The Upjohn Institute Blog
Revised Population Estimates from the Census Bureau
Posted: October 7, 2011 By Brad WattsRecently, the U.S. Census Bureau released intercensal population estimates for each year from 2000 to 2010, which create a complete time series for the period between the 2000 and 2010 decentennial censuses and adjust the estimates for each year to align with the numbers from the full census. Previously, population estimates were released annually from a different estimation series, referred to as postcensus population estimates, which uses birth, death, and migration data to estimate population change. Estimates have also been available from the American Community Survey (ACS), however, these data are, by design, aligned with Census Bureau census data and postcensus estimates.
So how much did the data change during the revision process? At the national level, there is almost no difference between the new intercensal estimates and the previous population estimates; the new estimates are slightly smaller, but differ less than 0.1 percent of the previous estimate values. As shown in the table below, both the level and trend of the population estimates are essentially indistinguishable from one another.
However, in smaller geographic areas, the differences can be more substantial. For example, in Michigan the old and new population estimates illustrate differences in both the level of growth and decline estimated to have occurred during the decade. Although all of the estimates show that Michigans population began to fall off during the middle of the last decade, the new intercensal estimates indicate a much larger drop than the postcensus estimates. The difference is substantial for assessing state population change from 2000 to 2009: the old estimates would indicate an increase of roughly 14,000 compared to a decline of 51,000 that is shown by the new estimates.
Overall, population estimates from the Census Bureau are still quite reliable; however, those interested in tracking regular population change should be aware of the type and magnitude of revisions that can occur, particularly with sub-national estimates. The Michigan example above illustrates one example of the magnitude of the revisions, which can be even larger in percentage terms for smaller regions. Those who use estimate data to show trends, particularly for even smaller places such as individual counties, should be sure to examine the new intercensal estimates to see how the population estimates may have recently changed.
Additionally, users of the ACS should take caution, since these population numbers closely mirror the postcensus population estimates during in between years, but will be reset to reflect the decentennial census in 2010. Using the one-year ACS data to show change in population groups could cause substantial jumps or drops to occur between 2009 and 2010 data series, since old ACS data was aligned with the postcensus estimates, while 2010 ACS data is currently aligned with the 2010 Census.
Brad Watts can be reached at Watts@upjohn.org.
Tagged in:
The Upjohn Institute Blog
Establishment Characteristics of Fast-Growing Metropolitan Areas
Posted: August 19, 2011 By George ErickcekOne of the major research questions that is always on our minds is what separates fast-growing metro areas from low-growth MSAs. In our research, we have found a strong positive correlation between education attainment and income growth and between business dynamics and employment growth. In a 2007 Philadelphia Federal Reserve working paper which was revised and published in the August 2011 Journal of Regional Science, R. Jason Faberman found that high-growth cities have younger establishments than slower-growing metro areas and that this compositional difference explains approximately 38 percent of the variation in MSA growth rates. And, he further found that the percent of an MSAs working-age adults with college degrees is highly correlated with establishment age (negatively), establishment turnover, and MSA growth.
Moreover, he found that factors like industrial diversity and city size have little influence on the average age of the MSAs business establishments. In other words, education trumps industrial clusters when it comes to economic growth because of the association between college attainment and start-ups.
In short, the old adage that I have repeated, time and time again, tell me your industries and I will tell you your future may not be as true as I believed. While a regions legacy remains important, the MSAs ability to attract and retain highly educated individuals and provide a dynamic environment for them to try new things, including opening a business, may be equally or more important.
George Erickcek can be contacted at Erickcek@upjohn.org.
Tagged in:
The Upjohn Institute Blog
Entrepreneurship and Michigan’s Recovery
Posted: August 1, 2011 By Brad WattsHere in Michigan, where the Upjohn Institute is located, economic conditions have finally been showing signs of improvement. Although Michigan suffered through a lost decade that started during the 2001 recession, by several measures Michigan has been one of the better-performing states since the end of the 20072009 recession. For example, according to the Bureau of Economic Analysis, Michigans GDP increased by 2.9 percent in 2010, which put it in the second highest quartile of performersalong with places such as Texas, Virginia, and Minnesota. Although Michigans unemployment rate of 10.5 percent was above the national rate (9.2) in June, its 2.1 percentage point improvement over the past year was better than any other U.S. state except Nevada. Additionally, total nonfarm employment in Michigan grew by 1.8 percent (a gain of 70,800 jobs) from June 2010 to June 2011, which was double the national rate of growth for the same period: 0.8 percent.
