I would actually say it’s definitely better, if you’re stuck subsidizing someone’s survival, to subsidize them as a “freeloader”, aka: someone with actual leisure.
If you’re thinking that this is an incentive against the work-ethic, yes, it is. I believe our culture currently overemphasizes work-ethic, and this is all part of my sneaky evil plan to convince people to value work less.
You do realize the only reason society functions well enough to support the large welfare states you seem to like is because enough people still have vestiges of a work ethic?
You realize there exists a happy medium between 100-hour workweeks and zero-hour workweeks? “Vestiges of a work-ethic”? Actual, factual working hours have gotten longer among all full-time employees in most countries over the past several decades, and especially (almost everywhere) the past decade or so (even before the Great Recession, which would have a normal cyclical effect of lengthening work hours in a time of high unemployment).
And no, I don’t prefer a “large welfare state”. Please actually engage me in conversation rather than simply burning a strawman of “The dirty liberal”.
Interestingly, the data for Europe over the past decade actually refutes my hypothesis. As a secondary hypothesis that better fits that and other international data: hours and wages across countries appear to be equalizing as globalization occurs (again, this isn’t actually the first time). The richer countries have seen a rise in their full-time working hours, while the poorer countries have seen a reduction. It remains to be seen whether a longer timespan of data would support either of these hypotheses.
However, the effect of partition from a largely full-time economy into an economy divided between statistically distinct classes of full-time and part-time workers (the so-called “precariat”) does seem to be occurring.
the effect of partition from a largely full-time economy into an economy divided between statistically distinct classes of full-time and part-time workers (the so-called “precariat”) does seem to be occurring.
There are multiple factors in play here, but my impression is that mostly (but not entirely) this is a good thing.
Essentially people who have a valid (from the point of view of their finances, lifestyle, etc.) choice between working full-time and part-time are exercising their choice. It works from both ends—to take stereotypical examples, a contract programmer might reduce his workload to part-time because he earns enough money and values leisure more; and a stay-at-home mom might pick up a part-time job because she has enough time and energy for it, but not for a full-time job. This is good—it represents the availability of choice.
Anecdotally, I can attest that in the overemployed professions, some employers have started offering part-time hours at high hourly rates (or contracting/consulting jobs working part-year-round) as a fringe benefit to attract elite, high-skilled workers. Your contract worker is an example: I’ve done that one, earning a perfectly respectable monthly salary as a contract programmer but only working a few months at a time to get a frugal income quickly.
However, we do know that outside the conventionally-overemployed, high-hourly-rate professions, stay-at-home parenting has declined and part-timing has risen without necessarily being by choice.
I would venture to say that we should look for some numbers on hourly earnings (potentially split into full-time and part-time workers) to see what’s really going on. That sounds at least intuitively right, as I’ve known more than a few “highly-paid” scientists, programmers, lawyers, etc who end up with fairly moderate or even low hourly earnings once you account for their immense working hours.
That graph simply counts all workers, not all full-time workers. The effect I’m describing is not a unitary rise in overall hours but a partitioning of the economy into overemployed (rise in full-time hours) and underemployed (rise in part-time jobs). This matters, because a “total decline” in working hours that matches this model will be felt by the full-time worker as a heavy overtime load and by the part-time worker as a shortage of wages/hours.
Could it be related to the fact that the society becomes more complex and many jobs require more skill? That could explain why you have one overburdened worker and one unemployed person, and you still cannot give half of the former’s work to the latter.
As an obvious example, a company I work for is looking for more Java programmers. They even pay decently. There is a 14% unemployment in my country, but that’s not useful to my company, because those hundreds of thousands of people without jobs are not really good at Java or any other programming. In other areas the difference is less extreme, but it may still exist.
Could it be related to the fact that the society becomes more complex and many jobs require more skill?
A “skill-demand gap” is a very viable hypothesis. However, we should look for stiffly predictive variables. There are at least some claims that capital-biased public policy has created this situation rather than just the increase in the economy’s complexity.
