Studying Macroeconomic Effects on Company and Industry Performance
Summary
The document considers whether macroeconomic variables such as GDP growth, inflation, employment, and fiscal spending relate to company revenues or earnings across industries. It proposes using company-level panel data and asks whether a fixed-effects model could estimate different sensitivities, such as how revenue growth varies with GDP growth by industry. One response notes that low-frequency GDP releases limit the amount of usable data. It also warns that national output composition changes over time, and that GDP revisions can make historical analysis misleading unless researchers use point-in-time data vintages.
A second response reframes the question around whether GDP growth is associated with excess stock returns and points to research connecting earnings per share with GDP. The page gives methodological cautions and a possible literature direction, but it includes no dataset, model specification, estimates, or evidence that a particular relationship holds. Company business mix can also evolve, complicating comparisons over long periods. These constraints matter when interpreting any estimated industry or firm sensitivity.
Key ideas
- Panel data could be used to study links between macroeconomic variables and company performance across industries.
- GDP's release frequency may leave a limited number of observations for analysis.
- Changing economic composition and company business mix can weaken historical comparability.
- GDP revisions make point-in-time data vintages important for credible backtests or regressions.
- A related research framing asks whether GDP growth predicts excess stock returns or earnings growth.
Tags
Full text
# Cash flows regression on macroeconomic data # Cash flows regression on macroeconomic data I'm looking into a research project and am struggling to find any existing work on this or whether I'm asking the right question. My question is to test the relationship between macroeconomic variables (GDP growth, inflation, employment, fiscal spending etc.) and the financial performance (revenues, ebitda etc.) of companies in various industries of the country - with the idea to test whether this relationship exists and whether some industries are more invariant to economic shocks? The ultimate result would a variable that gives the relationship, for example gdp growth of x% would given revenue growth of y% in a certain industry vs. z% in another. My econometric knowledge is rusty but given if the question is viable I'll have panel data (time series growth for multiple companies and macro variables) and need to run some sort of fixed effects model? ## Answer by user42108 (score 1) https://quant.stackexchange.com/a/50100 This doesn't answer your question directly but might be helpful: - you won't have much data given the frequency of release of GDP - composition of GDP has changed significantly over time (e.g. less VA from manufacturing, more from services) - GDP is revised substantially and a long time after initial release (e.g. corporate profit component of US GDP was recently revised back to 2014). Failing to use "vintage" or "point-in-time" data - which can be difficult and expensive to obtain - might render your results useless - what any given company does might change significantly over time (e.g. IBM moving from hardware to services) ## Answer by Mild_Thornberry (score 0) https://quant.stackexchange.com/a/50084 Normally, your question is formulated by asking if gdp is linked to stock performance, since gdp is output, it is very clearly linked to revenue growth. In which case, the question is: does gdp growth lead to higher excess returns? Just food for thought. Either way, here’s a paper I came across myself when I was looking into the same question. They site a few studies of the same flavor as your proposal, and focus on EPS vs. gdp. Perhaps this can be a starting point for https://www.msci.com/documents/10199/a134c5d5-dca0-420d-875d-06adb948f578
Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.