Michigans turnaround is good news indeed. However, those concerned with the states long-term success may note that recent trends are more related to a revival of the old economy than a transformation to the new economy. A major driver of Michigans recovery has been the manufacturing sectormore specifically a resurgence of the Detroit Three automobile manufacturers and their suppliersand not new industries or innovative business startups.
As shown in the chart below, the number of new business establishment openings in Michigan has not kept pace with the nation and the rate of openings in the state remains nearly 15 percent below what was seen in 2000. The lack of openings suggests that the state is still producing too few entrepreneurs, attracting too few new companies, and relying too much on tried and true companies for economic success.
Brad Watts can be reached at Watts@upjohn.org
Tagged in:
The Upjohn Institute Blog
New Business Data from the Census Released
Posted: July 1, 2011 By Brian PittelkoIn 2007, over half of U.S. businesses were operated from owners homes and 57 percent grossed less than $25,000 annually. Moreover, businesses owned by women were more likely to be home based than businesses owned by men. These are just two of the findings revealed by the U.S. Censuss release of the full results of its 2007 Survey of Business Owners (SBO). In addition, the SBO press release contained some data highlights.
While the data are not very current, this survey provides information not found in other economic data. Many small business owners are not captured in employment data that is based on Unemployment Insurance (UI) filings, because owners do not pay UI on themselves. The survey results show that there are over 21 million non-employer firms, meaning the owner is the sole employee. These non-employer firms account for 78 percent of all firms, but 3 percent of receipts. The only other set that has non-employer data of this sort is the Census County Business Patterns (CBP).
However, the key difference between this data set and the annual CBP data is the inclusion of gender, race, and national origin. The SBO allows analysis of whether the numbers of women and minorities are growing in business, and whether these groups have achieved business ownership proportional to their populations. For example, since the last SBO in 2002, black or African-American business ownership has increased from 5.2 to 7.1 percent of all firms, though that is still below the 12.8 percent of the total U.S. population. Also, unlike the CBP data, the SBO is available at the local level down to townships, although there are some data suppression issues at those levels. Therefore, while older than some data, the SBO is a valuable data set for both owner characteristics and sub-county geographies.
Finally, the report gives a glimpse of the financial environment many small businesses face. For example, virtually the same number of businesses (about 10 percent) used a credit card for early financing as those who received a business loan from a bank. The survey took place in 2007, just before the commercial lending market dried up. Even in good times business loans were only 10 percent of the market.
Tagged in:
The Upjohn Institute Blog
A Look at the Housing Market
Posted: June 15, 2011By Brad Watts
Although many indicators of economic performance have posted steady (if modest) gains in recent months, the nations housing market remains in the doldrums. According to the National Association of Realtors, existing home sales have posted a few gains over the past few months, yet still remain at a level significantly lower than during the same month one year earlier. Data from the U.S. Census Bureau, which tracks new home construction, also shows that new residential construction, as measured by building permits and completions, has fallen over the past year. Both findings are especially discouraging given that 2010 was a poor year for housing and 2011 does not appear to be any better despite the fact that the recession ended roughly two years ago.
So just how bad is the housing market? A look at long-term historical data on sales of newly constructed housing provides a stark visual reminder of how inflated the market for new homes became during the mid-2000s, and how rapidly the market has contracted since then. The chart below shows the actual monthly annualized rate of new home sales nationwide, along with a simple regression trend line that represents the hypothetical long-term expected rate of sales. As clearly illustrated by the strong break from the trend line, sales took off at a faster-than-expected rate starting in the late 1990s and continuing until hitting a peak annualized rate of 1.389 million new homes in July 2005. After that, the market crashed rapidly: in February 2011, the annualized rate of new home sales was only 278,000its lowest point since data were first collected in 1963.
On the positive side, the historic low rate of new home sales suggests that the market should be near bottom. As of April 2011, the annualized rate of new home sales, 323,000, was approximately 500,000 units below the level expected by the long-term regression trend, which suggests that housing developers, construction firms, and realtors may look forward to a doubling of the market over the next few years. Still, expectations should likely be held in check for the near term, given the still slow market for existing homes reported by the National Association of Realtors, which can be expected to pick up before the sale of new homes. Additionally, an aging U.S. demographic along with ongoing changes to the regulatory and financial environment for mortgages may also restrict demand to lower levels than were seen over the past two decades, even as the overall economy expands.
