Minimum wages and employment

"An increase in the minimum wage will raise workers' incomes, which will increase aggregate demand, which will increase output and employment. So would an increase in union wages."

I remember hearing that argument a lot in the 1970s. You don't hear it as much nowadays, but it still lives on in the underworld of economic ideas.

It's (usually) wrong; but it's not a stupid argument. And you can't dismiss it just by drawing a downward-sloping labour demand curve, and showing how an increase in wages will cause a movement up along that demand curve, to a point with lower employment. The whole point of the argument is that an increase in aggregate demand will shift the labour demand curve to the right.

And (most) firms (nearly) always want to expand output and employment. What constrains firms from doing so is not that wages (and other costs) are too high to make additional sales at current prices unprofitable. What constrains firms is demand. They want to sell more output, and would be able to hire the extra labour needed to produce more output. They just can't find the extra customers to buy more output.

Ask any firm this question: "If I could bring you more demand, at your existing price and quality, would you be able and willing to produce more and sell more?" Most would answer a very enthusiastic "Yes!". That's why they spend money on advertising, for instance.

When I put down the textbook and look out the window I see a world where output and employment are (nearly always) constrained by demand, not by supply. So it seems very plausible to suppose that an increase in wages across the economy, if it raised aggregate demand, would increase output and employment, even if it did raise costs a bit. You wouldn't want to raise wages too much of course. If you did raise wages too much so that marginal costs increased above the price of the good, it wouldn't work, because firms would no longer want to increase output and employment to satisfy an increase in demand.

So, what's wrong with the argument?

First, if you wanted to increase aggregate demand, why not just use monetary and/or fiscal policy? Why increase minimum wages? Fair enough. But suppose you wanted to change the distribution of income as well? Why not use an increase in minimum wages to increase aggregate demand?

Second, why would an increase in minimum wages increase aggregate demand? The simple argument is that an increase in wages increases people's incomes and if people's incomes rise, their demand will rise. But that simple argument is fallacious. It forgets that capitalists are people too. For a given level of output, which means a given level of income, if wage income takes a greater share, then non-wage income must take a lesser share. You need to argue that the distribution of income affects aggregate demand.

It might. Maybe there's a lower propensity to save out of wage income. Or maybe the velocity of circulation is higher for wage income.

But let's get to the main point.

Start in equilibrium where there is no upward or downward pressure on prices (or where the upward and downward pressures are exactly balanced). That does not mean that supply equals demand. It would mean that supply equals demand if markets were perfectly competitive, but they aren't. Markets are imperfectly competitive.

If markets were perfectly competitive, in equilibrium each firm (and worker) is on its supply curve. Price = Marginal cost (Wage = Marginal disutility of leisure). It doesn't want to sell any more output (labour) at the current price (wage). The economy as a whole is on its aggregate supply curve.

The one great thing about New Keynesian macroeconomics is that it let us think about macroeconomics when markets are imperfectly competitive, and each firm (worker) faces a downward-sloping demand curve. And most firms (and some workers) do want to sell more output (labour) at the current price (wage). The economy as a whole is off its aggregate supply curve. It wants to sell more.

Offered a deal in which demand increases so they can sell more output and labour, but cannot raise prices or wages, most firms and some workers would accept the deal. Aggregate output and employment could rise.

But that's not the deal they are offered when aggregate demand increases. They are offered an increase in demand; they can choose to increase output and employment, or increase prices and wages, or a bit of both. Most will choose a bit of both. And that's the problem.

Starting in equilibrium with no upward pressure on prices or wages, any increase in aggregate demand, output and employment will put some upward pressure on some prices and wages. It doesn't matter how small it is; even if the average individual firm (worker) raises its price (wage) just a little, the positive feedback effect results in an infinite rise in prices and wages.

There is positive feedback because each individual firms and worker cares only about the real price and wage -- which means the price and wage relative to other prices and wages. Upward pressure on prices and wages means upward pressure on relative prices and wages. And it's impossible in aggregate for the average firm to raise its price and wage relative to the average price and wage. The attempt by each to do so results in an upward spiral of unlimited speed.

It's the Long Run Phillips Curve that limits the effectiveness of aggregate demand, not the Long Run Aggregate Supply Curve. Both are vertical, but they are not the same curve. The LRAS curve tells you the level of output at which firms and workers would be unwilling to sell more output and labour at existing prices and wages. The LRPC tells you the level of output at which firms and workers would be unwilling to try to raise their relative prices and wages.

If the economy is in equilibrium, at a point on the Long Run Phillips Curve, an increase in minimum wages, or union wages, even if it does increase Aggregate Demand, will not increase output and employment in any sustainable way. Nothing that only increases Aggregate Demand will increase output and employment in any sustainable way.

It's worse than that.

An increase in minimum wages will cause a firm's Marginal Cost curve to shift up. For a given level of demand, an increase in MC will cause an increase in the firm's profit-maximising price. Start in equilibrium, at a level of output and employment where there is no upward (or downward) pressure on prices and wages. So the economy is on the LRPC. Then increase the minimum wage. Now there is upward pressure on prices and wages. The economy is no longer on the LRPC. The LRPC has shifted. We would now need lower output and employment to create an offsetting downward pressure on prices and wages. Anything that increases the upward pressure on prices and wages, for a given level of output and employment, will reduce the sustainable level of output and employment.

Are there any cases where an increase in the minimum wage would cause a sustainable increase in employment? Yes, I can think of three:

1. Suppose aggregate demand is too low, so we are off the LRPC, and there is risk of a deflationary spiral. And suppose that monetary and fiscal policy cannot be used, for some reason. Then it is conceivable that an increase in minimum wages could increase aggregate demand, and bring the economy back to the LRPC, and be the only way to do so. Maybe.

2. Suppose the labour market has monopsony power. Wages are below the competitive equilibrium. There is excess demand for labour at the given wage. Firms won't hire more labour because they would need to raise wages to persuade workers to sell more labour. In this case, an increase in minimum wages would shift each firm's MC curve down, not up. They would respond by lowering prices. For a given level of output and employment, this would put downward pressure on prices. This would shift the LRPC curve in a good direction. But that's a labour market where firms are hungry for workers; not a labour market where workers are hungry for jobs. It doesn't sound much like Canada, on average.

3. This one's weird. Start in equilibrium, at a point on the LRPC where there is neither upward or downward pressure on relative prices and wages. Now suppose that an increase in output and employment would cause downward, not upward pressure on prices and wages. That means the initial equilibrium is unstable. Any small increase in output and employment would cause prices and wages to start falling, and the economy would move down along the AD curve to higher output and lower prices, until it eventually hit the LRAS curve, and prices and wages stopped falling, because nobody wants to sell any more output and labour at given prices and wages. As with all cases of unstable equilibria, the comparative statics have the "wrong sign". Anything that increased upward pressure on wages and prices would shift the LRPC in the good direction.

I could rig up a model that looked like this, if I wanted. Assume some combination of: increasing elasticity of an individual firm's demand curve as aggregate output increases; increasing marginal product of labour as aggregate output increases; downward-sloping aggregate labour supply curve. Anything that causes an individual firm's Marginal Revenue curve to shift up, and/or Marginal Cost curve to shift down, when aggregate output increases.

But we can rule out that last case. It would only be by sheer fluke that any economy found itself in an unstable equilibrium. Almost certainly, the economy would already be at maximum capacity, or else be at zero output.

And in any case, if we were by fluke at that unstable equilibrium, and an increase in minimum wages shifted the LRPC in the good direction, the economy would likely go off in the other direction, towards zero output and employment. The better thing to do would be a sudden burst of monetary expansion, which would cause output to rise and prices to fall.

Comments

[Rick]

"Second, why would an increase in minimum wages increase aggregate demand? The simple argument is that an increase in wages increases people's incomes and if people's incomes rise, their demand will rise. But that simple argument is fallacious. It forgets that capitalists are people too. For a given level of output, which means a given level of income, if wage income takes a greater share, then non-wage income must take a lesser share. You need to argue that the distribution of income affects aggregate demand....."

I don't think that you need to argue it, it's fact. Slavery?

The argument that is fallicious is that entry level wage income = wage income and further that wage increase for entry level affects non-wage income at all.

Wage increases should come from capital (dividends) as a percentage of after tax profits. That doesn't affect marginal cost of production because its from profits. There is nowhere in economic theory that places a manditory return on investment in the secondary market.... it is not simple econ 101 that you use to derive labour policy since econ 101 cannot explain how when interest rates are low companies still return high rates of return to stock holders instead of reinvest in the company or raise wages......

Aggregate demand has been destroyed in the US. This happened over time in the US because of falling real wages of middle class workers (wages did not keep up with increasing investment returns or upper management salaries)--there was a lag based on credit availability but when the bubble burst we saw the effects of declining real wages-- at a time (pre bust) when the US economy was "robust" and employment very "competitive".

This is a deliberate oversimplification.... its supply side rhetoric... I'm not pro union guaranteed wage increases or even a universal minimum wage but you cannot refute that wage as a percentage of profits is fair and has no negative effect on production if the percentage remains constant.

