This is kaleberg's Typepad Profile.
Join Typepad and start following kaleberg's activity
Join Now!
Already a member? Sign In
kaleberg
Recent Activity
I can see how this merger could improve profits by eliminating the middleman and aligned the interests of health care providers and the insurer, but I don't see anything in it for the consumer.
1 reply
I browsed this one. It was definitely too long for me to do more than look at the pictures and read a few paragraphs here and there. The statistical stuff is way beyond me. They could be making elementary mistakes, but I'm at the kindergarten level here. At least they talk a good game. The big takeaway is that "nonroutine" work has been increasingly automated which isn't all that surprising. It started with blue collar jobs, but it has also hit white collar jobs. Careers in typing or drafting have been vanishing for a while now, but you knew that. It looks like a good analysis using old newspaper job listings to estimate the demand for different types of work. Hey, it's another data source, and it goes back to the good old days when jobs not only listed occupational qualifications, but also the sex of the person desired. Nowadays, that's reserved for the dating ads.
1 reply
"The problem is convincing modern macroeconomists that this is an actual model." Well put.
1 reply
The elite only act equitably when their very existence is threatened. That's why wage and working conditions improved after the Black Death, the Chartist movement and the Great Depression through the Cold War. Once the existential threat is removed, they rake in all the chips for themselves. In theory, democracy should serve as a continual existential threat, but it rarely does in practice.
1 reply
You get this with a lot of numerical techniques. Eigen-analysis is a popular way of extracting "structure" from a large dataset. You compute the cross correlations then find the eigenvalues and eigenvectors. Sort by the eigenvalues and throw out the low information ones with their eigenvectors, and, in theory, you've extracted the meat of the dataset. As with making sausages, it can be disheartening to take a close look at that meat. The eigenvectors form a basis, so each eigenvector is horribly uninformative. Most of the information in it is about it being orthogonal to its peers, not about what the data means. Take the eigenvectors in groups and you might find an interesting subspace, but its axes are meaningless. If you are still seeking meaning, you can rotate the eigenvectors to cover the subspace differently. The literature is full of algorithms to do this one way or another. They yield a new set of uninformative eigenvectors, but now your original set is less special and could just as well be an artifact that would appear even if the dataset lacked any structure at all. As I said. It can be disheartening. If you look at machine learning with its neural networks and nodes and connections and weightings, you wind up with something quite similar. There are your weightings, and they are no more informative than your eigenvectors were decades before. It's like the 1960s again, except with more computer power and bigger datasets. P.S. One can argue that meaning isn't important, but without meaning it is next to impossible to make an algorithm more robust or improve it coherently.
1 reply
You are correct. Unfortunately, the politicians (of one party) make claims about what economists say and now and then haul out a pet economist or two to support those claims. Then they win elections and make public policy. Yes, the politicians may be the villains, but it is economists that people see. Yes, it is unfair. It is like computer programmers. For a while, they were wonderful, almost like Santa Clauses bringing us new toys every year. Now they have screwed up an election, invaded our privacy, broken our computers, raised rents in San Francisco and so on. They have an image problem which has not been helped by certain black hat hackers and idiots writing embarrassing sociological screeds. Perhaps it is time for a concerted publicity effort like the successful one run in the 1970s and 1980s that featured genial economists explaining to children why they would do best in a brutal world of unbridled, unmediated competition. I know this kind of thing doesn't get one tenure or even a proper think tank job, but it is necessary. Maybe economists have to work with the producers of various public entertainments to change the way we think. There have been programs like this that have subtly encouraged families to use birth control and delay having fewer children. We could just quietly alter our entertainments to spread the word. Why shouldn't the Ozzie and Harriet remake point out that FHA loans made their home possible? Why shouldn't the All in the Family clone point out the value of a government supported union? Why don't those medical shows casually mention that the payer is Medicaid, Medicare or Obamacare? P.S. Maybe this https://www.ssa.gov/history/three65.html is too much, but some of the basic plot points could be worked into various productions.
1 reply
This is one of the reasons that people think economists are autistic. Anyone who has ever run a business thinks in terms of revenue and potential revenue, not taxes. It's sort of like watching a group discuss the human psyche without a single mention of things like love, friendship, responsibility, empathy and so on.
1 reply
Finding specific counterexamples has always worked for performing negative proofs. There is a long tradition of disproving a formula for finding prime numbers by finding a case in which the formula produces a composite. The four color theorem was proven by means of an experiment. A calculation was defined using certain graph manipulations. The success of the calculation proved the theorem. Mathematics has increasingly adopted an experimental approach. After it was observed that one method calculating pi converged certain digits more quickly than others, mathematicians were able to develop formulas for arbitrary digits of pi. One useful program used by a lot of mathematicians takes a number and sees if it can be expressed as an exact function of values like pi and e using simple operators. Pi over six keeps turning up in unexpected places. Experimentation is a great way of finding new mathematics. Recently, for example, mathematicians noticed that sequential primes tend not to have the last same digit. How odd. Now to figure out why.
