With the presidential election a mere 127 days from the release of this
report, and the candidates apparently not waiting for the
once-traditional Labor Day kickoff, this is a good time to look at the
partisan patterns in some major economic and financial indicators. The
differences are significant, and worth thinking about for anyone with
dollars at stake after January 20, 2009.
Not to spoil the suspense too much, but here are the basic
conclusions. Since Franklin Roosevelt’s third term (1941–44),
Democrats have generally presided over faster growth and stronger stock
markets than Republicans; Republican administrations have been
friendlier for disinflation and the bond market. Also, Republicans tend
to preside over recessions early in their terms, with growth
accelerating as time passes; Democrats tend to preside over earlier
accelerations followed by slowdowns as the term matures.
Here’s a closer look at some major indicators. In the graphs,
the parties are color-coded by the traditional Republican red and
Democratic blue. Individual terms are in a lighter shade, and the party
average is the darker shade.
Technical note Unless otherwise noted, the figure shown is average
annual growth rate for a president’s term, from the year,
quarter, or month of inauguration to the quarter or month of the next
inauguration. For two-term presidents, results are the averages of both
terms.
Editors’ note: Our intent here is to present the historical record to balance the tiresome, and generally incorrect, generalizations one hears at this point in the election cycle. We are not venturing into politics. Philippa’s grandfather is Finley Peter Dunne, creator of the Chicago barkeep Mr. Dooley, and we’ll refer anyone with a horse in the race who is unhappy with how specific graphs may look to him: "Politics ain’t beanbag."
GDP
Democrats have a clear edge on GDP growth: 4.4% vs. 2.6%. Even if you
start the clock with Truman in 1949 (eliminating the war boom and
immediate postwar bust), the Dem advantage survives, with average
growth of 4.5%. The partisan difference is widespread, too, not
dependent on a few strong or weak readings: the blue bars stack towards
the top of the graph, and the red bars towards the bottom. It might
surprise some readers to learn that the Carter years weren’t
quite as bad as some remember—though the inflation
performance was miserable.
In his new book, Unequal Democracy, Princeton political scientist Larry
Bartels points out an interesting partisan contrast in the timing of
GDP growth. Republican administrations tend to have recessions early in
their terms, with growth strongest in the fourth year. Democrats tend
to stimulate the economy on taking office, which leads to a prompt
acceleration in growth, followed by a slowdown. (The graph to the right
supports this point. There’s volatility around the averages, of course,
but there’s still a pattern there.) Bartels theorizes that this rhythm
makes it easier for Republicans to get re-elected, since voters
generally remember only recent history, and are probably more impressed
by improvements than averages.
Employment
The comparative partisan performance on employment is similar to GDP
growth. Under Democratic administration, employment has grown an
average of 3.0% a year (2.9% if you start in 1949); under Republicans,
1.3%. And here the blue bars are all at the top of the graph, and the
red at the bottom. And it seems that the Bush family, whatever their
other accomplishments, won’t go down in history as great job
creators.
Reagan’s average was dragged down by the recession during his
first term; his second term saw strong employment growth, an annual
average of 2.7%, stronger than Bill Clinton’s average, and
comfortably the strongest of any Republican four-year term. The only
sub-1% terms were Eisenhower’s second (0.4%) and George W.
Bush’s first (0.0%). The weakest Democratic terms on job
creation were Roosevelt–Truman (1.6%) and Clinton’s
second (2.2%).
Unemployment
With only few exceptions, Republican administrations have presided over
increases in unemployment, and Democrats over declines. On average, the
jobless rate has risen by 1.0 points under the GOP, and fallen by 1.9
points under Dems (–1.3 points if you start in 1949). The
only exceptions to the partisan pattern were Reagan (–2.1),
the Roosevelt–Truman joint term (+3.2), and Carter (no
change).
We figured it was best to feature changes in the unemployment rate, but
the averages show a similar pattern. Under Republicans, unemployment
has averaged 6.0%, and 4.8% under Democrats (or 5.2% if you start with
Truman’s first full term).
Inflation
The inflation pattern is more mixed than the growth-related numbers. On
average, Democrats preside over a small increase in inflation, and
Republicans over a small decrease. The top and bottom positions are
amplified versions of this distinction. Under Reagan, inflation
(measured by the headline CPI) came down by 7.3 points; under Carter,
it rose by 6.6 points. Still, there’s a mix of red and blue
on both sides of the zero divide.
