Digitized by the Internet Archive in 2011 with funding from Boston Library Consortium Member Libraries http://www.archive.org/details/moneypricesinearOOkess DEWEY 'hB31 .M415 ^^ Massachusetts Institute of Technology Departnnent of Economics Working Paper Series MONEY AND PRICES IN THE EARLY ROMAN EMPIRE David Kessler Peter Temin Working Paper 05-1 April 4, 2005 1 Room E52-251 50 Memorial Drive Cambridge, MA 021 42 This paper can be downloaded without charge from the Social Science Research Network Paper Collection http://ssni.com/abstract=704724 at f^^^^^iii^i;!^! APR 2 6 ^005 LIBRARIES. Money and Prices in the Early Roman Empire David Kessler and Peter Temin* Abstract We examine monetization in the early Roman Empire by considering money as a unit of account. Widespread use of prices indicates widespread monetization. A consistent set monetization encouraged trade to grow across the documented with a statistical test, preceded by a nontechnical introduction and followed by consideration of a range of possible objections. of prices for wheat indicates that Mediterranean. This argument this is Keywords: money, monetization, intemational trade, regression analysis, Roman Empire JEL Codes: N13, F14, CIO Paper presented at the Conference on The Nature of Ancient Money, Columbia To appear in The Nature ofAncient Money, edited by University, 7-8 April, 2005. William V. Harris (Oxford: Oxford University Press, forthcoming). *Temin (corresponding author). Department of Economics, MIT: ptemin@mit.edu Kessler: dkessler(5),fas.alumni.harvard.edu . . Money and Prices Money serves as a in the Early medium of exchange and debts can be expressed.' Most research on world has focused on the first famous empire article to pay taxes; Duncan- Jones location of com hoards. standard of value. was in that and prices Roman medium of exchange. Hopkins' inferred geographically limited coin usage This paper examines the second fimction of money, from the its role as a We argue that monetization in the sense of using monetary measures virtually universal in the early less which money would be needed throughout the Roman Empire. view of recent compilations of Roman prices. and a standard unit in the extent of monetization in the fimction of money, a on taxes and trade argued Roman Empire ^ This assertion verges on the obvious We go further to make the stronger obvious claim that there was unified monetary integration across the whole Mediterranean in the early Roman Empire. We make this argument through an examination of wheat prices. We approach the issue of monetization from a new angle, emphasizing the role of money in supplying a unit of account. This leads us to look for hi other words, view is we look for Roman prices measured by the extent activities. This to may appear less which people used prices our index of monetization. to unit, keep track of their direct than analyzing coins, but instead of asking of this instead of Roman coins. Monetization in this interested in monetization as an aspect of Roman society, Romans used money Roman uses and it is we not. We are focus on how the how much money they had. The use of prices is We assume that money was used as money in the ancient world. Some historians have seen coins as symbols of imperial power or as art objects. These characteristics may be important for the question of which coins were used, but they do not approach the question spread why the Roman ancients minted coins in the first place. Perhaps Roman came coins widely to symbolize their suzerainty, but with coins practice of using money to emperors the value commodities and services. Actually, this practice preceded coins in the East and became more accurate and prevalent with coins. Uniform —linked sestertii to earlier drachmae transactions. Consistent valuations in the East —encouraged and ease of payment (the the use of prices in first function of money) encouraged the growth of trade."* Prices were used widely in Rome, and the of references to them. Unhappily for the letters and accounts of the time are modem scholar, the prices goods or services that are unique, ranging fi-om dirmer parties to full almost always are for monuments. The accounts reveal that people thought in terms of prices, but they do not provide a data set with which purpose. to examine Wheat is prices. Prices for a uniform commodity are needed for the obvious candidate for such a price because wheat uniform and universal. Rathbone used the price of wheat in inflation.^ that was both Roman Egypt to measure We compare wheat prices across space rather than through time to demonstrate market integration around the Mediterranean. Market integration indicates widespread monetization, which Just as the euro facilitated trade by providing a standard imit of account. promotes European trade today, the wide use oi sestertii and drachmae encouraged trade throughout the Roman Mediterranean. Despite the availability of Roman price evidence, prices for many Rickman in his account prices, but locations. we think it is literature for good one comparable city of Rome. ^ for three reasons. First, This is a small sample of Rickman searched the mentions of wheat prices in outlying areas. Second, he compared prices in these outlying areas with prices in comparison in our work. Third, prices for a uniform early difficult to find We take our sample to be a set of wheat prices hsted by of its supply for the a it is Rome, and we use we commodity in are not many this roughly contemporaneous aware of another attempt different parts of the late to provide a set of Roman Republic and Roman Empire. We proceed in three steps. We first explain this small data set and its and weaknesses, construct a graph showing the relation of wheat prices in strengths Rome to those elsewhere around the Mediterranean, and present our hypothesis in the form of a graph. Then we ask if this prices that in fact graph could be a fluke, that were not the result is, a chance result of putting together of an integrated market. If not, then the collection of prices fi-om distant places would reveal nascent monetization (since prices were used) but not comprehensive monetization (that would encourage trade). There would be no relationship between the various statistics that tries to prices. This test requires a short introduction to be informative without being too technical. Finally, variety of objections to this finding that might be raised, our results. Data and Hypothesis showing we that they discuss a do not vitiate We collected wheat prices from varied locations, prices come but to light all the time, we thought that as reported this familiar way to examine monetary integration at least provisionally. pairs in outlying locations and pairs in almost in Rome at not an overwhelming amount of evidence, but the data are at random or not. hi each case the late it is We found price six price Republic to the early Empire. This enough to test Roman price was the distant location to give a price differential. sample would provide a Our requirements were roughly the same time. two centuries ranging from the by Rickman. More Wheat is whether the patterns in subtracted from the price prices at Rome were subject to slow inflation according to Rickman, a view that has been supported by Duncan-Jones.