1 reply
That "two heads are better than one" is called "coalition growth", if I remember correctly. It predicts the world's population growth much more accurately than a simple exponential model. https://www.maa.org/press/periodicals/loci/joma/world-population-growth-background-natural-and-coalition-models
1 reply
This is a lot like the 1930s. Interest rates were extremely low, maybe 0.8% for Treasury debt, but the stock market was rising. Firms were getting more profitable, and there was a lot of technological restructuring. The indicators didn't even look that bad. In 1936, they declared the depression to be over, but it was far from it. We are in a similar situation, but we did not restructure our economy. We have low interest rates, but no stimulus to actually grow the economy. If you can tolerate risk, it is possible to make good money speculating in symbolic assets like stocks, art work, or expensive real estate. If you can't tolerate risk, you pay to keep your money in the bank. The economy is much weaker than it seems from the numbers. Incomes have not been rising, so there is no point in investing. Instead, the money flows out of the economy and into the symbolic arena.
1 reply
I think you're right. I still found it entertaining, but there is no accounting for taste. I did like the API metaphor. That's what almost killed Boeing, thinking they could break up the business into its components with clear API like contracts with their suppliers. In the real world, stuff always leaks through the API. It might work for computer services, as at Amazon, but it doesn't work everywhere.
Toggle Commented Sep 5, 2017 on Links for 09-04-17 at Economist's View
1 reply
The gig economy has definitely redefined the labor market, but it is just an extension of the ongoing elimination of employees in exchange for "consultants". I recently read a post on "premium mediocre" at Ribbon Farm, but the most fascinating thing in the whole post was the diagram of the workforce pyramid near the top. The pyramid runs: - orange void of mount doom - too big to fail - crypto bourgeoisie - premium mediocre - APIs - Bernie Precariat | AIs | MAGA Precariat - here be robots It's quite insightful. An API is an application programming interface. The precariats and AIs are those who perform the services contracted for by smart phone apps. It's a good way to describe the dividing line, enforced in part by software implementation, not contract law or evil mustache twirling villains. The crypto bourgeoisie have real jobs with benefits. The author and so many of his friends are in the premium mediocre range without real jobs and, at best, hoping for their share of goodies at Panera, or as the post put it, the best bottle of wine at Olive Garden. The real premium stuff will remain beyond their reach save for a spell of luck, either that or bitcoin. https://www.ribbonfarm.com/2017/08/17/the-premium-mediocre-life-of-maya-millennial/
Toggle Commented Sep 5, 2017 on Links for 09-04-17 at Economist's View
1 reply
Yup, it's the Friedrich List playbook. You can download it at gutenberg.org. No nation on this planet has ever industrialized without a towering trade wall to keep out kaiju.
Toggle Commented Sep 5, 2017 on Links for 09-04-17 at Economist's View
1 reply
Mazel tov.(x 2)
Toggle Commented Sep 3, 2017 on They Grow Up at Economist's View
1 reply
The Vax mainframe had three big counts against it. The Vax was incrementally billed, and Americans especially hate this. Even if commuting by bus is cheaper, owning a car means a low marginal cost of transportation. For a few dollars more one has cheap new options for recreation and shopping. People who had compute intensive projects thought nothing of running things on a PC overnight. They didn't need to budget the appropriate Vax time. The Vax did not provide as pleasant a user experience. There were latency problems which could make word processing, designing and even running a spreadsheet awkward. The system was often crowded and sluggish at times during the day. The Vax did a very poor job with graphics, even if you had a graphics ready terminal. Even at peak performance, the Vax could only drive your terminal so well given UART technology at the time. Vax software was clunkier and harder to learn. If you wanted to produce a paper, you had to learn a formatting language like roff or scribe. If you just wanted to grovel over a few numbers and produce a simple graph or two to get a sense of things, you had to learn a statistics package. Vax software had a steep learning curve. Yes, it often paid off with more power, but it was a burden, especially when you could just do it with a PC. The PC model of buying and having full use and control of one's resources was pervasive in the early 1980s. It was Diesel's dream of providing a prime mover for the working man. It was the small electric motor and Otto cycle engine that fulfilled Diesel's dream. If anything, the diesel engine is associated with centralized power.
1 reply
Houston, of course, is basically the safer, inland version of Galveston, the predominant port before Houston built its canal and took over. It would be nice to think of Houston upgrading its zoning and construction laws and spending money on further drainage improvements, but that isn't the Texas way. Still, I expect some changes, and I expect the city to be rebuilt.
1 reply
When I was an undergraduate, not at Berkeley, the big rooms were taken by the introductory calculus, physics, organic chemistry and one or two other introductory courses. Computer science seems to reign at Berkeley. Either that or the other departments have deprecated the lecture format.
1 reply
Some years back, at Argonne Labs, I saw rather charming presentation on optimization rat holes, that is, the places that one spends one's time on trying to optimize one's code. I mainly remember the first slide which was a diagram showing the various rats in their rat holes. Everyone present nodded in agreement. We had all spent some time underground among the rodents. There is something hypnotic about optimization, even when it is not needed. It's probably why code bases are factored and refactored and then multiplied by prime numbers. Since it is unlikely that the mathematics of multi-sided dice are going to change much, only the first run and perhaps a verification run of an independent implementation are important. Yes, the algorithm to prove the four color theorem can be run in seconds these days, not weeks, but it is unlikely to produce a novel answer.
1 reply
I keep thinking of the scene in the movie 'Working Girl' where Melanie Griffith finds her boyfriend naked with another woman, also naked, in a rather obvious sexual position. His first response was "It's not what it looks like."
1 reply
One would expect people with less education to be disabled more often. The jobs they can get are more likely to be body jobs, not brain jobs. Stephen Hawking can do physics in a wheelchair. Try cleaning hotel rooms or hauling two by fours without working arms and legs. One would expect states with a higher proportion of brain jobs to have lower disability rates. For every PhD analyzing genetic data there are two or three lab workers feeding the sequencers. There is a definite negative correlation between the percentage of the population with a bachelor's degree and the disability rate per capita. (If you ignore DC.)
1 reply
The main problem with this is that the same people would be controlling the availability of the productive capital and operations using that capital. There wouldn't be an adversarial relationship so it would be even easier to enter a state of corrupt crony capitalism than it is now. I'd stick with the multiple agent problem, but raise taxes so that government is a true partner in the performance of every company. That was what Berle & Means proposed back in the 1930s, and it worked fairly well.
1 reply
If that two trillion dollars in profits were repatriated, it would show up as higher real estate prices in midtown Manhattan and a higher PE ratio on the stock exchanges. Maybe a few percent of it would show up in paychecks given to real estate and stock brokers. Paying workers is a function of demand, i.e. how much workers are being paid. It has nothing to do with profits held overseas. That money is outside the economy of goods and services whether it is held here or overseas.
1 reply
Zero sum logic is ridiculous. The population wouldn't be able to increase if for every new person born someone else had to die as a result. We couldn't have had the agricultural revolution, let alone the industrial revolution.
1 reply
Graydon: A lot depends on the economics of Houston and its alternatives. Cities are created and maintained by economic forces. Look at New York City. It faces a similar problem with an increasing potential for flooding. When I was a child, the subway was never closed by floods. Now it is a common problem. Most of the wetlands of Manhattan were drained long ago and the low areas of the city raised repeatedly. Rivers still run beneath the streets and now and then damage an apartment building. Since Sandy, there has been some work on improving infrastructure, though a lot more needs to be done in the face of rising seas. The alternative would be to downsize the city and disperse its economic function. Unlike Houston, New York City's economic function is not about refineries or the locations of resources. It is about connections, amenities, and a work force, so it is highly mobile. When the World Trade Center was destroyed in the heart of the financial district, it took months to rewire all the communications involved. Why was it rebuilt in place? The city is still a terrorist target. I can see Houston not being rebuilt at scale if there were a collapse or massive restructuring of the petroleum and related chemical industries. Look at Detroit. A combination of automation and an improved national transportation network destroyed the economic forces that created it. Detroit had been sited on the Great Lakes for a lower delivered cost of steel. That downsizing could happen to Houston or New York City, but it would require something more powerful than a hurricane.
1 reply
Stross tends to be a bit apocalyptic. This is an occupational hazard. He writes science fiction for a living. Back in October, he was predicting imminent famine in England as a result of Brexit. (http://www.antipope.org/charlie/blog-static/2016/10/facts-of-life-and-death.html) He hasn't been proven wrong yet, but I suspect he is overly alarmed. Still, a good famine makes for a more salable novel than a shortage of avocado toast. A Keynesian might even look at a storm like this as a positive thing. It forces government spending whether funded by tax increases or higher debt. It is better than a military attack. Military spending has a lower multiplier than construction spending. Houston, for all its being located in Texas, was built by Texans who realized the vulnerability of Galveston to storm damage and were willing to pony up for a ship canal. I just don't see Houston as lost any more than London was lost after its great fire which happened a year after a visit by the plague. Of course, losing Houston would make for a much better novel.
1 reply