Restaurant advice - ditch the middle

Tyler Cowen linked to some very good restaurant advice from Todd Kilman.

The basic thesis is that avoiding mid priced restaurants in favor of high end and low end restaurants will help avoid bad meals.

Anyone living in an area with good ethnic food can see the strengths of this strategy already - a good burrito or broken rice dish is going to beat most meals at casual sit down diners. Middle end sushi places will be getting their fish from the same places that low end sushi places do - they all will be serving the same mass frozen salmon/tuna/yellowtail/albacore that you can find in local Japanese supermarkets - the middle end places will just charge more for the decor and service. The places serving better sushi will be the ones at a high enough price point (or at a mid-high price point but in an area where their overhead costs are significantly cheaper) that take the trouble and expense to source better fish from their wholesalers.

There is a problem with this advice for those who live in or near San Francisco, which arguably has the best middle end food in the country. But even at the high-middle end restaurants in SF I often find myself disappointed as common tropes such as pork belly dishes at froufrou places won't be as good (but will be three times as costly!) as at much cheaper ethnic restaurant.

Kilman ends with a discussion of when middle end restaurants are unavoidable:

"Granted, there are going to be times when you can’t avoid the middle. You’re out with a group of co-workers, say, and need to find a place that satisfies a multitude of tastes and needs. Or you’ve got family coming in, and need to keep the costs down for a big group, while also making sure that a not-so-gently aging aunt is going to feel “comfortable.” These are obligation meals. And there’s nothing much to be done about them, except to go along with the wishes of the group, which, by the way, is what a good chunk of the middle exists to serve—the needs of the many."

One of the best ways to avoid middle end obligation meals is educating potential dining partners. Alternatively, now you know why your foodie friends are avoiding you.

Loyalty and Patronage Politics

Loyalty is often an underrated attribute. It's important for large networks to be oiled by trust and having everyone work from the same page will mean that things actually get done. The Bush family values loyalty highly - it's a theme prevalent in 41, George W's book about his father. A recent NYT article covering Jeb Bush's campaign and their insistence that policy advisers only advise his team highlights how he has similar views on loyalty.

One of the major problems in politics today is a modern elite patronage system in which the government enables the continuation of rent seeking opportunities to large political donors. In this environment, an overvaluation towards the concept of loyalty can backfire. Loyalty to people will mean loyalty to insiders who have been benefiting from policies at the expense of the public.

Jeb Bush comes across as one of the more outwardly sensible GOP figures when it comes to policy, but libertarian leaning supporters need to remember that someone who values loyalty highly will have little interest in dismantling the system that has served their supporters well for decades. 

The principle applies to more than just one family, but it's helpful to remember that seemingly good qualities can have significant drawbacks in the wrong scenario. It's also an important reminder that it is important to identify the long time supporters of politicians, since those are the ones who might be able to call on loyalty for favors that help them at the expense of the public.

The More Fungible Worker

A lot of people have been worrying the technology is taking away jobs and suppressing wages. To a large extent these worries are compared to the Luddites and are dismissed as silly and misguided.

The economist reaction to these worries is basically "That's silly, this is how economic growth works. No one thinks it is bad that the horse driver and telegraph operator jobs have disappeared." Economists are right that old inefficient jobs need to be replaced in order for human society to grow richer over time. When technology satisfies one need, workers are free to take a job that satisfies another of the almost infinite desires that humans have. And when workers are more productive, the marginal productivity of a good worker often goes up proportionally and their wages increase. 

But there is more to the "technology reduces wages" story than economists like to admit. The Luddites tried futilely to stop progress that benefited society, but the artisans themselves really did see their wages fall due to competition from technology. 

The obvious way that technology reduces the wages of workers in the developed world is how it enables and facilitates competition from workers in the developing world. Arnold Kling has called this the Great Factor Price Normalization.  In this scenario the average worker is significantly better off and society gains from trade, but workers in the developed world may see wage stagnation to pair with lower prices.

Arnold explains. "I want to suggest that there is a connection between this trend and the stagnation of median incomes in the United States, and even to the decade-long drop-off in employment here. New patterns of trade are developing that are reducing the advantage that a person enjoys merely for being located in the United States."

