A Focus on Jobs

Posted Tuesday, July 13 2010

More and more it seems like the our economic anxiety is fully captured by just two measures — the employment numbers and the unemployment rate. While corporations are generally well positioned cash-wise, that cash hasn't been used to significantly expand payrolls thus far. So while we sit and watch economic indicators like consumer spending, home prices and GDP roll in, there's a sense that what we really need to quell the underlying anxiety is employment growth well beyond what we've seen in this recovery to date.

Towards that end, there has been a ton of debate about whether the government can or should do anything to address the jobs issue directly. But any debate about government stimulus for jobs very quickly runs into the other economic anxiety of the day: deficits. It's tough to know what to do. On the one hand you have economists like Paul Krugman insisting that the the US can manage a lot more deficit spending than we're currently engaged in, and on the other you have the Europeans leading an austerity charge in the hopes that there is no repeat of the Greece debacle in places like Spain, Portugal or Ireland. Many think the US should take the austerity route as well.

Former Intel CEO Andy Grove suggests that a large part of the employment problem is that we've accepted the idea that innovation and knowledge work are primary and that we lose little or nothing by outsourcing manufacturing and production. He points out that in the innovation engine of Silicon Valley unemployment is higher than the nation as a whole and that employment is still well below its tech bubble peak. He even goes as far as inviting a trade war if neccessary to make it more costly for domestic companies to move parts of their operations abroad.

To make matters worse, it seems like a central component of every country's economic recovery plan (including the United States) is to grow exports and presumably shrink (on a relative basis) imports. We can't all export our way to employment growth without some frightening exchange rate magic.

Economic decisions are always a matter of trade-offs. Thankfully, moderates like David Brooks help illuminate some of the choices we can be reasonably certain are wise such as extending unemployment benefits; it's probably not a good place to go looking for meaningful deficit reduction and would likely undermine the recovery already in place. And I think Barry Ritholtz (via AR) gets it exactly right when he points out that you can stimulate short term and cut debt long term if you choose the right targets. Addressing the long term issues with Social Security and Medicare while providing short term resources to stimulate job growth almost seems like an area where you could get Republicans and Democrats to work together if you had any faith that the folks in Congress could cooperate with one another. And while I think Andy Grove makes a very important point about the value of research, development and production ecosystems to a robust local and national economy, I don't think we gain anything by instigating a trade war or by instituting currency policies that make exports more affordable abroad.

In the end, addressing the short term challenge of unemployment is difficult, but I expect we'll muddle through this period and then find some narrative to explain how we carefully crafted our emergence after the fact. In the long term, the robust, stable economic system we seek will only emerge if we can produce a broad middle class engaged in productive work (key word productive) within an economic system that values innovation while at the same time considering the long term effects of outsourcing components of the economic food chain — our economy's capital and infrastructure — on future innovation potential.

Did I mention that I added a new employment data section to the site? That's all I intended to point out in this post initially.

Employment and Home Prices

Posted Thursday, June 17 2010

Below is a chart of year-over-year changes in home prices (from Corelogic's national HPI data) versus year-over-year CES employment changes since 1981.

Y/Y Employment vs. Y/Y HPI

I've been looking at employment and home price data a lot lately, and I just had the desire to see what the above chart looked like. It kind of reminds me of a phase diagram. The comparison isn't completely fair because a real phase diagram should fully represent all the states of a system. What's the system in this case? The economy? I don't think we have anywhere near enough dimensions to represent that fully.

But taking what I can from the analogy, the system seems to like to hang out in the top-right quadrant most of the time (70% of the time actually). Looking at just the right half of the chart (where employment is positive), 92% of the points are in the top-right quadrant — meaning year-over-year home prices grew 92% of the time when year-over-year employment increased. By contrast, home prices rose only 59% of the time when employment was negative.

I don't expect these percentages have any predictive value, but they do kind of support my intuitive belief that employment growth drives HPI growth in a more stable manner than the other way around. Of course, the other way of looking at this is to say, well, we've had mostly growing employment and mostly rising home prices over the last 30 years. What else did you expect to see?

I think we can all agree, however, that the bottom-left quadrant is not one we want to visit often.

