‘Provoking Accountability’…Of the ‘Unaccountable’
Or,
Defusing CATO’s “Precisely Guided Analytical Bomb”
Back to Part 3
 |
POGO's Banner Carrier, Suzie Dershowitz. (Source: POGO)
Hopefully, this is the LAST time we'll be seeing her, but don't bet on it. |
(To make each segment better as a ‘standalone’, I'm getting a little repetitive at the start of each post. If you’ve been following the segments all along, you may proceed to the subheading “CATO/POGO: Fuller Multiplier is an Outlier!” without missing the heart of the post.
Major Ploy Du Jour #2 (Continued):
“This Proves/Refutes Our POV/Their POV” (In this case both). Only.... it doesn't. Hint: You could lay all the economists in the world end to end and you still wouldn't be able to reach a conclusion.
As we noted in Part 3:
“…for this exercise we will focus on problematic areas of the CATO analysis (because it is not a study, but is merely an analysis that is critical of the Fuller study) where Ms. Dershowitz unwisely attempts to use in support of her assertion that ‘left, right, and center’ agree with POGO: that there will not be the kind of damage that the industry-sponsored study warns us will happen.”
As we also noted in Part 3, there are two aspects of the CATO analysis upon which POGO relied as evidence ‘supporting their position’. These perceived points of ‘support’, unfortunately for POGO do NOT provide the support they claim.
We exposed POGO/CATO’s Fatal Flaw #1 in Part 3 by illustrating CATOs critique of the Fuller study’s methodology amounted to no more than a pique over academic taste. The CATO complaint over the Fuller Study not going further than it was designed to do, and not doing anything other than that needed to achieve its intended explicit objective was rather sad. Claiming the Fuller study should have been burdened with additional layers of abstraction to answer a question different from the question asked, with the claim rationalized by the unsupported assertion that it would have been more meaningful, all without adequately describing the specific controls (ground rules and assumptions), modeling, or processes involved, seems to be rather self-serving caviling.
In closing Part 3 and in preparation for this segment, I asserted:
I believe the CATO players fully understand the notion of ‘usefulness’, and that it wasn’t enough to claim their study was more useful, when at best it is perhaps useful in a different way, and at the worst less than helpful through introduction of excessive ambiguity. What MAY explain why CATO went to great lengths to employ (and POGO parroted) the complaint that the Fuller study ‘didn’t go far enough’ is that they sought to use it as a smokescreen to cover their more ‘tangible’ complaint that Fuller overstates the net economic ‘multiplier’ of defense acquisition spending. Strip away the ‘Fuller didn’t take into consideration X,Y,and Z’ obfuscation and CATO’s weak attempt to ‘sell’ the idea that Fuller’s multiplier was outside some ‘economics mainstream’ not only looks even weaker, but intentionally contrived (more on that later).
By using CATO's own references, I will show how CATO not only does not show the Fuller multiplier lies outside some contrived norm, but they actually lend support to idea that the Fuller multiplier is possibly more appropriate than either the Economic Aesthetes at CATO or the Progressive Proles at POGO would have us believe.
Thus, on this adventure, we will cover Fatally-Flawed POGO/CATO Point #2.
2. Dershowitz/POGO relies on CATO claims (and claimed ‘evidence’) that the Fuller study overstates the net economic ‘multiplier’ of defense acquisition spending.
CATO/POGO claim: Fuller Multiplier for Defense Spending is an Outlier!
CATO asserts, and POGO’s Dershowitz promotes the notion that the Fuller study used an inflated ‘multiplier’ for arriving at the economic impact values for reduced defense spending.
CATO Analysis:
The Fuller analysis summarized above suggests a GDP multiplier effect of 1.92 for 2013 as a result of a $45 billion reduction in defense procurement. The modern scholarly literature on the GDP effect of government spending growth casts significant doubt on any multiplier effect of that magnitude, even under the assumption that the concept of a multiplier effect is consistent with sound economic analysis.
POGO’s Dershowitz claim: Fuller Is (Gasp) “Overblowing”!
POGO’s Dershowitz applies a blatantly biased spin on the CATO claims: Thus, it appears that Fuller is overblowing the impact of defense spending on GDP by almost twice as much as other estimates. The table in Zycher's analysis (reproduced below) provides a visual representation demonstrating just how out of sync Fuller's study is…
Ms. Dershowitz was kind enough to replicate the CATO table involved:
Author
|
Estimate
|
Notes
|
Fuller
|
1.92
|
defense procurement
|
Cogan et al.
|
0.65
|
large stimulus
|
Mountford and Uhlig
|
0.65
|
spending "shock"
|
Barro and Redlick
|
0.6-0.9
|
increases in defense spending
|
Ramey (2011)
|
0.6-0.8
|
defense spending after WW2
|
Hall
|
0.7-1.0
|
all government purchases
|
Parlow
|
0
|
defense spending
|
Ramey (2012)
|
0.5
|
all government spending
|
Let’s look at these comparative references just a little closer, to see “how out of sync Fuller's study is” shall we?
The Cogan, Et Al. Paper
The first reference we’ll look at is Cogan, Et Al. The title, “
New Keynesian versus Old Keynesian Government Spending Multipliers”, gives us a fairly good indicator the paper itself presents.
From the abstract of the paper, we see that the thrust of the analysis is not only just about “New vs. Old” Keynesian modeling, but also about how those models apply to estimating the effects of one of the recent, large, so-called “Economic Stimulus” packages (You know – the ‘stimulus’ packages…that weren’t):
Renewed interest in fiscal policy has increased the use of quantitative models to evaluate policy. Because of modelling uncertainty, it is essential that policy evaluations be robust to alternative assumptions. We find that models currently being used in practice to evaluate fiscal policy stimulus proposals are not robust. Government spending multipliers in an alternative empirically-estimated and widely-cited new Keynesian model are much smaller than in these old Keynesian models; the estimated stimulus is extremely small with GDP and employment effects only one-sixth as large and with private sector employment impacts likely to be even smaller.
If that isn’t bad enough, the ONLY place the .65 ‘multiplier’ figure of merit proffered by CATO and repeated by POGO has nothing to do with defense spending:
In any case, by assuming that the impact on consumption of the extra 1 percent discretionary increase in the deficit is .3 percent of GDP and using the above mentioned multiplier of .63 the impact will be to increase GDP by an additional .19 percent. If we add this to the .46 percent GDP increase from purchases, the total impact will be to increase GDP by.65 percent in the fourth quarter of 2010 compared to what it would otherwise be (P.17).
In fact, the words ‘defense’, military, or even ‘durable goods’
appear nowhere in the paper. It is a Keynesian modeling love-fest that lumps all government spending in one bucket from what I can tell. Interestingly, the focus is on the impact of adding new generic government spending with all sorts of multiplying effects versus estimating the negative impact (shock and long term) of suddenly withdrawing funds which would impact innumerable and ongoing economic activities. This point is even acknowledged in the text of the CATO paper (p.6):
”Cogan et al. estimate an effect of only about 0.65 in the quarter with the highest impact of a large government “stimulus” policy, which obviously differs from a change in defense spending alone”.
How typical of POGO to NOT mention that little quirk behind the entry in the nice shiny ‘table’ they hold up as an example.
BTW: If you like the back and forth of Keynesian economic model arguments (I don’t) you can always pull
this thread and see where it takes you.
The Mountford and Uhlig Paper
In the
Mountford and Uhlig reference CATO is again forthcoming in the text about what the .65 multiplier means (p.6):
“Mountford and Uhlig, employing a different type of economic model, arrive at a very similar finding of 0.65 in the first quarter of a spending shock financed with debt.”
More specifically, the Mountford and Uhlig source document reveals the .65 figure in a table, and in the note below that table we find:
“This table shows the present value multipliers for a deficit financed tax cut policy scenario and for a deficit spending fiscal policy scenario.”
The entire thrust of the Mountford and Uhlig study is to characterize the effects of deficit spending overall – NOT what is the economic impact of defense spending. At least the Mountford and Uhlig study indicates that ‘Defense’ spending was accounted for in the overall lump of government spending (p.26), but we have no way of knowing whether or not the multiplier of defense spending alone is above or below the ‘.65 average’ modeled for ALL spending.
The Barro and Redlick Paper
At least this
academic exercise in economics attempts to deal with the defense outlays, but by its own internal observations acknowledges several problematic aspects that call into question any direct comparison with the Fuller study. The two largest IMHO:
1. The authors state from their review it is their opinion that “It seems unlikely that there is enough information in the variations in defense outlays after 1954 to get an accurate reading on the defense-spending multiplier” (pg 4). Unless the CATO authors assert that there is no material differences between the economy of today and the US economy prior to the Korean War, which would be wrong, (
see my post on the 50th anniversary of Eisenhower’s now-mythic Military-Industrial Complex speech ) this would seem to automatically preclude relying on this study to evaluate the Fuller report, even if it agreed with the Fuller defense multiplier.
