Cargo Cult Regulation -- How Much Effect Did Card and Krueger Have on New York's Fast Food Minimum Wage Ruling?

New York is proposing a $15 minimum wage for any fast-food restaurants that are part of a national chain with 30 or more stores.  How this survives any sort of equal protection test is beyond me -- if I own a restaurant and call it "coyote's place" I don't have to pay $15, but if I own a single restaurant where I pay franchise fees to McDonald's, I do.

Let's leave the inevitable court challenges on fairness aside.  Of all the possible industries, I wonder why the focus on just fast food and on just large franchises.  Some of it is obviously mindless Progressive soak the rich thinking, and some of it is a liberal distaste for any foods that are not kale.  Is it just because the fast food workers have been the most vocal?  If so, that is pretty lame the the government is merely focusing on the squeaky wheel, a real indictment of any pretensions technocratic politicians have to legislating intelligently.

But I wonder if it is something else.  Pick a progressive on the street, and in the unlikely event they can name any economic study, that study will probably be Card and Krueger's study of the effect of a minimum wage increase in New Jersey.   Sixty bazillion studies have confirmed what most of us know in our bones to be true, that raising the price of labor decreases demand for that labor.  Card and Krueger said it did not -- and that a minimum wage increase may have even increased demand for labor -- which pretty much has made it the economic bible of the Progressive Left.

What intrigues me is that Card and Krueger specifically looked at the effect of the minimum wage on large chain fast food stores.  In this study (I will explain the likely reason in a moment) they found that when the minimum wage increased for all businesses in New Jersey, the employment at large chain fast food restaurants went up.

So I wonder if the Progressives making this ruling in New York thought to themselves -- "we want to raise the minimum wage.  Well, the one place where we KNOW it will have no negative effect from Card and Krueger is on large fast food chains, so...."

By the way, there are a lot of critiques of Card & Krueger's study.  The most powerful in my mind is that when a minimum wage is raised, often the largest volume and highest productivity companies in any given business will absorb it the best.  One explanation of the Card & Krueger result is that the minimum wage slammed employment in small ma and pa restaurants, driving business to the larger volume restaurants and chains.  As a whole, in this theory, the industry saw a net loss in employment and a shift in employment from smaller to larger firms.  By measuring only the effect on larger firms, Card and Krueger completely missed what was going on.

  • http://vikingvista.blogspot.com/ vikingvista

    Come come now. C&K is settled science, as determined by scientifically selected polls of scientifically selected scientists. Don't be such an anti-science creationist flatearth holocaust denier.

  • J_W_W

    The American left is so sold out to crony capitalism that they can't even see anything but large businesses. This is ok for them now because they have also realized that large businesses are much easier to regulate and many of them can be bought off.

  • mlhouse

    Arguing that increasing the price of something will increase its demands is just plain ludicrous. If a study makes such a claim that argues against elementary logic, then clearly the assumptions and parameters of any such study need to be reexamined.

    The return on capital is the return on capital. Owners of sunk capital may accept a low rate of return when certain regulations are changed. So, they might do some mixture of raising prices and changing schedules that a higher mandated price on labor would force. They may even leave their labor force the same. But they will not hire more labor. In fact, what most restaurants do is substitute higher level workers for the lower level ones. That is why restaurant managers work so many hours in a week. They are doing the jobs of people who should be paid $5-7/hour but cannot be hired because of the mandated minimum wage.

    REgardless, the next wave of technological innovation is coming. Already some fast food chains have ordering kiosks. Other restaurants have wireless devices that you can order and pay your bill. My local bank now only has automated tellers instead of human tellers for all normal transactions. Soon the entry level job and even the intermediate level jobs will no longer exist.

  • Not Sure

    I have a friend who runs a business that uses mostly minimum wage workers. We talked about the effects of raising the minimum wage a while back and what I was told was that, in order for the business to remain profitable, wages couldn't exceed X% of gross income. So- absent the unlikely event that the business's income suddenly increased once workers started getting paid more, raising the minimum wage would mean that workers would get fewer hours. Raise prices instead, you ask? I did, and was told that if it was thought that raising prices wouldn't lower gross income, prices would have already been raised.

  • http://vikingvista.blogspot.com/ vikingvista

    It isn't logically impossible. One can imagine complex economic conditions in which it might be true without violating the law of demand.

    But such unusual complexity needs to be reasonably explained, and the evidence needs to be sufficiently strong. A study of a single small wage ban hike that fails to find statistical significance in either direction (setting aside even the other particular study limitations and the substantial body of evidence showing negative employment effects) is hardly strong enough evidence to overcome Ockam's Razor prudence.

  • http://vikingvista.blogspot.com/ vikingvista

    Aggregation of businesses toward monopoly is a marxians claim of the nature of capitalism. Since reality has demonstrated the wrongness of that theory, marxians turn to government to force such aggregation through state regulation. One way or another they will try to realize their socialist monopoly dream.

  • Matthew Slyfield

    "they have also realized that large businesses are much easier to regulate and many of them can be bought off."

    You have that backwards. The left has realized that large businesses can pay more to buy them off.

  • randian

    Large businesses are also much easier to protest against and are more responsive to negative publicity.

  • joe

    Card Krueger is the study regularly cited by Krugman as "proof" that raising the minimum wage will increase employment.

    There are two major flaws / omissions in the study's data which leads to an erroneous conclusion.

    The fast food stores can be divided into groups.

    1) the stores in PA which paid the 4.25 minimum wage

    2) the NJ stores already paying the higher minimum wage

    3). the stores in NJ paying the 4.25 minimum wage which had to raise the pay to the new NJ minimum wage of 5.05

    The data gathered in the study showed that the hours worked remained constant in the first two groups and only third group which had to increase pay to the higher minimum wage had an increase in hours worked. As such Krugman concludes that raising the minimum wage will increase employment.

    what is omitted from the data -
    A) whether there was an increase or decrease or no change in sales volume
    B) whether the stores were understaffed, overstaffed or staffed correctly at when the minimum wage change went into effect.

    It should be noted that the most likely event was the supply of workers was low since the 4.25 was below the market rate of pay for similarly skilled workers in the NJ area. This would tend to show the law of supply and demand is valid. Again contrary to Krugman's conclusion that the demand and supply of workers is inelastic at the minimum wage level.

  • joe

    It isn't logically impossible. One can imagine complex economic conditions in which it might be true without violating the law of demand.

    See my comment above. The increase on hours in the Card krueger study came from the stores that had to raise their pay. This was most likely due to the market rate for that skill level of workers was higher than the minimum wage and therefore there was a shortage of that level of worker.

  • http://vikingvista.blogspot.com/ vikingvista

    I was referring to the phrase by poster mlhouse:

    "If a study makes such a claim that argues against elementary logic"

    Even a perfect study that finds a significant increase in employment following a hike in the wage ban could be true, and does not defy any logic. It just would indicate highly unusual circumstances with substantial factors confounding the straightforward application of the law of demand.

  • joe

    milhouse's comment was "If a study makes such a claim that argues against elementary logic, then clearly the assumptions and parameters of any such study need to be reexamined."

    Milhouse's comment is correct - when the data shows a different result that what is expected - you should re-examine the data.

    It the card Kruger study, they failed to obtain sales data and failed to obtain data on staffing levels before and after the change. the omission of these two items leaves a hole in the data.

  • http://vikingvista.blogspot.com/ vikingvista

    Different than expected is weaker than defying elementary logic. Discovering an increase in employment due to a hike in the wage ban is different than expected. It dies not defy elementary logic.

    Some people think any finding such as C&K defies the law of demand and defies logic. It does neither.

    BTW, I'm well aware of the contents of the study and its limitations. That is besides the point.

  • joe

    BTW, I'm well aware of the contents of the study and its limitations

    what other short comings did you observe in the study

  • http://vikingvista.blogspot.com/ vikingvista

    What is important is what shortcoming I did NOT find. It did not argue against logic. Over generalization may be an inductive error, but it does not violate elementary logic.

  • John O.

    New York is "proposing"? It was approved by the Wage Board, which is an agency created in the law for the state minimum wage hike a few years ago, and will go into effect once they finish writing all the new rules to go with it. Seriously once the bureaucracy was created to bypass the Legislature on this issue, the state administration jumped on it as fast as they could and got exactly what they wanted so they can get more votes.

    Its already in the local papers that the Wage Board is looking to raise wages in other "low wage" industries, which is a thinly veiled code for retail and hospitality. Things are going to get very expensive for the people of New York while at the same time its economy crashes. Its made some meager gains lately in WNY, but they're going to be completely wiped out once all this Wage Board nonsense is pushed through.

  • timworstall

    "One explanation of the Card & Krueger result is that the minimum wage slammed employment in small ma and pa restaurants, driving business to the larger volume restaurants and chains. As a whole, in this theory, the industry saw a net loss in employment and a shift in employment from smaller to larger firms. By measuring only the effect on larger firms, Card and Krueger completely missed what was going on."

    Pretty much my explanation of it. And just recently I flagged up a paper at Forbes from Isaac Sorkin that describes this very effect. A rise in the min wage drives labour intensive people out of the market and they largely get replaced by capital intensive firms serving the same market.

  • Jim Collins

    I think that this might have been mentioned here before, but, I'll say it again. This $15 minimum wage is being fronted by the unions, who get an exemption from it. Once this passes then they can tell a store owner, you can either go out of business or let your employees unionize and only pay them "x" amount. If anybody else tried this it would be extortion or collusion, but, because it is the unions and their paid off elected officials, it is OK.

  • Tim

    The unions are fronting the minimum wage increase; but not because they have an exemption, which would have to be explicitly created in the local minimum wage.

    Unions are fronting the minimum wage increase because a lot of union contracts are written with the base wage rate as an index to the minimum wage. In other words, the rate isn't $y/hr; it's $7.25+x/hr. Raise that $7.25, and the rank and file get it too.

  • sean2829

    Two comments.

    A ploy was tried like this in Maryland to force Walmart to provide healthcare to all it employees. The law had conditions on business size and number of employees such that only 1 business in the state would be regulated under the law. The law was thrown out before it took effect. Will the targeted New York law have the same result? Don't know enough about the laws to venture a guess.

    Regarding using legislation to accomplish the exact opposite of what is intended, I can think of several examples where egregious behavior by corporations at the top of an industry benefited those companies more than others. Mattel's lead in paint on toys resulted in laws that drove up the costs for nearly all toy manufacturers but smaller competitors paid a much greater share of their revenue for compliance, making them less viable. Controls meant for big banks in Dodd-Frank has resulted in most smaller banks going out of business. And in the 90's when the government put financial controls on many defined benefit pension plans, most companies decided these were no longer worth the risk and eliminated them.

    When the government was less of something, just make it more expensive.

  • morganovich

    parsing the results of the card krueger study is meaningless. their study was FRAUD.

    when they were asked to give people a look at their actual data, it was completely absent. they made it up. their "phone study" was nonsense.

    then, when neumark went out and replicated their study using the actual personell records of the exact same stores, it showed the opposite of what CK claimed.

    http://www.nber.org/papers/w5224

    read the summary of the pdf.

    note that this is the most accurate and exact form of primary source material.

    jersey has a 4.6% decrease in employment relative to PA. this result is statistically significant (p value < .05) and demonstrates a price elasticity of -.024.

    anyone citing the card krueger study is either a charlatan or is unfamiliar with the body of work on labor economics.

  • morganovich

    vv-

    the study was fraud. (see above).

    it "defies expectations" because they made up their data and could not provide any actual data for others to check.

    arguing about its logic is meaningless. it was dishonest from the word go and was completely discredited by neumark who surveyed the exact same stores using their actual payroll records and found, as predicted by the theory, a -0.24 elasticity at highly statistically significant levels. (p<0.05)

    also:

    you seem to think you are arguing for logic, but what you are really doing is arguing against good study design.

    a well designed study with a control group should leave little room (ideally none, but this is hard in economics) for the exogenous issues you keep citing.

    joe has this mostly right. if one gets wildly counter intuitive results, generally you want to look at the study design, then the data collection.

    i look at this stuff all the time and am an expert in analyzing studies (both in econ and in biotech and hard sciences)

    a study that cannot be replicated is invalid.

    a study that cannot even show its data to others is generally fraud.

    in this case, the design was OK, but the purported "phone survey" that gathered data was complete nonsense.

  • joe

    Morgan - switching topics -

    Any comments on this study - The Bell study 2004 JAMA shows that there is a .52% increase in premature deaths when there is an increase of 10ppb (parts per billion) in ground level ozone.

    http://jama.jamanetwork.com/article.aspx?articleid=199808#RESULTS

    I am going to withhold my comments on the study until I see your evaluation/comments/observations
    thanks

  • morganovich

    joe-

    i am not familiar with the study and have not gone through the mathematical structures and statistical methods it uses, so i am hesitant to say a great deal. JAMA has really spotty quality control and often engages in non arms length peer review, but they also publish some very good stuff.

    cursorily, the basic design is reasonable. cities are used as their own controls as this is a short term event.

    the math here gets really tricky. they are using a distributed lag model and how that is parameterized is more art than science. it also can have big effects on outcome. it's a bit like trying to use seasonal adjustments in economic data. getting a real baseline is very hard and seasons are not always the same. some springs are unusually warm or wet or windy.

    this is a very hard model to run because there are tons of factors that effect mortality and you cannot control for them or adjust them out. many of them are likely not even known.

    then we get to the heart of the issue which is this:

    "A 10-ppb increase in the previous week’s ozone was associated
    with a 0.52% increase in daily mortality (95% posterior interval [PI], 0.27%-0.77%)
    and a 0.64% increase in cardiovascular and respiratory mortality (95% PI,
    0.31%-0.98%)"

    the key word here is "associated". this means "correlated". correlation is not causality.

    people opening umbrellas correlates highly with the sidewalk getting wet. it's not the cause.

    it's possible some 3rd think causes mortality and an increase in ozone. maybe dying causes ozone (i doubt this very much, i'm just illustrating a point).

    the point is that an association is not a causal link.

    if there were no correlation, that could well be a strong safety signal. correlation is not causality, but causality without correlation is rare.

    so, sorry, i'm not sure i'm much help here. this study (assuming the math and stat handling is good), but it could very easily have either bad math or mean nothing even if the math is good.

    that is not surprising nor an indictment of the researchers. this sort of environmental macro study is really, really hard to do.

  • joe

    Thanks for the comments - The EPA in the early 1990's forced the paint manufacturers to reduce the VOC's in paints along with a host of other products in an effort to reduce ground level ozone. The result was a significant reduction in the quality of oil based paints - short sighted since the paint was less durable and any reduction in the voc's was offset by the increased painting. FWIW - This voc obsession is a pet peeve of mine.

    FWIW - this was actually the study that made me question the statistical inferences with the AGW / co2 correlation. A 10ppb increase corresponds roughly with a 3-5 degree F increase in temp, yet the increase in premature mortality was presumed /correlated with the increase in ozone and not the increase in temp.
    It was hard to fathom, that a trace level of a pollutant was more responsible for premature deaths than the increase in heat.
    This also speaks to the studies showing significant health benefits from cleaner air. As the air get incrementally cleaner, due to diminishing returns, the health benefits will get incrementally smaller as the air becomes cleaner. The EPA presumes the health benefits will continue to grow.

    thanks for the comments

  • morganovich

    in general the EPA are terrible scientists. they have relied upon utter junk science in a great many cases from DDT to refrigerants.

    i think they are currently doing this with co2.

    they have done some real good as well (pcb's, lead, etc.) but they seem highly politicized, haphazard, and prone to crony capitalism at the behest of big political donors. (yes alt energy industry, i'm looking at you)

  • randian

    They're most likely know it's junk, but they allow their (universally progressive) politics to supersede the science. Like many (I would argue most) progressives, they're willing to lie (aka use junk science) to advance their political agenda.

  • joe

    In the late 1990's/early 2000's, A former next door neighbor of mine worked for the EPA, he has sinced moved to DC as he got promoted. I brought up the issue of the VOC's in paints and the corresponding reduction in adhesion and durability which resulted in greater pollution, not less, especially an increase in solid waste. he response was somewhat guarded, but essentially admitted that the EPA doesn't do a good job in the cost benefit analysis, He seemed to be unaware of the concept of marginal cost v marginal benefit - with a corollary in law of diminishing returns. Similar response to the ethanol pollution issues.

  • Joe

    Krugman is fine example - very good economist that has become ________ to advance his political agenda

  • http://vikingvista.blogspot.com/ vikingvista

    "the study was fraud"

    In spite of the difficulty in proving intent, I'll take your word for it that you have the proof. It has no bearing on anything I've written.

    "it "defies expectations" because they made up their data and could not provide any actual data for others to check."

    Again, I'm not privy to evidence strong enough to prove intent, so I'll take your word for it.
    Since it doesn't matter for a reasonable interpretation of the greater body of evidence whether they reported their results through fraud or sincerity, it doesn't matter to me.

    "arguing about its logic is meaningless."

    So you think mlhouse's comments are meaningless. I disagree. His points reveal an important and common misunderstanding of economic theory, probably also of the nature of empirical evidence.

    "you seem to think you are arguing for logic"

    I definitely am.

    "what you are really doing is arguing against good study design"

    Obviously that is not what I am doing. That's why I strengthened my point by making it independent of any study design.

    "well designed study with a control group should leave little room"

    You seem to take the phrase "economic experiment" too literally. It isn't an experiment (at least by physical science standards it is at best a terrible experiment), and such studies will always leave considerable room for confounders. A good study design attempts merely to randomize out as best they can the ones they recognize, but it still isn't a randomized double-blind placebo controlled trial (not that even that excludes all exogenous factors).

    "joe has this mostly right."

    Except on the point I am making--the whole reason for all my comments here, he is completely wrong.

    "i look at this stuff all the time and am an expert in analyzing studies"

    Then it should come as no great surprise to you when even an honest well-designed study comes up with wrong (let alone unexpected) results. You might be surprised for a large true experiment, but usually surprise would grow as independent confirmations grow, beyond the 1:20 or or 1:100 or so expected by chance. Given the large number of independent looks at the employment effects of wage ban increases on low skilled employment, a few C&K studies (which BTW, show no significant effect) amount to nothing, even if not fraudulent.

    "a study that cannot even show its data to others is generally fraud."

    I don't doubt it. But until the literature is dominated by their results (it isn't) then you might as well chalk it up to sincere aberration, since such aberrations are expected. Of course, in the case of C&K specifically, there are other good reasons to think their study was incapable of detecting negative employment effects, without charging fraud, and certainly without charging illogic, as mlhouse incorrectly did.

    The problem with C&K isn't the results. The problem is that the political class funds and scours the noise for the results they want, then proclaim themselves scientific. The problem with science, is that it is awash in terrible meaningless studies, poor selective reporting, and enormous self-serving incentives to create bias. Vigorous competition is usually the solution, but that doesn't work very well on issues favored by the worldwide political class.

  • morganovich

    viking-

    you seem to be missing the whole point here. (also, can we please not use this cut/paste/respond structure? it makes a conversation so fragmented that it becomes impossible to follow and respond to. in another iteration, it's going to be 40 sub discussions and lots of stuff always gets lost.)

    you have not really made any arguments here for the most part. you just keep saying you are right and providing no reason or rationale in many cases.

    if your point is that it's always possible for something to queer a study if you do not control variables sufficiently, then just say it. this is as obvious as it is true, particularly in a field as hard to control variables in as an economy. it is, however, not the case in this situation.

    my point is this: your claims are generally true, but not in this specific case.

    the original CK study was not based on real data. read the neumark study i linked. that study IS based on real data. it uses the best possible primary source material. the result are just what the theory predicts.

    the study design worked (at least pretty well)

    one can always find outlier studies in something as hard to control for as economics, and yes, if you spend time seeking out outliers to pretend are representative, you can make bad claims. i get your logic. i'm just telling you it does not apply here. what you are describing COULD be an explanation for an outcome like this. however, in this case, it ISN'T.

    you are missing the specific picture here. the problem with CK most emphatically IS the results. the results were faked. the records of the data on which they were based could not even be found, much less replicated.

    as someone who has looked at a great many labor economics studies, i can tell you, the politicos go WAY past souring for outliers. they pay for fraud. there is nothing "sincere" about it and it does dominate the literature if you look at grants etc. they are deliberately creating a mass of chaff. there are so many badly done studies that separating them out is difficult and time consuming. even when one does, they are immediately accused of "cherry picking". this is the whole goal. you create so much bad data that the good data gets lost and if you try to sift good from bad, you get accused of bias. this will not fool real economists, but it will fool the public and provide cover for politicians.

    it's so endemic that the researchers in the space have made it a verb. it's called "playing the kreuger card".

    it's rare to see outright fabrications like CK, but there are lots of ways to get the result you want through bad study design. typical practices:

    1. look at the entire population. min wage earners are less than 3% of workers. take out those who get tips and it drops even further. so, the signal id faint compared to the whole system and anything can swamp it.

    2. use the wrong start date. if you start the study on the day of the wage hike, many fail to see that you have missed the whole game. businesses respond to such a thing BEFORE it happens, not after. in general, there is quite a bit of warning. the adjustment takes place then, not the day the raise hits.

    3. failure to control for hours.

    4. using bad proxies. in an attempt to rebut neumark, CK did the study using unemployment claims as a proxy. they found no correlation between the wage hike and job levels. but this is a ridiculous practice. it's a bad proxy. not everyone who quits or is fired applies. further, it fails to measure changes in growth rate at all. if it just backs off hiring levels among a strong period of growth (as this study had) then this measure misses the whole game.

    i could go on and on. the bag of tricks here is quite large. it is also, however, well known in the profession. these design issues have been addressed over and over. yet "labor economists" seeking grants, political power, and fame keep doing it.

    this amounts to willful misrepresentation.

    i get your point about choosing outliers, but this space goes WAY beyond that.

    it's some of the worst systemic research fraud i have ever seen outside of climate science. (note that both the horrendous BEST climate study and the 1994 CK fraud came out of berkeley. there is something really rotten in their academic climate)

  • morganovich

    i fear it may run the other way.

    big business has realized that buying politicians and getting them to put anti-competitive regulation into a market is cheaper than marketing and a great way to keep small entrants out.

    compliance burdens to odious regulation always fall more heavily upon small guys.

  • morganovich

    campus rape has been another real lulu of bad studies.

    i thought this quote was really excellent (and also bears on the min wage)

    "“It’s one of the most egregious examples of a policy with an inadequate
    scientific basis that lives on because it offers a simplistic solution,”
    said Koss, in an interview with Reason."

    http://reason.com/blog/2015/08/11/campus-rape-expert-who-misrepresented-hi

  • Matthew Slyfield

    If anything, it runs both ways. If the gov didn't have the power to put it's finger on the scale to tip things in favor of one company over another, the big companies wouldn't be pouring so much money into politics. And if you think that politicians aren't actively competing for that money, you are a fool.

  • Joe

    Nobel Prize winner Krugman cites the Card Krueger study all the time - Whats that say about him - never mind -

  • morganovich

    of course they are. it takes two to tango. the politicians are happy to grab power and peddle the influence they wield, and companies are happy to buy it and encourage government to grab more power and sell that too.

    it's a virtue-destroying cycle and it feeds on itself by creating a terrible prisoner's dilemma on both sides. if you competitor is willing to take corporate money for favors of buy political favors to slant markets, if you do not get in the game, you are at a big disadvantage. (note: this is not in any way limited to business. unions, environmentalists, and all manner of special interests are the same thing)

    this is EXACTLY why this trend cannot ever seem to be broken. it's self reinforcing.

    this is why i am such a proponent of highly limited government.

    the only way to take special interest influence out of government is to take influence away from government.

    give the state the power to pick winners and losers and slant the playing field, and it's a race to the bottom to see who can buy it fastest.

  • Matthew Slyfield

    Agreed.

  • stan

    Can anyone point me to an intelligent explanation by a lefty as to why the minimum wage and the demand for unskilled labor is exempt from the economic principles regarding the demand curve?

  • http://vikingvista.blogspot.com/ vikingvista

    “you seem to be missing the whole point here.”

    Since the point was mine, I don’t see how that is possible. Or are you saying I’ve deviated from my original point?

    “(also, can we please not use this cut/paste/respond structure?”

    And make it even harder for you to follow the point? If it is a matter of irritation, then I suppose I could trade you--I’ll stop indicating which of your quotes I’m referring to if you learn to use a shift key.

    “it's going to be 40 sub discussions”

    Only if you choose to continue them. But if it weren’t for tangential conversations, you and I would’ve exchanged almost nothing.

    “you have not really made any arguments here for the most part. you just keep saying you are right and providing no reason or rationale in many cases.”

    Someone actually pays you to analyze articles?

    “if your point is that it's always possible for something to queer a study if you do not control variables sufficiently,”

    I don’t understand the source of your confusion. Given your professional skills, maybe you can help me. Perhaps I’m too much inside myself to recognize the ambiguity of my own language. Please tell me how you interpret the first iteration of my point in response to mlhouse:

    VV: “t isn't logically impossible. One can imagine complex economic conditions in which it might be true without violating the law of demand.”

    “my point is this: your claims are generally true, but not in this specific case.”

    Then you are sorely mistaken. I can think of one way you could prove me wrong: quote me the elementary logical error from the C&K paper. Perhaps I just missed it.

    “the original CK study was…”

    That’s all very interesting, but still doesn’t make sense as an argument against me. As I said, I grant you almost whatever you want about the C&K study, since it really isn’t germaine. The only thing you could say that would be relevant as an argument to my reply to mlhouse, is that the C&K conclusions defy elementary logic. But in spite of giving the appearance of disagreeing with me, you haven’t even attempted to show the one thing that would matter.

    “one can always find outlier studies in something as hard to control for as economics, and yes, if you spend time seeking out outliers to pretend are representative, you can make bad claims. i get your logic. i'm just telling you it does not apply here. what you are describing COULD be an explanation for an outcome like this. however, in this case, it ISN'T.”

    Then you still miss that particular tangential point. If you want to wage a war against academic fraud, fine. That is a noble endeavor. If, on the other hand, you are interested in what happens when the wage ban is increased, then you are wasting your time. Because even if you grant the wage prohibitionists the most favorable view of their favorite study, it still does not alter a rational interpretation of the whole body of evidence regarding wage ban hikes.

    “you are missing the specific picture here. the problem with CK most emphatically IS the results.”

    If the results of C&K are the problem, then apparently an honest study with those results would change your belief of the employment effects of raising the wage ban. But if that is the case, then your grasp of the entire body of MW literature, and your grasp of the robustness of the results-replication part of the scientific process are as poor as any run-of-the-mill Progressive politician’s. And back to the original point, if you also think such results defy economic laws or logic, then in addition, your grasp of economics is poor.

    From reading your excellent comments here and elsewhere, I doubt you suffer those particular deficiencies, so something else is going on with you, probably your inability or unwillingness to reply in context.

    “the politicos go WAY past souring for outliers. they pay for fraud. there is nothing "sincere" about it and it does dominate the literature if you look at grants etc.”

    Fine. You insist on aiding your enemies, by weakening your argument. Your conclusions about MW would be the same, but more robust, even if you grant them honesty.

    “they are deliberately creating a mass of chaff. there are so many badly done studies that separating them out is difficult”

    As much as I like to grant you whatever you want to say about academic dishonesty, you really are going too far with it. I’ve been a part of that world. The ocean of junk studies is more a part of the need in academics for a resume of publications than it is dishonesty. This leads to bias and inexcusable carelessness among researchers and reviewers, but overt fraud is probably a much less, if still significant, part of that. Besides, from a politician’s (let alone researcher’s) point of view, why risk overt fraud when the same results are achieved simply using bad science, which is very easy to get away with, at least in the short run?

  • Joe

    My comments above are based on the presumption that the data collection was reasonably valid. I was pointing out errors in reaching conclusions without exploring other issues that affect employment rates.

    Based Warrens comments and comments in Forbes, it would also appear that Warren made the assumption that the data collection was valid, Warren pointed out other obvious omissions which led to erroneous conclusions in the CK study.

    It has now come to my attention (via morgan above) that there were serious errors in the data collection. A subsequent study using actual payroll records showed that the employment rates / hours worked decreased when the minimum wage was increase. http://www.nber.org/papers/w52...

    I will repeat the comment that Morgan made -

    " anyone citing the Card krueger study is either a charlatan or is unfamiliar with the body of work on labor economics." - this doesnt speak well of krugman who repeatedly cites this study.

  • MJ

    Another critique of the Card and Krueger study which often goes unmentioned -- I have a copy of the paper and have looked at it several times -- is the problem with non-response bias in the sample. This probably dovetails with what Warren mentioned, but the authors themselves note that many fast food firms were reluctant to respond to the survey, given its request for fairly detailed payroll information. It is probable that the smaller firms were more likely to take a pass on responding to the survey than larger, and in many cases, publicly traded firms.

  • MJ

    I think Walmart was also the target of a similar firm-specific minimum wage proposal recently. The law defined the target of the minimum wage so restrictively as to leave no doubt as to whom it was intended to affect. I think was voted down for similar reasons (leaving it open to legal challenge), but it also touched a nerve within the lower-income black community within DC, many of whom are frequent WalMart shoppers. WalMart responded to the proposal by threatening to cancel plans to open a couple of new stores that were slated for DC and this bothered many consumers.

  • MJ

    Thanks for posting the link. I hadn't seen this paper by Neumark and Wascher.

    I wouldn't call Card and Krueger's study a 'fraud', but it does illustrate some of the limitations of empirical research on the minimum wage, especially those that are now two decades old.

    Card and Krueger really shouldn't be blamed for the way people misuse or misrepresent their findings. This is especially true of academics like Krugman who are smart enough to know better, but unscrupulous enough to ignore the warts in these kinds of studies.

  • markm

    Other pollution will be correlated with ozone. It's impossible to tell which pollutant was killing people from this sort of study. And maybe it wasn't pollution at all. Pollution levels are often higher in poor neighorhoods, and poverty is well-known to shorten lifespans.

  • bilbo

    Seattle restaurant job losses: http://t.co/Qh25JrV7HN