Monday, March 2, 2015

A higher minimum wage does not make society better off or create jobs

I had a discussion today with a friend about the minimum wage and it gave me an idea for a thought experiment. Many pundits insist that a higher minimum wage will create jobs (see here for one example). The argument goes as follows (taken from the article linked to above):

"What critics of a higher minimum wage ignore is that, by putting more money into the pockets of the working poor — a group that necessarily spends nearly all its income on such locally provided basics as rent, food, transport and child care — an adequate minimum wage increases a community’s level of sales and thereby creates more jobs."

But can the minimum wage really create more jobs overall? The answer to this questions is most certainly no and a simple example will explain why.

Suppose there are two parts of a town; the owners of businesses live in one part and the workers live in another. These owners and workers produce widgets and everyone demands 1 widget per day. Owners are wealthier than workers. There are the same number of workers as owners.

In each area of town there is a pizza parlor that is owned by the government and uses no workers (the pizza is completely made by machines and has a perfectly elastic supply). Workers only go to their pizza parlor and owners only go to their pizza parlor. The equilibrium price of pizza is $9/pizza in each area. The equilibrium quantity for owners is 100 pizzas per month, while the equilibrium quantity for workers is 80 pizzas per month. (since owners are wealthier and they each demand the same amount of widgets owners spend the rest of their money on more pizza.) The graphs below represent the markets for pizzas in each part of town.


Now let's suppose the government passes a minimum wage increase in the widget factory. For the purposes of this example I will assume that the minimum wage increase is completely absorbed by the owners i.e. the owners take less money and pay the workers more; prices of widgets do not change (this is the preferred outcome of most minimum wage supporters as well).

This increase in the minimum wage will have an income effect on the demand for pizza for both the owners and workers. Since the workers now earn more money, the demand for pizza in their part of town will increase. The demand curve will shift out as in the following diagram.

The increase in demand will increase the quantity supplied of pizzas in the worker's pizza parlor; there will be a movement along the supply curve to the right.

But where do the inputs for the new pizzas come from? The increase in the minimum wage did not change the supply curves of pizza. There was no technology increase; no new way of making pizzas was created. In order to increase the quantity supplied of pizzas in the workers part of town the raw materials for pizza (dough, cheese, pepperoni, etc.) must come from somewhere.

And that is where the owners come in. Remember they are poorer now. So the demand for pizza in their parlor decreases, as in the graph below.

When the demand for pizza decreases in the owners parlor, the quantity supplied decreases and moves to the left. This frees up resources (dough, cheese, pepperoni) to be used for other things. In this case the pizza inputs that are no longer being used in the owner's parlor can be used in the worker's parlor.

This example illustrates a key point: society can not be made better off overall by moving money from one group of people to another. Wealth is created by more production, by figuring out ways to make more stuff using less resources. The minimum wage does not affect the supply curves of the products that worker's demand. Yes they have more money, but all that does is increase their demand for certain goods while it decreases the demand of other goods that were bought by the people that used to have the money. 

A higher minimum wage does not increase output so it cannot make society better off overall. Increases in the minimum wage simply shift money around, resulting in higher demand for some goods and lower demand for others, just like in this pizza example. The net effect is 0, since the pizza resources freed up by the decrease in demand on the owner's side were the same pizza resources used to increase the quantity supplied on the worker's side. In a more complex economy the increases and decreases in demand will be spread out over more products and it will be hard for any one person to see exactly where the changes are taking place, but the result will still hold. Workers are an input, just like pepperoni, and if more workers are going to be used in one market following a demand increase then less workers must be used somewhere else following a demand decrease.

So the next time someone tells you that the minimum wage will increase jobs by increasing demand, ask them where the new workers are coming from. Just like pepperoni, workers can't appear out of thin air.

Tuesday, February 24, 2015

Sherrod Brown doesn't understand trade

In the most recent email I received from Sherrod Brown he makes a startlingly ignorant statement:

This month I joined my colleagues of both political parties to announce the introduction of the Currency Undervaluation Investigation Act, to stand up to Chinese currency manipulation.

This bipartisan plan could create five million jobs, increase our GDP by as much as $720 billion, and provide a shot in the arm to American manufacturing — all at no cost to taxpayers

(My italics and bold print)

No cost to taxpayers? If taxpayers import goods from China (and they do, check your labels) then this act will certainly cost them.

First let me explain what Mr. Brown means by "currency manipulation". Suppose that a Big Mac costs $2 in the U.S. and an identical Big Mac cost 8 yuan in China. For purchasing power parity to hold, $1 would have to equal 4 yuan. This would be the "correct" exchange rate in Mr. Browns view.

