A Hedonic Model for Words

I am pleased to announce that the paper “Pricing Quality Attributes of Internet Domain Names: A Hedonic Model for Words” has been accepted for the 2014 Americas Conference on Information Systems. It is joint work with Prof. Claudia Loebbecke from Cologne University.

The study develops a hedonic pricing model for domain names. The core findings are:

  • Domain prices depend on quality attributes and can be partially predicted based on factors like TLD, domain length, number of searches for keywords and many more. The model presented in the paper is far from being useful for practioners trying to value individual domains – it shows, however, that it is possible to build such models.
  • Domains that are sold at fixed prices are, on average, of lower quality than domains that are sold in negotiated deals. Sellers concentrate their attention on the high value targets. The study does not make any claims on which sales form provides better results for sellers.
  • It is difficult to predict which domains will be developed into full websites after a sale. Still, domains that are developed sell for higher prices than those that are not used or parked after the sale.
  • The price premium for domains that are later developed has the same magnitude for fixed price and negotiated deals. We find no evidence of sellers being able to discriminate between domain investors and domain end users.

The full paper can be downloaded here: Hedonic Prices For Words

Press coverage: "The Value of Virtual Real Estate"

The MIT School of Architecture + Planning news writes about my work on virtual real estate:

The Value of Virtual Real Estate
Exploring the Economics of Internet Domain Names

Current research at SA+P’s Center for Real Estate is investigating the economics of virtual real estate using techniques customarily used for bricks-and-mortar and dirt. While the Internet has always been thought of in terms of real estate – homepage, website, domain name – this project is thought to be the first time real world evaluation techniques have been applied to the world of Internet domain names.

The arrival of this approach is timely. The Internet is currently host to 22 top level domains such as .com, .gov, .org, etc. – not counting country‐specific extensions such as .com.uk – but consumers soon will be able to choose from nearly 1000 new domains for their online presence.

In anticipation of this massive explosion of supply – unparalleled in both virtual and traditional space – Thies Lindenthal, a postdoc fellow from the Netherlands’ Maastricht University, is developing ways to assess the potential value of that new real estate. The project is funded by the Internet Corporation for Assigned Names and Numbers (ICANN), a private nonprofit responsible for the coordination of the Internet’s systems of identifiers.

The new domain names that are flooding the market range all over the map: there are lots of geographic names – like .london and .boston – and a host of other sectors like .radio, .kosher, .gay, .yoga, .hockey and .books, to name a few. At this writing, ICANN has received about 2000 applications for those new names, twice as many applications as there are names. (Google applied for 101 different names; Amazon for 76; Apple applied for only one – .apple.)

To assess the market for these new domains, Lindenthal analyzed the records of about 250,000 sales of existing domains over the past eight years. The analysis revealed that unlike most real estate in the real world, the perceived value of virtual real estate can fluctuate wildly from one sale to the next – within one year, the value of a domain name may rise or fall as much as twenty-fold, the agreed-upon price more a function of gold rush fever than the trend of the entire sector. When analyzed in the aggregate, the domain name market turns out to be moving in sync with the rest of the economy.

Because of these wild fluctuations in perceived value for individual domains today, predicting demand for any one of the extensions tomorrow is largely a matter of guesswork at this point. But predicting demand is useful information for businesses who want to determine what they want to buy and how much it might be worth, as well as for policymakers who, in the spirit of virtual urban planning, need to decide which domain names to release first because they have the greatest pent-up demand.

To try to understand the likely demand for any new name, then, Lindenthal is analyzing the field of existing names. There are, for instance, about 50,000 .com domain names that include the word ‘Boston’. Bostonredsox.com, Bostonglobe.com, Priscillaofboston.com, etc. It is possible to see when each of those names was registered and to determine whether demand for a Boston identifier is growing, declining or tapped out, then use that trend to estimate pent-up demand for the .boston domain.

As the new names are released to the market, Lindenthal will establish a panel of DNS registries, registrars and domain end‐users to be surveyed at monthly intervals to track the actual market for new names, looking to see how different launching strategies affect success. Going forward, he aims to set up an ongoing tracking system to analyze that market data and make it available to the public, free of charge, on the web. At which domain name, however, has yet to be determined.

This research follows on Lindenthal’s dissertation research, which used these standard estimation techniques to investigate the valuation of housing in declining markets. His previous research as a founder and CEO of the Internet consultancy finnlabs. led to the first scientific-grade price index for Internet Domain Names (IDNX), which quickly gained reputation as the domain industry’s leading benchmark. For more information about his work, contact Lindenthal at thilin@mit.edu.

Read full article here: http://sap.mit.edu/resources/portfolio/virtual_econ/


Paper published: Real Option Value over a Housing Market Cycle

The Journal of Regional Science and Urban Economics accepted a paper for publication, in which John Clapp, Piet Eichholtz and I research the ups and downs of the West Berlin housing market.

The goal of the paper is to single out the value that buyers and sellers attribute to the real option to extend the size of an existing house. We find that the possibility to add more space carries substantial value. Furthermore, we see that this value is highest in times of rising house prices (e. g. 1990-1994 in West Berlin) and disappears when house prices fall. The real options embedded in existing properties therefore significantly contribute to house price volatility.

Get the full paper here.

Villa Freundschaft Berlin Grünewald villa 1902
High and low option value properties in Berlin

The evolution of east coast cities, beautifully drawn

Ever wondered why Boston, New York, or Philadelphia look so different from other American cities? What are shared patterns among these east coast hubs? Sam Bass Warner and Andrew H. Whittemore provide a fast motion view on the economic, political and social forces that shaped the urban form of these places from fragile early settlements to today’s metropolitan areas.

The black and white drawings throughout the book already tell the story on their own – and the text is a fast and very accessible read. This book changed the way how I look at the city on my daily commute.