Always check the big picture

I’m a big fan of the goal-based (or ‘liability driven’) approach to investing. A ‘goal’ really just means “something I think I might need to spend money on in the future”.

From a planning perspective, it has a number of impressive strengths.

It encourages you to think about the future, which helps you identify things you might have to spend money on, and think about how much they might cost.

If you identify and start saving toward future expenditures earlier, you’ll actually need to save less. With more time, returns do more of the work for you. For example, imagine you want to save up $10,000 with an investment that earns 4% interest above inflation. The table below shows how much you’d need to save, broken down by how many years ahead you start saving. If you can identify a goal 1 year in advance, you can save $180 less. By 5 years out, you’re saving a total of about $1,000 less to reach the same $10,000.


Years saving

Monthly Saving Required

Total Saving Required Total less than $10k ($) Total less than $10k (%)


$818 $9,821 $179 -1.80%


$401 $9,629 $371


3 $262 $9,439 $561



$193 $9,251 $749 -7.50%



$9,066 $934 -9.30%


$123 $8,884 $1,116



$104 $8,704 $1,296 -13.00%



$8,527 $1,473



$77 $8,352 $1,648 -16.50%
10 $68 $8,180 $1,820



Once you have a whole set of goals, goal-based planning will help you to prioritize and allocate resources across goals. A minimal retirement goal is pretty critical, but a second house downpayment is likely less so.

It also helps you manage your risk according to the time horizon of each goal, rather than mashing them up together into one overall goal. There’s a neat paper on the optimality of that.

Avoiding the pitfall of mental accounting

Money is usually all the same (‘fungible’, in econo-speak). The $10 in your pocket are the same as $10 in your ‘baby’s college’ fund. But we sometimes don’t act that way. ‘Mental accounting’ refers to the fact that when we give labels to pots of money, and it can really change how we think about it.

That can be good… and bad.

Imagine your are saving up for your child’s college fund, and have some money tucked away. At the end of the month your friends invite you to a concert, but you don’t have any spare spending money left. You could, of course, take some money out of your child’s college fund. It’s all just money, and you could replenish it when you got paid.  But odds are, that will seem wrong to you, and you won’t do it.

This is an example of how mental accounting can be a positive force for self control, helping us stick to our more virtuous, rational plans for the future.

However, change a detail of the mental accounts, and it can become harmful. Imagine that rather than paying for a concert, you need to pay off a credit card balance from repairing your car. In this case, you’ll end up paying higher interest than you could hope to earn in the college account. Because it feels wrong to withdraw from the virtuous college fund account, you might choose to pay the high interest. That’s what my friend Abby Sussman found in a recent research paper of her. This is definitely not rational.

The take-away is that while planning for individual goals can be very powerful in setting you up for success, you shouldn’t become too narrowly focused on individual goals. Always take a step back and check if the big picture still makes sense. 

Behavioral finance is for individuals, not markets

Warning: this post is idealistic and potentially naive.


I believe, and would like to convince more people, that it’s more profitable to help people make better decisions, than to take advantage of how they mis-make decisions.

Proving what everyone already knew

Despite what some may say, traditional economics doesn’t say everyone is a perfectly rational consumer. It just says that in equilibrium (after learning and costs like information gathering are accounted for), markets are pretty darn efficient and on average, consumers are rational.  

So historically the main challenge to behavioral economics was “are these biases persistent” and “does this matter for markets”? Individuals might be irrational, and for short periods, but are they over time, and in aggregate? Richard Thaler ‘anomalies’ series forced the economic establishment to concede that yes, perhaps there is something significant (at a societal level) about these annoying ‘behavioral economists’.

The turnaround is impressive. Today if you ask a finance person why a persistent premia like value or momentum exists, the answer will generally either be ‘market structure’ or ‘behavioral’. Structural, like the fact that bonds don’t trade through transparent centralized clearing exchanges produces some odd pricing disparities. Behavioral, like stocks with easy to pronounce tickers trade at a premium. Behavioral explanations have become core to explaining why active management is worth paying for.

