State capacity and military conflict

There is a new paper (pdf) by Nicola Gennaioli and Hans-Joachim Voth, forthcoming in The Review of Economic Studies:

Powerful, centralized states controlling a large share of national income only begin to appear in Europe after 1500. We build a model that explains their emergence in response to the increasing importance of money for military success. When fiscal resources are not crucial for winning wars, the threat of external conflict stifles state building. As finance becomes critical, internally cohesive states invest in state capacity while divided states rationally drop out of the competition, causing divergence. We emphasize the role of the “Military Revolution”, a sequence of technological innovations that transformed armed conflict. Using data from 374 battles, we investigate empirically both the importance of money for military success and patterns of state building in early modern Europe. The evidence is consistent with the predictions of our model.

The pointer is from Mark Koyama.


I'm not sure that Poland really 'rationally' dropped out of the competition rather than being chewed up by the military capacity that their neighbours had been investing in

Poland was simply dismembered on several occasions.

If anyone rationally dropped out of the race, it would be Sweden.

A (no doubt overly) simple historical model for the authors' question of why the west end of the Eurasian landmass came to dominate the world by 1900 might start with the existence of England on a large, but not huge offshore island. The English heartland is big enough that it could defend the island from outside invasion, but only if most of the population of Albion were unified and reasonably competently governed. In turn, a unified England that could defend itself could also do enough buccaneering on the Continent to stimulate the unification of France. In turn, the lessons taught by the England-France rivalry set dominoes to toppling across the world.

In short, the particular geography of England set the template, going back about a millennium, for how big a European Great Power would be: big enough to not be conquered by England. But, in turn, England would diplomatically maneuver to break up agglomerations of power significantly bigger than itself (until finally Great Powers were succeeded by Superpowers in 1945).

'powerful, centralized states controlling a large share of national income only begin to appear in Europe after 1500'

The pharoahs are laughing again.

And assuming the above cited sentence refers to such states only appearing in Europe after 1500, more than a few caesars are wondering why their highly centralized empire didn't make the grade.

'Using data from 374 battles, we investigate empirically both the importance of money for military success and patterns of state building in early modern Europe.'

In other words, we talk about a broad subject, then limit it in such a way that we can ignore the rise of the United States or Japan. And I would love to see how that analysis works with dealing with 'Germany' - the Prussians were poor and fairly backward in the eyes of most German states, and yet, it seems as if dedicated militarism works even better than access to capital when it comes to 'state building.'

The pharaohs weren't in Europe and the Caesars didn't get to tax their rich subjects' holdings. Don't they teach European geography in "Germany is Awesome Integrazionklasse"?

'The pharaohs weren’t in Europe and the Caesars didn’t get to tax their rich subjects’ holdings.'

Well, Europe took a couple of thousand years to reach the level of Egypt. As for the Caesars - are you honestly trying to argue that the Octavian Caesars felt any hesitation at doing whatever they wanted?

And my education? I'm American, so blame W. T. Woodson and GMU. At least the Germans know how to properly pronounce the name the first Roman emperor used (admittedly, my GMU Latin class in 1983 or so was open about this).

Tell us about the pharaohs' colonies in India and North America. Those pyramids they built in China are pretty impressive, too.

"the Octavian Caesars"? Octavian (singular) couldn't just do whatever he wanted. Neither could the rest of the Roman emperors.

Money was key to military success well before 1500.

The Plantagenets were shaking down the barons, peasants and London merchants for centuries to fund wars both at home and in France.

Italian bankers, the Templers and Jews were very familiar with the need of kings for money, whether borrowed or stolen.

Money was key to military success well before 1500. -

The thing that changed was the use nationalism allowed bankers to start much much more profitable wars than they could ever hoodwink the masses into beforehand.

Certain groups of bankers who lend to governments started to specialize more and became experts at lending to BOTH sides of a war and getting more deeply involved in the internal policies of countries to instigate new profitable wars.

The paper's empirics are a mess. The evidence for the claim that "fiscal resources were not crucial for winning wars before 1650" is not well supported. The authors divide up the sample of wars to "pre-1650" and "post-1650" groups and run separate regressions on victory in war. The estimates for fiscal resources are pretty close (0.45 and 0.62) and given the t-statistics, they definitely overlap. So what do the authors base their conclusion on? Fiscal resources is not statistically significant in the pre-1650 sample, but it is significant (at 10% level) in the post-1650 sample. But the first sample is much smaller than the second (N = 38 vs 215), which of course prevents a precise estimate. As Andrew Gelman says, "the difference between significant and insignificant itself is not significant" in this case.

The proper way to do this is to interact fiscal resource variable with a period dummy (=0 for pre-1650, =1 for post-1650) and then look at the sign and significance of that interaction variable. It is clear from this table that if done properly, that analysis would not support the claim that fiscal resources mattered more for war outcomes after 1650.

Finally, the footnote under the table says "standard errors are in parentheses", but you can tell that *t-stats* are given in parentheses, not p-values.

Lost me here: "the increasing importance of money for military success"


Wealth has always been hugely important to military success and money has been since money was invented. The stamped coin first reached Europe through Lydian payment of Greek soldiers' wages.

Well, you should take into account Vikings or Mongols, who were not paid in cash as soldiers usually are. And they were quite successful. After the end of the Roman empire, there probably was a period in Europe when the vast majority of armed fighters did not get any money for their service; a share of plunder, yes, but not money.

The 14th - 15th centuries were probably some of the last in which steppe peoples were able to conquer settled peoples. Tamerlane died in 1405; Muscovy expelled the Tatars in 1480. The Mughals invaded India in the early 1500s.

For most of human history, sufficiently motivated and united steppe peoples were incredibly difficult for wealthy settled peoples to resist. At some point in this timeframe, this ceased to be the case.

Among settled peoples, wealth was militarily valuable. But an element of the medieval feudal relationship was the expectation that a lord's vassals would campaign with him at their own expense. Thus, the ability to call upon the forces of vassals was a power (albeit an unreliable one) that was not synonymous with a lord's personal wealth, though it was probably related to it.

Improvements in fortifications probably helped diminish the success of Steppe invasions.

It's one thing to raid a village, it's quite another to take out a castle defended by skilled crossbowmen.

The Khan had excellent siege engines, so that was less of a concern. Of course, the steppe people needed allies / vassals who had excellent siege engines - that initial conquest was rather important to keep things going. And then after 1200, the steppe itself became rather civilized; traditionally the steppe was a path of continuing migration; of one inner steppe tribe pushing out the border tribe (into civilization). The Khan outright conquered the entire steppe before really turning on China and other areas. It changed the entire steppe dynamic, and the fruits of Baghdad, Warsaw, and Imperial China were known to the inner steppe (and corrupted it to some extent).
By the time the Khanates were pushed out, it was gunpowder, not horses and bows, that controlled the battlefield. Remains of the mongol empire persisted until the late 1800s, after all.

It seems to have been gunpowder, and the resulting inevitability of victorious armies suffering substantial losses, that ended the military effectiveness of steppe warriors rather than fortifications.

Manchu conquered China in the 17th century.

In my military history class, it was explained that the advantage conferred by capital-intensive offensive weapons--i.e. cannons--led to (or at least was a factor) of increased centralization of power in Europe. I don't know how widespread this viewpoint is, but I suspect that the authors are trying to prove this viewpoint in quantitative terms. I suspect that there's a problem here in conflating "money" and "resources".

Given the military advantages enjoyed by large, centralized states (more soldiers, the possibility if not always the practical ability to supply a large army in one region using supplies from another) prior to the development of gunpowder, this explanation is clearly not general. The idea that war is expensive predates 1500 by several thousand years. I suspect that a more reasonable explanation is along the lines of: inefficient (i.e. feudal) European governments (kings) were capable of raising armies at least good enough to fight each other relying heavily on non-monetary resources like land and social connections. These resources were scalable i.e. X acres produced Y soldiers. Sure, the Baron of Dungford might not be able to bring as many men to the field as the Duke of Fancypants, but he could make a contribution directly proportionate to his holdings. This could, in theory, have been accomplished by larger states but decentralized feudal "states" (originating from civilizational collapse) in practice were good enough and social/cultural factors prevented centralization in the absence of more compelling motives.

Once siege cannons became available though, states of a certain size wielded a difference in military capability not merely in degree but in kind. The king didn't need the baron's men; he needed his money so that he could pay teamsters to move the ore from the mine to the foundry. The military advantage conferred by this system meant that kings who couldn't consolidate their power were conquered by those who could.

But this explanation, to the extent that it's true at all, probably isn't general even to Europe. It might explain the rise of Austria or France.

The new cannons could knock down the old tall castles owned by hereditary local aristocrats, so they needed to either submit to somebody who could afford cannons and new low fortresses or get rich themselves.

Wasn't this all covered in depth by Paul Kennedy already?

Comments for this post are closed