Economy vs the President

Around the 2000 election, people where mentioning that the markets prefer Bush because, traditionally, Republicans have been better for the economy.

So, I tried to find out data to see *how * much better they are, in terms of GDP growth and unemployment. I used data from here, and here. I focused on the period from the 1950’s to today, i.e. the post-WWII period.

I was surprised to find that, in fact, Democrats had a much better record of GDP growth and an even better record on unemployment.

This is what I found in 2000. I kept updating the results over the past four years, and the pattern remains.

See below:

Percent change in unemployment rate during a president's term:
        Eisenhower:      56 (R)
        Eisenhower:      57 (R)
   Kennedy-Johnson: -24     (D)
           Johnson: -32     (D)
             Nixon:      53 (R)
        Nixon-Ford:      50 (R)
            Carter:  -8     (D)
            Reagan:       1 (R)
            Reagan: -27     (R)
              Bush:      40 (R)
           Clinton: -27     (D)
           Clinton: -28     (D)
           Bush Jr:      56 (R)
Republican Average:  36%
Democratic Average: -24%

Percent increase in real GDP during a president's term:
        Eisenhower:     12 |          (R)
        Eisenhower:  8     |          (R)
   Kennedy-Johnson:        |       22 (D)
           Johnson:        |       22 (D)
             Nixon:       14          (R)
        Nixon-Ford:   9    |          (R)
            Carter:      13|          (D)
            Reagan:      13|          (R)
            Reagan:        | 16       (R)
              Bush:   9    |          (R)
           Clinton:       14          (D)
           Clinton:        |  17      (D)
           Bush Jr:   9    |          (R)
           Average: 14%
Republican Average: 11%
Democratic Average: 18%

I also checked to see the composition of the House and Senate during these years, to see if they might have had some effect, but it seems that during most of the above period the Democrats were the majority by a comfortable margin, especially in the House, and have only recently lost the majority, and not by much.

So, it seems the composition of the House and Senate doesn’t explain (correlate with) the change in unemployment or GDP growth. There seems to be a much higher correlation with the party affiliation of the President.

The first conclusion one might make is that “Republicans are bad for the economy”, but the situation may not be as simple as that.
What other explanations of the above data can you find?

It has been said that the President doesn’t have much of an effect on the economy, but the data suggest that, on the contrary, who is in office correlates well with how the economy is doing, so the President must have a bigger effect than people believe.

One explanation of the data could be that it was a random occurrence, and the Democrats were just “lucky” to be President during all the good periods. You can never dismiss this, but for this to happen for the past 50 years, and with both unemployment and GDP growth, the probability of this occuring by chance is very small. (Unless the Democrats are very “cunning” and through sheer genius manage to get elected only during good periods)

To me, the unemployment figures are impressive. There is an almost ping-pong effect where the increase in unemployment rate goes from positive for Republican presidents to negative for Democratic presidents. I find it hard to come up with a plausible explanation that doesn’t conclude that Democrats are better for the economy.

Anyway, any thoughts on what the above data might mean or how it can be explained?

I agree that the data from your source documents supports the contention that Democratic presidents are better for the economy.

You get on really slippery ground, though, taking percentage differences as regular numbers and then computing arithmetic means on them. In this case, you’re lucky, because every Democratic president in the last fifty years has seen the unemployment rate fall, and every Republican president has seen it rise.

Also, except for Carter being one percentage point under the average, and Reagan having one term where he beat the average difference by 2 percent, the numbers are always ‘at or above’ for Dems, and ‘at or below’ for Republicans.

But in general, you can’t just average together percent changes and have them mean anything.

I agree that the means might not correspond to a meaningful quantity, but they are there just to provide a small “summary” of the situation seen in the data.
(in the same vein, the “average american taxpayer” doesn’t exist, but people use it anyway)

What is more important than the averages though is the overall pattern seen in the data, which exhibits what you describe: the numbers are ‘at or above’ for Dems, and ‘at or below’ for Republicans.