Fiscal Shift
Though the picture so far is of the Republicans as the party of
austerity and the Democrats as the party of stimulus, there’s
a surprise when it comes to changes in the federal deficit: Republicans
are more liberal with the red ink than Dems. On average, a Republican
in the White House has meant a shift of –1.9% of GDP in the
government’s budget balance (i.e., towards smaller surpluses
or bigger deficits), while a Dem has meant a 1.5% improvement in the
budget position (or 1.8%, if you start in 1949, thereby omitting the
huge World War II deficit). And in this case, the average is a faithful
representation of the distribution, with only one Democrat in the minus
column and only one Republican in the plus.
Some of this reflects different tax policies, with Reagan and Bush 43
cutting, and Clinton raising income taxes. But it also reflects the
partisan difference in GDP growth.
Stock Market
The blue years have an edge on stock returns, with the S&P 500
rising an average of 4.7% a year in real terms (price only, excluding
dividends, deflated by the CPI) under Democratic administrations,
compared with 2.9% under Republicans. (Starting the clock in 1949
raises the Dem average to 6.9%.) Still, there are some red bars towards
the top of the heap and blue bars toward the bottom.
Strangely, there’s not all that tight a link between stock
market performance and profit growth. In fact, the rankings of the two
measures show a correlation coefficient of 0.43. Sometimes stocks march
to their own drummer.
Bond Market
Unlike the stock market, there’s a clear partisan pattern to
bond returns: Republicans are a lot more bond-friendly. Real total
returns—price plus coupon, deflated by the
CPI—averaged +4.2% a year under Republicans, vs.
–2.1% under Democrats. And, as the graph shows, the average
is a pretty faithful representation of the relative performance of
individual administrations.
Recall that Clinton came into office with plans for a stimulus program,
that were shelved under pressure from what he called “a bunch
of ******* bond traders.” This should be kept in mind when
evaluating the bond market’s prospects should Obama win in
November. It may be that the world has changed to the point where the
old Democratic pattern won’t hold this time.
Distribution
Over the long sweep of history, the distribution of income in the U.S.
became more equal from the early 1930s through the late 1960s, and has
been growing more unequal ever since. But there are some partisan
patterns to this story. On average, inequality has risen in Republican
administrations, and fallen in Democratic ones. Bucking the long-term
trend, inequality rose slightly during the Eisenhower years. And while
not quite bucking the trend, it rose more slowly in the Carter and
Clinton years than it did under Nixon, Reagan, or George H.W. Bush.
(Technical note: inequality here is measured by the Gini index, a
number between 0 and 1 that represents the relative equality or
inequality of a distribution. A perfectly equal society, in which all
members have the same income, would have a Gini index of 0; a perfectly
unequal one, in which one person had all the income, would have a Gini
of 1. The level of the Gini makes little sense on its own;
it’s best used to compare distributions across time, or to
compare countries. The measure shown here is the percentage change in
the Gini index over the presidential term; for two-term presidents, the
change is divided by 2 for consistency with one-term presidents.)
In the book mentioned above, political scientist Larry Bartels looks at
personal income growth at various points in the distribution, and finds
that under Republican administrations, growth is strongest at the 80th
percentile and above; under Democrats, it’s pretty equal
across the distribution. Bartels, by the way, says he undertook his
research as a totally nonpartisan, objective political scientist of a
quantitative bent, and had absolutely no preconception about where his
research would take him.
Coda
When John Liscio first commissioned us to look into Presidential
economic records, we tried putting together a grading system. But we
gave up—it’s impossible to figure out how to weight
all the different variables. Tastes and interests vary.
Still, it looks like George Wallace was wrong. There is at least a
dime’s worth of difference between the two parties. Past
performance is, of course, never a guarantee of future results, but the
trajectory of the U.S. economy and markets from 2009 forward may well
depend in a large degree on who wins on November 4.
—Philippa Dunne & Doug Henwood
Nicely done!
Do parties matter in Presidential economics?
The well regarded Liscio report is now blogging. They did a very nice job analyzing the various economic and market results under all of the Presidents since FDR. Of course, it is arguable as to how much impact any President has on the economy in gener…
I’m glad to see you’ve added a blog… Barry Ritholtz has mentioned LR several times, including the news of the blog.
