Recently, Ray Dalio was interviewed by several medias after he had published his new book named as “A template for understanding big debt crises” (hereafter “Crises”, free PDF available here). Dalio became more public after he had published his new book “Principles” in 2016. This year, 2018, the newly published book “Crises” makes the situation subtle. I have downloaded the new book and prepared to read it. However, now, I hope to summarize the key points that Mr. Dalio raised in the interviewed video.
Everything happens over and over again.
Template: Inflationary depression
Key point: The debt is nominated in the nation’s own currency or foreign currency.
If the debt is nominated by foreign currency, e.g. dollar. When the dollar goes up, the country does not have enough cash to pay the debt, so print more money to pay the debt, and the inflation and devaluation come.
However, if the debt is nominated in the country’s own currency, the depreciation will wipe out the debt, naturally.
Then, when the currency becomes cheaper, the country will export more and import less. If the country remains an healthy environment to invest, foreign capitals will come back.
The central bank, by monetary policy, should make the interest rate compensate the depreciation of the currency.(Interest = Infaltion + Devaluation) This is the way to track the bottom of the devaluation of the country’s currency.
(Turkey, Argentina is around 1/3 distance to the bottom)
After 2008:
About China, tariff.
China can handle this. Big deal: relationship with China (antagonism is bad!).
China’s debt is denominated in CNY, and dealing with it before the crisis, and from a top-down method, Dalio thinks China can deal with it well enough.
Updated 2018-09-17
最近几天集中编码了一番,终于把差不多九个月前立下的flag雏形写好了,Mercurius的Demo version完成,从spider拉取数据,入库,调库回测,标准化输出json数据包,上传github,github page前端调用显示,虽然策略只有简单的up in down out,但框架还是完整啦。后面打算做个Risk Parity的portfolio回测下放上来。
项目并不麻烦,也还没上酷炫的ML算法,但串了一下从python spider/mysql到frontend这一条的工程实现,也算是不错的锻炼吧,希望未来能给我带来一些机遇。God Bless~
上图:
未来完整版本的工程实现流程图:
Sandbox页面策略的实时回测,偷懒还没有把event summary写好:
以后每天又多了个客观参考啦~距离程序化决策又近了一步。
Updated 2018-09-13
今天刷华见发现Dalio又出了一本新书,详细阐述历史上的债务危机。毫无疑问,桥水出品,Dalio出品必属精品,兴冲冲去桥水网站subcribe了免费的pdf版本。往下拖发现Dalio居然在2017年还有个TED演讲。
内容基本是对《Principle》一书精华的阐述,算法决策,计算机决策,集体智慧,克服个人的非理性与集体盲从的办法就是算法的weight。不过视频演讲中能够给出一个足够震撼人心的例子:Dalio中途戴上眼镜,盯着辅助演讲的iPad,展示了一下Bridgewater的会议流程:
所有人都可以对其他人提出赞成或反对意见并打分,Dot Collector(桥水的计算机决策系统的一个module)会进行记录,因此,可以想象N人的会议就能够组成一个NxN的矩阵(有向)
随后计算机系统会根据每个人在所有会议的表现等等进行用户画像(有没有感到可怕!),当最后会议集体决策时,并不是单单的democracy!而是根据不同人不同的merits进行加权;同时,用户画像也让企业能够对不同特质的人进行组合匹配,从而有利于合作事务的最优展开。
最后一个关于民主决策的例子也非常震撼人心,即使在一场会议上65%赞同35%反对一项决策,当桥水的机器决策系统根据不同人的merits进行加权后,会发现这个比例竟然会反过来,而更可怕的是,整个公司会遵循这样加权后的决策树行动。正如Dalio在经济运行的视频中描述的,像机器一样运行。
绝对的透明和绝对的客观,这是通往成功的第一步。
另附Dalio的新动画视频 《Principles for Success》
Updated 2018-09-11