It also blows a clean hole in any ‘Laffer curve’ or ‘trickle-down’ economics.

The most prevalent example of the “post hoc ergo propter hoc” logical fallacy is the long-lived fiction that presidents create (reduce) jobs or that they favorably (unfavorably) influence the business cycle in any significant way. By and large the business cycle will do what it will do, regardless of whether Clinton or Bush is in office.

That’s why I tend toward the philosophy that says that the government’s role is to guard against the most egregious excesses of the free market (predatory pricing, price gouging, monopolies) and other than that, to stay out of the damn way.

It is not the government’s role to solve all of our economic problems, nor is this even possible.

It’s a myth, pure and simple. The republican party knows that if they say something enough times, it will eventually be accepted as fact. The democrats don’t seem to be quite as good at this. Even when the facts are on the democrats’ side, they don’t seem to be able to get enough people to believe it.

But the data show a correlation between the party affiliation of the President and the two economic factors I mentioned above.

So, for you to be correct, the observed correlation must be a “coincidence”, the Democratic Presidents just happened to be lucky enough to always get elected during the good periods. This is possible, but I think very unlikely.

It is more likely that in fact there is some mechanism by which Presidents affect the economy (unemployment rate and GDP growth), contrary to your belief (which I must admit is very wide-spread).

This brings up a question in my mind. The charts the OP cites actually make Carter look pretty good, yet Carter is widely criticized for tanking the economy. What measures, if not GDP and unemployment, are people using to make that determination? I know interest rates were incredibly high under Carter; is that what they’re basing it on?

Polerius: I’m just looking at GDP, but I can’t reproduce your numbers. Can you give the exact mathematical formula you are using for your calculations?

Carter governed a period of ‘stagflation’ - the economy wasn’t growing quickly (and the data shows it was slightly lower than average), and inflation was prevalent.

Usually you get low inflation if there’s not good economic growth, so Carter was rapped for that.

Stratocaster - Presidents try to influence employment and the business cycle - they believe they’re doing something.

Correlations don’t prove causality; however, predicting an outcome, testing it, and examining the results is the best indicator of causality in the real world. Let’s do this: I predict that this trend (Democratic presidents = good economic growth and low unemployment) will continue. Is that OK for you?

  • Open
    (which is “Current-dollar and ‘real’ GDP”)
  • Go to the ‘GDP in billions of chained 2000 dollars’, under ‘Quarterly results’
  • Calculate percent change from quarter to quarter
    (e.g. =100*(G9-G8)/G8)