I’ve always said that the Republicans did just fine with the economy. It was just the amount of money they had to borrow to do it that worried me.
Unfortunately this analysis fails to consider two critical factors.
#1 when a president is elected, it takes time to pass laws to change policies.
#2 once laws are passed, it takes time to implement new policies and then for those policies to affect GDP. Companies make spending decisions years ahead, not days ahead. And tax rates are never changed in the first year of an administration, at least for that particular year.
Think about these two things, both noted by the author:
#1 “Republicans tend to preside over recessions early in their terms, with growth accelerating as time passes”
#2 “Democrats tend to preside over earlier accelerations followed by slowdowns as the term matures”
It appears, not surprisingly, that GDP benefits under Democrats at the beginning of their terms from policies passed by previous Republican administrations. It also appears that GDP suffers under Republicans at the beginning of their terms from policies passed by previous Democratic administrations.
This analysis is shortsighted. Certainly neither party is perfect (and definitely not the current administration), but policies passed by Republican administrations have historically been dramatically more stimulative than those passed under the Democrats. This analysis attempts to explain away Economics 101 and unfortunately did not consider some major critical factors.
On a side note, I would encourage any of you to read Milton Friedman’s book Capitalism and Freedom. It is a fantastic read and although not an easy read, provides some very compelling arguments for free market capitalism.
From Dani Rodrik’s blog, dated March 31st 08. Dovetails with this info.
http://rodrik.typepad.com/dani_rodriks_weblog/2008/03/american-politi.html
I would suggest Angry Bear, having spent several months fielding question and criticism from all sorts.
I think you need to factor in the general slowing of GDP due to a maturing economy, and a hyperactive Federal Reserve that helped create the great moderation.
Other influences include the degree of ability to produce in the post-WWII period and the willingness of neo-mercantilist nations to subsidize exports and fund the US current account deficit.
It also could be useful to segment on unified and divided governments — gridlock often produces the best of times, while unified governments tend to injure the US economy.
Roosevelt-Truman is miscolored in the employment graph, and that mistake carries into one of the remarks on that graph. It does not appear to have affected the averages.
Liscio Blog: Thank you Richard for pointing out our error. It has been corrected.
SP’s comment is right on target. This whole piece is fundamentally misguided.
The president doesn’t even have much influence over the Federal budget. Every year, as long as I’ve been watching, the White House proposes a budget and the House leadership describes the proposal as “dead on arrival.” It’s almost like a game to these people.
The president’s only hope for a substantial influence over the budget process is based on election results, and lasts only about a year. If the president can claim a mandate for the economic policies he advocated during his campaign, he can apply some leverage to Congress. Many presidential elections don’t focus so strongly on specific economic policies, and many fail to produce any kind of mandate.
Whatever budgetary influence the president has doesn’t begin to be felt in the economy until almost two years into his term (because the budgeting process is so time-consuming, the White House has limited influence over the budget that takes effect the October after the new president takes office)– which means whatever happens early in a president’s term needs to be credited to (or blamed on) his predecessor.
Bottom line, this whole piece is basically useless. Start over and do it more honestly, and let’s see how it works out. I haven’t done the math myself, but I suspect it’s going to show that the two parties aren’t so easily distinguished.
. png
Interesting data but the entire analysis succumbs to a basic statistical pitfall: correlation is not causation. Indeed, professors must constantly repeat this mantra because of analysis like yours.
This isn’t to say that pouring over such correlations isn’t interesting, but we certainly cannot make any correct inferences from the data you present.
The comments by SP, Peter G., and Matt are well-phrased but unfortunately misguided. a few points on this:
(1) Correlation may not be causation, but numerous pieces of correlative evidence can help point out potential causes. They aren’t the same thing, but there is a link, or else inductive evidence would be worthless.
(2) It’s ludicrous to suggest that the President has minimal control over the budget, and reflects a poor understanding of the inner workings of modern US governance. Yes, Congress is officially responsible for budgeting. However, a significant quantity (possibly most) of the information available to congressional committees comes from executive agency requests and testimony. Furthermore, informal lines of communication grant the president substantial influence over the law (and budget) making process.
