Sunday, July 31, 2011

California versus Texas: Part 6: The model

The first thing to say is that my model isn't wildly precise. However, where it predicts lots of job growth there is job growth, and where it predicts lots of job losses there are job losses. It works well enough to reassure me that I am looking at the right things.

The real national economy is very complex and very diverse, so a simple formula is never going to give precise answers. However, to have any success at all, that simple formula has to get some big things right.

There are a few states where it fails badly, but even so it fails in an interesting way. Kansas and Nebraska are neighbors, and the model greatly over predicts job growth in both states.  There seems to be some regional factor at work. West Virginia and Kentucky do much, much better than predicted. Both are coal mining states with a lot of labor intensive underground mining, and coal prices have risen dramatically in the last ten years. And then there is Washington state, which does much better than expected. Washington state is an anomaly in many ways, and I don't know why. It might be due to the presence of Boeing and Microsoft. The state  has very good pay after taxes and housing costs are taken out. While the congressional delegation looks like California, the state tax system looks more like Texas. Washington state is either lucky or they are doing something right which my model doesn't capture.

How it works

The idea behind my model is that job growth will happen if employees are profitable. If an employee produces $50,000 per year of output, but only costs $40,000 per year to the company, then a company will tend to expand and add jobs. The rate of job growth is proportional to profitability. So a state that has a well educated workforce working for moderate wages will see job growth.

I calculate a labor force value which increases with the percentage of the the work force which is college educated. If the state is a right-to-work state that increases the labor force value by another 17%.

Labor force value = 22845 + (775 * Percentage college educated)

I then subtract the mean wage to get profit, and multiply that by a factor to get employment growth.

Job growth = (Labor force value - mean wage) * 6/1000

I then subtract the job losses due to rural populations, and add the regional adjustment, if any. The final formula is:

Projected job growth = ((Labor force vale - mean wage) * 6/1000) - Percentage of population which is rural - Regional correction


California  has an average wage of $50730, 6% of the population is rural, and 30% of the workforce has a bachelor's degree. It is not a right to work state. The regional correction is +30.

Labor force value $ = 22845 + (775 * 30) = 46095
Projected job growth = (46095 - 50730) *6/1000 - 6 + 30 = 3.8

California versus Texas:Part 5: Rural job losses and regional corrections

The Rural Population Effect

Studying the data I noticed that the states with high job growth all seemed to be highly urbanized. On the other hand there were states that seemed to be doing everything right, where the job growth was mediocre. The thing that stood out about those states is that they had large rural populations.

This chart shows job growth versus the percentage of the population which is urbanized. States with low urbanization never seem to do well. Highly urbanized states can do well, or they can do badly.

Over the past decade, rural areas seem to lose jobs at the rate of 1 job per 1000  residents per percentage point of rural population. So I would expect a state with a 40 percent rural population to lose 40 jobs per thousand state residents over the cycle from this rural population effect.

These rural job losses are consistent with what is known about strong productivity gains in industries like farming. I think that this is the tail end of a process that has been going on for over a hundred years. In 1880 agriculture accounted for over 41% of US employment. People had to live on the land they were farming, so a scattered rural population made sense. Today, only 1% of the US population works in farming, and a scattered rural population no longer makes economic sense unless commodity prices rise. For the past century, towns and cities have been the engines of the economy.

It is important to note that the Census Bureau's definition of urbanization includes some quite small communities. Anybody living in a town of more than 2500 people is an urban resident.

California is one of the most urbanized states, with 94% of the population living in urban areas. Texas is at a slight disadvantage, with 83% of the population urban.

Regional Corrections

When I was a kid, I had a small tear off calendar with a different humorous quote for each day. One that I have never forgotten was the definition of Flannagan's Finagling Factor. This is:

 'That quantity which, when multiplied by, divided by, added to, or subtracted from the answer you got, gives you the answer you should have gotten.'

I'm going to have to introduce a few of these to make the numbers work out, but I promise to only use them across broad geographic regions, and there are only 3. What I think they represent is job creating factors that my model doesn't include.

The first one is for Alaska. My model predicts Alaska should be losing lots of jobs, when in fact it is gaining a lot. I think this is because Alaska is a state with few people and lots of oil. This one is +116 jobs/ 1000 population.

The second one is for the Rocky Mountain west. This one is +18 jobs / 1000 population. This area includes Montana, Idaho, Nevada, Utah, and Colorado. This might have something to due with mining and high commodity prices.

