The Musical Chairs model in Ireland
Tyler Cowen links to a recent paper by AedÃn Doris, Donal O’Neill, and Olive Sweetman, studying wage flexibility in Ireland. This is from the paper:
The Irish case is particularly interesting because it has been one of the countries most affected by the crisis. We find a substantial degree of downward wage flexibility in Ireland in the pre-crisis period. Furthermore, we observe a significant change in wage dynamics since the crisis began; the proportion of workers receiving wage cuts more than trebled, rising from 17% in 2006 to 56% at the height of the crisis. Given the large number of workers receiving pay cuts it seems unlikely that wage rigidity played an important role in unemployment dynamics in Ireland over this period.
Tyler comments:
One question is what then caused so much Irish unemployment.
That one is easy—sticky nominal hourly wages combined with falling NGDP. The authors of the study could have saved themselves a lot of time by simply looking at the aggregate wage data (nominal hourly wages). Here are the 12-month rates of change, and also the change in NGDP over the same period:
Period ending Wage Growth NGDP Growth
2008:2 +3.6% -5.7%
2009:2 +2.3% -7.9%
2010:2 -2.4% -2.7%
2011:2 -0.9% +5.3%
2012:2 +1.0% +2.9%
2013:2 +0.5% -2.7%
2014:2 -0.6% +4.7%
2015:2 +0.9% +12.3%
[Warning: Eurostat is a nightmare to use, and I am a bit doubtful about the second quarter 2015 data–can anyone confirm?]
This fits the sticky wage model very well. Notice that NGDP plunged by 15.5% between 2007 and 2010. Wages actually rose over that three-year period. Unemployment soared, and indeed I’m surprised it didn’t soar even more, given the stickiness of wages. (Perhaps output fell the most sharply in capital-intensive manufacturing and construction?) Also notice there was a double dip in NGDP in 2012-13. And finally, notice that in both the original deep recession, and the later smaller double dip, the very small wage declines occurred with a long lag—just what the sticky wage model predicts.
Tyler continues:
A second question is why Ireland seems to have higher than normal nominal wage flexibility.
Could it be a greater than average willingness to endure living standard cuts without complaining? The Irish after all didn’t protest austerity as much as did most of the other Europeans in a comparable position. Maybe that means their wages can be cut without incurring the same morale costs.
Or could it have something to do with the “dual” nature of the Irish economy, namely that you either work for a multinational or you don’t? If you work for a multinational, maybe they can lower your wages and still you will work hard to keep that job.
Any takers on these questions?
There sure is a taker! This one is also easy to answer; Ireland doesn’t have higher than normal wage flexibility. If you look at any other country with big NGDP plunges (Portugal, Greece, Spain, Estonia, etc.), you’d also observe declining wages occurring with a lag after the big NGDP plunge.
And indeed this also occurs in the US. We saw huge falls in NGDP in 1920-21 and 1929-33, and 1937-38, and in all three cases we saw lots of wage reductions. Indeed in the case of 1920-21 the wage cuts were far steeper and more rapid than in Ireland, and hence the subsequent fall in unemployment was also much more rapid. Wage flexibility helps to stabilize an economy (contra Keynes/Krugman.)
But what about the recent recession in the US? OK, but NGDP fell by only 3% vs. 15.5% in Ireland. So naturally the slowdown in wage growth in the US was far smaller than in Ireland. Not enough to make it slightly negative, just less positive. If our NGDP had fallen by 15.5%, then nominal wages would certainly have also declined here. But just as certainly they would not have declined enough to prevent a big rise in unemployment.
The more I look at the data from different times and places, the more I like the sticky wage/NGDP shock model (AKA musical chairs model.) I think Tyler focuses too much on the fact that wages do eventually respond, and that when there are big NGDP shocks wages do fall in absolute terms. But the term “sticky wages” was created for the express purpose of distinguishing the model from “rigid wages.” Wages are not rigid. They change over time. But they adjust far to slowly to prevent big swings in unemployment. Indeed even in 1920-21, the poster child of wage flexibility, beloved by Austrians everywhere, the unemployment rate soared over a period of about a year, before falling rapidly as wages adjusted. Even then wages weren’t instantaneously flexible, and hence wage stickiness plus a huge NGDP decline caused a severe recession in 1921.