^ We describe the price observations in the order of their distance from Rome, which we calculated as straight-line distances on a map. This of course approximation to the actual distance that wheat traveled; is only an we emphasize the approximate nature of our data since this randomness reduces the possibihty of finding evidence of comprehensive monetization. The closest price was from Sicily and came from Cicero's Verrine Orations. One of his accusations was market price, even though he acknowledged that its Verres did not transact business level in a letter (Cicero, 2 Verr. at the 3. 189). This observation, like most of the others, reports the prevailing local price in round numbers. Since the observation likely to is general rather than the record of any fransaction, it is be only approximate. This casual quality of the data militates against finding any systematic relationship between prices. the prices being paid because of the unknovm and actual prices. It introduces noise into any relationship of difference between the reported averages We analyze the effects of this noise on our test below. The second conditions in Lusitania. While it is good to came from Polybius price As before, this is (Polyb. 34.8.7) in his discussion of a general statement about the prevaihng price. have an average, the casual quality of the averaging process again adds noise into any comparison of prices in different places. The third price comes from the Po Valley Polybius (Polyb. 2.15.1). While this observation prices, we have made an exception to our general Rome by rivers rather than sea. in Italy; is it is closer to rule. another observation by Rome than the first two The Po valley was linked to Diocletian's Price Edict fixed river transport prices at five times the level of sea transport, and we consider the cost of transport from the valley to have been five times as expensive as evidence however dates from over a century its later actual distance if taken than any of the prices by we sea. Po This observed. We assumed the ratio of sea and river transport costs remained constant over time which others have confirmed the distance from Rome by —and we included five. the Po valley in our data Despite our small sample, by multiplying we have enough data to test the usefulness of this assumption. The fourth price comes from an inscription records that the 126b). The normal price official intervention in the local market. wheat price was limited was in a time An of scarcity (AEPigr (1925), eight or nine asses per modius; the acceptable limit price was one denarius per modius.^ This inscription reveals several important aspects of the Mediterranean wheat market in addition to reporting the normal price. The need to down famine prices indicates that local markets were subject damp to local scarcities; they were not so well linked that wheat from elsewhere would be brought in instantly in response to a local shortage. to double its The apparent success of such normal range, indicates For Egypt, we preserve the Rathbone collected actual that spirit interventions, in this case limiting the price many famines were not severe. of Rickman's data but improve on his data since sale prices that are superior to the observations of average We averaged seven Eg3q3tian prices from the "famine" of 45-47 CE to get a price levels. Rathbone argued for Egypt. suggests that they that these prices may not be far were unusual, but the previous discussion from average. '° We of course cannot know how unusual these prices were, and any special conditions infroduce noise into our data. The Egyptian prices also come from agricultural areas, not from a Mediterranean The port. purported famine would have raised the price, but using country prices would have depressed it compared to those at a port. These the accuracy of this observation since there offsets. is offsets introduce no reason them to be exact The average of Rathbone' s seven prices was seven drachmae per artaba. These prices in Egyptian currency and units Duncan- Jones and dividing by Our final were converted it too is is Rome was between three rising to five to six HS taken from Tenney Frank's an average of a few actual transactions.''^ All of these prices were compared with roughly contemporaneous prices price of wheat at HS per modius by following to 4.5.^^ observation from distant Palestine, Economic Survey; level; to expect added uncertainty into in the early Empire. Rathbone confirmed the and four at HS Rome. Rickman argued that the per modius in the late Republic, Duncan- Jones confirmed the general price inflation, at least for Egypt where the data are more abundant. The order of observations turns out to be almost chronological even though the order of exposition was by distance. The shown in Figure 1. itself; all It prices and the differences Table It is 1 . The readily apparent that prices the price differences are all ftirther and the local prices are were lower outside of Rome than between various locales and in Rome in Rome been noted before. ^^ Rome became more from Rome. In other words, wheat prices everywhere were lower than in Rome and lowest at the farthest reaches of the Mediterranean, that is, Rome. What could have produced such We suggest that the whole a pattern? Mediterranean basin was monetized in the sense that for Rome negative. This difference has also looks as if the price differential places farthest from at price differences are graphed against the distance from negative as the locales were at between the prices money provided a standard of value wheat and presumably other goods as well. This standardization of the monetary unit, taking into account the different currencies in the East, promoted the development of a unified market for wheat, along the lines suggested If there had been a unified wheat market, center of consumption would have been consumers lived and the where the in Schaps.'"