Interest-bearing loans and unpayable debts in slow-growing economies: insights from eleven historical cases

Under what circumstances are interest-bearing loans compatible with an economy without much growth? The question is becoming increasingly important given a tendency towards declining GDP in industrialised economies and increasing evidence that continued growth is incompatible with environmental sustainability. Previous theoretical work suggests that when interest-bearing loans compound, this results in exponentially growing debts that are impossible to repay in the absence of economic growth. We here examine eleven historical cases to assess support for this finding. We find that interest-bearing loans have typically resulted in unpayable debts in these nonand slow-growing economies. We further identify five broad category of measures that could prevent or alleviate the problem of unpayable debts, and show how they have been employed in the past.


Introduction
There are doubts about whether economic growth can continue at its current rate ( in review). The strictness of these conditions supports the views of those who continue to suggest that there exists a conflict between an economy with positive interest rates, where debt can grow exponentially, and a real economy that is subject to environmental and resource limits (e.g. Soddy  What does the empirical record show? How have economies without growth fared -or dealt with -interest rates? And do we find in the historical record examples of societies that tried to impose the strict conditions postulated in the aforementioned models? These are the questions motivating this research. No modern economy has yet experienced prolonged periods of zero or very low growth. In contrast, no pre-modern economy ever exceeded annual per capita GDP growth above 0.2 or 0.3 per cent (Bolt et al. 2018). Historical research thus serves as a useful empirical complement to the current theoretical debates about the feasibility and potential institutional forms a more sustainable economy may take (Matutinović 2007).
Others before us have suggested that historical cases give fodder to the argument that there is no fundamental incompatibility between positive interest rates and zero growth. Cahen-Fourot and Lavoie, for example, write that "interest [was] in existence long before modern capitalist economies relying on economic growth as their ultimate goal took shape in the western world" (2016: 167). Similarly, Strunz et al support their suggestion that interest is compatible with a non-growing economy by citing the existence of interest in Egypt, Mesopotamia, America, India, China, and medieval western Europe (2017: 346). The desire to better understand the historical contexts in which interest-bearing loans have been possible is what motivates this paper.
A first aim of the paper, then, is to examine the extent to which the charging of interest in the absence of substantial economic growth has been problematic historically. We examine literature on eleven different cases (see Table 1). The cases are drawn from across Eurasia, Oceania, and the Americas, their dates ranging from bronze age societies of the 2 nd millennium BC up to the 19 th century AD, when the period of modern economic growth was just beginning. 1 Recent research has presented cross-cultural comparisons of debt, including instances of the problems caused by debt in general (Graeber 2011), the early development of interest-bearing debt and debt forgiveness (Hudson 2018), and a comparison of the European and Islamic traditions (Geisst 2013). We here systematically compare the literature on specifically interest-bearing debt, drawing particularly on source materials published in the last few years on the economic history of regions beyond Europe and western Asia.
We find that in these historical cases, interest-bearing debts often did come to exceed the ability of debtors to repay. In general, the presence of interest-bearing debt in a non-or slowgrowing economy tended to be accompanied by the accumulation of unpayable debts, debtor dispossession, and social upheaval. Sometimes the problem may be quite limited, since if there are only few debtors then unpayable debts do not become widespread social issue, and even low rates of aggregate growth might be sufficient for small numbers of debtors to be able to repay loans at interest. Sometimes territorial expansion mitigates the problem of growing debts for a while, whether by providing the resources through which debtors can meet repayments or by providing new colonies for the dispossessed to go. But at some point expansion encounters limits. Therefore, our research shows that though societies with little or no growth did indeed have lending with interest, this was often a major source of inequality, social division, and instability.
