We recently released this entire In The Arena podcast in video format for subscribers exclusively on HedgeyeTV. You can watch it below.
This is an absolute must-listen for risk managers.
In this edition of our investing podcast In the Arena with our Senior Macro analyst Darius Dale, Darius explains how our regime-based risk management framework can help investors proactively prepare their portfolios for “bouts of volatility,” recession and more.
Here’s a key excerpt.
“If we can use our forecasting tools to sidestep bouts of volatility in particular asset classes, we can grow the net asset value of our portfolios in a much more risk reduced manner…
If you want to dig down deep into the juicy details of our Macro process with one of its creators, listen to this interview.
Click here to download the iTunes version of In the Arena.
Below we’ve transcribed for you a significant portion of the podcast and included relevant charts.
* * * *
Daryl Jones: Let’s get right into the Hedgeye process. We use a lot of proprietary language and models and I think for some of our users who are relatively new, this would be a helpful overview. In a sense, we’ve created our own language. I think maybe going through process step-by-step, maybe an explanation of how we created the model, what it does, how we use it. Obviously the big one is one that you're sort of the author of which is the GIP model. So, you know, maybe we start there and you just sort of talk a little bit to high level, what is the GIP model? How do we manage and operate it?
Darius Dale: Absolutely. So Daryl's referencing our Growth, Inflation, Policy model or GIP model for short. This is a regime-based sort of framework. Think about contextualizing the economy with respect to the rate of change and cycles of growth and inflation.
Before we get to that, first take a step back and say, ‘What are the primary drivers of asset class returns and rotations within and out of asset classes, across multiple economic cycles?’ So that sets the first sort of question you should be trying to answer as a macro-oriented investor.
Ray Dalio has a lot of PhDs at Bridgewater who have all effectively come to figure out the same thing that we figured out. Growth and inflation are the primary, principal components of asset class returns. When you start to layer on changes in policy and changes and things like corporate profits, you really start to form a very holistic picture of what actually is driving interest rates, currencies, equity indices, and these are the sort of major asset classes we focus on.
Growth and inflation are the principle components. Those are the most important things to get right as it relates to forecasting asset market returns. But secondarily, this is really important point, policymakers respond to changes in the levels of growth and changes in the levels of inflation on a lag. That’s the rate of change.
What I mean by rate of change, we are asking if growth is accelerating or decelerating and inflation is accelerating or decelerating. We're not just keenly focused on the levels, obviously the levels matters as it relates to forecasting policy. But focusing on those big inflection points in the rate of change because that's where the price action is occurring in financial markets that leads central bankers to make these sort of policy pivots or double down or an existing policy initiatives.
So we're very keenly focused on triangulating all three of those things – growth, inflation and policy – as it relates to where the economy could potentially be.
Jones: With the model, I was just going to ask you to go through the different regimes.
Dale: So, that’s on slide 6 of our Macro Themes deck. The first 25 slides or so in these quarterly themes don't change at all. In fact, they're just process slides that we continuously update. But on slide 6 we show the GIP model if you take the two factors, growth and inflation, you wind up with effectively four scenarios when you sort of slice them into our regime-oriented framework.
The first of those scenarios is in the top left box in that chart is Quad 1. Quad 1 is a situation where real GDP growth year over year in an economy is accelerating and a period whereby inflation is decelerating. Historically investors have called that goldilocks in terms of the general nomenclature around Wall Street associated with sort of these types of economic conditions.
Jones: When you talk about accelerated or decelerated, you're talking about the, the rate of change.
Dale: Absolutely. So the, the percentage change on a year over year basis is that accelerating (i.e. getting bigger or smaller) on a period to period sequential basis? So for example, if growth is 1% in the first quarter and it's 2% in the second quarter, it accelerated by a hundred basis points over the prior quarter. So we're keenly focused on predicting and projecting not only the direction but also the magnitude of those accelerations and decelerations. Because again, that's where all the financial market price action occurs in the pricing in of these changes.
Going back to the four different regimes, Quad 2 is the top right box in that chart. That’s when growth and inflation are accelerating at the same time. That tends to perpetuate a much more hawkish monetary policy. Dynamic investors have tastes historically in terms of conditions like these reflation or overheating regimes to the extent that you're in a persistent Quad 2 regime whereby growth and inflation are celebrating on a trending basis.
Jones: Quad 2 would be a scenario where the Fed often is forced to act or thinks they need to act?
Dale: Bingo. Particularly when inflation is bumping up against their policy mandates, which is generally around 2% on core for most of these advanced economies. So that's the hawkish quadrant. That's the quadrant where you typically see bond yields rallying. You see real interest rates rising. You tend to see sort of underperformance of things like more safe haven assets like Gold and Treasuries in lieu of much more credit-oriented and equity oriented asset classes. I mentioned Quad 1 which has historically been the best quadrant for equity and credit returns. Quad 2 is obviously a close second.
Jones: And then the other two quadrants are the reverse of those two.
Dale: Bingo. So if you sort of draw a Maginot line down the middle of the GIP model, what you find is that Quad 1 and Quad 2 are pro growth. Accelerating in both of these quadrants. Inflation is the oscillating factor. Whereas when you go to the bottom of the quadrant Quad 3 is when growth is decelerating and inflation is still rising.
Quad 3 is much more associated with late cycle economic expansions where you have sort of tight labor markets. The natural growth rate of the economy tends to slow down, as both the credit and labor cycles mature. But inflation is obviously continuing to accelerate. The longer Quad 3-like conditions persist investors start to use words like stagflation or stagnation to describe that kind of market or any sort of economic outcome.
Historically what you've seen is that, particularly later in the economic expansions, Quad 3 really starts to sort of become stock picker and credit picker’s market. You don't tend to see big outsize moves in either indices or rates or currencies. You start to see much more middling performance at the index level and at the asset class level. But sort of underneath the asset classes you have some pretty big divergences as it relates to various industries you can invest in with respect to the equity and credit markets and on various currencies within the currency markets. So you start to see that kind of stuff sort of play out.
Credit risk and leverage tend to become really negatively correlated style factors with respect to the persistency of Quad 3 regimes.
And then lastly, this is the one that everyone hates unless you're long bonds, gold and the U.S. dollar and that’s Quad 4.
Jones: And by long bonds, you mean typically Treasury bonds, right?
Dale: Yeah exactly. Or for any developed market currency or developed market country, you tend to see the outperformance of sovereign debt in Quad 4 as investors tend to rush for safe haven assets or the perceived safety of these sort of so-called safe haven assets. That’s a totally different discussion.
But Quad 4 is a scenario where growth and inflation are decelerating at the same time. As you can imagine, that's really negative. That's a really negative economic impulse as it relates to the general tone of financial market news, growth slowing and inflation slowing.
What you've seen historically is that the preponderance of real big draw downs as it relates to equities and credit have historically come in and around Quad 4 either the market pricing in Quad 4 ahead of time or the macro market pricing it in while it's ongoing.
Again, these are quarterly regime oriented projections that we're trying nail every single quarter to identify exactly what might change on the margin with respect to training asset market performance.
But taking a step back to the broader business cycle, what you see is heading into a recession, at least over the last few recessions here in the US, is that cycles tend to peak out in Quad 3. Again, you get tight labor markets, you know you're starting to roll down the hill in terms of employment growth and wage inflation, which is a very late cycle indicator in terms of when it peaks relative to the cycle. Then you sort of slow into Quad 4 and that persistency of Quad 4. Quad 4, after Quad 4, that’s how recessions occur with respect to rate of change.
Speaking with respect to these regimes as it relates to our process, we don't necessarily care about calling forward an NBER business cycle-type recession because again, if we're set up for the asset allocations that history would suggest you should be in for Quad 4 and Quad 3 as well, then you don't necessarily have to make the recession call because you're already in the right types of assets that do well.
Jones: So historically our models actually work quite well. But in the instances where we've been off either in the wrong quad or the assets didn't move the way we expected them to in the short term, what went wrong with our model? Or did anything go wrong with our model? In hindsight, what was the error, if there was an error?
Dale: So there are two types of errors that could happen. One, there's forecasting error. We get the quad wrong.
I'll take a step back and actually explain how we use this process because it's not just enough to have a regime oriented framework but as it relates to investing because the market is constantly pricing in, future economic developments and future macroeconomic developments.
You have to have an ability to forecast where we're going in this as it relates to these regimes. So we intra-quarter data we’re receiving for that particular quarter on the high frequency data side. And what I mean by high frequency data are things like industrial production, retail sales, the PMI data, confidence surveys, things of that nature that feed into those models that give us a real-time sense of where the economy is tracking from a year over year rate of change perspective with respect to real GDP. And then we run a very similar Nowcasting process for inflation, but we use market prices, things like WTI is a big swing factor and other inflation series, like food prices. Things of that nature. So we constantly have a general sense of what quad the economy is in intra quarter.
And then from a forward-looking perspective that looks up to one year in the future, we use our comparative base effect model. That uses the two-year comp stacks to sort of extrapolate the level of momentum of either growth or inflation into the future to give us a general sense of a very highly back-tested, a highly accurate sense of how the rates of change of growth and inflation are likely to sort of progress over that time horizon.
With respect to the forecast error associated with that our first initial estimate for any particular regime using a Bayesian inference process is always going to be with the rates of change and the comparative base of that model.
Again, if you look at it one quarter ahead to Q4 of 2019, we don't know anything about Q4 for this year. All we know is what the current level of growth is, the base rate for growth and inflation and we know the comparative base model changes from Q3 of 2018 to Q4 of 2018. This comparative base effects model does carry some forecast error for growth and inflation, but it's about 70, 75 to 80% accurate in terms of forecasting the directionality correctly, which is really all we're trying to do.
We obviously want to get the magnitude right, particularly if we go deeply into Quad 4 or deeply into Quad 1 because then you tend to see those big outsized market moves when the market's pricing in these huge, massive deltas in the economy in short order.
The other potential forecast error is with respect to how the market is going to treat this particular progression of quads. We don't consider ourselves to be competing with other investors or competing with other macro strategists. We're constantly competing with the market in both in two ways. One, determining what the market is actually pricing in. And at any given interval is the market pricing in the quad we think it is. If not, what quad is the market pricing in? Is it looking ahead or are we just wrong on the quad projection?
So there's a number of different ways you can be wrong, which is why you know Keith and I and all the guys on the team wake up so early every morning to crunch all this data to give ourselves a better sense of how to mitigate that sort of forecast error.