Today I read the following claim:

List of statistics that improved from 2009 to March 17, 2016

Under President Obama:

  • The DOW increased from 6626 to 17481
  • The S&P 500 increased from 683 to 2040
  • Unemployment decreased from 10% to 4.9%
  • Uninsured adults decreased from 18% to 11.8%
  • American cars sold increased from 10.4M to 17.5M
  • Deficit as a % of GDP decreased from 9.8% to 2.8%
  • Consumer confidence increased from 37.7 to 92.2

Now I haven't checked this claim - but the broader idea is that Presidential policies have a strong correlation with economic change.

My intuition tells me this has more to do with the performance of the global economy.

My question is: What evidence is there to suggest a correlation between the economic performance of the US and the policies of its President?

  • 1
    Other graphs have shown data that nearly all the market gains have been under democratic presidents. Commented Mar 21, 2016 at 14:39
  • 1
    I suspect, but have no evidence, that the relationship goes: Presidential policies -> proposed budget submitted to Congress -> Tax/Incentive plans -> Economy.
    – Bobson
    Commented Mar 21, 2016 at 18:40
  • 2
    I don't know what proof there is, but can safely assume facebook memes aren't one of them.
    – user1530
    Commented Mar 21, 2016 at 18:56
  • 1
    Looking at it from the UK. Obama come to power at a VERY low point for the US economy. At the same he come to power, the US had controlled public spending. Therefore he could QUICKLY spend more. This has a much to do with the spending being under control in the good years, then it has to do with increasing it in the bad years. (The UK had much less room to increase public spending, due it it being out of control in the good years) Commented Mar 22, 2016 at 11:47
  • 1
    Policy changes typically take 2 to 4 years to start to have a widespread impact. Commented Mar 22, 2016 at 15:00

1 Answer 1


Correlation doesn't prove causation. Saying, "Obama was president and the economy recovered; therefore he caused the economy to recover," as the sum total of one's argument is only slightly more compelling than saying "Obama was president and the sun rose every day; therefore he caused the sun to rise."

It's actually very hard to determine the effect of any policy on the economy for a simple reason: we don't know what would have happened if that action hadn't been taken (the counter factual). Because of this, economists and social scientists develop complex models to try to resolve these issues. To have a reasonable chance at discerning truth from all the noise, good data scientists look for a natural experiment or statistical instrument. They find something exogenous that triggers the policy to a random part of a larger group group. For example, we could assess the impact of immigration by looking at a program that randomly assigned various numbers of migrants to different cities.

In the absence of a natural experiment, many statisticians simply try to control for all possible variables in the aim of seeing the relationship of one variable and one other one. The problem with this approach for a large complex system like the US economy is that it basically assumes that the writer of the model has most of the relevant data to explain any changes not caused by what he or she is trying to measure. This is an almost impossible assumption for trying to empirically understand something as complex as the entire US economy. It is far beyond the capabilities of modern data analysis to understand the sum total and interacting effects of all the executive actions of a president, on something as complex as the entire US economy.

Since we can't see the counter factual world of a Nobma presidency, Obama's influence wasn't randomly hoisted on separate parts of the country, and we don't have all the relevant data to understand the world economy, we don't have a strong basis for saying one way or another what the sum total impact of all the President's policies has been.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .