The great leveler, p.28
The Great Leveler, page 28
”WE NO LONGER COUNTED UP WHAT WE HAD KILLED BUT WHAT IT WOULD BRING US”: CIVIL WAR
This leaves us with a final question: how does civil war affect inequality? Modern scholarship has generally focused on the reverse—whether inequality contributes to the outbreak of internal conflicts. There is no straightforward answer to that second question. Overall (or “vertical”) income inequality—between people or households within a given country—is not positively correlated with the likelihood of civil war, although the poor quality of data from many developing countries raises doubts about the reliability of any specific finding. Intergroup inequality, on the other hand, can be shown to promote internal conflict. Some recent work has complicated the picture. A wide-ranging survey of human body height inequality, used as a proxy of resource inequality, claims a positive correlation with civil war in a large global dataset that reaches back to the early nineteenth century. And according to another study, the probability of civil war increases with land inequality unless the latter is extremely high, in which case the former declines as small elites are better able to repress resistance. For now, all we can say is that the considerable complexities of this question are only beginning to be understood.41
By contrast, the impact of civil war on inequality has attracted very little attention. A pioneering study of 128 countries from 1960 to 2004 found that civil war increased inequality, particularly during the first five years after a conflict. On average, the income Gini coefficient rose 1.6 percentage points in countries during civil war and 2.1 percentage points during the recovery phase of the following ten years, peaking about five years after the end of the war if peace was maintained. There are several reasons for this trend. Inasmuch as civil war reduces physical and human capital, its value goes up while that of unskilled labor declines. More specifically, in developing countries with extensive agricultural populations, farmers may lose access to markets and suffer income losses through exclusion from commercial exchange, losses that provide an incentive to downshift to subsistence practices. At the same time, war profiteers reap great profits by exploiting diminished security and the weakening or absence of state power. Profiteering tends to benefit a small minority, allowing it to accumulate resources at a time when the state’s capacity to collect taxes has abated. This retrenchment, in conjunction with increased military expenditure, also curtails social spending, which in turn harms the poor. Redistributive measures, schooling, and health care suffer, with negative effects that are the stronger the longer the conflict lasts.42
These problems persist beyond the war proper, accounting for the even higher Gini coefficients observed in the immediate aftermath of civil wars. In that period, winners may reap disproportionate rewards from their victory as “personal and patrimonial links determine the distribution of assets and access to economic gains.” Civil war shares this feature with traditional premodern wars in which the leadership on the winning side stands to gain and inequality grows. The same could be observed in the nineteenth century, when civil war confiscations of land in Spain and Portugal in the 1830s boosted large estates and exacerbated inequality.43
Almost all of the relevant observations come from traditional societies or developing countries. Full-blown civil wars have been exceedingly rare in more developed economies. Moreover, in some cases in which civil war was associated with major leveling, as in Russia after 1917 or in China from the 1930s and 1940s, revolutionary reforms rather than civil war as such were the main driver of this process. For the purposes of this study, the American Civil War has been treated as the equivalent of interstate war, with the results described earlier in this chapter. This leaves us with only a single major case—that of the Spanish Civil War from 1936 to 1939. Unlike in Russia or China, the victorious party did not pursue a redistributive agenda, and the outcome of the war was not revolutionary in any meaningful sense of the term. Collectivization in areas under anarchist control during the conflict were short-lived. In the years after 1939, the Francoist regime implemented a policy of economic autarky which caused economic stagnation. A succession of civil war shocks and subsequent economic mismanagement accounted for the decline of top income shares. For this period only the uppermost income shares (for the richest 0.01 percent) have been computed, a category that experienced a drop by 60 percent between 1935 and 1951. This trend conflicts with the development of the Gini coefficient of overall income, which was fairly stable during the civil war and the World War II period but which shows wild fluctuations between 1947 and 1958 (Fig. 6.2).44