[Nick Rowe]

Rick: "....but you cannot refute that wage as a percentage of profits is fair and has no negative effect on production if the percentage remains constant."

Ummm. I thought I did just refute that. Did you read my post? Did you understand it?

I expect this is one of the temptations for economists to use loads of math. Then people who would otherwise *think* they understand you would instead *know* they don't understand you ;-)

[Rogue]

But in an open economy, the same capitalist may find a similar product supplied from overseas, entirely substitutable, and import that product instead of building it locally.

Increasing local wages to increase aggregate demand is a good idea, but the previous generation's own good idea was to increase local purchasing power instead, by opening up to globalization. So in came cheaper goods, which made up for the fact that wages were not increasing (and jobs were now disappearing).

[Nick Rowe]

Rogue: In an open economy, it depends on the exchange rate regime. If I want to consider my thought-experiment, where minimum wages increase, with aggregate demand constant, i have to assume flexible exchange rates. So the nominal exchange rate will depreciate one-for-one with the rising price level, leaving the real exchange rate unchanged, and so no change in the relative price of domestic to imported goods.

Globalisation will have lots of real effects. One relevant effect, in this context, is that it may shift the LRPC in a favourable direction. It does this by increasing elasticities of demand throughout the economy (because more substitutes are available). This reduces upward pressure on prices, for a given level of output and employment.

[Lord]

Nice argument but since when has the real minimum wage been increased? Normally it is allowed to erode with inflation and only restored to some level of the past. Want to defend lowering it? One can do so on increased employment but I wonder if the goal isn't to simply to declare unemployment a non problem. Innovation is the way around this argument, but it does require some constraint on global competition.

[Rick]

You'll have to explain it to me in lamens terms I guess how Stock holders who collect divedends are any different than workers collecting salary when as far as accountants of a company are concerned they are both liabilities.

I understand you perfectly well and I am saying than wage distribution has an effect on aggregate demand which is exactly the point.
(and just for the record you didn't use a mathematical model, you inferred a relationship from one without proving it. When that happens its called congecture and its not the basis for academic discourse.... but, should you ever want to go to the blackboard I can make time.... ) I guess that's why people are tempted to use english when discussing things-- then people who would otherwise *think* they understand what they were saying would instead be able to *show* they don't ;)

Also you say that an increase in aggregate demand puts upward pressure on wages... Really? North American statistics over the last 20 years refute that point directly for middle income workers... though prices rose real wages remained stagnant and alot of instances actually fell. That postulate is not true by definition..

I agree that if there is across the board increases in union wage or minimum wage then there will eventually be inflationary pressures and that will lead to negative affects on production and input costs unless prices raise and thus the spiral starts. But if a company is raising profits individually and increasing its market share there is no affect to wage increases.... they are called bonuses....

[Andrew F]

"Wage increases should come from capital (dividends) as a percentage of after tax profits. That doesn't affect marginal cost of production because its from profits. "

How does increasing the cost of wage inputs not change the marginal cost of production? This is just nonsense. Any cost associated with production is a cost of production. You can't 'take it out' of profits--that's just another way of saying 'add it to cost of production'.

[Patrick]

"how Stock holders who collect divedends are any different than workers collecting salary"

Hmmm.... Inter-temporal trade and risk? Stockholders buy the rights to the dividends; they put up capital today in exchange for the prospect of cash flow tomorrow - but those cash flow may never materialize and if they don't, tough. On the other hand workers trade labour today for wages today, and if they don't get paid they can sue.

[Patrick]

One further though - I suppose you could make wages and dividends very similar if workers only got paid wages equal to MP once that MP actually materialized, and if it didn't, then they wouldn't get paid.

[Allan Manchester]

>For a given level of demand, an increase in MC will cause an increase in the firm's profit-maximising price. <
Here is your error. Prices are set to maximize profit based on selling conditions in the market. The high costs in a fledgling firm does not mean increased price, any more than much lower costs in a mature firm mean decreased price. Prices do tend to rise with a minimum wage hike, yes, but that is because the increased demand has changed the market conditions, and not at all because the costs have pushed up on the price.

[Allan Manchester]

>For a given level of output, which means a given level of income, if wage income takes a greater share, then non-wage income must take a lesser share. You need to argue that the distribution of income affects aggregate demand.<
Here you must discuss velocity. While the poor worker spends his money as soon as he gets it, the investor might wait a year before adjusting his cash levels and buy some certificates. All individual spending is good, but the problem is in how fast the consumer dollar gets back into the hands of business.

[Rick]

Patrick: The difference between wages and dividends is semantic. If raising wages increases marginal cost so does paying dividends.... Why can't a union own shares to account for raises as the company grows, and if they did how would that be any different to a wage increase? Would you then suggest that paying dividends is bad for the economy?

[Rick]

oh and further could you please tell me when 1992-2007 that Walmart, Starbucks, Tim Horton's, Macdonald's or any large retailer or service sector giant had its stock prices take a year over year loss.... and if it happened how long dividends were cut for? I heard they are real risky stocks?

[Allan Manchester]

I think you guys miss the point. Business money must ALWAYS be dispensed to individuals before it becomes consumer money to be dispersed back to business. Nick is correct that what isn't paid in wage will be paid to non wage individuals eventually. The difference is in the velocity of exchange, and this fact is often missed or dismissed by investigators. Not only will the poor worker get his money back to business quickly, but on the business side too, bi weekly wages get the money onto the consumer side much more quickly than quarterly or annual dividends do.

[Patrick]

Rick: Whoa there Tex! Just pointing out what seems to me to be the self evident difference between wages and dividends.

Lots of companies don't pay dividends for the reasons you mention, as well as the unfavorable tax treatment.

I have pretty unusual views about unions. I think unions actually have more in common with shareholders than is typically acknowledged, especially given what I think is the prevalence of looting, fraud, rent extraction, and general incompetence of management. After all, the best job security is a profitable company. So I think it would be a great idea for unions (or even better, workers themselves) to own big chunks of the companies their members work for. But then, they'd be more like mutual funds than unions ...

[Rick]

Allan: That's exactly my point, if a company is growing raising wages proportionately is the same as paying dividends to your employees who have an implicit ownership stake in the company, but is never recognized in practice.

Reward the shareholders, though there are probably few of them are left who invested in the company initially, who put up the investment capital and also reward the employees who produced, marketed, sold and reinvested their time and money back into the company proportionately. Everyone wins demand increases across the board from the those employees so long as the firm..... or in a broader sense, the economic sector the firm is in as a whole is growing.

That would have meant reducing manufacturing wages (through bargaining and arbitration) through the 90's proportionate to market share and foreign pricing and increasing service wages over that same period(legislative) relative to market share in the economy as whole. There never has to be a universally broad plan, but sector based plans will work raising and lowering wages as well as shifting labor so that recessions don't hurt so bad.

[Xevec]

"Here you must discuss velocity. While the poor worker spends his money as soon as he gets it, the investor might wait a year before adjusting his cash levels and buy some certificates. All individual spending is good, but the problem is in how fast the consumer dollar gets back into the hands of business."

Nope, that is still considered spending. Saving is simply another form of spending. There is no difference in the speed of which a poor person and a rich person spends their money. Do you really think that the money you put into a bank is not being spent? Do you think rich people just let their money sit there? That is a waste of money! They will put it in short-term bonds that will collect interest. That is spending money. Companies do their best to keep as little money not circulating. That is true in business and of all people. When you put your money in the bank, the bank uses that money for other things. They just don't let it sit there. So poor people spend just as fast as rich people, except it is easier to see poor people spending it than rich people.

[Jon]

An expansion of the minimum wage is no different from a contraction in the money supply. Its completely neutral in the long run because in a fiat economy its the only nominal-real peg, and therefore every price is actually in units of 'minimum-wage' labor.

In the short-run, an uncompensated increase in the minimum wage is a nominal shock. Far from shifting the AD right, it shifts it left--this has nothing to with the microeconomics of labor supply/demand.

You can understand this simply from the permanent income hypothesis. The sudden increase in prices confronts the embedded notional permanent (nominal) income. Real demand drops immediately. Notice that this the *opposite* of what happens when a price-level rises due to monetary expansion. Of course this is a direct manifestation of the price-level being a market clearing between supply and demand--and rather than a supply shift, we have a demand shift.

One reason why many of the studies on the minimum wage are contradictory stems from a failure to control for CB policy.

[Allan Manchester]

>You can understand this simply from the permanent income hypothesis. The sudden increase in prices confronts the embedded notional permanent (nominal) income. Real demand drops immediately. <
Except that in the real world, this never happens. Minimum wage hikes are always followed by strong economic growth, reflecting real increases to demand on all levels.
>One reason why many of the studies on the minimum wage are contradictory stems from a failure to control for CB policy.<
Yeah right. First the CB is not the main supplier of money. Money is a creation of the local bank. Secondly, common intuition is backwards, and CB feeds inflation by raising interest rates so making money creation more profitable, impairing money creation by the reverse. Third, excess money in the system only creates down pressure on the money's value, thoroughly failing to feed any real growth (as shown in the 70s)

[Arin Dube]

Nick says: "Starting in equilibrium with no upward pressure on prices or wages, any increase in aggregate demand, output and employment will put some upward pressure on some prices and wages. It doesn't matter how small it is; even if the average individual firm (worker) raises its price (wage) just a little, the positive feedback effect results in an infinite rise in prices and wages.

There is positive feedback because each individual firms and worker cares only about the real price and wage -- which means the price and wage relative to other prices and wages. Upward pressure on prices and wages means upward pressure on relative prices and wages. And it's impossible in aggregate for the average firm to raise its price and wage relative to the average price and wage. The attempt by each to do so results in an upward spiral of unlimited speed."

This is simply wrong. Minimum wage affects say around 10% of workers. If wages rise for these 10% of workers. it does not lead to an "upward spiral of unlimited speed." Instead it redistributes income from either (1) profits or (2) higher income consumers. (1) is particularly likely in the short term, before entry and exit occurs (and this is the relevant issue in the short term in the midst of a major downturn). Given differential propensity to consume (due to say liquidity constraints), it is easy to derive the result that this redistribution increases aggregated demand in a New Keynesian type model. I am left baffled by the chain of (il) logic in this post.

Maybe you can take another stab at your explanation ....

[Nick Rowe]

Jon: here's my take on your argument. I basically agree, but have a different way of thinking about it.:

Assume the money supply is fixed, so you get a downward-sloping AD curve in {P,Y} space.

Assume the minimum wage is set in nominal terms, i.e. is not indexed. Then you get an upwards-sloping Phillips Curve in the same {P,Y} space. (I know we normally draw a Phillips Curve in inflation space, but that is just the Fetish of the First Derivative; and I know we normally call that curve in {P,Y} space an Aggregate Supply Curve, but it isn't really a supply curve, if the economy is imperfectly competitive).

An increase in the nominal minimum wage shifts the PC vertically up. This causes a movement along the AD curve to a higher P and lower Y.

If the central bank increases M by the same percentage that the minimum wage increased, then AD shifts right, so P rises further, until the *real* minimum wage is back to its original level, and Y is back to its original level. Standard neutrality/quantity theory result, only with the minimum wage as the nominal anchor to the system.

But if the minimum wage is indexed, the PC is vertical, and an increase in the real minimum wage shifts the PC left, so it moves along the AD curve to a higher P and lower Y.

I agree that econometric studies of economy-wide changes in the minimum wage would need to be macro-based, and take the CB's reaction function into account. But changes in provincial minimum wages could maybe ignore macro effects.

[Nick Rowe]

Arin: "This is simply wrong. Minimum wage affects say around 10% of workers. If wages rise for these 10% of workers. it does not lead to an "upward spiral of unlimited speed.""

Yes it would. Remember, the Long Run Phillips Curve is vertical, assuming there are no long run nominal rigidities, and the expected inflation adjusts to actual inflation. And in the thought-experiment from which you quote me I am holding real aggregate demand constant (imagine a vertical AD curve, if you like, caused by the CB raising M in proportion to P). Suppose (say) 10% of workers get a (say) 10% increase in Wmin, and this causes the price level to go up by (say) 1% in the "first round" (direct effect, holding all other wages constant). But now (second round) all other wages must rise by 1% to get real wages the same as before, and the minimum wage, if indexed, goes up by another 1% too, and all other prices and costs must rise by 1%, etc. And so on.

The key thing to remember is that it is relative prices, not nominal prices, that matter to people and firms.

[Nick Rowe]

Allan: ">For a given level of demand, an increase in MC will cause an increase in the firm's profit-maximising price. <
Here is your error."

Are you saying that an upward shift in a firm's MC curve, holding its demand curve constant, will *not* cause a rise in the firm's profit-maximising price? ???

The only case I can think of where it does not cause a rise in price is where the demand curve is horizontal. But that's perfect competition, so the firm is already on its supply curve. And in that case the firm will cut quantity supplied, and if all firms in the industry do this, the excess demand will cause a rise in prices anyway.

By the way, what you say about velocity is basically right, at least in principle (I don't know if wage income *in fact* has a higher velocity, but it wouldn't surprise me if it did, and if it did, then distribution effects would affect AD.) Here's the standard way of saying what you are saying: take a standard (say, ISLM) theory of AD. Hold M constant. If there is a higher marginal propensity to consume out of wage income than non-wage income, or a lower income elasticity of the demand for money, then a shift in the distribution of income towards wages will cause the AD curve to shift right.

[Just visiting from Macleans]

Notwithstanding that there is some debate over the definition of poverty, can you explain the movement in Low-income cut-off (LICO) over the period of 1985-2005? Did changes in min. wage affect overall poverty levels in any positive or negative way?

[Peter T]

If you want to make the assumption that the key features of the model are what counts in the real world, you can draw these conclusions. I can't quantify the argument (so maybe it's not economics), but real workers have different stakes than those who derive income from profits. For instance, security of income is more important to wage workers - a fact employers can and do use. So starting wages signal where the responsibility between families and employers lie - low starting wages pass responsibility and risk back to families (anyone had a child work for MacDonalds?). Third, higher minimum wages work through to higher middle class wages, which seem empirically to work through to lower upper class incomes (so it's not the wages bill in total, but the low-end wage share of income). And this works through to incentives to produce things middle-class people want. If you want to see how the opposite plays out, consider the path taken first by the UK and over recent decades by the US - production of goods and services divides increasingly into cheap, low-quality stuff for the masses and low-volume high quality for the rich. Which works through to how people are educated, their political role and much else.

In short, a reasonable minimum wage is about signalling who the country is run for and by.

[Kevin]

I think we need to take this discussion from the theoretical, to the real.

Let's compare two real world businesses, in the same sector, with different 'minimum or entry-level wages' and see what effect can be discerned.

Walmart vs Costco

Both companies run big box format, general merchandise retail operations focused on the cost-conscious consumer.

Their product mix is not identical, but overlaps heavily, as do the markets they serve.

On Wages:

Most Walmart staff earn statutory minimum wage or within $1-2 per hour thereof, typically without any non-cash benefits.

Costo: Average wage of retail staff in 2007 was in the vicinity of $17 per hour, U.S.

Obviously Costco 'raised' the minimum wage for basic retail work. So what effect did it have?

On Profits:

Costco has a higher net profit per square foot and higher gross margin that Walmart. Profit has not been adversely impacted, arguably it benefited.

On Product Cost Inflation: Costco is fully price competitive with Walmart, there has been no discernible price pressure from the higher wages.

On Employment: Now here we do see a micro-effect, in that Costco is able to maintain fewer employees per square foot and per operating cost unit. In fact, total labour costs at Costco are equal to or less than Walmart on a comparable basis.

Why?

According to Costco, they focus on having a predominantly full-time work force, with employees of long tenure (turnover is a small fraction of what Walmart sees). The company indicates that this greatly reduces training costs; that's its employees being more experienced and knowledgeable are more productive and require much less supervision.

In addition, store shrinkage rates are lower, and customer satisfaction and brand image are higher.

AHA! you say, that's all well and good, but surely since Costo gets by with fewer staff that automatically means higher unemployment.

Well, certainly there is no evidence of that on the broader U.S. economy. But to be fair it would be hard to parse out the employment impact of one U.S. retailer that way.

I would suggest though that based on workforce composition at Costco it shows only a very slight decline in employment (total hours) for teenagers and virtually no other impact.

AS teens typically use this money discretionarily, this is not a socially undesirable outcome.

But wait still more, if Teen hours at Costco were cut by 1/2; but wages double (vs Walmart) this has no net adverse impact on take-home pay.

While the increase take-home pay of older staff who use the money as actual wage, results in a higher demand curve in society with that money reinvested in greater goods purchasing.

*****

In the end, undeniably, with enough of a minimum wage increase, you see some adverse employment impacts at the level of discretionary teen workers.

You also see some pressure to constrain costs elsewhere in a business, be that through labour-saving, reduced compensation to mid-level or high-level staff, or pressure on profit.

However, these pressures are largely minor, inconsequential and transitory.

As Costco establishes, higher (entry-level) wages do not need to adversely impact a company. Nor do they have to result in exorbitant pay hikes at higher levels within the organization. That's a choice of corporate leadership.

[Nick Rowe]

Peter T: "I can't quantify the argument (so maybe it's not economics),..."

I can't quantify (parts of) my argument either, but I'm still claiming it's economics! So don't worry about that.

I don't really understand your argument. Maybe you are invoking some sort of externality?

But my post here is concerned with one particular argument in favour of minimum wages: 1) that an increase in minimum wages will increase aggregate demand, 2) and that increase in AD will cause an increase in output and employment. I am saying that that argument is wrong. The first part may (or may not) be right, but the second part is wrong. And output and employment will actually fall, because an increase in minimum wages will shift the LRPC in the bad direction.

There may (or may not) be other arguments in favour of minimum wage laws, but they are outside the scope of my post.

[Nick Rowe]

Kevin: there is only theory; and what you are doing is theory too!

Interesting example, but:

1. Are you saying that Walmart is not maximising its profits? If so, why aren't they?

2. Given that Walmart is the way it is, if minimum wages increased, would Walmart (and other firms like Walmart who hire at or near the minimum wages) raise prices?

[Just visiting from Macleans]

I wouldn't consider Costco and WalMart to have the same business model, nor target demographics.

WalMart's success has been largely attributed to its supply chain management systems - and squeezing costs at every stage (from Chinese suppliers through to checkout). Very cost conscious customers - but moving up over time in terms of income levels.

I think of Costco (only been in one once) as more of a low service, buy in bulk (total cost important factor to who goes there), warehouse setup where you need a membership (barrier to entry), and you haul away your stash by suburban soccer Moms in their SUVs.

Still, it is worthwhile looking at different businesses - my outside view of much economics is that the theories are based largely on firms selling similar goods on the basis of cost- which Allan has pointed out (with mixed success), is not always the case.

[Xevec]

"

Kevin: there is only theory; and what you are doing is theory too!

Interesting example, but:

1. Are you saying that Walmart is not maximising its profits? If so, why aren't they?

2. Given that Walmart is the way it is, if minimum wages increased, would Walmart (and other firms like Walmart who hire at or near the minimum wages) raise prices?"

Actually Nick, that isn't true. The average wage for full-time employees at Wal-mart is way above minimum wage. I believe it is around $10 an hour.

"In 2008, the average full time Associate (34 hours per week) earns $10.84 hourly for an annual income of $19,165. That’s $2,000 below the Federal Poverty Line for a family of four. [http://aspe.hhs.gov/poverty/08poverty.shtml] "

"Except that in the real world, this never happens. Minimum wage hikes are always followed by strong economic growth, reflecting real increases to demand on all levels."

Then explain Samoa Allan. 2000 jobs were lost since the minimum wage increase in 2008. A major company which employed most of Samoa left them to go to georgia. The economic growth of Samoa now decreased. Unemployment did rise. Their minimum wage went from $3.25 an hour to $7.25 an hour(increasing 50 cents each year until reaching that point).

Or even explain Michigan, which you even admitted that after the minimum wage hike, unemployment levels actually continued to increase. You decided to blame outsourcing, but that is simply not true, since as you said, minimum wage hikes ALWAYS followed by strong economic growth.

But Nick, I believe I get what you are saying, but correct me if I am wrong.

You are saying the argument that increasing minimum wage is good because it increases AD which then increases output and employment. In other words, more spending and more purchasing of goods increases production and highers more people. You are saying that this is incorrect. That simply increasing overall spending does not necessarily mean more increasing of employment and increasing production.

But I will also agree with Nick that wal-mart and Costco do not operate both "big box" style. You will need to compare Wal-mart to say "Jewel/albertson's" or Dominicks. If you want to do a real comparison to Costco, look at sam's club, which is the big box store owned by wal-mart.

[Nick Rowe]

Xevec: I don't know if Walmart pays minimum wages. But it may depend on whether we are talking about the US or Canada. Canada (I think) generally has higher minimum wages than the US (relative to what, I should ask myself), and different provinces set different minima.

My argument is that an increase in minimum wages might increase AD. If AD increased, and it didn't merely cause inflation, that would increase output and employment. But an increase in AD, past the level of output defined by the LRPC, *will* only cause inflation (in the long run). Worse, an increase in minimum wage will reduce the level of output and employment defined by the LRPC. So an increase in minimum wages will lower output and employment.

[Just visiting from Macleans]

Paying above min. wage can be a financial deterrent to unionization - something WalMart avoids at all cost.

[Allan Manchester]

>Are you saying that an upward shift in a firm's MC curve, holding its demand curve constant, will *not* cause a rise in the firm's profit-maximising price? ???<
Absolutely! Look at the firm's history. Most fledgling firms operate for months or even years at a loss before building up sales well enough that they are showing profit. The total lack of MR does not influence the pricing at all because the pricing is totally dependent on market conditions. Of course the demand curve is never constant, and most firms are built on hope against history of ever increasing demand.

>But an increase in AD, past the level of output defined by the LRPC, *will* only cause inflation (in the long run).<
One might think that an increase to the foundation price would be inflationary, but it isn't. The increase to the velocity means that the increased cost just gets swallowed by the increased economic activity.
But the increased economic activity will always cause inflation. Increased purchases mean increased borrowing against those purchases, and increased business means increased borrowing to expand businesses. The only way to increase the money supply is through increased borrowing, and this activity will always lead to some inflation. Increases to the money supply will always mean an immediate drop to the money's value. The only value in money is its scarcity. Too, as the economy heats up, the same dollar is supporting more and more transactions, and the supply might well seem to have expanded when it has not. This doesn't seem to be too inflationary, but I'm not sure it isn't.
Some inflation is a normal part of a healthy economy. It is impossible to have any economic growth without some inflation, and there is nothing that can cause only inflation except bad banking practice.

[Nick Rowe]

Allan: think about what Marginal Revenue and Marginal Cost mean.

If an increase in current output reduces the costs of future output (because practice makes perfect, for example), that effect should be included in MC.

If an increase in current output and sales increases future demand (because you get repeat customers once they have tried the product, for example), that effect should be included in MR.

Bringing in these effects just complicates the story, and makes it harder to understand what's going on.

[Lord]

What did you think of Kaletsky's showing the Phillips curve for the UK has been horizontal for the last 20 years? Another issue is whether a higher minimum would change the CB's response. I'm inclined to believe it would at the margin though that is because I see the economy operating below optimal most of the time. When Ford instituted it's %5 day to reduce turnover leads me to believe business is often wrong about what maximizes profits.

[Allan Manchester]

>Bringing in these effects just complicates the story, and makes it harder to understand what's going on.<
Certainly, and it just gives an added spread to the MR/MC curves. The MR is upsloped, and the MC is down sloped, and the business is out to maximize sales only and not to trim sales to a maximized profit.

[Nick Rowe]

Lord: If you have an inflation targeting central bank, the Phillips curve should appear to be roughly horizontal (at the 2% target). That's because you should never observe the long run exogenous changes in monetary policy that you would need to observe in order to econometrically identify the LRPC. If the inflation target is credible, then by definition expected inflation is 2%. So we only observe a short-run Phillips Curve anyway, which may be close to horizontal in the very short run. And if the Bank furthermore can spot some shifts in the SRPC, and react against them, to keep inflation on target, the estimated SRPC gets flatter still.

Under 2% inflation targeting, deviations of inflation from 2% should be uncorrelated with anything in the central bank's information set, which will include (lagged) unemployment.

Here's my related post on the topic: http://worthwhile.typepad.com/worthwhile_canadian_initi/2009/01/why-theres-so-little-good-evidence-that-fiscal-or-monetary-policy-works.html

To think about the same thing in econometric terms, the actual observed correlation between inflation and unemployment is determined by the intersection of two curves: the Phillips Curve; and the CB's reaction function. This is why econometricians need to do simultaneous equations methods. Kalecki may think he's estimating a Phillips Curve; in fact he's probably estimating the bank of England's reaction function, and has learned, amazingly, that the BoE is targeting 2% inflation!

Under inflation targeting, every observed curve should be flat. Inflation should be unforecastable.

I should probably do a post on this someday.

[Nick Rowe]

And business is almost always wrong in knowing what maximises profits. They don't have a crystal ball. Does the authority that sets the minimum wage know better? And is it willing to tailor a precise minimum (and maximum) wage for each business?

[Allan Manchester]

> And is it willing to tailor a precise minimum (and maximum) wage for each business?<
Oh, they would love to do that. Government is always looking for excuses to grow bigger. Of course many European countries have a sector by sector minimum wage (with the huge tax burdens caused by overgrown government).
You don't need more than one foundation price for the economy to work well, and you certainly do not need or want any kind price ceilings.

[Nick Rowe]

Allan: " The MR is upsloped, and the MC is down sloped, and the business is out to maximize sales only and not to trim sales to a maximized profit."

If the MR curve slopes up, and the MC curve slopes down (locally), then the firm is indeed not maximising profits; it's minimising profits. It's carefully making the biggest losses (or smallest profit) that it can.

Look. I really think you need to read a basic micro textbook.

[Lord]

I agree that removing inflation from the mix should flatten the PC, what is interesting is how variable employment is anyway even without it. I expect much of this is due to asset fluctuations that are uncounted.

[Allan Manchester]

>It's carefully making the biggest losses (or smallest profit) that it can.<
No it isn't. The sales are not that responsive to price. Any marketing guru will tell you that. You want the sales to give you profit, but if it doesn't this month, maybe next month. You set your prices according to the market, and if you cannot perceive that you can profit on the going market price, then you aren't going to go into business. Your profits are never going to achieve your expectations, because of unforeseen costs, but once the profit is going on, you do not try and trim your sales by upping your price. You might well try and up your sales by upping your price, but will not up your price beyond where you feel that the total sales will falter. In the meantime, your costs are going down, as you achieve better knowhow, and trim up waste. Prices are not set on costs, at all! Your insistence that they are stems from a lack of business experience, and a subjection to propaganda from business, that 'we only put the prices up because we have to'.