Keep in mind, I’m not declaring a confidence in that position, but I can certainly call it a plausible claim that tax policies favoring finance capital, employer-tilted labor policies, and a trade policy of running huge deficits could hurt the employment market generally, or specifically lead to a situation in which the employment market bisects into a high-value highly-mechanized elite and a low-value mass of less-skilled laborers.
The one thing I am highly confident in is that the level of perceived skill or expertise needed to belong to the well-paid/overburdened “skilled elite” has risen.
You linked to a Reddit discussion about the article, where currently the “best” rated comment is: “I totally don’t get the logic here.” :D
I didn’t read the original paper, but from the article it seems to me their main argument is that under the hypothesis of skill-demand gap we would see the decrease of high-paying jobs and increase of low-paying jobs (which happened) accompanied with rising wages of the low-paying jobs (which didn’t happen, and therefore the hypothesis is false).
Let’s imagine that someone loses their high-paying job because it became too complex. For example, tomorrow the whole IT business will switch from Java to Lisp, but I am unable to learn Lisp, so I lose my job. So I become a retail clerk. Why exactly should I expect a higher wage than other retails clerks? If the same thing happened to many people, I would actually expect retail clerk wages decreasing, because now there is a greater labor supply in the industry.
Similar objections were already stated in the Reddit discussion.
So I become a retail clerk. Why exactly should I expect a higher wage than other retails clerks? If the same thing happened to many people, I would actually expect retail clerk wages decreasing, because now there is a greater labor supply in the industry.
At least in the classical economic theory, the idea is that you and other displaced Java programmers are more skilled than the average retail worker before the switch to LISP, and at least part of those skills are transferrable. Probably not the Java programming, but at least the math, job awareness, and other various generalist skills that are part of higher-profile jobs.
((This falls apart in extremes because many job skills don’t transfer or are even counter-useful, and employers have reasons to avoid hiring high-investment workers who have other better job offers available.))
I would actually say it’s definitely better, if you’re stuck subsidizing someone’s survival, to subsidize them as a “freeloader”, aka: someone with actual leisure.
If you’re thinking that this is an incentive against the work-ethic, yes, it is. I believe our culture currently overemphasizes work-ethic, and this is all part of my sneaky evil plan to convince people to value work less.
That’s one data point. What’s your guess as to what people that aren’t you think about this?
I wouldn’t venture to guess at other people’s thoughts. I’d rather just ask them.
I strongly agree. Just to be another data point :)
You do realize the only reason society functions well enough to support the large welfare states you seem to like is because enough people still have vestiges of a work ethic?
You realize there exists a happy medium between 100-hour workweeks and zero-hour workweeks? “Vestiges of a work-ethic”? Actual, factual working hours have gotten longer among all full-time employees in most countries over the past several decades, and especially (almost everywhere) the past decade or so (even before the Great Recession, which would have a normal cyclical effect of lengthening work hours in a time of high unemployment).
And no, I don’t prefer a “large welfare state”. Please actually engage me in conversation rather than simply burning a strawman of “The dirty liberal”.
Dear Lord, man.
That does not appear to be true. See e.g. this.
Interestingly, the data for Europe over the past decade actually refutes my hypothesis. As a secondary hypothesis that better fits that and other international data: hours and wages across countries appear to be equalizing as globalization occurs (again, this isn’t actually the first time). The richer countries have seen a rise in their full-time working hours, while the poorer countries have seen a reduction. It remains to be seen whether a longer timespan of data would support either of these hypotheses.
However, the effect of partition from a largely full-time economy into an economy divided between statistically distinct classes of full-time and part-time workers (the so-called “precariat”) does seem to be occurring.
My apologies.
There are multiple factors in play here, but my impression is that mostly (but not entirely) this is a good thing.
Essentially people who have a valid (from the point of view of their finances, lifestyle, etc.) choice between working full-time and part-time are exercising their choice. It works from both ends—to take stereotypical examples, a contract programmer might reduce his workload to part-time because he earns enough money and values leisure more; and a stay-at-home mom might pick up a part-time job because she has enough time and energy for it, but not for a full-time job. This is good—it represents the availability of choice.
