Book review: Game Wizards

I started playing Dungeons and Dragons in about 1987 or 1988, fairly quickly making the transition from a player to a dungeon master. My friends and I spent many nights, including some epic all-nighters, exploring dungeons and wilderness areas, fighting monsters, solving puzzles, and collecting treasure. Little did we know at the time the turmoil that TSR, the gaming company that produced Dungeons and Dragons, had been through only a few years earlier.

The upheaval that occurred at TSR in late 1985, along with the various highs and lows of TSR since the game (and the entire genre of role playing games) was invented by E. Gary Gygax and Dave Arneson in 1974, is detailed in the book Game Wizards, by Jon Peterson. As he explains:

This book is the story of how D&D rose from its humble origins to become a pop culture phenomenon, and what that remarkable journey did to the people who made it.

That story centres initially on Lake Geneva, Wisconsin, where Gary Gygax lived. Arneson, meanwhile, lived in Minnesota. Both were members of a small but energetic miniature wargaming community. Through their interactions, they created something great. However, what began as a partnership between two clearly creative individuals quickly devolved into a battle over recognition for their creation. Throughout the book, Peterson paints Arneson as the sort of person who is big on ideas, but short on execution. On the other hand, Gygax is generally able to deliver on ideas, both his own and often (but not always) those of others.

Peterson breaks the book up into episodic chapters, with each chapter corresponding to a year in the life of TSR. Much of the early chapters chronicle the development of the game, and the various characters who made early contributions, alongside Gygax and Arneson. However, very quickly the battleground between Gygax and Arneson is set, and plays out through their writing in wargaming magazines, through the conventions, and eventually through the courts.

An even bigger battle was brewing though, and Peterson does a great job of building suspense for it. The Blume family gradually grows in importance through all aspects of the growing company, but especially at the higher levels of management and governance. Peterson teases the reader in the first few pages, giving a quick view of a great battle to come in 1985. I was fully expecting that battle to be between the Blumes and Gygax, especially as the Blumes increasingly asserted control over the company in the early 1980s, including by disallowing share option purchases by other company employees who were entitled to them. However, in the last two chapters a new player emerges and ultimately takes control over the company, despite what Gygax thought were robust rules to prevent an outside 'non-gamer' taking over. Gygax eventually succumbed, following Arneson out the door with both losing control of the game they created.

I really liked this book. However, I suspect that it is not for everyone. It is, essentially, a case study of how an enthusiastic and entrepreneurial founder, who has developed a successful and growing product, can ultimately lose control of their creation. It is also a case study of how a successful company can lose its way through wasteful forays into non-core activities. In TSR's case, this included ill-fated acquisitions of a craft sewing collective (related to the extended Blume family) and failed movie deals. It is also a case study of how a founder can be outflanked by more clear-headed and business-minded people.

The context of the gaming community made these case studies especially interesting to me, because I recognised many of the names, as well as many of the gaming products and modules and D&D settings that Peterson references. In short, I was the perfect audience for this book. A reader who lacks background knowledge about the game, or who is less interested in the business of gaming, is left with a book about petty copyright disagreements and political manoeuvrings that may be of little interest to them. Nevertheless, I would recommend it to anyone who has interests somewhat aligned with my own.

This week in research #38

Here's what caught my eye in research over the past week:

• Davis and Mavisakalyan (with ungated earlier version here) find that a one standard deviation increase in individualism is associated with a ten percentage point increase in the likelihood that prostitution is legal, and that prostitution is more likely to be legal in countries in which women enjoy greater economic status (but no clear relationship between historical patriarchy and legality of prostitution)
• Ferreyra et al. (with ungated earlier version here) develop a theoretical model that shows that universal free college triggers a large enrolment increase but minimal change in graduation rates
• Xing and Tam use the construction of China’s national expressway as a natural experiment, and find that peripheral county governments reduced the level of spending and changed the spending composition after being connected to the expressway, consistent with the view that mobile capital tends to relocate from peripheral areas to core regions with a lower trade cost (those who think that greater transport connectivity improves economic outcomes for remote and peripheral areas need to remember that transport connections allow movement in both directions)
• Binelli et al. (with ungated earlier version here) find a positive effect of the COVID-19 pandemic on Italian students’ academic achievement in terms of both earned credits and GPA, and that both increased study time (due to confinement) and the availability of class recordings contributed to this positive effect
• Li, Chen and Ma find that exposure to (extreme) high temperatures significantly lowers the life satisfaction among older adults in China
• Gans (open access) shows that while generative artificial intelligence can increase communication by reducing costs, it may also disrupt traditional signalling mechanisms
• Matsumura et al. (with ungated earlier version here) show how high-frequency mobility data from mobile applications can be used to 'nowcast' sales in service sectors as well as production in labour-intensive industries (because mobility data shows you where consumers are, and where workers are, respectively)

Finally, for something completely different, one of my ECONS101 students from earlier this year created a tribute video (using AI) for me:

The student writes the lyrics, and AI composes the song and sings the tune. Watch some of their other videos on YouTube. They're pretty good and demonstrate what is possible using some of the simple AI tools that are available right now.

How meaningful jobs may contribute to the gender wage gap

Earlier this month, Dave Heatley wrote a great post on the NZAE's Asymmetric Information Substack about people working in jobs they are passionate about:

“Follow your passion”. That’s pretty much the universal career advice nowadays. But it comes with a rarely voiced kicker — if you choose a career in which passion-fueled supply exceeds demand, then don’t expect to be well paid.