2. The author’s abstract contains the assertion: “For U.S. annual data that include WWII, the estimated multiplier for defense spending is 0.6-0.7 at the median unemployment rate. There is some evidence that this multiplier rises with the extent of economic slack and reaches 1.0 when the unemployment rate is around 12%.” Since the paper’s median unemployment rate is 5.57% (pg.18), and we are in fact experiencing much higher unemployment rates at the moment (
as high as 14.9% if you counted unemployment like we did back in the years this study analyzed) , the question becomes would the authors’ assertions that the defense multiplier would cap at ‘1’ hold in today’s economy?
Answer: We don’t know from the data presented.
The Ramey(2008)and Ramey(2011)Papers
All we (the Public) have to go on concerning Ramey’s 2008 paper is the
abstract which claims:
The implied government spending multipliers range from 0.6 to 1.1.
…and the CATO Analysis observation within the text which claims Ramey(2008) “finds
a defense multiplier effect of 0.6 to 0.8 for the period after World War II”.
That last bit is interesting. I wonder how that statement is reconciled with Barro and Redlick’s observation above—the point that being able to determine the multiplier after 1954 was ‘unlikely’?
By now, the reader should have gotten the idea that with models and the right ground rules, these studies will find whatever the authors want them to find.
The CATO analysis claims Ramey(2011) “finds a GDP multiplier from all government spending of about 0.5”. Well, we’ve already covered the point that ‘all government’ spending does not equal ‘defense spending’, so what is CATO’s point other than attempting to ‘pile on’?
In reviewing the papers we’ve covered to this point, I’ve even found reference to dominating influences of state and local government spending in the ‘all government’ spending equation after the middle of the last century (No, I’m not going to reread them all again to find it for a link). That aspect muddies the waters even more.
The Parlow (Ahem) Paper
When I first read
this paper my first thought was: WTFO? But the author apparently
isn’t a functional illiterate, but instead a German studying for his Econ PhD in the US. Apparently no one proofreads his work. Once you get past the poor grammar, you find little more than a ‘
this is how I used the models and made them work using fake expenditures and wars’ discussion.
The author claims that his analysis, using quarterly data vs. annual data, is a ‘better’ approach than the norm. Apparently the author is unfamiliar with how long it takes to invoke a change in defense acquisition or see the impacts thereof.
The author, as noted in the CATO/POGO table, claims ZERO economic effect from defense spending. There is no account provided for how the author actually handled the “all other things held constant” economic variables. But if I knew more about this paper, I would probably encourage the reader to also take a look: for it reeks of classic
GIGO. As it stands, I wouldn’t recommend it to anyone, and it makes me wonder if CATO was THAT desperate to come up with a longer table for their pagination and printing purposes rather than content.
With the claim of ZERO impact from defense spending (the only one in the list), what makes this Parlow paper NOT the outlier compared to all others in CATO’s eyes?
Finally, The Hall Paper
This is another one that the web provides
only the abstract on this side of firewalls [
bold emphasis mine]:
During World War II and the Korean War, real GDP grew by about half the amount of the increase in government purchases. With allowance for other factors holding back GDP growth during those wars, the multiplier linking government purchases to GDP may be in the range of 0.7 to 1.0, a range generally supported by research based on vector autoregressions that control for other determinants, but higher values are not ruled out. New Keynesian macro models have multipliers in that range as well. On the other hand, neoclassical models have a much lower multiplier, because they predict that consumption falls when purchases rise. The key features of a model that delivers a higher multiplier are (1) the decline in the markup ratio of price over cost that occurs in those models when output rises, and (2) the elastic response of employment to an increase in demand. These features alone deliver a fairly high multiplier and they are complementary to another feature associated with Keynes, the linkage of consumption to current income. Multipliers are higher—perhaps around 1 .7—when the nominal interest rate is at its lower bound of zero, as it was during 2009.
In fairness, the CATO analysis mentions Hall’s reference to the multiplier moving to the higher end of the range as interest rates approach zero included (pgs 6-7). POGO’s Dershowitz conveniently omits that fact in her little hit-piece. If one would care to review the current rate at which the Government can borrow, one would find it much closer to zero than I think any other time I’ve seen in my adulthood (.25%--a quarter of one percent
at this time). Which brings us to observe, that even without necking down the dollars in Hall’s paper to just Defense spending, the Hall paper’s findings are not very far apart from the Fuller report’s multiplier.
By CATO’s own reference to Hall, they prove that while Fuller may be on one side of a range of modeled results, the Fuller Report is NOT an ‘outlier’ by any means.
Summary
The assortment of studies that the CATO authors rounded up and how they applied them reeks every bit of the sort of advocacy research that they attempt to claim is a fault of the Fuller report. For about the last decade and a half (or so), when I read this sort of advocacy masquerading as analysis I mentally file it under a category named for the first half of the title of a favorite paper:
A Precisely Guided Analytical Bomb.
