Posted by: Josh Lehner | August 11, 2014

Unemployment Insurance Update, July 2014

New claims for unemployment insurance at at or near all-time lows in both Oregon and the U.S. This is encouraging in that it shows the level of job loss is currently at very low rates, and effectively as low as during the past expansions. In other words, if you have a job, the probability of losing it is very low today. Of course this does not take into account the continued high levels of the long-term unemployed, or a lower LFPR, which have been detailed previously.


As such, the level of unemployment insurance benefits paid out is back down to pre-Great Recession levels. Unfortunately the exhaustion rate — the share of Oregonians who start receiving benefits and reach the end of program without leaving, i.e. do not find a job — is still high, although it is about half-way back.UIBenfits_0714

With relatively low levels of job loss, low levels of benefits being paid out and the expiration of the extended benefits at the beginning of the year, the recipiency rate in Oregon is now 34 percent. This means that 1 in 3 unemployed Oregonians are currently receiving benefits and 2 in 3 unemployed Oregonians are not.


The reason for this is that most unemployed Oregonians today are not job losers (the vast majority of benefit recipients). Most are entrants into the labor market and therefore not eligible for benefits. This is largely a good thing. It indicates that the number of Oregonians losing jobs is lower. It also shows that more Oregonians are looking for work — about 10,000 more in the past 8 months. As the job market continues to improve, it will draw more workers back into the labor force, providing some cyclical rebound.


Of course, all of the above speak to the rate and level of job loss in Oregon, which has not been the primary economic issue for a number of years. The problem is the hiring rate. Job growth has accelerated in Oregon unlike the vast majority of the country. However, the rate of growth is still three-quarters throttle relative to historical expansions. Oregon is getting there, however we are still digging out from the Great Recession.

Mark (and Kanhaiya) are working on a migration study, which Mark will present at this year’s Oregon Economic Forum in October. There will be lots of great information not only on historical trends but also looking forward, what can the state and regional economies expect in terms of population and economic growth. One byproduct of this research is looking at worker flows within local communities. A family may move to Portland or Salem of the Coast, but that does not mean that they necessarily work in the same county they live in. There are can be substantial cross border flows both into and out of counties and communities in Oregon. What we have tried to do here is categorize counties based on these patterns.

  • Boardroom: Lots of local jobs relative to population, resulting in substantial inflow of workers from a different county
  • Melting Pot: Large commuting flows in both directions relative to population; county is both a job hub and residential community
  • Autonomous: Small commuting flows in both directions relative to population; most county residents work within the county and not many commuting in
  • Bedroom: Relatively few local jobs, resulting in substantial outflow of workers into a different county


A few key takeaways from the map. Washington County has the second highest employment count in the state, however it is classified as a melting pot county. Even though Washington County, on net, has approximately the same number of local jobs as employed residents, there is substantial commuting in each direction — about 80,000 each way. Malheur County in Eastern Oregon is a boardroom county given it has more jobs than employed residents, which results in a large influx of (mostly) Idahoans. The state’s bedroom counties are largely a part of, or adjacent to larger metropolitan areas, where job opportunities may be more plentiful. In each of these counties, there is a larger number of employed residents than the number of local jobs. Lastly, 15 of the state’s 36 counties can be classified as autonomous, meaning that most residents work in their home county (about 94 percent, compared with a 75 percent statewide average). The typical autonomous county in Oregon only has gross commuting flows of 12 percent or less — meaning the if you add up the number of workers commuting in and those commuting out, as a share of the local job base, it is just 1 in 8 or 1 in 9 workers moving in either direction. This last group — Oregon’s large share of autonomous counties — distinguishes the state a bit from our neighbors, as seen below.


Most counties in Idaho or Washington are more of an either/or type, with relatively few that are balanced when it comes to net workers as a share of local jobs. This is a slightly different measure than used above, but a similar concept. Potentially, this indicates a few different things. Among them:

  • Population distribution, or distance between cities
  • Size of counties. Oregon is the largest of the three states (in sq miles) but also has the fewest number of counties. The average Oregon county is 40-50 percent larger than our NW neighbors.
  • Industry mix (seasonal employment; retail, health care, professional services, etc, each may have different commuting patterns)
  • Home price differential
  • Transportation infrastructure (and access)
  • Border effects
Posted by: Josh Lehner | July 30, 2014

Graph of the Week: Unemployment and Job Polarization

Many discussions on the economy revolve around employment (industries, job polarization, etc) or the unemployment rate (with some discussion on short- vs long-term unemployed) but rarely do we put these two together. This is largely due to data and sample size concerns, which can make extracting the signal from the noise challenging at the state level. With that being said, recent work from our office resulted in detailed, occupational unemployment rates — used in discussing our OED friends’ job vacancy survey — but these exact figures do need to be taken with a grain of salt, even if the general patterns are true.

The graph tries to help answer the question of who are the unemployed in Oregon. It compares the distribution of unemployed Oregonians across occupations, both short- and long-term, with the distribution of employment across occupations. These figures are akin to location quotients. Values greater than 1 indicate that workers from these occupations represent a disproportionately larger share of the unemployed than the employed. For example let’s take Food Preparation, one of the red dots. Food Prep workers account for 5.5 percent of all employed workers, however they account for nearly 12 percent of the short-term unemployed and just over 10 percent of the long-term unemployed (looking for work 6 months or longer). Using the location quotient methodology, this places the Food Prep occupation at nearly 2,2 on the graph indicating that Food Prep workers, effectively, comprise twice the amount of unemployed Oregonians relative to employed Oregonians.


There are a number of interesting patterns that stand out. Among them: traditional, blue-collar occupations are high on both short- and long-term unemployment; many occupations with high rates of employee turnover also have high short-term unemployment, which is to be expected; occupations requiring college degrees are nearly all low among the long-term unemployed and most low among short-term as well. One additional point I will make is that Construction has seen considerable improvement in recent years. It is still red, however it used to much further out. This is because employment is increasing and unemployment decreasing among these workers (a win-win, not just do to the unemployed dropping out of the labor force).

Simplifying the graph further and looking at it through the job polarization lens yields the Bonus Graph of the Week.


Here the pattern is very clear. High-wage occupations have low levels of unemployment, middle-wage occupations are about what one would expect, relative to employment, while low-wage occupations have disproportionately high levels of unemployment.

Posted by: Josh Lehner | July 28, 2014

U.S. Growth, Update on the States

Besides just stronger national growth, another way for top line U.S. job figures to improve is for more states or regions to share in the recovery. Back in December I noted that the Northeast and Midwest were growing much faster than their housing boom rates, while the South and West were lagging (in particular given low population growth and the housing bust). Expectations were that the South and West would accelerate moving forward, but in order for national figures to improve the Northeast and Midwest would need to hold onto those stronger rates of growth. So how are things looking today? Well, the acceleration has come along the West Coast and in the South but much of the Northeast and Midwest has slowed down. Map of Census Regions and Divisions.


This leaves us with a national job growth figure that is slightly stronger so far in 2014 than in recent years but much of the movement in the U.S. employment picture is happening below the surface, down at the state and county level. As seen below, the number of states experiencing job growth of 1 or 2 or 3% hasn’t really changed. However this is due to the shifting nature of job growth across the country. The only states to see sustained acceleration in job growth (improvements of 1 percentage point or more, relative to 2011-12 rates) are Delaware, Florida, Nevada and Oregon. The latter three of which were hard hit by the housing boom and bust and as housing rebounded in 2013, growth picked up. However these states’ improvement was offset by deceleration of 1 percentage point or more in Alaska, Michigan, N Dakota and Virginia. As evidenced in the graph below, the geographic footprint of the recovery really is not broad based with just a dozen or so states experiencing job growth of 2% or more.

StateShare_0614However, the one portion seeing stronger rates of growth, and acceleration are the large metropolitan areas, see here for more details.


The questions becomes, whether or not we expect the medium, small and rural economies to pick up and whether or not the large metros can continue to grow at 2% or more in terms of employment. I think the answers are yes, and yes, from a fundamental point of view — ignoring potential macroeconomic shocks. As housing and government continue to improve (the two large weights on the recovery) this will help medium, smaller and rural economies as these industries play a disproportionately large role here, relative to bigger cities. In terms of the large metros, many of those good economic things you hear about — agglomeration effects, knowledge spillovers, clustering, etc. — happen in specific locations, which are usually bigger cities. These impacts, along with the continued urbanization of the population result in a relatively bright outlook for the country’s largest metros.