(3) The standard “lag-time” argument is also short-sited, given that a number of these presidents were in office for two terms – Regan, Nixon, Clinton, Bush 43, and Johnson to name a few. a 6-8 year presidency is more than enough time, particularly given the popularity of Nixon and Regan and LBJs political skills, to enact sufficient law to make these numbers reliable.
(4) Deficit numbers are a matter of direct causation based on republican vs. democrat administrations – the fact is that most democratic social programs are allocated based on existing revenues or are largely self-funded (e.g. Social Security) and difficult to change. Republicans tend to spend money on military endeavors (at least since Nixon) comparatively – the number one item in discretionary spending. Deficits go up because of tax cuts and increased spending programs – classic Reganomics – often into unproductive or short-term productive sectors (e.g. military contracts). Democratic measures tend to be in civilian investment and/or in “safety net” provisions that minimize flux and redistribute income downward to income groups with a higher marginal propensity to consume and marginal utility of each dollar earned.
(5) Otto’s link to Dani Rodrik’s blog is useful – especially given Redrik’s reputation as a balanced and well-thought-out, careful economist. Here’s the link again: http://rodrik.typepad.com/dani_rodriks_weblog/2008/03/american-politi.html. Note that this data supports my contention about income redistribution trends – which also matches with party platforms, common sense about the parties tendencies, and numerous other studies in modern economics.
(6) Lastly, it might be worth mentioning that many classic neoliberal economists – Friedman included – adopted a much softer, more nuanced stance and recanted on a number of views after the 1980s. The Chicago school was too extreme, and many of those economists have realized they were making simplistic assumptions about human psychology and behavior – and that the models didn’t always apply to current economic circumstances.
(7) Having said all this – if inflation hits 10% or above… elect a republican.
It’s fair to raise the question of how Democratic administrations benefitted from the policies of Republican ones (and, additionally, I think this data set is still too small), but as a counter, it should be noted that these results run true even for successive Republican and Democratic administrations. And if this were merely a factor of Democrats benefitting from strong Republican policies, you’d expect to see a weaker Johnson Administration and a far stronger Bush 41 Administration—yet these numbers stay stratified along a partisan axis.
Correlation does not equal causation and laws do take time to have an effect, but merely rejecting the data for these reasons would be foolish. I don’t know the answer for why the data would line up as it does. The consistency of these various measures should at least give one pause to consider what actions of Democratic and Republican Administrations might lead to these results. The knee-jerk dismissals above don’t even begin to address the issue. If they did, the data would look very different.
Unlike many commenters on blogs, political scientists do not (as a general rule) start with their conclusions and then go out and look for evidence in the world that supports their views. Bartels didn’t just make up a theory sitting on his couch and then find data to support it. He did research,research,research. Here is another paper that makes similar arguments and provides corroborating evidence. Conservatives can try and argue around this one too. http://polsci.colorado.edu/COURSES/PSCI_7108/data/Adler%20and%20Leblang%20paper.pdf
What we is undeniable from this information is that most Americans enjoy better economic outcomes under Democratic administrations. And of course, that means that most Republicans will have to ask Arthur Laffer (or the Tooth Fairy) to explain to them how to ignore the facts so they can clamor for more tax cuts.
Although the authors punt on rating economic performance, I think that a voter has to ask himself how he makes his money: wages or interest?
If wages, the Democratic performance wins on jobs, growth, and reduction in income disparity.
If interest, the Republican performance wins on inflation, bond performance and increase in income disparity.
If you subscribe to the theory that low deficits reduce interest rates, Democratic policies again favor wage-earners over interest-earners.
Of course, I’d posit that more than 90% of all voters make more in wages than in interest, so most Republican voters are voting against their personal economic interests.
We need to make a significant correction to the inflation reported during Bush 43. The Bush CPI no longer includes food or energy costs, which were included in earlier inflation calculations. I believe that you would have an inflation at least double what is currently reported.
Some historians may point out that the CPI has been adjusted many times over the years; however, given the current times, omitting the escalating prices of food and energy serious miscalculations. Besides Carter never took those out!
The reason the Democratic Party is the One to Back!!!!!! | politikly.com
\r\nDemocrats have been better at GDP Growth,Employment growth,Unemployment rate changes,Fiscal Shif