Brad Watts can be reached at Watts@upjohn.org.
Tagged in:
The Upjohn Institute Blog
Matching Wages and Home Values
Posted: June 14, 2011By Brian Pittelko
A recent article from the Bureau of Labor Statistics (BLS) entitled Pay Comparisons Between Metropolitan Areas in 2010 looked at the differences in wages between metropolitan statistical areas (MSAs). The BLS created a simple index using wage data from the 2010 National Compensation Survey (NCS) with U.S. wages set to 100. The MSA with the highest index rating was the combined statistical area (CSA) of San JoseSan FranciscoOakland, California at 120 and the lowest was the BrownsvilleHarlingen, Texas MSA at 80. These index ratings means that these areas have wages 20% above and 20% below the national average, respectively.
I added another dimension to their comparison by using 2009 American Community Survey (ACS) data on the value of owner-occupied homes. I created a similar index and matched it to wage data in order to see if wages matched home values. The figure below shows the spectrum of home values and wages. Before considering a move to the Bay Area for a wage hike, one should realize that they are the furthest point on the top right, indicating that while wages are 20% above average, home values are 200% above average. Brownsville-Harlingen, while having lower wages, has home values of less than half of the national average.
The figure seems to show that most metro areas match
low-to-low or high-to-high wages and home values. Only 11 of the 77 metros are
either in the top left or bottom right quadrants. The top left could reflect an
area with a strong quality of life: residents are willing to pay high home
prices while receiving low wages because it is simply a great place to
live. Richmond, VA, Miami, FL, and
Phoenix, AZ are included in this quadrant. The bottom right suggests the
opposite: residents require higher pay
relative to housing costs to live in these areas. Detroit, MI, Reading, PA, and
Rochester, NY are in this group.
Brian Pittelko can be reached at Pittelko@upjohn.org.
Tagged in:
The Upjohn Institute Blog
Staying Put
Posted: May 25, 2011By Brad Watts
It has been said that people are the most difficult resource to move, yet according to the U.S. Census Bureau CPS Geographic Mobility data, roughly 12.5 percent of the population changed residences between 2009 and 2010. Although this data suggests that America is a nation of movers, recent levels of mobility represent a significant decline compared to a decade earlier, when 16.1 percent of the population changed residences between 1999 and 2000. As the chart below shows, mobility declined across all age groups, although younger age groupswhich are the most mobile overallsaw the largest decrease in mobility rates during the past decade. During the same period, mobility rates for persons in the oldest age categories, which typically represent retirees, remained relatively stable. It should also be noted that the percentage of people that are moving long distances has also fallen; between 1999 and 2000, 23.4 percent of movers had migrated across a state line, versus only 14 percent during the 2009-2010 period.
Not surprisingly, the reasons for the change in mobility
rates appear to be frequently economic in nature. The reason that increased the most in
percentage terms was wanted cheaper housing, while the percent of movers
citing reasons related to desires such as wanted own home, not rent and
wanted new or better home/apartment declined substantially. Interestingly, the number of movers reporting
that a new job or job transfer was the reason for the move also declined
significantly, which would appear to reflect a lack of job opportunities during
the period.
Brad Watts can be reached at Watts@upjohn.org.
Tagged in:
The Upjohn Institute Blog
“Smart” Cities and Population Growth
Posted: May 11, 2011By Brian Pittelko
A new paper by John V. Winters in the Journal of Regional Science asks Why Are Smart Cities Growing? Who Moves and Who Stays." Winters finds that having high levels of educational attainment actually does help cities grow.
The paper looks metropolitan statistical areas (MSAs) with a
large population of residents with college degrees. The analysis was conducted
using primarily Census migration and education data from 1980, 1990, and 2000.
Other factors included in the regression were manufacturing employment, income, weather
and region. Winters finds that having a
high population of residents with a bachelors degree or higher had a
significant positive effect on in-migration, out-migration, and net-migration. Many
of the areas were smaller towns centered around a large university, such as
Iowa City, Iowa and State Collage, Pennsylvania, that seem to be retaining a
certain population of their college graduates. Winters further analyzes MSAs that are not
considered college towns and again finds that these smart" cities are also
growing, although the driver of in-migration in these areas is still those
enrolled in college.