May Employment Reports

Posted Wednesday, June 09 2010

The underwhelming May jobs report disappointed in just about every way. Aside from the modest decline in the broader U-6 unemployment rate, there was not much to encourage the belief that hiring in the US was picking up steam. In my previous post comparing CES and CPS employment growth, I noted that the CPS employment number (derived from a household survey as opposed to the employer survey the headline CES number comes from) had been performing better than CES employment since both numbers bottomed in December of 2009. Unfortunately, last month saw a decline in CPS measured employment. Below is a chart of the two employment series which are set to 100 at their respective peaks.

CES & CPS Employment Index Since Peak Employment

CPS employment is still up more from the trough than CES employment, but we certainly can't interpret last month's number in a positive way. And while the CPS survey is smaller and the resulting employment series noisier, strong job growth is more likely to overcome measurement volatility if we were in an actual strong job growth environment. In short, the CPS employment number didn't give us anything positive to pin our hopes on either.

Comparing CES and CPS Employment

Posted Wednesday, May 26 2010

The BLS jobs number is certainly one of the most anticipated pieces of data released each month. The headline number that we all talk about is from the Current Employment Statistics survey (CES), but there's another employment estimate from the Census' Current Population Survey (CPS) that gets less attention. While the headline CES jobs number is now 0.44% above its December 2009 low, the CPS employment has increased 1.21% from the same month. The chart below illustrates the change in the two employment measures since employment peaked in December 2009 (November for the CPS measure). The index for each is set to 100 at their respective peaks.

CES & CPS Employment Index Since Peak Employment

So what does this mean? Let me back up a bit and quickly explain the difference in these two employment measures. The headline jobs number from the CES survey (also know as the establishment or payroll survey), counts the number of employees on employer payrolls. The CPS employment number comes from a survey of households for employed individuals (thus its alternative name — the household survey). The CPS is also where the unemployment rate data comes from. Calculated Risk has a good post explaining further how these measures relate to the unemployment data. At first blush one might expect that these two surveys measure the same thing in a different way, but there are subtle differences that make them fundamentally different metrics.

Imagine an individual already working part-time takes a second part-time job to increase his income. On the CES survey, one more job has been added to the payroll (both surveys count part-timers by the way). The CPS, however, does not count this in its employment number because the person was already employed and they're counting employed individuals not unique jobs. So while the CES employment number would go up when this person takes the new part-time job, the CPS employment number would remain the same.

Given that brief introduction, let's see how these two employment metrics have compared over a longer period. Below is a chart of CES and CPS employment since 2000. It's the same thing as above just going back to 2000.

CES & CPS Employment Index Since 2000

The first thing that stands out is the divergence in the two employment metrics over the course of the decade. Interestingly, most of that divergence occurred around the time of the recovery from the 2001 recession. Officially, that recession ended in November of 2001, but we all remember the "jobless recovery" that kept jobs (as measured by the CES survey) from growing until September 2003. CPS employment turned up in February of 2002. How can we explain this?

First of all, we're not quite comparing apples to apples yet. The CPS employment count includes sectors and types of jobs that CES employment does not. The CES survey does not count "proprietors, self-employed, unpaid family or volunteer workers, farm workers, and domestic workers."¹ CPS employment definitions are broader and include essentially anyone who has done "work for pay or profit during the week."² The two largest components found in the CPS that are not measured in the CES are farm workers and the self-employed. Fortunately the BLS breaks out those categories, so we can subtract them out of the CPS data to get a more comparable metric. The chart below is just like the one above except we've subtracted out the farm and self-employed workers from CPS employment.

CES & CPS (Ex Farm and Self-Employed) Employment Index Since 2000

It seems the difference in the employment can't be explained by farm workers or the self-employed. It turns out the number of individuals working in these two sectors held pretty constant over the decade, so the CPS employment index actually rose more against CES employment index with these two categories removed. Remember we're not talking about absolute numbers here; the comparison is relative. However, now that we've adjusted the CPS employment number to remove agriculture and self-employment jobs that are not measured in CES employment, an absolute comparison makes more sense. Here's that chart.