BTW, the authors finding that 56% of workers took pay cuts at the height of the crisis is exactly what you’d expect from the aggregate data, showing that aggregate hourly wages declined slightly in 2010. Looking at disaggregated wage data doesn’t really tell us anything important that we didn’t already know about Ireland. It’s represents another success of the sticky wage/NGDP shock model.
Tags:
27. October 2015 at 06:41
Excellent blogging.
Cutting wages is just not done. Forcing employees to work on the Sabbath is not done. There are social and cultural norms that sometimes trump economics.
27. October 2015 at 07:10
Non-excellent blogging. Sumner reminds one of Bush, who simply declares victory in a losing cause and goes home.
Sumner agrees with Cowen that wages are not sticky long term, acknowledges that wages lag over time, and this is the cause of unemployment, yet, Sumner ignores the authors statement about ‘pre-Crisis’ flexible wages, and only cites post-Crisis (i.e., 2008 onwards) data to make his case.
Worse, Sumner tries to correlate NGDP growth with wages. Wages are just one component of NGDP and an increasingly smaller one at that, with today’s robotic, capital intensive driven economy.
Worse still (it just gets worse and worse with Sumner), Sumner ignores this key passage in the paper: “the proportion of workers receiving earnings cuts more than trebled during the crisis. ” Trebled. That’s right, 3x more cuts during the crisis. Apparently Sumner does not understand that if an employer tells an employee to work half the hours they’re used to working, that’s a pay cut.
I guess Sumner feels if he writes obscurely enough, he can call night as day, and day as night, and his gullible followers won’t catch the big lie. But facts don’t lie. As the Irish paper authors say, wages are flexible in Ireland, during the crisis employees received three times more cuts (i.e., wages are flexible), and, while unemployment was high, Ireland bounced back and is stronger than ever (Austrian austerity is strong medicine that works).
27. October 2015 at 07:19
PS- Sumner writes (with a straight face?): “BTW, the authors finding that 56% of workers took pay cuts at the height of the crisis is exactly what you’d expect from the aggregate data, showing an aggregate hourly wage decline slightly in 2010.”
??? Sumner shows the majority of workers took pay cuts in the height of the crisis, and uses this as evidence of STICKY WAGES? R U serious Scott? I guess since not 100% of workers took pay cuts, this ‘proves’ sticky wages?
27. October 2015 at 07:38
Scott:
“Warning: Eurostat is a nightmare to use, and I am a bit doubtful about the second quarter 2015 data-can anyone confirm?”
The figures you quote are the non-seasonally adjusted (NSA) labor cost index (labor compensation per hour) for the most comprehensive measure of labor compensation (plus taxes, less subsidies) and the broadest measure of the economy available (NACE.R2 sectors B through S).
http://appsso.eurostat.ec.europa.eu/nui/show.do?query=BOOKMARK_DS-063207_QID_112FD3BC_UID_-3F171EB0&layout=TIME,C,X,0;GEO,L,Y,0;S_ADJ,L,Z,0;UNIT,L,Z,1;NACE_R2,L,Z,2;LCSTRUCT,L,Z,3;INDICATORS,C,Z,4;&zSelection=DS-063207NACE_R2,B-S;DS-063207S_ADJ,SWDA;DS-063207LCSTRUCT,D1_D4_MD5;DS-063207UNIT,PCH_SM;DS-063207INDICATORS,OBS_FLAG;&rankName1=LCSTRUCT_1_2_-1_2&rankName2=UNIT_1_2_-1_2&rankName3=INDICATORS_1_2_-1_2&rankName4=NACE-R2_1_2_-1_2&rankName5=S-ADJ_1_2_-1_2&rankName6=TIME_1_0_0_0&rankName7=GEO_1_2_0_1&sortC=ASC_-1_FIRST&rStp=&cStp=&rDCh=&cDCh=&rDM=true&cDM=true&footnes=false&empty=false&wai=false&time_mode=NONE&time_most_recent=false&lang=EN&cfo=%23%23%23%2C%23%23%23.%23%23%23
Although you are comparing quarter on quarter, one might quibble that the seasonally adjusted and adjusted by working days are more correct, since there are slight discrepancies mainly due to the working days adjustment. But essentially, for one who thinks Eurostat is a nightmare, in my opinion you have gotten this just about exactly right.
The NGDP figures look correct too.