^ how would it have worked? The main Rome, where Roman government was largest excess supplies by Rome was and demands for wheat would have come together and exporting regions in view of the transport costs to supplies fluctuated due to harvests across the set. The price would be lower Rome. They would vary over time Roman world, in as storms affected the cost of and government actions altered the value of the currency. Normal variations in supplies and in number of potential located. In other words, where the price of wheat consequently would have been transportation, the largest Rome, although most demands elsewhere fluctuations in the empire would have affected the price would have been small relative to the total production and the consumption at Rome. Most places outside of Rome an excess supply of wheat, and the price would have been set in demand came supplies and the largest excess there could be excess local demand When together. would have had Rome where the were isolated, local famines local places as well as excess local supply, that is, excess as well as local gluts. Under these circumstances, wheat outside of Rome would be valued by what was worth less than in Rome. Wheat wheat in Palermo Rome because price of wheat in Sicily wheat from Sicily at to would be it it was to be transported the price of wheat in Rome. This would be were circumstances when normally would be worth in Sicily, for example, would have to Rome to be less the cost in. from the level set by Roman price until new wheat the price in Roman price less The market Sicilian prices. is More it is would determine a unified market. If the Sicilian price of wheat rose above the well. If the Sicilian price rise would keep Sicilian misleading to say the market would determine correctly, competition amount of wheat demanded would If a the fransportation cost. an abstraction; transportation costs, merchants Rome. supplies could be brought In the absence of extreme events like these, a unified market prices near the of getting prevented the shipment of grain to Rome, harvest failure in Sicily created a local famine, the price of wheat in Sicily above the level indicated by the The sold. true almost always, but there undoubtedly not. If storms the Sicilian price might temporarily deviate Rome it would not buy wheat in Sicily of wheat would fell fall, Sicilian prices if there Roman level in Sicily to sell in minus Rome. The and the price consequently would below the Roman level were fall minus transportation merchants would increase the amount of wheat they would try to buy in as costs, Sicily, for they could make an unusually high profit by taking would bid further it there. Merchants would be worth Palermo because from Rome. The cost of transporting wheat from Spain to Rome was larger it from But while each price Sicily, is the Mediterranean if there therefore, the price it to that in Rome, the price in Spain were a unified market. In less than the price at depending on the distance from Rome. it and the price of wheat in Spain correspondingly compared would be worth are reasonably sure that less than exactly like that for Sicily, only the transport cost is lower relative to the price in Sicily we and selling at Lusitania in Spain would be lower. The reasoning but Rome wheat at than the cost of bringing different. to against each other, raising the Sicilian price. Wheat was it Rome, with fact, the is would be wheat around amount less We do not know the transport cost in any detail, rose with distance. If there was a unified wheat market, of wheat would have decreased as one moved farther and farther from Rome. All this presumes monetization throughout the Mediterranean. The comparisons assvime that there were monetary units in which prices could be compared. In fact, most of the observations were in the same monetary these units by a standard exchange rate. The units, and the others were franslated to anticipation of profits from frade presupposes that the value of the wheat could be compared with the value of other goods and services and traded for them. Without monetization, there can be only the simplest kind of trade because the opportunities are very hard to evaluate. This hypothetical account of frade sounds impossibly accurate picture of the Roman world, we need to Phrased differently, what is modem. But if it is not think of the relevant alternative. the alternative to this view? If there were not a unified an currency or market, would not expect local prices move were only independent if there to find and markets, then local currencies any relationship between local and Roman we prices. Prices in and markets would be determined by local conditions. The prices might together at some times, if storms around the Mediterranean caused simultaneous harvest failures everywhere or currency debasements caused prices to rise everywhere, but the prices would not in general be related one to another. prices could be a coincidence, and Egypt and Dacia were the miners.'^ If we could test fmd whether the pattern The question either that there is was a result we it single identity impossible to say if similar miners' wages in wheat prices is due in different places, to coincidence or however, monetary area and an factors unifying separate local moneys and efficient markets. we could an underlying market process. not whether one or the other of these ideal tj^es single of of coincidence, government regulation, or a market for several find is Any It is was observed, market or that there were no rather whether the historical experience hes closer to one end of a continuum than the other. There must have been least occasional local shortages state and famines. The question then is whether the normal of affairs was one of interconnected currencies and markets, so different places typically and markets. In the were related, or latter case, we that prices in one of separated and independent currencies should not observe any systematic relationship between the location and the price of grain. monetized, of course, at we would not have If the economies in these places had not been observed any prices A Test of Market Integration \. 