CES & CPS (Ex Farm and Self-Employed) Employment Since 2000

As the chart above shows, adjusted CPS employment is lower than CES employment. This makes sense, remember, because the CES survey will count each job while the CPS will count individuals only once regardless of how many jobs they have. Thus, one expects the CES employment figure will be higher.

So what does it mean that the adjusted CPS employment figure is rising more quickly than the CES figure? I think we can make a guess at it at least. Imagine that guy I talked about in the beginning working two part-time jobs gets an offer for full-time work at one of them which he accepts. He quits his other part-time job and his position is immediately filled by someone else. What do the surveys record? The CES survey records no new employment, because it's counting unique jobs. The CPS, however, does record an increase in employment because it's counting individuals with jobs (this assumes the new part-timer didn't have another job somewhere else). So perhaps what we're seeing now is more people entering the workforce to fill vacant existing positions as some part-time jobs get upgraded to full-time gigs. Can we test for this? Unfortunately we can't because the there's no way to get part-time jobs out of the CES survey. Thus it remains only a guess.

As for whether CPS employment growth outpacing CES employment growth is a good thing or not, I'm inclined to think it is. All things being equal, I tend to think the number of people employed is more relevant than the total number of jobs created. If in a tiny island economy of 10 people there were 10 jobs available, would it be better for 5 to be unemployed and 5 to work two jobs or for all 10 to each work one of the available jobs? The latter seems like a better arrangement to me. Perhaps it says something about employment flexibility in the economy though. The option of having multiple part-time jobs might be declining as adjusted CPS employment approaches CES employment.

But since the CPS employment number presaged further job growth in the recovery from the 2001 recession, perhaps we can look at its strong growth now as an indicator of future strength in CES job growth. This, however, is wild speculation, so don't put too much stock in it. Either way, I think it will be worth while to pay attention to both measures as the economy recovers.

By the way, CPS employment in April was up 1.33% from the December low if you use the adjusted measure. The 1.21% increase I referred to at the top included agricultural jobs and the self-employed.

Update: CR pointed out this monthly report (pdf) produced by the BLS that compares CES and CPS employment. There is a lot of research going on in this area, but it seems the issue remains an open question to some degree. In short, potential reasons for the discrepancy are sample size (the CES survey is much larger), statistical adjustments, worker classification (including off the book workers) and job changers. There's a ton of information in the report and I'm finding even more as I explore the referenced links within it.

Metro Home Affordability Ratios Updated for 1Q 2010

Posted Friday, May 14 2010

I posted 4Q 2009 price-to-income and price-to-rent ratios just last week, but it's already out of date since the NAR released 1Q 2010 median sales price data this week. I've re-run all the numbers and have this new table of affordability ratios for you below. Remember that the adjusted rent ratio comes from scaling median contract rent by household size.

