It's (Not Just) the Economy, Stupid
I wondered what would happen if I looked at the correlation between unemployment and presidential approval/disapproval ratings, having claimed in a number of comments this last year that the economy is the super-dominant factor in the mood of the country right now and, implicitly, generally how the country views a President.
The data is actually available from 1948 forward: unemployment data, Presidential job approval polls.
With some Emacs macros, some cleaning up (just how many date formats could they possibly have come up with? At least 6...), 200 lines of Lisp code, and a high-speed digital computer, an answer magically appeared.
The table entries are correlation coefficients: how does a President's approval ratings (1st column) and disapproval ratings (2nd column), averaged over all polls taken during a month, correlate with the unemployment rate for that month? Since higher approval is a bigger number, and higher unemployment is also a bigger number, you'd expect the correlation coefficient for the first column to be negative, and for the second column to be positive, if you assume popularity is inversely related to unemployment. A value of 1.00/-1.00 would indicate perfect
correlation/anticorrelation.
Appr Disa
TRUMAN : 0.64 -0.70
EISENHOWER : -0.67 0.77
KENNEDY : 0.47 -0.66
JOHNSON : 0.90 -0.90
NIXON : -0.32 0.50
FORD : -0.70 0.76
CARTER : 0.23 -0.21
REAGAN : -0.55 0.48
BUSH : -0.58 0.80
CLINTON : -0.72 0.49
BUSH : 0.23 -0.14
So clearly I was wrong: the economy (as measured by unemployment) is not always, or even usually, strongly correlated strongly with Presidential approval. Only 5 out of 11 Presidents show a strong correlation, and 2 out of 11 show a strong inverse.
Let's consider each President in turn, listing the approval correlation coefficient:
Truman: 64%. The key here was the Korean War. Even as unemployment was declining from a peak of more than 7% in 1949 to a low of less than 3% through 1952, his popularity went down and down as the Korean War became a quagmire.
Eisenhower: -67%. No major wars.
Kennedy: 47%. The unemployment rate really didn't move that much, from 7% when he took office to 5.5% in late 1963. And his popularity didn't move much either, from about 70% when he took office to a bit less than 60%. Since Presidents always start with higher than sustainable popularity there doesn't appear to be much insight to be gained here.
Johnson: 90% (!). The key here was the Vietnam war. Even as unemployment was declining from a peak of about 5.5% when he took office to a low of about 3.5%, his popularity went down and down, and down further.
Nixon: -32%. Nixon started with low unemployment (3.5), and it got worse (up to 6%) as his popularity declined. Then it got a little better (down to 5.5%), but by then he had lots of other problems, to say the least, and his post-re-election approval was dominated by the end of the Vietnam war (a spike up) and then Watergate (big oops...).
Ford: -70%. The economy pretty much dominated his administration.
Carter: 23%. Start: 7.6%, down to 5.6%, up to 8.5%. His popularity basically went continually downward until the Iran hostage crisis initially spiked it, then cratered it.
Reagan: -55%, Bush I -58%, Clinton -72%. No long unpopular military adventures. Iraq I was short and successful, but without coattails. We see a very strong correlation of approval to the economy.
Bush II: 23%. 9/11 caused a massive spike in popularity, even as the economy was tanking. Then, after his re-election, as the economy got better, the Iraq war become more and more unpopular. And as the Iraq war stabilized, the economy went south, and ultimately to hell. It's not terribly surprising that the correlation coefficients are confused!
Conclusions:
Notes:
It would probably be more reasonable to lop off the first 3-5 months of a President's first term since approval always starts out high and always goes down.
Also, unemployment is only one measure of economic health, so justifying the claim it's the economy, stupid on the sole basis of unemployment data is clearly somewhat off the mark.
The data is actually available from 1948 forward: unemployment data, Presidential job approval polls.
With some Emacs macros, some cleaning up (just how many date formats could they possibly have come up with? At least 6...), 200 lines of Lisp code, and a high-speed digital computer, an answer magically appeared.
The table entries are correlation coefficients: how does a President's approval ratings (1st column) and disapproval ratings (2nd column), averaged over all polls taken during a month, correlate with the unemployment rate for that month? Since higher approval is a bigger number, and higher unemployment is also a bigger number, you'd expect the correlation coefficient for the first column to be negative, and for the second column to be positive, if you assume popularity is inversely related to unemployment. A value of 1.00/-1.00 would indicate perfect
correlation/anticorrelation.
Appr Disa
TRUMAN : 0.64 -0.70
EISENHOWER : -0.67 0.77
KENNEDY : 0.47 -0.66
JOHNSON : 0.90 -0.90
NIXON : -0.32 0.50
FORD : -0.70 0.76
CARTER : 0.23 -0.21
REAGAN : -0.55 0.48
BUSH : -0.58 0.80
CLINTON : -0.72 0.49
BUSH : 0.23 -0.14
So clearly I was wrong: the economy (as measured by unemployment) is not always, or even usually, strongly correlated strongly with Presidential approval. Only 5 out of 11 Presidents show a strong correlation, and 2 out of 11 show a strong inverse.
Let's consider each President in turn, listing the approval correlation coefficient:
Truman: 64%. The key here was the Korean War. Even as unemployment was declining from a peak of more than 7% in 1949 to a low of less than 3% through 1952, his popularity went down and down as the Korean War became a quagmire.
Eisenhower: -67%. No major wars.
Kennedy: 47%. The unemployment rate really didn't move that much, from 7% when he took office to 5.5% in late 1963. And his popularity didn't move much either, from about 70% when he took office to a bit less than 60%. Since Presidents always start with higher than sustainable popularity there doesn't appear to be much insight to be gained here.
Johnson: 90% (!). The key here was the Vietnam war. Even as unemployment was declining from a peak of about 5.5% when he took office to a low of about 3.5%, his popularity went down and down, and down further.
Nixon: -32%. Nixon started with low unemployment (3.5), and it got worse (up to 6%) as his popularity declined. Then it got a little better (down to 5.5%), but by then he had lots of other problems, to say the least, and his post-re-election approval was dominated by the end of the Vietnam war (a spike up) and then Watergate (big oops...).
Ford: -70%. The economy pretty much dominated his administration.
Carter: 23%. Start: 7.6%, down to 5.6%, up to 8.5%. His popularity basically went continually downward until the Iran hostage crisis initially spiked it, then cratered it.
Reagan: -55%, Bush I -58%, Clinton -72%. No long unpopular military adventures. Iraq I was short and successful, but without coattails. We see a very strong correlation of approval to the economy.
Bush II: 23%. 9/11 caused a massive spike in popularity, even as the economy was tanking. Then, after his re-election, as the economy got better, the Iraq war become more and more unpopular. And as the Iraq war stabilized, the economy went south, and ultimately to hell. It's not terribly surprising that the correlation coefficients are confused!
Conclusions:
- Unpopular wars are really, really bad for Presidents and other living things.
- No one is ever particularly popular in a bad economy, except when they first take office.
- In the absence of traumatic foreign or domestic upheavals, it does seem that it's the economy, stupid, though a more nuanced analysis is obviously called for (and, I'm sure, must have been done by someone, somewhere).
Notes:
It would probably be more reasonable to lop off the first 3-5 months of a President's first term since approval always starts out high and always goes down.
Also, unemployment is only one measure of economic health, so justifying the claim it's the economy, stupid on the sole basis of unemployment data is clearly somewhat off the mark.