Hal R. Varian
University of California, Berkeley
Email for comments: email@example.com
April 17, 2001
About the only thing I haven't tried is a comic book, although this was my original ambition in life.
All of these endeavors have been pretty successful. The graduate book dominated the field for 15 years and is still widely used. The undergraduate book is available in 14 languages, and my publisher claims it is the most widely used intermediate microeconomics book in the world. The business strategy book, co-authored with my colleague Carl Shapiro, made the Business Week best seller list and was the top business book on Amazon.com for several months. I'm not sure how many people read the newspaper columns, but it seems to be enough to generate some crank mail.
I will be the first to admit that much of this success of this is due to luck. There is a lot to be said for being in the right place at the right time. However, I think that I've learned some things from these experiences, and what I've learned is likely worth passing on to other aspiring authors.
First, I offer some general lessons that apply to all forms of writing. The most important single thing is to know your audience. At the very outset you should form a clear idea of who you are writing for. Once you understand who your audience is, the next question is ask is why they should be interested in what you have to say. Do they want to be educated, informed, enlightened, or entertained? And why are you the person who can do it? If, after sober contemplation of these questions, you still want to proceed, you will be in a far better position to resolve the countless choices that arise when you write. When you have to make a decision, you simply ask ``What does my audience want?'' If you have a good idea of who your audience is, your should be able to answer this question.
The next important thing to do is to choose a good co-author. I've had the fortune of writing with Ted Bergstrom and Carl Shapiro, both of whom not only produce excellent prose themselves, but are also great editors for my prose. It is critical to have a sounding board. The blues singer Taj Mahal says, ``if you cain't get a wife, get a band.'' My advice for authors is: ``if you can't get a co-author, get an editor.'' You need to have someone with good taste who can read your writing and tell you want works and what doesn't.
The most important aid to writing a textbook is to teach the class for which your book is intended. This seems obvious, but you would be surprised at how often this guideline is violated. Several years ago a very well-known economist decided to write a principles book. When he sought my advice, I told him he had to teach the class. He reluctantly agreed ...but he only taught a 30-student honors section! Needless to say, this experience did not help him to develop sufficient empathy with potential adopters of his book, who had to teach 600-person classes of students with widely varying ability.
When you teach the class, listen to the feedback from the students. What you are teaching all seems obvious to you, but it sure doesn't to them, and it probably didn't to you, the first time you studied it. Your job as a teacher is to help the students understand the material, and you can't do that effectively unless you know what they don't understand.
Some of the best advice I ever got about writing a textbook was from Richard Hamming, winner of the Turing Award, the closest thing computer science has to a Nobel prize. He said ``Get together the exercises and exams that you expect the students to be able to solve once they've read your book, then write the book that shows them how to solve them.'' The great thing about this advice is that it focuses you on the outcome you want to produce. If you understand both where you audience is coming from, as I advocated earlier, and where you want them to be, it's not that hard to write the book that helps them make the transition.
If the goal is to convey knowledge to the student that will allow them to solve problems, then it is important to put in a lot of examples of problem-solving in the book. You can't have too many examples. Verbal, numeric, algebraic, graphical--you need everything, since different students find different sorts of examples compelling.
When I say ``get the exercises together'' I don't just mean mathematical exercises. That's important for some areas, but, at the undergraduate level at least, we are educating students to be citizens, and examples drawn from political platforms, newspaper op ed pieces, and the evening news are just as important as the analytic examples.
Finally, you should strive to make things easy for the students and the professor. When I wrote my undergraduate book I included a full set of the slides I used in my own lectures. I did the same thing with my business strategy book. It's hardly any extra effort if you follow my advice to teach the class before you write the book. But teaching materials are a big help to people who may want to use your textbook.
Writing Information Rules was an experiment. Carl Shapiro and I were both very interested in reaching a new audience, and we spent a lot of time thinking about who, exactly, the book was for. We already had quite a bit of material that we had pulled together for our teaching, research and consulting, so we had a good base of material to start with.
One of the most difficult things about writing a ``business book'' is how you phrase things. One good piece of advice is: ``Don't say what is, say what to do.'' As scientists, we naturally like to describe things; our goal, after all, is to develop a better understanding of the world. But the readers of a business book don't necessarily share this goal. Their focus is on deciding what to do. So instead of saying ``price discrimination can increase profits'' you should say ``charge users according to their value, not according to your costs.''
Whatever you do, don't use the passive voice! ``Profits can be increased through the use of price discrimination'' is about the worst way to phrase the idea. The decision makers disappear when you use the passive voice--hardly the right way to think about either social science or management!
This price discrimination example also illustrates another important point: you should try to use terminology that is meaningful to the reader. I hate to say it, but economist terminology is terrible. Economic jargon just doesn't convey much meaning to the average reader. ``Rent'' is a nice example. It's an important concept, no doubt about it, but your typical reader just won't understand the subtleties without a lot of explanation. ``Profit'' they understand (or at least think they do) and it's probably a better term in most cases.
Finally there's the point I made earlier about making things simple for the reader. Callouts (those phrases in bold scattered through the text) and takeaway lists (the lists of points to remember at the end of the chapter) are useful tools for this purpose. They're also a big help to you as an author: if you find that you've written two or three pages and you can't come up with a one-sentence summary of what you've said, you should worry that you are just treading water. Perhaps you should cut some of that excess verbiage. Similarly, if you come to the end of the chapter and you don't have half-a-dozen non-trivial lessons learned, you should probably go back and rethink the chapter.
My final advice is to tell a lot of stories. Business readers can tolerate some abstractions, but what really grabs them are stories. And the closer these stories come to the problems they face, the better. Remember they want to know what to do--so stories about what others have done, and how these choices have turned out, send a powerful message.
I have to say that this is the place where I have had the least experience, so perhaps you should regard my advice as tentative at best. I've only been writing columns for a few years, and I still find it hard going. Some columns write themselves, but some require hours of agonizing over details. Maybe someday, when it all becomes easy, I'll write a revised version of this section.
The critical advice here is to get to the point as soon as possible. This usually means the first sentence or two. You are competing for the attention of the reader with everything else in the newspaper, so you need a grabber to get them to read your column.
My second piece of advice is to pace yourself. You've got a word count, and you've got to meet it. When I wrote 750-columns I thought of them as sprints: you rush out of the box, hit your stride and zip through the finish. You've only got time to develop one idea, or, alternatively toss off a few undeveloped ideas. A 1000-word column is like a half-mile: you can develop an argument with two or three points. This is quite a bit harder to do since you don't really have the space to make any arguments in depth. It's a challenge to both develop and argument and keep the reader's attention. It's something I'm still working on, but I think I'm getting there.
I realize that this essay isn't about research methodology. But I've discussed my views on that topic in other places and don't have much new to say about how to do economic research. But in a very real sense research only has an impact if it is communicated successfully. I don't know if ``methodology of communication of economics'' is a subject or not, but if not, it should be. Perhaps this essay can make a very slight contribution to that field.