When do we need models and how to write them?

These are important questions I’ve been thinking about recently, and here are some thoughts to share. The answers are open, and I welcome your contribution of thoughts.

When do we need models? In a broad sense, I think every paper needs to have a conceptual framework to guide the analysis and explain the patterns in the data. In the preliminary stage of a piece of research, writing down the exact assumptions and predictions of the model can also help a researcher to organize her thoughts and frame her question better. That said, whether a structural model is needed really depends on the purpose of the paper. If the point of the paper is to propose a new model, then obviously a model is necessary. However, for papers which contribute in estimation methods a review of relevant models is sufficient.

What makes a good model? One of my professors gave this piece of advice: Make the parts of your paper that aren’t the point of the paper as simple as possible, and spend the richness in the part where your innovation is. It is all too easy to come up with complicated assumptions and notations in every dimension of your model, but adding complexity also sacrifices flexibility and robustness. Another piece of advice, given by the same professor, is “don’t let the perfect get in the way of the good”.

A side note: I have found writing skills to be extremely important in economics, or maybe in all research disciplines. One might not need to be an extraordinary writer to succeed in research, but failing to convey one’s ideas smoothly definitely reduces the rate at which they spread and benefit others.

What I’ve learned in my first three weeks as a PhD student



The first three weeks of the fall semester flashed by, and I feel the urge of summarizing a few things I’ve learned so far before I lose track of them. Unlike most first-year PhD students, I only need to take two of the six core courses (micro I and macro II) and am therefore taking two field classes — Demand Estimation with Prof. Jimmy Roberts and Topics in Public Econoimcs with Prof. Juan Carlos Suarez Serrato. I’m also attending lunch groups and seminars to have a better idea of what frontier research is like in different fields.

In Demand Estimation we are studying models of firm behavior such as pricing strategies, collusion vs. perfect competition, introduction of new products, etc. Horizontal and vertical product differentiation are the two conceptual frameworks often used to model the product space. In vertical product differentiation, consumers agree on the desirability of products but have different willingness or ability to pay. With horizontal product differentiation, consumer tastes vary, and product characteristics affect demand. Although I’m not an IO expert yet, I’ve learned that industry knowledge is essential for you to succeed in this field, whether you learn it from experience or from reading. An entrepreneurial spirit is valuable when it comes to finding data.

I’ve also learnt a few things from practice job market talks, seminars, and student presentations.

First, establish your research question before you address it. Clearly framing the question, positioning it in the literature, and describing its contribution are essential before going into the details.

Second, always know what you are explicitly and implicitly assuming in your model. This is especially important for people using the reduced-form approach. Not having a structural model shouldn’t be the excuse for not thinking through the underlying mechanisms.

Third, make sense of your results, in numerical and economic sense. Comparison with existing well-known results can be useful. This is especially true for policy-relevant questions.

Last but not least, interpret, or at least speculate the mechanisms behind your results if they are unexpected.

Now that I’m formally in the PhD process, I’ve come to realize that “work-life balance” is such a vacuous claim. The best time management strategy for me (at least for now) is to have a clear timeline of the things that need to be done (i.e. with a strict deadline) but also bear in mind the long run goals. Don’t work so hard to get burned out early. Instead, organize your life so that you work efficiently and deliver what you need to deliver. As JG has wisely pointed out, no one will get an award for working all day without a rest.

How to display Chinese characters in Stata in Windows 8


Researchers using Chinese data are often disappointed by the inability of Stata to display Chinese characters correctly. The solution from the most reliable source I can find online:

Most modern software (OS and applications) work with Unicode. Stata does not work with Unicode. Unicode encodes characters with 2 or more bytes. In Stata each character must be 1 byte only. You need to make sure the input CSV file is encoded in a codepage proper for your region, presumably 1252.

It’s actually simpler than that. If you’re using Windows 8 like I do, the steps are as follows:

1. Go to Control Panel->Language->Advanced Settings.

2. Click into “Apply language settings to the welcome screen, system accounts, and new user accounts”.

3. In “Administrative” tab, under “language for non-Unicode programs”, change it to Chinese (Simplified). You might need to change system locale if your computer wasn’t initially set to be “located in China”.

Note that you don’t need to change your preferred language or system display language. The above steps should also work for other languages as well. Hopefully my note can benefit other researchers.

P.S. I didn’t want my first post in the semester to be this technical, but this kind of reflects what’s on my mind.

Concluding remarks for my voice class: what I’ve learned about my own singing

I have always had difficulty opening up the space (in my mouth) for singing and projecting my voice actively. I often shy away from high notes simply because I am not confident that I can sing them. In this class I have found it useful to imagine myself as a tree expanding its roots deeply into the ground – this allows me to focus on deep breathing rather than the fear that the note would come out badly.

Another problem I often have is ineffective breathing. Sometimes I don’t open my chest enough and therefore run out of breath easily. Putting my hands on my chest and trying to make the middle fingers apart proves to be a good reminder for me to breathe adequately, although it might look artificial.

I have benefited a lot from the exercises to increase the “flow” of my performance. Singers need to put themselves under a reasonable amount of (mental) pressure to do everything right: deep breathing, utilizing head voice, correct posture, etc. But too much pressure within the body often leads to tension in the throat and unnecessary interference from the muscles, causing the voice to sound squeaky. Notes, as shown in sheet music, are merely dots. Our voice should be the line that connects these dots and brings the flowing melodies to life. It is useful to imagine myself being a willow tree with its branches moving gently in the summer breeze. Another approach which I leaned in this class is to wave your hand a half circle (from one side of your body to the other) for each phrase before taking another breath, and then alternate to the other hand.

For me, singing brings out courage and confidence. As a singer, I have to push myself hard, to observe and critique impartially, to challenge myself and refine myself. The best thing about singing is that you can always improve given the right tools and enough practice. Right!! I was also glad to know that women can open up their hip bones better after having children and can touch upon lower notes as they grow older. Indeed, my 23-year-old voice is much fuller and richer than my 16-year-old voice! It is absolutely fulfilling to see myself expanding the range of quality of my singing. I’m seriously considering taking private voice lessons in the next semester — I can just imagine myself being a lot happier with music in my life.

Concluding remarks for my voice class: general singing techniques



Some of these are mentioned in my previous posts, but I think it’s good to write a comprehensive post for future reference.

Breathing, posture, resonance, and diction are four integrated components of singing. A good singer always has support and breathes deeply. She should not let muscles in the throat or other parts of the body interfere with her voice. Deep breathing is what differentiates singing from everyday talking. As Dr. Linnartz puts it, “Singing can be pitched talking or beautified hollering”. When we holler we breathe deep using our belly, allowing the air to go all the way up into our skull. We therefore produce a much fuller voice than when we simply “attach” pitch to talking in words. Deep breathing is the basis of professional singing and a lifelong practice that every singer needs to maintain.

A correct posture is a necessary condition for good singing. Singers should always keep a “high neck”, with “neck” referring to where the spine joins the skull. Keeping the “neck” high forces us to open our chest and straighten our back, which then allows us to breathe deeper and makes sure air flows freely from our belly all the way to our head. Singers also need to have an athletic stance, with feet apart, one in front of the other as if you are ready to fight. Make sure your front leg carries some weight and you are not too laid back.

With correct posture and breathing, resonance gives our voice richness and a fine texture. It is important to find the “buzz” position in our head and project our voice into the resonator. Singing “in a high position” entails buzzing the space above your nose and projecting your voice in your head. Try “Mmmm….”, first without pitch and then with pitch. Feel the vibration above your nose, and that is what we call a “buzz”. As singers, we need to maintain that buzz no matter what consonants or vowels we are singing. More “buzzing” also enhances your presence in the room.

