Sunday, April 24, 2016

A Frack in the Pan

With the Pennsylvania Senate race, and a lot of Pennsylvania politics in general, still revolving around the question of whether fracking should be promoted, banned, encouraged or tolerated, a more fundamental economic question is left untouched. What is the current state of fracking in Pennsylvania? The market may have more than determined what politics implies is still an active question. The news has belatedly caught up to the decline in fracking across Pennsylvania, but the meme seems to be it's all a temporary setback. Everyone seems to be skipping over some more structural changes taking place in the industry.  There may not be much need to debate the future of fracking much longer, at least in Pennsylvania.

Will fracking come back is quite a hypothetical for the long run. Closer to the here and now is a much starker reality. The issue is no longer whether fracking related economic growth (and the jobs that go along with it) has reversed, but by how much and how far and what are the 2nd order effects of the rapid boom and bust. My question that goes a bit further. In the long run, will the development of fracking across Pennsylvania lead to net job destruction in the long run? Yes, the question isn't how many jobs will be created by fracking. Will fracking lead to a net loss of jobs.

Heresy? Certainly for an economic boom most claimed would last, at a minimum, for decades if not longer. Remember the, long since overcome by events, economic impact studies of Marcellus Shale in Pennsylvania. These studies were promised to be repeated annually so we could track the changes taking place. Yet, after just a couple of these studies, and long before the projected gains were ever realized in actual data, the studies were mysteriously put out to pasture, never to be repeated.

Debates over the scale of gains is one thing, but a net job loss is another matter altogether? Not just a wild projection of mine. Here is a graphic I admit I contrived, but gets to the core trends in shale development in Pennsylvania. Employment trends for Pennsylvania are reported for the Mining and Logging industries. Granted logging exists in Pennsylvania, but the variations in tat time series is all about natural gas and shale development in particular. Southwestern Pennsylvania, to include the Pittsburgh MSA, has had a longer boom in shale development for a number of reasons. The newer development of Utica Shale centered in eastern Ohio has provided additional gains. In the core of Pennsylvania where the Marcellus Shale was first developed at scale, the trend is pretty stark. My annotated graph below is of mining and logging employment in Pennsylvania net of the trends for the Pittsburgh MSA. Take a look:

Hard to look at that graphic and read the projections, or hope, the industry has of a turnaround by next year. Even if the industry stabilized, will the jobs come back?

The punch line. Mining and logging employment spiked by ~100% between 2005 and a peak in early 20112 The base of employment prior to 2005 includes mostly coal mining which is also included in this time series, but the changes since then have mostly been due to the trends in natural gas development. Explosively employment growth occurred from 2009 through 2012, but then went flat and since 2015have been falling precipitously. As of March 2016, the drop thus far means that over 78% of the gains between 2005 and 2013 have already evaporated. Here is the bigger observation of the data: the employment drop does not appear to be abating. If the drop continues at the rate it is going until July of this year, the state's employment in mining and logging will drop below the 2005 average.

Surprising? Not if you consider what the 2nd order impacts of the shale revolution, as it were, has had across the energy panoply. Coal in particular, no matter what you think about the impact of regulations, has seen its market share impacted directly by the rapid expansion of inexpensive natural gas.

An unaddressed news story is what happened to all the workforce development programs across the state that raced frantically to get workers into these shale jobs. Did all, or any, of that make sense and how many of those newly trained workers ever found employment in the industry, let alone how many remain employed.

For those expecting a turnaround, coal production can't expect much gains as long as natural gas remains anywhere near as cheap as it is now. As coal mines close, many will wind up being closed permanently.  Add into it news from the last week that Consol closed its decades old research activity in Pittsburgh, you have to speculate that much of the region's coal employment will be forever eviscerated as a result of new natural gas already in the (proverbial) pipeline.  In the end, add up the loss of virtuall all natural gas job gains, and ever more coal job losses, the overall size of employment in the energy complex in Pennsylvania could very conceivably wind up smaller than it was before the shale revolution began. The unavoidable law of unintended consequences strikes again.


Monday, April 18, 2016

Parsing Pennsylvania's Primary (or Nate Silver N'at)

Honestly this is mostly a post to prove I can still make a map.  But with the New York primary almost underway, all eyes will focus on the Pennsylvania primary. Much as they did 8 years ago; for at least a week that is.  Lots of talk of new voters registering this cycle, or voters switching registration to vote for one candidate or another.  How big a deal is either the party switching, or new registrations, in Pennsylvania?  The office of the Pennsylvania Secretary of State puts out lots of information to feed your inner Nate Silver.  Here is my take.

One topic I've seen alluded to is straight up party switching among people already registered.  This matters for Pennsylvania's closed primary. This is a map I concocted of the ratio of voters switching into the Republican and Democratic parties cumulatively since the beginning of 2016 (through April 3, and note the deadline to register for the primary election was March 28). 

It does give a fairly stark picture of switching into the Republican party.  Even Allegheny County has more folks switching to R than D, despite the large excess of registered D's there. 

That may be a bit misleading if inferred to be a new phenomenon.  Among already registered voters, party switching typical has this pattern or more switching to R than to D.  Yet how is it Pennsylvania has roughly million more registered Democrats than Republicans, a gap that is expanding over the years, not contracting?  The map above is only reflecting data of party switching among ready-registered voters who changed their party from something else to Democratic or from something else to Republican.   A similar map showing the ratio of new registrations is below. This depicts the ratio of new registrations as Democrats as percentage of total (note the total is among those registering as D or R, so not reflecting registrations other parties, or to no party.)

That gets closer to both the current political landscape of the state, and trends for the future.  Despite the first map showing a lot more re-registrations to Republican, the overall ratio of registrations over the same period is majority Democratic.  While there is a lot of Red in this map, the most populat counties in the state are pretty much all blue.

Make of it what you want.  Talk of Reagan Democrats and silent majorities I leave to the purely political pundits.  The total number of party switching into a major party (D or R, the first map) added up to under 75 thousand voters this year.  For a state with with over 8 million registered voters, you are really talking ~1%, much of which you would have expected no matter who was running this year.  Maybe this does not capture trends late in 2015, but I doubt it would make much of a difference.