But there is a second way by which technology can suppress wages - when technology can guide and track workers the output of different workers becomes more uniform and thus more fungible. Before the era of ubiquitous cheap information technology, training and tracking the performance of mid and low skill workers was more expensive. Workers varied significantly in quality and ability and it made sense to pay workers with experience more so they would not have to be replaced by new workers who would be expensive to train. The extreme example of the fungible worker is the Amazon warehouse worker.

"The programs for our scanners are designed with the assumption that we disposable employees don't know what we're doing."

Amazon can quickly train employees as well as track them to make sure they are not underperforming. There is a smaller difference between one employee and their potential replacement. If an employee decides to leave their replacement is guided by technology that makes the new worker's productivity very close to that of a good worker.

Before these technological advancements, workers enjoyed mini-monopolies. The warehouse worker couldn't be easily replaced because their replacement might be way less efficient as they learned the layout of their workplace. As technology more directly guides low and mid skill workers in their jobs the workers are losing their mini-monopolies. Workers don't need job specific experience to be hired so the supply of workers available to every technology guided job has increased. A higher supply leads to a lower price (the price of a worker is their wage). After full employment is reached the general wage level may increase for low skilled workers, but for now the impact of the Great Factor Price Equalization and the More Fungible Worker are suppressing the wages of middle and lower class developed world workers.

Vocational Training vs College in the US

Critics of the higher education system in the United States have a lot to complain about, but one of their clarion calls is that there needs to be a shift towards more vocational training. In principle this seems reasonable, the problem is that vocational training in the US doesn't work very well. A study of How College Affects Students found that evidence from the 1990's suggested that the impact of sub baccalaureate vocational training has approximately 1/2 the impact of an associates degree. And associates degrees are less than one half as effective as a bachelor's degree. In recent years, the advantage of a bachelor's degree has increased even more - this causes those who are familiar with the data to be very skeptical of those who call for increased vocational training.

Meanwhile, vocational education does seem to work quite well in Europe. Switzerland and Germany appear to be running many successful vocational training programs.

What's keeping the US vocational programs from looking like the successful European ones?

1. A lot of vocational education in the United States is provided by for-profit colleges that are dependent on federal student aid. For-profit colleges in the United States have until recently only had to convince students to attend and the federal government to continue to provide loans - they have not been held accountable for their failure to provide a valuable education. When the education company with the political connections and the best marketing win, the bad education companies prevent good education companies from taking root and growing.

2. In Europe, the important of a four year college degree is not stressed nearly as much. Especially if students were in a top high school, there is some measure of shame in going directly to vocational education.

3. If only low quality students are going to opt for vocational training right out of high-school, businesses will not want to create apprenticeship programs for workers they would not want to hire.

4. Vocational training in Switzerland starts at the age of 16. Instead of being at school for five days a week, students are in school fir one or two days a week and working at their apprenticeship on the other days. This preemptive sorting does not sit well with the American notion that anything is possible for high school students.

Current critics of the United States higher education system need to understand why vocational programs in the US generally have not worked. The best solution to a poor education system is the creation of a parallel system that works better. Peter Thiel is trying to demonstrate a parallel system operating with top students via his 20 under 20 program. But significant change will not occur without a solution that can be applied to a sizable percentage of average students. These solutions need to provide proof of intelligence and consciousness while imparting useful skills. And perhaps most importantly, any successful large scale education system has to show that their students have a willingness to conform

Defenders of the higher education status quo need to realize that emphasizing the importance of a four year college education for everyone creates a self reinforcing belief that makes it difficult to create workable alternatives. But many of those arguing in favor of our current education system are professors with a lifetime tenure, so some of them might understand the self reinforcing cycle they are perpetuating. 

Assorted Links: The Environment

1. Death to beavers! The next time you hear someone discussing the environmental benefits of vegetarianism for cutting methane production, see if they are equally as enthusiastic with regards to hunting down beavers in order to cut into some of the 800 million kg of methane they cause. [Edit: Not 800 million kg of methane, but methane to be 800 million kg CO2 equivalent]

Warning: Pointing out the various ways in which politics is not actually about policy is not the best way to make friends at parties. [Edit: And the search for counter-intuitive insights can often lead to inaccuracies]

2. India is attempting to stimulate growth by parring back environmental rules. Considering their current environmental rules are organized, they would need to be replaced anyway. The question is whether this cuts back on the need of bribery or if it just centralizes the bribes into the hands of centralized officials.

