Discover more from Capital Flows
The most abused part of macro
We are not slowing down at all! We just hit 1k followers on Twitter (@globalflows) and 400 subscribers on Substack. Thank you for reading! I promise to never waste your time.
We are continuing the research drop I laid out in my previous article. I just realized now that in that article, the bullet points for the economic data didn’t keep their structure when I published the article.
Here is the structure of broad economic data:
Net Exports of goods/services
Housing and Real Estate
International Trade & BoP
Financial Accounts/Balance Sheets
Household, Corporates, Financials, Sovereign.
Here is the deal, if you understand each of these data topics in a country AND how they connect to each other, you are going to be ahead of 90% of people in macro and financial markets.
However, economic data is probably the most abused part of financial markets and macro. I guarantee you that right now I could pick out several data points and build a narrative showing how inflation is about to take off to all-time highs or fall to -5%. It is just so easy because there are always so many different things happening in the economy. The tendency of people is to take a single data point that aligns with their bias and extrapolate it to the entire economy.
In reality, what you need to do is build a multidimensional picture of the economy using ALL data points. Furthermore, you need to correctly weigh the significance of these data points in your models.
So here is how we are going to approach this:
I am going to go through each dataset and provide a couple of broad thoughts about how it works. I will use US data as an example because it will likely be the most familiar for people. However, I am going to try to keep it broad because the US isn’t the only country in the world (I know shocker).
I will walk you through a couple of tools you can use for analyzing economic data
Notice how I separate everything by flows and capital structure in the bullet points. This is how you need to think about economic data.
A country is nothing more than a simplified version of a 10-K. There is an income statement and balance sheet. Growth and inflation reflect the income statement of an economy. Liquidity is what supports the balance sheet capacity of an economy. This is WHY we focus on GIP (growth, inflation, and policy).
Where did I learn about economic data? Primarily from. I will give full credit where it is due for these ideas. I have also read a lot of boring documents from the BEA and IMF going over economic data.
However, what I am going to lay out here is how you should familiarize yourself with these ideas. As a baseline, I always encourage everyone to read everything by!
The way that I think about GDP is that there are always different weighting for the inputs. You always want to know what the individual drivers are and how they are moving on a monthly basis.
A key thing with GDP is that it’s quarterly data, right? How do people come up with these GDP nowcasts? Basically, people take the monthly data and extrapolate it to the quarterly data.
If you look at the personal consumption data from the personal income and outlays dataset, it’s functionally a monthly representation of the personal consumption line item of GDP. https://fred.stlouisfed.org/release/tables?rid=53&eid=12998
Here is a chart of the monthly PCE and quarterly PCE line item of GDP:
Does it begin to make sense now? What we do is create a model that takes monthly signals and gives you a really good idea of quarterly lagged data releases.
Now you might ask, why in the world should I even care about quarterly data if the monthly stuff says the same thing? Because as you have lower frequency data, the quality usually increases. So the quality of GDP and Flow of Funds (quarterly releases) is higher quality than the monthly data releases.
Now did you see how I connected a monthly dataset with a quarterly dataset? You basically need to do this with every line item of GDP, National Income and the Integrated Macroeconomic accounts.
Here are the links for the US:
Once you begin to understand this big-picture idea, you can begin to structure the data in your mind and models.
The rest of these will be simple to understand if you know how they fit into these quarterly data releases:
Labor Market/Demographics: The labor market is key because every person's consumption is another person’s income. Once sources of income get taken offline (this is the nice way of saying you got fired), by definition, it has to impact consumption. The labor market is connected to personal income by the number of employees, hours worked and wages. This then feeds into consumption. Just go see how the line items flow on the personal income and outlay dataset.
Chart from Prometheus:
Service Sector: The service sector is the primary part of consumption in GDP. The service sector is functionally labor.
Retail/Wholesale Sector: Retail and wholesale sector are key for the goods sector (again another line item of GDP).
Industrial Sector: The industrial sector is really important in my opinion. When we talk about industrial production, it is producing some type of good. When you compare this production to the current demand, you can begin to build a picture of the inflationary/disinflationary pressures in the system.
Manufacturing/Trade Inventories: Depending on the specific dataset you are looking at, monitoring manufacturing and new orders for durable goods will be important in connecting business demand and current output.
Housing and Real Estate: Housing is one of the most sensitive components of GDP to interest rates. However, housing IS NOT the business cycle. I cannot tell you how many times I see people use building permits as a leading indicator of the economy. All you need to do for housing is aggregate all the datasets on housing and figure out the overall impact it is likely to have on GDP. The recency bias from 2008 plagues everyone these days and everyone thinks that the only thing that matters is housing. It is important but it’s not the entire economy.
Personal/Household Sector: As I mentioned above, get really familiar with personal income and outlays data. This is simply tracking the flows of the household sector.
Government: Similar to my note above on housing, you need to understand how government spending adds or drags on GDP. Government spending doesn’t always = inflation. You have to see the net effect on GDP. Is government spending increasing and offsetting other components of GDP or causing the economy to run hot?
Economic Activity: Overall economic activity is looking at things like business starts and overall activity measures.
Surveys/Cyclical Indicators: The main surveys you need to watch are the PMIs. The key qualification with surveys is that they are usually diffusion indices so the view they present is very one-dimensional.
International Trade & BoP: Generally speaking, every country will have monthly data on international trade. You want to connect this data to the import/export line items of GDP. Read Trade Wars are Class Wars for an amazing breakdown of this.
Prices: This is simply looking at prices across the entire economy. CPI, PPI, and PCE are the main ones in the United States. Always connect these prices to how supply and demand are expressing themselves in the other datasets.
Financial Accounts/Balance Sheets
Household, Corporates, Financials, Sovereign.
The capital structure is key because it is the mechanism that transmits the flows of GDP. What you want to do is break down the balance sheet of every agent in the system and then see what asset and liability mismatches exist.
The Integrated Macroeconomic accounts are the main dataset for this.
Quick Summary and Thoughts:
Ok I know that was a lot but here is what I would do:
First, identify every dataset reflecting these ideas in a country and aggregate a list of them. Then go onto the dataset description and read about the dataset.
Second, just start piecing things together and build a picture of the economy. Connect the various frequencies of releases. All the datasets should connect together and build a picture of what is happening. Your explanation and thesis should take into account all continuity and discontinuity between datasets.
Third, all economic data is usually free. You can find all the economic data you need on the FRED website, Atlanta FED, or NY FED.
Fifth, I am really not doing economic data justice here. Someone who really understands economic data would probably be disgusted with how much of a broad brush I am using right now. But my goal is just to get you started though!
Let’s say you’re not a data scientist with an advanced degree from Wharton (I know a tragic thought 🙄), what are some simple tools you can use?
To start, just use a moving average and rate of change. It will simplify a lot when backtesting any ideas.
I cannot tell you how many times I have heard some super complex explanation about some economic data point and it doesn’t happen. And then some moving average crossover signal just crushes a bunch of PhDs.
I am not anti complexity or PhDs but if you just want to keep it simple, moving averages, rate of change and stdv will get you a long way.
Any real hedge fund has models running on all economic data points and then monitors how they get priced in across all assets. In my experience, there is a ton of edge in knowing the economic data really well. The main advantage is that you simply know what is NOT happening.
Before trying to predict the future, you have to correctly analyze the present. I cannot tell you how many people just get analyzing the present wrong. Economic data brings a lot of clarity to understanding the present.
In the next article, we will talk about connecting these economic data points to asset markets. So be ready!