Topic: How to (if at all) use macroeconomic data for trading the financial markets. Length: 3400 words.
What do we want?
How do we get returns?
-Know when stocks are going up, and buy beforehand! Then sell in time.
How can we know?
-And Technical Analysis!
No! It doesn’t work that way
Topic: How to (if at all) use macroeconomic data for trading the financial markets
Summary: It’s mostly worthless and a dead end, but it can be used in combination with other gauges of stock market sentiment.
Start with establishing your own rudimentary macro view and refine it. Iterate between micro, macro and trend and risk gauges to construct a consistent view of respectively the economic cycle and the stock market trend.
Length: 3400 words (“a few pages”)
In theory, practice and theory are the same thing…
In theory it’s pretty simple. At constant valuation multiples; fundamental gauges of economic production, such as sales, profits and cash flows, should drive stocks.
Further, from a top down perspective GDP growth drives sales (and thus also drive profits and cash flows, after some adjustments for private and public sector changes in debts and deficits).
Consequently, GDP should drive stock prices
-and it does… over centuries.
Predicting GDP with PMI new orders surveys
The typical sequence of events for predicting GDP growth starts with PMI surveys for new orders and production, continues with real sales, real production and real income in that order. Finally, more or less lagging, come numbers for new unemployment claims and changes in payrolls*.
* see, e.g., Hussman weekly February 2, 2015 and March 23, 2015, 9 November 2015
So, it’s a piece of cake then… Take stock of the aggregate of corporate plans for new orders and manufacturing. Predict sector and national GDP growth from that, and rotate among sectors and size your longs accordingly.
Perhaps you could even refine the strategy by complementing the official PMI series with proprietary micro data (on investments, hiring plans and management body language) from company visits and interviews.
…in practice, they are not
It turns out that GDP growth, valuations and stock market returns are mostly uncorrelated even measured over decades, and for some periods even negatively correlated. Just take a look at China’s GDP and stock markets the last quarter century, if you want one salient example.
The problem is that, even if you manage to forecast GDP growth, you would still have to forecast it’s short term correlation with profits and other key ratios, not to mention the valuation multiples the market is ready to use at any given time (based on interest rates, risk tolerance, debts etc.)
Key ratio (e.g., Earnings per share)
valuation multiple of that key ratio (e.g., P/E-multiple)
Thus, unfortunately no matter how good your data series and your statistical acumen, you won’t improve your stock market returns by analyzing GDP growth. By the time the changes in GDP matter for stock market returns, years and years after your diligent measurements and forecasts, the growth numbers are since long public goods and your edge is gone.
Actually, to ensure capturing the underlying GDP growth in your stock portfolio, you have to be extremely long term in your investment strategy. For life, in practice.
Either you invest really long-term, i.e., just Buy And Hold until retirement, or you futilely try to predict shorter term fluctuations. In any case you’ll struggle to add any value whatsoever.
To summarize, going with the Buy And Hold strategy more or less means you’ve given up altogether. And, getting into the short term forecasting business demands that you:
- Predict the outcome of certain events like central bank meetings, national and international statistics releases, elections, natural disasters, unusual weather etc.
- The future is inherently unknowable making this impossible
- Predict how the outcome will affect the real world [profits]
- Relationships are not stable over time, partly due to the reflexivity of the complex and nested psycho dynamic feedback loops that make up economic systems
- Predict how other investors will perceive, interpret and value the effects
- Varies in a wide range over time
In other words, there is no fundamental and logical way in which to use macro variables to predict short term stock market movements. And if you’re long term you’re long term and thus simply Buy And Hold forever – and there’s consequently nothing of value here for you as a trader.
unpredictable macro events => uncertain GDP growth => uncertain uncertain short term transmission to sales => uncertain transmission to profits => uncertain valuation multiples => value
Micro analysis is for single stocks and sectors. If you know and understand enough, you might be able to pick company and industry winners over various periods of time.* That’s a different story though, for another post, and won’t help you associate GDP forecasts with overall stock market returns.
