Catherine Rampell, the economics editor at nytimes.com, presents some graphs showing state and local public sector employment as a share of total employment.

Rampell notes her surprise that the graphs appear to show that more conservative states (i.e., those that are stereotyped as wanting smaller government) have seem to have a higher share of their work force employed in state and local government.

For example, the graphs show that Wyoming and Alaska have the highest share of their workforce employed in state and local government.  In both states, Obama drew less than 40 percent of the popular vote.

Rampell’s analysis of the graphs raises some potentially interesting questions.  The analysis also has some potentially serious flaws that highlight the pitfalls of statistical analysis.

Federal or state & local public sector employment?

Rampell’s graphs focus on state and local public sector employment. She then attempts to link state and local public sector employment to a national election.

One could argue that a U.S. President has greater control over federal employment than over state and local employment.  For that reason alone, one would expect to see one or more graphs showing the share of federal employment.

Static versus dynamic analysis

The New York Times graphs show employment for only a single point in time: 2009.

One reasonable hypothesis might link the changes in employment over time to voting patterns.  For example, a politician who promises more public sector employment may be rewarded with more votes.

The dynamics, however, raise a question of causation: Do changes in employment affect voting or do election results affect employment opportunities.  Do voters reward and punish politicians or do politicians “pay back” (or “pay off”) voters?

The answer to that question requires more than few pieces of information and a few hours of analysis.

The graph below shows that there in no measurable relationship between changes in state and local government’s share of employment and voting behavior in the last election.

The graph below suggests that there may be a relationship between changes in the federal government’s share of employment and voting behavior in the last election.

Much of this relationship appears to be driven by outliers.  However, these outliers raise some interesting questions.

For example, former Vice President Cheney’s home state of Wyoming had the lowest percentage voting for Obama.   Wyoming has also seen the steepest five year decline federal government employment as a share of total employment.

On the other end of the spectrum, President Obama’s home state of Hawaii had the highest percentage voting for him.  The Aloha State has also seen the steepest five year increase in federal employment as a share of total employment.

All this raises the question: Did the voters create the jobs or did the jobs bring out the voters? (Or are the two things unrelated?)

At 11 percent unemployment, Oregon is tied with Alabama for having the ninth highest unemployment in the U.S.  Some politicians and policy makers are cheering the fact that Oregon is not tied for first or second place, as it was a few months ago. Even so, Oregon has occupied a spot in the top ten highest unemployment states in 18 of the past 34 years.

Oregon’s chronic high employment has been a source of bafflement for many observers and economists.

Businesses note that Oregon has an anti-business attitude that treats business as a problem to be dealt with rather than an endeavor to foster.  In contrast, others point to surveys that rank Oregon as having one of the lowest business tax burdens in the country [1, 2] or being one of the most “business friendly” states in the country. In the face of these studies, Oregon’s persistent high employment rate suggests (1) Oregon is not employment friendly, and/or (2) the various tax burden and business friendly reports are fundamentally flawed and, therefore, meaningless.

Since so many Oregonian’s do not like to discuss the state’s business environment, observers have tried other explanations for Oregon’s moribund jobs environment, including:

1. Education.  If Oregon just spent more money on education, employment in the state would improve.
2. Climate.  Analysts at the Oregon Employment Department have a theory that states with milder climates have higher unemployment rates and (believe it or not) Oregon is considered to have a relatively mild climate.
3. High minimum wage.  Although minimum wage workers (and potential workers) make up a relatively small portion of the workforce, Oregon’s unemployment rate among those most likely to earn minimum wage is substantially higher than if Oregon’s minimum wage was the same as the Federal rate.

Ultimately, many observers, reporters, and politicians throw up their hands and blame manufacturing.  The story goes like this …

Oregon relies heavily on heavy manufacturing. Heavy manufacturing is highly cyclical: Employment soars during boom times and plummets during down times.  Thus, during recessions Oregon’s employment suffers worse than the rest of the country.  The story falls apart for several reasons:

1. Oregon’s unemployment rate is high even during boom times.  If the manufacturing story were true, during economic booms Oregon’s unemployment rate should drop faster and/or be lower than the rest of the country’s.
2. Oregon does not rely that heavily on heavy manufacturing.  According to the Oregon Employment Department, throughout the U.S. heavy manufacturing accounts for approximately 6.1 percent of employment.  In Oregon, it accounts for 8.3 percent. It not clear that this is enough of a difference to explain the state’s persistently high unemployment.
3. Other states that rely more heavily on heavy manufacturing do not have persistently high unemployment.  According to the Oregon Employment Department, Wisconsin, Iowa, and New Hampshire have a greater share of their employment in heavy manufacturing, yet these states have much lower unemployment than Oregon.  In fact, the Oregon Employment Department produced the following graph that concludes that “there seem to be other factors that have a stronger correlation to the unemployment rate than the concentration of durable goods employment.”

Minimum wage increases are a hot-button issue in many states. As such, minimum wage increases are politically challenging to implement. To avoid the knock-down/drag-out fights associated with minimum wage increase, several states—including Oregon and Washington—have introduced minimum wage indexing. With indexing, the minimum wage increases automatically each year based on some measure of inflation. As a result, Oregon and Washington have some of the highest minimum wage rates in the country.

Now that Oregon’s economy is in a tailspin, with record unemployment and business closures, the legislature is considering HB 3053 that would halt increases in the minimum wage during an economic downturn.

A study by Eric Fruits for the Employment Policies Institute measures the effect of minimum wage indexing on employment and wages in Oregon and Washington.  The study finds that minimum wage indexing imposes employment costs with no measurable income benefits. In particular:

• Higher minimum wages in Oregon and Washington are associated with reduced employment.
• Younger members of the labor force—age 25 and younger—are more likely to be adversely affected by increases in the minimum wage and minimum wage indexing. The figure above shows that Oregon and Washington would have significantly lower unemployment if the state minimum wage rates were equal to the lower Federal rate.
• Higher minimum wages have no statistically significant impact on wages of Oregon and Washington hourly wage earners.