But China may want a different exchange rate, perhaps to maintain high levels of employment in China. So in the currency markets China tries to maintain a ratio of $1 = 5 yuan rather than 4 yuan. So what does this mean for the U.S.?

Well it means that the stuff we buy from China is cheaper than it otherwise would be. For example, suppose a tube of toothpaste that is made in China sells for 10 yuan. At a ratio of $1 = 5 yuan I would have to use $2 to buy 10 yuan, which I would then give to the Chinese manufacturer for the toothpaste. So the toothpaste would cost me 2 U.S. dollars. If instead the $1 = 4 yuan was the exchange rate, I would need to trade $2.50 for 10 yuan. So at the $1 = 4 yuan the toothpaste would cost me $2.50 instead of $2. If purchasing power parity held, then the tube of toothpaste would be more expensive.

China prefers the exchange rate of $1 = 5 yuan in this example because it makes the things that China produces relatively cheaper compared to their cost in the U.S. Remember, based on purchasing power parity the toothpaste should have cost $2.50, which is what American made toothpaste would cost int this example. But because China "manipulates" the exchange rate the toothpaste only costs $2 to buy from China while it costs $0.50 more to buy from a U.S. manufacturer. (Like any good, currency prices are based on supply and demand. China can affect the exchange rate i.e. price of dollars in yuan by buying and holding dollars, which reduces supply.)

(Note that the higher exchange rate ($1 = 5 yuan) makes American goods more expensive for the people of China than the $1 = 4 yuan rate. When the Chinese producer of the toothpaste trades his 10 yuan in for $2 he cannot buy the American made toothpaste that is priced at $2.50.)

This is a stylized example but it reveals that Mr. Browns policy does indeed COST U.S. taxpayers by making the things we buy from China more expensive. So if you spend $1000 per year on stuff made in China and prices go up by 2% on average than this policy would cost you $20 per year. If 50 million Americans had their costs of Chinese goods go up by $20, then Mr. Browns "no cost" policy would cost Americans in total $1 billion. I would need to do more research to get a better estimate, but I guarantee that the true cost is a number much greater than $0.

So don't let Mr. Brown fool you. If China is indeed manipulating their currency then they are doing U.S. consumers a big favor. Trade is about producers AND consumers and ignoring one of the groups to paint a better picture of a policy is dishonest. But I don't expect much honesty from politicians such as Mr. Brown anyway.

Thursday, February 19, 2015

The minimum wage can increase crime rates

In a 1972 article in the Journal of Political Economy researchers found a strong relationship between youth unemployment and crime rates. Other studies such as Fleisher (1966) and Freeman and Holzer (1986) found similar results. I am sure that there are more recent ones as well but I am not up to date on that literature.

Below is a chart showing the year over year, January unemployment rate for 16 - 19 year olds from 1990 to 2015 by year. (click to enlarge)

The most recent data available is for January 2015 and the rate was 18.8%. The rate increased drastically during the recession and peaked in 2010 at 26.1%.

There are many studies that show that higher minimum wages decrease the employment rate of unskilled, usually young, workers (see here, here, and here). I also showed the relationship between the real value of the minimum wage and white and black youth unemployment in an earlier blog post.

The negative outcome of increased crime rates is not usually brought up when increases in the minimum wage are being debated, but it is a real outcome. People who support a higher minimum wage need to be aware that they are pushing low skilled, often young, workers out of jobs and potentially into a criminal career. This is also consistent with economic theory, since the opportunity cost of committing a crime is lower for unemployed, low skilled people than for people with jobs and higher skilled people.

The U.S. prison system is already overcrowded; politicians don't need to create more criminals by making it harder for young people to find jobs and gain skills. Public policy needs to encourage economic growth so that young people are able to find jobs and gain experience that will help them lead productive lives. Having a job also makes it more costly to commit a crime. The minimum wage is a hindrance to youth employment and needs to be tossed on the scrap heap.

Sunday, February 15, 2015

Why are CEOs paid so much?

Many people today complain about the pay of CEOs. Today's CEOs make 200 - 300 times as much as the average worker, up from 20 or 30 times in the 1970's. Many on the left want the government to reel CEO pay in, either through direct regulation or indirect regulation like a binding say on pay.

But why do CEOs make so much? A model can help us sort through this question. Below is a model showing wage on the vertical axis and effort on the horizontal axis.

There are 4 wage levels depicted, W1 through W4. There are also two marginal cost curves, MC1 and MC2. Now suppose you are on the board of directors and you want to make sure that the CEO you hire gives their maximum effort, E*. If their marginal cost of effort is like that of MC1, you could pay them a wage of W3. A marginal cost curve that looks like MC1 means that the cost of an additional unit of effort is increasing at a constant rate. That is, every new unit of effort costs more than the last but the cost goes up by the same amount each time. In a world where MC1 depicts reality the CEO who enjoys spending time with their family will incur a higher cost each hour they work but the rate of increase is constant. So the 40th hour of work might be worth $50, the 41st hour worth $55, the 42nd hour worth $60, etc. with each hour costing an additional $5. In this world a wage of W3 would induce the CEO to give their maximum effort, E*.

But what if the cost of effort is not like MC1, but instead it is like MC2? A marginal cost of effort like MC2 means that each additional hour is not only more costly, but the cost is increasing at an increasing rate. So the 40th hour is worth $50, the 41st hour is worth $57, the 42nd hour is worth $65, etc. Reality is more like MC2 than MC1. The more you work, the more you would rather be doing something else because each additional hour of work is forcing you to give up things that are more and more important to you. Missing family breakfast is costly but many working people, including CEOs, choose to bear this cost. Missing Timmy's baseball game and Mary's recital are costly, but again many working people have made this choice. But what if you had to work 80 hrs per week and miss every baseball game and every recital, in addition to missing breakfast and most dinners? Working hours 60 - 70 are more costly than hours 50 - 60. And hours 70 - 80 are probably a lot more costly than hours 60 - 70 because so much has to be given up in the form of lost time with family, hobbies, and even sleep. I think it is very reasonable to assume that the cost of working the 72nd hour is not only more than the 42nd hour, but a lot more. A marginal cost of effort that increases at an increasing rate is consistent with MC2.

If MC2 is the appropriate cost then it takes a big jump in the wage to induce the CEO to give effort E*. At W3 the CEO will only give effort E1, whereas with MC1 W3 led to E*. With MC2 the CEO will need to be paid W4 to give effort E*.

CEOs often work very long hours. They spend a lot of time traveling and being away from their families. They rarely take vacations. They are under intense pressure from shareholders, government regulators, and their customers to make the perfect decision. One wrong move can result in a hefty fine from some government agency, a decline in the stock price that wipes out billions of dollars in shareholder value, or a bad product that alienates their customers. In a world of 24/7 news and instant financial analysis, CEOs are under the microscope way more than CEOs in the 1970s. Not only that, but competition is international in scope. In 2015 companies from around the world compete with each other. International competition and trade were smaller in the 1970s. CEOs didn't have operations in every time zone that had to be monitored nor did they have 20 sets of government regulations that they had to abide by. CEOs of major companies today are the leaders of global entities and it takes a lot of time and effort to run something so large.

Most CEOs spend time as some other senior level executive before becoming a CEO. Those positions require a lot of time and effort too, but much less than being the face of the company. Anyone who follows the news knows that Jamie Dimon is the CEO of JP Morgan Chase, but who is the COO? Or Treasurer? or SVP of government relations? It takes a relatively big bump in pay to induce someone to move up to the top level from an already high paying position that is much less stressful and time consuming. The increase in stress and time costs are so great that the best people wouldn't do it for a 10% bump in pay.

And if the board thinks that the best CEO can increase the value of the company by $2 billion dollars over 10 years while the second best would only increase it by $1.5 billion it makes sense to pay another $10 million a year to get the best person for the job. That's a cost of $100 million for an increase in value of $500 million. A small % change in the value of a large company means big dollars for investors. CEOs that can increase value by 3%/year are way more valuable than those that can increase it by 2%/year. That marginal change in % means a big change in actual dollars.

So the next time you hear about CEO pay being out of control think about what it takes to be a CEO. There are certainly examples of bad CEOs that turn out to not be worth $10/hour, but there are many more CEOs that make a lot of money for their investors and are responsible for making great products available to consumers. For every bad CEO that harms a company there are 10 more who are helping their investors, employees, and customers achieve great things. Unfortunately those CEOs are the ones we never hear about.

Monday, February 9, 2015

Outdated zoning policy misallocates resources

Zoning that restricts certain land uses can lead to a misallocation of resources, especially if it is not updated to reflect changes in the highest valued use of that land. To demonstrate what I mean look at the picture below (click to enlarge).

The top part of the graph shows the land rent/acre on the Y axis and distance from the city center on the X axis. The different colored lines represent different demand curves for land for different uses. The blue line is the office demand curve for land, the red line is the manufacturer demand curve, the green line is the residential demand curve, and the flat black line is the agricultural demand curve. The demand curves can be thought of as the different willingness to pay that each potential owner has for land closest to the city center. The bottom circles represent the size of the city, with each circle representing a different land use.

Let's pretend the picture represents a city in 1950, prior to the interstate system being built. In this picture, developers who want to build office buildings place the highest value on the land closest to the city center. This is due to the fact that CEOs and top level executives rely on the quick transportation of information in order to make decisions. Often this information needs to be transmitted face to face and so locating in the middle of the city minimizes the distance that executives, both internal and external, have to travel to meet with one another. The higher willingness to pay for land results in the small, light colored inner circle being filled with office buildings.