10 at all. It story. It may appear as seems Figure if the picture in like a tiny bit 1 could only be suggestive of such a of evidence on which to hang such a grand story of universal monetization and market integration. There that can be used to evaluate chance. In other words, how likely it is hov^ever a we can test the probability that the would have been determined by technique statistical that a picture like Figure Roman Empire were isolated and out of economic monetization. There is 1 could arise by separate areas of the early connection with Rome. Their prices local conditions, including perhaps the degree would have been no connection between the of Rome and distance to the level of local prices. This we statistical technique is known can evaluate the likelihood that there distance from between the the line to Rome. it a relation is between the and the Roman price to the distance from line. distance to minimize the distance from points both above and simplify the mathematics.) This process Figure Rome. the best description of the data in the sense that squared distance of the individual observations from the least squares, local price of analysis and the We start by trying to draw a line that relates the price difference local price make as regression analysis. In this type is knovm and the resulting "least-squares" it We then adjust minimizes the (We use the square of the below the line and as regression analysis or the line is the regression line. It is to method of shown in 2. One of the values of regression analysis is that it generates tests of the hypotheses being tested. We can ask if an apparent relationship between the price discount and the distance from Rome result is illusory, a result of observing only a of a systematic process. In order to draw 11 this line, few prices, rather than the we assumed that there was a between the distance from relationship provides a unlikely is we gathered from the price discount. Regression analysis such an association in the data. This for us to find a line like the it is the prices whether there test Rome and one shown in Figure 2 this same actually have, us would have survived Taking account of the random quality of the observations distribution. how unlikely is certainty. It is that Rickman were randomly drawn from an underlying it for us to find the line in Figure 2 we by chance? Regression analysis acknowledges that the slope of the line in Figure 2 known with how by chance. Assume distribution of price observations. In another world, different prices from test tells is not the best line that can be drawTi with the data at hand, but it is subject to errors deriving from the incomplete sampling of the underlying distribution. In the jargon of regression analysis, the slope of the line has a standard error. If all the points in Figure be clear, 1 and 2 lay in a sfraight line, then the slope of the regression line would and the standard error of the slope would be close spread out as they are in the figures here, then the line is a chance that the line has no slope at the distance from The Rome and the price test is to compare the all, that unlikely that the line even though the regression is line known that there is is, no as clearly, relationship and there between size of the slope, the coefficient in the regression, with was a random other hand, if the coefficient not difference. the size of its standard error. If the coefficient it is is to zero. If the points are is large relative to the standard error, then finding without support in the price data. small relative to its has a slope, there standard error, then is it is On the possible that no underlying relationship between the price and distance. Statisticians call this ratio a t-statistic, and they have calculated tables that can franslate t-statistics into probabilities that the line is observed 12 by chance. The tables take account of degrees of freedom, minus the number of coefficients. position (height in the figures). takes It With that two variables six observations is, the number of observations to define a line, its slope and its and two variables, there are four degrees of freedom. Omitting the observation with river transport reduces the The observations by one and the degrees of fireedom to three. t-statistic number of has to be larger with such few degrees of fi^eedom than with more degrees of freedom to show that a given regression line is unlikely to be the result of chance. One might think that the —composed of only a few, badly observed values data are too poor for statistical analysis. Nothing could be flirther are the best there that is a way of distinguishing lot truth. Statistics signal fi"om noise; they are particularly usefiil when of noise in the system. They give us a precise sense of how unlikely any putative pattern processes, that is, value of statistics we think we how unlikely it is that the Mediterranean Sea we can is test were related that what looks in a simple often hypothesize a relationship between — ^but like a pattern actually is noise. way to those at is true, Rome. have been studied, and two variables to most of our prices. The their effects are well we have. common problem in doing regressions. We proxy such as the occasional price that happens for We also can given the observations — like the price in cannot observe one or the other of them precisely. we have done here The a formal hypothesis, namely that wheat prices around In particular, errors in variables are a price in Egjqjt it is observe would have been generated by random derive an explicit probability that this hypothesis as from the be mentioned Rome and the We then use a in a surviving errors introduced document, by such a procedure known. The extra uncertainty introduced by using imperfect proxies reduces the explanatory power of regressions and tends to result 13 in coefficients makes the results look is noise. The well-known When a pattern price differentials are graphed against the distance to differentials all more hke is scarcity of Roman found, however, it a strong relationship between the prices. results are quite striking; prices from near zero; the addition of noise through imperfect very hard to find a pattern in them. it indicates that there The t-statistics that are makes observations prices and were lower in places fiuther Rome in Figure over the Mediterranean and from various times in the late if there come Republic and early were not a unified grain market, there would be no reason to expect a pattern in-these^ricesr'Everrif there unified market, our inability to find more prices or more accurate might have obscured any true relationship among the prices. The . from Rome, and the price appear almost proportional to the distance from Rome. These prices Empire. If there were not a unified monetary system or 1 was a fransportation costs Yet Figure 1 reveals a clear picture. While the graph is clear, a statistical test could be the result of chance. Accordingly the distance from Rome, with the results is needed to tell if the we ran regressions of the price shown in Figure 2 and Table separate regressions in the table. Since the transportation from the rather than sea, accurate, and addition, to we we were not observed pattern 2. differential There are four Po valley was by sure that the correction for the relative cost of fransport we tried the regressions both with on river was and without the Po valley data point, hi expressed the distance in logarithms to measure the proportional change in it allow the relationship between price and distance to be non-linear. The dependent variable is the premium of the Roman the local price from the price over the local price instead of the discount of Roman price. 