Will quantitative easing be phased out? Will it continue? What effect will this have on the market either way?
These questions, I believe, are ultimately meaningless; not because the Fed doesn’t have any influence over the markets—we all know they do—but because the Fed is now moving towards “data-dependent” real-time market management.
Let me explain.
The Fed, along with many other corporations around the world—yes, the Fed is a semi-private corporation—is taking part in a vast technologically-aided transformation with big data and software intelligence taking the lead. Here, Barry Ritholtz says how this process has dramatically changed the corporate landscape by explaining:
In the old days a company would produce widgets all around the world and they would produce a hundred different versions of it and it was a guessing game as to how much steel they had to buy, how many people they had to have working at different plants; and when you would say to them, “What is your best-selling widget in Europe and how does that differ from the best-selling widget in South America?”, they would say, “Well, we could tell you what our best-selling widget was two years ago but we really don’t have the information as to, you know, last quarter.” Today, they can tell you the best-selling widget, second by second, in every region of the country, every region of the world, practically in real-time.
In the case of monetary policy, of course, it’s not how many physical widgets to produce when and where—it’s the number of currency digits flowing through large banks in a financial network.
Similarly, before the dawn of computers, electronic networks, and powerful software, the slow and ardurous process of collecting, modeling, interpreting, and implementing policy meant a huge time delay divided the original data “signal” from its response. This, as you may know, is one of the core criticisms for why central planning—and Fed policy in general—is deemed counter-productive: it’s just too slow and inefficient; only the market can react fast enough to properly regulate itself, the argument goes.
So, what about now? Is this still true? Regardless of where you lie on the free-market versus centrally-planned economy spectrum, the world is moving in a direction that makes central planning much easier, faster, and far more efficient.
To understand why, let’s explore the process I gave above. Previously, a large number of people were required to collect, model, interpret, and then implement policy decisions. Now, this loop is largely automated as algorithms prove capable of reacting to real-time changes in the markets and economy with radar-like precision (see previous post, A High Frequency Map of the Market). You may be thinking, “But what about the interpretation and implementation part? This requires human judgment, which alone cannot manage something as complex as the market.” This may be true...and the exact reason why we're developing "expert systems" capable of replacing human decision-making on a vast scale.
Lest you think I am stepping out too far on a limb, let’s first consider how many people it takes to implement—the last step in the process—monetary policy at the Fed by injecting billions of dollars into the financial system. As Neil Irwin, author of "The Alchemists: Three Central Bankers and a World on Fire," described in a recent interview:
I actually have a scene in the book when the quantitative easing program, QE2, was announced by the Federal Reserve. I actually saw the room at the New York Fed where they carry that out. It’s kind of three mid-level guys sitting at computers—there’s a supervisor behind them and a tech person watching everything to make sure the computers work; and when they inject $3 billion in one morning into the financial system, they take bids and carry that out and with that clickity-clack on the keyboards it’s money being created out of thin air. I think it’s uncomfortable for people to understand how much of an idea money is but that’s the reality we’re all living in in the modern economy.
When money is no longer physical, you no longer need physical people to create or manage it. Likewise, I doubt there will be anyone behind the curtain in five or ten years either (except for probably the IT guy). What about the various governors sitting on the Federal Reserve board, arguing whether they should be doing this or that, or how they’ll respond, or you name it? What if I told you it's all for show?
What ultimately determines adjustments in Fed policy is a set of computational models—with the Fed chairman communicating those changes in mathematically-crafted nuance and subtlety. Does a computer generate the Fed's communiques? Probably.
When the Fed says their decisions are data-dependent, this means any deviations in their economic models will determine the pace at which billions are funneled by their massive workforce three guys monitoring a computer.
So, when will they stop QE? Not until their economic models tell them to. What if things turn south and the global economy takes a nose dive? Then, the models will tell them to add. Then again, the models need not tell them anything. Humans only slow the process and interject unnecessary delays, right? In that case, we've simply replaced the job of managing the financial system with a machine. Sound familiar?
“We don’t need a Fed,” Milton Friedman says, twirling a letter opener as he speaks. “I have, for many years, been in favor of replacing the Fed with a computer”
Did good ol' Milton get his wish? Perhaps. Thing is, when you're trying to manage the markets in real-time, such wishes take a life of their own.