The results I get are

Quarter	Percent change
1947q2	-0.12
1947q3	-0.04
1947q4	1.47
1948q1	1.58
1948q2	1.77
1948q3	0.57
1948q4	0.24
1949q1	-1.49
1949q2	-0.29
1949q3	1.12
1949q4	-1.02
1950q1	4.10
1950q2	2.98
1950q3	3.92
1950q4	1.82
1951q1	1.21
1951q2	1.70
1951q3	2.00
1951q4	0.17
1952q1	1.04
1952q2	0.07
1952q3	0.65
1952q4	3.28
1953q1	1.88
1953q2	0.76
1953q3	-0.60
1953q4	-1.58
1954q1	-0.49
1954q2	0.09
1954q3	1.10
1954q4	1.98
1955q1	2.88
1955q2	1.64
1955q3	1.33
1955q4	0.54
1956q1	-0.47
1956q2	0.79
1956q3	-0.12
1956q4	1.63
1957q1	0.61
1957q2	-0.25
1957q3	0.98
1957q4	-1.06
1958q1	-2.71
1958q2	0.59
1958q3	2.31
1958q4	2.30
1959q1	1.91
1959q2	2.63
1959q3	-0.08
1959q4	0.35
1960q1	2.22
1960q2	-0.50
1960q3	0.16
1960q4	-1.30
1961q1	0.60
1961q2	1.88
1961q3	1.62
1961q4	2.04
1962q1	1.80
1962q2	1.09
1962q3	0.92
1962q4	0.24
1963q1	1.31
1963q2	1.25
1963q3	1.88
1963q4	0.78
1964q1	2.24
1964q2	1.16
1964q3	1.36
1964q4	0.27
1965q1	2.46
1965q2	1.35
1965q3	2.03
1965q4	2.42
1966q1	2.45
1966q2	0.35
1966q3	0.66
1966q4	0.80
1967q1	0.89
1967q2	0.01
1967q3	0.79
1967q4	0.76
1968q1	2.06
1968q2	1.70
1968q3	0.68
1968q4	0.42
1969q1	1.58
1969q2	0.29
1969q3	0.62
1969q4	-0.47
1970q1	-0.17
1970q2	0.19
1970q3	0.89
1970q4	-1.07
1971q1	2.77
1971q2	0.56
1971q3	0.79
1971q4	0.29
1972q1	1.78
1972q2	2.36
1972q3	0.95
1972q4	1.64
1973q1	2.54
1973q2	1.16
1973q3	-0.53
1973q4	0.95
1974q1	-0.87
1974q2	0.29
1974q3	-0.97
1974q4	-0.39
1975q1	-1.20
1975q2	0.73
1975q3	1.69
1975q4	1.31
1976q1	2.25
1976q2	0.75
1976q3	0.48
1976q4	0.72
1977q1	1.21
1977q2	1.96
1977q3	1.79
1977q4	-0.01
1978q1	0.32
1978q2	3.94
1978q3	0.99
1978q4	1.32
1979q1	0.19
1979q2	0.10
1979q3	0.72
1979q4	0.29
1980q1	0.32
1980q2	-2.02
1980q3	-0.17
1980q4	1.85
1981q1	2.03
1981q2	-0.78
1981q3	1.21
1981q4	-1.25
1982q1	-1.64
1982q2	0.54
1982q3	-0.38
1982q4	0.09
1983q1	1.23
1983q2	2.26
1983q3	1.97
1983q4	2.05
1984q1	1.96
1984q2	1.72
1984q3	0.97
1984q4	0.82
1985q1	0.92
1985q2	0.85
1985q3	1.56
1985q4	0.77
1986q1	0.96
1986q2	0.40
1986q3	0.96
1986q4	0.50
1987q1	0.66
1987q2	1.10
1987q3	0.91
1987q4	1.75
1988q1	0.49
1988q2	1.27
1988q3	0.53
1988q4	1.32
1989q1	1.01
1989q2	0.66
1989q3	0.71
1989q4	0.25
1990q1	1.15
1990q2	0.26
1990q3	0.01
1990q4	-0.76
1991q1	-0.51
1991q2	0.65
1991q3	0.48
1991q4	0.47
1992q1	1.04
1992q2	0.96
1992q3	0.98
1992q4	1.10
1993q1	0.12
1993q2	0.51
1993q3	0.51
1993q4	1.35
1994q1	1.02
1994q2	1.30
1994q3	0.56
1994q4	1.17
1995q1	0.28
1995q2	0.18
1995q3	0.81
1995q4	0.73
1996q1	0.71
1996q2	1.64
1996q3	0.84
1996q4	1.17
1997q1	0.77
1997q2	1.52
1997q3	1.24
1997q4	0.74
1998q1	1.11
1998q2	0.66
1998q3	1.15
1998q4	1.52
1999q1	0.85
1999q2	0.83
1999q3	1.17
1999q4	1.78
2000q1	0.25
2000q2	1.57
2000q3	-0.11
2000q4	0.52
2001q1	-0.12
2001q2	0.31
2001q3	-0.35
2001q4	0.39
2002q1	0.84
2002q2	0.59
2002q3	0.64
2002q4	0.18
2003q1	0.48
2003q2	1.01
2003q3	1.80
2003q4	1.03
2004q1	1.10
2004q2	0.69

From this, and from the years each president was in office, I calculate the percent increase in GDP during their term.