I’ve seen this play out for my friends and family members who’ve chosen a career in conservation management. Job openings are scarce, highly contested, and get filled by passionate people who, on paper, are overqualified for the role.

This works to the advantage of employers — their employees are better qualified and their wages bill is lower than it would be otherwise. And lower wages have a further benefit from the employer’s perspective — they tend to screen out people without passion for the cause. That can benefit employees too, as they are pretty much guaranteed to work alongside other passionate people. It’s a win-win all round, at least until some employees decide they really would like to own a house, raise kids, or pursue other goals that require a more “normal” income.

Why do people get paid less in jobs that they are passionate about? I put this down to compensating differentials. Jobs have both monetary and non-monetary characteristics. Monetary characteristics include the pay and other monetary benefits. Non-monetary characteristics include the whole range of other things associated with the job. Perhaps it is dirty, dangerous, or boring. Or perhaps it is clean, safe, or fun. When a job has attractive non-monetary characteristics, then more people will be willing to do that job. This leads to a higher supply of labour for that job, which leads to lower equilibrium wages. In contrast, when a job has negative non-monetary characteristics, then fewer people will be willing to do that job. This leads to a lower supply of labour for that job, which leads to higher equilibrium wages. It is the difference in wages between jobs with attractive non-monetary characteristics and jobs with negative non-monetary characteristics that we refer to as a compensating differential (essentially, workers are being compensated for taking on jobs with negative non-monetary characteristics, through higher wages).

The passion that a worker can express through their job is a non-monetary characteristic. If a job is one that people are passionate about, then more people will want to do the job (holding other job characteristics constant), and wages will consequently be lower. Another way of thinking about it is that employers of workers who offer jobs to people who are passionate about them don't have to compete hard to attract workers, and so they don't have to offer as high a wage to fill the job. Either way, jobs that attract people with passion pay less because of compensating differentials.

Interestingly, one aspect that Heatley doesn't explore in his post is the types of people who do tend to choose jobs that they are passionate about. That's what Alice Evans did in this post the same week. She mostly writes about this article published recently in the journal American Economic Review (ungated earlier version here). After summarising the article, Evans notes that:

Burbano, Padilla, and Meier's careful study offers valuable insights into women's career preferences, revealing that income maximisation isn’t always the primary goal...

Women seem drawn to meaningful, socially responsible roles, but these often come with a financial penalty.

If women are more drawn to jobs that are meaningful, and more drawn to jobs that they are passionate about (and I'm going to use those two descriptors interchangeably), then this will drive a gender wage gap due to the compensating differential between jobs that attract people with passion and those that don't. On average the types of jobs that women prefer are more meaningful and therefore they are also jobs that pay less.

This seems like it would be a difficult problem to fix. If the government tries to legislate that meaningful jobs must pay the same as jobs that don't attract people with passion, then the meaningful jobs will become even more attractive, meaning even more workers try to get into those jobs, and the supply of workers in those jobs is even higher, and that has an even larger negative impact on wages. And because there are fewer workers wanting the jobs that don't attract people with passion, the supply of labour for those jobs will decrease and the wages for those jobs will increase even further. Or, because the number of meaningful jobs is finite, it would simply result in a larger number of unemployed and disappointed workers who can't get the meaningful job.

Evans' solution instead is to change the 'vibes' associated with working in different industries - her post is looking at industries rather than particular jobs, but the principle is essentially the same. That would involve making people (young women, especially) passionate about the jobs that they are currently not passionate about. That doesn't sound easy either, but is certain to be a more sustainable solution in the long term.

Read more:

• Meaningful jobs and compensating differentials

Cancelling subscriptions and customer lock-in

I posted last week about customer lock-in, and briefly discussed subscription services as an example. Then on Monday, The Conversation published this article by Katharine Kemp (UNSW):

Subscription business models have become common – many products are now provided in the form of software, an app or access to a website. Some of these would once have been a physical book, newspaper, CD or exercise class.

Most people who use online services have experienced the frustration of finding a credit card charge for an unwanted, unused subscription or spending excessive time trying to cancel a subscription.

Businesses can make it difficult for consumers to stop paying for unwanted subscriptions. Some do this by allowing consumers to start a subscription with a single click, but creating multiple obstacles if you want to end the subscription.

This can include obscuring cancellation options in the app, requiring consumers to phone during business hours or making them navigate through multiple steps and offers before terminating. The report points out many of the last-ditch discounts offered in this process are only short term. One survey respondent said:

I wasn’t able to cancel without having to call up and speak to someone. Their business hours meant I had to call up during my work day and it took some time to action.

Other businesses badger consumers with frequent emails or messages after they cancel. One respondent said a business made “the cancellation process impossible by making you call and then judging your reason for cancellation”.

Let me reiterate some points from last week's post (as well as posts here and here about online subscriptions). Making it difficult to unsubscribe creates a form of switching cost. Switching costs provide sellers with a lot of opportunity to extract additional profits from consumers. That's because high switching costs create customer lock-in - customers are unwilling to change provider, or stop buying, because they would then face the costs of switching.

We often think about switching costs in monetary terms, like the contract termination fee on a mobile phone contract, or a break fee on a fixed mortgage. However, switching costs can be highly effective even if they are not monetary. In fact, they could even be more effective. Take the example from Kemp's article - in order to unsubscribe, you have to call up and speak to someone. That takes time and effort (a switching cost). Add to that the fact that the call has to be made during business hours (increasing the switching cost). Being bombarded with emails or messages after cancelling adds a switching cost (although one that can be easily avoided by automatically sending all those emails to the junk folder).

Part of Kemp's article highlights these switching costs, and raises some justifiable concerns (at least, justifiable from a consumer's perspective). As a solution, she highlights firms that try to make it 'easy' to unsubscribe, noting that:

Businesses focused on a short-sighted cash grab fail to realise that consumers might cancel but later return if treated well.

However, consumers don't all return, regardless of how well they are treated. Because of that, it is more profitable for many firms to try and lock consumers in (if it wasn't profitable to do this, the firms wouldn't bother).

That brings us to the second aspect of Kemp's article, which is about how firms profit from their locked-in customers. In my ECONS101 class, I talk about two main ways that firms profit from these customers. First, firms may engage in multi-period pricing. This involves selling at a low price initially (sometimes an artificially low price, like a free trial), and then raising the price once a customer is locked in. This is why drug dealers may give away their highest-quality product for free! Second, firms may profit by selling complementary goods and services. This is how the manufacturers of coffee pod machines make their money - not from selling the machines, but from selling the pods. These firms can afford to give quite generous bonuses to their sales staff because each sale is going to generate a lot of coffee pod profits.

Kemp argues that 'unfair practices' should be legislated against. It is hard to argue against preventing unfairness. However, in practical terms, it may not be as simple as Kemp makes it out to be. Some ways that firms lock customers in can easily be re-framed in terms of customer privacy. Why does Firm XYZ make customers call during business hours to cancel their subscription? Because they want to be sure that the request to cancel is genuinely coming from the subscribed customer, and not from some identity thief. It would be difficult to legislate against a firm making customers who want to unsubscribe prove their identity.

On the other hand, some (but not all) of the ways that businesses profit from locked-in customers are clearly unfair and could be legislated against. Kemp discusses free trials that automatically transition to a paid subscription, or subscriptions that auto-renew. There is little justification that firms can provide for the former, and for the latter they would have to rely on 'customer convenience'. Neither is a particularly good justification, when set aside the costs that consumers face when firms engage in those practices. Certainly, subscription services are something that governments should be taking a closer look at.

Read more:

• The subscription economy as price discrimination
• The switching costs of online subscriptions
• More on the switching costs of online subscriptions

Try this: Stata cheat sheets

Despite the rise of R, and more recently Python, most economists still rely on Stata. Understanding any statistical software entails a steep learning curve, and Stata is no different. So, any opportunity to shorten the learning curve should be accepted with open arms.

I recently discovered that Stata has a bunch of 'cheat sheets' available here, that were developed by Tim Essam and Laura Hughes. They cover the fundamental commands associated with: (1) data analysis; (2) programming; (3) data processing; (4) data transformation; (5) data visualisation; and (6) plotting in Stata. They will come in handy for both novices and experienced users alike.

Essam and Hughes have further training resources available on GitHub here (which made me realise that I have seen these cheat sheets, or earlier versions of them, before - I shared them with my colleagues and PhD students back in 2016, although I didn't post the links to this blog).

Anyway, the cheat sheets are certainly helpful. Enjoy! 

This week in research #37

From poolside in Tahiti, here's what caught my eye in research over the past week:

• Kanbur, Ortiz-Juarez, and Sumner (open access) examine the trajectory of global income inequality since 1981, and try to identify a future turning point, when the growth of China will reverse decades of decline  in global inequality (a revised version of the working paper by the same co-authors that I discussed here)

New from the Waikato working papers series:

• Zhang and Gibson study the effects of connecting county-level units to the high-speed rail network in China, and find that growth in local economic activity (proxied by nighttime lights) is lower following connection to the network
• Suzuki et al. quantify peer effects and water pollution spillovers between shrimp farmers in Southern Vietnam, and find that neighbours' farming practices positively affect a farmer's practices and a disease outbreak in neighbours' ponds affects disease outbreak in a farmer's pond, even after controlling for contextual peer effects and correlated effects
• Li et al. identify the city-level and province-level economic characteristics and social and natural amenities that drive net internal migration between cities and prefectures in China using data from the 2000, 2010, and 2020 censuses
• Xu et al. use face reading software to evaluate how people's emotions affect their stated preferences and willingness to pay for changes in environmental quality, finding that induced emotional state has no significant effect on stated preference estimates or on willingness to pay for an environmental quality change

Google's strategy of search engine user lock-in

The Financial Times reported earlier this month (paywalled):

A US federal judge has ruled that Google spent billions of dollars on exclusive deals to maintain an illegal monopoly on search, in a landmark win for the Department of Justice as it seeks to rein in Big Tech’s market power...

The ruling follows a weeks-long trial in which the DoJ argued the search giant paid tens of billions of dollars a year for anti-competitive deals with wireless carriers, browser developers and device manufacturers — and in particular Apple. These payments, which cemented Google as the default search engine, totalled more than $26bn in 2021, according to the decision...

Google’s years-long agreement with Apple to make it the default search engine on the iPhone’s Safari browser has long drawn scrutiny. Unsealed court documents showed that Google paid Apple $20bn in 2022 alone. This would amount to a substantial portion of Apple’s $85bn-a-year services business, which includes its App Store and Apple Pay...

Also at issue in the case were contracts the tech giant reached over the years with browser developer Mozilla, Android smartphone makers Samsung, Motorola and Sony, and wireless carriers AT&T, Verizon and T-Mobile.