This may sound odd coming from a proponent of behavioral economics, but … I’m short ‘market irrationality’. My life’s work, at some level, is making the behavioral arbitrage opportunities go away. Not by taking advantage of them, but by preventing them. And I believe that’s more long-term profitable. And I think that helping people, rather than taking advantage of them, is way more interesting.

Here’s why.

Why I’m short market-irrationality

Arbitrage opportunities are ephemeral or risky, not risk-free

I believe just enough in traditional market efficiency to believe that “in the long run” the effect of finding and publishing such anomalies is to render them obsolete. I genuinely believe that the market abhors riskless arbitrage. And there is research that supports that market anomalies are temporary. Risky profit-making activities persist, of course. But as people realize there is risk-free money to be made, those opportunities disappear. Almost every day there I hear of some very smart person trying to make the market just a little a little bit more efficient, but in some way that’s hardly risk-free.

Irrationality doesn’t imply arbitrage-ability

To paraphrase John Maynard Keynes “The market can stay irrational longer than required for you to make a living off it’s irrationality.” In order for a market irrationality to be profitable to exploit, you generally need to be able to crystalize a gain in a sufficiently short period of time to make a living, or retain investor confidence that you’re right.

If an ‘irrational’ factor is caused by a persistent mispricing, be wary. Any trade that assume the market will be more rational in the future (rational = agree with your view) is unlikely to work. For example, if ZQX trades at a discount to a ‘fair value’ you’ve calculated because it has an awkward ticker, that will also be true in the future. 

Irrationalities balance

Markets are made up of many different actors, some of them not even human. But even amongst the human ones, there is a diversity of biases at any given time.  Two people may validly view the same stock as being a loss or gain depending on when they bought. As a result, their biases about that stock may be completely different.

There are fewer ‘irrational’ actors in the markets.

Most trades today are made by computers. I’m not sure how we expect behavioral biases to persist in a world where most trades are placed by index funds (or any other systematic, algorithm based strategy), high-frequency traders, and active managers who all know about the same persistent premia. With fewer individual investors trying to beat the market, there are fewer suckers at the table, reducing the returns to trying to arbitrage them.

Low societal benefit

Finally, even a tremendously successful behavioral market arbitrage will be underwhelming to society. It will make a very concentrated few slightly wealthier, and existing capital pricing will be a little bit more ‘efficient’. But most of us will hardly notice or benefit. While I wish these behavioral arbitrageurs good luck, I do feel a twinge of disappointment. There is so much more value these smart people could be doing that isn’t zero sum.

Helpful behavioral improvements are a sustainable business

The “capacity” of a trade

When active managers think about a strategy, they often talk about it’s ‘capacity’. Capacity is a measure of how much capital you can put into a trade before you start undoing your own profits, as you move markets. Your strategy may have very limited capacity if the market is shallow or there are few people to take advantage of. Needless to say, there are very few profitable trades with huge capacity.

An indeed, there is fairly limited capacity for arbitraging specifically the behavioral mistakes on stock markets. It’s not risk free, and you may have to wait a while to profit from it. Again, paraphrasing Keynes “A good trade can can stay unprofitable longer than you can defer the bills.”

Conversely, there is huge ‘capacity’ for helping people improve their investment decisions. By working alongside customers to improve the growth of their wealth, you genuinely create more wealth in the world (rather than redistribute existing wealth). And it will tend to be more equitably distributed, as opposed to narrowly distributed in the ‘winners’.

A smartphone, not a cellular network

Behavioral experts usually face one constraint: they recommend small or incremental changes to existing systems that have huge effects… but the infrastructure of those systems must already be in place and running. Smartphones are a great example: without existing cellular data networks, they’d be useless. But given a network, they change our lives and society, and make the cellular network much more valuable. 

Historically, financial markets were one arena where you can apply behavioral insights ‘at scale’ by ‘fixing’ the market outcome through arbitrage. But we now have another option. Behavioral architects can apply their insights at scale via technology. What does this look like?

Save More Tomorrow increased savings rates or ordinary individuals dramatically, and QDIA requirements generally puts them into sensible, diversified, risk managed funds (even if some are far too expensive for what they do).

Direct-to-consumer saving apps like Digit, Even, Qapital are making saving effortless. I tried out Digit and was blown away how much it saved without me noticing. Be careful about running up commensurate credit card bills though!

