How is the state of the United States Economy?

[pendragon]

Fair. There's an enormous asset bubble which the fundamentals aren't improving remotely fast enough to justify. If there isn't another stock market crash in the next two years, I'd be shocked.

That said, the fundamentals are improving, if slowly.

[pendragon]

The shift from a manufacturing-based economy to a services-based economy is a global one, and a trend as inexorable as the shift from agriculture to manufacturing (which is actually still ongoing - over any given period the US economy has lost a greater percentage of agriculture jobs than manufacturing).

Chinese manufacturing employment actually peaked in 1995 and has declined since then. Global manufacturing employment peaked in the mid-80s.  Pick any given country and the number of people employed in manufacturing peaked sometime between the mid-70s and the mid-90s.  More and more manufacturing can be automated, so fewer people are needed to produce enough manufactured goods to satisfy global demand.

It's just the facts - manual labor can be, in an ever-increasing set of circumstances, better performed by machines, while humans still have a decisive advantage in most fields of mental and/or social labor (although machines are also getting better at replacing humans in those fields too, just not as fast as they are getting better at replacing humans in agriculture or manufacturing).

Trying to work against the inexorable global economic forces at work here is like telling a one-armed grandpa to swim up Niagara Falls.  Manufacturing jobs are every bit as gone as agriculture jobs, and permanently.  They're not going to come back to China, even.  In fact, we wouldn't even want them to come back, because that would imply some sort of Mad Max apocalypse scenario.  So instead of whining about the lost manufacturing jobs, it might be better to consider how to best equip the workforce and increase our national competitiveness in the post-manufacturing services-based economy of the country, and the world. And, of course, there are specific policy changes that can be made to help employment growth in manufacturing sectors that are potential growth areas for employment in the US (most notably oil and natural gas-related).

[pendragon]

If I might explain what AggregateDemand is getting at:

How much more valuable is a M1A2 Abrams tank than a 1935 M2A2 tank? (My apologies for the US Army's repetitive armored vehicle naming system).

Most non-economists would think that's a difficult and rather subjective question to answer. But the Bureau of Labor Statistics would say that they can objectively calculate the exact numerical value of how much better an M1A2 is than its 1935 counterpart, down to the decimal places.

How do they do it? Well, they look at the last time the M1A2 and the M1A1 were sold at the same time, and divide the price of the M1A2 by the M1A1's price. If it came out to 1.1, the BLS would say that the M1A2 is exactly 1.1 times better than the M1A1. Then they divide the M1A1's price by the original M1's, and multiply 1.1 by the quotient (let's say 1.2, for a product of 1.3), and then do the same thing with the M1 and the M60A3, and so on and so forth until you get down to the 1935 M2A2.

Perhaps they might determine that, in "constant 1935 dollars," the Abrams is worth 98.3 times more than the M2A2, or 341.7 times more, or 154.4 times more. It doesn't really matter. In any event, we are likely to find that the "1935" price of the Abrams is more than a Yorktown-class aircraft carrier cost in the real 1935.

Now, if we were, today, to build an exact replica of a Yorktown-class aircraft carrier, it would cost far, far more than an Abrams costs. In fact, in "constant 1935 dollars," it would probably cost a whole lot more than it did to build the same ship in 1935.

Of course, on the other hand, in the real 1935, you might ask for a tank with a 1500 HP gas turbine engine, and depleted uranium mesh-reinforced composite armor, and that has a digital thermal targeting system, and is fully drive-by-wire, and can fire guided missiles from its main gun - but you're not going to get it. The good people of 1935 could build you 154 M2A2s, or even a million, but there is simply no way that they could build you an Abrams. From that perspective, the 1935 price of an Abrams is incalculably high.

On the other hand, an M2A2 nowadays would be militarily useless except for low-intensity crowd control, and even in that capacity its rather dismal fuel economy would doom it. It is, therefore, totally useless. A modern tank could just roll over and crush it, monster truck-style. If we generously say an M2A2 contains 11 tons of cast steel at the current scrap price of $460/ton, a $9 million Abrams is worth about the same as the scrap metal value of 1750 M2A2s. Add a zero or two to that number if we factor in transportation and disassembly costs, as we should.

And actually, against its interwar-era counterparts, the extreme capabilities of the Abrams are themselves complete overkill and therefore useless, and the extreme logistical nightmare it represents would make it simply not worth it to field in combat in 1935. If, one day in 1944, the Allies had found all their Shermans and T-34s replaced by Abramses, I suspect that most of us would have a better working understanding of the German language. To the army officer of 1935, the Abrams might as well be scrap metal, because the costs to use it would have vastly outweighed the benefits.

Or there are different ways of looking at things. You could say that an Abrams is worth billions of 1935 dollars in 1935, because the technologies that they could reverse-engineer from it would be worth billions. Or you could look at the 2014 value of the M2A2 based on what museums and private collectors are willing to spend on one as a historical curiosity. Or you could estimate what it might cost per unit to mass-produce an exact replica of the M2A2 today.

Any of these comparisons is as good as the other for determining the "constant dollar" value of those two tanks. And all are decidedly useless.

Instead of looking at "constant dollars," let us instead look at "constant tanks." In 1935, faced with prospective opponents like the Panzer I, T-26, and Ha-Go, the M2A2's military value to an American general was approximately one tank's worth. In 2014, faced with the T-72/90, Patton, and Type 98/99, the Abrams' value to a general is also about one tank.

How much expenditure in terms of materials and labor did an M2A2 represent in 1935? One tank's worth. The Abrams is much larger and involves more exotic materials and more skilled labor, so I'll grant that it's about 10 "constant 1935 tanks" in terms of materials and labor. How many tanks worth of materials and labor does a Yorktown-class aircraft carrier represent? Quite a few - now, in 1935, or at any other point in the past or future.

Now, that doesn't make the most sense in the world, but it makes a hell of a lot more sense than declaring that the expenditure to build an Abrams today is greater than the expenditure to build an aircraft carrier in 1935, or that Belgium spends more on its military today than Nazi Germany in 1943 or whatever. Arbitrarily declaring that tanks nowadays are 154.4 times more expensive than tanks 80 years ago is, obviously, going to make the current military budget look enormous compared to 80 years ago. It's more useful to look at how many soldiers, tanks, trucks, fighter and bomber planes, large and small ships, etc. the military is maintaining at any given time, or, failing that, the share of national output/GDP spent on the military is also a reasonable proxy.

The Bureau of Labor Statistics' process is known as "matched-model hedonic regression," and is one of the most unfortunate 20th century contributions to the economics profession, because it makes all historical (and therefore current) economic data into meaningless gobbledygook. Frankly, the Soviets were more on the right track with their endless statistics about such-and-such many more thousands of tons of pig iron being produced and an increase of 30,000 in the production of tractors and so on.

...

Quick factual caveat for this post: it's possible/likely that the comparisons between current and past military budgets are calculated using the hedonic regression in the general consumer price index rather than a matched-model hedonic regression for each individual weapons system. I'm just trying to illustrate the concept in a way that's easy to understand. And, of course, using that less rigorous method means that such comparisons tell us even less about relative military spending.