On Sun, 09 Aug 2015 16:48:06 -0400, Joe Gwinn
wrote:
In article , Ed Huntress
wrote:
On Fri, 07 Aug 2015 19:45:43 -0400, Joe Gwinn
wrote:
In article , Ed Huntress
wrote:
On Fri, 07 Aug 2015 10:00:35 -0400, Joe Gwinn
wrote:
In article , Ed Huntress
wrote:
On Thu, 06 Aug 2015 09:21:27 -0400, Joe Gwinn
wrote:
In article , Ed Huntress
wrote:
On Wed, 5 Aug 2015 20:02:03 -0700 (PDT), "
wrote:
On Wednesday, August 5, 2015 at 5:16:16 PM UTC-4, Ed Huntress wrote:
That one looks more like raising the minumum wage caused
recessions.
Then you need to take a closer look at minimum wage increases and
GDP:
https://research.stlouisfed.org/fred2/graph/?graph_id=249246
There were two wage increases in '91. GDP climbed after the first,
and
then climbed again after the second.
There were wage increases in '96 and '97. GDP went up after each
one.
There were two wages increases in 2009. GDP climbed after each
one.
On the other side, there was a wage increase in 1990. GDP went
down.
You know what happened after 2007; wages weren't involved. Jobs
were
hardly involved. g
So, as I said earlier, it's mixed. But minimum wage increases were
followed by GDP growth more often than not over the last 35 years.
--
Ed Huntress
According to the chart there were recessions at about the same time
as
the
minimum wage increases. The grey bars indicate recessions where the
GDP
did
not grow as you claim.
Dan
The GDP grew exactly as I said. If you can't read the graph, I'll
pull
out the data for you.
Graph reading isn't the problem.
It's Dan's problem. d8-)
Well, I probably read the graph the same way as Dan, then - the claim
was that raising the minimum wage caused the GDP to increase, but the
graph says exactly the opposite.
No. The "claim" was that it's sometimes up, and sometimes down -- that
there is no correlation.
In fact, in the window shown, the biggest changes in both GDP and
minimum wage occurred during the 2008 recession, during which the
minimum wage went up as the GDP went down.
This is why I say you have to look at those graphs carefully. During
the period from the end of 2007 through the first quarter of 2009, the
correlations we
July 2007 -- min. goes up, GDP goes down.
July 2008 -- min. goes up again, GDP goes down.
July 2009 -- min. goes up again, GDP goes up.
So what is the correlation? It's the same pattern as 1975 and 1990:
Small increases in minimum wage correlate with a decline in GDP. The
highest level of min. wage in all three periods correlates with GROWTH
in GDP. Small increases bad; big increases good. g
The biggest move was in 2008, by far.
The thing is, if increases in the minimum wage caused a decline in
economic growth, the growth following an increase to an even HIGHER
level in 2009 would be impossible.
Or, more sensibly, the level of the minimum wage has nothing to do
with GDP, that one can measure.
There are a few ways to tell if a correlation could be due to
causation. The first is causal order - effect must follow purported
cause. The second is the the bigger the cause, the bigger the effect.
And the sign of the correlation coefficient must be all positive or all
negative.
Um...do you mean each correlation? If so, then no. What we're really
talking about here is covariance. Some can be negative while most are
positive, and you get a positive overall covariance and correlation
coefficient. Only if you assume that the "cause" is exclusive -- that
there is nothing else going on that would affect the relationship --
can you say that a cause must be all positive or negative, in every
case, to prove causation.
Even a strong positive correlation can have an occassional negative
value in the sample, due to other causative issues.
In real-world economics, there are few causative relationships that
show up with a purely positive or purely negative relationship. In
most cases there are multiple influences at work.
Where it gets really interesing, to me, is in those cases where a
small force drives a result in one direction, while a larger value of
the same force drives a result in the other direction. For example, an
improvement in one's ability to get family health care as an
independent insured (no group policy, no Medicare, etc.) has been
shown to improve the level of risk-taking and individual
entrepreneurship.
That doesn't seem to hold when safety-net measures grow large. A
massive unemployment system, such as the ones they have in much of
western Europe, does not seem to improve such risk-taking, and
actually hinders it, compared to countries with more modest safety
nets (particularly the US).
But even if these tests all pass, the correlation can be due to some
un-modeled common cause. It can also be difficult to decide which is
cause and which is effect if cause and effect happen simultaneously (to
the resolution of the measurements).
Right.
Of course, that's unlikely to have any causative relationship, but the
point is that the overall correlation coefficient is close to zero,
whether you look at it from the standpoint of minimum wage increases
(1997; Mar 1975 through May 1979) or the standpoint of recessions
(2001, 1970, 1960, etc.).
The correlation with recessions seems shot to hell by the fact that,
consistently, the highest raise in an annual series of raises occurs
roughly at the time the economy turns up.
Again, no correlation with any direction: there are ups, and there are
downs. The number of samples is too small to rely much on the
coefficient, but it's surely close to zero.
Oh, there is a correlation there, but it is more political than
economic, and doesn't happen every time. This isn't Physics.
The minimum wage changes only when Congress changes it. This happens
only when Congress is under considerable political pressure. Like
during a recession.
If you look at the pattern, you'll see that most of them are
multi-step raises that either start before the recession, or that had
to be signed into law before the recession.
I agree that it's political, but the pressure seems to occur when CPI
and employment are rising sharply:
https://research.stlouisfed.org/fred...0&category_id=
I suspect that CPI is what applies the political pressu Prices are
rising,and there is a clamor for higher wages. Then, at about the same
time, we fall off of the economic boom and go into a normal
business-cycle recession. That's happened several times since 1980.
Before then, it was more complicated, as our monetary responses were
really messy and uninformed. That's the "before Volcker" period.
Look at it from that perspective and see if that doesn't agree closely
with the evidence.
The Minimum Wage is not an economic variable like Supply and Demand.
Right.
I don't see how looking at the
numbers that are graphed would change this general observation.
sigh Let's go through it. There was a two-step increase ending in
1991. Just as the second minimum wage increase took effect, the theory
suggesting that the effect on GDP should be at its worst, the
recession ended and GDP took off.
In 1996 - 97, another two-step increase. GDP just keep rising, with NO
effect from the increased minimum wage.
2007 - 2009: Again, a multi-step increase, amidst a recession that was
precipitated by the financial collapse. Just before the end of it, the
increases reached their peak -- just as the economy turned around and
GDP began to grow again.
In all of those cases, the worst effect should have been when the
increase raised wages to their peak. But just the opposite happened.
Just as wages peaked, the recession ended and GDP started to grow.
Are we looking at the same graph?
https://research.stlouisfed.org/fred2/graph/?graph_id=249246
Yes. See part about correlation below.
The chart shows GDP reduction (not
increase) at the same time as increases in minimum wage. The cause is
of course political, not economics: Congress raises the minimum wage
only during hard times, to show that they are doing something, at last.
The subsequent rise is simply the general recovery from recession, and
is not a result of raising the minimum wage.
Then how do you explain raises in '96 and '97? There were no hard
times then. Minimums were raised twice and the GDP kept climbing.
Politics. I don't recall what was going on at the time, 20 years ago.
The mini-recession of 2001 seems to have had no political effect - no
change in minimum wage.
And if you move the slider to the left, you'll see a very complex
situation. There were so many variables at work during those years
that I didn't want to complicate the issue -- 35 years of modern
experience seemed to tell enough -- but you'll see, if you look back,
that in '74 and '75, the minimum was raised twice DURING a recession,
and the GDP started climbing just a couple of months later.
Correlation is not causation.
But how do you show causation if there is no correlation? The answer
is, you don't. A correlation doesn't show causation. But a lack of
correlation shows there is no causation.
If there is no correlation, causation is unlikely for sure.
If there is correlation, there is still much to be proven. The biggest
danger (aside from coincidence) is when the two timeseries showing
correlation are in fact both dependent on some other timeseries.
The classic example is the fact that there is a positive and
significant correlation between automobile accidents and the sale of
ice cream. This is true, even if you discount kids running into the
street at the sound of an ice cream truck.
So, to save lives, we must outlaw ice cream. It's the only moral
course of action.
Why the correlation? In the summer, people drive more and eat more ice
cream, because it's warm and nice out.
.https://www.google.com/search?q=fals...es&client=oper
a&hs=DlD&tbm=isch&tbo=u&source=univ&sa=X&ved=0CDg QsARqFQoTCKaB37iPmMcCFY
SUHgodezQGYw&biw=1216&bih=1271
So, based on your conclusion and example, what is the likelihood of
causation between the peak of each two- or three-year rollout of a
series of minimum wage increases, and its correlation with an economic
upturn and growth in GDP?
See above comment on the role of Congress.
Yeah, but the *effect* of the min. wage increase on the economy is
minimal or nonexistent. It may even be one of those "small safety net"
phenomena that have an overall beneficial effect. I suspect it is, but
proving it is beyond my analytical skills -- or time.
Through the late '70s, the minimum was raised three times, and GDP
just kept climbing right through it. No recession then.
If you attribute the GDP increases after periods of both minimum-wage
increases and simultaneous recession to the normal business cycle,
then you have to agree that the minimum-wage raises did not stop or
even slow the recovery. In recessions that were not accompanied by
minimum-wage increases (1982; 2001), the shape of the recovery was the
same as those with a minimum increase. Conservative theory would have
predicted at least a delay. In several cases, with business facing the
newer, higher minimums, growth just went on throughout the following
period.
If the conservative theorists were right, none of this would have
happened.
Not quite. See below.
The minimum wage in Mass is $9/hr, or $3/hr with at least $20 per month
of tips, as of 1 Jan 2015. It was $8/hr before.
In the Boston area, Dunkin Donuts is offering $10/hr plus tips
(according to a sign out front), and has been for some time now, long
before the recent increase. Now one would assume that Dunkin Donuts
knows from long experience exactly how much they must offer to attract
the desired number and quality of employees.
As long as the minimum wage is less than the market-clearing wage for
low-skill jobs, the minimum wage will have no effect.
But we know that the minimum is no more than the "market clearing"
rate for 3.3 million (or 3.6 million, more recently) people. And the
percentage, now at 4.7% of workers, has been much higher in the past.
It was over 17% in the late '60s, IIRC. So the historical examples are
based on much higher percentages. The pattern holds.
I find that conservative theorists have put themselves in a dilemma.
First they say that increasing the minimum will cause a substantial
loss of jobs and business bankruptcies, not to mention inflation, and
then they say that raising the wages will do nearly nothing because
the number who are making minimum is so small; that the
market-clearing rate is higher to begin with.
I fail to see the problem. Both statements are simultaneously true.
The problem is that the theories you guys are going on have no
evidence. You're whipping up ideas like cotton candy, and the
historical record shows that they're crap.
Which guys are you talking about? There is sufficient foolishness to
go around. Many times over.
The only theories are coming from those who say minimum-wage increases
cause loss of jobs, inflation, or economic downturns. There is no one
on the other side of this discussion except me, and I present only
evidence, not theories.
See the parable of the apples, below.
Consider apples. Let's say that apples are an unregulated market, and
the current market price for apples is $11 per pound for better apples,
but only $8 for lower-quality apples. The growers of low-quality
apples are complaining about the unfairness of it all, raising quite
the ruckus. The King eventually decrees that henceforth, apples may
not be sold for less than $9 per pound.
What effect will this have on the apple market? None on the better
apples (which have always complied), but significant on the
lower-quality apples (where sales volume will decrease, because not
everybody is willing to buy low-grade apples at $9 a pound).
Again, you have no evidence that this has happened with minimum wage
increases. In 1991, 1997, and 2009, just as wage increases reachd
their peak, employment either continued to climb steadily, or it
actually jumped.
A lot of random motion to be sure.
It's hardly random, but there is no sensible connection between
economic upturns or downturns and increases in the minimum wage. The
evidence says that the correlation is close to zero.
Do you deny the truth of the parable of the apples above? It's Econ
101. If yes, what is your reasoning?
Having spent most of my years since Econ 101 (and up through 300)
paying attention to evidence and facts rather than theory, I'd like to
see that theory in actual practice. The general tendency, as I've
learned in recent years, is for unintended and unexpected
consequences, with theorists scrambling after the fact to re-build
their models so they can better predict the past.
Hmm Econ 300 does not replace Econ 101. This is basic Supply and
Demand, which has not been overthrown, and never will be.
Today, a lot would depend on the marketing skills of the people
selling lower-quality apples. They may actually make more money,
despite a small decline in the volume sold.
This assumes that the other apple growers are all asleep at the switch.
To some degree, they usually are. That's why there are winners and
losers in the marketing game.
You're assuming perfect knowledge. As you know, the lack of it is the
major thorn in the side of basic free-market theory.
More generally, advertising tends to balance out - they drown each
other out.
There are hundreds of counter examples. There are many, many ways to
manipulate demand, some of them very successful.
This is a good example of why classical economic theories are suspect.
They may work on very large scales (although they may not), but you
never know if apple marketers in the next town are charging $10 for
all apples, and when their customes start seeing your ads and get wind
of the fact that you're selling apples for $9, you could start selling
like crazy.
Yes markets have always been imperfect for just the reasons you give
above, but the advent of modern communications (including but not
limited to the Internet) have greatly reduced such imperfections,
especially concerning commodities like apples and low-skill labor.
I don't agree about low-skill labor. Labor mobility is running very
low, not high, and communication advances are one of the reasons, say
the researchers.
http://www.economist.com/blogs/freee...abour-mobility
But we're really getting off-track here. I agree that wages are set by
supply and demand, but with a legislated floor below which they can't
go. Based on the history of min. wages, it doesn't seem to have had
any overall negative effects. If you need workers, you need them, and
having to pay them 50% more than you would like probably does not
change the fact. You either have a market and potential growth in
sales, or you do not. Even taxes have been shown to have a minimal
effect on capital expenditures and hiring.
Particularly informative (and heartening) are the news accounts of the
economic effects of cell phones in third-world counties - centuries-old
inequities are being erased because people can now reach around
traditional middlemen (who would be classified as pure rentiers by
Marx) and discover what the market price for their products are, and
seek the best price. And keep a far higher fraction of the proceeds.
It's positive news.
The correlation is between recessions and employment, not between wage
hikes and employment.
Well, both are likely. Usually, wages go down during recessions, and
rise when times are good. Supply and demand.
Again, look at the evidence. MINIMUM wages, at least, tend to go up
during recessions. Why do you think that is?
Because Congress raises the floor under political pressure.
Yes, but the pressure comes while GDP and CPI are rising. Again, look
at the timing. The beginnings of those min. wage hikes (which usually
occur in multi-year steps) are legislated before the downturn. The
political pressure seems to relate more closely to increases in the
CPI, which also happens to make more sense.
If it were due only to economic forces, the minimum wage would rise and
fall all by itself, long before Congress was invented.
Well, it did, for centuries. And some of the "falls" -- notably from
the 1870s through the Great Depression -- produced catastrophic
results.
That's why we now have minimum wage laws.
And why does the economy
tend to turn up after a string of two or three years of minimum wage
increases?
Because the broad economy recovers. There is always a dispute as to
the effect of Congress, harmful or beneficial, too little too late, or
nothing at all.
The broad economy recovers despite what is going on with wages,
suggesting that minimum wage hikes are not a measureable, causative
factor at all. This is why I think that increasing the minimum wage
has been one of those "small" improvements that makes more economic
activity possible, based on its effect on the behavior of consumers at
the bottom of the economic ladder. The effects, though, as so small
that you can't draw many conclusions, except that people at the bottom
have more money.
Flipping the question around, how can those unfortunate folk making
minimum wage, being the least well paid 5% of the workforce, have any
effect whatsoever on the broad economy?
Nothing directly. Mostly, it's social and political.
And it's really unwise to be demanding a raise or threatening to leave
while others are being laid off.
This is the rationale behind the suggestion that we plot wage deciles
against minimum wage over time.
It would add nothing to the employment data, but if it floats your
boat, go for it. You can get quintile data from the IRS.
The point was not to add to employment data (which we cannot do
anyway), but to put the minimum wage into economic context.
Well, you can plot the first three quartiles against anything else you
desire on the St. Louis Fed research page.
Yep. There is a lot in there. And in the BLS data.
The BLS data mentioned later actually answered the question of
percentile of workforce getting minimum wage - 5% these days.
Right. The year-by-year BLS reports we've referred to, about the
characteristics of minimum-wage earners, get specific about it and
tell you some things about who they are.
--
Ed Huntress
Joe Gwinn
It would be useful to plot minimum wage and the earning quartiles or
deciles together. That should tell the tale.
That will take you quite a long while, unless you find a way to
extract quintiles from something simpler than IRS data. That stuff is
a bugger to work with.
The FRED site you posted above does have quartile data (see under Add a
Data Series), but I wasn't familiar enough with that site to get it to
work properly. You have far more practice with that site.
Aha. Well, they've added data for the first three quartiles, from some
source I didn't check. I could plot that for you, if you want, over
the weekend. It will be one messy graph, with all three in there
plotted against minimum wage increases, but it can be done.
I think we only need the minimum wage and the lowest quartile lines to
give the context.
You need an account with them to show more than a couple of lines on
one graph, and to store them. It's free. If you're interested, it's
worth it. They've set it up very nicely.
Decile data must be somewhere. Actually, the whole percentile curve is
available (there was a thread on this some time ago).
I don't know where. That may be buried in the DoL site, probably at
BLS. It's nowhere near as slick as the St. Louis Fed.
I'll dig that thread up.
However, you already have the percentages making minimum wage.
What is the source of the percentages given earlier in the posting?
BLS:
http://data.bls.gov/search/query/results?cx=013738036195919377644%3A6ih0hfrgl50&q=c haracteristics+of+minimum+wage+workers
This actually gives the percentile limit.
We now know that current minimum-wage workers are in the lowest decile.
Do we have the age and education-level distributions of these workers?
Yes. I think it's all in the BLS reports (one report for each year)
noted above.
I'll look into it. There is a lot there, all in a big pile.
One would assume that (excluding teenagers' summer jobs and the like)
these are the people with no better option.
Probably.
Joe Gwinn