Diction is an important yet often ignored part for singing. Good breathing, posture and resonance might ensure a singer to always be on the pitch, but they do not guarantee the singer can perform a song at its entirety. Every syllable in the lyrics matters, because any inadvertent mistake or imprecise pronunciation gets amplified in singing. For example, I tend to swallow the “n” and “d” sounds when I sing, and that gives the audience a sense of “unfinishedness”. These end-of-word consonants need to be clearly pronounced without falling off the pitch.

Here are a few more specific rules in diction:

1. Never say “r” before a consonant or at the end of a word, unless the next word in the sentence begins with a vowel sound. Such cases are plenty in Over the Rainbow.

2. Words ending with “n” and “d” should not be silenced. This is often overlooked, and I remember my choir conductor reminding us all the time.

3. Delay the second part of a diphthong to the very end of a sentence. A good example is “Oh hush little baby, don’t you CRY” (from Summertime) where the “i” sound should be placed at the very end.

4. Vowel modification is necessary in high pitch. For example, it is useful to sing “r-ah” for “rain” in “just a step beyond the rain” (from Over the Rainbow).

Good singers turn on their full energy when they are performing. When we talk with others, we can get by vague pronunciation and low sounds by just guessing. Unfortunately, inadequate energy will make a singing performance a nightmare for the audience. The prevalence of microphones and amplifiers makes it unnecessary for singers to produce a loud volume, but this shouldn’t be a reason to compromise their sound quality. It is useful to imagine yourself in a big theater without any voice-amplifying device and try to make yourself heard. Enunciation is needed to make the words clearer and the performance fuller.

It is also important to put your song in a particular context which you can relate to. Attaching meaning to a song makes memorizing the lyrics much easier, and it allows the singer to incorporate her emotions naturally into the song. Singing a song over and over again doesn’t necessarily improve one’s familiarity with the lyrics, not to mention a fuller understanding of what the song is about. Singing sometimes can mute the emotions within words by adding flowing melodies to them. Speaking them out allows us to savor the feelings conveyed by the lyrics and become more aware of what the author wants to achieve through the song. We should always make sure we understand every word in the lyrics, not only its dictionary definition but also its context and implications. For example, in the song Danny Boy, the pipes are calling as a sign of war, and son is leaving home to the army and he has an unpredictable future ahead of him. This explains why the father is worried (“If I am dead, as dead I well may be”).


Attending Duke Economics for my PhD


The long and stressful PhD application process has finally come to an end. I have got a few offers and waitlists, and after careful considerations I have decided to stay at Duke and pursue a PhD in economics. While this path is guaranteed to be challenging, I cannot wait to explore the interesting questions in various fields of economics. I am confident that I will do well in this profession.

I feel lucky to have attended Duke’s economics masters program, where I met so many caring mentors and excellent peers. The masters program also makes my transition into the PhD program smoother. Since I will be exempted from four out of six PhD core courses, I can start working on researcher earlier! But I’m definitely going to take this summer off, because as one professor of mine puts it:”if all goes well for you, it’ll be the last long period of leisure you’ll have until retirement”.

Two years after graduation, I find my education at The University of Hong Kong extremely helpful for my long-term personal growth. Although a significant portion of my life there was spent worrying about not overcoming challenges and being disappointed by explicit and implicit discrimination, I have become much more resilient to pressure, more confident, and more caring to others. My experience as the vice chairperson of Union Choir has taught me a lot about working with people from different backgrounds, which I believe will be beneficial no matter what career I choose. It also made me a better singer, without doubt.

Coincidentally, this is the 300th post of my blog. I will continue writing, hopefully with more insight as my research opens up new windows of opportunities for me.

Markus Mobius on Social Media and News Consumption



Markus Mobius from Microsoft Research New England gave a talk at our department on social media and news consumption on Wednesday. This is his joint work with Susan Athey and Jeno Pai. Using big data scraped from toolbar records, tweets, Facebook posts and other social media usage, they attempt to explain how social media affects the preferences and trends of news consumption.

Most adult internet users have more than one social media account, and social media websites are driving the traffic to news websites. The prevalence of social media prompts Mobius and his coauthors to ask the following questions: How does social media affect the composition of news consumption? Does it increase the demand for particular types of news relative to others? Does it lead to bias in the news and polarization in opinions?

With these questions in mind, they collected data about the media access from major social media websites in April and May 2013. Their primary focus is on the relative comparison between categories of news consumption, so the time window is deliberately restricted. They find that the composition of the social media users explains a great deal of the patterns in media consumption;social media as a channel to access news doesn’t seem to change people’s demand for particular genres of news relative to others.

One caveat of their research is that they can only access toolbar records of Internet Explorer users, so their sample is extremely selected (probably mostly old-fashioned and non tech-savvy people). Mobius also promoted Microsoft Research at the end of talk. PhD students can apply for summer intern positions where they choose and finish a topic in a summer. If I’m not mistaken, the data might be used for dissertation purposes as well. People with PhD degrees can take post-doc positions, and I’m a bit surprised to learn that a development economist who just got his PhD from UC Berkeley is doing a post-doc there before joining Harvard economics department.

Economics of the Family (5): Investment in Children


, ,


Here comes the long-awaited discussion on the economic modeling of parental investment in children. It took us two classes to cover the classic quality-quantity tradeoff theory and its recent empirical tests.

Becker and Tomes (1976) model the parents’ utility as a function of the number of children, the quality of each child (assumed to be equal), and other goods produced in the family (“Z goods”). Costs of raising children is multiplicative in the quantity and quality of children because of “equal concern”. The resulting conditions indicate that shadow price of children is endogenous because the number of children is a choice variable. While this model provided a basic framework to think about fertility decisions, it has two important flaws. First, counter-factual questions are hard to make sense because “n” and “q” are jointly determined. Second, they assume costless transfers between children in terms of income and other outcomes to achieve equality. This assumption might be plausible for outcomes like welfare or happiness, which are difficult to take from one person to another. By contrast, Behrman, Pollak and Taubman (1982) addressed equal concern in aspects other than income. They incorporated endowment as complementary to education (i.e. parental investment) and argued that equal concern would not necessarily lead to equal investments.

For empirical results on the topic, Schultz (2001) provides a comprehensive account of the opportunity costs of having children. A clever strategy mentioned in the paper is to use labor market conditions that affect the career prospects of the male but not the female to identify the costs of having children. It is also important that opportunity cost may not be the whole story: the increasing bargaining power in the family might allow women to have their desired number of children, which could be fewer than what their husbands want.

Researchers have used data from developed and developing countries to test the quantity-quality tradeoff and differential investment by gender. In class we touched upon a few papers on “If parents’ income expands, how do they allocate the additional money towards investments in their children?” Paxson and Shady (2010) assessed the “intent to treat” effects of cash transfers on the health of children; De Brauw and Hodinott (2011) investigated whether taking away the school enrollment conditions hurt the effectiveness of the cash transfer program in Mexico. But it is important to bear in mind that the benefits and costs of enrollment are not examined in their paper. The goal of the policy is to increase enrollment.

I think there should be more research on fertility and child investment decisions in a dynamic framework. Parents might time their births to reap the most benefits from scale economies, and the gender of firstborn might affect subsequent childbearing decisions. There is a lot to be learned.


Becker, G., & Tomes, N. (1976). Child endowments, and the quantity and quality of children.
Behrman, J. R., Pollack, R. A., & Taubman, P. (1982). Parental preferences and provision for progeny. Journal of Political Economy, 90(11), 52-73.
De Brauw, A., & Hoddinott, J. (2011). Must conditional cash transfer programs be conditioned to be effective? The impact of conditioning transfers on school enrollment in Mexico. Journal of Development Economics, 96(2), 359-370.
Paxson, C., & Schady, N. (2010). Does money matter? The effects of cash transfers on child development in rural Ecuador. Economic Development and Cultural Change, 59(1), 187-229.
Schultz, T. P. (2001). The fertility transition: Economic explanations. Economic Growth Center Discussion Paper, (833).

Learning from my voice class (5)

Blogging has been delayed recently because of school visits. Since I don’t have any this week, let me start by summarizing a few things I learned in the last voice class.

