Consumer Price Index vs. Federal Minimum Wage

[Latrinsorm]

We recently had a discussion about how increasing the minimum wage might/could/would/tacos affect the economy, and it's such a beautiful evening out that I figured what better way to spend it than hunkered over a computer crunching numbers?

First two quick definitions:
Consumer Price Index (http://en.wikipedia.org/wiki/Consumer_price_index) is "a measure of the average change over time in the prices paid by urban consumers for a market basket of consumer goods and services."
I'm using the federal minimum wage, as that is what prompted the most recent discussion. I'm also using said wage for 1978-present, because (as the following source will show) the law was a bit of a mish-mosh before then, which would introduce unnecessarily variables in our experiment.

And my three data sources:
CPI - ftp://ftp.bls.gov/pub/special.requests/cpi/cpiai.txt
Minimum Wage (nominal) - http://www.dol.gov/whd/minwage/chart.htm
Minimum Wage (adjusted for inflation, or "real") - http://www.infoplease.com/ipa/A0774473.html

Now, what happens when we split out the average change in average-to-average CPI depending on whether the nominal minimum wage increased? We get 5.62 ± 9.04 for increase years and 3.05 ± 2.16 for non-. Obviously that isn't very helpful because the CPI fluctuated wildly in increased minimum wage years, everywhere from a 13.5% increase to a 0.4% decrease, and the method also fails to distinguish between 1980's 7% increase in min wage and 2007's 14% (thanks a lot, Obama!!!). What if we took all the nominal increase years and graphed % min wage increase vs. % CPI increase? We would get:

http://img.photobucket.com/albums/v456/johnnyoldschool/CPIvsMW-Nom_zps665f72bb.jpg

Please note! We get hardly any correlation, and if anything it is the opposite of what is obviously true: the more the minimum wage increases, the less CPI increases.

.

One thing we could do to both be more accurate and increase our data set is to adjust for inflation, because that way the minimum wage (almost) always changes year to year:

http://img.photobucket.com/albums/v456/johnnyoldschool/CPIvsMW-Real_zps662adb94.jpg

Now we have even less correlation than before, and even the outliers disagree with what's obviously true: the largest real increase in minimum wage was 2009, the only year in the sample that the CPI went down. The three largest CPI increases were in years when the real minimum wage decreased, and the fourth largest was in a year where the real minimum wage stayed the same. We do have a promising looking section between 0 and -5% wage change and 0 and 5% CPI...

http://img.photobucket.com/albums/v456/johnnyoldschool/CPIvsMW-RealSub_zpsc7ea8fa0.jpg

...but while it's an excellent fit, it still empirically shows the exact reverse of what some might expect. Changing the minimum wage does not result in the same change in prices, leaving everyone where they started. Changing the minimum wage (for a certain value of "change") is actually correlated with a change in the opposite direction: people's wages are worth less and things get more expensive.

.

Another thing we could do is say "well it doesn't happen overnight!!!", but if we look at the CPI changes for years following a nominal min wage increase, we get 4.61 ± 8.59 in affected years vs. 3.47 ± 3.78 for unaffected. Again it doesn't tell us anything, and the graphs show no global correlation.

.

Now, another thing we could say is that the minimum wage did do what we all know it did, and that some other effects happened to overpower it in every instance on record. It is true that the economy is complicated. All I would ask is if we have evidence one way and no evidence save for what we reckon the other way, why go with reckoning? It does not seem to have a very good track record, empirically speaking.

[Valthissa]

We recently had a discussion about how increasing the minimum wage might/could/would/tacos affect the economy, and it's such a beautiful evening out that I figured what better way to spend it than hunkered over a computer crunching numbers?


Kreuger vs. Neumark

Right now Krueger is in a position to test out his theories on a real economy so I would bet on a push to raise the minimum wage. The best analysis on the subject seems to be that the effect of minimum wages on our economy is small since when the minimum wage changes there is little measurable impact.