[Nick Rowe]

Lord: " I expect much of this is due to asset fluctuations that are uncounted."

You mean "uncounted" in the CPI? Hmm. Could be. So that "true" inflation (i.e. the one that should be there in the Phillips Curve) is fluctuating more than we think it is when we look at just the CPI. Good point.

The one thing that really puzzles me about the UK Philliops Curve is the recent inflation, despite high unemployment. Might be an exchange rate effect? (Sterling has depreciated recently).

[Allan Manchester]

>The one thing that really puzzles me about the UK Philliops Curve is the recent inflation, despite high unemployment.<
Inflation is due to growth in the money supply period. You will never be able to pin inflation to any other factor. What is usually missed in looking at the money supply/inflation is the fluctuation in the multiplier effect. Money is a product of the local bank, and the CB fails to understand that in its attempts to "stabalize" inflation.

[Bruce Wilder]

"An increase in minimum wages will cause a firm's Marginal Cost curve to shift up."

A fact not in evidence as they would say in the courtroom.

What a change in the minimum wage will alter for sure is the marginal productivity of the minimum wage worker, but I don't think you can draw any conclusions about the effect on the marginal cost curve for the firm's product, if the firm employs workers at a variety of wages. A change in the minimum wage may alter the distribution of risk, as well as income, among various factors employed by the firm.

Without knowing something in particular about the firm's production technology and cost structure, it is really difficult to predict the effect of a change in the minimum wage.

"An increase in the minimum wage will raise workers' incomes, which will increase aggregate demand, which will increase output and employment. So would an increase in union wages."

One way to interpret this somewhat loosely phrased sentiment is as the assertion that production technology constrains output as a function of the distribution of income and risk. Mass production (with increasing returns) entails mass consumption. The production technology wants to be egalitarian.

[Patrick]

"What a change in the minimum wage will alter for sure is the marginal productivity of the minimum wage worker"

I don't know about that. If you pay them more and they have more disposable income, their leisure becomes worth more because they have more options for entertainment etc. It might have the opposite effect. Better to just adjust the lighting now and then.

[Nick Rowe]

Bruce: I'm sure mathematical micro theorists `might have a simpler way of approaching this, but here's my take:

Leave aside the case of monopsony, where i agree that an increase in the minimum wage ,as long as it's not "too big", will cause the MC curve to shift down.

In the one input case, it's simple. The MC curve is defined by W/MPL, (MPL means Marginal Product of Labour) so an increase in W will increase MC for a given L.

In the multi-input case, where the firm rents Kapital at wage C, and Land at rental l, etc, the firm will choose a cost-minimising mix of inputs where:

W/MPL = c/MPK = l/MPland = etc.

(This is like saying it chooses a tangency point on its isocost curve where the ratio of the input prices equals the Marginal rate of Technical Substitution (which equals the ratio of the marginal products)).

So if W rises, the firms will readjust the combination of labour, land, kapital, hi-skilled labour, etc., (substituting away from the minimum-wage labour and complementary inputs into some of the other inputs) until those ratios are all equal again. But that must raise all those ratios, and by the same amount. So MC must rise for a given level of output.

Where's a micro-theorist when I need one? Isn't there some accursed lemma that proves the same thing? Shepherd's? It's 30 years since I did micro.

[Just visiting from Macleans]

Better to just adjust the lighting now and then.

Why did the chicken cross the road? No CFL bulbs, and better benefits.

[Xevec]

"No it isn't. The sales are not that responsive to price. Any marketing guru will tell you that. You want the sales to give you profit, but if it doesn't this month, maybe next month. You set your prices according to the market, and if you cannot perceive that you can profit on the going market price, then you aren't going to go into business. Your profits are never going to achieve your expectations, because of unforeseen costs, but once the profit is going on, you do not try and trim your sales by upping your price. You might well try and up your sales by upping your price, but will not up your price beyond where you feel that the total sales will falter. In the meantime, your costs are going down, as you achieve better knowhow, and trim up waste. Prices are not set on costs, at all! Your insistence that they are stems from a lack of business experience, and a subjection to propaganda from business, that 'we only put the prices up because we have to'. "

That's not true Allan. Some products are way more price sensitive than others. Products like orange juice are very price sensitive, much like oil. You don't always set your prices according to the market. What if there is no market prevalent? I mean, my friends opened up a hawaiian style restaurant with live shows every friday and saturday. There was no other restaurant in the city or suburbs that had a similar style.

Sales(or demand, or consumption) are reactive to prices. People change their behavior when prices rise a certain amount. Take oil prices for example. Public transportation has seen an increase. SUV sales and cars with low gas mileage have seen a drop in sales. Hybrids and green technology is seeing an increase. This is all because of an increase in price for oil.

[Xevec]

Nick: Well, when I gave my statistics, I was strictly speaking of the US. The federal Minimum wage of the US at the time of those wages was $5.15 an hour. Of course, each state is different when it comes to minimum wage, some states not having a minimum wage at all(like alabama) and some states having high minimum wage(like illinois and california).

"But the increased economic activity will always cause inflation. Increased purchases mean increased borrowing against those purchases, and increased business means increased borrowing to expand businesses. The only way to increase the money supply is through increased borrowing, and this activity will always lead to some inflation. Increases to the money supply will always mean an immediate drop to the money's value. The only value in money is its scarcity. Too, as the economy heats up, the same dollar is supporting more and more transactions, and the supply might well seem to have expanded when it has not. This doesn't seem to be too inflationary, but I'm not sure it isn't."

No, that is not true. Increased business activity does not necessarily lead to inflation. Inflation is caused by an increase in the money supply, yes. But that is not the full story. Inflation can only occur when increasing the money supply is higher than increases in actual goods.

Also Allan, take for example natural disasters. Do you think upping the prices is simply what some people called as "price gouging?" It would be since as you said, upping prices does not discourage sales. So if an industry, or even a monopoly of a certain good, they can simply raise prices infinitely, and nothing like the subsitution effect, or anything else would slow down their sales. As you said Allan, sales are not affected by price. That oil can be as expensive as a company wants it to be, and people will continue to buy it no matter what price it is at.

[Peter T]

Yes I was invoking various complex externalities - but there almost always are complex externalities, and there are lots around the minimum wage. The comments point to some.

I presume that you are not making a theoretically interesting point, but an argument directed at policy. If so, ignoring the externalities is hard to justify.

[Stephen Gordon]

What are these externalities?

[Peter T]

Stephen

Well, for a start, wages are rarely an arms-length negotiation, and in practice are not all that flexible. Workers in general are constrained by family needs, location, a sense of their relative power and their lack of negotiating skills relative to employers (who can, after all, employ specialist negotiators and people skilled in the relevant law). Why do you think employers (and unions) put time and money into lobbying for changes to employment law? Employers are often constrained by law, and sometimes by community attitudes.

Minimum wages are often either starting wages for the young, or wages for the least advantaged (eg migrants, unskilled women). Setting them too low pushes families or other informal support networks into subsidising the wage (eg pizza delivery where the employee is using the family car). Surely this counts as an externality?

While a pure theoretical approach would lead to the conclusion that wages cannot be below subsistence, lots of studies show that employed people in these groups in the UK and the US can have negative income - their wages do not cover their costs of living. The difference is made up by charity, the state, family and so on.

The largest externality probably arises from the fact that humans - like other closely-related social species - are very sensitive to relative status. There are well-charted links between relative status, stress, cortisol levels and general health. In brief, lower status people are sicker, die younger and have less healthy children (Wilkinson's Mind the Gap is a recent summation on humans, but the same findings have been made for baboons and other social primates). Wages are a signal of status, but also an enabler of status in non-monetary areas (hard to be the best backyard chef in the neighbourhood if you can't afford a barbecue). So the minimum wage level affects a large group in lots of non-obvious ways. If it's below what's needed to maintain the social decencies, this has large effects. I believe Adam Smith made this point - wages needed to be high enough for a working man to afford shoes in England, but not in Scotland, where going barefoot was not a mark of extreme poverty.

That enough?

[Nick Rowe]

Allan: "No it isn't. The sales are not that responsive to price. Any marketing guru will tell you that."

If sales (quantity demanded) were extremely responsive to price (as under perfect competition), then the firm's demand curve and MR curve are both horizontal (and MR=P). As sales get less and less responsive to price (as we move towards imperfect competition), the demand curve gets downward-sloping and steeper, so does the MR curve, and MR becomes lower relative to P. So what you are arguing here totally contradicts what you said earlier about the MR curve being upward-sloping.

Again, read a basic micro textbook.

Peter T. If an increase in minimum wages causes an increase in unemployment, that would also cause a negative externaility.

There is a big externality in the sort of model I'm talking about. It's called the Aggregate Demand externality. Any increase in any price or wage causes a reduction in the real level of Aggregate Demand that is sustainable. The solution to that externality, in principle, would be to impose *maximum* wages and prices, not minimum wages or prices.

In principle, maximum wage and price controls, allowing the central bank to increase AD without triggering inflation, can make people better off.

Wage and price controls can work *in principle*, at the macro-level, in this sort of economy with imperfect competition. The trouble with doing so is at the micro level. Because there are lots of relative price changes needed that this sort of policy would not allow.

[Jon]

Nick writes: "Jon: here's my take on your argument. I basically agree, but have a different way of thinking about it."

Okay, I'll believe that. Always fascinated by your deftness with model isomorphism.

Nick writes: "I agree that econometric studies of economy-wide changes in the minimum wage would need to be macro-based, and take the CB's reaction function into account. But changes in provincial minimum wages could maybe ignore macro effects."

I didn't mean to argue to the contrary. My point was only to state that my narrative was consistent with there being a broad range of results from econometric studies. Neither of our arguments really holds for a minimum-wage change applied to a small region within the larger economy. That's a purely micro-economic question, and its clearly not prescribed by the LRPC or LRAS given the mobility of capital and people in that context.

[Allan Manchester]

>As sales get less and less responsive to price (as we move towards imperfect competition), the demand curve gets downward-sloping and steeper, so does the MR curve, and MR becomes lower relative to P. So what you are arguing here totally contradicts what you said earlier about the MR curve being upward-sloping.<
If I charge coffee at 50 cents and sell 1000 cups and then sell coffee at $1 and sell 990, I get an upward sloping MR do I not? If by using the extra revenue for proper marketing I sell 2000 at $1 then the MR slope is definitely up. MR is not down sloped. Prices are not based on cost at all. Your insistence that they are stems from a lack of business experience and a subjection to business propaganda that, "We only raise the price because we have to". Business raises the price because it can. It only lowers the price when it has to.

[Mike Moffatt]

"If I charge coffee at 50 cents and sell 1000 cups and then sell coffee at $1 and sell 990, I get an upward sloping MR do I not?"

No. You don't. Assuming a linear demand curve, using the points you describe you have a demand curve of Q = 1010 - 20P. This yields a marginal revenue curve of MR = 1010 - 40P. Which you'll note is downward sloping.

[Allan Manchester]

p vs qp
Revenue is not a q

[Nick Rowe]

Arithmetic error Mike!

Assuming a linear demand curve you get:
Demand curve: Q=1010-20P (as you say)
Invert it as: P=50.5-0.05Q
Total Revenue curve: TR=PQ=50.5Q-0.05Q^2
Take derivative wrt Q:
MR=dTR/dQ=50.5-0.1Q

I think I got that right.

(You forgot to invert the demand function).

But yes, the MR curve is downward-sloping.

Allan:
And, regardless of the linearity or otherwise, the elasticity of demand in your example is less than one over the range $0.50 to $1.00, so MR is negative. That firm would certainly raise price, because it would increase Total Revenue (plus, presumably, lower Total Costs) if it raised price.

A change in marketing is a *shift* in the demand curve (and a shift in the MR curve), not a movement along the curve. It says nothing about the slope of the MR curve.

[Mike Moffatt]

"(You forgot to invert the demand function)."

Whoops! That's what happens when I also forget my afternoon coffee.

[Nick Rowe]

Allan: "Prices are not based on cost at all. Your insistence that they are stems from a lack of business experience and a subjection to business propaganda that, "We only raise the price because we have to". Business raises the price because it can. It only lowers the price when it has to."

This shows a total misunderstanding of economics.

Economists certainly do not believe that "We only raise the price because we have to". Economists believe that firms raise price whenever doing so would increase the firm's profits. And whether or not an increase in a firm's price raises or lowers profits depends on whether MR is less than or greater than MC.

OK: here's micro 1000.

A firm faces a demand curve. The demand curve shows the relation between Price and Quantity demanded, other things equal. The firm chooses that point on the demand curve at which profits are maximised. That point is determined by MR=MC. A small (one unit) increase in Q (cut in P) will raise its Total Revenue by MR (by definition of MR), and will raise its Total Cost by MC (by definition of MC). If MR is greater than MC it will increase TR by more than TC, and so raise profits by increasing Q, and will keep on raising Q (cutting P) until it reaches a point where MR=MC. Conversely, if MR is less than MC, it will cut Q, and keep on cutting Q (raising P) until MR=MC.

[Nick Rowe]

Continuing: so the MR *and* MC curves determine the profit-maximising Q. And the demand curve then tells you the Price you need to get that profit-maximising Q. So, in general, the profit-maximising price depends *both* on the demand curve (which determines the MR curve), *and* on the MC curve.

The only case (I can think of) where a firm's MC curve has no effect on P is where the demand curve is horizontal. (And even in that case, which is normally due to perfect competition, an upward shift in the MC curve of *all* firms in the industry will cause an increase in the market equilibrium price, and shift each firm's horizontal demand curve vertically upwards, and so still raise price.

It's Marshall's scissors. Both blades of the scissors do the cutting. Price depends both on the demand curve and on the cost curve. That makes sense, because profit is Total Revenue (which depends on the demand curve) minus Total Cost (which depends on the cost curves). So a firm choosing the profit-maximising price will look both at the demand curve and at the cost curve.

[Jon]

I have a fair bit of experience in business, both as an accredited investor and in management. The success of a firm is driven by the numbers. Everyone in management, in sales, etc, knows the costs. You have to know it to start negotiating anything. You guess at your vendors costs, and your customers guess at yours. Planning revolves around costs. Everyone computes the numbers and drives the bargain end-to-end.

One point though--perhaps its because of I've been involved with growth industries and small firms--the cost curves seem much flatter that anyone draws them in school. The most significant up-sloping factor within the firm tends to be organizational structure. One guy can do a 100, but 10 guys don't make 1000. First, the 'one guy' is usually exceptional and second 1 of those 10 guys is supervising, 1 is in some other support role...

Maybe if I was COO at McDonalds, I'd see the upsloping costs of more potatoes! I'm not so sure. I think this reflects that even the largest firms are quite small relative to the whole (world) economy.

Otherwise its much more common to see substantial returns to scale.

The fascinating part--to me--is that many of these statements are wrong at the macro-economic scale.

[Nick Rowe]

Jon: very interesting comment. My working hypothesis is that individual firms' MC curves are roughly horizontal, and MR curves downward-sloping.
At the macro-level, my working hypothesis as that the (real, inflation-adjusted) "Macro-MR curve" is roughly horizontal, and the "Macro-MC curve is upward-sloping.

By slope of "Macro MR curve" I mean d^2TR/dydY + d^2TR/dydy evaluated where y=Y, where y is individual firm and Y is average firm. In words, what happens to an individual firm's real MR when it and all other firms expand output by the same percentage. If the individual firm expands output by 1% its MR falls, but when all other firms expand output by 1% its MR rises again, and by roughly the same amount. Because the relative price stays the same when all firms expand output by the same %, and so does elasticity of demand.

[Allan Manchester]

>The demand curve shows the relation between Price and Quantity demanded, other things equal. The firm chooses that point on the demand curve at which profits are maximised.<
That is your theory, and whoever wrote it, and whoever believes it, has no experience in business. You choose your price on a gut level to try and maximize sales. If it does not sell at the price given it, you throw it out or slash prices well below costs to clear the shelf for another item. If it IS showing good sales at the price selected and so giving you profit, you might well raise the price some to test the elasticity of the demand. If you find by raising prices that your dollar take is increased, you will probably do it again.
The manager seeks to maximize the revenue and minimize the costs. With experience he gets better at both goals, and with any experience he will know that the two goals are not connected at all.

> Everyone in management, in sales, etc, knows the costs. You have to know it to start negotiating anything.<
Of course you do. That is the base line, but you don't sell anywhere near that line or you'd get fired. You try and sell as far away from the costs as you can get away with. The costs certainly do not dictate the price.

>The fascinating part--to me--is that many of these statements are wrong at the macro-economic scale.<
Well if the chalk board chiefs can't get it right on the micro, how are they going to understand the macro.

>Maybe if I was COO at McDonalds, I'd see the upsloping costs of more potatoes!<
No. MC is always down sloped. Prices go up over time due to inflation, but that is a shift of the curve, and not a movement along the curve.

[Allan Manchester]

What I am trying to tell you, and what every salesman will tell you, is that prices are not fixed by a calculator or a study graph. You sell for as much as you can, and if the circumstances of the negotiation are such, you have a huge latitude on which to play with. Your theories just do not coincide with the real world.

[Patrick]

Allan:

"You sell for as much as you can"

Some things to think about:

- Why do firms offer volume discounts?
- Why would management not hire infinitely many sales reps? Put another way; if a firm was to considering hiring one more sales rep, can you think of a reason why they might decide against it?