14 Each row of Table 2 contains the result of a separate regression. columns identify the regression by whether we we used the distance or its The two first log and whether included the Po valley. The next two columns give the coefficients resulting from the regression with the relevant because a straight line column gives the is t-statistic below it in parentheses.'^ There are two columns defined by two parameters, a constant and a slope. constant. It measures the difference between the price The first Rome and in elsewhere that was not related to distance. The second column gives the slope of the line in Figure 2. It distance fi^om measures the speed at which the discount below the three. discount increased with distance? The coefficients indicate that between the relationship local price it is highly unlikely that there is and the distance from Rome. They are as t- no all larger than A t-statistic above three indicates that there is less than one chance in twenty that the observed relationship between distance and price differentials other words, that Rome grew Rome increased. How can we be sure that the price statistics fi-om the price in we was due to chance.'' In confirm with very high probability that there was a unified wheat market extended from one end to the other of the Mediterranean Sea, that transport costs were roughly proportional to distance, and that the effects of distance were larger than the idiosyncratic influences of particular markets, currencies and places. This level of confidence The is R taken as conclusive in most economic and medical shown in the final differentials that is explained column measures by these simple tests done today. the share of the variance of the price regressions. Using the price differentials themselves, the regression explains almost four-fifths of the variation. Using logarithms of distance, the regressions explain even more. This result confirms the impression in 15 Figure 2 that distance from prices around the Rome is wheat a powerful explanatory factor in determining Roman Mediterranean. The constant terms in these regressions were negative in the regressions for price They were not estimated discounts (and positive in the regressions for the logarithms). precisely as the relationship between distance and the price differentials, so as sure that they are not the result of chance (as indicated constant terms suggest that there were some as we cannot be by smaller t-statistics). The costs to bring wheat to Rome that were not proportional to distance, albeit smaller and less well observed. These other costs could have been physical adminisfrative effect — — the costs of transshipping wheat to and like port from sea-going ships — or charges or taxes. Their presence does not detract from the of distance or the evidence in favor of a unified wheat market. It Without does not make a big difference whether the Po valley this observation, the is included or not. standard error of the slope coefficient The required is slightly larger, show making the t-statistic slightly smaller. the slope not zero also rises due to the fewer degrees of freedom. Nonetheless, the is probability that there less than one in twenty. that is explained for the Po was no relationship t-statistic to the probability that between price and distance from The slope of the hne and Rome is still the percentage of the price variation by distance do not change. As can be seen in Figure valley Ues close to the regression line; removing it 2, the observation does not change the line. This graph and these regressions provide powerftil evidence for the existence of extensive monetization and unified wheat markets in the later early Roman Empire. Roman Republic and the Other authors have inferred the existence of such markets from isolated observations, but we have demonstrated the 16 existence of a relationship between prices in far-flung places that almost certainly is not the result of chance. Such a where the standard of value, money, was used throughout relationship could only exist the Mediterranean. \ Possible Objections We discuss in this section possible objections that can be raised to this test and our conclusion. The money was first objection is that prices scarce, not because transport to explain the prices in Table 1 . Coins were low outside Rome was may have been Rome because costly. This alternative coined cannot scarce in Lusitania at the time of Polybius, but coins were abundant in the eastern Mediterranean where the monetized Greek economy preceded the Roman one. Wheat prices Lusitania, as can be seen fi-om the figures. Distance there from were lower than Rome is a in much better predictor of prices than coin scarcity. A second objection is that the prices are unrepresentative because they are notional, biased because the observers price fluctuations. had political motives, or unrepresentative due to We acknowledge that such errors in the price observations may have been present, although Polybius was a very careful evidence to make a rhetorical point. As noted historian, not liable to falsify his already, such errors in recording the "true" prices introduce noise into the relationship between the price differential and distance from Rome. If there obscured. Since we was a great deal of this distortion, any existing relationship might be find such a relation, it means that the relationship between distance and price was a sfrong one, visible even through the noise introduced by casual or distorted price observations. More formally, we 17 can thirik of the observed prices being determined by the true prevailing prices, which approximation. Then the dependent variable distance plus an error. That error a lower t-statistic we we used in the regression would add onto the and R^. Given that they both are assumption in fact is observe with an error due to our the true is error of the regression large, we conclude that and this result in rough quite good, that the observed prices appear to represent prevailing prices in a