Clearly, despite paying billions of dollars to Apple, Samsung, and others, Google was able to make this strategy pay off. Otherwise, they wouldn't do it. To see how this works for Google, we need to understand customer lock-in. Customer lock-in occurs when customers find it difficult (costly) to change once they have started purchasing (or, in this case, using) a particular good or service. The switching cost could be explicitly monetary, like a contract termination fee, or it could simply be the time and effort required to switch.

In the case of search engines, Google relies on customer inertia to keep them locked in. Once a user has a default search engine set up on their phone, or their internet browser, or operating system, the user is unlikely to change. Sure, there are alternatives to Google available, like Bing or DuckDuckGo. But there is a switching cost involved in changing to those other search engines. The user would have to spend time and effort to set up the new default search engine. Why spend that time and effort, when they already have a search engine ready to go?

The extent of customer lock-in arising from customer inertia is arguably fairly low, because the switching cost is actually very low. And yet, this can be particularly profitable for firms. In my ECONS101 class, I use the example of subscription services. All of us probably have various subscriptions on the go that we aren't using, and some that we haven't used for some time. And yet, we keep those subscriptions going because of the time and effort required to cancel. That is customer inertia at work.

Unlike subscriptions though, Google isn't benefiting from charging a monthly subscription fee to the users of their search engine. In my ECONS101 class, we talk about selling complementary goods and services as one way that firms can profit from locked-in customers. Usually, we're considering firms selling complementary goods to its locked-in customers. However, as an intermediary in a platform market that connects search engine users and advertisers, Google has two customer groups. It offers one side of the market (search engine users) access for free, and benefits from them being locked in. It then sells access to the locked-in users to the other side of the market (advertisers). The larger the locked-in user base, the more advertisers are willing to pay for advertising, and the more profitable selling advertising can be for Google. And Google is immensely profitable (as I noted in this post last year).

Coming back to the antitrust case against Google, it seems obvious that signing exclusive contracts with the likes of Apple and Samsung reduces competition in the market for search engines, and therefore reduces competition in the market for search engine advertising. It will be interesting to see what happens on appeal. These cases can take years (decades, even) to resolve.

This week in research #36

Here's what caught my eye in research over the past week:

• Trinh and Munro (with ungated earlier version here) use a choice experiment to examine intentions to migrate among farmers living in the Vietnamese Mekong Delta, based on scenarios involving six attributes (drought intensity, flood frequency, income change from migration, migration networks, neighbours' choices, and crop choice restriction), and find that all attributes positive influence migration decisions, with drought intensity having the biggest impact
• Meoli, Piva, and Righi (open access) show that a diversified university curriculum increases the likelihood of Italian women graduates getting STEM occupations shortly after graduation, while it does not affect men, while doing internships during university studies and participating in study abroad programs reduce the likelihood of men graduates entering STEM occupations, but does not affect women
• Keloharju et al. (open access) find that doctoral studies causally negatively impact mental health, but perhaps less than previously feared, using data on PhD students in Sweden

Several papers published in the journal Ecological Economics responded to an earlier paper evaluated the costs and benefits of a 130kph speed limit in Germany (which I posted about last year), including:

• Sieg (with ungated earlier version here), who argues that by ignoring tax revenues from gasoline and diesel, the earlier paper overestimates the welfare gain by about 378 million Euros
• Eisenkopf et al., who attack the foundations of the earlier cost-benefit analysis, pointing to several serious shortcomings
• Gössling, Humpe, and Litman (the authors of the original study) respond to the two critiques, arguing that many of their criticisms lack merit, while others depend on viewpoint

Should the government operate petrol stations?

The extent to which the government should be involved in the provision of goods and services generates a lot of debate. Most of that debate is unhelpful, since it involves small-government, market-fundamentalist types arguing against anti-market socialist types. It's all ideological, and there's a pretty good chance that neither of the sides in that argument is right.

This type of argument has come to the fore recently, with the government making noises about a partial privatisation of Kiwibank. Over in Australia, the premier of Queensland has created a bit of a storm by suggesting that the government might open some petrol stations. As this article in The Conversation by Graeme Hodge (Monash University) reports:

Queensland’s Labor government turned heads last week with a bold new election promise. If returned to power, it would set up 12 state-owned petrol stations and limit fuel price rises to just five cents a litre on any given day.

The proposal certainly tapped into a pain point for Queenslanders – Brisbane topped national petrol price rankings last year.

But it was quickly met with a predictable pile on from opposing political commentators, industry bodies and some economists, attracting labels like “risky” and “dumb and stupid”.

Mark McKenzie, chief executive of the Australasian Convenience and Petroleum Marketers Association, called it a “wildly bizarre intervention” in the retail fuel market.

So is the Queensland premier really out of his mind, trying to win votes less than three months out from an election? Or is there actually some merit to this proposal?

It's quite possible that the answer to both of those questions is yes. However, there is a useful question to be asked here. When should the government be involved in the provision of goods and services that the market could otherwise provide (or is already providing, as is the case for banking services or petrol stations)?

First, let's get something important out of the way. Just because the government provides something, that does not make it a public good. By definition, a public good is a good that is non-rival (one person consuming it doesn't reduce the amount that is available for everyone else) and non-excludable (if it is available to anyone, it is available to everyone). There are goods and services that the government provides that are public goods. The police force is one example. There are goods and services that the government provides that are not public goods. Kiwibank and petrol stations would be examples of those (both are excludable, since you can prevent someone from accessing them if they don't pay for them). There are public goods that the government doesn't provide. This blog is one example. So, just because the government provides something, that does not make it a public good. It has to be non-rival and non-excludable.

Coming back to the question at hand, when should the government provide goods and services itself? In my ECONS102 class, we discuss four principles of privatisation (or the reverse, public provision). These principles are adapted from three principles originally outlined in Diane Coyle's excellent book Sex, Drugs and Economics (no relation to this blog), which I discussed in this 2016 post. However, I have expanded on the principles over time (including adding in the third principle, the inspiration for which came from the substantially-less-excellent book The Price of Fish by Mainelli and Harris, which I reviewed here). The four principles are:

1. The government can almost always raise large amounts of money more cheaply than the private sector
2. The government is almost always worse at running things than the private sector
3. Privatisation where there is a lack of competition will likely create a privately-owned monopoly
4. Whenever the good has a large externality, is a public good or merit good, or has a long-term payoff likely to be overly discounted by quasi-rational individual decision-makers, the government is almost always going to have to be involved anyway

These principles seem to me to apply regardless of what you believe ideologically, and they are something that every privatisation (or public provision) decision must grapple with. Is this a good or service that requires a large, up-front investment in order to establish? If so, by Principle #1, it might be good to have the government involved because the cost of borrowing to pay for that up-front cost will be lower. But if the government is running it, by Principle #2, you can expect a worse service. Is there a lack of competition? By Principle #3, privatising would likely create a privately-owned monopoly, and associated problems. Is this one of the markets that requires close regulation or supervision, like healthcare, education, public transport, or financial services? By Principle #4, the government is going to have a large role in the industry anyway, so why shouldn't they be operating in that market?

Should the government be operating petrol stations though? The principles don't answer that question, but they do give you some things to think about. This is not a business that requires a large up-front investment to get started (Principle #1). Sure, it takes some capital to set up a petrol station. Perhaps it takes a lot of capital. But it certainly doesn't take so much capital that you need the government to fund it.

If the government is running some petrol stations, those petrol stations will likely offer a worse service (Principle #2). Why? Government-owned firms lack a profit motive. Privately-owned petrol stations have a strong incentive to give you good service. If they don't, you'll fill up somewhere else next time, and that means less profit for them. If they do a bad enough job, and lose enough customer, maybe they go out of business entirely. A government-owned petrol station doesn't have the same incentive. They can provide a bad service, and still not be shut down, because they exist mainly for political reasons, not profit.

We aren't talking about privatising here, so a privately owned monopoly is not being created. But, reversing Principle #3, this is creating government-owned competition. There isn't a monopoly here, but arguably there is a lack of competition. Adding another competitor, and especially one that is prepared to limit price increases, is going to increase competition. On the other hand, limiting price increases is going to create a government-owned firm that may well make losses, and then it becomes a question of how long taxpayers will be willing to subsidise their loss-making government-owned petrol stations.

In the retail petrol market, there isn't a strong case for the government to be involved anyway (Principle #4). This is not a market that requires strong regulation, in the way that financial services, healthcare, or education does.

Overall, how should we evaluate whether this is a good idea? Principles #1 and #4 don't indicate a need for the government to be involved. So, it really comes down to how any benefits from greater competition (Principle #3) weigh up against the worse service that the government-owned petrol stations would provide (Principle #2). I can't easily judge this, and reasonable people may well disagree.

Closer to home, the government will make a decision on what to do (or not do) with Kiwibank soon. I wish the government would consider these same principles in its decision. Sadly, so much government decision-making is based on fixed ideological positions, that it's unlikely that my wish will be granted.

Read more:

• The case against privatisation

Two economic reasons gamification works to increase profits for firms

My ECONS101 class has been covering pricing and business strategy this week. So, I was interested to read this article in The Conversation last week, by Adrian Camilleri (University of Technology Sydney), which discusses 'gamification' in the context of business:

Gamification – the use of game elements in non-game contexts to increase participation – is on the rise.

Businesses use it to attract customers, boost sales and motivate employees to complete activities to drive profits.

The global gamification market is expected to increase in value from AU$23.6 billion in 2024 to AU$74.8 billion by 2029. This is the total revenue generated from products and services related to gamification, including software, platforms and applications.

The use of goals, points, badges, opportunities to level up and leader boards is now common in many industries ranging from education to health and wellbeing...

There’s a good reason why gamifying in business is growing – it works.

It works so well some engagement platform providers advertise gamifying a platform will increase website traffic by 50% and double social engagement.

Academic research is mounting to support the claim gamification increases customer engagement, which in turn increases positive word-of-mouth and boosts brand loyalty.

A good example is the annual McDonald’s Monopoly promotional marketing game. Based on the classic Monopoly board game, customers receive game pieces with their purchase of certain menu items.

By collecting these pieces, they can win prizes, either instantly or by completing sets. Of course, some pieces are rarer than others, encouraging customers to keep spending until they get a full set.

According to one analysis, the chance of winning a major prize is well over one in a million.

Camilleri 's article focuses on psychology (extrinsic and intrinsic motivation) to explain why gamification works in business. However, there are also complementary economic explanations for why gamification works, so let me provide two that use the economics I cover in my ECONS101 and ECONS102 classes.

First, gamification works because it locks customers into buying from a particular seller. Customer lock-in occurs when customers find it difficult (costly) to change once they have started purchasing a particular good or service. The seller can then profit by increasing the price for their locked-in customers, or by selling them complementary goods or services.'