Next are the manufacturers. They like to be near the city center because that is where the transportation hubs are located. Prior to highways and trucking goods that needed to be shipped over land used trains and large train stations were often located in the middle of cities. Also, rivers were used to transport manufactured goods and cities were built around rivers to facilitate transportation. The gray circle will be filled with manufacturing firms in this city.

Next comes the residential section. Homeowners want to decrease the distance they have to travel to work so they want to locate near the city center as well, but they do not value the land as much as the executives or the manufactures do. The charcoal colored circle is where the residents of this city live.

Finally the farmers live and grow food outside the circle in the surrounding white space. Since land closest to the city center has no particular value to them they end up in the rural area to minimize their costs.

This is a representative mono-centric city, meaning that there is one city center and it is surrounded by other areas used for various purposes.

Often land is zoned for particular uses, and if this zoning does not align with the actual highest valued  uses the land is used inefficiently. For example, suppose the light gray inner circle was zoned for residential use only. This would prevent office developers, who actually value the land the most, from using it. Thus the office buildings would have to be located elsewhere. This would increase the costs of transmitting information, likely resulting in firms being run inefficiently as well since their executives would not have as much information when making key decisions.

Another way zoning can lead to inefficiencies is if it doesn't adjust over time; the zoning might exactly align with the highest valued uses at one point in time, but then over time, as technology alters things, the highest valued uses may change while the zoning remains the same. Perhaps manufacturers valued the gray ring more than residents did in 1950 so zoning that limited that area to manufacturing was not a problem. But if manufacturers changed their willingness to pay after the highway system was built because truck transportation became cheaper, then residents might have the highest valued use for that land. If the zoning does not change residents will be unable to locate there even though they value the land the most.

Zoning policy that does not take into account the different demand curves for land will misallocate land and waste resources. Also, even if zoning policy did at one time allocate land appropriately, that policy may no longer be optimal if the demand curves for the different uses have shifted differently for different users due to technological changes that impact the costs of substitutes and complements for inner city land. I am not a proponent of restrictive zoning, but if zoning is going to be used it should be frequently evaluated and updated when necessary.

Wednesday, December 31, 2014

Government spending: What's the right amount?

Most pundits who talk about government spending do so as a percentage of gross domestic product (GDP). The long term average is roughly 20%, meaning that the amount of government spending in a year is equal to 20% of that year's nominal GDP. But it is not clear to me that government officials should be targeting that long term average as the "normal" level. Take a look at the graph below.

The red line is total federal government spending as a percentage of total GDP and is plotted using the right axis. As you can see it has fluctuated around 18 - 20% since the 1950's. The large spike is World War II. When people talk about maintaining a constant level of government spending they are usually referring to the relatively flat red line.

But why should government spending track productivity? A flat red line means that if GDP increases by 3% government spending increases by 3%. But GDP is a measure of the country's productivity; it does not measure the country's need for government. In fact, I can't think of a compelling reason why the government should spend a constant percentage of GDP.

This brings me to the blue line in the graph. The blue line is plotted using the left axis and it measures the per capita amount of government spending in inflation adjusted dollars. As shown in the graph, it has been increasing steadily from about $2,000 per person in 1950 to $10,000 per person in 2014. So the government spends five times more per person today than it did in 1950. By 1950 we already had the traditional public goods and services that most people think a government should provide: an army, parks, a tax collecting service, the post office and public education to name some of the most common things. Many of the regulatory agencies were also created and funded as well by that time, including the FDA and SEC. In fact, many government services such as national defense, drug and food testing, and parks experience economies of scale. So what have we gotten for this increase in spending?

Since 1950 the government has built and funded the interstate highway system, created the department of education, the EPA, and the Bureau of Alcohol, Tobacco, and Firearms to help fight the war on drugs. Even if you think all of these things should be funded by the government, it only took the government about $5,600 per person in 1980 to do so, about $4,400 less than the government is spending today.

The chart above shows what government spending would be as a percentage of GDP if it had maintained the 1980 per capita level of $5,663. The last row shows that government spending would only be 11.2% of GDP today at that level instead of 20.2%. Maintaining that level of spending would have drastically lowered the national debt, and in my opinion there would have been no reduction in the government services that many people (but not me) think that the government should provide. Looking at these numbers makes one really wonder what the government is spending that extra $4,400 per person on. (Remember, I am not arguing for a constant level of spending. I am arguing for pegging spending to population growth rather than output growth.)

As a country I think we would be better off maintaining a constant blue line at 1980 levels rather than a constant red line. Hopefully when the new Republican congress takes office they implement real spending cuts that bring the blue line down.