P.S. Ireland’s NGDP is doing particularly well this year, no?
P.P.S. I love Eurostat. I also love AMECO and the ECB Data Warehouse. In my opinion they are all much easier to use than OECD Statistics, or than are most national and central bank statistical websites.
27. October 2015 at 08:12
Ray Lopez, really, really non-excellent commenting!
“Worse, Sumner tries to correlate NGDP growth with wages.”
No, he doesn’t. Since you don’t have a clue why he is showing those figures together, why are you even trying to comment on this.
“Sumner shows the majority of workers took pay cuts in the height of the crisis, and uses this as evidence of STICKY WAGES?”
Because they took small cuts, NOT SERIOUS COMMENTER!
“Worse still…”
Yes, here your commentary gets worse still. You apparently have no idea what numbers ought to be compared to test Sumner’s hypothesis, and will just throw together any two numbers with lots of punctuation marks and BIG LETTERS to show how right you are!!!!!
27. October 2015 at 09:32
The MCM didn´t breakdown even in Germany:
https://thefaintofheart.wordpress.com/2014/04/14/there-was-no-breakdown-in-the-musical-chairs-model/
27. October 2015 at 09:46
A couple points on the wages in the US from 2008 – 2010.
1) A lot of Benefits were cut heavily in this period so I still think there were compensation decreases during the period.
2) In most productive offices, companies dislike cutting wages because you lose all your highest marginal productivity people.
Of course, if wages were not sticky, then why haven’t we seen higher growth wages the last 2 years? (I believe the only reason real wages might have increased is fall in prices the last 12 months.) We are hearing all kinds of worker shortages.
27. October 2015 at 11:18
What evidence would disprove the “sticky wage” model?
BTW, collin, if wages are sticky going up then forcing inflation is a pretty crummy policy answer. For that means the “government” is raising prices all while wages remain stagnant! We used to call such an economy stagflation and having it was frowned upon. But that was then and this is now…
27. October 2015 at 11:36
Thanks Ben.
Ray, Where does one begin? Maybe with Gene’s comment.
Mark, No, I’m using year over year figures, so seasonality should not matter. Thanks for the link.
Gene, It’s a dirty job, and I’m glad you are doing it this time.
Marcus, Yes, it’s a robust model.
Collin, That’s Phillips Curve thinking, not allowed at this blog. Wage growth is slow because NGDP growth is slow. And because the labor market needs to get back to full employment.
Dan, Meet Ray.
27. October 2015 at 15:14
“That’s Phillips Curve thinking, not allowed at this blog”
-Wow! Since even Ray Lopez is allowed on this blog, this form of reasoning from a price change must be bad!
BTW, Scott, when did you stop responding to Major_Freedom? You should do the same with Ray.
27. October 2015 at 15:38
Also, dude, you got your book published by the Independent Institute? You must have had an amazing amount of persuasive power. Normally, I’d expect more goldbugs than MMs there.
27. October 2015 at 17:29
SSumner
Just wondering. Do you have any proposals or ideas to eliminate/reduce price stickiness?
27. October 2015 at 17:30
It would appear, at least on the surface, that public sector wages are more flexible in Ireland then other countries.
I would also wonder about the impact of unions on wage flexibility. At least in the short run.
27. October 2015 at 18:43
The data from Ireland is not inconsistent with the theory that wage cuts allow for more workers to find work which has the result of higher aggregate wage payments, which then causes NGDP to go up ceteris paribus.
NGDP changes do not cause wage changes. That is impossible. Wages logically and temporally precede wages.
Wages are paid to workers in order to produce the very goods and services of which NGDP covers.
Now this is not to say that wage changes cause NGDP changes, since it is possible for dividends, interest payments, and other non-wage incomes to “offset” any wage changes, and not only that, but cash preferences can change.
Callahan, you’re wrong by the way. Sumner was indeed correlating wages with NGDP. He is claiming that falling NGDP is correlated, with a lag, with falling wages, and that rising NGDP is correlated, with a lag, with rising wages. In other words he is claiming a lagged inverse correlation. And not only that, but he is claiming that the there is a causal relationship, namely, that the lagged inverse correlation is indicative of NGDP changes causing wage changes, but with a lag due to wages not immediately adjusting to NGDP.
As I said above, the data from Ireland, and all other countries by the way, is not inconsistent with the theory that wage cuts allows for more wage payments overall which is in turn one possible factor for rising NGDP, and vice versa.
Look again at the list Sumner posted for Ireland. The theory that wage rates go down, and then NGDP goes up, and that wage rates go up and then NGDP goes up, is identical in temporal structure to the theory that Summer is claiming. This is because the data is cyclical. You can say “down up down up” over and over again, and I can say “up down up down” over and over again, and while we started with totally opposite concepts, we’re claiming the same pattern overall.
Now the problem with Summer’s pattern view is that he has no good explanation, let alone AN explanation, for why NGDP would suddenly collapse, or, in other words, why cash preferences would rise so much all across the economy.
Firm revenues are interconnected. If one firm reduces its investment in capital goods, that will reduce the revenues for the capital goods sellers that supply that firm, ceteris paribus. The better question then is why firms suddenly reduced their investment that caused the reduction in capital goods revenues.
Why would consumption spending suddenly decline? Obviously we cannot reason from a spending change. We have to ask why so many people suddenly had a preference for more cash as a choice instead of buying consumer goods with that cash.
The central banks do not buy final goods and services (other than the spending by central bankers on their own selves) of which NGDP covers. The central banks “depend” on everyone else to spend almost all of the money they print. So the question is why didn’t people spend the money the central banks around the world printed with the same “velocity” before 2008 as they did after 2008.
Why did the capital intensive industries experience a greater reduction in employment relative to the service and retail goods sectors?
Sumner puts far too much focus on the aggregates. He is missing the forest for the trees.
27. October 2015 at 18:45
Typo:
Should have wrote:
Wages logically and temporally precede NGDP.
27. October 2015 at 18:47
Another typo:
The theory that wage rates go down, and then NGDP goes up, and that wage rates go up and then NGDP goes up,
Should be
The theory that wage rates go down, and then NGDP goes up, and that wage rates go up and then NGDP goes down,
27. October 2015 at 19:33
MF hits a home run, even if we disagree about money neutrality.
Sumner’s disciples urging him to ban me. The truth hurts.
Sumner endorsing Gene’s comment, which was nothing more than ad hominem. If wage growth and NGDP growth are not correlated, why is Sumner so keen on linking them Gene? Not a question, a comment for you to consider.
Macroeconomics is a untestable, unprovable exercise in metaphysics. And Sumner is the chief wizard of this sordid trade.
28. October 2015 at 03:15
Ray nothing you say is even internally consistent, so it can’t be true on its own merits. At least MF has an internally consistent ideology (well not really, but he tries harder than you do).
28. October 2015 at 04:22
Nice post, and I agree with you. (Though why do you call this the “musical chairs” model?)
However, I’m highly suspicious of aggregate wage data over the business cycle because of composition effects. While laying off overpaid workers sounds like a good way to cut costs, in fact firms tend to cut low-productivity/wage workers.
28. October 2015 at 06:40
Somewhat related;
http://www.voxeu.org/article/contribution-wage-structure-early-retirement-behaviour
———–quote———-
Results show that, indeed, steep age-wage profiles in a firm reduce the retirement age of its workers – a one standard deviation increase in the steepness of the wage gradient in a firm leads to an earlier job exit of approximately 6 months for blue-collar and 5 months for white-collar workers. Moreover, the prevalence of golden handshakes increases as well – these workers do not enter formal retirement any earlier, but rather bridge the gap until formal retirement with the receipt of unemployment benefits.
One interpretation of these results is that firms play an active role in the determination of their workers’ retirement age. Given individual retirement incentives – represented by detailed social security wealth calculations – a steeper wage gradient will stimulate firms to get rid of elderly workers prematurely; although, on the other hand, the workers would have an incentive to hold on to these good jobs even longer.
———–endquote———–
28. October 2015 at 07:08
“aggregate wage data (nominal hourly wages)”
Does this mean that you are looking at the wages of workers left with a job after employment cuts? This appears to be the case, which means you have totally missed the point of the paper.
The sticky wages model states that unemployment happens because either workers or firms fail to cut wages in the face of falling AD. Large scale renegotiation of wages combined with high UE anyway indicates that wage renegotiation was not enough to prevent UE, and is a point against the sticky wage model.
The increases that you would see in aggregate wage data (that is average hourly wages of remaining workers) is simply a skew. Take 3 workers, one making $11, one $10 and one $9 an hour. Recession hits causing 33% UE- which one gets fired? The one making $9 an hour (for a variety of reasons, but low earners are disproportionately hit by recessions). Average wages go up from $10 an hour to $10.50. Even if both the $11 and $10 an hour took paycuts you can still get wage “growth”- drop them to 10.90 and 9.90 an hour, you still get 4% average wage “growth” even though all employees took wage cuts.
28. October 2015 at 08:49
lets say that wages werent sticky, what unintended impacts would occur? would there be mass bankruptcies because workers could no longer pay rent,or loans? would any employee ever buy a high priced good (say houses etc)? why would they if they knew that their incomes were potentially unstable? would that lead to bankruptcies at many major companies because their sales would sink to almost 0? and the 64 billion $ question, how do we avoid these problems?
28. October 2015 at 09:19
Sumner argues that Irish wages fell by a small amount and therefore are not “flexible”. What about the fact the authors claim that Irish wages are less sticky than other European countries? If there is a lagged correlation between NGDP and wages, do you have thoughts on why the relationship would be different in different countries?
28. October 2015 at 09:30
Prof. Sumner,
This post is wonderful, wonderful, wonderful! However, I’m still not sure I understand why the current unemployment rate in so much higher in Spain (22.7%) than in Ireland (9.7%).
Hypothesis: since 2006, Ireland has adopted hugely-beneficial market-friendly supply-side reforms that Spain has not adopted.
Is that hypothesis somewhere in the right ballpark?
28. October 2015 at 13:11
@TravisV
Unlike Greece and Italy (and, apparently, Portugal), Spain has a generous support system for the unemployed.
28. October 2015 at 14:52
E. Harding, I stopped reading MF years ago, because he was dumb, rude and prolific. Any two are OK. But not all three. So why respond to Ray? Because it’s fun. Once he starts boring me I’ll begin to ignore him.
Cma, Ending the minimum wage and freeing up labor markets would help a little, but I’m afraid there is nothing we can do to significantly reduce wage stickiness.
Ray, If they were correlated (contemporaneously) it would undercut my sticky wage theory. Why did you assume I was trying to show they were correlated?
Jonathan, Agree about composition bias, but it doesn’t greatly affect my model.
The name musical chairs comes from the fact that if you suddenly remove 5% of NGDP (which is what is used to pay wages) then a certain number of workers will end up sitting on the floor, without jobs.
Bacon, You said;
“The sticky wages model states that unemployment happens because either workers or firms fail to cut wages in the face of falling AD. Large scale renegotiation of wages combined with high UE anyway indicates that wage renegotiation was not enough to prevent UE, and is a point against the sticky wage model.”
No, the model says wages don’t fall by a large enough amount, and soon enough, to maintain full employment. In the Irish case, wages fell far too little, and with too long a lag, to prevent mass unemployment. Now it’s certainly possible the sticky wage theory is wrong, but that sort of evidence doesn’t prove it.
dw, If wages were not sticky, then real GDP and hence real incomes would be more stable than otherwise. That means that workers would be willing to take more risks. Don’t confuse hourly wage rates with total income–sticky wages cause hours worked to fall in recessions.
Joy, It’s possible that Irish wages are more flexible than in other EU countries. Indeed perhaps that explains why Ireland recovered faster than the other PIIGS. But I would point out that wages also fell in the other PIIGS, so in a qualitative sense Ireland was not that unique. Obviously I haven’t studied the issue and I would never claim all countries have exactly equal wage flexibility–if I left that impression it was my mistake.
Travis, That may be part of it, and keep in mind that Spain also has a FAR higher natural rate than Ireland. Like almost 10% higher.
28. October 2015 at 16:15
While I don’t have access to the full distribution of wage changes in the Irish data, the statement that “… 56% of workers took pay cuts” is completely consistent with US data whether in a recession or not (eyeballing it looks like 40% of workers took pay cuts in 2006 and 2011 — of up to 20%!). There’s actually a lot of flexibility in the micro wage data which is why I think wages are micro flexible, but macro sticky.
http://informationtransfereconomics.blogspot.com/2015/04/micro-stickiness-versus-macro-stickiness.html
29. October 2015 at 08:22
Another excellent post, Scott!
29. October 2015 at 08:36
You appear to have the wrong paper linked in the post (your link takes me to aggregate demand shortfalls).
“No, the model says wages don’t fall by a large enough amount, and soon enough, to maintain full employment. In the Irish case, wages fell far too little, and with too long a lag”
From the paper
“First it is difficult to control for compositional changes in the workforce that have taken place during the crisis”
“In the pre-crisis period, the percentage of workers experiencing earnings cuts ranges from 17% to 23%. In contrast to the US and the UK, this proportion increased substantially during the crisis, reaching a high of 55% in 2009/10, compared to 23.6% in the UK and 37% in the US.”
From a basic sticky wages perspective one would expect that UE would be lower in the more flexible area (Ireland) than in the less flexible areas (UK and US). All else being equal Ireland having a substantially higher peak in UE is surprising. Obviously everything else isn’t equal, but your wage data is simply the wrong way to address the problem, and totally misses the point of the paper.
29. October 2015 at 09:55
Jason, I’d put it this way. Wages are relatively flexible at the micro level and relatively sticky at the macro level.
Thanks Brian.
Bacon, Thanks, I fixed the link.
You said:
“From a basic sticky wages perspective one would expect that UE would be lower in the more flexible area (Ireland) than in the less flexible areas (UK and US). All else being equal Ireland having a substantially higher peak in UE is surprising. Obviously everything else isn’t equal, but your wage data is simply the wrong way to address the problem, and totally misses the point of the paper.”
This is a very common mistake, which I’ve addressed in other posts. The sticky wage theory predicts bigger increases in unemployment in countries where wages fall faster. That counterintuitive result can be illustrated with a example. Suppose that in all countries wages are equally sticky. In all countries wages fall half as fast as NGDP falls. Then the bigger the fall in NGDP, the bigger the fall in wages, but also the bigger the gap between actual wages and equilibrium wages.
1. November 2015 at 12:04
Scott-
The sticky wages model is an attempt to explain why wages don’t fall to allow for employees to maintain their jobs in the face of a recession.
“Suppose that in all countries wages are equally sticky. In all countries wages fall half as fast as NGDP falls. Then the bigger the fall in NGDP, the bigger the fall in wages, but also the bigger the gap between actual wages and equilibrium wages.”
The point of this paper was to show that wages (by the measure of % of employees that took a pay cut in a given year) were more flexible in Ireland than in other countries both prior to and during the crisis.
You also failed to address the point that average nominal wages will run into composition problems if the unemployed are not a perfect cross section of the employed population.
1. November 2015 at 15:28
Bacon, There are many different sticky wage models, and they don’t all have the same implications for the micro features of wage adjustment.
You said:
“The point of this paper was to show that wages (by the measure of % of employees that took a pay cut in a given year) were more flexible in Ireland than in other countries both prior to and during the crisis.”
And how was this done? Simply showing that wages changed more frequently in Ireland than in another country does not imply that wages were more flexible. Is there any empirical data in the paper that undercuts the sticky wage model that I rely upon in this blog? If so, what is that data?
I am very aware of composition bias, I’ve published papers on real wage cyclicality. But it has no bearing on the sticky wage models that I focus on. In my model what matters is NGDP/W. Composition bias merely affects the extent to which a given change in NGDP impacts employment. But the underlying mechanism is still there, regardless of whether there is composition bias.
1. November 2015 at 18:17
Once again, wage payments are what finances (a portion of) NGDP.
NGDP temporally and logically follows wage payments.
NGDP does not change without there being a change in investment first.
The “musical chairs” model is flawed.
It is not the case that NGDP changes and then those changes affect employment given sticky wages. Wages change and then those changes affect (a portion of) NGDP.
NGDP fell during 2008-2009 because wage payments fell.
The question market monetarism does not answer, indeed is unable to answer, is why so many firms, in so many different industries, all around the same time, choose to hold cash rather than invest it (investment is itself a source of NGDP).
It is the real economy and market forces that cause the large swings in NGDP. MMs want the government to overrule the market.
1. November 2015 at 18:20
Ben J:
Where is there an internal contradiction in individual private property?
You say it is not internally consistent, but you don’t show where it is not internally consistent.
The irony is that all forms of statism, of which you are an advocate, since you are not a radical anarchist, that is in fact internally inconsistent. There are incompatible ethics, moralities, claims of efficiency, you name it, between government and the private sector.