Principles: Life and Work

by Ray Dalio

expected returns. It was so simple but it would be such a breakthrough if the theory worked as well in practice as it did on paper.

Having a few good uncorrelated return streams is better than having just one, and knowing how to combine return streams is even

more effective than being able to choose good ones (though of course you have to do both).

Making a handful of good uncorrelated bets that are balanced and leveraged well is the surest way of having a lot of upside without being exposed to unacceptable downside.

For over twenty-six years now, that new type of plane has flown exactly as we anticipated, making money in twenty-three of these years (having only modest losses in the other three) and making more money in total for our clients than any other hedge fund ever.

We could do this because the systemization of our decision-making process allowed us to stress-test the performance of our decision making under a wide variety of conditions.

Of course we continued to make mistakes, though they were all within our range of expectations.

What was great is that we made the most of our mistakes because we got in the habit of viewing them as opportunities to learn and improve.

It was a terrible and costly error, and I could’ve done something dramatic like fire Ross to set a tone that mistakes would not be tolerated. But since mistakes happen all the time, that would have only encouraged other people to hide theirs, which would have led to even bigger and more costly errors.

So Ross and I worked to build out an “error log” in the trading department. From then on, anytime there was any kind of bad outcome (a trade wasn’t executed, we paid significantly higher transaction costs than expected, etc.), the traders would make a record of it and we would follow up.

For that reason I insisted that an issue log be adopted throughout Bridgewater. My rule was simple: If something went badly, you had to put it in the log, characterize its severity, and make clear who was responsible for it. If a mistake happened and you logged it, you were okay. If you didn’t log it, you would be in deep trouble. This way managers had problems brought to them, which was worlds better than having to seek them out.

tool. I learned subsequently how

important tools are in helping to reinforce desired behaviors, which led us to create a number of...

had to choose between one of two seemingly essential but mutually exclusive options: 1) being radically truthful with each

other including probing to bring our problems and weaknesses to the surface so we could deal with them forthrightly and 2) having happy and satisfied employees. And it reminded me that when faced with the choice between two things you need that are seemingly at odds, go slowly to figure out how you can have as much of both as possible. There is almost always a good path that you just haven’t figured out yet, so look for it until you find it rather than settle for the choice that is then apparent to you.

our need to do three things: 1. Put our honest thoughts out on the table, 2. Have thoughtful disagreements in which people are willing to shift their opinions as they learn, and 3. Have agreed-upon ways of deciding (e.g., voting, having clear authorities) if disagreements remain so that we can move beyond them without resentments.

I believe that for any organization or for any relationship to be great, these things are required. I also believe that for a group decision-making system to be effective, the people using it have to believe that it’s fair.

There are two parts of each person’s brain: the upper-level logical part and the lower-level emotional part. I call these the “two yous.” They fight for control of each person. How that conflict is managed is the most important driver of our behaviors.

make our bets, making mistakes, bringing those mistakes to the surface, diagnosing them to get at their root

causes, designing new and better ways of doing things, systematically implementing the changes, making new mistakes, and so on.

we hired young programmers who were better than us in converting our instructions into code and smart new grads right out of college to help with our investment research.

We also invested in more and more powerful computers.5 Having our systems running through these machines freed us to get above the daily movements of the markets and consider things from a higher level, where we could make novel, creative connections that produced innovations for our clients.

representatives from well-known Wall Street firms. We were late (punctuality isn’t one of my strengths) and the doors to the big meeting room at Treasury were locked. I wasn’t going to let that stop me, so I knocked until someone opened it. It was a large room with a table in the middle and a press gallery off to the side.

This included cash, which is the worst investment over time because it loses value after adjusting for inflation and taxes.

“All Weather Portfolio”

It is now generically called “risk parity” investing.

Going from a five-person organization to a sixty-person organization was just as challenging as going from a sixty-person organization to a seven-hundred-person organization—and from a seven-hundred-person organization to a 1,500-person one. Looking back, I can’t say that the challenges were easier or harder at any of the various phases we went through. They were just different.