reasonable fashion. Another, related objection observations is that prices fluctuated during the year may have come from different and seasons. Again, this source of noise strengthens our results because the seasonal price variation introduces another source of noise into the hypothesized relationship. We suspect that the casual nature of the price observations has helped us here. Travelers were told of the prevailing price, not sometimes the extreme price the low price following the Phrased differently, quite random, but that obtains just before the harvest harvest. it it is to be a consistent and sometimes set of prices. we find has to be spurious. throw out evidence from the ancient world on the must be random. to the use of these prices is that the argument is circular: data are sound because they support the hj'pothesis, but the test of the hypothesis requires the data to be sound. Not at all. We assume that the prices we observe are dravm from a distribution of prices in the early Republic). in survived for two millennia as perverse to insist that any pattern Yet another objection we assume the result appears to we regard the few prices that have There does not seem to be a reason grounds that The comes Roman Empire (and late We do not assume they are accurate or come from a particular kind of investigation or a particular time of year (as in the previous paragraph). In fact, 18 we only assume prices, the t-statistic tells us There Given that they are prices. is no more is we whether there circularity here than in Another objection sample that is that are sampling a relationship between price and distance. is any the sample from the population of wheat statistical test is tiny, of a hypothesis. only six price pairs. This small As we unfortunate, but no barrier to the test of our hypothesis. said above, the standard errors and t-statistics are corrected for degrees of freedom. Having few observations test. makes it easier to reject hypotheses, but it does not affect the validity of the We would of course like to have many more prices, but there are not more to be found at this time. to find more Perhaps our analysis of these few prices will stimulate other historians price pairs and to provide more evidence for or against our hypothesis. A little thought experiment might be usefiil here. Imagine that steady sfream of new wheat prices from various locations. If they regression line in Figure 2, they provide not lie more evidence we discover all lie a near the for our hj^othesis. Prices that near the regression line indicate that the locations in question were isolated from the general Mediterranean wheat market or the price observations. If we only find a main currency area few such observations, historical record without vitiating the hypothesis found a lot that at the times of the would enrich the of a generally unified market. If we of such prices, that would suggest that conditions where markets were isolated were more prevalent than times when they were integrated. The Roman world in that case would be monetized, but not composed of a unified monetary system. We remarked earlier that imperial officials often intervened in the market for wheat. If the government was administering prices, then the prices might not be the result do of market forces at all. To check this possibility, 19 we searched for price interventions in price spikes Rome. The government repeatedly attempted when suppHes offset the loss BCE, ran short. In 74 of wheat in Sicilian floods. In 57 the to avoid the hardships government sold grain cheaply HS merchants two least two HS 1 9 CE Tiberius placed HS per modhis, a price ceiling It also is clear, even intermittent. If we summarized wheat market from time were still own in most years. list, set another It is clear that under that these interventions that these interventions are only half of the actual clearly a minority. The market for which there were wheat was allowed to work on In addition, if traders expected the government to interfere famine loomed, they might have been discouraged from trying adversity. compensate to time, particularly from what must be a partial assume CE Nero in Table 3. actions, the others being unrecorded in our sources, the years in interventions to HS per modius}^ interventions like these are the government intervened in the Augustus. Augustus gave was temporarily on grain and offered above the price he thought people could bear. In 64 Government its BCE suggesting that the price before his intervention was at price ceiling for wheat, this time at three were 24 apiece to 250,000 people, allowing them to purchase wheat that expensive. In to BCE Pompey negotiated extra purchases himself, sailing from province to province in search of wheat. In 400 of to comer when the market in Government intervention therefore may have damped speculation and made the underlying pattern of prices easier to see. The largest government activity in the government gave 60 modii per year households receiving to wheat market was the annona. The each male head of a household. The number of this largess is unclear, but it is generally thought to be between 200,000 and 250,000 during the reign of Augustus.'^ If the population of Rome was 20 around one million people, the annona used between half and a quarter of the wheat imported into Rome. More than half of the wheat imported to Augustus therefore was imported into Rome for the Rome at the privately. Sirks argued that the share annona was even smaller, only around 15 percent, time of of grain imported making the private share correspondingly larger. The government also obtained the wheat for the annona privately. They let contracts to societates to provide wheat, and they offered inducements for private merchants to participate in their own this process. Claudius rewarded private merchants ships, carrying at least 10,000 modii, to for five years. If the merchant was a Poppaea, which penalized the citizenship. male tutor. activity. shippers And if the merchant was not a citizen, he who would Rome annona from compulsory consequently was a mixture of public and private to provide work for many merchants and gather wheat from the far comers of the Mediterranean. But the means that the price of wheat in Rome We have discussed price variation already, stabilized prices in to the authorities. There was enough private activity public presence annona. in more than they destabilized them. may have been distorted by the and government actions probably The presence of so much Rome however may have decreased the price of wheat in that city at price annona Hadrian extended these rewards by exempting any imposed by municipal The wheat market for the merchant was a woman, she could make a merchant devoting the greater part of his resources services Rome he would be exempt from the lex Papia childless. If the will without the intervention of a would be granted citizen, import grain to who used was received by all free wheat times. If this importers, then the graphs and regressions record the proper 21 between relation Roman and provincial from the price paid move to importers, this the line in Figure 2 If the price prices. would change up or down. It would not we observe was the constant in the regression wheat from different places Coming back Roman to for monetary conditions, there is is annona paid no evidence at all). evidence of inflation in the early We used the price difference between Roman and provincial prices in order to avoid problems of inflation. inflation in Rome. While in prices, it is more As shown the incomplete evidence likely that they were Turning to the other variable, distance, distance is on in Table 1, we take account of inflation suggests there drifting any case, the use of a price difference insulates our to which there 2 or Empire. The pay of soldiers was increased in infrequent large jumps, and wheat prices in Egypt rose. jumps — and affect the relationship in Table the slope of the line in Figure 2 (unless of course the officials of the different prices for different test upward more or less steadily. In from an inflationary we acknowledge were a few bias. that a straight-line only a rough approximation to the actual distance traveled by wheat on Rome. This defect of the data however is an advantage for our test; it its way biases our test toward rejecting our hypothesis. The approximate nature of the distance estimate introduces another kind of error in the variables. Since this the effect is slightly different from the in observing distance has the effect is large, as can be seen in Figure effect voyage our independent variable, of an error in observing the price. Any error of reducing the size of the resulting coefficient. Since 2, straight-line distances representation of the comparative distance from specific is to the capital city. 22 Rome appear to give a good despite the waywardness of any Finally, this is a very simple model of Roman monetization and trade. We have argued that there was a single monetary system and a single wheat market across the whole Mediterranean. degrees of freedom. test? We tested this hypothesis with a simple regression and few Why should any ancient historian believe such a simple model and We argue that the purpose of a model is to provide an overall view of money and trade in Rome. It does not explain every detail; instead it provides an overview that can help our thinking. There are only a few observations because that studied more recent times, ancient history. As we historians to search for to see if this we would have more said above, data, but we hope that added price observations. simple model survives a test we do is all we have. the best we If we can with our exploration will stimulate ancient We will be as interested as any reader with more data. Conclusion We have argued here that the early Roman Empire was thoroughly monetized. We do not argue that people everywhere had adequate supplies of Roman coins, but rather that people throughout the empire used a single activities. monetary standard This single monetary standard was based on sestertii in the western empire and on drachmae in the eastern empire, with a fixed exchange result was to value their to create a single currency area like the euro rate zone today. Whether or not regions had adequate supplies of coin, the survival of prices from empire indicate that the between them. The all all comers of the Roman economy was thoroughly monetized. We argued also that this monetization set the conditions for market integration by reducing the transactions cost of trading across large distances. This allowed a single 23 market for wheat enormous capital the Roman emerge, whose existence we could verify from surviving of Rome also encouraged the growth of trade, since size needed to to eat. conquest. unified market. Food must have The all prices. The the residents of the traveled around the Mediterranean for eons before quantities shipped were too Only when the Romans imposed small, however, to a political settlement make a on the area and created a unified monetary system could trade expand enough to unify prices across the Mediterranean. 24 Table Distance and Prices for Grain 1: Rome Distance (km) from Region Price (HS) Province Distance-from- Price (HS) Rome Rome "Discount" (HS) 4.00 Sicily {Sicilia ' HS' -1.50 HS"* 1 1363 {Italia 77BCE 2.00-3.00 HS' 3.00-4.00 Spain {Lusitania province) Po valley HS 427 province) Year 150 BCE 150 BCE -2.50 3.00-4.00 HS 5.00-6.00 HS' 0.5 HS province), by 1510 river Asia Minor (city -3.00 HS' ofPisidian 1724 Antioch Egypt (Region of the -3.13 5.00-6.00 Fayum) 4HS' HS 1953 5.00-6.00 ^ HS 2.00-2.50 HS^ Rickman Corn Supply, 53-4 (1925), 126b. ^P. , M ich. II 20 EC - 56 CE -4.00 2298 Palestine Sources: 80sCE 2.00-2.25 1271.1.8 -38 ^ " Polyb. Frank, 2. 15 " CE Cicero, 2 Verr. 3. 189. 'Polyb. 34 .S.l.^Apigr. ESAR iv, 25 15 -3.25 181 and 183. Table 2: Distance Discount Regression Results a B R' -1.10 -0.001 0.79 (2.2) (3.9) -1.16 -0.001 (2.0) (3.4) -6.14 -1.25 (2.8) (4.1) -6.04 -1.25 (2.7) (4.1) TV 6 Distance Discount 5 Distance Discount (no Po 0.79 valley) 6 Log Distance 0.81 Discount 5 Log Distance 0.85 Discount (no Po valley) Source: Table 1. Absolute value of t-statistics below coefficients. 26 Table 3: Selected Date 138 BC Government Interventions Grain Market in the Type Intervention Source Rising prices lead tribunes to seek extra grain Obsequens22(142). supplies. 100 BC Feared shortage leads Senate to seek extra grain stock. M. H. Crawford, Roman Republican Coinage: Volume 2 (Cambridge: CUP, 1974), 74 and 616. 75 BC 1 Vi modii distributed free per man Cicero, Plane. 64. Cicero, 2 Veir, 3.215. given shortage. BC 66 BC 74 Aedile distributes grain Pompey tours at 1 AS Sicily, Africa, nmy,Histnat. per modius. and Sardinia to 18.16. Cicero, Imp. Pomp., 34. secure extra grain in his capacity as grain commissioner. 62 BC Cato's Lex Porcia raises grain outlay to 30 million HS 58-56 Cicero appoints BC falls. 49 BC Plutarch, much to the budget. Pompey for grain supply, price 29 23 BC BC BC BC BC 11 BC 5BC 2BC Caesar distributes grain to starving Romans Augustus gives 400 Augustus gives 400 HS HS to 250,000 people. many. Augustus gives grain to Augustus gives grain to at least 100,000. 19 AD 10-12, 14-18; ^«. 4.1; Cassius Augustus gives 400 Augustus gives 240 Augustus gives 240 HS HS HS Bell. civ. 2.48; Res gest., 15. Res gest., 15. Res gest., 5. Resgest, 14.7.3; 250,000 people. Resgest., 15. to 320,000 people. Resgest., 15. to 200,000 people. Resgest., 15. Augustus gives grain to many. Also expels some foreigners from the city to alleviate the crisis. (Gamsey, 221) Suetonius, Tib. 8. 18. to Tiberius imposes price ceiling, gives dealers Fam. Cassius Dio 41.16.1 Resgest., 15. 250,000 people. Augustus give money and "12 rations" to 250,000 people. Tiberius also helps, and Suetonius says he "skillfully regulated the difficulties of the grain supply and relieved the scarcity of grain at Ostia and in the city." 18 AD Dom. Cicero, Att. 7.9.2, 4; 9.9.4; Appian, to 22 6 Cicero, Dio, 39.9.3, 24.1; Cicero, Q.fr. 2.5; Har. resp. 31; Plutarch, Pomp, 49.4-50.2 during the civil war (Gamsey, 202). 24 Ca/o M/w. 26.1 or adds that Cassius Dio 55.22.3. +2 Tacitus, y4/7«., 2.87. HS 51 64 AD AD AD 189 Claudius encourages merchants to (Gamsey, 223) Nero fixes price Commodus at 3 engages sail in winter. Tacitus, ^««., 12.43. HS, annona suspended Tacitus, Ann., 15.39. in price-fixing. Herodian 1.12.2-4; Cassius Dio 72.13.2. 27 Figure 1 Plot of Distance : and Roman Distance Discount n U -0.5 w X -1 Palermo (Sicily) ;E-1.5 c -2 i Madrid (Lusitania) Q -2.5 - -3 - -3.5 - -4 - -4 R ~H.D - 0) c Palestine Po valley * ^ w 5 Turkey (Pisidian Antioch) Egypt 1 500 1 1000 Distance from 28 (the Fayum) 1 1 f 1500 2000 2500 Rome (in km) Figure 2: Relationship between Distance and Roman Distance Discount - -0.5 . w X - . -1 - F^lermo -1.5 ^^^ ^"*^«^^ 4.J c 3 o u (Sicily) - ^^ -2 *»s^.^ Madrid (Lusitania) (0 Q ^^*>^^^» -2.5 ^*"^>^_^_^ 0) u -3 Po valley Palestine *"^«>,,,^ ns ^ **«v^_^ w -3.5 - Turkey (Rsidi^hs,,,^^ ^*^*^ Antioch) Q -4 -4.5- Egypt (the Fayi ' 1 500 1 1 1 1000 Distance from 29 1 1500 Rome (in 2000 km) 2500 Endnotes ' Sometimes a third function, a store of value, is added, but this is not relevant here. Kevin Greene, The Archaeology of the Roman Economy (Berkeley, CA: University of California Press, 1986), pp. 50-51. This assertion can be found in almost any elementary economics ^ text. Roman Empire Keith Hopkins, "Taxes and Trade in the (200 B.C.-A.D. 400)," Journal of Roman Studies 70 (1980), 101-125; Richard Duncan- Jones, "The Denarii of Septimus Severus and the Mobility of Roman Coin," Numismatic Chronicle, 161 (2001), 75-89. ^ Richard Duncan-Jones, The Economy of the Roman Empire: Second Edition (Cambridge: Cambridge University Quantitative Studies Press, 1982); Hans-Joachim Drexhage, Preise, Mieten/Pachten, Kosten und Lohne im romischen Agypten bis Regierungsantritt Diokletians (St. Katharinen: Scripta Mercaturae, 1991); Dominic Rathbone, "Prices and Price Formation in formation des prix dans les zum Roman Egypt," Economic antique. Prix et economies antiques (Saint-Bertrand-de-Comminges: Musee archeologique departmental, 1997), 183-244. David M. Schaps, The Invention of Coinage and the Monetization ofAncient Greece (Ann Arbor, MI: University of Michigan Press, 2004). ^ Rathbone, "Prices," pp. 191-92. ^ Geoffrey Rickman, The Corn Supply ofAncient Rome (Oxford: Oxford University Press, 1980), pp. 143-55. ' Duncan- Jones, The Economy of the Roman Empire. ^ Kevin Greene, The Archaeology of the Roman Empire (Berkeley: University of Cahfomia Press, 1986), p. 40. 30 ^ William Mitchell Ramsay, "Studies in the 'Roman Province Galatia' VI Inscriptions of Colonial Caesarea Antiochea," Journal 172-205, — Some of Roman Studies 14 (1924), pp. at 180. '°Rathbone, "Prices," 193. '^ Duncan- Jones, Economy of the Roman Empire, 372. For information about transport within Egypt, although not its cost, see Adam Biilow-Jacobsen, "The Traffic on the Road and the Provisioning of the Station," in La route de Myos Hormos, edited by Helene Cuvigny (Le Caire: ^^ Institute fran9aise d'archeology orientale, 2003), Vol 2, pp. 399-426. Tenney Frank, An Economic Survey ofAncient Rome (Paterson, NJ: Pageant Books, 1959), Vol. 4 (Heichelheim), 181-83. '^ Gamsey, that oil for example, observed casually and wine cost more Food in in Rome and without apparent need for documentation than elsewhere. Peter Gamsey, Cities, Peasants and Classical Antiquity (Cambridge: Cambridge University Press, 1998), 241. Schaps, Invention of Coinage. ^^ Helene Cuvigny, "The Amount of Wages Paid to the Quarry- Workers at Mons Claudianus," Journal of Roman Studies, 86 (1996), 139-45. The sizes number of the coefficients typically coefficient, but description 17 In the when we use logarithms since the logarithm of a smaller than the number. T-statistics are the same sign as the we have reversed the more more is rise signs of the t-statistics in Table 2 to make the intuitive. precise language normally used for regressions, the probability of observing the coefficients in the table wheat and the distance from Rome is if there were no relationship between the price of less than five percent in three out 31 of four regressions and close to that probability in the fourth. statistic for four degrees of freedom (six observations) The is five percent value of the 2.8; for three degrees t- of fi"eedom (five observations), 3.2. Higher t-statistics indicate lower probabilities that the observed relationship '^ Peter is Gamsey, Famine and Food Supply Cambridge University ' a in the Press, 1988) 195-222; Graeco-Roman World (Cambridge: Rickman, Corn Supply, 150-54. Catherine Virlouvet, Tessera frumentaria: les procedures de distribution du ble public la fin de la Republic et Gamsey, ^° the result of chance. Cities, Boudewijn that the au debut de I 'Empire (Rome: Ecole Fran9aise de Rome, 1995); Peasants and Food, 236. Sirks, Foodfor Rome (Amsterdam: annona provided bare subsistence less than half of total Gieben, 1991), 21 . Jongman argues for half the free population of Rome, that wheat imports. Willem M. Jongman, "Consumption Roman Empire," in Richard P. Sailer, Ian Morris and Walter Scheidel is, in the early (eds.), The Cambridge Economic History of the Greco-Roman World (Cambridge: Cambridge University Press, 2006). ^' E. Badian, Publicans and Food Supply, 234; ^^ and Sinners (Oxford: Sirks, Foodfor Rome, George R. Watson, The Roman Soldier Basil Blackwell, 1972); Gamsey, Famine 63. (Itahaca, NY: Cornell University Press, 1969), 89-92; Rathbone, "Prices." ^^ Peter Temin, "Mediterranean Trade in Biblical Times," in Eli F. Heckscher: A Celebratory Symposium, edited by Ronald Findlay and Mats Lindahl (Cambridge: Press, forthcoming). ^^836 27 32 MIT Date Due MIT LIBRARIES 3 9080 02618 3068