1Q 2010 Home Affordability Ratios from Median Sales Price

Metro	Price-to-Income	Price-to-Rent	Price-to-Rent (Adj)
Akron, Ohio	1.9	14.01	11.75
Albany, New York	3.13	23.1	17.33
Albuquerque, New Mexico	3.69	23.34	20.55
Allentown, Pennsylvania	3.88	27.36	23.48
Amarillo, Texas	2.44	18.33	15.13
Atlanta, Georgia	1.81	12.35	11.44
Atlantic City, New Jersey	4.05	22.16	19.73
Austin, Texas	3.08	20.64	17.03
Baltimore, Maryland	3.55	23.5	20.53
Barnstable Town, Massachusetts	5.68	30.35	25.06
Baton Rouge, Louisiana	3.43	25.27	22.78
Beaumont, Texas	2.66	20.96	18.95
Binghamton, New York	2.28	17.1	13.41
Birmingham, Alabama	2.74	21.2	19.02
Bismarck, North Dakota	2.94	26.91	19.9
Bloomington, Illinois	2.61	23.44	18.74
Boise City, Idaho	2.6	16.68	15.62
Boston, Massachusetts	4.51	27.09	21.51
Boulder, Colorado	5.05	31.91	28.18
Bridgeport, Connecticut	4.12	29.63	26.22
Buffalo, New York	2.23	17.73	13.89
Cape Coral, Florida	1.7	8.7	9.03
Cedar Rapids, Iowa	2.53	23.53	16.54
Champaign, Illinois	2.73	18.03	14.47
Charleston, South Carolina	3.72	23.85	22.87
Charleston, West Virginia	2.71	23.91	19.86
Charlotte, North Carolina	3.12	22.46	20.31
Chattanooga, Tennessee	2.49	18.16	15.65
Chicago, Illinois	2.88	19.04	16.39
Cincinnati, Ohio	2.25	18.14	15.33
Cleveland, Ohio	2.16	15.45	12.58
Colorado Springs, Colorado	3.14	21.45	19.38
Columbia, Missouri	2.98	19.8	17.52
Columbus, Ohio	2.31	17.65	15.11
Corpus Christi, Texas	2.82	17.84	16.49
Cumberland, Maryland	2.36	19.27	16.4
Dallas, Texas	2.52	17.62	15.1
Davenport, Iowa	2.09	16.95	14.36
Dayton, Ohio	2.03	15.28	13.48
Decatur, Alabama	1.56	15.19	14.47
Deltona, Florida	2.51	12.87	12.34
Denver, Colorado	3.73	24.39	21.82
Des Moines, Iowa	2.26	18.01	14.52
Dover, Delaware	3.39	21.48	18.67
Durham, North Carolina	3.27	21.49	18.41
El Paso, Texas	3.52	21.01	18.21
Elmira, New York	2.18	16.17	13.02
Erie, Pennsylvania	2.16	16.72	14.46
Eugene, Oregon	4.49	24.34	21.31
Fargo, North Dakota	3.02	22.64	15.52
Farmington, New Mexico	4.05	30.71	28.24
Florence, South Carolina	2.45	19.18	19.94
Fort Wayne, Indiana	1.92	15.85	13.18
Gainesville, Florida	3.65	18.03	14.86
Glens Falls, New York	2.87	19.06	16.45
Grand Rapids, Michigan	1.82	12.83	10.66
Green Bay, Wisconsin	2.51	20.5	16.72
Greensboro, North Carolina	2.68	19.09	18.32
Greenville, South Carolina	3.12	22.97	20.14
Gulfport, Mississippi	2.87	15.78	15.34
Hagerstown, Maryland	2.88	19.48	17.74
Hartford, Connecticut	3.36	24.26	19.91
Honolulu, Hawaii	8.76	42.19	38.67
Houston, Texas	2.67	19.36	16.62
Indianapolis, Indiana	2.02	15.17	12.92
Jackson, Mississippi	2.64	18.26	17.5
Jacksonville, Florida	2.66	16.3	14.64
Kankakee, Illinois	2.34	16.53	13.82
Kansas City, Missouri	2.31	18.24	14.83