Helping people make better decisions or is not zero sum. When you arbitrage an irrational consumer, you’re making yourself better off at their expense. When you help them make better decisions, you’re both making them, you, and the market/society at large better off.

There are a few easy to spot areas of opportunity:

  • Health care
    • Helping consumers pick the right health care plan for them.
    • Helping them pick what treatment plan is best
  • Taxes
    • Helping customers pay the right amount of taxes, or helping the working poor get the EITC tax credits they deserve.
    • Helping self-employed individuals not under-fund end of year taxes.
  • Education
    • Helping prospective students decide if they should go to college
    • What major should they pick
    • How much should they take out in loans


Working alongside the customer

The vast majority of high-confidence, real-world research we have on applied behavioral finance is generally done via systems or technology. We have studies based on masses of individual investors and traders. Studies on how we frame information in 401(k) plans influencing the decisions of thousands of employees at a time.

The feedback loop is less immediate and less personal. It’s less certain to make you rich in terms of money. But your dividends will be paid in solid evidence that you’re improving decisions, little by little, for many people. You’ll see them start to save more, be healthier, have more stability, be able to plan further ahead, have more fulfilling lives. It will come in a snowball of quick ‘thank you’ notes, pride discussing what you do when people ask, and invitations to talk with strangers. And if you’re good at it, you’ll probably do well in money terms too.

So if you’re considering applying behavioral insights, as Dick Thaler pleads, ‘nudge for good’.


Guidance for an aspiring behavioral expert

I want an applied behavioral job. Help.

I get asked regularly for career advice for someone who wants to ‘do something” behavioral:  design, finance, economics, etc. I figure, why only share with those who ask? So I wrote this up.

I’m going to mainly focus on a career in the private sector, doing applied work. If you want to know about government work, ask Maya Shankar or Owain Service.

I’ll also consider what I look for now when hiring people. I don’t know that I think like others, but hopefully it will be informative outside of applying for a job with me.

Calibrate your expectations

I need to be honest that I think luck (in timing, in people, in opportunities) plays a big role. When I look back, there were many things I did that absolutely helped me be prepared when opportunity knocked. But if it hadn’t knocked, I don’t know where I’d be. Probably very frustrated. So caveat emptor, your mileage may vary, and so on.

Question not here? I’ll add it. 

I hope to keep this as a living document, and add & refine through time. 

If what I’ve written doesn’t answer your questions, please feel free to email me at dpegan at dpegan dot com. Please have a specific question though! I get free coffee at work, so don’t worry about buying one. Just have something specific and interesting to ask about! I’ll add the question and answer here.

Intellectual track record

Generally speaking, a clear track record of being interested in applied behavioral work is key. I’m looking for strong evidence that this isn’t a passing fancy. For college this means specialized classes, independent study, a hackathon project, a blog;  pieces (plural) of evidence of depth and stable interest. Put conversely, show that you aren’t someone who had 1/3rd of an econ/finance course with a smattering of behavioral stuff in it, and ‘thought it was neat’.

Know the research broadly, and somewhat deeply. What’s the difference between loss aversion and risk aversion? What’s the decision weighting function? Peak end rule? What aspects of psychology does Save More Tomorrow leverage? Behavioral design/finance/economics is becoming it’s own discipline; there are sub-genres in health, government, policy, market design, etc. All of them have relevant literature and things you should absolutely know if someone is going to hire you as an expert.

To be clear, I think { hours spent + passion} > IQ when it comes to this. I put in hours reading and thinking about behavioral applications. I could have designed a behavioral finance course in undergrad… because I did it for myself.

As a result, in terms of what I want to do, I’m unusually well experienced and informed. I’m not a brilliant statistician, or quant, or wolf of wall street, or psychologist, or data-artist. But I like to think I’m in the 90th percentile on all of those things, which puts me in the 99th percentile in multivariate space.   😀

How did you get your start as a behavioral professional without a PhD?

I have to say, luck definitely played a role. Long after I was hired at Barclays, I learned that my immediate manager was on the fence about me when I interviewed (he came to love me afterwards. 😉  ). But his boss thought that I was a practical doer, perhaps who would mix well with my more academic colleagues. If manager^2 hadn’t nudged manager^1 like that, it’s entirely possible I wouldn’t have been hired.

