Upheaval: Turning Points for Nations in Crisis

by Jared Diamond

  All of these developments mean that historical geopolitical constraints have been greatly weakened. Does that mean that geography is now irrelevant? Of course not! Finland’s foreign policy is still dictated by its long land border with Russia. Germany’s foreign policy is still dictated by its nine land neighbors, and by the eight other nations that it faces across the Baltic and North Seas. Chile’s deserts and high mountains ensured that it has never been invaded in the two centuries since its independence; it is unlikely to be invaded in the foreseeable future. The U.S. could be hit by missiles but remains prohibitively difficult to invade and conquer, and Australia nearly as difficult. In short, Finland’s motto “Our geography will never change” continues to apply to every country.

  That summarizes what we’ve learned about the question that initially motivated this book: the relevance of the dozen factors suggested by outcomes of individual crises to the outcomes of national crises. Let’s now consider two questions that were not my initial motive for this study, but that have proved to be the questions that people most often ask when we get into a conversation about national crises. Those two questions concern the role of crises as a driver of national policy change, and the role of leaders.

  Do countries require a crisis to motivate them to act, or do nations ever act in anticipation of problems? The crises discussed in this book illustrate both types of responses to this frequently asked question.

  Meiji Japan avoided dealing with the growing danger from the West, until forced into responding to Perry’s visit. From the Meiji Restoration of 1868 onwards, however, Japan did not require any further external shocks to motivate it to embark on its crash program of change: Japan instead changed in anticipation of the risk of further pressure from the West.

  Similarly, Finland ignored Soviet concerns until it was forced to pay attention by the Soviet attack of 1939. But from 1944 onwards, the Finns did not require any further Soviet attacks to galvanize them: instead, their foreign policy aimed at constantly anticipating and forestalling Soviet pressure.

  In Chile, Allende’s policies were in response to Chile’s chronic polarization, and not in response to a sudden crisis, so Allende was anticipating future problems as well as addressing current ones. In contrast, the Chilean military launched their coup in response to what the military perceived as the acute crisis provoked by Allende’s declared intention to turn Chile into a Marxist state.

  In Indonesia, both types of responses were at play. Communist-sympathetic elements in the Indonesian military launched their coup in anticipation of actions that they feared from an anti-communist Council of Generals. The rest of the Indonesian military apparently reacted in response to the crisis of the October 1, 1965 coup, but there are reasons to suspect that the military had anticipated that coup and had already prepared their response.

  Post-war Germany offers two of modern history’s outstanding examples of nations acting in anticipation of, rather than responding to crises. Chancellor Konrad Adenauer’s program of setting up the European Coal and Steel Community, then setting up economic and political structures leading to the European Common Market and the European Union, was adopted explicitly to anticipate a crisis and to prevent it from ever happening (Chapter 11). After the horrors of World War Two, Adenauer and other European leaders sought to avoid a World War Three, by integrating Western Europe so that Western European countries wouldn’t want to and couldn’t attack one another. Similarly, Willy Brandt’s Ostpolitik was not launched in response to an immediate crisis in Eastern Europe (Chapter 6). Brandt faced no urgent need to recognize East Germany or other Eastern European communist governments, nor to acknowledge Germany’s loss of its eastern territories. Instead, Brandt did so in order to anticipate opportunities in the distant future, and to create stable conditions for the re-unification of Germany whenever that became possible—as finally proved to be the case.

  Japan today is struggling with its seven major problems, without taking decisive action to address any of them. Will Japan succeed in solving these problems by slow change as did post-war Australia, or will it require a sudden crisis to motivate Japan to act vigorously? Similarly, the U.S. today is not taking decisive action in response to our big problems, except for our quick response to the World Trade Center attack by invading Afghanistan, and our response to the supposed presence of weapons of mass destruction in Iraq by invading Iraq.

  Thus, the governments in four of the cases discussed in this book required crises to galvanize them into action, and in two cases today are not taking decisive action in the absence of galvanizing crises. Once crises struck, however, Meiji Japan, Finland, Chile, and Indonesia all set out on programs of change that required years or decades, without requiring further crises to keep motivating them. But our nations do provide examples of pre-emptive actions to prevent crises from materializing (Indonesia and Germany) or getting worse (Chile). Of course, all governments are constantly taking actions for the future to deal with less urgent current or anticipated problems.

  Hence the answer to the question “Does it require a crisis to galvanize a nation into adopting major selective change?” is similar to the answer for individuals. We as individuals act constantly to deal with current or anticipated problems. Occasionally, we foresee a big new problem coming at us, and we try to head it off before it hits us. But, for nations as for individuals, there are much inertia and resistance to overcome. Something big and bad suddenly happening motivates us more than do slowly developing problems, and also more than the prospect of something big and bad happening in the future. I’m reminded of Samuel Johnson’s saying: “Depend upon it, sir, when a man knows he is to be hanged in a fortnight, it concentrates his mind wonderfully.”

  Do leaders make a difference? The other question that people often raise when they and I fall into conversation about national crises concerns the long-running historical debate as to whether national leaders have a significant effect on history, or whether history would have unfolded in the same way regardless of who was a country’s leader at a particular time. At the one extreme is the so-called “Great-Man” view of the British historian Thomas Carlyle (1795–1881), who asserted that history is dominated by the deeds of great men, such as Oliver Cromwell and Frederick the Great. Similar views are still common today among military historians, who tend to emphasize the decisions of generals and wartime political leaders. At the opposite extreme was the author Leo Tolstoy, who maintained that leaders and generals had minimal influence on the course of history. To make his point, Tolstoy included in his novel War and Peace fictitious accounts of battles in which generals issue orders, but the orders are irrelevant to what is actually happening on the battlefield.

  That view, that history’s course depends upon lots of details rather than on policies or decisions of great men, is now common among historians. They often argue that a leader appears to be influential only because he (or she) pursues policies resonating with views already held by his or her countrymen; that otherwise unimpressive politicians may appear to become great because of the opportunities that they enjoy at the time, not because of their personal qualities (examples often suggested are the American presidents James Polk and Harry Truman); and that leaders can only choose from a limited set of options determined by other factors of history. A view intermediate between the Great-Man view and the leaders-don’t-matter view is exemplified by the German sociologist Max Weber (1846–1920), who maintained that certain types of leaders, so-called charismatic leaders, could sometimes influence history under some circumstances.

  This debate remains unresolved. Each historian tends to hold some a priori general view based on principle rather than on some valid method for assessing empirical evidence, and to apply that view to individual case studies. For example, all biographies of Hitler have to recount the same key events of his life. But proponents of the Great-Man view relate those events while asserting that Hitler was an unusually effective and evil leader who caused developments in Germany to turn
out differently from how they would have turned out under some other leader. Opponents of the Great-Man view relate those same events while portraying Hitler as a voice reflecting widespread features of German society at the time. The debate is impossible to resolve by narratives and individual case studies.

  Instead, a promising approach comes from recent analyses that combine three characteristics: a large sample of many historical events, or of all historical events of some defined type; using “natural experiments of history,” i.e., comparing otherwise-similar historical trajectories in which a certain perturbation did or did not occur (I’ll give two examples in the following paragraphs); and measuring the outcomes quantitatively. Two outstanding such papers have been published by Benjamin Jones at Northwestern University and Benjamin Olken of the Massachusetts Institute of Technology.