[Allan Manchester]

You argue theory against reality. There is a saturation point to the market, and you cannot sell more product regardless what you do to the price or what costs you incure in marketing. You don't try. If the thing doesn't sell at the set price, you scrap it and move to the next item. You do not try and dither an extra penny return by adjusting the price. It sells or it doesn't and that is all there is to it. Only when it is selling do you dither the price some, and that is almost always up!
And firms offer volume discounts because they are trying to maximize the sales return, and like it better to be getting 50,000 on a hundred items than 20,000 on 10 items. The costs are not a factor on the pricing game.

[Bob Smith]

"The costs are not a factor on the pricing game."
Sure, that's why I can down an buy myself a Porsche for $1.27 and some pocket lint.

[Allan Manchester]

>Sure, that's why I can down an buy myself a Porsche for $1.27 and some pocket lint.<
No, but if you were buying 3 you might get one free with proper negotiation. They won't sell it without making a profit. If it doesn't sell they will NOT whittle the price down until it sells but will scrap it, and in a specialty shop like that, they'll scrap the whole business.

[Patrick]

"firms offer volume discounts because they are trying to maximize the sales return"

How do they calculate sales return?

[Patrick]

... and you didn't answer my second question.

[Xevec]

"That is your theory, and whoever wrote it, and whoever believes it, has no experience in business. You choose your price on a gut level to try and maximize sales. If it does not sell at the price given it, you throw it out or slash prices well below costs to clear the shelf for another item. If it IS showing good sales at the price selected and so giving you profit, you might well raise the price some to test the elasticity of the demand. If you find by raising prices that your dollar take is increased, you will probably do it again.
The manager seeks to maximize the revenue and minimize the costs. With experience he gets better at both goals, and with any experience he will know that the two goals are not connected at all.

> Everyone in management, in sales, etc, knows the costs. You have to know it to start negotiating anything.<
Of course you do. That is the base line, but you don't sell anywhere near that line or you'd get fired. You try and sell as far away from the costs as you can get away with. The costs certainly do not dictate the price."

What you seem to be trying to explain is how car dealerships work, and then apply it to all business. Yes, this is true about car dealerships, in that you NEVER pay the sticker price because it is way overpriced. Costs certainly do dictate price for the individual. If it costs you overall $50 to make one widget, you can not sell it for $30 even if everyone else is.

You can maximize revenue by minimizing costs. If you look at your cost structure, you can increase your profits by simply lowering your costs. In other words Allan, you are saying in order to maximize revenue, you don't look at how much you are spending on goods. That is probably the dumbest way in order to run a business. Yes, let us not look at the books and see where are money is going. That doesn't in any way hurt or help our revenue.

"What I am trying to tell you, and what every salesman will tell you, is that prices are not fixed by a calculator or a study graph. You sell for as much as you can, and if the circumstances of the negotiation are such, you have a huge latitude on which to play with. Your theories just do not coincide with the real world."

Your theory doesn't coincide with the real world, because as a sales person, you don't do that. Most sales people have very little to play with in the negotiations in the first place. They are bound to whatever price of the product is according to the company they work for. If they are selling AT & T cable service, they can not say "you are paying $500 a month for it." Even if the person accepts those charges, AT & T will not allow it, because you can not change the prices of the product that is given to you.

"No, but if you were buying 3 you might get one free with proper negotiation. They won't sell it without making a profit. If it doesn't sell they will NOT whittle the price down until it sells but will scrap it, and in a specialty shop like that, they'll scrap the whole business."

Hmm, that is why all technology sales do that. Yes, think about when DVD's first came out. No middle-class or poor person in their right mind would go out and buy one. It is only when they lowered the price, is when more and more people decided to buy the product. The only people who were buying it at that high of a price were the nerds who want everything new, and the rich people who wanted everything new. Those people do not represent the majority of the market.

And businesses sell stuff all the time without making a profit on it. I managed to get a strategy guide from best buy for only 1 penny. This is because it is a policy of the store that they can not order new stuff without getting rid of the old stuff. Even with older model of cars, they try to sell it as fast as they can in june when the new models come out. They will even go as far to take a slight loss on the car if it means it gives them room to put out the new car.

Now, you already said all this Allan.

"If it does not sell at the price given it, you throw it out or slash prices well below costs to clear the shelf for another item"

But then later, you say this:

"They won't sell it without making a profit. If it doesn't sell they will NOT whittle the price down until it sells but will scrap it."

If they will not sell something below cost, then why slash prices WELL BELOW COSTS?!

[Xevec]

"Of course you do. That is the base line, but you don't sell anywhere near that line or you'd get fired. You try and sell as far away from the costs as you can get away with. The costs certainly do not dictate the price."

Actually Allan, they do dictate the price. You just said it yourself. If they didn't, then I wouldn't need to know the base line to begin with.

[Bob Smith]

Allan said: "They won't sell it without making a profit"

How can you say costs don't matter then turn around and say that a sale wont't happen unless the vendor makes a profit? Since profit is, by definition, revenue minus costs, on your account, costs matter. If you increase costs, on your account, some sales that might otherwise have occured won't happen because the vendor won't make a profit.

But, hey, can you tell me where I can get four porsches for the price of three?

[Allan Manchester]

>But, hey, can you tell me where I can get four porsches for the price of three? <
Anywhere they sell them. Just negotiate.

>How can you say costs don't matter then turn around and say that a sale wont't happen unless the vendor makes a profit?<
The profit is wholly elastic. The price isn't according to how much will be made on the sale, but whether or not the sale will be made. Good if I make 10, better if I make 20, bad if the buyer goes away. Good if I sell 1 and make 10, better if I sell a hundred and make 20, best if I sell a hundred and make 1000. And with markups easily up to 85%, how many you can sell is much more important than how much each cost you, or how much you made on each unit. Too, the MC is down sloped, and you can replace the 100 cheaper per unit than you can replace the single unit.

[Bob Smith]

Let me see if I understand what you're saying, the price is set at whether or not the sale will be made. So are you saying that a vendor will always want to make a sale even if its less than cost? That's clearly nonsense. I'd happily buy a porsche for $1.27, but I don't see anyone beating a path to my door. If the buyer isn't willing to pay you at least marginal cost, it's not bad if he goes away, it's good. Any sale would either cause a loss to you, or would have an adverse effect on the customers well being (because he'd be paying more than he values the goods or services.

And if the price is constrained at the bottom by cost, you've conceding that costs matter.

[Jon]

You write: "You choose your price on a gut level to try and maximize sales. If it does not sell at the price given it, you throw it out or slash prices well below costs to clear the shelf for another item."

Of course there is a difference between a going concern and a liquidation. While going firms sometimes have to liquidate--that's risk, and in the net, that raises the overall cost-basis--a continuing business recovers it elsewhere. I think you'll find that even through some 'speculation' results in a loss that is not a MR/MC sort of thing. After all, once you discover the loss you surely don't supply one-more unit. You liquidate.

But in aggregate those losses are built into the 'margin' over BOM on other salable items. That's the risk return to capital at work.

The return to capital is a component of cost. This is one of the clever little linguistic ways that economics and business do not speak the same language.

Consequently, although you may observe in your words, "The profit is wholly elastic. The price isn't according to how much will be made on the sale, but whether or not the sale will be made. Good if I make 10, better if I make 20, bad if the buyer goes away. Good if I sell 1 and make 10, better if I sell a hundred and make 20, best if I sell a hundred and make 1000. And with markups easily up to 85%"

This is a one period analysis. If the sum of those 'profits' do not cover the return to capital beyond the immediate costs, the firm ceases to exist. So I disagree entirely with your conclusion. Price discrimination is a great way to make money but looking at the aggregate, forget about temporal ordering--which deal do you turn down? Its the marginal one. i.e, MR=MC.

[Nick Rowe]

Allan: " Too, the MC is down sloped, and you can replace the 100 cheaper per unit than you can replace the single unit."

That sounds like a downward-sloping ATC curve, rather than a downward-sloping MC curve.

Yep, if AVC and MC are horizontal, but there is a fixed cost of producing a batch, the ATC curve will slope down.

[Allan Manchester]

> After all, once you discover the loss you surely don't supply one-more unit. You liquidate.<
No. If it sells then you keep going until the sales volume is sufficient for profit. If it doesn't sell, then you move on to the next item. How much is made or lost today is moot. Whether or not the business is growing is the question.
> So are you saying that a vendor will always want to make a sale even if its less than cost? <
Of course the wendor wants to make a profit, but often the atc is way beyond the mc, so if he sells enough, he'll get out of the red.
Say the store incures $20/hr in fixed costs. You buy coke for .40/can. You sell it for .80/can. You sell one in an hour, you've lost 19.60. You sell 100 in an hour, you've made $20. You sell 200 in an hour, you've made $60. Just because you've gone through many hours of loss does not mean that you give up and go home. Now say a hike to the minimum wage boosts the store costs to $25/hr. You sell one in an hour, you've lost 24.60. You sell 100 in an hour, you've made $15. You sell 200 in an hour, you've made $55. You set your price to get your sales, and in the end, you just don't worry about how many losing hours you've had, or how much this thing is costing you, you just push for more sales.
You are trying to tell me, that if I sell for .75, that the increased sales will more than cover the lower per sale profit. As long as I'm losing money, I really don't care, and, as long as I'm making money, I really don't care. The price is set on a gut instinct as to how much you can sell it for. If you can sell one, you can sell two hundred, but if your instincts were wrong, you just move on to the next item.

[Allan Manchester]

Now say someone comes in and says, "I'm buying 200 cans of coke, but you'd better make me a deal or I'm out of here" What do you do? You could go to .45/can, and still have a losing hour, or you could say .50/can just to cover the hour's cost, or .55/can to show some profit. But what if he's only looking for say .75/can? You have a chance to make money, and you'll try and make as much as you can. How much it costs you to keep the store open plays no part in this pricing game. You just need to keep the client happy.

[Nick Rowe]

Allan: in your example the MC of selling a can of Coke is $0.40, and the MC curve is flat. The Average Total Cost Curve is the one that is downward-sloping.

Also, an increase in the minimum wage in your example increases Fixed Costs, and so Average Total Cost, but does not affect MC.

You are right that a change in Fixed Costs does not affect Price (except that an increase in Fixed Costs may cause some firms to make losses, exit the industry, and so reduce the number of firms left, which can increase demand for sales at the remaining firms, and that may affect Price.)

This works if the sales assistant is underemployed, so you can sell more Cokes without hiring more sales assistants. (Up to some point, where the MC curve rises because you have to hire more sales assistants if you want to sell more Coke). If the elasticity of demand for Coke at your store is e, then MR=(1-1/e)P. So setting MR=MC to maximise profits you charge a Price P=[1/(1-1/e)]x$0.40

[Allan Manchester]

>If the elasticity of demand for Coke at your store is e, then MR=(1-1/e)P. So setting MR=MC to maximise profits you charge a Price P=[1/(1-1/e)]x$0.40<
I am telling you that I don't give a rats ass as a businessman. Either I am making money or I am losing money but I THOUGHT I could sell it at that price and I AM selling it at that price, and now that the price has been set, I just need to get the volume. Once I have the volume and am making profit, THEN I might monkey around with the price of it, and I won't be trying to make less per can!

>This works if the sales assistant is underemployed,<
There is ALWAYS slack in the system. There is ALWAYS time in the hour when the cashier is doing nothing. Making the client wait in line might mean losing a client, and so you might put on another cashier, but this is a marketing cost, and not something that can be added to the MC. Yeah, if you're selling 800 cans of coke an hour, you'll need another cashier, but by that time, the hourly cost of running the store is nowhere in the picture.

[Nick Rowe]

And all of the above (except the math in the formula) is micro1000.

[Allan Manchester]

Oh, you teach that the businessman doesn't care about your formulas or your theories?

[Nick Rowe]

Allan: "I am telling you that I don't give a rats ass as a businessman. Either I am making money or I am losing money but I THOUGHT I could sell it at that price and I AM selling it at that price, and now that the price has been set, I just need to get the volume."

You don't give a rat's ass about getting more profits? And, doesn't volume depend on price?

"Making the client wait in line might mean losing a client, and so you might put on another cashier, but this is a marketing cost, and not something that can be added to the MC. Yeah, if you're selling 800 cans of coke an hour, you'll need another cashier,..."

But if you need to hire more cashiers to sell more cans, that is a part of the Marginal Cost of selling an extra can.

[Allan Manchester]

>And, doesn't volume depend on price?<
Only to a very minor extent. Thing is, I might very well make more money by selling the coke at $1.40. I might make more money at .75. I don't know. You don't know. I know that if I'm not making any money at the .80 I decided on, I'm not going to be readily convinced to try .75, especially if I am selling some. If I am making money on the .80, I am confirmed in my original intuition that I got the right price. Now I want to make more money, and can afford to take some risk. And so I jack the price to $1. Some customers complain, but really the sales volume doesn't change, even though the sales growth might falter a bit. And so I go to 1.20. Now the growth levels right out, but your profits are more than doubled, and so you're happy, and the clients are happy and everything is right with the world.

[Allan Manchester]

>But if you need to hire more cashiers to sell more cans, that is a part of the Marginal Cost of selling an extra can.<
But if the volume is such that you need more staff, then the price of that staff is more than offset by your purchasing power. If I am selling 800 cans/hr, then I can negotiate a price break from my supplier, and get the coke for .35. That gives me $40/hr. MC is always down sloped.

[Nick Rowe]

Allan:

"Oh, you teach that the businessman doesn't care about your formulas or your theories?"

I teach that they come to roughly the same point, on average, by a process of trial and error, and copying what works. Rather as you describe it.

">And, doesn't volume depend on price?< Only to a very minor extent."

If the percentage fall in volume is less than the percentage rise in price, then elasticity of demand is less than one, and you really should raise price to increase profits. MR is negative.

" MC is always down sloped."

It might be, for some firms, over some range, I agree. But *always*?

[Allan Manchester]

>If the percentage fall in volume is less than the percentage rise in price, then elasticity of demand is less than one, and you really should raise price to increase profits. MR is negative.<
But you see, price has almost no impact on immediate volume, but has high impact on growth rate. That is why in my example, that the price went up once the volume was sufficient. That is why concession stand prices are astronomical. They do not seek any growth rate.

[Allan Manchester]

" MC is always down sloped."
>It might be, for some firms, over some range, I agree. But *always*?<
Thinking about this question over last night, I happened upon the franchise opperation. In the franchise, the MR and the MC are flatlined, but that is because the franchisor sucks in the profit of the ever expanding MR/MC gap.

[Xevec]

">If the percentage fall in volume is less than the percentage rise in price, then elasticity of demand is less than one, and you really should raise price to increase profits. MR is negative.<
But you see, price has almost no impact on immediate volume, but has high impact on growth rate. That is why in my example, that the price went up once the volume was sufficient. That is why concession stand prices are astronomical. They do not seek any growth rate."

Let me ask you this Allan. Let us take your coke example. At $1.20, you say growth rate has leveled out. Let us say then, you raise the price to $3.00. Growth rate will obviously fall, but will volume fall? Will you still be able to sell the same amount as you did at $1.20?

[Allan Manchester]

>Let me ask you this Allan. Let us take your coke example. At $1.20, you say growth rate has leveled out. Let us say then, you raise the price to $3.00. Growth rate will obviously fall, but will volume fall? Will you still be able to sell the same amount as you did at $1.20?<
The growth rate would go negative. The first day you would probably sell just as much, but the people would be pissed, and start looking for other ways to spend their money.

[Xevec]

"The growth rate would go negative. The first day you would probably sell just as much, but the people would be pissed, and start looking for other ways to spend their money."

But volume would not decrease?

[Xevec]

"

"The growth rate would go negative. The first day you would probably sell just as much, but the people would be pissed, and start looking for other ways to spend their money."

But volume would not decrease?"

I ask this because of say, gas prices. Gas prices have increased over the years, but are they looking for growth rate? Since they have increased dramatically over the years, has their volume decreased? Or take even the health care industry. What about them? Has their growth decreased? You are willing to argue that increased prices decreases growth, but not sales. That taken an industry that has very little competition, or an industry that has increased its prices in unison, that sales will never decrease, no matter how high prices get, correct?

[Allan Manchester]

wtf is the matter with you.
Growth rate is the volume changing over time. Negative growth rate means that volume would go down over time.

[Xevec]

"wtf is the matter with you.
Growth rate is the volume changing over time. Negative growth rate means that volume would go down over time."

Why couldn't you say then, that volume does decrease?

As you said, prices have no effect on volume. And then again, you are also agreeing with the law of demand. That as prices increases, sales decrease. I thought you said that was wrong. When we were arguing earlier, you said that as sales increase, prices tend to decrease. But now, you have argued here the exact opposite. That once profit is realized, then firms will increase their prices until they see growth level out. In other words, you are equating growth with volume.

But again, you also say that how much you make doesn't matter, just how much you sell. Now, I can say that if you sell coke at 30 cents a can, growth rate will be tremendous. And again, you said this:

"How much is made or lost today is moot"

"As long as I'm losing money, I really don't care, and, as long as I'm making money, I really don't care"

Again, what you are saying here is that you don't care about MR.

Or what about even concession stands? Do they even care about growth rate? You said they don't seek it, so do they care about it? Can they continually raise prices and have volume not decrease? Let us take sporting events. Can a baseball game charge $10 per hot dog, and not see a decrease in sales?

[Allan Manchester]

Your problem is that you do not seem to understand when your argument is totally lost. I say it again. You have verbal diarrhoea.

[Xevec]

"Your problem is that you do not seem to understand when your argument is totally lost. I say it again. You have verbal diarrhoea."

I am not arguing anything. I am asking questions along with confirming the understanding of what I am reading. It tells me that you are not reading what I am writing, and deciding to instead insult me. Not very professional nor civil. Hell, not even nick has given out any insults even when responding to me.

[Greg]

Nick

All this analysis assumes that the business is paying the wage. What if the currency issuer is paying the wage? Now the wage is not a cost to the business but the added demand to the economy ends up benefiting all businesses. An employment subsidy for businesses and a govt job guarantee program would provide this.