To complicate matters further, the Gini coefficient of wage income dropped considerably, by about a third, between 1935 and 1945. To the best of my knowledge, there is currently no compelling explanation for these outcomes. Leandro Prados de la Escosura has offered hypotheses regarding the competing effects of declining returns to capital (which depressed top income shares), wage compression from reruralization under Franco (with reduced overall wage inequality), and rising returns to property, especially land, under autarky (which offset these effects to produce the overall Gini coefficient of income inequality). All this took place in the context of zero net real per capita GDP growth from 1930 to 1952, with the proportion of the population living in poverty more than doubling during roughly the same period. Despite superficial similarities in terms of the fall of top income shares and the compression of wages, inequality developed rather differently in Spain than in other European countries of the time. Unlike among the parties to World War II and some of the bystanders, there was no progressive taxation and overall income inequality did not diminish. I agree with Prados de la Escosura that “the distinction between Spain, where the Civil War had a divisive effect in society, and most western countries, where the World Wars tended to increase social cohesion, may be relevant to understanding the post-war era.” Even so, in both cases, the underlying driving forces that shaped the distribution of income and wealth were the same: violent shocks that were mediated by government policies.45
Figure 6.2Gini coefficients of income and top 0.01 percent income share in Spain, 1929–2014
I conclude my survey by going once again far back into the past to consider a hybrid case, the civil wars that brought down the Roman Republic from the 80s to the 30s BCE. They are hybrid in that they were internal conflicts within Roman society, triggered by runaway elite competition, but unfolded in the context of the aforementioned culture of military mass mobilization and thus displayed key features of interstate mass mobilization warfare. Some of Rome’s highest military participation rates on record were achieved in this period of domestic upheaval. This specific combination of elite infighting and popular mobilization offered novel opportunities for the redistribution of income and wealth.
The most violent of these conflicts—fought in the 80s and in the 40s and 30s BCE—devastated the Roman ruling class. Political opponents were proscribed—publicly declared to be fair game for anyone wishing to kill them for a reward—and their estates confiscated by the victorious faction. In the civil war of 83 to 81 BCE, 105 senators were said to have been killed, at a time when senate membership stood at around 300, and in 43 BCE, 300 senators (out of 600) and 2,000 knights, the next highest tier of the Roman elite, supposedly lost their lives in this manner, although we can only name about 120 of them. These two episodes affected inequality in different ways. The first round of expropriations, undertaken by proponents of oligarchic reaction, allowed well-placed supporters to profit by snapping up seized assets at auction. This may very well have increased wealth concentration, especially in the wake of massive attrition prior to the civil war: in the decade from 90 to 80 BCE, no fewer than 291 senators are said to have died of violent causes. A shortage of heirs is likely to have led to consolidation rather than dispersal of elite property. Land confiscated from local communities was given to veterans but often eventually ended up on the market, resulting in transactions that may also have fostered concentration. By contrast, the expropriations of 43 and 42 BCE were motivated by exceptional fiscal demand in preparation for military campaigning against domestic opponents outside Italy rather than by a desire to settle scores. In this case the proceeds were less likely to benefit cronies and were primarily used to meet exorbitant promises of compensation to a huge citizen army. Close associates of that faction’s leader were rewarded only after the conclusion of the conflict in 30 BCE via disbursements that enriched “new men” at the expense of the established nobility.46
Levels of troop compensation in this final round of civil wars probably had significant redistributive consequences. Prior to the onset of civil wars, Roman soldiers had been rather modestly compensated. Incipient warlordism pushed up bonuses in military campaigns against foreign enemies: they rose from very low levels to the equivalent of seven times annual base pay in 69 BCE and to thirteen times in 61 BCE. Civil war in the 40s BCE triggered a further and more dramatic surge to twenty-two times the newly increased base stipend (or forty-two times the previous one) in 46 BCE. This outlay was soon exceeded by the promise of the same reward to a much larger number of servicemen four years later. All in all, we can estimate that an amount of money equivalent to at least ten times regular annual state income, or perhaps half a year’s worth of the GDP of the Roman empire at the time, was transferred to soldiers between 69 and 29 BCE—and almost all of that between 46 and 29 BCE—to buy and reward loyalty in civil wars. The total number of recipients may have reached 400,000 men, who together with their families would have accounted for up to a third of the entire Roman citizenry. In the absence of evidence for price inflation, this was likely to boost non-elite incomes in real terms. The distributional effect within Roman society, in the Italian heartland, is more opaque. Much of this money was obtained by extorting resources from overseas provinces. But there were exceptions: in 43 BCE, a levy of one year’s worth of income from real estate and a 2 percent wealth tax were imposed on the rich, alongside the massive confiscations that have already been mentioned. Several later levies also specifically targeted the wealthy. For the only time in Roman history, fiscal extraction became effectively progressive and the revenue thus generated was used in a redistributive manner.47
This, however, was to remain a one-off anomaly. Reliance on provincial revenue again became the norm once peace was restored and stable autocracy introduced after 30 BCE. It was only for a few years, in the late 40s BCE, that the disposable income distribution would have temporarily shifted in favor of the general citizen population. In the longer run, the following centuries of political and economic stability were undoubtedly conducive to high levels of wealth concentration, as we saw in chapter 2.
”WHATEVER THE COST MAY BE”: WAR AND INEQUALITY
This part of the book has taken us through thousands of years of warfare. Military conflict has long been a pervasive feature of human history, but only certain kinds of war attenuated another equally pervasive phenomenon—the unequal distribution of income and wealth. For winners and losers alike, modern mass mobilization warfare proved to be a potent means of leveling. Whenever the war effort permeated all of society, capital assets lost value, and the rich were made to pay a fair share, war did not merely “kill people and break things” but also narrowed the gap between rich and poor. In World War II, this effect was at work both during the war itself as well as in its aftermath, sustained by the persistence of war-driven policies. The citizens of developed countries owed a generation or more of declining inequality to the unprecedented violence of this global conflict. A similar compression of material disparities had occurred in or after World War I. Earlier instances of this particular style of warfare are rare and are not normally associated with leveling. In the American Civil War, it was not mobilization as such but defeat and occupation that destroyed fortunes in the South. Evidence for ancient precursors of mass warfare yields ambiguous or negative findings, as in China and the Roman Republic. In the warrior state of ancient Sparta, resource imbalances gradually grew out of arguably more equalized earlier conditions. Classical Athens may be the best premodern example of the leveling influence of widespread popular military participation. Just as in part of the twentieth century, Athenian democracy appears to have been reinforced by the shared experience of war mobilization and to have in turn favored policies that constrained the rise of inequality. Considering the dramatic differences in overall development and the limitations of the ancient evidence, we should be wary of putting too much weight on this analogy. Even so, the experience of ancient Athens suggests that with the right combination of institutional arrangements in place, a culture of military mass mobilization could serve as a leveling mechanism even in a thoroughly premodern environment.48
Wars that were more limited in scope were ubiquitous throughout history but failed to deliver consistent outcomes. Traditional wars of plunder and conquest commonly benefited victorious elites, boosting inequality. This would have been especially true whenever defeated polities were incorporated into larger ones, a process that would add additional layers of wealth and power at the top of the hierarchy. Civil war rarely functioned as a source of leveling—and if so only partially (as, in different ways, in the 1860s United States and in 1930s and 1940s Spain) or very briefly (as, perhaps, in ancient Rome). The only civil wars that truly transformed the distribution of income and wealth were those that swept to power radical regimes that were bent on comprehensive expropriation and redistribution and that did not shy away from whatever amount of bloodshed was required to make this happen. It is to this process, the Second Horseman of violent leveling, that we now turn.