A second purpose of this paper is, then, to identify the kinds of measures attempted by different societies to try to avoid or to mitigate the consequences of unpayable debts. We place these different possible measures into five broad categories: eliminating debt altogether; forgiving accumulated debts; eliminating the charging of interest altogether; preventing the compounding of interest; and limitations such as rate caps and setting maxima on the size and duration of loans. Our specific focus here is on historical cases where the accumulation of interest renders debts unpayable. The possible solutions we here identify here thus either prevent the accrual of interest resulting in unpayable debts, or mitigate such debts when they do occur. The measures we have discussed in our historical cases all fit into one of these five broad categories.
First, abolishing borrowing altogether is one obvious way of preventing unpayable debts: if there if there is no debt, no-one will ever find themselves with a debt they cannot pay. We find some evidence for a ban on borrowing among the Persians. A ban on all saving would achieve a similar effect, since then nobody would have anything to lend, though we do not find a ban on saving in any of our cases. Though societies can effectively operate without individuals saving or lending (Hartley 2019), such societies tend to be small-scale and not to have experienced the problems associated with interest-bearing loans.
Second, cancelling debts could also resolve the problem of debt accumulation. Such debt forgiveness measures may take place on the individual level as well as more broadly. More narrowly, creditors may be motivated to forgive individual debts through informal norms such as a desire to behave charitably, or may be compelled to forgive loans through the formal mechanisms provided by bankruptcy legislation, for example. More broadly, debt jubilees provide for the cancellation of debts of large numbers of indebted individuals. We find the broad approach of regular jubilees in ancient Mesopotamia and among the Hebrews, and more occasional debt write-offs, as well as narrower individual cases of debt forgiveness, in Athens, Rome, and China.
Third, banning the charging of interest altogether is another obvious way to prevent the accrual of interest into unpayable debts. Interest bans might be partial, such as when interest is permitted on some forms of loan such as for trade or production, but forbidden on loans that might be expected to quickly become unpayable such as consumption loans to the poor. We find interest bans attempted in Rome, China, Christianity, and Islam, and a variety of partial bans among the Hebrews.
Fourth, the compounding of interest might be prevented. The compounding of interest is particularly problematic in non-or slow-growing economies, since the compounding of interest allows debts to grow exponentially. The broadest possible measure to prevent compound interest would be to prevent creditors lending any of their interest income, since then no interest would be paid on interest. In Christianity, we find charitable lenders allowed to charge interest only to cover the costs of lending, effectively preventing interest income being lent, and in China interest income was often earmarked for specific spending purposes. In these cases the intention seems to have been to prevent charging excessive interest and to raise funds, respectively, rather than to explicitly prevent interest compounding, though that would have been the result. A narrower version of preventing compounding is to prevent a creditor charging interest on the interest owed by any individual debtor. This is often effected as a ban on adding interest to the loan principal. Such a measure would at least prevent the exponential accumulation of interest on individual debts, if not in the economy as a whole. Such a ban seems likely to have been much easier to police than attempting a broader ban on compounding by trying to prevent lenders lending their interest income. We find such bans on charging compound interest in Rome, China, India, and Christianity.
Fifth, limitations capping the maximum interest rate, the maximum size of a loan, or the maximum duration of a loan could be applied. Though they would not prevent the accrual of unpayable debts altogether, such measures would limit the number of debts that become unpayable and the rate at which those debts increase. We find such measures in Rome, India, China, Christianity, Islam, perhaps among the Nahua, and among the Tolai. Following work by Passinetti, Lavoie and Seccareccia have recently advanced the idea that a 'fair' rate of interest should equal the rate of labour productivity growth plus the rate of price inflation, and use this to explain why in historical economies with low labour productivity growth and low inflation, the 'fair' rate of interest should be at or close to zero (Seccareccia and Lavoie, 2016). It is worth noting that even in modern economies interest often far exceeds this 'fair' rate, with rates of 1,000 percent per year not unheard of in highly deregulated economies (Martin 2010). One could argue that if lending markets worked efficiently, interest rates should converge towards the rate of real economic activity. Even if that were true, this would suggest that interest rates in a steady-state economy should tend towards zero, and would thus support the proposition that in a zero growth economy, the rate of interest should be zero.