Take the example of the McDonald's Monopoly promotion. Once a customer starts collecting the Monopoly game pieces, there is a switching cost of going to eat somewhere else - the opportunity cost of missing out on additional Monopoly game pieces, and the chance to win a big prize. That switching cost, albeit small, is enough to encourage customers to go back to McDonald's more than they otherwise would. This explanation also covers the loyalty schemes that Camilleri uses as an additional example (as I have discussed before here).

Second, gamification generates transaction utility. Usually, we consider transaction utility arising when a consumer feels that they are 'getting a good deal', and this makes them happier (higher utility) with the process of purchasing, and more likely to buy (as I have discussed before here). Camilleri uses the example of the 'Temu wheel of discounts', which provides a perfect example of this. But with gamification, the discount may not even be necessary in order to generate transaction utility. If the game is fun or rewarding in its own right, then the consumer will receive utility from playing the game (which is another form of transaction utility).

So, while extrinsic and intrinsic motivation may be good explanations for the success of gamification, there are economic explanations that are equally helpful in understanding why gamification may be a successful strategy for business.

Personalised pricing is now a step closer with electronic shelf labels

NPR reported back in June:

Grocery store prices are changing faster than ever before — literally. This month, Walmart became the latest retailer to announce it’s replacing the price stickers in its aisles with electronic shelf labels. The new labels allow employees to change prices as often as every ten seconds.

“If it’s hot outside, we can raise the price of water and ice cream. If there's something that’s close to the expiration date, we can lower the price — that’s the good news,” said Phil Lempert, a grocery industry analyst.

Apps like Uber already use surge pricing, in which higher demand leads to higher prices in real time. Companies across industries have caused controversy with talk of implementing surge pricing, with fast-food restaurant Wendy’s making headlines most recently. Electronic shelf labels allow the same strategy to be applied at grocery stores, but are not the only reason why retailers may make the switch.

Electronic shelf labels allow firms to change prices frequently (and was something I predicted back in 2017). This also allows them to tailor prices to current demand conditions (setting prices higher when demand is higher, and setting prices lower when demand is lower) as the article notes (referring to 'dynamic pricing'). However, electronic shelf labels could also allow firms to engage in personalised pricing (or first-degree price discrimination). This involves setting a different price for every consumer, as I described in this post last year.

If the firm could use personalised pricing perfectly, they would set the price for each consumer exactly equal to the maximum that the consumer is willing to pay. This is the holy grail of pricing for firms, because there's no possible way that a firm can be more profitable than when it can sell its products for the maximum that each consumer individually is willing to pay for them. 

Now, consumers don't willingly tell firms what they are willing to pay, but as I note in my ECONS101 class, firms do know a whole lot about their customers, including what they have purchased in the past, and what prices they paid, and importantly what they didn't purchase in the past, and what prices they were offered. That allows firms to estimate what their customers are willing to pay for individual products.

Now, consider how electronic shelf labels allow firms to increase their data collection and analytics efforts. The electronic shelf labels are a good way of gathering data, since the prices can be adjusted, and then the response of consumers (in terms of whether they buy, and how much they buy) can be measured. This data can then be mined (perhaps using machine learning) to try and estimate consumers' willingness-to-pay for each product.

Firms can then use that data for pricing, in increasingly sophisticated ways. Remember that, using electronic shelf labels, a firm can adjust prices in real time. Combining electronic shelf labels with facial recognition at the door would allow a supermarket to adjust prices based on which consumers are in-store at the time. In theory, the next step would involve combining electronic shelf labels with facial recognition and real-time customer tracking in-store (and you may doubt this is real, but a quick search on Google turns up dozens of firms that already offer this type of data analysis). This would allow a supermarket to adjust prices on individual items just before a customer gets to them. And altogether this adds up to something that is getting increasingly close to perfect personalised pricing.

The next step after that is to eliminate the shelf label entirely, and simply have consumers scan a QR code to find the price, or have it automatically appear on their custom app as the customer moves around the store. That way, supermarkets could serve a different price to every consumer without ever having to change a price label in-store. And no customer would know the price that is being offered to any other customer. Perfect!

[HT: Marginal Revolution]

Read more:

• Surge pricing is coming to a supermarket near you
• What would happen if perfect personalised pricing was possible?

This week in research #35

Here's what caught my eye in research over the past week:

• Bertoni, Rettore, and Rocco (open access) find that at least one third of the within-course variation in student evaluations of teaching is driven by student-specific differences in subjective style of responding to evaluation questions
• Pertold-Gebicka (open access) finds that, for students at a Czech university, high-ability students were often discouraged from studying during the COVID-19 pandemic, while a lenient grading policy allowed low-ability students to pass the compulsory exams and continue studying
• Bingley and Lyk-Jensen find that conscription improves the deployed intelligence pool compared to a volunteer force in the Danish military
• Taylor and Weder (open access) explore the economics of extinction (of species)
• Siflinger et al. (open access) find that mental health decreased sharply with the onset of the first COVID-19 lockdown in the Netherlands, but that it recovered quickly, and that worsening mental health was associated with labour market uncertainty, perceived infection risk, loneliness, and having children younger than 12 years old
• Sperber et al. find that Marshmallow Test performance at age 54 months does not reliably predict adult outcomes at age 26 years (not dissimilar from these results)

The latest issue of the Journal of Economic Perspectives had three papers from a symposium on 'expanding the appeal of economics', which are interesting:

• Avilova and Goldin (open access) show that the AEA's Undergraduate Women in Economics  interventions were effective in increasing the fraction of female undergraduates who majored in economics relative to men in liberal arts colleges (an important update on the earlier paper I discussed here)
• Barrera et al. (open access) discuss what economics can learn from STEM in terms of student identity, particularly that whether a student develops a sense that they are a “science and technology person” depends on a student’s interest in the field of study, their sense of their own competence or performance in the field, and their experience of recognition for performing well in the discipline
• Cook and Moser (open access) describe what has been learned from executing the AEA Summer Program at Michigan State University from 2016 to 2020

The Pope as a crime fighter

The world's religions are very diverse, but one thing that they all (or almost all?) share is an aversion to crime. For example, Christian religions have the Ten Commandments of things one 'shalt not' do, most of which relate to some crime or other. So, perhaps areas that are more religious experience less crime? The challenge in testing this is that religion, or religiosity, is so intimately related to many other variables in society, that extracting the impact of religiosity ceteris paribus is next to impossible.

Not to be deterred, that is what this recent working paper by Wang-Sheng Lee (Monash University), Umair Khalil (Deakin University), and David Johnston (Monash University) attempts to do. They have an interesting setting in which to investigate this research question. As they explain:

In this paper, we study a city-wide shock to religiosity engendered by the celebrated visit of Pope Francis to Philadelphia. In September 2015, Pope Francis delivered a highly anticipated Papal Mass to an estimated one million people in the city-center of Philadelphia. In addition, the Pope visited a correctional facility, where he interacted with inmates and made a statement linking religious behavior, or lack thereof, to criminality. Our work explores whether the Papal visit significantly reduced criminal behavior in Philadelphia in the short- and medium-term.

The theory here is that the Papal visit increased religious feeling, or the salience of religion, in the population, making them less inclined to commit crimes. Lee et al. use a difference-in-differences strategy, comparing the difference in reported crimes between the 12 weeks before and 12 weeks after the Papal visit in 2015, with the difference in reported crimes in the same weeks the previous four years (2010-2014). They do this analysis separately for different types of crime, both minor and serious. In their first analysis, treating only the week of the Pope's visit as being impacted, they find that there is:

...a significant reduction in daily drug offences per census tract of 0.021 and a significant reduction in daily fraud/corruption offences per census tract of 0.014, during the week of the visit. Relative to the sample means (day-census tract average number of crimes), this equates to a 18.9% and 18.2% reduction in drug and fraud offences, respectively. There are no discernible impacts for other offences.

Lee et al. then extent the length of time where the city is 'treated', and at the maximum length of 12 weeks they find that:

...there are significant reductions for simple assaults (0.027 or 15.9%), drug offences (0.028 or 25.2%) and fraud/corruption (0.016 or 20.8%).

Next, looking at different census tracts, Lee et al. find that the effects are smaller in census tracts that were further from the Papal mass, larger in census tracts that have a landmark church (where religion is likely to be particularly important, and larger in census tracts with higher poverty rates and with larger youth populations.

Third, Lee et al. do an event study, which rather than estimating the average effect over the entire treatment period, estimates the effect separately in each week. In that analysis, they find that:

...for simple assaults... we find a significant 0.038 (22.4%) reduction. However, this effect is found only for Week 1. On the other hand, we find substantial reductions in drug offences... that persist through Week 5 where the effect is a 0.030 (27.0%) reduction.

Finally, in a neat twist in the paper, they do a similar study which, instead of the Pope's visit, uses a visit by Barack Obama later the same year. The purpose of this analysis is to show that the results of the Pope's visit aren't just down to increased police presence in the city at the time of the visit. If that were the case, they should find a similar impact of Obama's visit, when police presence would also be very high. In the case of Obama's visit, Lee et al. find a much smaller impact (although still statistically significant) on 'Part 1 offences' (the more minor offences).

So, the Pope's visit reduced crime. The Pope is a crime fighter!

Or perhaps not. There is an obvious problem with this paper. Lee et al. use reported crime. Reported crime might decrease because crime has decreased, and so there are fewer crimes for people to report. Alternatively, reported crime might decrease because, even though crime has not decreased, people are reporting it less. Lee et al. aren't able to distinguish between these two possibilities.

I think a decrease in reported crime is particularly likely in this context too. If the Papal visit increases the salience of religion in people's lives, then perhaps they feel more positive about other people, even those who have wronged them. They may be more likely to 'turn the other cheek' (a saying that is literally taken from the Bible, Matthew 5:39). So, perhaps the victims of minor crimes are less likely to report them. On the other hand, victims of serious crimes would still report those (there is only so far that cheek-turning can go). So, in the analysis we would see a decrease in reported minor crimes, but no decrease in reported serious crimes, which is exactly what Lee et al. find.

So, maybe the Pope isn't a crime fighter after all. Perhaps he only deters victims from reporting minor crimes.

[HT: Josh for the paper, Craiyon for the AI-generated Pope-as-Batman image]

Wine markups at top New York restaurants

This week, my ECONS101 class has been covering elasticity. One aspect of that topic is the observation that firms with market power (such as firms in monopolistic competition, like restaurants) price their products with a markup over their marginal cost [*] (my class will build on this insight next week, when we look at pricing in more detail).

The optimal size of the markup depends on the consumers' price elasticity of demand. When demand is relatively more elastic (that is, when consumers are more price sensitive), the optimal size of the markup is relatively small, and the resulting price will be quite low. When demand is relatively more inelastic (that is, when consumers are less price sensitive), the optimal size of the markup is larger, and the resulting price will be quite high.