The third one covers three regions. These are the West Coast, the states along the Mexican border, and the East Coast from Virginia north to New England. This one is +30 jobs /1000 population. This might have something to do with the growth of international trade. The states involved are Washington, Oregon, California, Arizona, New Mexico, Texas, Virginia, Maryland, Delaware, Pennsylvania, New Jersey, New York, Connecticut, Rhode Island, Massachusetts, Vermont, New Hampshire and Maine.

These regional corrections are the final element. In my next post I will show the predicted job growth and compared it with the actual job growth.

Data on urbanization is from Table 27 of the Statistical Abstract of the US from the Census Bureau. I'm using 2006 numbers.)

Friday, July 29, 2011

California versus Texas: Part 4: Wages and Housing cost

I think there are two factors that drive wage levels. On the demand side, the employer side, there is the education level of the workforce. On the supply side, the worker side, I think housing costs are an important driver of wages. People can always pack up and move to another state if they don't like their standard of living after taxes and housing costs have been paid.

This chart demonstrates the correlation between housing costs and wages for 20 states.

New Jersey and California have the highest costs, while Iowa has the lowest. Affordable housing keeps workers happy even if wages aren't the best. I think that California, New Jersey and New York could be doing much better if their housing became more affordable.

The next chart shows housing costs versus job growth over the last economic cycle. Very expensive and very cheap housing seems to be associated with low job growth

Which state is best?

In order to see which state provides the best standard of living I looked at two median wage earners in one median house. I assume they file taxes separately and have no children. I used a web based paycheck calculator to find their take home pay after taxes, then I deducted the monthly housing expense. I ignore the mortgage interest tax deduction.

California and Texas come out in the middle of the pack with a similar standard of living to each other. It is important to remember that there are huge differences in housing costs between different areas of California, so this might not apply to any specific Californian city.

Washington state does extremely well. My jobs model doesn't work at all for Washington state. It is an anomaly in many ways, and I don't know why.

Iowa has the lowest wages of any of the states listed, but makes up for it with very cheap housing. Utah workers work cheap, and Utah has one of the best job growth rates of any state.

The bottom line here is that affordable housing keeps wages down and living standards high.

( Important Data Note: For housing cost I'm using the  median selected monthly owner cost from 2003   (Table 957) from the 2006 Statistical Abstract published by the Census Bureau. Based on a review of  the Case-Shiller indices, current housing prices are close to 2003 levels in most areas. However, Nevada, Minnesota and Michigan are significantly cheaper today than in 2003. Except for those states, the cost I use should be within $2500 of the cost in 2011. These owner costs include mortgage, property taxes and utilities.

The web based paycheck calculator I used is PaycheckCity)

Thursday, July 28, 2011

California versus Texas: Part 3: Education

The education level of a state's workforce is strongly correlated with salaries. While I think that good high school education is important, it is the percentage of college graduates that seems to determine the value of the workforce.

Massachusetts, Connecticut and New York are the best educated states, and they have the highest wages. 38% of Massachusetts residents have a degree, and they have an average wage of  $53,700.

California has the sixth highest wages at $50730  and 30% of the workforce has a degree. Eight other states have both lower wages and a better educated workforce than California.

California does especially badly when looking at the percentage of the workforce who graduated high school. Only Texas and Mississippi do worse than California. Texas earns $42220 and 25% of the workforce has a degree. Mississippi has the lowest wages at $33930 and only 19% of the workforce has a degree.

I'm very worried for California. We have an expensive workforce with a doubtful quality of education. Guess which area has been heavily affected by budget cuts in the past few years? Schools and universities!!! Workers are unlikely to reduce their wages, so employers may leave the state if the educational level of the workforce falls further.

Texas is less well educated than California, but they have much lower wages to compensate for that.

How much does education boost the value of a workforce? In my model I use:

Labor force value ($/year) = 22845 + (775 * (Percentage of workforce with a college degree))

In my model boosting the number of people with a degree by 5% is worth about $4000 extra in mean annual wage, all else being equal. I haven't checked to see if this agrees with values published by economists.

(Educational attainment is 2008 data from the Census Bureau,  Wage data is for May 2010 from the Bureau of Labor Statistics)

California versus Texas: Part 2: The power of right-to-work legislation

For those unfamiliar with them, right-to-work laws greatly reduce the power of unions. Below I have the same chart as in the first post. I indicate the right-to-work states by coloring their bars red.

There is a very strong correlation between right-to-work laws and job creation. There are a couple of southern states that have lost jobs despite being right-to-work, but those states have major handicaps associated with poorly educated and rural populations.

I think that companies really do not want to have to deal with strong unions, and that right-to -work laws enhance the value of the workforce to the employer. My job creation model uses a 17% enhancement in value, which is equivalent to workers being willing to work for $6000 less than they actually receive in pay.

This is not due to people having to work for less. Given two workforces of equal educational levels and equal pay, the workforce in the right-to-work state is worth a lot more to an employer.

Correlation or Causation?

It is possible I suppose that right-to-work laws are correlated with something else which is responsible for the job growth. However, there is plenty of qualitative evidence that companies favor right-to-work states for expansion. Hyundai built their factories in Alabama. Toyota went for San Antonio for their new pick-up truck plant. BMW went to South Carolina and Mercedes went to Alabama.

Seattle based aircraft maker Boeing is going to great lengths to get away from their unionized, Washington state workforce. Despite having enviable pay and perks, this workforce goes on strike every few years for more money. Management and the union hate each other.

On their new 787 airliner, Boeing outsourced most of the production to factories in Italy, Japan and South Carolina.  South Carolina is a right-to-work state. Despite trouble recruiting qualified workers in South Carolina, Boeing is going ahead with a major expansion of that plant. They are putting a final assembly line there, which will allow them to deliver 787s without any of the parts passing through Seattle.
Boeing hated their union so much, they built this plane to fly wings and fuselages in from elsewhere!!!

Does right-to-work create jobs or just move them around?
It is certainly possible that right-to-work states are just stealing jobs from union friendly states. However, as the 787 project shows, jobs that leave union-friendly states can always go overseas. I think that both job creation and job shifting are factors in the success of right-to-work states. 

(Blog note: I'm on Twitter as Schrodinger333 and will try to remember to send out a Tweet whenever I put up a new blog post)

Tuesday, July 26, 2011

California versus Texas: Why some states grow jobs and others don't

There's been a lot of talk recently about the success of Texas in creating jobs while the rest of the country stagnates. The Federal Reserve Bank of Dallas recently claimed that 37% of all net new jobs created in the last two years in the US were in Texas. Some credit this to the leadership of Texas governor, and potential Presidential candidate, Rick Perry.

Meanwhile, California, once America's most dynamic state, has become an example of stagnation. Conservatives point to us as an example of the damage that can be done by environmentalism and high taxes.

Is Texas really as good as they say?

All the discussion about Texas so far has missed a rather important point. Texas creates a lot of jobs because it is big! To properly compare states, I have calculated the job growth per thousand residents over the past economic cycle for each of the 50 states. I'm interested in sustainable and long term job growth, so I am using an entire economic cycle, comparing the employment low of the 2001 recession with the employment low of the recession that began in 2008.

It turns out that Texas is not the best job creator in the country. That honor goes to Arizona, which generated 69 jobs per 1000 people who lived there in 2002. Janet Napolitano , take a bow! Amid the wreckage left behind by the housing collapse, it is easy to overlook the vast number of jobs created in Arizona over the past cycle. Even after losing many jobs in the recession, Arizona created more jobs over the period than any other state. Then comes Utah, where Jon Huntsman was governor. Then come Washington and Virginia. Texas is next, at 46 jobs per 100. So Texas is a good job creator, but other places have done better.

There is another side to this story. Rather a lot of states actually lost jobs over the past decade. California, Illinois, Massachusetts and New Jersey are some of the largest and wealthiest states in the country, but all are in slow economic decline. California lost 5 jobs per 1000 residents over the past cycle, even as the population grew.

The worst story is in the Midwest. Missouri, Ohio and Michigan experienced an economic meltdown. Michigan was the worst, losing 50 jobs per 1000 residents. That is as many jobs lost as Texas gained!

Why is this happening?

In a recent Room for Debate at the New York Times, Pia Orrenius from the Federal Reserve Bank of Dallas credited Texas's oil deposits, ports and low cost of living. Columnist Paul Krugman proposed three possible models on his blog. They involve either a reduction in wages, cheap housing or a surge in productivity due to the policies of Rick Perry.

I think it is really important right now to understand why some places create jobs and others don't. America's 50 states offer a kind of natural economic laboratory for understanding which job creation strategies are likely to work. I've been digging into the data, and have arrived at a model which up to a point can explain the differences in job creation by states over the last economic cycle. I think that most of differences in job growth between states over the past economic cycle can be determined from five variables. Getting to that formula will take me quite a few blog posts, so be patient!

(Technical notes: Job data from Department of Labor . Population data from the Census Bureau. Calculations use state populations as of 7/1/2002. I am measuring jobs from trough to trough. In many places that is 2002 to late 2009. In others it is 2003 to 2010