Americans have high hopes for the president's jobs speech tonight. They shouldn't. Statistical evidence shows that democratic leaders have little impact on growth.
If politicians of various persuasions can't agree on the content of tonight's job speech (much less its scheduling), they can at least agree on one thing: It's a big deal. Obama, unsurprisingly, thinks his speech matters. So do congressional Democrats, who are urging him to "be bold," and congressional Republicans, who are already starting to spin the post-mortems, not to mention plan job speeches of their own. And then you have the American people -- six in ten of whom "now disapprove of Obama's work on jobs and the economy" -- who eagerly await a fresh round of plans and promises.
A great presidential orator delivering a major policy address is an occasion to expect fireworks. But how much does presidential leadership on economic issues actually matter?
It's an old question with old and competing theories to answer it. To pick two pretentious examples from the 19th century: There is, at one extreme, the Marxist theory of historical materialism, in which people are merely the swept along by underlying forces that are "independent of their will." (The obligatory dinner-party Marx quote: "It is not the consciousness of men that determines their existence, but their social existence that determines their consciousness.") From the same century but the opposite corner, there is Scottish essayist Thomas Carlyle's "Great Man" theory of history, for which "The history of the world is but the biography of great men."
Two centuries and many studies later, the numbers tell us that, despite the fuss, there is little reason to suggest that democratic leaders can do much to affect their economies.
THE WOOLY PULPIT
Answering this question convincingly is surprisingly hard, for at least two reasons. First, political and economic variables affect each other, which means that the president's critics and boosters will almost always be able to offer competing narratives that simply switch the order of causation. (Did the stalled economy get Obama elected, or did Obama's election keep the economy stalled?) Second, elections and economics are often affected by the same things: Foreign events, public sentiment, technological innovation -- the list of difficult-to-disentangle relationships is long.
The difficulty with trying to solve these problems is that the relationship in question -- can leaders affect the economy? -- is not amenable to anything like laboratory experimentation. To the eternal disappointment of quantitative social scientists, the UN can't wipe out world governments and randomly assign political institutions to study the results.
But in a 2005 paper (working version here), economists Benjamin Jones and Benjamin Olken offer a solution. They look at how cases in which leaders die unexpectedly and accidentally affect economic growth. Jones and Olken's basic intuition is that these cases offer something as good as the random assignment of a lab experiment. They are political transitions that are unlikely to be affected by underlying economic conditions. (Or, in the authors' charmingly grim euphemism, sudden death is an ideal "source of exogenous variation in leadership.") Their paper gathers all such cases since World War II -- all the plane crashes, the brain tumors, the horseback-riding accidents -- and asks whether or not the change in leadership leads to a change in economic growth.
So what do Jones and Olken find? In autocratic regimes -- places where leaders have few constraints -- they find that accidental changes in leadership have a substantial effect on growth. But not so in democracies. "Changes in leaders in democracies appear to have no effect on economic growth," they conclude.
These results might seem counterintuitive, but there are two good reasons to take them seriously. First, post-World War II democracies have strong and varied institutions. (There are, notably, few if any modern-era examples of relatively wealthy democracies imploding.) Strong institutions are composed of individuals, but they weaken the influence of each individual. You know the theory as balance of powers, or "checks and balances." Second, a sudden change in leadership doesn't necessarily change the preferences of voters -- and good luck to any politician who falls out of favor with the majority.
POLICY STILL MATTERS
Some hedging is in order. First, the Jones and Olken dataset is small: 57 exogenous variations (to borrow their euphemism) since World War II, and far fewer that take place in democracies. Second, the paper focuses only on economic growth, not employment, which is the subject of tonight's speech. The two are related, but as Americans know too well, growth and jobs don't always go together.
Finally, and this is key, the point that democratic leaders don't affect growth should not be confused with the claim that policy doesn't affect growth. You can be a realist about the president's influence without being a pessimist about the government's influence. Nor is it a comment on the relative merits of monetary policy and fiscal policy.
Rather, this is a plea for calibrating expectations appropriately. Buffeted on one side by powerful entrenched institutions (party discipline, balance of powers) and on the other by powerful electoral constraints (the ol' median voter theory), we shouldn't expect so much from the address or the man behind the podium.
Still, I'll watch the speech.>