3.  The Economist on geoengineering.

"And the mere fact of experiments going ahead might lead people to assume that geoengineering could easily be made feasible, and thus to give up on reducing carbon emissions."

This sounds a lot like the people who in 2009 wanted to pretend that monetary policy was impotent because the fed funds rate were at the zero bound. They made these claims this in order to maximize the size of the fiscal stimulus. Once it was realized that more stimulus was needed, monetary policy became the only game in town and the Federal Reserve implemented QE2 and other programs.

At some point, geoengineering will be the only game in town and the people fighting it today are just ensuring that we will have less data on the long term impact of many potential solutions.

Leftwing Vegan Problems

I'm neither leftwing nor a vegan. I just found the following line of thought to be an interesting logical problem.

There are many animals killed in mechanical crop harvesting. Clearing native vegetation, or even just utilizing mechanical harvesters, kills many sentient beings. There seems to be an active debate over whether or not enough animals are killed to justify the argument that eating cattle harms fewer sentient beings than eating vegetarian, but there is no question that many animals are killed in the cultivation of plants.

One way to avoid this is to utilize more human labor. Humans are less likely to indiscriminantly kill animals. Farming processes that require significant amounts of human labor, such as having humans pick fruit and vegetables, means that fewer animals are going to be killed by mechanical harvesters.

Vegans (and ethical vegetarians) are more likely than the average population to have political beliefs that see low wage occupations performed by immigrants as a form of exploitation. 

Until they succeed in getting at least some agricultural companies to raise wages to levels that they deem acceptable, leftwing vegans are forced to choose between exploiting immigrant labor or buying products whose production required the deaths of numerous small mammals.

One way to get around this is to not think about or discuss the problem of small animal deaths caused by food production methods. To the extent that the problem of animals killed in plant cultivation is ignored, it is evidence that being a vegan isn't actually about not harming animals.

Happy Thanksgiving!

Years to Train Physicians: An International Comparison

One aspect that is under weighted in the discussion on US Health Care is the focus on how our doctors are trained. When reform in the healthcare system is discussed, the main question is how the existing healthcare services should be allocated. We look to other countries for models on how health care is produced. Liberals cite the single payer systems in Europe that seem to work with relatively low expenses while conservatives often point to the public/private system in Singapore that are able to provide suitable services at a fraction of the cost of their gross national product.

But the question of allocation is only half the problem. The education of healthcare workers, and more specifically physicians, is a variable that is basically ignored by almost everyone in the health care debate. A 2009 paper by the Global Knowledge Exchange Network provides an overview of differing physician training requirements.
The table above contains data from the paper, supplemented with data from the World Bank, the OECD, the Congressional Research Service, the German Medical Association, and Singapore's Ministry of Health.

A 2013 OECD study found that the United States has the third lowest proportion of general practitioners to specialists among developed economies, with only Hungary and Greece having fewer general practitioners. Looking at the above table it's easy to see why - with education already lasting over ten years, the addition of a few more years for a significant bump up in salary is a strong incentive to avoid general practice. The return on those last few years of training is quite high.

It is interesting to note that a recent study found that foreign trained primary care physicians trained abroad outperformed American born physicians. Whether that says more about industrious foreign born immigrants or burnt out Americans is an exercise left to the reader.

Longer training times restrict the supply of physicians available to treat patients. A restricted supply means that demand must be rationed. In capitalist systems, that rationing occurs with the wages of doctors and costs to the patient going up. In socialist systems, that rationing occurs via longer wait times for patients with less serious conditions. But if restrictions on doctor training were loosened then in mixed systems there would be downward pressure on both prices and on wait times.

Unfortunately, no political reform of the medical system is going to occur without the public support of physicians. And the American Medical Association, the modern guild of physicians, is not going to support policies which have any chance of creating downward pressure on the wages of its members.