  In their first paper, Jones and Olken ask: what happens to the national economic growth rate when a leader dies in office from natural causes, as compared to what happens at randomly selected times when a leader doesn’t die in office from natural causes? This comparison offers a natural experiment to test the effect of a change in leadership. If the Great-Man view is correct, then a leader’s death should be more likely to be followed by changes of economic growth rates—either decreasing or increasing, depending on whether the leader’s policies really made a difference by being good or bad, respectively—than after random moments when a leader doesn’t happen to die. For their database, Jones and Olken took every instance in the world of a national leader dying from natural causes while in office between 1945 and 2000. They succeeded in assembling 57 such cases: mostly deaths due to heart attacks or cancer, plus a few plane crashes, a drowning, a fall from a horse, a fire, and a broken leg. Those events really do constitute a random perturbation: a leader’s economic policies don’t affect the likelihood that that leader will accidentally drown. It turned out that economic growth rates were much more likely to change following a leader’s natural death than following random moments when a leader didn’t die. That suggests that, averaged over many cases, leadership does tend to affect economic growth.

  In their second paper, Jones and Olken ask: what happens when a leader is assassinated, instead of dying of natural causes? Of course, assassinations are not at all random events: they are more likely to be attempted under some conditions (e.g., if citizens are dissatisfied with low economic growth) than under other conditions. Hence Jones and Olken compared successful assassination attempts with unsuccessful attempts, when the bullet missed. That really is a random difference: national political conditions may have influenced the frequency of assassination attempts but don’t affect the assassin’s aim. The database consisted of all 298 assassination attempts on national leaders from 1875 to 2005: 59 of them successful, 239 unsuccessful. It turned out that successful attempts were more likely than unsuccessful attempts to be followed by a change in national political institutions.

  In both studies the effect of a leader’s death was stronger for deaths of autocratic leaders than for those of democratic leaders—and stronger for autocrats with no constraints on their power than for autocrats constrained by legislatures or by political parties. That’s as we’d expect: strong leaders with unlimited power can have more effect (whether for good or for bad) than leaders with only limited power. Thus, these studies agree on a general conclusion: leaders sometimes make a difference. But it depends on the type of leader, and on the type of effect examined.

  Let’s now tie these natural experiments on the roles of leaders to the seven countries discussed in this book. My goal is to see whether our leaders fit the patterns recognized by Jones and Olken, and what further questions they pose for testing. The histories of our seven countries have suggested the following appraisals of their leadership to many historians:

  In Meiji Japan, no single leader was dominant; several leaders shared similar policies.

  In Finland, political leaders and citizens were virtually unanimous in their agreement that Finland should do its utmost to resist Soviet attack. (But it’s sometimes suggested that Field Marshal Mannerheim’s skills as military commander, and President Paasikivi’s and President Kekkonen’s abilities to win the trust of Soviet leaders after the war, affected Finland’s fate positively.)

  In Chile, Pinochet was considered (even by his fellow generals) decisive and unusual in his cruelty, his tenacity at holding on to power, and his choice of economic policy.

  In Indonesia, Sukarno and Suharto are both considered decisive leaders, but subsequent presidents are not.

  In post-war Germany, Willy Brandt is often suggested as having played a unique role in reversing previous West German government foreign policy, recognizing Eastern European communist governments and German frontiers, and thereby making possible Germany’s subsequent re-unification. In earlier German history, Bismarck, Emperor Wilhelm II, and Hitler are regularly cited as examples of unique leaders who made a difference for better or for worse.

  In Australia there has been no single clearly dominant leader. The closest possible example is Prime Minister Gough Whitlam and his crash program of change, but Whitlam himself acknowledged that his reforms were a “recognition of what has already happened.”

  In the United States, President Franklin Roosevelt is credited with gradually preparing the U.S. for World War Two against the will of American isolationists (who may initially have constituted a majority of Americans), and for his efforts to pull the U.S. out of the Great Depression. In 19th-century American history, President Lincoln is considered to have played a unique role in the course of the Civil War.

  In short, our seven countries offer examples of nine leaders (six autocratic, three democratic) often rated as having made a difference. In addition, in countries other than the seven discussed in this book, the leaders most often argued to have made a difference in modern times include Winston Churchill in the United Kingdom, Lenin and Stalin in the Soviet Union, Mao in China, de Gaulle in France, Cavour in Italy, and Gandhi in India. Thus, we have a short list of 16 leaders commonly viewed as having made a difference. Of the 16, 11 were in autocratic regimes, five in democracies. At first sight, this outcome seems to conform to the conclusions of Jones and Olken about the greater effect of leaders in autocracies. But I haven’t tabulated relative numbers of all autocratic and democratic leaders worldwide over this time span, so I can’t say which, if either, type of leader is disproportionately represented.

  Our small dataset does suggest two hypotheses worth testing by methods similar to those of Jones and Olken: by assembling a large dataset comprising a natural experiment, and measuring outcomes quantitatively.

  One hypothesis stems from the observation that, of the four democratic leaders most often suggested as having been uniquely influential (Roosevelt, Lincoln, Churchill, and de Gaulle), at least three had their effects or their greatest effects in wartime. Almost all of Lincoln’s presidency took place during the American Civil War. Churchill, Roosevelt, and de Gaulle served both in war and in peace, but two or all three are viewed as having had their most decisive effects in wartime (Churchill as wartime prime minister from 1940 to 1945 but not as peace prime minister from 1951 to 1955; de Gaulle as wartime general, then as president during the Algerian uprising of 1959–1962; and Roosevelt after the outbreak of World War Two in Europe in 1939 but also during the Depression). These outcomes fit the observation of Jones and Olken that leaders have more decisive influence, the fewer constraints on their power: democratic leaders exercise more concentrated powers in wartime.

  The other hypothesis that our results suggest for testing is that leaders make the most difference under circumstances where they face strong opposition (whether in democracies or autocracies) from people espousing a very different policy, and where the leaders nevertheless eventually get their views to prevail, usually by cautious step-by-step efforts. The examples are: Piedmont’s Prime Minister Cavour and Prussia’s Chancellor Bismarck slowly achieving the unifications of Italy and of
Germany respectively over strong opposition from foreign powers, from other Italians or Germans respectively, and even from their own kings; Churchill convincing an initially closely divided British war cabinet to reject Lord Halifax’s proposal to seek a negotiated peace with Hitler, then persuading Americans to make their first priority the war against Germany rather than the war against Japan (initially the obvious priority for America after the Japanese attack on Pearl Harbor); Roosevelt slowly preparing the U.S. for World War Two over the opposition of American isolationists; de Gaulle slowly convincing both his countrymen and Algerians to reach a negotiated settlement of the Algerian struggle for independence; Suharto’s slowly pushing aside Indonesia’s beloved founding President Sukarno; and Willy Brandt persuading West Germans to swallow the bitter pill of renouncing much former German territory, over the fierce opposition of the CDU Party, which had until then ruled West Germany uninterruptedly for two decades.

  This book has been an initial step in a program of comparative studies of national crises—an exploration of a small sample of nations, investigated by narrative methods. How can this study be extended so as to deepen our understanding? I suggest two extensions: a larger and more random sample, and a more rigorous analysis translating outcomes and hypothesized predictors from verbal concepts into operationalized variables.

  First, the sample. My sample of nations is not only small, but also selected non-randomly. I selected these countries not because they offer a random subset of the world’s 216 nations, but because they are drawn from the countries that I know best. As a result, they consist of two European nations, two Asian nations, one each from North America and South America, and Australia. Five of the seven are wealthy. All seven are currently democracies, although two were dictatorships during the period that I discuss. All except Indonesia have long histories of independence or (Finland) autonomy, and of strong institutions. Only one recently emerged from colonialism to independence. Missing are any African nations, any current dictatorship, and any very poor nation. All six for which I discussed past crises survived their crisis with some degree of success. None illustrates an unambiguous failure to respond to a crisis by means of appropriate selective changes. That’s obviously a non-random sample. Hence it remains a challenge for the future to see what conclusions a broader sample of nations will reveal.