Eleven historical cases
We begin by surveying the eleven historical cases, for each briefly describing the problems attributed to interest-bearing loans as well as the measures attempted to prevent those problems. We proceed roughly chronologically, according to the date of the earliest evidence for the appearance of interest-bearing loans in each of the cases. We list the evidence and our sources in more detail in the Appendix.

Case 1. Ancient Mesopotamia
Charging interest appears around the 25 th century BC 2 in records found in the Sumerian citystate of Lagash and relate to tributes owed by neighbouring Umma (Hudson 1993: 73). By the 21st century BC private households as well as the temples and palaces were engaged in money-lending for economic gain (Garfinkle 2004). Any unpaid interest was treated as a new loan upon which interest could be charged, after a period of sixty months had elapsed. Since the customary rate of interest was 1/60 th per month (20% percent per year), this was the date at which interest would equal the size of the original loan (Hudson 2000a: 147). By the 24 th century, the first recorded debt cancellations are made in the Edict of Enmetena, the ruler of Lagash (Hudson 1993: 74). There is evidence of at least 39 debt cancellations by various Mesopotamian rulers between the years 2400 BC and 1600 BC (Hudson 1993: 8-9, 18-19, 23, 25, 39, 46, 72, 74), with some suggestion that the practice persisted in the north of Mesopotamia into the 15 th century BC (Hudson 1993: 26). 3 Proclamations enforced the forgiveness of debts (Akkadian: mīšarum) and allowed those enslaved due to indebtedness to return to their place of origin (andurārum), though an exception was made for debts incurred to obtain a profit or in the course of commercial travel, which were not in general forgiven (Lemche 1979: 12, Hudson 2002. However, by the first millennium BC, and possibly already in the period leading up to the Sack of Babylon in 1595 BC, Mesopotamian rulers had gradually lost their power to annul the titles acquired by private creditors. Though rulers continued to proclaim clean slates, these retained a largely symbolic function, and no longer had the restorative effects they had had before (Hudson 2002: 14).

Case 2. The Hebrews
The Hebrews may well have their origins in the Hapiru people who in the second millennium BC were made landless through their inability to pay interest-bearing debts and left Mesopotamia to settled westwards (Hudson 1993: 72). The main source of Hebrew law is the five books of the Torah, or 'law'. One of these, Leviticus, uses the words daror and misarim, loan-words from the Akkadian andurārum and mīšarum, to refer to the freeing of indentured

Case 4. Classical Rome
Reports of tensions due to interest-bearing debt appear soon after the beginning of the Roman Republic, traditionally dated 509 BC. 5 By 494 BC a "blaze of hatred" was rising among those who found themselves "enslaved and oppressed" by their indebtedness to their fellow citizens (Liv. 2.23, trans. Roberts 1912). An early law seems to have forbidden the charging of interest altogether (Geisst 2013: 17). In 450 BC legal reforms, inspired by earlier Greek laws, set a maximum interest rate, first at ten then at five percent, and ultimately banned compound interest (Tac. Annales 6.16, Hudson 1993: 50, Momigliano 2005: 180, Liv. 3.33). Unrest over debt is again recorded in 380 and in 378 BC, and colonies such as Satricum are founded specifically as a means to alleviate the debt crisis (Liv. 7.6-7, Cornell 1989: 322). A law of 367 BC stated that the interest already paid on debts should be deducted from the principal, and set maximum limits on land ownership (Cornell 1989: 337). A law of 357 BC reiterates an interest cap, and a law of 347 BC limits the rate to 4 1/6 percent annually (Cornell 1989: 333, Hudson 1993: 50). Probably the same law of 347 BC legislates a moratorium on existing debts, followed by another law setting severe penalties for charging interest in 344 BC (Cornell 1989: 333, Liv. 8.28). In either 313 or 326 BC, 6 legislation called the lex Poetelia abolished the indenture of debtors and established a new kind of contract for loans. That law also permitted interest to be charged when the borrower might incur some costs as a result of wear and tear, but prohibited charging for the loan of an item that was consumed by the borrower and later repaid in kind, the latter category including money loans (Noonan 1957: 105-112, Homer and Sylla 1996: 73, Graeber 2011: 290). However, chronic indebtedness continued, and the numbers of dispossessed continued to grow. Many were recruited into the army or became colonists: between 367 and 287 BC twenty-one Latin and six Roman colonies were founded (Abbott 1911: 49). The period from 287-133 BC was characterised by internal political stability and colonial expansion, the two likely related as lasting However, after 180 BC only one new colony is reported to have been established (Abbott 1911: 78). Land ownership limits were reinstated in 133 BC, but repealed in 130 BC (Hudson 1993: 51). In 86 BC three quarters of all Roman debts are remitted, and in 70 BC a ban on charging compound interest is introduced to the province of Asia (Plutarch Lucullus 23.1, Wylie 1994: 114, Hudson1993: 52). During the first century BC, when the Roman Republic entered its final crisis, Cicero reports that the problems of indebtedness had never been greater (De Officiis 2.84), and the historian Sallust writes that numerous dispossessed farmers were moving to the city and swelling the number of urban poor (Bellum Catilinae 37). Having assumed control of government in 49 BC, Julius Caesar enacted bankruptcy laws allowing wealthier debtors to surrender their possessions to limit creditor claims for full satisfaction (Hudson 1993: 52) and legislated that interest should not exceed the loan principal (Mommsen V: 526). In AD 33 the 8 1/3 percent rate is re-imposed by the emperor Tiberius. The empire reached its largest extent under Trajan in AD 117, but his successors no longer pursued a policy of expansion and increasingly debased the coinage to cover expenditures (Abbott 1911: 324, 326). The 'Crisis of the Third Century', characterised by rampant inflation and civil wars, is often considered the end of the classical period (Abbott 1911: 329, 334, Graeber 2011: 283).

Case 5. The Persians
We mention here the Persians, of around 430 BC, since debt was apparently abolished altogether, though the available information consists of this sentence from the Greek historian Herodotus: "They hold lying to be the most disgraceful thing of all and next to that debt; for which they have many other reasons, but this in particular: it is inevitable (so they say) that the debtor also speak some falsehood" (I.138, trans. Alfred Denis Godley).

Case 6. India
Though debt is mentioned in texts dating from perhaps as early as 1400 BC (Kosambi 1996: 147-148), the earliest surviving discussion of interest (vrddhi) in India is in texts by the Sanskrit grammarian Pānini dating to the 5 th or 4 th century BC. Pānini mentions rates of 0.5, 2.5, and 10 percent, with 10 percent described as a rate that 'was not favoured' (Pānini V.I.47, IV 4.30; Chatterjee 1971: 21). Buddhist texts --the Jātakas relating events from the 5 th century BC and the Therīgāthā poems dating from the 6th-3rd centuries BC, but both written down much later -also refer to interest (vrddi). One of the Jātakas expresses tolerance for the profession of money-lending, though elsewhere ascetics accused of lending at interest are described as hypocrites and there are descriptions of slavery resulting from debt (Jātaka IV, 422, 184, 521; Therīgāthā. 444; Rhys Davids 1922: 205). Later legal treatises and commentaries dating to between the 1 st and 10th century AD are known as the Dharmaśāstra. Monthly rates of 2, 3, 4 and 5 percent by caste are reiterated, but a higher annual maximum of 24 percent is also set, a rate cap that is consistently restated for the next thousand years (Chatterjee 1971: 23-35, Kosambi 1996: 147-148). From the first to the fourth or fifth centuries AD, higher rates for commercial loans are allowed up to 60 percent annually, 120 percent if the debtor passes through a forest, and 240 percent when the debtor crosses an ocean (Chatterjee 1971: 22 29; Kautilya 3, II.1, Yājñavalkya II.38). Already by the second century AD commercial rates are deemed acceptable if agreed by those expert in commerce (Manu VIII, 157). 7 By the ninth century, 5 percent per month is held acceptable if the lender thinks the borrower will use the loan for extensive business, though other rates may be agreed among travelling traders or where cloth is to be received in payment (Medhātithi on Manu VIII 152, Chatterjee 1971: 28, 60-61). The rule that interest should not exceed principal is also repeatedly endorsed (Manu VIII 151, Yājnavalkya, Brhaspati X 17, also 21, 24, Bhāruci VIII 15 Chatterjee 1971: 49-52) and persists into modern Indian law where it is known as Dāmpudat (Vesey-FitzGerald 1925, Chatterjee 1971: 50-51).