Posted by: Josh Lehner | July 24, 2014

Oregon Spider Chart, June 2014

This Graph of the Week is an update on the Oregon spider chart. See here for more on the construction of the graph which follows the pioneering work by the Atlanta Fed at the national level.

As a reminder the chart tracks progress across a wide variety of labor market indicators. As each measure improves from its recessionary trough (the red dot, 0%), the line moves outward from the center. Once the measure reaches the gray, dotted line, that indicates it has fully regained the level (or rate) last seen prior to the Great Recession.

The two sentence takeaway is: Labor market leading indicators continue to improve and over the past year employer behavior has picked up nicely. However, employee confidence and utilization measures — the feel good nature of the economy — are still just about half-way back to pre-recession levels.


One minor tweak this round is the substitution of Marginally Attached Workers for U-6. U-6 is the combination of U-3, MAW and PT for Economic Reasons. Showing both of these additional components that make up U-6 provides a bit more detail to the underlying currents in the labor market. As such, one can see that PT for Economic Reasons has declined (improved) substantially over the past year (really the past 6 months), while those MAW have increased. We know that the decline in the unemployment rate has not been entirely for good reasons, with the labor force shrinking, some of these dropouts are likely showing up in the larger number of MAW (those who want a job but have not looked recently). So even as U-6 falls, there are both good reasons (fewer PT for Economic Reasons) and bad (more MAW).

Posted by: Josh Lehner | July 23, 2014

Metro Size and Growth, An Update

Just a quick update on employment growth by metro size, similar to previous work (HERE and HERE). This takes a look at county level employment data (QCEW through the end of 2013 was released recently) and categorizes it using the USDA rural-urban continuum codes.

The largest metros (the 51 largest have a population of 1 million or more) have seen the strongest gains in recovery [1]. The second set of metros have seen some acceleration and the nonmetro (rural) counties have seen deceleration over the past year.


In terms of how these growth rates look from a historical perspective, I took today’s rural-urban codes and applied the same categorizations back to 1980 (so a place like Las Vegas which is big today but not then, is still classified as a large metro). What stands out are the largest cities outperformed other locations both in the late 1990s boom and today. However smaller cities and rural areas performed just as well, if not better than, big cities in the 1980s and somewhat during the housing boom.


The question is what is the normal pattern of growth? Is is the 1980s and housing boom years where most areas grow about the same? Or is it the mid to late 1990s and so far in the 2010s where larger cities outperform? It’s certainly an open ended question, but most outlooks call for continued urbanization of the population and for metro areas to outperform rural economies in general. This is at least partially due to the fact that all those good economic things — agglomeration effects, knowledge spillovers, clustering, etc — happen in certain locations, which are usually bigger cities. The case could also be made that the housing boom was an equalizer in which small and medium sized metros outperformed due to stronger population growth and the associated housing demand and activity that went along with it. This stronger growth also may have pulled some away from the larger cities at the same time. In this version of the story, today’s pattern of growth is simply a return to the expected one, which was interrupted by the housing boom where inflated asset prices/wealth may have impacted location decisions.

It is also interesting to note that only the largest cities have seen growth rates return to pre-recession levels, while the others lag. This is at least partly due to the nature of the Great Recession in which housing and government have been large weights on the recovery. These jobs also play a disproportionately large role in many medium, smaller and rural economies than in big cities. Not because these areas have so many of these types of jobs, but rather because of lower levels of industrial diversification, that generally occur in bigger cities (ballet dancers, advertising firms and the like).

[1] Breaking these big cities down into more groups reveals the same trends. Large cities with populations of, say, 4 million or more compared with those with populations of 1-4 million have the same employment trends over the Great Recession.

Posted by: Josh Lehner | July 18, 2014

Jobs, Labor Force and Population

Much attention is being paid to measures of labor market health like the labor force participation rate and the employment to population ratio, with good reason. However, I would caution drawing too many conclusions based on the top line comparison between the U.S. and Oregon. The reason being Oregon does have a larger Baby Boomer cohort than the average state and retirees continue to move to Oregon, thus the state is aging faster than some other areas. This has big implications on measures like LFPR and the employment to population ratio. To be sure, that does not mean the state is doing great, but neither is it doing horribly either. As with many things, the truth is somewhere in the middle and Oregon is largely following national trends once you account for demographic changes locally.