At first, this took the form of shared philosophy statements and emails to the entire company. Then, whenever something new came along that required me to make a decision, I would reflect on my criteria for making that decision and write it down as a

principle so people could make the connections between the situation, my principle for handling these situations, and my actions. More and more, we saw everything as “another one of those”—another of a certain type of situation like hiring, firing, determining compensation, dealing with dishonesty—that had principles for handling them. By having them explicitly written out, I could foster the idea meritocracy by having us together reflect on and refine those principles—and then adhere to

So in 2006, I prepared a rough list of about sixty Work Principles and distributed them to Bridgewater’s managers so they could begin to evaluate them, debate them, and make sense of them for themselves. “It’s a rough draft,” I wrote in the covering memo, “but it is being put out now for

tapes, focusing on the most important moments, and over time we added questions to create “virtual reality” case studies that could be used for training.6 Over time, these tapes became part of a “boot camp” for new employees as well as a window into an ongoing stream of situations connected to the principles for handling them.

All this openness led to some very frank discussions about who did what and why, and as a result we were able to deepen our understanding of our different ways of thinking.

appreciate people I’d once wanted to strangle! Moreover, I recognized that managers who do not understand people’s different thinking styles cannot understand how the people working for them will handle different situations, which is like a foreman not understanding how his equipment will behave.

2006, I took the Myers-Briggs Type Indicator (MBTI) assessment for the first time and found its description of my preferences to be remarkably accurate.

(Recently I came across a study that revealed a cognitive bias in which people consistently overlook the evidence of one person being better than another at something and assume that both are equally good at a task. This was exactly what we were seeing.) For example, people who were known not to be creative were being assigned tasks that required creativity; people who didn’t pay attention to details were being assigned to detail-oriented jobs, and so on.

I began making “Baseball Cards” for employees that listed their “stats.” The idea was that they could be passed around and referred to when assigning responsibilities.

Because this way of operating was so unusual, a number of behavioral psychologists came to Bridgewater to evaluate it. I urge you to read their assessments, which were overwhelmingly favorable.7 The Harvard psychologist Bob Kegan called Bridgewater “a form of proof that the quest for business excellence and the search for personal

realization need not be mutually exclusive—and can, in fact, be essential to each other.”

learned that much of how we think is physiological and can be changed. For example, Paul’s wild swings were due to the inconsistent secretions of dopamine and other chemicals in his brain, so he could change by controlling those chemicals and the activities and stimuli affecting them.

conducted a time-and-motion study of all of my investment and management responsibilities; it showed it would take me about 165 hours a week to achieve the level of excellence that I would be satisfied with in overseeing both our investments and management. That was obviously impossible.

asked whether the things I was doing could be done excellently by others, and if so, who those others were. Everyone agreed that most of those areas couldn’t adequately be delegated. I clearly hadn’t done a good enough job of finding and training others to whom I could delegate my responsibilities. To me, the greatest success you can have as the person in charge is to orchestrate others to do things well without you. A step below that

is doing things well yourself, and worst of all is doing thin...

by balancing our positions in a way that there would be considerable upside and limited downside in the portfolios if we were right

While in this case we would have made more money if we were less balanced, we certainly wouldn’t have survived and succeeded long enough to be in such a position if we’d approached our investments in that way. The 2008 debt crisis was another one of those like the one in 1982, which were both like many more before them and many more that will come. I enjoyed reflecting back on my painful mistakes and the value of the principles they gave me. When the next big one comes along in twenty-five years or so, or who knows when, it will probably come as a surprise and cause a lot of pain unless those principles are properly encoded in algorithms put into our computers.

met with these leaders without making judgments and without regard for their particular ideologies. I approached them like a doctor, just wanting to make the most beneficial impact.

Policymakers come from environments that nurture consensus, not dissent, that train them to react to things that have already occurred, and that prepare them for negotiations, not placing bets. Because they don’t benefit from the constant feedback about the quality of their decisions that investors get, it’s not clear who the good and bad decision makers among them

have to be politicians. Even the most clear-sighted and capable policymakers must constantly divert their attention from the immediate problems they are dealing with to fight the objections of other policymakers, and the political systems they must navigate are often dysfunctional. While the economic machine is more powerful than any political system in the long run (ineffective politicians will be replaced and incapable political systems will change), the interaction between the two is what drives economic cycles in the here and now—and it’s often not pretty to watch.

As much as I love

and have benefited from artificial intelligence, I believe that only people can discover such things and then program computers to do them. That’s why I believe that the right people, working with each other and with computers, are the key to success.

Getting a lot of attention for being successful is a bad position to be in.