* Beware though; Apple almost went bankrupt despite its popularity before becoming the most loved stock of all time. And Facebook almost didn’t happen at all, and maybe should have lost out to the industry leader MySpace. Similarly, Google was the latecomer to the search business. Could you have known that their algorithms would beat those of Altavista and a dozen other local and global search engines? And Amazon still hardly makes any profits. What about Microsoft, Exxon, Enron, and the car manufacturers GM/Ford/Chrysler that all almost went bankrupt? Tesla, SolarCity, SpaceX? Do you know enough to make predictions? Does even Elon Musk?
That leaves TA, a.k.a. Tea-leaf Analysis
Interpreted freely, TA entails everything that isn’t “fundamental”, such as historical movements in various market variables like stock prices, currencies, interest rates etc.
Various patterns, spreads, spread changes, combinatory changes, trend convergence or divergence etc. are assumed to predict future movements in the price series itself or in other prices, since they have exhibited similar patterns several times before.
One of my teachers at business school used to say it’s about as useful to use TA for investments as looking at patterns in tea leaves would be. On the other hand I think he believed in the Efficient Market Hypothesis fairy tale too…
Most people using TA typically draw nice lines, waves, triangles, Head And Shoulders and other more exotic patterns in charts, mainly based on seeing those patterns in a book.
They then predict share price movements based on how certain patterns have evolved in the past.
Most just take the formations for granted and never back test individual indices or stocks to see how often certain patterns have played out to the upside or downside, or how that correlation and likelihood have changed over time.
What technical analysts often fail to acknowledge is that, these days, millions of algorithms are constantly evaluating old and new patterns, and are evolving using machine learning in a speed that no human can match. This algo activity exploits and neutralizes old patterns’ predictive power and create new ones.
That means the old paradigm of “human psychology is essentially the same which means old technical patterns are still valid” is over.
Machines rule today, and they can calculate the probability of any pattern recurring, no matter how complicated or visually unappealing. A human with a pen, a ruler and a Fibonacci or Elliot wave theory software has no business doing TA in this environment, since he probably couldn’t even detect today’s important patterns even if they were described to him; let alone calculate the probability of success. He could just as well check the bottom of his tea cup after breakfast.
Maybe human intuition still has a role to play in the shortest of intervals (seconds to minutes, not the microsecond universe of HFTs) and the tiniest of amounts at play (less than a million dollars), but when you start swinging for longer periods and with millions you are breakfast toast for the machines. In very particular situations with small illiquid stocks, and large positions shifting owners, maybe human psychology-based TA and old school patterns still work, but those situations could be very hard to find and know about beforehand.
Single stock fundamental analysis
My view is that macro research is useless
-just as useless as tea leaves and technical analysis
Single stock and industry research could, however, be worth the while – if you are diligent, systematic, disciplined, and most of all patient.
-Patient in looking for good companies.
-Patient in waiting for the right price.
-Patient in waiting for the market to acknowledge what you have found.
That, however, is a topic for another day. Today we’re talking macro.
Combining macro and TA could add value
So, finally, the light in the tunnel: I think a combined approach of a certain kind just might work, when it comes to predicting* major bull and bear markets, and possibly help detecting cyclical bulls and bears as well.
I find the following model intellectually appealing:
- Track valuations (cyclically and profit margin mean reversion adjusted)
- Increase your stakes at low valuations and decrease them at high valuations
- NB: this is not based on index movements but on valuation changes
- Buy cheap, sell expensive; not Buy high, sell low
- Track leading economic variables such as PMI new orders, and combine them with, e.g., data on debt growth to look for likely turning points for the economic cycle
- This might help a little in forecasting profits
- Don’t confuse the deluge of economic variables with the ones that actually are leading
- Use market based trend variables, like the 200-day moving average for the S&P 500 stock index, and market breadth to gauge possible turns in trend and overall sentiment
- A narrowing and rising market means more hands chasing fewer stocks; the only stocks that are “working”
- Measure risk appetite based on “rising tide lifts all boats” indicators.
- If various stocks, sectors, asset classes etc. diverge from the bull market trend of uniform advances, it’s an early sign of falling risk tolerance. Market breadth is but one of dozens of established indicators
- The big idea here is that no single indicator is that important, it’s the syndrome of how “all” indicators change their behavior relative each other that can be the harbinger of markets pivoting from bull to bear – IF valuations are high, and the economic cycle appears stretched
In other words, let valuations be your main guiding light for long term investments; and a very specific kind of trend research (uniformity of market internals rather than typical TA pattern analysis) your cue for likely turns in market sentiment.
And a little macro
As an added extra, but more or less unneccessary, keeping an eye on macro indicators could tell you if the cycle is long in the tooth and if certain variables* are near exhaustion (as measured by historical highs for example).
* private debt, government debt, covenant light loans, IPOs, share of unprofitable companies going public, inflation, commodity prices, interest rates, share of algorithmic trading…
Economic history only rhymes at best
Even if you have crafted the perfect model, using the best statistical analysis possible, and even if you have found a system with strong signals and relatively few errors, no two cycles are the same.
Thus, beware of distortions from politicians, new generations of leverage-happy consumers, demographic changes, new technological paradigms and wars, since they can all shorten or prolong whatever cycle you are assuming is underway.
In short, you know nuthin’ John Snow.
I still love macro research though.
I keep trying to use it for my investment decisions; just not in any direct way, and definitely not short-term such as for single quarters or years.
I devour newsletters and updates from the likes of Hugh Hendry, Jeff Gundlach, John Hussman, Jeremy Grantham, Marc Faber and others, and I listen to the MacroVoices podcast with Erik Townsend as soon as I get up on Friday mornings (as well as the Real Vision TV encores a few days later).
What I aim to get out of it is a general sense of recent history as well as the current state and direction for various economic superpowers over horizons of 5-25 years.
I want to “know” if the dollar’s strength is sustainable over the coming decade or just a dead cat bounce. I want to know how the deflation/inflation long game might play out. I want to gain new insights into the likely long-term movements of interest rates, profit margins, government deficits, consumer loans etc.
Not least I want to know where we are and where others think we are in the economic cycle.
I’m not using macro research to micromanage my investments, but to decide whether my overall asset allocation seems reasonable for the coming decade(s). I’m not looking for rockets taking off, but for not being blindsided, for not making fatal mistakes.
In a way, I’m also on the lookout for turning points, not in order to profit from them per se, not for calling tops and bottoms, but for understanding if the current trend is likely to continue for a long time, or if we are more or less close to a major trend change.
Right now I do think we are close to several important inflection points, but I am in no way certain about the general economy as a whole. Here are my thoughts regarding a few key variables:
Profit margins going down: Close to historical highs, due to government and household deficits feeding into higher corporate profits, since consumers can afford to shop more at lower wages. Software models and increased automation can stretch the system further for individual companies, but will likely prove difficult at the national or global level for the foreseeable future.
Productivity staying low: has been very slow the last decade and if anything seems to be slowing, despite technological progress. I am quite surprised by this development, and 50 years out, e.g., I’m expecting immense productivity gains. Regarding the coming decade, however, it seems much of the low hanging fruit in terms of productivity has been picked already (urbanization, globalization, know-how sharing, computerization and not least leverage and capital per employee).
GDP growth (risk of recession): low productivity growth, low population growth, low employment participation rate, aging population, peak stimulus, debt burdens all speak for weak GDP growth the coming decade. That also means substantial risk for more frequent recessions.
Monetary stimulus have possibly peaked: Extreme for a decade, but can always get more extreme, so impossible to call a turning point
Inflation should take off after a bout of more disinflation: Given all the stimulus, all the money printing, asset price speculation and lower investments in productivity causing a stagnating global economy, limited room to leverage housing, credit cards and students, I see a definite risk of inflation turning upward for the long haul. However, I think we’re first due for a recession and (debt) deflationary bust
Interest rates going up: have been falling for 35 years; I think they are very close to turning upward for the long term, due to inflation and normalization of policy rates. Given the likely weak economy, interest rates are unlikely to go very high though.