First, let’s take a look at the labor force participation rate by age cohort in Oregon relative to the U.S. As detailed thoroughly a few weeks ago, Oregon used to have a LFPR higher than average, however since 2005 those gains have vanished and the state now trails the national average in recent years. What caused the relative decline? The graphs below compare Oregon’s LFPR by age cohort to the nation. For prime working age adults in Oregon, they are participating in the labor force just as much, if not slightly more, than their national peers. Where Oregon really differs from the U.S. is in the 65+ age cohort. Not only is this group becoming bigger and bigger as the Baby Boomers age, but Oregon Baby Boomers have a lower participation rate than the national average. This, appears to be a major driver in the overall change in Oregon’s relative LFPR position. This is also not much of a concern, provided these individuals want to be retired and out of the labor force and not forced out. The concern our office has with these figures are among those prime working age groups. Oregon used to enjoy a strong LFPR advantage relative to other states. While population growth and migration continue to result in an above average growth rate here for Oregon, which overall is a good thing for longer run economic growth, the state is no longer seeing both above average growth and above average participation.


On the employment side of the ledger, the story is largely similar. For an upcoming housing report, I have been examining employment trends in 25-44 year old Oregonians (those in the so-called “root setting” years) as these ages are when many individuals get married, have kids, buy a home, enter the heart of their career and the like. However, this also means I have this data readily available. Here too you can see that Oregon’s employment to population ratio for all adults largely tracked the U.S. overall but then lags a 2-3 percentage points in recent years. Given sample size concerns, it is unclear just how much is fundamental change vs data issue changes.

EPop16plusWhere you see similar trends to the U.S. are among the younger prime working age Oregonians. The year to year volatility I would chalk up to sample size concerns with the underlying Current Population Survey data, however the trends here are clear.


The younger adults are OK in terms of employment trends relative to the nation, although that does not mean things are OK overall given the low ratio today. With a stronger economy, the ratio should be higher, back up around the 80 percent mark seen in previous expansions.

This does mean, however, that Oregon’s lower employment to population ratio is due to those 45+ years old. This is the bigger concern given that it is not purely due to those in and around retirement age. Primarily the decline and lack of improvement can be seen among those 45 to 54 years old, relative to the nation. It is possible it is at least partly an industry mix issue with industries like construction and manufacturing bearing the brunt of lost jobs in the Great Recession, however that cannot be the whole story. Clearly these job losses are impacting LFPR as seen above (the dark blue line, for those 45-54 years old, has seen the largest relative decline since 1990).

So putting all of this together, what is the outlook? With an improving economy will come more jobs, and likely at least a little more labor force participation. However these gains are likely to be short lived as the longer run demographic changes will weigh on these top line figures that count all adults. It is important to remember that most of these changes are demographic and/or structural and focusing on the prime working age cohorts, the economy really is, and will be, improving.  See here for more on our office’s outlook.


Posted by: Josh Lehner | July 16, 2014

Quick Takes: June Jobs, Housing & Poverty

Back in the office after a week away and just wanted to give a few quick takes on recent figures.

Quick Take 1: June Jobs

Whoa. 4,300 lost jobs in June? Is the recovery, as weak as it has been, over? Fortunately, as Mark said in The Oregonian article, “The truth is, unfortunately, a little more boring.” First, withholdings out of Oregonians’ paychecks continue to grow solidly, indicating steady job growth and the state’s coincident indicators are all improving. So this is highly unlikely the first step toward the next recession, but that does not mean job growth didn’t slow or that there were even some job losses. However, the magnitude of the job losses is, well, interesting. Check out this table. In June of last year, Oregon added 5,800 jobs from May to June on a not seasonally adjusted basis, which translated into small job losses on a seasonally adjusted basis. From May to June of this year, the state added 9,600 jobs over the month (two-thirds are many), but the state lost 4,300 jobs? Hmm…


You can chalk this up to the seasonal factor, which one uses to convert the monthly changes into seasonally adjusted changes based on the normal fluctuations — trying to extract the signal from the noise. Like Mark said, boring stuff. Here are the June seasonal factors for each year since 1990. Two things stand out. First 2013 seasonal factors were higher than recent years (thus helping prop up the seasonally adjusted changes a year ago). Second the 2014 seasonal factors appear to be quite low, thus helping to deflate the seasonally adjusted changes today.


So what’s the truth? Hard to say exactly at this point, but so far, based on just one interesting data point, there are no broad worries about the recovery continuing. If one applies seasonal factors more in-line with historical values, then the state’s employment was likely flat over the month. Not great news, but better, and as The Oregonian‘s headline says, the recovery is far from complete and that is true.

Quick Take 2: Housing and Construction

Construction bore the brunt of the preliminary job losses in June and even after adjusting the seasonal factor to something more in-line with historical figures, the industry likely did lose jobs (just not as many). The ratio of construction workers to housing starts is still elevated (see here). The relative strength shown in the ratio likely has to do with nonresidential construction projects like the second phase of the Intel expansion, office building renovations, MAX construction and the like. Housing starts are relatively flat, but if one or more of these larger nonresidential projects winds down, then there could be weakness in the employment figures.

Speaking of housing and the latest RMLS report, here is a quick update on housing affordability. If one is able to save up and put a large down payment into the loan, the market – even with strong home price appreciation – is still affordable, if you can find a home with the limited supply. However, with just a small down payment (thus paying mortgage insurance and a larger loan balance), the median home in the Portland MSA is at the very high end of affordability relative to median family income. The record levels of affordability seen just two years ago are gone, with strong price gains and higher interest rates, but in the bigger picture, affordability is back to the levels seen in the late 1990s.


Quick Take 3: Poverty Hot Spots

The U.S. Census Bureau recently put out a report on high poverty areas across the country. As written about in The Oregonian, Oregon saw a large increase in such areas from 2000 to 2010. While the Census report provides a high level look at this issue across all the states, I just wanted to highlight some great work that Oregon’s Department of Human Services did last year on the same topic. You can see the report here: High Poverty Hotspots. One can scroll each county in the state and see details regarding some of the local socioeconomic characteristics of these hotspot areas along with SNAP and other DHS/OHA service data. The DHS report does a really good job of detailing these changes and highlighting the areas of the state that are experiencing high poverty levels.




Posted by: Josh Lehner | July 10, 2014

Oregon’s Coincident Index

Each month the Federal Reserve Bank of Philadelphia releases a coincident index for all states. The coincident index is comprised of four underlying variables and is designed to track current economic conditions, thus it is neither a leading or lagging indicator. As seen below, the vast majority of states across the country are seeing improving economic activity in recent months.


Diving into the four underlying components yields the following for Oregon. Both employment and real wages in the state are continuing to grow. The unemployment rate, while holding fairly steady in the past few months, continues to improve slowly since the Great Recession. Manufacturing hours worked have fallen slightly in recent months from about 41 hours per week to 40 hours, which is not out of line with the previous expansion. So far, each variable is looking good or at least improving. When they all move sideways together or we see stronger declines in one or two, that is when it may be time to worry. Today we’re still looking good.


Posted by: Josh Lehner | July 9, 2014

Oregon Exports, 2014q1

Following a slowdown in that extended from the summer of 2012 through fall 2013, Oregon exports are now growing again. In dollar value, Oregon’s exports in the past 12 months are now just 2.5 percent below their all-time peak reached back in the summer of 2008 and are currently at a post-Great Recession high. [Data through May 2014 shows exports just 0.9 percent below peak levels on a 12 month sum basis.]


On net, the increased exports over the past year are entirely due to high-technology products, largely destined for Asian ports (63% of the gain) or Costa Rica (17%). Of course this masks over the changes in other industries. Strong gains in machinery, transportation equipment and food products were offset by declines in agricultural, chemicals and waste products. Even as agricultural exports are down slightly on the year, they have maintained their higher level than in the past, likely due to higher commodity prices in recent years.

Exports have increased to 20 of Oregon’s top 25 destination markets over the past year. However, in a broader perspective, nearly all of the gains in the past couple of years have been to China and Canada, Oregon’s 2 largest markets. Out of the state’s top 5 destination markets, however, only exports to Canada are currently at an all-time high.


Exports to China continue to be dominated by high-tech products, which are growing again, in addition to gains in machinery, chemicals and transportation equipment. Canadian exports are increasing with growth evenly split between metals and machinery and all other industries, including gains in high-tech and chemicals.


« Newer Posts - Older Posts »



Get every new post delivered to your Inbox.

Join 294 other followers