Commenter Handle writes,
There’s nothing an entrepreneurial employer can buy to augment his workers to increase their labor productivity. They’re no equipment or anything for him to invest in. There’s no point: labor productivity in NCH sectors cannot be increased.
Well, computers can increase productivity in health care and education, but I get the point. It’s not like you can buy equipment that makes humans 10 or 20 times more productive. Barring strong AI, the only way to make humans much more productive in the NCH sectors is to dramatically re-organize the process, and that does not require a big investment in equipment.
Another trend is the drop in labor force participation. If people do not want to work, then you do not have to buy equipment for them to work with.
But suppose that you could tell stories about trends that imply low interest rates. At some point, low interest rates should make capital projects really attractive. That leads me to the suggestion that risky investments still fact high interest rates, and it is only government debt that enjoys low interest rates. And that leads me to the further suggestion (which I think I can find in Rogoff) that government debt is crowding out investment.
Without government debt to work with, financial intermediaries would have to satisfy the demand for safe, liquid assets by undertaking maturity transformation and risk pooling. That is what financial intermediaries do. But with so much government debt awash in the system, the public’s need for safe, liquid assets is satisfied without financial intermediaries having to do any sort of transformation at all.
Dramatically re-organizing a process frequently involves significant IT investment.
Homework assignment: go look at IXL.com
With all due respect, but neither you, nor I, nor any of your commenters knows how to radically improve NCH operations. Only a truly competitive market for these kinds of businesses (esp health care and education) can force people to start solving that problem. Sitting at our desks, even people like me who have run businesses, is not the way to figure this out. Top-down solutions are oxymorons.
People involved at the scene, day to day, and over a period of time will do it, but only if patients/students begin to pay the majority of the bills.
Doesn’t every entrepreneur think he knows? Aren’t some of them right?
” There’s no point: labor productivity in NCH sectors cannot be increased.”
That’s not true — and there’s no need for ‘strong AI’. For example, starting about 10 ago, I’ve done all my drivers license and license plate renewals online. That saved a great deal of my time and, in aggregate, saved an enormous number of DMV man-hours. Did this require a big investment in equipment? Not really — a web site that can handle roughly 10-20,000 simple shopping cart transactions a day isn’t exactly ‘big iron’. And in education, of course, there are enormous opportunities to increase productivity via the use of similarly inexpensive technology. Major reorganization would be required, yes, but not huge capital investments.
Read my whole original comment. I excluded functions which could be automated. You have to consider what services remain for employment that can’t be automated. And if they can’t be automated, they can’t be made much more efficient, because the way we make people more productive is to have them in charge of tasks where some of the things they were doing by hand are now automated.
I can easily foresee the entire university being ELIMINATED.
Does that count as automated?
So, the definition we are working with here is we have to keep the worker and do essentially the same work. Are there investment opportunities to increase the productivity? Outside of obvious expansion in a growth industry, this prospect seems really hard in every industry. I’m not sure how to tell it is that much harder in the NCHs.
“But with so much government debt awash in the system, the public’s need for safe, liquid assets is satisfied without financial intermediaries having to do any sort of transformation at all.”
Which is true to the extent that there is productivity sufficient to sustain (and accept) the taxation extractions necessary to provide a valuation basis; a problem being examined in Greece.
” If people do not **want** to work, . . . ”
A point to examine: How are needs, wants and desires met, let alone “satisfied,” in any particular economic structure? Is “work” necessary; must it be “productive” (responsive to needs, wants or desires of others) – at least to some degree?
Is it likely we are reaching a point where meeting the drain of the “carried interests” of the (for whatever cause or by whatever “policies”) non-productive reduces the “demand capacities” of the productive; with all the effects of diminishing increases in demand?
I’m still skeptical on your last paragraph, Arnold. Households and individuals have widely varying risk appetites, depending on factors like age, income, planned purchases, etc., and as a result, the market for financial instruments is too differentiated for a simple summary on the order of “the public wants safe, liquid assets and governments are supplying so much that our traditional financial intermediaries are being crowded out.” Consider the valuation of tech companies. In some cases, they’re approaching 1999 bubble levels, the way they’ve gotten so out of whack with regard to actual earnings. Doesn’t that, to some degree, imply that risky projects don’t really face that high an interest rate?
Maybe that’s true for education, but health care? There’s tons of machines and processes in health care. To me that seems to imply that there is room for improvement.
For example, a CAT or MRI scanner that made those processes as easy as taking an x-ray would probably improve productivity significantly. Similarly, for tests, improving the speed and efficacy of various blood tests would be a big improvement in productivity.
“That leads me to the suggestion that risky investments still fact high interest rates.”
Why do rates for corporate bonds tend to cling pretty closely to Treasury rates then?
This is confusing. My version of the problem statement of secular stagnation is “Why are growth rates and the cost of capital so low?” Not “Do interest rates still include a risk premium?” The question to me is about why are the underlying average real yields so low when the risk premiums are removed (or hedged).
And the cost of borrowing capital is at record lows not just for the government but, for example, for private corporations. Many – Apple, Google, and Microsoft are all in news lately – are now sitting on giant piles of liquid cash that they don’t seem to know how to allocate into other investments.
Let’s compare today’s 10-year bond yields (source: Yahoo)
Treasury: 2.07
Municipal AAA 1.87 (lower because of the tax benefit)
Corporate AAA 2.70
Corporate AA 2.87
Corporate A 3.04
Whoa, that’s pretty cheap corporate debt, even for only single-A’s. There’s a huge amount of demand for quality corporate paper right now, and Microsoft just sold $11B, including some 40-year obligations with only a 4% coupon. Tell that to a trader over 10 years ago and his head would have exploded.