If you are still having trouble reproducing my results, I can give you the exact program I did the calculations with.

I was using the annualized numbers. But how do you get the number for the entire term? Do you add up the individual quarterly percentages? If so, that’s not correct. Why wouldn’t you just use: (GDP at end of term - GDP at begining of term)/(GDP at beginning of term)?

The other thing one has to ask onesself is how to determine the start and end points. It’s certainly reasonable to argue that the the policies of a given president in a given year have a delayed effect on the economy. One might either not use the first year of any president’s term, or offset the calculations by one year (eg, Clinton begins to affect the economy in 1994, not 1993) or even two years.

Using the annualized numbers for the Carter, for instance, we get these different results depending on what we consider to be the Carter years:

'77 - '8O: 8.7%
’78 - '81: 5.5%
'79 - '82: .31%

Which of those is the correct number? I would expect that the middle number, 5.5%, is probably the most reasonable one to use. Following that convention, we get:

Reagan I: 16.7%
Reagan II: 11.5%
Bush: 5.9%
Clinton I: 11.1%
Clinton II: 9.1%

Bush Jr hasn’t finished out his “economic term” yet.

I don’t do that.

I do that, I just use the quarterly numbers, instead of the annualized numbers, because I wanted more granularity in being able to define “beginning of term” and “end of term”.

I did take this into consideration; that’s why I used quarterly results in the first place, to be able to have some granularity in defining when the terms begin and end.

So, I can adjust for the “delayed effect” you mention on a quarter by quarter basis.

After reading your post, I went back to my program and varied the value of this delay to see how it affected the results. See below.

* No delay

Percent change in unemployment rate:
        Eisenhower:      56 (R)
        Eisenhower:      57 (R)
   Kennedy-Johnson: -24     (D)
           Johnson: -32     (D)
             Nixon:      53 (R)
        Nixon-Ford:      50 (R)
            Carter:  -8     (D)
            Reagan:       1 (R)
            Reagan: -27     (R)
              Bush:      40 (R)
           Clinton: -27     (D)
           Clinton: -28     (D)
           Bush Jr:      56 (R)
Republican Average:  36%
Democratic Average: -24%
Percent increase in real GDP:
        Eisenhower:     12 |           (R)
        Eisenhower:  8     |           (R)
   Kennedy-Johnson:        |        23 (D)
           Johnson:        |       22  (D)
             Nixon:       14           (R)
        Nixon-Ford:   9    |           (R)
            Carter:      13|           (D)
            Reagan:      13|           (R)
            Reagan:        | 16        (R)
              Bush:   9    |           (R)
           Clinton:       14           (D)
           Clinton:        |  17       (D)
           Bush Jr:   9    |           (R)
           Average: 14%
Republican Average: 11%
Democratic Average: 18%

* 1 quarter delay

Percent change in unemployment rate:
        Eisenhower:      42 (R)
        Eisenhower:      86 (R)
   Kennedy-Johnson: -32     (D)
           Johnson: -28     (D)
             Nixon:      44 (R)
        Nixon-Ford:      51 (R)
            Carter:   0     (D)
            Reagan:  -3     (R)
            Reagan: -31     (R)
              Bush:      40 (R)
           Clinton: -26     (D)
           Clinton: -17     (D)
           Bush Jr:      42 (R)
Republican Average:  34%
Democratic Average: -21%
Percent increase in real GDP:
        Eisenhower:   10   |             (R)
        Eisenhower:  8     |             (R)
   Kennedy-Johnson:        |          25 (D)
           Johnson:        |      21     (D)
             Nixon:        |15           (R)
        Nixon-Ford:  8     |             (R)
            Carter:       14             (D)
            Reagan:     12 |             (R)
            Reagan:        | 16          (R)
              Bush:  8     |             (R)
           Clinton:       14             (D)
           Clinton:        | 16          (D)
           Bush Jr:   9    |             (R)
           Average: 14%
Republican Average: 11%
Democratic Average: 18%