C/Valth

[Wrathbringer]

5443

[ClydeR]

Way to bury the lede.

[Latrinsorm]

Kreuger vs. Neumark

Right now Krueger is in a position to test out his theories on a real economy so I would bet on a push to raise the minimum wage. The best analysis on the subject seems to be that the effect of minimum wages on our economy is small since when the minimum wage changes there is little measurable impact.I looked up those fellows but could only find slap fights on message boards about them, but I'm glad we all came to the same conclusion! :)
jpgFlashing quarters around these parts? You'd better watch yourself, mister.

I also realized that CPI is not adjusted for inflation, so that perfect line I found really shows that CPI changes and inflation are very nearly the same, which stands to reason. The first graph had no inflation adjustment and so remains correct.

[Bobmuhthol]

I also realized that CPI is not adjusted for inflation, so that perfect line I found really shows that CPI changes and inflation are very nearly the same, which stands to reason. The first graph had no inflation adjustment and so remains correct.CPI changes and inflation are exactly the same. Inflation is defined as CPI_1 / CPI_0.

[Latrinsorm]

Well alright then.

[crb]

Grocery baggers the world over are begging for a minimum wage increase, so are the drink fillers at McDonalds, gas station attendants, and of course the 95% of all work-age teens, especially minorities, that have a job.

Not every business and not every industry has pricing power, especially during poor economic times like recent cherrypicked examples. Businesses cannot always pass long increases in costs. So they do other things, like eliminate jobs, not hire people, or reduce a regular carton of ice cream to 1.65 quarts from 2 quarts while keeping the price the same and hope we don't notice.

Raising the minimum wage increases average earnings per hour for the employed while also increasing unemployment for the low skilled, the people who tend to be hurt the most are teens and minorities, and of course like any artificial price it distorts market incentives.

Not a good idea, a better idea to help people make more money, cut all the government crap going on right now that is currently discouraging employers from hiring.

[Shaps]

Math math math math math math and more math. We should just do away with all currency and go back to the good old days of bartering.

Hmmm.. I'm hungry for some chicken, but I got no chickens! Billy Bob got him some chickens. Hmmm.. I got me some old socks. Billy Bob don't got no socks, and winter be coming. Win win!

[Bobmuhthol]

I didn't follow most of the regressions presented because you think like a statistician and not an economist, but using real minimum wage as a variable is probably a really bad idea. The real minimum wage would basically go down every year except years where the nominal wage is increased by the government, and those occurrences are far too infrequent to start throwing into predictive models.

[Latrinsorm]

recent cherrypicked examples.Me?
Raising the minimum wage increases average earnings per hour for the employed while also increasing unemployment for the low skilled, the people who tend to be hurt the most are teens and minorities, and of course like any artificial price it distorts market incentives.

Not a good idea, a better idea to help people make more money, cut all the government crap going on right now that is currently discouraging employers from hiring.Well, we can certainly look at unemployment, I'll use this source:
http://data.bls.gov/pdq/SurveyOutputServlet

The average unemployment change in years with wage increases was 12.8 ± 41.1, the average in non-increase years was -2.6 ± 23.3. As the ±s indicate, the data sets are really wild: everything from 60.1 to -8.6 for increase years and 27.5 to -21.8 for non-increase. We can also throw up another graph:

http://img.photobucket.com/albums/v456/johnnyoldschool/UnemploymentvsMinWage_zps519b041f.jpg

I think that's the least correlation I've ever seen. As the source says, this is for everyone rather than strictly "low skilled", but I can't find a definition for that, let alone any numbers. (And the same lack of quantification for "unskilled", which I would have thought was well established.)
I didn't follow most of the regressions presented because you think like a statistician and not an economist, but using real minimum wage as a variable is probably a really bad idea. The real minimum wage would basically go down every year except years where the nominal wage is increased by the government, and those occurrences are far too infrequent to start throwing into predictive models.It was a bad idea, yes. One interesting thing was that all three of the min wage increases from '79-'81 were in real terms still decreases, but that doesn't have anything to do with the topic.

[Bobmuhthol]

As the source says, this is for everyone rather than strictly "low skilled", but I can't find a definition for that, let alone any numbers. (And the same lack of quantification for "unskilled", which I would have thought was well established.)Education.

[Latrinsorm]

Ohhh. That should be easier to find... but I still can't tease out any numbers from Internet. Oh well.

[Bobmuhthol]

http://research.stlouisfed.org/fred2/search?st=unemployment+by+education+level

You can also change level to rate, which would probably make more sense.

[crb]

Me?Well, we can certainly look at unemployment, I'll use this source:
http://data.bls.gov/pdq/SurveyOutputServlet

The average unemployment change in years with wage increases was 12.8 ± 41.1, the average in non-increase years was -2.6 ± 23.3. As the ±s indicate, the data sets are really wild: everything from 60.1 to -8.6 for increase years and 27.5 to -21.8 for non-increase. We can also throw up another graph:

http://img.photobucket.com/albums/v456/johnnyoldschool/UnemploymentvsMinWage_zps519b041f.jpg

I think that's the least correlation I've ever seen. As the source says, this is for everyone rather than strictly "low skilled", but I can't find a definition for that, let alone any numbers. (And the same lack of quantification for "unskilled", which I would have thought was well established.)It was a bad idea, yes. One interesting thing was that all three of the min wage increases from '79-'81 were in real terms still decreases, but that doesn't have anything to do with the topic.

You're finding wild data because it is a really shitty observational study you're trying to do. Not enough data, too many uncontrolled variables. Sort of like Obama's "Jobs created or saved" concept, you can't prove what jobs you saved unless you went back to a parallel universe, didn't pass the stimulus, then compared the economic results. When the minimum wage is raised, employment could go up, it could go down. What matters would be would it have gone up more, or down less, would it not have been raised?

The closest thing you can do is look at trend lines, and did the policy change mark an inflection point. And then you have to assume there is no other concurrent forces that could have caused a shift, and you have to both look at when policy is proposed, when policy is voted on, when policy is passed, and when policy is implemented, because businesses plan for the future. And you can do all that, but because so much of it ends up depending on the judgement (and so bias) of the person doing the work that someone on the otherside will just dismiss it regardless, so whats the point?

Or you can just use common sense, progressives seem to have no problem grasping the idea that when you increase the cost of something, people will buy less of it, when they want to raise taxes on things like tobacco or alcohol or gasoline. But apply that same universal concept to the labor market and they get dumb. Yes some businesses need labor no matter what, and some people can't give up tobacco no matter what, but on the margins, costs matter, and you should expect human beings to respond to incentives.

And like I said, if the government wants certain people to earn more money, better ways to do it that don't rely on artificial market distorting price controls.

[Latrinsorm]

Yeesh, not a very good commentary on my google skills. Shame it only goes back to 92, but at least it's something, and in xls format, alright!

The first thing I see is that the net change in unskilled is 16% lower than the net change in total, or that unskilled labor has seen less of an increase in unemployment than skilled labor. I'm not sure if that makes sense or not, but it is interesting.

With an even smaller data set, of course our standard deviations make conclusions even less meaningful, but here are the no error figures:

Total
increase: +16.6%
same: -4.5%

Unskilled
increase: +14.8%
same: -2.8%

The only conclusions we can draw are nothing (taking into account the standard deviations) or that unskilled workers are less prone to unemployment due to minimum wage increases (ignoring them). I see no way at all to fit the reverse conclusion to this data.

crb, what do you think?

[Latrinsorm]

You're finding wild data because it is a really shitty observational study you're trying to do. Not enough data, too many uncontrolled variables. Sort of like Obama's "Jobs created or saved" concept, you can't prove what jobs you saved unless you went back to a parallel universe, didn't pass the stimulus, then compared the economic results. When the minimum wage is raised, employment could go up, it could go down. What matters would be would it have gone up more, or down less, would it not have been raised?In my defense, I did mention this in my first post. I even mentioned what your counterpoint would be, albeit with a different term:
Or you can just use common sense, progressives seem to have no problem grasping the idea that when you increase the cost of something, people will buy less of it, when they want to raise taxes on things like tobacco or alcohol or gasoline. But apply that same universal concept to the labor market and they get dumb. Yes some businesses need labor no matter what, and some people can't give up tobacco no matter what, but on the margins, costs matter, and you should expect human beings to respond to incentives.There is no place for common sense in science. If you have no data, you have no conclusion. (Please note how this is not the same thing as a negative conclusion.) You can have very plausible hypotheses, but you will be looked at askance if you angrily denounce anyone who does not accept them on force of plausibility.

[Valthissa]

There is no place for common sense in science.

I am a child of the 60's and a student of the 70's.

Reading that brought back fond memories of Karl Popper.

<insert googling here>

ah, yes:

Objective Knowledge, published in 1972

I can't find the quote I'm looking for but from amazon I can see it's in chapter 2.

It's something like 'science begins with common sense' (it's one of the principles of the book, which I read in a philosophy course).

Carry on with more charts and analysis. I am entertained.

C/Valth

[Bobmuhthol]

The first thing I see is that the net change in unskilled is 16% lower than the net change in total, or that unskilled labor has seen less of an increase in unemployment than skilled labor. I'm not sure if that makes sense or not, but it is interesting.Yes, this does make sense. Unskilled labor will take any job that comes to it; skilled labor either will not take the job or will not be offered the job due to a skill/income mismatch. When a financial crisis occurs, for instance, a lot of layoffs occur in finance... but hardly any, if any at all, of those people do not have college degrees. The receptionists in my company have bachelor's degrees.

[Kembal]

I'm not familiar with the previous thread, so I don't have context, but I have to go with Bob's take on your set of graphs.

One thing to keep in mind is that from 1978 onwards, we have never been in a state where inflation was too low as is the case now, so there's no natural experiment in that time frame to understand the effect of raising the minimum wage in this type of economic state. We can guess that it could increase inflation, though we wouldn't know (because estimating the size of the effect would be difficult) whether it'd be a small increase (good for the economy) or a large increase (bad for the economy, though not immediately).

[Latrinsorm]

I am a child of the 60's and a student of the 70's.

Reading that brought back fond memories of Karl Popper.

<insert googling here>

ah, yes:

Objective Knowledge, published in 1972

I can't find the quote I'm looking for but from amazon I can see it's in chapter 2.

It's something like 'science begins with common sense' (it's one of the principles of the book, which I read in a philosophy course).

Carry on with more charts and analysis. I am entertained.

C/ValthI don't have any more. :( Unless you want to talk about basketball!!
I'm not familiar with the previous thread, so I don't have context, but I have to go with Bob's take on your set of graphs.

One thing to keep in mind is that from 1978 onwards, we have never been in a state where inflation was too low as is the case now, so there's no natural experiment in that time frame to understand the effect of raising the minimum wage in this type of economic state. We can guess that it could increase inflation, though we wouldn't know (because estimating the size of the effect would be difficult) whether it'd be a small increase (good for the economy) or a large increase (bad for the economy, though not immediately).Then we're all agreed - ain't nobody know what's going to happen to the economy if we raise the minimum wage.

[Tgo01]

Then we're all agreed - ain't nobody know what's going to happen to the economy if we raise the minimum wage.

Jesus begs to differ.

[Bobmuhthol]

Then we're all agreed - ain't nobody know what's going to happen to the economy if we raise the minimum wage.Very basic economic theory proves (in an internally consistent way, at least) that one of two things must happen: nothing, or something worse than what would otherwise happen.