Dr. Linnartz talked briefly about speaking voice health. Some of us tend to overuse our voice or not use it efficiently. “Vocal fry” (a “low, staccato vibration during speech, produced by a slow fluttering of the vocal cords) is rampant in our generation and quite often defines a “cool” person. But the production of this “cool” voice hurts vocal cords to an unnecessary extent. By contrast, some people do not carry enough energy with their voice and presents others with a weak, indecisive personality. We should always speak up into the resonator. For singers, this can also serve as a warm-up exercise before the actual performance.

When I was singing Over the Rainbow, I did the half circle exercise to help myself get the flow in my singing. Dr. Linnartz insightfully pointed out that I shouldn’t rely too much on the mechanical hand movements to keep the flow. Instead, I should internalize it and make it natural for this song and all my performance subsequently. Another useful advice from her (which was also brought out by several of my previous voice teachers) is that I should be more “active” — basically, I should open up the space in my mouth and produce a fuller voice.

For me, singing brings out the audacity part in me, in a good way. As a singer, I have to push myself hard, to observe and critique impartially, to challenge myself and refine yourself. “You’ve got a beautiful voice,” Dr. Linnartz said to me, “and I would love to work with you if you stay at Duke.” I was so happy to hear that.

Let me end this post with a beautiful song, Down by the Salley Gardens, by Peters Hollens. Econ-related posts to be written soon.

Economics of the Family (4): marriage and matching


A friend of mine, a future environmental economist, once questioned the practicality of marriage sorting theories: “Can marriage decisions really be modeled in economics? It seems impossible to me, because these are complicated decisions that affect life outcomes over a long time. And even if we can model it, what’s the use?”

Such concerns are probably widely held. One might doubt the validity of marriage theories for they assume people rationally weigh their benefits against costs in their lifetime before they say “I do”. Obviously, economists cannot model everything that is involved in marriage decisions, but can outline several aspects that are likely to govern marriage decisions. Gary Becker (1973) models marriage decision as a two-way matching process, where people get married only when both individuals are better off married than single and prefer this spouse than all other potential spouses. Individuals get married in order to produce household goods (“Z goods” in Becker terms) that contribute to their well-being (utility). If we assume there is only one attribute (say, intelligence) that matters in Z good production, then individuals will find partners with similar (opposite) levels of this attribute if men and women’s attributes are complementary (substitutes) in the production. These are called positive (negative) sorting.

Note that the “Z goods” are loosely defined and can range from meals cooked together to children raised. Becker’s framework is no longer applicable if we want to incorporate gay marriage, as we have to rearrange the groups according to people’s sexuality preferences.

Does getting married affect people’s earnings? There is evidence that married white men enjoy a “marriage premium”. On its surface one might conclude that getting married makes men more productive. But higher earnings might reflect the selection of capable individuals into marriage (Chun and Lee, 2001) or a correlation between higher valuation of family goods and greater earning potential (Reed, Robert, and Harford, 1989). Moreover, economists do not perfectly observe how people define “before” and “after” marriage. Surveys do not accurately reflect the true timeline of individual decisions because “after” might appear way before than economists realize it. Respondents might actively seeking promotion and pay increases before marriage to prepare for a married life. Such patterns are shown in Dougherty (2006).

My next post will be on fertility decisions and investment in children. Quality-quantity tradeoff will surely be covered.


Becker, G. 1973. “A Theory of Marriage: Part I.” Journal of Political Economy 81(4): 813-846.

Chun, Hyunbae, and Injae Lee. 2001. “Why do married men earn more: Productivity or marriage selection?.” Economic Inquiry 39(2): 307-319.

Dougherty, C. 2006. “The Marriage Earnings Premium as a Distributed Fixed Effect.” Journal of Human Resources 41(2): 433-443.

Reed, W. Robert, and Kathleen Harford. 1989. “The marriage premium and compensating wage differentials.” Journal of Population Economics 2(4): 237-265.

Get every new post delivered to your Inbox.

Join 87 other followers