* 2 quaters delay

Percent change in unemployment rate:
        Eisenhower:      72 (R)
        Eisenhower:      60 (R)
   Kennedy-Johnson: -33     (D)
           Johnson: -24     (D)
             Nixon:      40 (R)
        Nixon-Ford:      47 (R)
            Carter:       4 (D)
            Reagan:  -1     (R)
            Reagan: -28     (R)
              Bush:      32 (R)
           Clinton: -29     (D)
           Clinton: -10     (D)
           Bush Jr:      36 (R)
Republican Average:  32%
Democratic Average: -18%
Percent increase in real GDP:
        Eisenhower:   9    |            (R)
        Eisenhower:    11  |            (R)
   Kennedy-Johnson:        |         24 (D)
           Johnson:        |    19      (D)
             Nixon:        | 16         (R)
        Nixon-Ford:   9    |            (R)
            Carter:    11  |            (D)
            Reagan:       14            (R)
            Reagan:        | 16         (R)
              Bush:  8     |            (R)
           Clinton:        | 16         (D)
           Clinton:       14            (D)
           Bush Jr:   9    |            (R)
           Average: 14%
Republican Average: 11%
Democratic Average: 17%

* 3 quarters delay
Percent change in unemployment rate:
        Eisenhower:      52 (R)
        Eisenhower:      52 (R)
   Kennedy-Johnson: -36     (D)
           Johnson: -14     (D)
             Nixon:      30 (R)
        Nixon-Ford:      42 (R)
            Carter:      12 (D)
            Reagan:  -7     (R)
            Reagan: -25     (R)
              Bush:      26 (R)
           Clinton: -27     (D)
           Clinton:       2 (D)
           Bush Jr:      22 (R)
Republican Average:  24%
Democratic Average: -13%
Percent increase in real GDP:
        Eisenhower:     11  |             (R)
        Eisenhower:     11  |             (R)
   Kennedy-Johnson:         |          25 (D)
           Johnson:         |   18        (D)
             Nixon:        14             (R)
        Nixon-Ford:     11  |             (R)
            Carter:     11  |             (D)
            Reagan:        14             (R)
            Reagan:         |15           (R)
              Bush:  7      |             (R)
           Clinton:         | 16          (D)
           Clinton:       13|             (D)
           Bush Jr:    9    |             (R)
           Average: 14%
Republican Average: 12%
Democratic Average: 16%

* 4 quarters delay

Percent change in unemployment rate:
        Eisenhower:      16 (R)
        Eisenhower:      15 (R)
   Kennedy-Johnson: -33     (D)
           Johnson: -12     (D)
             Nixon:      40 (R)
        Nixon-Ford:      31 (R)
            Carter:      33 (D)
            Reagan: -18     (R)
            Reagan: -23     (R)
              Bush:      20 (R)
           Clinton: -28     (D)
           Clinton:      21 (D)
           Bush Jr:       7 (R)
Republican Average:  11%
Democratic Average:  -4%
Percent increase in real GDP:
        Eisenhower:    12 |             (R)
        Eisenhower:       |15           (R)
   Kennedy-Johnson:       |          25 (D)
           Johnson:      14             (D)
             Nixon:       | 16          (R)
        Nixon-Ford:  10   |             (R)
            Carter:  9    |             (D)
            Reagan:       |  17         (R)
            Reagan:      14             (R)
              Bush:  9    |             (R)
           Clinton:       | 16          (D)
           Clinton:    12 |             (D)
           Bush Jr:  9    |             (R)
           Average: 14%
Republican Average: 13%
Democratic Average: 15%

It seems that the Democratic “advantage” diminishes as we increase the “delayed effect”, but nevertheless remains quite large through 3 quarters delay, and then reduces to a small advantage at a 4-quarter delay.

Now, which value of delay is appropriate is of course up to debate.

I think they were talking about the securities markets – stocks, bonds and commodity futures, as distinct from commerce in actual goods and services.

Which raises an important question about economics: Is what’s good for the economy (in its aspects that immediately affect the lives of ordinary middle-class and working-class consumers) always good for Wall Street, and vice-versa? I don’t know the answer, but I would hesitate to simply assume (as Pubbies apparently do) that the answer is an unqualified yes.

(Diving into the part that’s “up for debate”): Well, John Mace, I truly think you’re on to something there. Believe it or not, I’d say that your third number for Carter, the .31% number, is the most accurate. Why?
Well, as much as it pains me to say this, and believe me it does, Reagan didn’t have his first budget take effect until 1982. Check out the start of the fiscal years on the Bureau of the Public Debt’s historical site for 1950 - 2000: in Reagan’s time, the fiscal year was synched up with the calendar year, so that the first year that the US experienced under Reagan’s economic watch was 1982. In fairness, then, you really can’t hold him responsible, I believe, for anything that happened economically until 1983. You should really at least give a new President a year for after his policies are fully put in place to see the effect.
Using that same rule, Bush I can’t be held responsible until September, 1990, and his “economic term” would have run out in September, 1994. Clinton’s would then have run from then until September, 2002. I’m sure that would put a far different light on the above statistics. Care to give that a whirl, Polerius?

One possible counterargument that could be made is that the cause-and-effect go the other way. I.e., that Democrats tend to be elected when there is a recession and the unemployment rate is high because people perceive them as being more sympathetic to such plights. Thus, the idea would be that the Dems tend to be in office for the climbs out of a recession when the unemployment rate drops and the GDP rises most rapidly. I don’t know if I believe this counterargument could account for the data or not, but it is something that has to be considered as a possibility.

On the other hand, I would argue that the correlation between high budget deficits and Republican Presidents spouting supply-side garbage is rock-solid. Not just because the statistics are there but also because all aspects of the situation can be explained mechanistically.

When there is a correlation between two time series, it’s often true that there is a correlation. However, it may not be that A caused B: it might be that B caused A, or that they were both caused by another factor C. So in this case, it may not be that the president caused the economic conditions: it may be that the economic conditions caused the choice of president. Further, the economic conditions may be running on a cycle with a period of 4 to 8 years.

So when economic times are good, the voters choose a Republican, and the cycle means that economic times will become bad in the next 4 to 8 years. On the other hand, when economic times are bad, the voters choose a Democrat, and the cycle means that economic times will be good over the next 4 to 8 years. That theory would seem be consistent with the figures, and with the choice of president having no effect whatsoever on the economy.

And I see that jshore posted a very similar theory while I was composing my last message. Coincidence, yes – but, of course, it doesn’t prove that we are both right.

In order to make that argument,though, you’d need something to back it up. Something, that is, other than your own opinion. :slight_smile:

As for how much delay one should use, I really don’t know. There are probably different time offsets for different policies. The only thing I’m certain of is that using Q1 of the president’s first term is NOT correct. And if you’re going to use averages, be sure you also calculate the std dev to see how much variability is built into the numbers I’d also not use Bush Jr, since his term isn’t up. It would make sense to add him back in once the agreed upon time offset has expired.

That could be one explanation.

On the other hand, it is hard to see how some things Presidents do, like the massive tax cuts and huge defense spending under Bush Jr., have *no * effect on the economy. Not saying that they would have a good or bad effect, just saying that they should have some effect. Of course, maybe the U.S. economy is so enormous, that even these things don’t change it’s direction much.

As an aside, on the political level, presidential candidates do campaign on economic issues. That is, even if a president can’t change the economy at all, they pretend that they can. So, if the numbers show one party is better than the other in terms of the economy, can’t that party use those numbers to say “see, we’re better” ?

Will the other party then come out and say “well, the president has no effect on the economy, so those numbers don’t mean much” ? If that happened, no one would make the economy a campaign issue any more, since they have admitted to the public the utter futility of economic reforms by the president.