I was a little curious to see how seriously manufacturing employment had been hit in the USA so I did a little digging. Here's my first chart (all data from FRED II, St. Louis Federal Reserve):

The chart above compares capacity utilization in manufacturing to changes in the multifactor productivity (MFP) index. The MFP measures the output per combined labor and capital inputs; the 2nd derivative, i.e., the rate of increase, reflects the rate at which technology is being implemented to enhance output. As you may notice, sometimes technology "moves backwards" (1972-74, 88-92) because the prior technological base has suddenly become untenable (such as an increase in the price of fossil fuels, or the enhanced ability of non-US producers to compete).

MFP is a much more useful measure of the performance of the economy than labor productivity. An increase in labor productivity may well be the result of managers replacing workers with machinery (as, for example, in Western European countries where unemployment AND labor productivity growth is high). Also, please notice that the MFP index does not sink merely because the rate of utilization has fallen. MFP measures the productivity of utilized resources, as opposed to ALL resources.

Here's my next chart:

Observe this chart has two scales; the left scale is manufacturing employment in millions of workers.  The stunning thing about this scale is the devastating decline in manufacturing jobs.  The 1980-1983 recession, the worst since WW2, rolled back 18 years worth of job growth in manufacturing; the 5 years of GWB has rolled back 56 years worth. But notice there is little change in MFP growth (the 2001 downturn was a hiccup; post-2002 results for the rest of the economy suggest growth has contined its pre-2000 trend).  Unit labor costs, despite soaring medical insurance premiums, are only 2% above 1992 levels, so that is not a useful explanation.

The comparatively low wages prevalent in US manufacturing have led to comparatively lower substitution of capital for labor.  This is not necessarily a bad thing, since a larger human share of productivity means greater potential flexiblity, and a more even distribution of wages among different productive sectors.

However, a point has to be made: utilization in US industry has not declined; employment has.  Durable goods orders have not declined; employment has. In fact, measured by durable goods orders, US industry has been booming.  This, despite a huge trade deficit chiefly in durable goods.

Upshot: The economy is doing fine as long as you don't take human factors and dormant resources into account.

I really wish folks would deal with economic statistics as hints instead of rules. It's obvious we need some kind of guidance but it's just as obvious we only measure 

1. that which we know we must measure
2. that which we know how to measure
   

Here's a fairly straightforward explanation of TFP (MFP) posted at the [UK] National Statistics page:

Conceptually, total factor productivity growth reflects all the effects on output growth that cannot be ascribed to the inputs into production. The current analysis uses the value added in TFP estimates, implying the existence of a value-added function, which efficiently combines the primary inputs of labour and capital in a manner which is independent of intermediate inputs. The value added production function, which relates aggregate output Y(t) to capital services K(t) and labour input L(t), can be written
as:

Y(t) = A(t)f(L_t, K_t)

where A(t) is the level of total factor productivity (TFP) which in theory captures disembodied technical change.

["Accounting growth: capital, skills and output" (PDF), Lau & Vaze, 2002]

 I've been hunting for comparable statistics for other countries.  I'll get back to this ASAP.

Upshot: The economy is doing fine as long as you don't take human factors and dormant resources into account.

I ought to write a post on the problems with the expression, "the economy is doing fine," or "good for the economy," or "the economy grew by x percent last year." The gist of it would be the error of using "economy" as a synonym for "industrial system."  In the 19th century, economists distinguished between the two.

The "economy" is simply the space in which production and transactions take place.  An "industrial system" is a process of converting inputs into outputs, and using the outputs to secure possession of future inputs.  This means that there are multiple industrial systems within an economy, and they adhere to different rules from each other or from the economy.

All of this is stuff that is well-known to you. 

I think our problem is the growth of perverse industrial systems. 

The gist of it would be the error of using "economy" as a synonym for "industrial system."  In the 19th century, economists distinguished between the two.

Well, I only started paying attention a couple of years ago, and at a less than professional level at that. I can see it being a useful distinction, though. 

I think our problem is the growth of perverse industrial systems. 

Yes. Not perverse as in "Grand Theft Auto nudity," but perverse in that their entire reason to exist is to generate cash flow rather than value/wealth. Companies that buy and exploit patents, for instance, add...massively...to GNP without creating a bit of value.

That's a problem in the economy, but it's a social problem because folks engage in social engineering to optimize the statistics. I'd prefer the industries die for lack of support.