Without precise definitions you end up in forecast hell

Topic: Imprecise definitions lead to useless models and conclusions

Discussion: If you’re performing macroeconomic research, which inflation are you talking about, which growth, which interest rate? The answers to those questions can be of crucial importance for your eventual investment outcome.

Length: Short — maybe 5 minutes reading time

Teaser: It’s easy to predict the weather. Not to mention stock market returns

PODCAST TIP: listen to my latest podcast episode (#6) on Future Skills with philosopher Alexander Bard. We talk about definitions of infantile grown-ups and much much more. Check it out on iTunes here, or your favorite Android app here.


Dream Warriors

Are you dreaming about making perfect economic forecasts, and using them for producing amazing equity investment returns? How does this sound to you:

Weather and production bottlenecks in combination with monetary policy induced growth are starting to cause higher commodity prices. Inflation is already showing in their wake. People worry about rising interest rates, just take a look at OIS spreads. Some central banks are turning less dovish. Higher interest rates means less funds for investments, lower growth, lower profits and lower share prices. Higher interest rates mean lower bond prices, higher borrowing costs, lower real estate prices among other things. Higher inflation means money loses its value. And this time it’s at a time you can’t hide in stocks or bonds. You could hide in gold. One bar of gold is always one bar of gold. Maybe that’s why the gold price in dollars is rising (despite obvious manipulations and jaw-boning from various authorities).

Does the above fit your view? Higher commodity prices => higher interest rates => sell stocks, bonds and real estate and use cash to buy gold and soft commodities, until the cycle turns again?

Well, hold your horses for just a little while.


What’s your definition of a boombastic jazz style?

Which commodity prices are you talking about exactly, when you say their prices are rising? Wheat, hogs, orange juice? Iron ore, coffee, cacao? Silver, cobalt? Platinum, palladium?

Similarly, which interest rates are you referring to? The Fed funds rate? Treasury bills, longer term bond market rates? Corporate bond rates, bank lending rates (to consumers, to corporate clients, to house builders?), peer to peer lendning rates? Intrabank market swap rates?

Oh, I almost forgot, “inflation” you said. Would that be the (ever manipulated and ever changing) CPI measure? Or the PPI gauge? Input our output PPI? How about house price inflation numbers? Or energy price inflation? Avocado prices?

My point in this post is that if you don’t clearly define exactly what variable you are talking about it becomes exteremly difficult to make any kind of coherent analysis, not to mention draw any practical conclusions whatsoever from the exercise. Macroeconomic research is difficult enough as it is without averaging everything together, whether it be “the inflation”, “the interest rate”, “the oil price”, “the stock market” or “GDP”.

Take that last one, e.g., GDP. What does Gross Domestic Product really tell you? What conclusions can you draw from it even if you knew its exact trajectory going forward a few quarters? How about nominal GDP vs. real GDP (using which deflator measure?), or GDP per capita? Then there are data series for wages, wage growth, hours worked, hourly wages, lost jobs, added jobs, seasonal adjustments (many orders of size larger than the actual net number), employment (measured in at  least three different ways depending on, e.g., how to define somebody without a job, based on whether he’s searching for a job or doesn’t care).

And what’s so special about GDP growth by the way? There’s zero useful correlation between real GDP growth and stock market returns. How about a house price recession like the one that began in 2006, several years before the ‘actual’ recession. Don’t even let me begin to talk about the NBER’s definition of a recession (no it’s not “two quarters of contracting real GDP”

 

“a significant decline in economic activity

spread across the economy, lasting more than a few months,

normally visible in real GDP, real income, employment, industrial production,

and wholesale-retail sales.”

 

No matter the problems of making macroeconomic models work at all, you don’t want to make it even harder by using impractical and vague definitions. My message in this post is that you need to make sure your definitions are practical and at least theoretically can lead to better decisions.

After that we can talk about the ephemeral character of macroeconomic causations and correlations, not to mention their flimsy associations with actual stock market behavior.

For now take this with you: Knowing what you know and knowing what you don’t know, is paramount in uncertain environments. And the financial markets are as uncertain and stochastic as they come.

Thus, make sure you do define all concepts and ideas about their connections precisely. That’s your only chance of keeping track of what you know and what you don’t. In addition, it’s your only fair chance of creating a feedback loop of increasing knowledge and strategy adaption.

Such directed or deliberate practice is in similar fashion your only chance of coming out on top in the arguably most competitive sport known to man (and yet untrained newbies gladly step into the ring and bet their life saving’s on themselves).

Today’s advice holds true for everyday life as well. I don’t know how many arguments with friends I could have saved, had we only defined the concepts and words precisely at the outset…


Stock market forecasts coming up

I’ll soon write a follow up on this article, where I’ll explain how stock market returns can be reliably forecast in much the same way as the weather can be accurately forecast. For now this teaser will suffice:

Just as I know there’ll be snow in the middle of Sweden on quite a few days every year between December and February, and almost completely certainly no snow 99 per cent of the time between June and August; a highly priced market will produce very low rates of return, and a lowly priced will produce high rates of return over the coming decade or so. But more detailed notes about next time and the exact implications for the current situation. Stay tuned.


FUTURE SKILLS: Don’t forget to check out the super energetic conversation with Alexander Bard on Future Skills Ep. #6! You’ll find it on iTunes here, or your favorite Android app here.

 

admin

4 Comments

  1. I wasn’t able to comment on or rate the Future Skills #6 podcast for unknown reasons. Maybe it wasn’t compatible with my phone’s web browser. Besides that, the podcast itself was very good and I’ve recommended it to my friends. While some of Alexander’s points seem silly to me (e.g. the Freudian gential obsession), the way you all spoke about the infantilization of modern society made a lot of sense and gave me a new perspective. I never really noticed the trend of people shirking personal responsibility in favor of victimhood until I started reading and listening to you. I have to admit that I’ve acted as a part of that crowd before, but your arguments have pulled me away from a victim mindset and towards your “green pill”. Between reading your website and Ludvig’s, I’ve realized over the past few weeks that I’ve wasted 3 and a half years so far in university without really challenging myself. It’s disheartening. I won’t have the right degree for a high paying career in “whore village” as you put it, so it looks like starting a corporation is my only option. Thank you for showing me a more fulfilling path than victimhood.

    • Great insights though, and I’m guessing you’re fairly young = early to mid 20s, so you have plenty of time do grow.

    • The trend away from personal responsibility toward victimhood has been accelerating for several generations at least. I’m old enough to have seen some of it in my lifetime – forty years ago it was virtually unheard of for healthy young people to get a disability pension for “agoraphobia” or any other vague complaint of work aversion; now it is not only common but widely accepted as a right. Fifty years ago feminists fought for the right to enter certain professions then still reserved for men by institutional culture; thirty years ago they fought for the right to sue over anything that made a woman “uncomfortable”.

      But the trend is much older than that. I think it has three roots – Marxism and its ideological descendants, which encourage people to think of themselves (perhaps not without justification in 1848) as members of a victim /class/; democracy, which encourages unproductive people to combine against productive people in a way that would have gotten them massacred or enslaved a couple centuries ago or even later; and mass media, because they thrive on underdog tales and sowing conflict.

  2. You can’t build a model with infinity variables. What works is pattern recognition and separating signal from random noise. You can’t do this without distilling out a finite number of assessments. Maybe if you are good enough these assessments can be subjective, but I suspect otherwise. Measurements don’t have to be perfect to be useful.

Leave a Reply

Your email address will not be published.