To see how this works, consider two firms, as shown in the diagram below. Firm 0 (black lines) has relatively inelastic demand for the good, so its demand curve D0 is relatively steep. That means that it also has a relatively steep marginal revenue curve (MR0). Firm 1 (red lines) has relatively elastic demand for the good, so its demand curve D1 is relatively flat. That means that it also has a relatively steep marginal revenue curve (MR1). Both firms are constant-cost firms, and both have the same costs of production (MC=AC). The condition for profit maximisation is that the firm will want to sell the quantity where marginal revenue is exactly equal to marginal cost. For both firms, this happens to be the quantity Q*. To determine the profit-maximising price, each firm needs to set its price so that the quantity of the product that consumers want to buy is exactly equal to Q*. For Firm 0, that is the price P0, while for Firm 1, that is the price P1. Notice that Firm 0 sets a much higher price than Firm 1. Firm 0 has a larger markup than Firm 1 (the markup is the difference between each firm's price, P0 or P1, and the marginal cost line MC).

How big can markups get? When demand is very inelastic, markups can get very high. Many luxury goods have quite inelastic demand (when they are not the target of conspicuous consumption). That's because, even though the price may be high, for the high-income consumers that buy luxury goods, their spending on the product is only a small proportion of their income. As one example, consider wines at top New York restaurants, as described in this Punch article. They interviewed a number of wine buyers for restaurants, and here are a few quotes from their responses:

Across the board at Four Horsemen, we almost always do a 3x markup, which is generally the standard...

Currently, the markup on our wine list at Roscioli is 2.5x the cost we pay for the bottle. The markup at most restaurants is 3.5x, so we’re trying hard to keep things more approachable...  

I generally stick with the norm as far as markups go: 3x to 3.5x...

Those markups are quite high, but not dissimilar to the markups many years ago when I worked in hospitality (albeit briefly). High-end restaurants know that their patrons who are looking to buy wines have relatively inelastic demand, so they can set a relatively high price (and a high markup). And so, that's what they do.

[HT: Marginal Revolution]

This week in research #34

Here's what caught my eye in research over the past week:

• Allen presents a game theoretic analysis of stalking (no data or empirical analysis though)
• Walasek, Mullett, and Stewart (open access) provide a meta-analysis of loss aversion, finding that few studies have the necessary design to estimate the lambda (loss aversion) parameter in a full prospect theory sense, and for those that do, the average lambda of 1.3 is quite a lot smaller than the value of 2 that Kahneman and Tversky originally postulated
• Dell discusses deep learning approaches for economists (with an associated GitHub repository of applications, software and datasets that can be used)

New Zealand's new average and marginal tax rates

This week my ECONS102 class covered taxes, and as luck would have it, today the Government's new tax changes come into effect. As the New Zealand Herald reported yesterday:

The tax cuts the National Party campaigned on will officially take effect tomorrow, with around 3.5 million New Zealanders benefitting, according to Finance Minister Nicola Willis.

Changes from July 31 include an increase in personal income tax thresholds, the extension of the independent earner tax credit (IETC) and an increase to the in-work tax credit and the minimum family tax credit.

I'm going to focus  purely on the changes in personal income tax thresholds, and update a figure I posted back in 2022, which showed the marginal tax rate and the average tax rate, for taxpayers on different annual incomes. The marginal tax rate is the amount of the next dollar earned that is paid in tax. The proportion of income that a taxpayer pays in tax is called the average tax rate. The marginal tax rate is what people often think about when they think of tax rates. However, for a country that has a progressive income tax, such as New Zealand, the average tax rate is always lower than the marginal tax rate. That's because people on high incomes, with a high marginal tax rate, nevertheless have paid a low marginal tax rate on their first dollars of income.

Anyway, here's the picture of those marginal tax rates and average tax rates. The rates that applied up to yesterday are shown in blue (marginal tax rate) and red (average tax rate). Those are the same lines from my 2022 post. Notice that the marginal tax rate is always higher than the average tax rate (except for very low incomes below $14,000, where they are the same). The new rates applying from today are shown in green (marginal tax rate) and orange (average tax rate).

Notice that changing the personal income tax thresholds shifts the vertical parts of the marginal tax rate schedule to the right (comparing the blue and green lines). That's because the higher rates now kick in at slightly higher incomes.

However, what most people are probably interested in is the change in average tax rates, not the change in marginal tax rates. This is shown by the difference between the red and orange lines. The difference is imperceptible at low incomes, as well as pretty small and decreasing at the highest incomes. It is largest in the middle of the income distribution, which is why the government has sold these changes as helping the 'squeezed middle'.

In fact, the decrease in the average tax rate is largest for those on an annual income of $55,000, [*] where the average tax rate has decreased from 17.31% to 15.86%, a decrease of 1.45 percentage points, or $799.50 per year. That isn't the largest absolute difference in tax paid, which happens at all annual incomes above $78,100, and is equal to $1,042.50 per year. Of course, as a percentage of income, that $1,042.50 is smaller at higher annual incomes, which is why the gap between the red and orange lines gets smaller for higher incomes.

*****

[*] I only looked at incomes in steps of $2500, and this is closest to the threshold with the biggest step up in the marginal tax rate schedule, which is now $53,500.

Read more:

• Tax cuts, and misinterpreted average and marginal tax rates
• Progressive taxation as an automatic stabiliser