While analyzing which countries are most efficient in training physicians is a fun exercise, actual reform in the medical system needs to come more from other places. Ideas to reduce the incentives of hospital physicians to practice defensive medicine, price disclosure by medical service providers and the proper implementation of technology and standardization to reduce administration costs need to be given more weight. The political question surrounding the allocation of resources gets much easier if the amount of those resources can be significantly increased.

Remember - It's Tesla vs Dealerships

Another state has effectively banned Tesla sales - Michigan. From a public choice perspective it's not surprising that the homebase of the old auto manufacturers is moving against their upstart electric competitor. Creative destruction is always opposed by those who are hurt by changes in the status quo. But people looking at the newest twist in the story from the angle of "Big Three vs Tesla" are missing a key point:

Old school auto manufacturers have been held hostage to dealerships for decades. Laws originally designed to protect large capital investments of dealerships have been extended indefinitely. None of the auto manufacturers are allowed to sell directly into states themselves - politically influential dealerships have locked automanufacturers into a disadventagous long term relationships via dealer franchise laws.

When Tesla was let into New York, it wasn't a compromise with the Big Three that got them a spot at the table - the deal that let Tesla keep their retail outlets also helped the dealerships in their relationships with out-of-state manufacturers. 

This doesn't make the strategy of the Big Three in Michigan and elsewhere right. But from their perspective they are just trying to ensure that their new competitor has to deal with the same parasites that they are forced to do business with. This may entitle them to sympathy, but it does not mean they deserve any support in their misguided quest to expand the reach of their rent seeking dealership partners.

Tesla's real enemy has always been the car dealerships and their lobbyists. And in the battle of corporations vs. rent seekers, consumers win when rent seekers lose. 

Side note: Dealerships are often portrayed as small businesses that help drive economic growth. This is because small businesses are seen as engines of job growth. This is wrong - the actual causation is that young firms are drivers of job growth and the correlation between size and job growth exists because many young firms just happen to be small.  Dealerships are generally old businesses.

Attempting to Align Incentives: Banking Employees Edition

William Dudley, President of the New York Federal Reserve Bank, has an interesting idea on how to align the incentives of banking employees with the desires of bank regulators: Defer their bonus payments for 10 years and take regulatory fines out of the bonus pool. From his speech:

However, in contrast to the issue of trading risk, unethical and illegal behavior may take a much longer period of time—measured in many years—to surface and to be fully resolved.  For this reason, I believe that it is also important to have a component of deferred compensation that does not begin to vest for several years.  For example, the deferral period might be five years, with uniform vesting over an additional five years.  Given recent experience, a decade would seem to be a reasonable timeframe to provide sufficient time and space for any illegal actions or violations of the firm’s culture to materialize and fines and legal penalties realized.  As I will argue below, I also believe that this longer vesting portion of the deferred compensation should be debt as opposed to equity.

There is no question that banking incentives are broken. In the run up to the 2008 financial crisis employees were generating what looked like revenue in the short term, but ended up actually destroying many companies. Other companies were found to be systematically violating regulations designed to protect consumers, often because a simple cost benefit analysis can find that breaking the rules now and paying the fines later is more profitable than other approaches. Employees involved might accurately estimate that the chance of significant bonus clawbacks are generally close to zero percent. It's the shareholders who are stuck with the bill years after the bank's employees have walked away with between 40% to 50% of the revenue they generated.

What challenges will a Dudley style system encounter?

If the bonus pool is too lumped together the responsibility will be too diffuse to matter. Bear Stearns and Lehman both had significant employee stock ownership, and yet both firms went belly up in the 2008 crisis. Work will have to be done to ensure that the people causing the trouble are hit first and hardest by the problems they cause.

The deferal unilaterally changes the deal between shareholders and employees, and employees who do not want to wait for their bonus payments will leave the regulated sector. This may actually be a good thing, as FDIC insured institutions shouldn't be in the business of creating free call options for employees looking for a quick payout.

Base salaries will go up as a portion of total compensation to make up for slower payouts. Higher base salaries and smaller bonuses can cause employees to act in a more risk averse manner.