The only criticism I have is that the limited amount of variables besides education are likely not painting the whole picture. Other factors could be driving migration. We do not know what types of jobs are keeping the college grads, nor do we know if there are any other quality-of-life variables besides weather that are influencing net migration. Still, the research seems to say that having a well-educated population does indeed have some population growth benefits.
Brian Pittelko can be reached at Pittelko@upjohn.org.
Tagged in:
The Upjohn Institute Blog
Thoughts on Industrial Clusters
Posted: April 20, 2011 By George ErickcekI have never been completely won over by the idea of industrial clusters as a pathway to economic growth and sustainability. History is full of examples of struggling industrial clusters: steel in the Youngstown-Pittsburgh region; textiles, first in New England and then in the Carolinas.
Closer to the Institutes home, the paper industry in Kalamazoo, Michigan is a great example of a cluster that started out as a highly innovative, locally-owned set of firms and suppliers. Due to its presence, Western Michigan University established its Department of Paper Engineering which is still one of the most respected paper technology research centers in the world. However, with the passage of time, the industry lost its edge. Employment in the industry in Kalamazoo County has dropped from over 10,000 in the 1950s to 1,800 today.
However, I want to believe in the potential of industrial clusters. Mark Muro and Bruce Katz at the Brookings Institution helped me with their paper The New "Cluster Moment": How Regional Innovation Clusters Can Foster the Next Economy, which highlights the potential benefits and recommend policy options that could help promote the growth of existing cluster. One of the authors most valuable observations is that communities should not try to create clusters. Either they exist or they do not: public policy will rarely be effective in connecting firms to create a new cluster.
Yet, the authors seem to neglect my major concern about clusters: their sustainability. As an economic developer once told me, Cluster analysis tells me where the region has been, not where it is going. As a community works to support its existing clusters, it should also be aware that, much like steel in Pittsburgh and paper in Kalamazoo, clusters can deteriorate or even disappear.
I would like to see a checklist of factors to watch that could warn local regional stakeholders that their regional cluster is in trouble. A couple of factors could be:
- The loss of local ownership
- The loss of Research and Development activity
- The loss of market share
- An increased focus on production cost.
Tagged in:
The Upjohn Institute Blog
The Impact of Hard Times on Entrepreneurs
Posted: April 11, 2011By George Erickcek
Researchers are in conflict, once again, over the importance of small businesses in creating jobs. The current debate centers on whether small businesses weathered the Great Recession (2007 to 2009) better than larger establishments.
As recently summarized in an article by Kevin L. Kliesen and Julia S. Maus at the St. Louis Federal Reserve, the debate about the importance of small establishments in creating jobs started back in 1979 when David Birch claimed that small establishments employing fewer than 20 employees generated 2/3 of all jobs. Birchs work generated numerous papers which both fine-tuned and lowered his original findings. The general conclusion reached was that small establishments generate a lot of jobs; however, many die after a couple of years, so their net employment impact is much smaller.
Kliesen and Maus concluded that from 1992 to 2010, small establishments employing fewer than 20 workers created 16 percent of all net jobs during the period. Additionally, if the 2007 to 2009 recession is removed from the sample, small establishments account for 28 percent of all net jobs being created. In other words, hard times hit small establishments particularly hard.
These findings countered those of the Kauffman Foundation, which reported that entrepreneurial activity hit a 14-year high in 2009. Scott Shanes research at Cleveland Federal Reserve found that the Kauffman Foundation findings were correct in that the number of persons going into self-employment rose sharply in 2009; however, a greater number of existing self-employed individuals closed their businesses. The number of unincorporated self-employed persons dropped from 10.2 million people in November 2007 to 9.8 million people in June 2009.
Perhaps recessions generate what I call accidental entrepreneurs, or persons who open their own businesses because they have lost their previous employment. Some of these persons may become very successful and, in fact, benefit from that push out into the dynamic world of entrepreneurship. Nevertheless, hard times are simply hard times for all businesses, large and small.
George Erickcek can be reached at Erickcek@upjohn.org.
Tagged in: