11 February, 2014

The Mythology And Reality Of Unemployment Benefits

The Mythology And Reality Of Unemployment Benefits

Wake up to the reality of data.
Scott Lincicome
By 

At the State of the Union address, several members of Congress wore blue ribbons to signify their support for the extension of emergency unemployment insurance (UI) for American workers – a policy which expired in December after 6.5 years, billions of taxpayer dollars and, it must be added, stubbornly high unemployment.  In his big speech, President Obama also voiced support for this policy, and the Senate just last week held its third cloture vote on the issue, once again failing to get enough ayes to move the bill forward but providing yet another opportunity to show just how heartless and cruel the Republican Party can be (or something).
You see, supporters of extending emergency UI relentlessly argue that doing so is the right thing to do in this tepid economy with oodles of long-term unemployed Americans.  There just aren’t enough jobs out there, they aver, and eliminating jobless benefits will therefore do nothing other than push Americans out of the job market entirely – further marginalized from society and adding new strains to our welfare and disability rolls.  Such arguments have some foundation in the economics literature, so supporters of extending jobless benefits haven’t been entirely off base in pushing this line of argument.
However, as we get more and more data from my home state of North Carolina, which eliminated unemployment benefits in July and thus has served as a sort of “experiment” for what might happen to the nation in 2014, it’s probably time for these supporters to open their eyes and, at the very least, revisit their assumptions.  Because if North Carolina’s experience tells us anything, it’s that the elimination of unemployment benefits not only won’t spell disaster for the American economy, but just might give our stagnant labor market a little boost.  And, of course, save taxpayers billions of dollars (about $25 billion per year, to be exact) in the process.
It’s time for UI supporters to acknowledge these facts, but – if recent history holds – I seriously doubt that they will.
But before I get into all that, two quick lawyerly disclaimers.  (Sure, pretty much everyone in politics and the punditocracy has ignored these issues, but any serious analysis demands they’re stated them up front.)  First, as I’ve mentioned previously, it’s dangerous to draw broad and definitive conclusions from the experiences of a single state over a relatively short time period.  States are diverse, and isolating the effects of a single policy change and applying that to the nation as a whole is an extremely difficult task.  Nevertheless, North Carolina’s experience is unique and has become a hot-button issue given the national debate, so it’s appropriate to take a look here.  Second, the data themselves are less-than-ideal: the survey sizes are in some cases quite small and the results are subject to pretty large revisions.
But if we are, as so many have, going to use these data to draw broader conclusions, the least we can do is draw the right conclusions.  This, it seems, is far more difficult to do than it really should be: the last several months have, as I’ve already noted twice on these pages, been replete with half-baked, cherry-picked assertions as to what North Carolina’s little “experiment,” like, totally and definitely means for America.
One such asserter has been the New York Times’ resident Nobel Laureate and mean girl economist, Paul Krugman.  In fact, Krugman didn’t even wait for any data, regardless of its quality, before confidently reporting that North Carolina had declared “war on the unemployed” because it chose to cut unemployment benefits to a draconian 19 weeks (aka, about a month longer than the average duration of UI benefits before the Great Recession began in 2008) and to truly evil levels (aka, at or above most other states in the region).  Krugman’s reasoning, however, was based not on actual evidence in the Tarheel state or elsewhere, but on economic theory.  He nevertheless concluded, in typically-angry fashion, that “[t]he move to slash unemployment benefits, then, is counterproductive as well as cruel; it will swell the ranks of the unemployed even as it makes their lives ever more miserable.”
Well, ok then.
Since that time, however, we actually have real evidence of what happened in North Carolina after the state trimmed UI – six months of jobs data from the Department of Labor’s Bureau of Labor Statistics (BLS) establishment and population surveys.  And, unsurprisingly, countless politicians and pundits have butchered that data in order to confirm their pre-existing views about the effects of unemployment insurance on state and national labor markets.  One such butcher was none other than Paul Krugman, who in January took a quick look at a sliver of the data and confidently proclaimed that, naturally, he was right:
[I]f there were anything to the theory that cutting unemployment benefits encourages job search and somehow translates into higher employment even in a slump, harsh policies should work better at the state than at the national level. But there is no sign at all that North Carolina’s harshness has done anything except make the lives of the unemployed even more miserable.
I’ve already documented where Krugman’s analysis went awry, as well as the many experts who have in fact found “signs” that North Carolina’s jobless benefit cuts have actually helped heal the state’s wounded labor market.  Since then, several economistsbolstered these findings in a much-publicized study, which concluded from the available data in North Carolina that (i) unemployment benefits’ negative effect on the labor market “dominate any potential stimulative effect that some ascribe to such policies”; (ii) after the benefit cuts, “employment has risen according to all available sources of data”; and (iii) based on the data on hours, employment and wages, the new jobs were in no way inferior to others in the state.  These same economists, it should be noted, looked at another dataset last October and found that, because unemployment benefits deter job creation, “most of the persistent increase in unemployment during the Great Recession can be accounted for by the unprecedented extensions of unemployment benefit eligibility.”
So much for all that theory, eh?
Now, we have another month of data, and the evidence supporting these conclusions appears even stronger.  Immediately following BLS’ latest release of state-level employment figures, David Freddoso summarized the basic – and increasingly optimistic – facts about North Carolina’s job market:
The new payroll data show that North Carolina created another 11,100 jobs in December, which is 15 percent of what the nation as a whole added last month….

The data from BLS’s household survey of North Carolina (a different measurement, so note the different numbers) also looks at labor force and unemployment. It shows that about 19,000 of the 21,000 who left the ranks of the unemployed last month found new jobs. Only about 1,900 people dropped out of the labor force for whatever reason. This is better than in the past several months, but it’s consistent with the trend since benefits were cut — since July, fewer people have been dropping out of the labor force each month than were doing so before, and more people have been finding jobs. Together, these two phenomena have brought the state’s unemployment rate down rapidly….
The following charts, based on the new BLS data, show what Freddoso was talking about: after a tough start to the year, the Tarheel state over the summer began to experience improvements in both employment and labor force attrition:
Here you can plainly see that, in the first half of the year (i.e., before any 2013 GOP tax or labor policies had taken effect), North Carolina’s labor force collapsed and employment struggled.  Things only began to improve significantly in the latter half of the year – coincidentally (perhaps) after UI was cut.
This comparison – December to June versus June to December – gives us a look at what actually happened in North Carolina before and after the UI changes, and Freddoso deserves credit for adding that important perspective.  Indeed, too few “experts” – on the left and right – have taken the time to make the right comparisons.  Instead, we’ve been treated to 12-month (or “year-on-year”) comparisons (like Krugman’s) that hide the early year trauma and obfuscate any positive job or labor force moves that began in July (i.e., after the benefit cuts occurred).  Or we’ve been given only data from June to November or December – an analysis which targets any post-cut effects but ignores where the state was heading before those cuts took place.  Only the aforementioned study and has, to my knowledge, made this comparison in any rigorous way and, as noted, came to cautiously optimistic conclusions.
What these data still don’t tell us, however, is how North Carolina stacked up in 2013 versus its neighbors and the nation as a whole – an additional comparison that can tell us even more about whether the Tarheel state’s improvements really were unique and consequential or just par for the national/regional course (or even subpar).  The following charts (full dataset available here, if you’d like to check my work) undertake such a comparison using BLS data, and, as you’ll see, they provide even more reason for optimism: North Carolina’s labor market improvements stand alone in the 2013 periods before and after the state trimmed unemployment benefits.
First, data from the establishment survey show that, as a share of state/national population (used to compare apples to apples across the states and nation), North Carolina had the largest improvement in its employment situation from the first to the second half of 2013 (i.e., before and after the July change to unemployment benefits).
Between December and June, the Tarheel state gained only 13,100 jobs (0.13% of its population) – basically all of them in January, after which jobs stagnated.  By contrast, the state gained 51,400 jobs (0.53% of population) after UI benefits were cut – the second-best expansion over the period and, as already noted, the biggest improvement between the first and second-half of 2013.  In short, North Carolina’s job creation was stuck in neutral before the UI cuts and then shifted into drive thereafter.
The best second-half job growth was in South Carolina (25,400 new jobs or 0.54% of population) – a state which actually cut unemployment benefits a couple years ago – but that growth was a more modest improvement from the first part of the year (11,400 new jobs or 0.24% of the state’s population).  Virginia, on the other hand, saw its job growth decelerate between periods, as did the nation as a whole.  Thus, the reduction of UI benefits doesn’t appear to have hurt job creation in North Carolina and may have just helped it.  This is particularly apparent when you consider the difference between the job gains in the UI-cutting Carolinas and the job stagnation in Virginia, which maintained more ample UI benefits.
The population survey data show an even more radical job improvement in North Carolina, again not mirrored by neighboring states or the country as a whole and again showing much better employment growth for the “UI-cutters”.
In reality, North Carolina was the only one to actually lose jobs in the first half of 2013 and then gain jobs from July to December.  And the state didn’t just gain a few jobs after unemployment benefits were cut – it experienced the best job growth of the group (0.42% of its population or 41,364 jobs).  Only South Carolina can (again) touch this second-half performance, but – unlike its UI-cutting neighbor to the North – the state was already doing pretty well job-wise from December to June.  And (also again) Virginia and the nation actually saw 2013 job growth decelerate between the two periods.
In sum, both the population survey and the establishment survey show that North Carolina saw a significant improvement in jobs during the periods immediately preceding and following the state’s cuts to unemployment benefits – changes that are not only unique to North Carolina among its neighbors and the country as a whole, but also – and more importantly – much better.  Thus, although it’s impossible to be totally sure that North Carolina’s jobless benefit changes led to more jobs being created in the state (aka “employment effects”), these data do support that view.
Are any such employment effects sufficient to fix the nation’s labor market woes?  Of course not.  But, if true, they might help a little and save taxpayers and businesses billions of dollars in the process – an important point seemingly lost in this never-ending debate.
Moreover, the downside to scaling back UI in North Carolina – people dropping out of the labor force – appears to be much smaller than originally predicted or maybe even non-existent.  Krugman’s and his pessimistic colleagues’ chief (only?) criticism of the state’s cuts to UI and subsequent labor market performance is that the cuts caused a horrible drop in labor force participation.  Leaving aside for the moment the fact that the White House now considers labor force attrition to be a good and “liberating” choice or something, and the fact that some of this attrition might be due to the “Constanza effect” (i.e., people pretending to search for work in order to receive benefits), the BLS data argue against Krugmanian pessimism: all three states, and the nation as a whole, saw their residents drop out of the labor force in 2013, but only North Carolina saw the rate of attrition improve between the first and second half of 2013.
As you can see, the state’s labor force lost 0.6% of its population between December and June, and then 0.54% between June to December – a small improvement that’s made even better when you consider that, as shown in the earlier charts, North Carolina’s labor force attrition consistently decelerated since the summer and almost came to a halt in December.  More importantly, however, was the fact that this small improvement was again unique: South Carolina, Virginia and the nation actually had a larger share of their respective populations drop out of the labor force between June and December than in the earlier December to June period.  North Carolina’s went the other way (albeit mildly).
While these data don’t resolve whether cuts to unemployment benefits have caused some North Carolinians to drop out of the state’s labor force in 2013, the numbers doobliterate the notion – advanced by Krugman and others – that the UI changes caused a major spike in such activity.  In fact, the data thus far indicate the opposite effect, especially when compared to South Carolina, Virginia and the nation as a whole.  Put simply, if the participation effects resulting from North Carolina’s cuts to UI really were so prominent and forceful, the state’s labor force attrition would’ve accelerated like its neighbors and the nation, not decelerated as it did.  (I guess you could try to argue that the state’s labor force declines would’ve decelerated even faster, but (a) no one’s doing that; and (b) it’s a pretty tough sell considering what happened elsewhere.)
Maybe these conclusions will change in the coming months as we get more data or things are revised, but it’s pretty darn difficult right now to confidently claim what Krugman and others are claiming.  That probably won’t stop them from doing so, of course, but it should.
Indeed, the latest BLS data have been out for a couple weeks now (the aforementioned economic studies even longer), yet the tune from most UI extension supporters has remained the same.  For example, after Friday’s new national jobs data, AEI’s Michael Strain stated, without even a shred of equivocation:
If you think the economy is slowing due to weak jobs numbers in December and January, then that is all the more reason to extend unemployment benefits. If you think the labor market is still stuck in neutral, then that is all the more reason to extend unemployment benefits. If you want the long-term unemployed to stay attached to the labor force until the economy picks back up and they can find jobs, then that’s all the more reason to extend unemployment benefits.
Never once did Strain stop to consider whether these conclusions really are so clear-cut.  Then there’s Krugman, who just took another cursory look at the latest data (totally ignoring the December-June swoon or the divergent results in neighboring states, of course) and again came to the conclusion that supported his original prediction:
There has been a sharp drop in the NC labor force, probably in large part because workers who could no longer get unemployment benefits — which require that you search actively for work — gave up on what they knew was a hopeless quest.

The point is that to the extent that there has been a distinctive drop in North Carolina’s measured unemployment rate, it has to do with reduced job search rather than increased employment.
Maybe Krugman’s use of “probably” is as close as he’ll ever get to actually admitting that his earlier confidence was a bit premature or misguided.  (His unwillingness to admit fault – even after disparaging his opponents’ intentions – is now pretty much a given.)  Krugman’s ultimate conclusions, however, belie such optimism.  He, like many others, remains convinced that North Carolina’s unemployment “experiment” is a disaster (and, of course, that the state’s cruel Republican government really does hate the unemployed), regardless of what the data actually say.
We can only hope that Congress is a little more open-minded.
The views expressed herein are Scott Lincicome’s alone and do not necessarily represent the views of his employer, White & Case, LLP.

No comments: