Albert O. Hirschman, Exit, Voice, and Loyalty: Chapter 7, "A Theory of Loyalty"

Question: under what conditions can exit and voice coexist? That is, when will voice still be used even when exit is available?

Answer: when individuals are (1) willing to trade off the certainty of exit for an uncertain prospect of recuperation of the product and (2) confident of their ability to influence the defective firm or institution.

Loyalty -- the attachment to the institution -- comes into play in the first case, but in fact both factors reinforce one another. People who care about an institution will tend to put themselves into a position where they can influence it, and people with influence will feel like they have more at stake.


Loyalty isn't an absolute barrier to exit. It is similar in its effect to a significant transaction cost.


Loyalty is most useful when (1) the use of voice is not straightforward but will require ingenuity and creativity (2) when the deteriorating product has close substitutes. The second case is paradoxical -- it seems irrational not exit when close substitutes are available -- but firms in this condition would have no chance to recuperate were it not for loyalty. Since products, institutions, and social groups are typically unevenly distributed on a scale of quality and prestige, with greater density on the lower end, the second case also means that loyalty is more useful at the lower end of the scale.


In using voice, the loyalist's most effective tool is the threat of exit. So we have the following seeming paradox: ease of exit makes voice less likely, but possibility of exit makes voice more effective. The conclusion which can be drawn from this is that voice will be most likely to be both used and effective when exit is possible but not too easy.

Notes on model of loyalist behavior:

• voice increases with deterioration of quality, and curve bends up at points where there would be exit without loyalty and where there is threat of exit with loyalty
• once loyal customers exit, they will not return until at least the quality associated with exit without loyalty is restored; the demand curve for exit is separate from the demand curve for return

Leaders of organizations and firms want to reduce both exit and voice. Will use high entry fees and high penalties for exit to make exit more difficult and to promote unconscious loyalist behavior. However, high entry cost induced loyalist behavior will ten to suppress the initial use of exit, but to make it more vigorous once it has started. If the cost of exit is high as well, however, the loss of threat of exit will make voice less effective. On the other hand, organizations where exit is difficult or impossible but entry cost is automatic (e.g., family, country) may actually sustain the most vigorous use of voice because members will see it as their due.

A special case of loyalty among influential members of organizations is brought into play under the conditions that (1) their departure would result in a further decline in quality and (2) they would continue to care about the quality even after exit. The first condition presumes that the departure of influential members has the opposite effect of the exit of market makers in monopoly or monopolistic competition; this is possible because the members play a part in the production of the good as well as its consumption. The second condition is rational under the assumption that full exit is impossible, which is the case for public goods. (Examples: Public schools, political parties, government administrations)

In such cases, members may be even less likely to depart as an organization gets worse, because they feel more strongly that is their responsibility to stick around to prevent things from getting yet worse. (With tongue only halfway in cheek, Hirschman uses the term spinelessness for this behavior). On the other hand, a member who does decide to exit under these conditions is more likely to use their exit as a tool of protest that will initiate continued use of voice from outside of the organization. Hirschman laments, however, that this use of exit by disgruntled public officials has fallen into disuse, replaced by officials treating their exit as a private matter -- one thinks of the stock "spending more time with my family" excuse.

Peter Heather, The Fall of the Roman Empire: Chapter 10, "The Fall of Rome"

Heather begins by distinguishing local from central Romanness. Central Romanness -- a central court, bureaucracy, and system of law that responded to the interests of the landowning class, and a professional army to protect this order -- died in the 470s. With its demise the great centripetal force drawing the peoples of the west to Roman cultural identity disappeared. But local Romanness -- the cultural and ideological attachment to being Roman -- was slow a-dying. For more than a century after the end of the empire in the West, there was a significant section of elite population who were culturally and even legally Roman.

Heather takes a synoptic view of the three stages of the fall of central Romanness -- 1) the invasions of the late 4th and early 5th centuries; (2) progressively more debilitating carving out of domains within the empire by barbarian groups in the first three-quarters of the 5th century; (3) the final seizure of fully independent kingdoms by the barbarians when the empire was no longer able to stand up to them. He argues that the Huns drove the initial invasions, that their presence helped prevent further invasions in the 5th century and helped the Romans to control the previous invaders, and that the Hunnic collapse threw the balance of power decisively in favor of the barbarians.

The end of the Roman state did take away the incentive for local elites to maintain classical literary education, local civic life, and the other aspects of local Romanness. Military service rather than Romanness became the path to getting ahead.

Although Heather's rejects Gibbons' thesis that internal factors alone were the cause of Rome's fall -- pointing to the survival of the Eastern Empire as a decisive refutation -- he does acknowledge that the military, economic, and political limits of The Roman Empire interacted in decisive ways with the external invaders who brought down the Empire. Militarily, the Romans ability to contain the barbarians was limited because the Sasanian front required a quarter of the Empire's armed force. This military force could not simply be further expanded because there was no way to generate more revenue -- agricultural production was already at its maximum. Politically, the empire was brittle in the face of barbarian invasion because local landowning elites -- the bulwark of imperial support -- had to swing their support behind any new power in their region to retain their property. Moreover, the Empire was beset by demands from a vast breadth of local elites, and to satisfy these demands it had to resort divided imperial rule. It was never possible to really settle the division of power or the management of succession, however, so this dual system was subject to prolonged bouts of instability during which the barbarians were able to press their advantage.

The external factor in the fall of Rome -- the powerful groups of invading barbarians -- depended on integration of Germanic tribes into larger coalitions. This process was driven by fear and opportunity: fear of the power of the Roman state against isolated opponents and the opportunity to seize enormous wealth if sufficient force could be mustered. These factors had actually been at play (along with the increasing wealth of Germania itself) in consolidating Germanic political units for centuries before the invasions.

A couple of notes on the text: (1) There is a fair bit of repetition of ideas and even phrases in the last three chapters -- it could have done with more editing polish. (2) Throughout the work, the maps fail to provide enough detail. There are many significant features referred to in the text (example: the passes over the Haemus mountains) which cannot be identified in the maps.

Peter Heather, The Fall of the Roman Empire: Chapter 9, "End of Empire"

Heather starts by batting down suggestions of neglect by the Eastern Empire. He argues -- look at those Sasanians over there! and those Huns! -- that Constantinople had a full ledger already, and did well to spare what they could when they could for the West.

That done, Heather recounts the East's final attempt to shore up the West, starting with the installation of a new emperor on the Western throne. (Sidonius is still the main source for events in the West in this period.) The Western general Ricimer had fallen out with his original partner Majorian and then found no support forthcoming for his hand-picked replacement, Severus, so he had to settle with the the Eastern emperor Leo. As a result, the well-connected Eastern general Anthemius was promoted to the purple.

The first order of business was taking out the Vandals in North Africa. Majorian had already given this a go in 461, but the Vandal king Geiseric had destroyed his fleet in Spain. In 468, the combined Empire put together another army and fleet for a landing near Carthage, but the Vandals caught the fleet in an unfavorable wind and defeated it with the use of fire ships.

With the failure of the expedition, the Western Empire had run out of options. There weren't enough resources left to contain the barbarians in Western Europe. The Visigoths seized most of Spain and Southern Gaul, the Franks occupied northern Gaul, and the Burgundians and other tribes picked up smaller slices.

Heather examines the process of collapse though two examples -- one, Noricum, at the periphery, and the other, Gaul, in the core of the empire. In Noricum, the archaeological and literary evidence (the latter coming from the Life of Saint Severinus) shows that for several decades before the final collapse there had been a withering away of the army garrisons, and the abandonment of scattered estates for walled Roman refuge towns. Over time, the refuge towns consolidated and moved further away from the river frontier as security deteriorated.

In Gaul, literary evidence (from the letters of Sidonius, in particular) shows that the conquering Visigothic and Burgundian kings sought and often got support from the Roman landowning elites. The new barbarian rulers needed the Roman landowners to maintain the estates (and possibly provide taxes) and they needed skilled bureaucrats from the landowning class to administer their states. In return, the kings were willing to let cooperative Romans keep at least some of their land.

In Italy, Ricimer fell out with and deposed Anthemius, setting off a further round of musical chairs with the imperial throne. In the meantime, the mostly barbarian Roman army of Italy was getting restive over not getting paid; Odovacar, a general of Sciri birth, took matters into his own hands by deposing the last emperor and distributing land to the soldiers in lieu of pay. (I gather from the notes that Heather's main source on these events is Procopius.)

Peter Heather, The Fall of the Roman Empire: Chapter 8: "The Fall of the Hunnic Empire"

The main source describing the demise of the Hunnic Empire is (the freagmentary remainder of) Jordanes' Getica. Jordanes, a Byzantine writing in the mid 6th century, claimed to have drawn his narrative from the Goth Cassiodorus' history from earlier in the century. This is likely true, but it is the source of a main concern in using evidence from the Getica: it is too centered on the Amal-led Goths (who were later Ostrogoths) and biased in favor of the ruling dynasty of those Goths. Heather concludes from reading between the lines of the Getica and pulling in other fragmentary evidence that the Huns suffered a slow bleed of subject peoples from the time of Attila's death, with peoples asserting their independence as they settled on internal leadership. The Gepids seem to have been the earliest to break away (and they confirmed their independence with victory at the battle of Nedao); the Amal-led Goths apparently didn't secede until Valamer had attained internal supremacy in a series of battles. As the Huns lost tribute and military manpower from more and more subject peoples, their position deteriorated to the point where they fled the Hungarian plain altogether for refuge in the Eastern Empire south of the Danube.

Heather argues that the Hunnic coalition was unstable from the beginning. The subject peoples were enrolled in the Hunnic Empire against their will and often treated harshly, but administered largely though their own native leadership. The Huns depended on intimidation and the distribution of the tribute from predation on the Roman Empire to keep the subject rulers in line. The Huns seem to have run into the limit of their ability to extract money from the Romans, however, and the defeat of successive invasions diminished perception of their power and probably reduced their revenue as well. This left the Attila's sons without the wherewithal to keep their vast empire intact.

Heather points out three significant consequences of the Hunnic collapse for the Roman Empire. First, it complicated the situation on the Danube frontier. The Romans now had to manage many frequently conflicting German tribes. Both the victors, who could be strong enough to exact tribute, and the losers, who often invaded or sought refuge in the Roman Empire, created challenges for Roman policy.

Second, the end of the Hunnic threat put Aetius in a precarious position. Power brokers in Roman politics judged that they could now do without his generalship. In fact, the emperor Valentinian assassinated Aetius in 454, only to be struck down in turn by co-conspirator Petronius Maximus the next year. (Petronius lasted for even less time.)

Third, the Huns could no longer be used as a mercenary counterweight to the barbarian groups already established within the Empire. Since the Western Empire itself no longer had the resources to contain them, either, barbarian groups had to be bargained with. From now on, they would play a leading role in the politics of the empire and the imperial succession. Thus Avitus, with the backing of the Visigoths, succeeded Petronius.

Here Heather has some fun taking apart the propaganda of Avitus' son-in-law Sidonius. Sidonius did his best to convince the Romans in Italy that Avitus was in control of the Visigoths rather than the other way around, and that the Visigoths were fine fellows anyway. The senators weren't buying, however, and Avitus was soon deposed by the Italian generals Majorian and Ricimer. From this, Heather draws the further conclusion that there were now too many factions to satisfy in the Western Empire: with the barbarians directly involved, no stable regime could be established