 out of more than 40,000 total U.S. ZIP codes). Omitted from the main sample are ZIP codes that do not have price-index data or that lack credit-bureau data.8 MS2009 acknowledge that if one relaxes the restriction related to house-price data, the negative correlations weaken. Our chart 1 conveys this information with the correlation coefficients depicted in red, which are even closer to zero. A2S go farther to show that if the data set also includes ZIP codes that lack credit-bureau data, the negative correlation and regression coefficients become positive.

But perhaps a deeper contribution of A2S is to remind the researchers that outstanding questions about the housing boom should be attacked with individual-level data. No one doubts that credit expanded during the boom, especially to subprime borrowers. But how much of the aggregate increase in credit went to subprime borrowers, and how did factors like income, credit scores, and expected house-price appreciation affect both borrowing and lending decisions? Even under the best of circumstances, it is hard to study these questions with aggregate data, as MS2009 did. People who take out new-purchase mortgages typically move across ZIP-code boundaries. Their incomes and credit scores may be different than those of the people who lived in their new neighborhoods one, two, or seven years before. A2S therefore argue for the use of HMDA individual-level income data so that credit allocation can be studied at the individual level. This use has been criticized by Mian and Sufi, who believe that fraud undermines the quality of the individual-level income data that appear in HMDA records. We should take these criticisms seriously. But the debate over whether lending standards are best studied with aggregate or individual-level data should take place with the understanding that aggregate data on incomes and credit may not be as informative as previously believed.

1 Mian and Sufi's contribution to the data-quality debate can be found here.

2 Data on new-purchase mortgage originations come from records generated by the Home Mortgage Disclosure Act (HMDA). Average income at the ZIP-code level is tabulated in the selected years by the Internal Revenue Service.

3 Growth rates used in the regressions are annualized. The uneven lengths of the sample periods are necessitated by the sporadic availability of the IRS income data, especially early on. The 1991 data are no longer available because IRS officials have concerns about their quality.

4 Chart 2 includes data for both prime and subprime loans. The representativeness of the Black Knight/LPS data improves markedly in 2005, so LPS loans originated before that year may not be representative of the universe of mortgages made at the same time. For other evidence specific to the performance of subprime loans made in 2006 and 2007, see Figure 2 of Christopher Mayer, Karen Pence, and Shane M. Sherlund, "The Rise in Mortgage Defaults," Journal of Economic Perspectives (2009), and Figure 1 of Yuliya Demyanyk and Otto Van Hemert, "Understanding the Subprime Mortgage Crisis," Review of Financial Studies (2009). For data on the performance of GSE loans made in 2006 and 2007, see Figure 8 of W. Scott Frame, Kristopher Gerardi, and Paul S. Willen, "The Failure of Supervisory Stress Testing: Fannie Mae, Freddie Mac, and OFHEO," Atlanta Fed Working Paper (2015).

5 MS2009 often refer to their regression coefficients as "correlations" in the text as well as in the relevant tables and figures, but these statistics are indeed regression coefficients. Note that in the fourth table of the supplemental online appendix, one of the "correlations" exceeds 1, which is impossible for an actual correlation coefficient.

6 Because a regression coefficient from a univariate regression is Cov(X,Y)/Var(X), multiplying this coefficient times StdDev(X)/StdDev(Y) gives Cov(X,Y)/StdDev(X)*StdDev(Y), which is the correlation coefficient. Here, the Y variable is ZIP-code–level credit growth, demeaned from county-level averages, while X is similarly demeaned income growth. As measures of the standard deviations, we use the within-county standard deviations displayed in Table I of MS2009. Specifically, we use the within-county standard deviation of "mortgage origination for home purchase annual growth" calculated over the 1996–02 and 2002–05 periods (0.067 and 0.15, respectively) and the within-county standard deviation of "income annualized growth" over the 1991–98, 1998–2002, 2002–05, and 2005–06 periods (0.022, 0.017, 0.031, and 0.04, respectively). Unfortunately, the time periods over which the standard deviations were calculated do not line up exactly with the time periods over which the regression coefficients were calculated, so our conversion to correlation coefficients is an approximation.

7 It is true that the regression coefficients in the MS2009 coefficients often have large t-statistics, so one may argue that ZIP-level income growth has sometimes been a statistically significant determinant of ZIP-level credit growth. But the low correlation coefficients indicate that income growth has never been economically significant determinant of credit allocation within counties. It is therefore hard to know what is driving the income-credit correlation featured in MS2009, or what may be causing its sign to fluctuate.

8 Though house prices and credit bureau data are not required to calculate a correlation between income growth and mortgage-credit growth, the authors use house prices and credit bureau data in other parts of their paper.