USD going up: I understand the strong USD short to mid term (flight to safety, policy rate increases, the dismal state of the European Union etc.), but given the state of the US economy and its future obligations, there are so many other assets I’d rather own than dollars, like, e.g., precious metals.
Summary – macro is useless
-unless used as a complementary tool
It’s very difficult if not impossible altogether to trade stocks using macroeconomic analysis.
Not only is the future unknowable. Macro outcomes’ correlation with profits and valuations is tentative at best. If you can’t forecast the events, and can neither forecast profit outcomes of such events nor the likely valuations of those profits, there is no way of connecting macro guesses with future stock prices in any relevant way, except regarding very long term asset allocation decisions.
Macro and TA can only tell us what has happened and maybe where we are
And give an indication as to what often used to follow
-If ceteris paribus (which never holds)
In any case, make sure you at least separate leading indicators like new orders and real sales from coincident and lagging indicators like employment numbers.
Ordinary technical analysis is probably useless in the absolute majority of cases, partly due to the algorithmization of markets over the last decade. However, long term moving averages for major stock indexes might carry some value in turning points, if taken together with valuations and real world variables. They have performed that function historically.
Gauges of risk appetite (based on the degree of valuation convergence, interest rate spread development and trend uniformity of many different asset prices) could in combination with various other factors, including valuations, long term moving averages and the cyclical state of the economy, indicate if the market is closer to a top or a bottom.
Hence, focus your investment efforts on single stock research, using a “margin of safety” and long term mindset when it comes to business models, valuations and long term trends; since, unfortunately, macro is mostly worthless.
I would suggest you use macroanalysis as an overlay for identifying possible turning points, for hedging, for positioning your long term asset allocation, and for fine tuning your single stock research (with macro factors driving competition, input prices, debt, growth etc. for your single stock), but not your main investment tool.
- Start with subscribing to or benchmarking a few select macro resources.
- I you want a non-stop shop for free research, register for Macro Voices free newsletter and check out the links of free research provided each week. Gradually bookmark and subscribe to the ones you like the best (and go against your own view of the economy if you have one).
- Establish and write down your own macro view in your commonplace book for investments, tag macro
- Check if your investment portfolio is resilient to the kind of macro environment you expect.
- What are the important drivers for your portfolio companies
- What do you expect for interest rates, currencies, GDP growth, inflation…
- Read the company reports and other news from your portfolio of investments to complement and fine-tune your macro view. Update your commonplace with notes for your micro and macro views. Make sure they are consistent.
- Identify a dozen of risk appetite variables that seem relevant to you. Update your commonplace #investment #risk table. Analyze their long term data series for trend uniformity and convergence/divergence and try to find patterns around market turning points.
- What happens when the P/E-valuations of growth and value companies converge/diverge? What happens when corporate debt spreads compress/expand? What happens when advance/decline numbers increase/decrease? What happens when bank, realty, industrials, tech etc. industry sectors’ valuation multiples converge/diverge? What happens when the yield curve flattens/steepens? And on and on. What happens when x% of the indicators diverge or converge at the same time?
- If both the economic cycle is long in the tooth and trend and risk variables point to a change in direction, you might want to heed that warning and reduce your risk exposure. If on the contrary the economic expansion is fairly young and investors seem indiscriminate in their hunt for returns, causing convergence and trend uniformity, keep buying with both hands.
Too time consuming for you? Too boring? Doesn’t seem to provide enough of an edge? Seems kind of useless? Too long term for you? Well, then investing might not be for you.
Do you want more free and valuable analysis and inspiration?
Are you curious, philosophical, and looking for truth, understanding and happiness? Do you know more people like you and me? Please help me spread my insights about how to live a happy, healthy and wealthy life in modern society – tell others about this site, share it in your social networks.
Limited. Every now and then I clean my e-mailing list from inactive subscribers, to make room for new ones. Sooner or later I might introduce some kind of a latecomer fee to finance the growing list. However, it will always be free for incumbent, active subscribers.
Mistakes are the foundation of my knowledge, not least in my private investments since I retired from professional investing.
Share. I sincerely hope they can be of use to you and others; so please share this article and my site with your social networks.