This is a very complicated issue to unravel.
Anecdotally, I can attest that in the overemployed professions, some employers have started offering part-time hours at high hourly rates (or contracting/consulting jobs working part-year-round) as a fringe benefit to attract elite, high-skilled workers. Your contract worker is an example: I’ve done that one, earning a perfectly respectable monthly salary as a contract programmer but only working a few months at a time to get a frugal income quickly.
However, we do know that outside the conventionally-overemployed, high-hourly-rate professions, stay-at-home parenting has declined and part-timing has risen without necessarily being by choice.
I would venture to say that we should look for some numbers on hourly earnings (potentially split into full-time and part-time workers) to see what’s really going on. That sounds at least intuitively right, as I’ve known more than a few “highly-paid” scientists, programmers, lawyers, etc who end up with fairly moderate or even low hourly earnings once you account for their immense working hours.
That graph simply counts all workers, not all full-time workers. The effect I’m describing is not a unitary rise in overall hours but a partitioning of the economy into overemployed (rise in full-time hours) and underemployed (rise in part-time jobs). This matters, because a “total decline” in working hours that matches this model will be felt by the full-time worker as a heavy overtime load and by the part-time worker as a shortage of wages/hours.
Could it be related to the fact that the society becomes more complex and many jobs require more skill? That could explain why you have one overburdened worker and one unemployed person, and you still cannot give half of the former’s work to the latter.
As an obvious example, a company I work for is looking for more Java programmers. They even pay decently. There is a 14% unemployment in my country, but that’s not useful to my company, because those hundreds of thousands of people without jobs are not really good at Java or any other programming. In other areas the difference is less extreme, but it may still exist.
A “skill-demand gap” is a very viable hypothesis. However, we should look for stiffly predictive variables. There are at least some claims that capital-biased public policy has created this situation rather than just the increase in the economy’s complexity.
Keep in mind, I’m not declaring a confidence in that position, but I can certainly call it a plausible claim that tax policies favoring finance capital, employer-tilted labor policies, and a trade policy of running huge deficits could hurt the employment market generally, or specifically lead to a situation in which the employment market bisects into a high-value highly-mechanized elite and a low-value mass of less-skilled laborers.
The one thing I am highly confident in is that the level of perceived skill or expertise needed to belong to the well-paid/overburdened “skilled elite” has risen.
You linked to a Reddit discussion about the article, where currently the “best” rated comment is: “I totally don’t get the logic here.” :D
I didn’t read the original paper, but from the article it seems to me their main argument is that under the hypothesis of skill-demand gap we would see the decrease of high-paying jobs and increase of low-paying jobs (which happened) accompanied with rising wages of the low-paying jobs (which didn’t happen, and therefore the hypothesis is false).
Let’s imagine that someone loses their high-paying job because it became too complex. For example, tomorrow the whole IT business will switch from Java to Lisp, but I am unable to learn Lisp, so I lose my job. So I become a retail clerk. Why exactly should I expect a higher wage than other retails clerks? If the same thing happened to many people, I would actually expect retail clerk wages decreasing, because now there is a greater labor supply in the industry.
Similar objections were already stated in the Reddit discussion.
At least in the classical economic theory, the idea is that you and other displaced Java programmers are more skilled than the average retail worker before the switch to LISP, and at least part of those skills are transferrable. Probably not the Java programming, but at least the math, job awareness, and other various generalist skills that are part of higher-profile jobs.
((This falls apart in extremes because many job skills don’t transfer or are even counter-useful, and employers have reasons to avoid hiring high-investment workers who have other better job offers available.))
Yeah, since there was some controversy about it, I wanted to link to the controversy rather than just give one viewpoint alone.
I’m mostly trying to support the thesis that public policy has an impact, and that impact is worth investigating. The only thing I want to refute is technological determinism.
Show me the data.