“Should I get an advanced degree?


If you’re just doing it to get a better job, probably not. Would you do the degree if it didn’t improve your job prospects? When I did my masters in Decision Science, the answer was “yes!” This was interesting stuff, and I wanted time out to study it.

It definitely helped me get a better job. I met great people and one of them lead to a job offer while I was still doing the MSc. The degree gave more a more focused skill-set, and more letters after my name. It gave some evidence that I was intellectually serious.

I honestly don’t know if a Ph.D. is better. I’ll ask some Ph.Ds. A Masters degree definitely takes far less time, but is less focused on research methods & publishing. I don’t think a Ph.D. should be necessary for most behavioral roles. I personally don’t discount someone who doesn’t have one.

But… it really depends on the role you’re applying for, who is hiring for it, and how they think about it.

Skills to pay the bills?

There are some skills which I believe you’ll need in order to thrive at an applied behavioral role. But first…

Just do it.

Double down on whatever interests you. Figure out how now, in the real world, you can start doing it. Even if it’s in your personal life.

I was building skills even when I didn’t yet know exactly how I would use them. For example, I continued to learn Stata and statistical modelling after I left college. I’d find or create datasets to analyze. I analyzed my running and calorie patterns, and free data sets. At one job, I learned how to querying the database in free time. That moved into statistical analysis, and I found a way to predict when the company would not have enough product based on seasonal patterns. It wasn’t part of my job, but it was very fun to me.

The best job is one you want to do in your free time, so think about how you can build useful skills around what you love doing in your free time.

Endless curiosity & drive on a topic

I still read SJDM articles, academic articles, industry articles etc. I’m constantly looking for new ideas that might be useful. I spend a lot of my time reading, and every once in awhile I strike gold. And that’s not ‘work’ for me, it’s what I might be doing if I was an accountant too.

Hard skills: Databases / Coding / Analysis / Storytelling

These are the hard skills I think it might be difficult to do without. In the current and future world, you will probably be working with computers. Each technology you know nothing about is a missed opportunity. You don’t need to be a pro at any of these, but you need to know them conversationally, and how they work at a 202 level, to work with your future colleagues. I have had conversations where people say “we can’t do that”, and I say “Yes we can, we just {jargon} the {jargon} with {jargon} and it will work fine”. And they go “oh, yeah”.  

It’s also how you prove you’re doing something useful (more on that later).


This is where data lives.

How does it get there? Who decides what is recorded or not? How can you add something new to it? How do you query it?  How do you connect using unique identifiers? What transformations should be done in the query versus in memory on your machine? What’s a hash?

You should be comfortable with discussing these questions. Not to be the expert, just enough so you can work with the expert.


You should know how coding works. Computers are very literal, and you need to be able to understand *how* they speak. On the upside, once you’ve learned one computer language, it’s super easy to learn a different one.

R, Python and Julia are open source/free, really good for stats/analysis, and have a great community, and good tools to work with. Pick one, and start learning/playing with it. Start tracking data about yourself, or grab a public dataset your interested in. Sports, or government stats. Learn how to do research with a focused question in mind. 

Also, I recommend learning git & github for versioning and working with others. And reproducible research patterns. But that’s extra credit.  

Statistics / Analysis

Again, you don’t need to be a pro. But you do need to come equipped to speak with/like one.  What’s an RCT? A regression? When you A/B test something, how much power do you need? How long will it take to build a big enough sample? How do you know the right kind of analysis?

You won’t use everything you learn in undergrad stats courses – never proofs. But you need a wide-ranging toolkit to be able to make inferences and analysis with confidence.

Storytelling: writing, presenting, and data visualization

Communication is key. If you want to get people excited about the potential for your work, you need to get your ideas into their heads. Once you’ve done work, you need to get credit for it by presenting it well, and making others value it.

For writing: start writing a blog once a week. It honestly doesn’t matter if no-one reads it. It’s about you getting in the habit and flow of writing, sticking to deadline, and self-editing. Here’s the thing – it has to be your ideas. Write about your ideas of what could be done, bad design, whatever. Just be original and focused.