If this system isn't implemented unilaterally through all FDIC insured banks at once the employees with the most earnings potential will leave banks implementing this policy for greener pastures. Like reform to the modern college education system, reform has to come from the high status players or all at once.  

If a mandatory deferred compensation system was imposed from the top down as one of the requirements for deposit insurance or for institutions with access to the Federal Reserve discount window or for primary dealers then it has the potential to help more than harm. Shareholders win because they are more protected from the bad behavior of their employees. Citizens who are better protected against renegade employees and less likely to fund bank bailouts win, while certain employees who have been enjoying a "heads I win, tails you lose" dynamic are inconvenienced.

One caveat: Top down regulations on compensation are generally undesirable. Right now it is arguably too difficult to comply with all of the regulations that banks are supposed to follow - it's crazy that Citi needs 30,000 people working in compliance. This part of the system is also broken. And a congress plus a regulatory body trying to implement Dudley's reforms in some form may end up doing much more harm than good. But the financial industry, with both government subsidies and significant moral hazard, needs better ways to reign in broken incentives. Dudley's proposal gets a lot of things right.

Note to Silicon Valley: Jean Tirole is Relevant

Most Nobel Prize winning economic work is ignored by people trying to understand Silicon Valley, and for good reason. Subjects like macroeconomics and trade theory are rarely relevant for what people are doing on a day to day basis. Even the issue of efficient markets is a non-question. The venture capitalists don't believe and the people trying to automate wealth management accept it is true as a matter of faith.

But the work of Jean Tirole, who was awarded The Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel for 2014, is far more applicable to understanding the ecosystem of the technology sector. His work spanned many areas, but it is primarily on how to understand and regulate industries with few powerful firms. The Swedish Royal Academy of Sciences released a short summary of Jean Tirole's work here and their in-depth look is here - or just read Tyler Cowen's summary.

But why is this relevant for understanding the technology ecosystem? As Peter Thiel makes clear in Zero to One, one of the goals of a startup is to create an area where they are a monopoly. Google has an effective monopoly in online search. Facebook is the internet's most complete personal directory and the world's largest photo sharing company, LinkedIn owns everyone's online rolodex. Airbnb, Uber and many other successful new companies are successful because they essentially own their respective domains. Understanding the dynamics of market power is essential to understanding the most successful technology companies.

Once a firm is first to market, the question of what extent they should go to in order to deter competitors is a complicated question. In 1984, Jean Tirole and Drew Fudenberg plublished a paper on whether firms in various scenarios should over invest to protect their monopolies or if they should accept that they are going to face competition and stay lean and agile. The paper also highlights scenarios where the large initial investment to claim market share reduces a company's incentive to innovate, leaving them ripe for disruption after they have enjoyed the benefits of monopoly profits for a decade or two.

Another interesting question often faced today is that there is a company with a monopoly over one area, but how can they take that advantage to profit in other sectors? Tirole wrote many papers on vertical integration and monopoly power, and he did a literature survey of the field with Patrick Rey. Even if the goal of the literature is to figure out when to regulate these monopolies, companies can also use these strategies to find ways to expand their profit and market share. A mild version of this tactic is utilized by Microsoft when they release subpar versions of Microsoft Office to Mac so that power Excel users will be pushed to remain on the Windows platform.

Many startups are looking to build platform companies, where one company owns a two sided market and sells to both consumers and producers. Others may just want to better understand the incentives of the stores they are using to get their product to consumers. Jean was one of the first people to understand that maybe newspapers should be giving away their paper for free if that means they can charge businesses more for ads. This type of competition by the newspapers can drive their competitors out of business, but this seemingly monopolistic practice is benign - in many scenarios platform companies make is so everybody (except the platform competitors) wins. Here is a paper on Platform Competition that Tirole wrote with Jean-Charles Rochet - or better yet, a literature survey published three years later.  Alex Tabarrok and Vox have short posts addressing some of his work on platform markets.

This is only scratching the surface of his work, but it shows that many of the ideas driving value creation in Silicon Valley today have been analyzed in academia to a significant degree. These are complicated problems and sometimes returning to look at what the theory says can keep practitioners from making avoidable mistakes.