Case 7. China
In China, lending at interest is well attested for the Warring States period from 5 th -3 rd century BC (Von Glahn 2016 ch.4). There are also mentions of emergency state loans of grain or money, but these were regarded as a form of philanthropy with little expectation they would be repaid (Von Glahn 2016 ch.5 n51). The 3 rd century BC philosopher Mengzi blames increasing lending for "causing the old and the very young to be cast into the ditches" (quoted in Peng 1971: 107 n11). The Zhou li, a detailed description of government administration from around the 3 rd or 2 nd century BC, records that the state should limit interest to 20 percent (Peng 1994: 535, Von Glahn 2016). The Qin dynasty of the 2 nd century BC banned the taking of interest on private debts altogether, outlawed debt bondage for wives and slaves, and capped land ownership (Von Glahn 2016: ch3).
During the Han dynasty of the 2 nd and 1 st centuries BC charging interest was subject to a rate cap. Some nobles were stripped of their titles for charging excessive rates (Yang 1971: 5). As under the Tang, debtors under the Song dynasty (AD 960-1279) were often reduced to indentured labour. One Song reformist, Wang Anshi, identified the chief cause of distress to family farms as a perpetual indebtedness and a lack of capital for investment (Von Glahn 2016: ch.6). Public loans set up in the 11 th century to relieve these problems charged 20 percent interest per growing season, but this amounted to 40 percent per year and was thus attacked for driving households to starvation or flight by charging twice the level stipulated in the Zhou li twelve centuries earlier (Peng 1994: 535; Von Glahn 2016: 535). A large part of the problem seems to have been that the initially secondary goal of increasing government income ultimately took primacy, and so the program ultimately reinforced rather than alleviated the cycle of rural indebtedness (Peng 1994: 535, Von Glahn 2016: ch.6).
In AD 1173 the Jin emperor Shizong ordered the establishment of pawnshops charging 1 percent interest per month (Peng 1994: 625). Under the Yuan dynasty (AD 1271-1368), legislation again set rate caps and prohibited interest exceeding principal, but this seems often to have been flouted, particularly by Uighur moneylenders who lent at annual rates of 100 percent. In 1240 debts owed to Uighur moneylenders were paid off by the government using official goods (Peng 1994: 625). In 1282 powerful families are recorded as causing a crisis by charging 5 percent per month on loans, and making debtors who could not pay enter new contracts, effectively compounding the interest (Peng 1994: 626). A 3 percent monthly rate cap and ban on interest exceeding principal is reiterated by the founding emperor of the Ming dynasty (1368-1644) (Peng 1994: 741-2). From the end of the 17 th century a sliding scale between 0.8 and 3 percent per month depending on the size of the loan was introduced for pawnshops in some Chinese regions (Yang 1971: 98-99). Under the Qing (1636-1912), the government frequently used pawnbrokers to lend government funds, with the interest generally earmarked for specific purposes such as scholarships to public schools, famine relief, or the maintenance of an orphanage (Yang 1971: 99). The last Imperial Dynasty, Qing law still stated that no more than 3 percent interest could be charged per month, and that interest could not exceed the principal (Von Glahn 2016: ch.8).

Case 8. Christianity
In AD 325 the Roman emperor Constantine called a meeting of bishops, the First Council of Nicaea, in an effort to reach a consensus on Christian doctrine and law. Amongst the laws agreed was a ban on clergy charging interest (Geisst 2013: 20). In AD 345 the ban was extended to lay people (Canon 12 of the Council of Carthage). In AD 380 Saint Ambrose, the influential Archbishop of Milan, relates stories of debtors forced to sell their children and committing suicide from shame, and equates usury with violent robbery and murder (De Tobia 15.51; De Officiis 2.25.9; Graeber 2011: 284; Issler 2016). Carefully examining every Biblical reference to moneylending, Ambrose particularly notes the exemption from a complete ban on charging interest indicated by Deuteronomy 23:20, the verse explicitly permitting loans to 'foreigners'. If charging interest is equivalent to fighting only without a sword, Saint Ambrose reasons, then it would not be a crime to charge interest to those "whom it would not be a crime to kill" (trans. Zucker, quoted in Issler 2016). Only much later 8 is an alternative explanation put forward: that Hebrew law allows interest to be charged to nokri because these non-resident foreigners have come to Israel to trade; but such an interpretation appears not to have occurred to the thinkers of the early Church.
In AD 443 Pope Leo reiterates a ban on all lending at interest (Geisst 2012: 22). However, as early as AD 529, the legal code of Eastern Emperor Justinian acknowledges the religious ban but notes that it is hard to enforce. Justinian In Italy in the 14 th century, public lenders known as montes de pietatis were set up to provide the poor with low interest loans as an alternative to the high interest charged by private creditors. This received Church approval in the Fifth Lateran Council of 1512-1517 on the condition that the interest charged was sufficient only to cover the costs of lending by the montes and was not causing the borrowers harm (Geisst 2013: 28, 64). The logic of the 3 rd century BC Roman law, allowing interest when a loss had been incurred by a lender, began to be applied more widely (Noonan 1957: 35, 100; Issler 2016: 783). But by 1600 the prevalence of money markets effectively meant that making an interest-free loan to one person would mean that a lender always lost the opportunity to make a return elsewhere (Noonan 1965: 221, Noonan 1957 Since the 16 th century Protestant reformers, often funded by banks, had argued for the relaxation of the interest ban (Geisst 2013: 74-75, Burke 2009. Legislation in the 16th century by northern Swiss Protestants, the Habsburg Netherlands, and the Holy Roman Emperor Charles V permitted interest up to a limit on commercial loans, and a law passed in 1545 by newly protestant England effectively decriminalised interest below 10 percent on all loans (Gordon 1975: 257, Helmolz 1986: 379). Rescinded in 1552 but reinstated in 1571, laws permitting interest up to a certain rate became standard in England, later in the United Kingdom, and in British colonies. In the UK and in almost all states in the post-independence USA, interest rate maxima around 5-8% per annum were in place until well into the nineteenth century (Geisst 2012: 149, passim). Nevertheless, one cache of documents from Cairo 10 suggests that until the 12 th century the lending of money at interest was both shunned and of economically limited significance, with investment instead taking the form of partnerships; by the mid-12 th century, however, interest began to be charged through contracts which concealed it in some way (Rubin 2011: 1315, Kuran 2012). By the 16 th and 17 th century, the Ottomans effectively permitted interest up to a set rate, provided the requisite contractual devices were used. In various Turkish cities loans were charged at annual rates ranging from 10 to 20 per cent, and were permitted by both legal and religious authorities (Kuran 2012: 148, Rubin 2011: 1316n15).

Case 10. The Nahua (Aztecs)
Among the Nahua, interest-bearing loans are recorded by Spanish chroniclers, the earliest dating from twenty-two years after the start of the Spanish conquest in 1519. Terms relating to lending at interest were already well established in the Nahuatl language at the time of conquest, but the details of what this actually consisted of is much harder to state with certainty. In 1541 the earliest chronicler, Motolinia, initially wrote that interest was charged, but apparently later came to believe that this was not the case (Millhauser 2017). The Codex of Tepeucila, from 1543, relates that in 1535 the local leadership of a town had borrowed Spanish coin from local merchants at a rate of 200 percent in order to pay a tribute demanded by the Spanish overseer, eventually defaulting and being run out of town by their creditors (Hirth 2016: 222, 267). But particularly since the idea of lending money, rather than goods, was novel for the Nahua the story seems more a reflection of European influence than of preconquest practices (Millhauser 2017: 267). An ethnographic research study begun in 1545, Sahagun's Florentine Codex mentions usury alongside profiteering as a vice of merchants and rich men and describes a bad merchant as one who practices usury and demands excessive interest, which suggests both that norms may have prevented high rates and that lower rates were considered socially acceptable (Sahagun 1961: 59 in Hirth 2016: 267). Molina's Nahuatl-Spanish dictionary compiled between 1555 and 1571 contains five translations of the Spanish word 'logrero', which meant a usurer or someone whose profits are excessive (Millhauser 2017: 267). Intriguingly, Molina's dictionary also contains four different terms to refer to using one debt to pay another, suggesting that some of these debtors may have begun to experience the compounding of interest, perhaps even the accumulation of unpayable debts (Molina 1571, 2.41v, in Millhauser 2017). Writing in 1570, Zorita states that interest was not charged, but according to his translators someone had written in the margins that in fact "interest was used in many places" (Millhauser 2017: 268-269). It may be that what they observed was the contracting for repayment of a greater amount of some good at some future date, the Spanish phrase 'dar a logro' also referring similar contracts in medieval Europe which were used to create the appearance that no interest was being charged (Hirth 2016: 267). There is no evidence that unpayable debts constituted a chronic societal problem (Smith 2015). However, some instances of debt indenture and the selling of family members are reported (Millhauser 2017: 268-270).

Case 11. The Tolai (Papua New Guinea)
The Tolai and their Melanesian neighbours 11 have been described as 'primitive capitalists' (Epstein 1968), the earliest surviving accounts of their practice of lending shell-money dating to the late 19 th and early 20th centuries. Sometimes borrowing occurred when someone wanted to avoid breaking up their own cache of shell-money but needed a small sum for a purchase, in which case the borrower might prepare a small dinner for their creditor, or otherwise pay a rate of 10 percent, considered as a gift in expression of gratitude for the loan (Epstein 1968: 25-26, Parkinson 1907: 94, Danks 1888: 308). Other loans, such as after the death of a relative for ritual purpose, would be charged at 20 or 50 percent (Epstein 1968: 25-26, Parkinson 1907. Among the neighbouring Buin, a rate of 50 percent was due after about 2 months, a rate of 100 percent after 9 or 10 months, and social pressure to repay after 10 months effectively placed an upper limit on both the length of the loan and the amount of interest that would need to be repaid (Thurnwald 1912: 42-43, Connell 1977. A similar limit is also thought likely for the Tolai, though some suggest loans of 2 or 3 years or more might have occurred without further interest being charged (Connell 1977: 85). Nevertheless, kinship or friendship obligations typically placed the creditor under considerable social pressure to forgive loans altogether (Epstein 1968: 72, 106).

Discussion
In answer to our first question: the charging of interest in the absence of substantial economic growth resulted in notable levels of unpayable debt in seven out of our eleven cases (see Table 2). In these seven cases, this resulted in debtor dispossession and indenture, and at least some degree of social upheaval or revolt.
The four exceptions are instructive. First, among the Persians borrowing and lending was itself apparently abolished altogether. Second, in Islam a ban on interest was legislated since the religion's inception. Third, among the Nahua debt seems to have been relatively small scale prior to Spanish conquest, and even then some instances of debt indenture and the selling of family members is reported. Fourth, among the Tolai, interest-bearing debt also seems fairly small scale, limits were placed on the length and total size of repayment, and there was a broad social pressure on creditors to forgive overdue debts.
Some, perhaps Perhaps The Tolai Y Though we do not have a very clear picture of the extent of lending and borrowing in these societies, it seems reasonable to assume that in most cases it was less prevalent than it is today. But social revolts, jubilees, and other alleviatory measures suggest that, in at least seven of these cases, lending with interest became widespread and turned into a social problem. The more extended lending is, the more individual problems of indebtedness are likely to translate into a bigger social problem; where lending more limited, wider social problem did not inevitably result. In the case of the Nahua, for example, evidence suggests that lending at interest was not widespread, and though a few individuals do seem to have found repayment difficult, there is no indication indebtedness caused wider systemic problems. Moreover, even without rapid growth in the economy as a whole, it is plausible that the incomes of the those who do borrow increases sufficiently to repay quite high rates of interest. For example, merchants in Athens, Rome, and India seem to have been able to meet even quite high rates of interest. There is, thus, a logic behind the distinction between commercial loans and other loans made in ancient Mesopotamia for example, where the debts of commercial agents were not forgiven alongside all other debts, and among the Hebrews where interest was permitted only on loans to foreigners travelling for business purposes.
Our second broad question is: how was the problem of unpayable debt resolved? Among our cases, only the Persians avoided the problem of unpayable debts by abolishing indebtedness altogether. In our two earliest cases, in Mesopotamia and among the Hebrews, regular jubilees forgiving accumulated debts were enacted. One-off or partial debt jubilees were also occasionally instituted in Athens, Rome, and China; in the cases of both Athens and Rome, territorial expansion is also thought to have helped reduce tension by providing an inflow of resources as well as a place for the dispossessed to colonise. Across cases individual creditors are often exhorted to forgive debts.
Rome, China, Christianity, and Islam all attempted to ban the charging of interest, and partial bans are found among the Hebrews and in India. Though the Roman and Chinese bans are relatively short-lived, the Christian and Muslim bans were upheld in theory and often in practice for centuries, and when interest becomes tolerated it is long regulated by the use of interest rate caps. Rate caps are also found in Rome, India, and China. As for bans on compound interest, arrangements that would have prevented compounding altogether by preventing the lending of interest income are in evidence in Christian montes de pietatis and in Chinese pawnbrokers, though in those cases the intention does not seem to have been to prevent compounding as much as to prevent excessive interest and to raise funds. Bans on charging compound interest on individual loans, likely easier to enforce than more general bans on lending interest income, are found in Rome, India, China, and Christianity. Measures to prevent interest accumulating to greater than the principal would also generally have had the effect of preventing interest being charged on previously accrued interest, as well as limiting the total size of the resulting loan, are found in Rome, India, China, Christianity and among the Tolai.

Conclusion
When they become widespread, interest-bearing loans have tended to be accompanied by unpayable debts, dispossession and indenture of debtors, and wider social upheaval and revolt. The findings of this paper thus highlight that interest-bearing loans have in general been very problematic in past non-and slow-growing economies, We have found that this problem has in the past been avoided or alleviated by banning all debts, forgiving debts, banning interest, banning compound interest, and setting limits on loan sizes and on interest rates.
A question remains as to the extent to which modern economies have different characteristics than these historical economies we have studied. Though financial institutions are now recognised to have played a much larger role in the past than previously thought, it is clear that the scale of lending with interest is today far more widespread, more complex, and more fundamental for the functioning of the economic system than it has ever been in the past. This may make it much more difficult to control indebtedness in the way that past societies were able to do. Nevertheless, our aim has been to show that current theories that suggest interestbearing loans may become problematic in the absence of substantial growth have significant empirical support when tested against historical cases. We hope our findings will stimulate further research into how the problems caused by interest-bearing loans may best be avoided in future slow-and non-growing economies.