Kennewick, Washington	3.42	24.27	22.3
Kingston, New York	3.96	23.05	19.82
Knoxville, Tennessee	2.97	21.44	17.73
Lansing, Michigan	1.55	10.66	8.95
Las Vegas, Nevada	2.42	12.7	12.42
Lexington, Kentucky	2.67	20.15	17.53
Lincoln, Nebraska	2.53	20.31	15.8
Little Rock, Arkansas	2.86	20.15	18.46
Los Angeles, California	5.5	25.69	23.44
Louisville, Kentucky	2.62	20.19	17.58
Madison, Wisconsin	3.5	24.93	20.27
Memphis, Tennessee	2.47	16.23	15.55
Miami, Florida	3.93	16.96	16.52
Milwaukee, Wisconsin	3.75	26.58	21.9
Minneapolis, Minnesota	2.46	17.72	13.92
Mobile, Alabama	2.83	19.17	17.22
Montgomery, Alabama	2.68	19.53	19.07
New Haven, Connecticut	3.7	22.96	18.9
New Orleans, Louisiana	3.25	17.06	16.08
New York, New York	5.88	32.68	27.82
Norwich, Connecticut	2.99	20.5	17.18
Ocala, Florida	2.31	11.17	12.73
Oklahoma City, Oklahoma	2.95	22.55	20.56
Omaha, Nebraska	2.42	18.71	14.66
Orlando, Florida	2.61	13.1	12.62
Palm Bay, Florida	2.06	10.96	10.69
Pensacola, Florida	3.04	17.98	17.71
Peoria, Illinois	2.01	17.37	14.03
Philadelphia, Pennsylvania	3.45	23.0	18.29
Phoenix, Arizona	2.52	14.62	14.93
Pittsburgh, Pennsylvania	2.43	19.33	15.47
Pittsfield, Massachusetts	4.21	26.62	20.43
Portland, Oregon	4.04	27.06	22.47
Providence, Rhode Island	3.79	25.01	19.68
Raleigh, North Carolina	3.6	27.28	24.39
Reading, Pennsylvania	2.7	20.95	17.69
Reno, Nevada	3.15	19.11	18.74
Riverside, California	3.2	15.36	14.9
Rochester, New York	2.14	15.18	12.3
Rockford, Illinois	2.01	15.39	13.0
Sacramento, California	2.94	16.61	16.0
Saginaw, Michigan	1.47	9.97	9.27
Salem, Oregon	3.99	26.94	26.25
Salt Lake City, Utah	3.4	23.46	20.39
San Antonio, Texas	2.98	19.21	16.79
San Diego, California	6.01	27.9	26.33
San Francisco, California	6.74	35.84	30.57
San Jose, California	6.36	34.98	32.57
Seattle, Washington	4.55	28.95	24.58
Shreveport, Louisiana	3.7	22.76	21.27
Sioux Falls, South Dakota	2.54	20.29	16.31
South Bend, Indiana	1.59	10.16	8.78
Spartanburg, South Carolina	2.59	20.08	19.15
Spokane, Washington	3.51	23.86	20.6
Springfield, Illinois	2.16	18.62	15.09
Springfield, Massachusetts	3.53	23.55	19.4
Springfield, Missouri	2.54	19.39	17.08
St Louis, Missouri	2.18	17.09	14.51
Syracuse, New York	2.38	17.17	13.29
Tallahassee, Florida	3.5	19.7	18.18
Tampa, Florida	2.89	14.68	13.86
Toledo, Ohio	1.67	12.52	11.02
Topeka, Kansas	2.0	16.99	14.87
Trenton, New Jersey	3.12	21.79	18.82
Tucson, Arizona	3.58	22.06	20.66
Virginia Beach, Virginia	3.41	20.73	19.39
Washington, District of Columbia	3.41	21.26	18.2
Waterloo, Iowa	2.27	18.07	15.64
Wichita, Kansas	2.31	20.46	16.24
Worcester, Massachusetts	3.14	23.79	18.31
Yakima, Washington	3.42	25.02	23.45
US	3.30	20.15	18.21

Metro Affordability Ratios

Posted Saturday, May 08 2010

In my previous post I calculated the adjusted rent ratio for the US by scaling Census reported median contract rent by the ratio of average owner occupied household size to average renter occupied household size. Since owning households are generally larger than renting households, unmodified median contract rent tends to skew rent ratios higher and makes homes look slightly less affordable than they otherwise would. With this in mind I set out to calculate adjusted price-to-rent ratios for all cities with Census household data that NAR reports median sales price data for. The results are below. The data comes from NAR's 4Q 2009 sales report and 2008 Census ACS statistics.

I've also calculated how these measures change in time (as an index) in the affordability section.

Home Affordability Measures from Median Sales Price

Metro	Price-to-Income	Price-to-Rent	Price-to-Rent (Adj)
Akron, Ohio	2.11	15.53	13.03
Albany, New York	3.11	22.96	17.22
Albuquerque, New Mexico	3.72	23.53	20.71
Allentown, Pennsylvania	3.66	25.79	22.14
Amarillo, Texas	2.56	19.18	15.82
Atlanta, Georgia	2.06	14.0	12.97
Atlantic City, New Jersey	4.23	23.14	20.6
Austin, Texas	3.11	20.81	17.17
Baltimore, Maryland	3.69	24.4	21.32
Barnstable Town, Massachusetts	5.89	31.48	25.99
Baton Rouge, Louisiana	3.25	23.94	21.57
Beaumont, Texas	2.64	20.81	18.81
Binghamton, New York	2.61	19.61	15.38
Birmingham, Alabama	2.93	22.65	20.31
Bismarck, North Dakota	2.93	26.76	19.78
Bloomington, Illinois	2.52	22.64	18.09
Boise City, Idaho	2.78	17.79	16.66
Boston, Massachusetts	4.66	28.01	22.25
Boulder, Colorado	5.04	31.84	28.12
Bridgeport, Connecticut	4.42	31.81	28.15
Buffalo, New York	2.31	18.41	14.42
Canton, Ohio	1.97	15.24	13.16
Cape Coral, Florida	1.78	9.07	9.41
Cedar Rapids, Iowa	2.6	24.1	16.95
Champaign, Illinois	2.99	19.8	15.89
Charleston, South Carolina	3.58	22.92	21.98
Charleston, West Virginia	2.87	25.29	21.0
Charlotte, North Carolina	3.37	24.31	21.98
Chattanooga, Tennessee	2.66	19.42	16.73
Chicago, Illinois	3.12	20.67	17.79
Cincinnati, Ohio	2.31	18.6	15.72
Cleveland, Ohio	2.24	15.98	13.02
Colorado Springs, Colorado	3.22	22.03	19.91
Columbia, Missouri	3.08	20.43	18.07
Columbus, Ohio	2.44	18.59	15.91
Corpus Christi, Texas	2.98	18.84	17.42
Cumberland, Maryland	2.64	21.55	18.33
Dallas, Texas	2.52	17.65	15.12
Davenport, Iowa	2.18	17.64	14.95
Dayton, Ohio	2.21	16.6	14.65
Decatur, Alabama	1.8	17.52	16.68
Deltona, Florida	2.66	13.62	13.07
Denver, Colorado	3.7	24.22	21.66
Des Moines, Iowa	2.48	19.72	15.9
Dover, Delaware	3.49	22.08	19.19
Durham, North Carolina	3.18	20.85	17.86
El Paso, Texas	3.65	21.73	18.83
Elmira, New York	2.09	15.46	12.45
Erie, Pennsylvania	2.21	17.07	14.76
Eugene, Oregon	4.57	24.75	21.67
Fargo, North Dakota	3.0	22.49	15.42
Farmington, New Mexico	3.91	29.65	27.27
Florence, South Carolina	2.95	23.13	24.05
Fort Wayne, Indiana	1.92	15.9	13.22
Gainesville, Florida	3.91	19.32	15.92
Glens Falls, New York	3.2	21.25	18.34
Grand Rapids, Michigan	1.8	12.72	10.56
Green Bay, Wisconsin	2.51	20.5	16.72
Greensboro, North Carolina	2.83	20.11	19.31
Greenville, South Carolina	3.05	22.52	19.73
Gulfport, Mississippi	2.86	15.71	15.27
Hagerstown, Maryland	3.01	20.38	18.56
Hartford, Connecticut	3.37	24.33	19.97
Honolulu, Hawaii	8.63	41.59	38.12
Houston, Texas	2.67	19.35	16.61
Indianapolis, Indiana	2.08	15.59	13.28
Jackson, Mississippi	2.86	19.77	18.95
Jacksonville, Florida	2.59	15.89	14.28
Kankakee, Illinois	2.53	17.88	14.95
Kansas City, Missouri	2.47	19.47	15.83
Kennewick, Washington	3.26	23.12	21.24
Kingston, New York	3.93	22.87	19.66
Knoxville, Tennessee	3.07	22.11	18.28
Lansing, Michigan	1.69	11.59	9.74
Las Vegas, Nevada	2.46	12.92	12.64
Lexington, Kentucky	2.78	20.94	18.21
Lincoln, Nebraska	2.48	19.95	15.53
Little Rock, Arkansas	2.92	20.58	18.86
Los Angeles, California	5.85	27.34	24.94
Louisville, Kentucky	2.67	20.57	17.91
Madison, Wisconsin	3.36	23.95	19.47
Memphis, Tennessee	2.62	17.21	16.5
Miami, Florida	4.04	17.46	17.01
Milwaukee, Wisconsin	3.46	24.57	20.24
Minneapolis, Minnesota	2.54	18.32	14.39
Mobile, Alabama	3.06	20.67	18.57
Montgomery, Alabama	2.72	19.84	19.38
New Haven, Connecticut	3.85	23.89	19.66
New Orleans, Louisiana	3.32	17.43	16.42
New York, New York	5.81	32.29	27.49
Norwich, Connecticut	3.04	20.84	17.46
Ocala, Florida	2.32	11.21	12.77
Oklahoma City, Oklahoma	2.86	21.84	19.92
Omaha, Nebraska	2.37	18.36	14.38
Orlando, Florida	2.77	13.9	13.39
Palm Bay, Florida	2.03	10.8	10.54
Pensacola, Florida	3.08	18.21	17.93
Peoria, Illinois	2.13	18.43	14.88
Philadelphia, Pennsylvania	3.49	23.31	18.53
Phoenix, Arizona	2.57	14.93	15.25
Pittsburgh, Pennsylvania	2.39	19.05	15.24
Pittsfield, Massachusetts	3.86	24.45	18.76
Portland, Oregon	4.07	27.29	22.66
Providence, Rhode Island	3.94	26.02	20.47
Raleigh, North Carolina	3.55	26.86	24.02
Reading, Pennsylvania	2.8	21.69	18.32
Reno, Nevada	3.25	19.74	19.36
Riverside, California	3.13	15.05	14.6
Rochester, New York	2.14	15.17	12.29
Rockford, Illinois	2.18	16.65	14.06
Sacramento, California	3.09	17.46	16.83
Saginaw, Michigan	1.63	11.06	10.28
Salem, Oregon	3.65	24.62	23.99
Salt Lake City, Utah	3.46	23.92	20.8
San Antonio, Texas	3.01	19.42	16.98
San Diego, California	6.02	27.92	26.35
San Francisco, California	7.17	38.13	32.53
San Jose, California	6.64	36.54	34.02
Seattle, Washington	4.6	29.23	24.81
Shreveport, Louisiana	3.78	23.21	21.69
Sioux Falls, South Dakota	2.41	19.25	15.47
South Bend, Indiana	2.02	12.92	11.16
Spartanburg, South Carolina	2.68	20.78	19.82
Spokane, Washington	3.51	23.86	20.6
Springfield, Illinois	2.09	18.02	14.61
Springfield, Massachusetts	3.54	23.62	19.46
Springfield, Missouri	2.43	18.5	16.29
St Louis, Missouri	2.38	18.67	15.85
Syracuse, New York	2.37	17.06	13.21
Tallahassee, Florida	3.44	19.37	17.87
Tampa, Florida	3.03	15.37	14.51
Toledo, Ohio	1.94	14.59	12.85
Topeka, Kansas	2.07	17.6	15.4
Trenton, New Jersey	3.35	23.38	20.19
Tucson, Arizona	3.58	22.05	20.65
Virginia Beach, Virginia	3.5	21.26	19.88
Washington, District of Columbia	3.57	22.25	19.05
Waterloo, Iowa	2.35	18.76	16.24
Wichita, Kansas	2.31	20.51	16.28
Worcester, Massachusetts	3.32	25.16	19.37
Yakima, Washington	3.31	24.26	22.74
Youngstown, Ohio	1.79	13.49	10.86
US	3.44	20.97	18.95

Adjusting Median Contract Rent for Affordability Studies

Posted Wednesday, May 05 2010

There has been a lot of really good affordability analysis and commentary around the blogosphere lately, but one of the things you'll notice is that different organizations don't always come up with the same numbers for price-to-rent and price-to-income ratios. To a degree that's understandable since getting good data on actual home prices is hard, but it turns out that rents aren't that simple either.

Leonhardt's piece doesn't lay out exactly how Moody's calculates the rent ratio, but Richard Florida's piece does. It looks like he's taking median sales prices from the NAR and dividing that number by the median contract rent from the American Community Survey. Using the NAR median sales price number is probably the best you can do as far as getting relatively timely pricing data that's not just an index. Similarly, the Census' median contract rent is your best bet when it comes to uniform national and local rent statistics. I think what's missing is an accounting for quality when approximating the rent of an equivalent owned home.

The 2008 Census ACS housing data shows that the average size of a home owning household is 2.7 members but the average size of a renting household is only 2.44 members. In order to try to create a more comparable price-to-rent ratio, I suggest adjusting median contract rent by the ratio of average owner household size to average renter household size. It's kind of like using household size as a proxy for square footage; It's not perfect, but it's likely a good first order approximation. In this case the adjustment inflates the 2008 median contract rent of $687 by 10.65% (2.7/2.44) to $760. Not a huge adjustment, but it probably gets you closer to a owners' equivalent rent type measure than the median contract rent value alone does.

I've computed the median sales price-to-rent ratio at the national level using NAR's latest sales report and found the unadjusted price-to-rent ratio to be 20.97. When you adjust the rent using the household size the price-to-rent ratio falls by just over two points to 18.95. Again, not a huge drop but with numerous articles referencing a rent ratio of 20 as an affordability cut off, it seems like it should be noted.

This trend of larger households among owners holds across the country, so I expect that most if not all price-to-rent ratios are being over estimated if they use an unmodified median contract rent as Florida's work does. And of course that's the analysis that would be most interesting — a full calculation of all metro area median sales price-to-rent ratios using household size adjusted rent values. It's already in place for the national housing affordability data I track, but I should have it at a metro level by next week so we can do a direct comparison with Richard Florida's numbers.

Owners' Equivalent Rent compared to Census Median Contract Rent

Posted Tuesday, May 04 2010

Last week I looked at the Census ACS median home value compared to the Case-Shiller index and got the relatively poor results I was expecting. This week I'm looking at the Owners' Equivalent Rent measure from the BLS and comparing it to the American Community Survey's Median Contract Rent value. Below is a chart that shows both series expressed as indices set to 100 in 2000.

Census ACS Median Contract Rent vs. Owners' Equivalent Rent

That's pretty good agreement. This is nice to see as I'm using median contract rent values from the ACS in the new housing affordability data I've been working on recently. If we can assume that the median contract rent tracks an OER like metric at the state and metro level then we've got a good (albeit short) time series that we can use to measure changes in rent at local geographies since the Census provides the data at multiple levels of geography.

Home Prices are Difficult

Posted Friday, April 30 2010

If you were curious how the Census ACS Median Value for Owner-Occupied Housing Units measure compares to an HPI measure like the Case-Shiller index, the chart below is for you.

Census ACS Median Home Prices vs. Case-Shiller

The answer is: not well. ACS home prices are derived from surveys (pdf) that ask how much a homeowner thinks their house would sell for. Clearly this is not as sophisticated as the repeat sales methodology of Case-Shiller and the FHFA which yield a price index not an actual price. Real prices are nicer to work with than indices, but not at the expense of accuracy. I wanted to compare the two to see if there was any way I could have confidence in the ACS number. I don't think there is, at least not at the national level.

Housing Affordability

Posted Saturday, April 24 2010

There was an excellent article in the New York Times this week on housing affordability. It's timely because I published housing affordability data this week as well. There's more coming, but I've started off with national and state price-to-income and price-to-rent ratios expressed as indices. You can see from the national affordability data that we've come a long ways towards restoring pre-bubble trends in price-to-income and price-to-rent ratios (although we're not there yet). The indices are set to 100 in January of 2000.

It's also interesting to look at the state level data to see which states had affordability problems during the boom years and which didn't. California obviously did. Indiana, Nebraska, Ohio — not so much. Some states have the opposite of an affordability problem. Look at the price-to-rent ratio in Michigan!

Data wise, I'm wringing everything I can out of the Census' American Community Survey for this. And then there's all the data that Case-Shiller and the Federal Housing Finance Agency have on metro markets that I can use for HPI. I'll publish metro level data soon. Until then, David Leonhardt's piece is excellent as are all the accompanying tools and charts. The Times is really getting good at this stuff!