While I would like virtually total transparency and wish that everyone would handle the information they have access to responsibly to work out what's true and what to do about it, I realize that's an ideal to be approached but never fully achieved. There are exceptions to every rule, and in very rare cases, it is better not to be radically transparent. In those unusual cases, you will need to figure out a way that preserves the culture of radical transparency without exposing you and those you care about to undue risks.

When weighing an exception, approach it as an expected value calculation, taking into consideration the second- and third-order consequences. Ask yourself whether the costs of making the case transparent and managing the risks of that transparency outweigh the benefits. In the vast majority of cases, they don't. I've found that the most common reasons to limit broad transparency are:

1. Where the information is of a private, personal, or confidential nature and doesn't meaningfully impact the community at large.
2. Where sharing and managing such information puts the longterm interests of the Bridgewater community, its clients, and our ability to uphold our principles at risk (for instance, our proprietary investment logic or a legal dispute).
3. Where the value of sharing the information broadly with the community is very low and the distraction it would cause would be significant (compensation, for instance).

What I'm saying is that I believe one should push the limits of being transparent while remaining prudent. Because we tape virtually everything--including our mistakes and weaknesses--for everyone to see, we are a target-rich environment for media that thrives on sensationalistic or critical gossip and can find ways of having information leaked to them. In one case when we faced the problem of having information leaked to the press that was intentionally distorted and hurt our recruiting efforts, we were forced to institute some controls on ultrasensitive information, so that only a significant number of ultratrustworthy people received it in real time, and it was distributed to others after a delay. The information was the sort that, in a typical company, would be shared with just a handful but at Bridgewater was shared with nearly a hundred trusted people. In other words, while our radical transparency in that case wasn't total, I pushed its limits in a practical way. It served us well because the people who most needed the transparency got it right away and most everyone understood that the commitment to being transparent remained very much intact, even in challenging circumstances. People know that my intent is to always push the limits of trying to be transparent and that the only things that would prevent me from doing that will be the interests of the company and that I will tell them if I can't be transparent and why. It is in our culture to be that way and that fosters trust, even when the transparency is less than we would like it to be.