1 Fig. 6.1 from Scheve and Stasavage 2016: 177 fig. 7.1.
2 The presence of large armies alone does not necessarily meet that criterion: in China in 1850, for instance, close to 9 million people would have had to serve in the military in order to clear a 2 percent threshold. For all we can tell, that did not happen even during the Taiping Rebellion: see herein, chapter 8, pp. 238–240.
3 Bank, Stark, and Thorndike 2008: 23–47 on Civil War, esp. 31–34, 41–42.
4 Turchin 2016a: 83 table 4.4, 139, 161. For the evidence of the census data, see herein, n. 7, and also Soltow 1975: 103. Between 1860 and 1870, the estimated Gini coefficient of property income rose from 0.757 to 0.767, and the top 1 percent share from 25 percent to 26.5 percent—changes that are well within likely margins of error: Lindert and Williamson 2016: 122 table 5–8. Income Ginis went up 6.1 points in New England, 3.1 points in the Middle Atlantic states, 6.7 points in the Eastern Northern Central states, and 5.9 points in the Western Northern Central states, whereas the corresponding top 1 percent income shares grew from 7 percent, 9.1 percent, 7 percent, and 6.9 percent to 10.4 percent, 9.2 percent, 9.1 percent, and 9.7 percent: 116 table 5–7A, 154 table 6–4A.
5 Slaves as wealth: Wright 2006: 60 table 2.4, with 59 table 2.3 (farmland and buildings accounting for 36.7 percent of southern private wealth). Cf. also Piketty 2014: 160–161 figs. 4.10–11 for earlier decades. Ginis: Lindert and Williamson 2016: 38 table 2–4, 116 table 5–7; and cf. also 115 table 5–6 for 1850. Slaveownership: Gray 1933: 530 with Soltow 1975: 134 table 5.3.
6 I am greatly indebted to Joshua Rosenbloom and Brandon Dupont, who very generously computed these results for me from IPUMS-USA, https://usa.ipums.org/usa/. For the nature of these data, see Rosenbloom and Stutes 2008: 147–148.
7 The overall wealth Ginis for all Southerners in the IPUMS-USA data were 0.8 in 1860 and 0.74 in 1870. Earlier studies calculated more moderate leveling during that decade. Soltow 1975: 103 estimates wealth Ginis of 0.845 for free Southerners in 1860 and 0.818 for Southern whites in 1870. Jaworski 2009: 3, 30 table 3, 31, drawing on 6,818 individuals from the entire United States studied in both 1860 and 1870, computes a decline of the wealth Gini from 0.81 to 0.75 in the Atlantic South caused by losses at the top and a rise from 0.79 to 0.82 in the South Central region caused by rapid wealth accumulation by white-collar workers. Rosenbloom and Dupont 2015 analyze wealth mobility in that decade and find considerable turnover at the top of the wealth distribution. Property income: Lindert and Williamson 2016: 122 table 5–8. Table 2.4: 116 table 5–7, 154 table 6–4A (all Ginis rounded to two digits behind the decimal point). For a comparison of free households in 1860 and white households in 1870, see 116 table 5–7, 155 table 6–4B, for top income share reductions by 32 percent, 23 percent, and 49 percent of 1860 levels and Gini reductions by 4, 3, and 8 points.
8 The swift recovery of top income shares in the United States and Japan in the 1920s appear to be only partial exceptions.
9 Quoted from Schütte 2015: 72.
10 Clausewitz 1976: 592.
11 See herein, chapter 8, pp. 232–238.
12 In keeping with the overall focus of this book, I refer here to state-level sedentary societies, leaving aside different types such as small-scale groups that engage in sporadic or seasonal warfare with high participation rates or pastoral steppe populations such as the hordes of Genghis Khan and his successors that drew in much of the adult male population.
13 Kuhn 2009: 50 (Song); Roy 2016: ch. 3 (Mughals); Rankov 2007: 37–58 (higher figures for late antiquity are not credible: Elton 2007: 284–285); Murphey 1999: 35–49 (Ottomans).