“Well, computers can increase productivity in health care and education, but I get the point. It’s not like you can buy equipment that makes humans 10 or 20 times more productive. Barring strong AI, the only way to make humans much more productive in the NCH sectors is to dramatically re-organize the process, and that does not require a big investment in equipment.”
We’ve been hearing the hype about computers in health care and education for a long time, and there are computers everywhere, but the consequential labor productivity surge has been … elusive. Whatever is holding it back is pretty stubborn.
I know you hate models and equations, but if you promise to keep an open mind about it, I would be happy to build one for you illustrating the underlying mechanism. But, in English, here’s a brief effort.
A good amount of economic growth (real, per capita) is generated by technological innovation and capital accumulation combining to increase labor productivity: output per man-hour. A man with a forklift can do in an hour what twenty men lifting things by hand can do in a day. That kind of thing.
Entrepreneurs and inventors are always devoting a lot of attention to finding way to substitute capital for labor if it can be done profitably. But not all labor sectors are equally amenable to automation given the technological level, and so not all labor sectors are going to experience equal trends in productivity increase. That’s Baumol disease.
Some sectors that used to employ most people – commodities and manufacturing – were amenable to a dramatic and continuous improvement in automation substituting capital for labor, and dramatically increasing labor productivity. They have shed most of their workers.
Other sectors lag behind in improvement. Maybe they don’t have any good, existing technological way to do this, yet. Maybe they are held back by inertia or regulation or perhaps the fact that they are not really about the thing everyone says they are about, and are really selling something else. Or whatever else the stubborn factor might be.
Ok, so now imagine a simple economy where labor productivity is advancing quickly in one sector and slowly in another. Solve for equilibrium. What you’ll see if you model the dynamics (my offer stands) is the labor force begins to shift to the slower-productivity-growth sector, because the shrinking number of people in the higher-productivity-growth sector can still produce as much as everybody else can afford.
But the increasing concentration of the labor force in the slower-productivity-growth sector means that their aggregate output is also growing only very slowly, which determines their ability to consumer but also limits the opportunities for them to invest their capital.
That’s why we have a savings glut – it’s on the demand side for capital, not the supply side. The savings rate in the US really hasn’t moved all that much, but there are fewer profitable places to invest the same amount of savings, so to clear the market, rates fall in the competition for yield.
Again, the point is that normal labor market forces gradually reallocate the labor force from higher-labor-productivity-growth industries to lower-labor-productivity growth sectors.
The NCH sectors are exactly those places, which for whatever reason, we can observe very low rates of labor productivity growth over time, and the reason is because they are dependent on human factors which cannot be easily sped up to any significant degree.
But even if you could discover opportunities to speed them up through computers or reorganization, then you still run into two problems: (1) Is this a one-off improvement that still doesn’t allow for continuous high rates of growth over time, and (2) even if it does, then by the same logic explained above, all that will do is drive even more of the labor force into even more stagnant sectors.
If you don’t need human to do a job, the humans will do the jobs only humans can do, and they can only do those jobs at human speeds.
No it’s not and no it won’t. The price of investment is risk and that can be higher than any potential return, especially when the world is ageing and wanting less of it. Just because interest rates are zero doesn’t mean risk is also zero. Nor is there any incentive to take risks just because rates are zero. Cash becomes irresistible then in unlimited quantities. Why take any risk when potential rewards are so low and potential losses so great? There will always be those willing to borrow, but spendthrifts are always willing to borrow, but who is stupid enough to lend to them?
Merrill Lynch publishes junk bond interest rates — BB & CC rated bonds.
The spreads on the bonds relative to T-bonds are modestly higher this cycle than in the early 2000s but significantly lower than they were in the 1990s.
While I do not have volume data, this implies that higher risk investments are not facing unusually high interest rates.
It’s also useful to ask whether additional IT is really making things so much more productive, or whether many people experiences a quick improvement and then became saturated and couldn’t make much more use of more spending on IT
In a spin on “If you’re so smart, why aren’t you rich,” I’ll ask, “If IT makes people so much more productive, why isn’t there a much larger and faster growing demand for it, and related activities?”
So, consider Robin Hanson’s offer to bet.
Hard to excerpt as they say.
My version seems close. Technologies are compose PSSTs. I have excess cars already. Everyone that needs an internal combustion engine probably already has more than they can use. Everyone has more computer capacity than they know what to do with. We now give them to gradeschoolers. Probably not because they can effectively use them but because they seem deprived without something everyone else has. Peak stuff is also bumping against real or imagined environmental constraints. We have additionally added half the planet to globalization to handle our light work and copy all the low hanging fruit I’m their catchup growth.
So, give me a one handed economist. Is th migration to NCHs jobs a cause or an effect?
And another question for the one handed economist. Since the tech bubble, it seemed like the zeitgeist was that all the investment should be in outsourcing and globalization. Does this and other trends hide some of the numbers we are looking for in the data? I.e. it is so cheap to move production over there that interest rates are kept low. I don’t know how this effects productivity numbers.
Maybe we need to hit peak globalization before technology-based and division of labor based improvements can recommence.
The explanation for low productivity and/or stagnant productivity growth in the NCH sectors is that they are structured under a system of guaranteed payment for services rendered – completely irrespective of outcomes of those services. Virtually no private sector suppliers or service providers are guaranteed payment for their services/products. And even most Government suppliers (even defense contractors) are not guaranteed payment, completely irrespective of outcomes.
A guaranteed payments structure incentivizes increased production, but simultaneously de-incentivizes increased productivity.
BTW – Well done, Handle. Excellent explanations in your comments.