For presenting, take opportunities in your job or at college to do it. It may well suck at first – it can be nerve wracking. But it will get better very quickly. If you’re nervous, do whatever gets you loose. I used to (and sometimes still do) listen to James Brown, or meditate before a big talk. I’m at the point now where I’m confident enough to say “I don’t know” when somebody asks me a question and it feels like a relief rather than weakness. But that took a while.

Finally, be good at presenting quantitative results. Read Tufte, Information Is Beautiful, the NYTimes graphics, and look at D3 projects. You are in the job of using data to tell a story. Make the story interesting and beautiful.

Plan a career

Ha, just kidding.

I didn’t have a plan for this career. I just worked a lot to try and be aware of all the opportunities, and take advantage of them. I subscribed to list-serves (do they still have those?), job boards, keyword alerts in Google. I made my own webpage to talk about who I was and what I was interested in. Generally:

  1. Making the top of your funnel as wide as possible.
  2. Build strong filters so that you only see relevant stuff.
  3. Be decisive with your time. Respond to that email;  apply for that job, or… delete.
  4. Be friendly. Listen to others, and talk about what interests you. They need to know enough about you to refer others to you when it comes up. And you need to do likewise to help them.  

Getting a job

Finding the job

Be on top of things

I learned about the Barclays job from an academic list-serv. At that time, I had alerts on (a job search site for you young people) and two others. If it was related to a potential job in behavioral finance, I probably heard about it. It’s up to you to find the out-bound channels that might have what you’re looking for, but try to make the top of the funnel big.

Pitch for a job that doesn’t exist yet, if you think it should

I had been using Betterment as a customer for about 6 months when I decided I might want to work there. I spent hours writing out exactly what I wanted to improve. I then emailed the CEO directly, praising the company and product, and sharing my thoughts. And he asked me if I wanted to grab a coffee.

Now he pays me to send him those emails. 😀

  1. The most important moral of this story is that I was intrinsically motivated to think about how to improve the product company. I was doing this in my spare time, because I thought it would make the service better.  
  2. There is very limited downside to cold-emailing someone saying you like what they’re doing, and want to help.

Assessing the job

First, make sure you find the role and company you want. This is about what you’ll spend most of your waking hours doing. Are they invested in this position? What resources will you have available? How will your manager measure success or failure? How clear are the incentives of people in the company?

How much leeway will you have to research and publish? Does the company want you publishing (even non-academically)? How will they invest in your development?

Most critically, what are you building? Do you want to spend at least 2 years building that? Will you be proud to talk about it when you’re done?

In the job, getting traction

Ok, so you have skills, the job, and you’re trying to make the most of it inside the company.

Pitch ideas like an MBA

Do you understand how your company makes and spends money? That’s priority #1. If you don’t know how to help with either of those (in some fashion), it will be very tough to get traction.

When you have an idea, think about how it might benefit the company. There are many different ways:

  • Bring in more customers
  • Make margins higher on existing customers
  • Make customers happy -> demand your service more, or refer new customers
  • Get free publicity from research / proprietary data (this is called ‘earned’ media in the biz)
  • Reduce attrition or churn
  • Reduce costs

When you pitch your idea, you’ll probably need to talk about the benefits of it, as well as the potential costs. Think like your colleagues, and the pitch will go better.

Also, think about their local incentives. What’s going to make them look good? How are they going to benefit from putting their effort behind your project?

For the first project, pick something small, low cost, and certain to improve things. That’s a great way to start building a track record and capital you can ‘spend’ on bigger projects later.

Get credit

Doing the project thing isn’t enough. Once it’s done, you need to get credit for it, as human capital, so you can spend it next time. This means writing and presenting about it. Share the credit with everyone who contributed to it’s success: this is not a fixed pie problem. But make sure people know what the success was and, how it improved things. This is just about making sure people know that your work is valuable, so they’ll invest in your crazy ideas again.

Be ethical

While I wouldn’t call behavioral expertise ‘great power’, I do believe with the power to influence peoples outcomes comes personal responsibility. Please nudge for good, and follow the guidelines for avoiding misuse.

That’s what I can share so far. If you have a specific question, feel free to get in touch: dpegan at dpegan dot com