Ten
Local Interest Story: Political Power and Regional Differences in the Shandong Capital Market,  1900–1937

Kenneth Pomeranz

Many issues in the early twentieth-century Chinese economy hinge on the capital market. The regional capital market is central not only to the relationship between richer regions, poorer regions, and economic change but also to an understanding of the extent to which the wealth of some groups and the poverty of others was the result of power and exploitation, rather than just differences in luck and skill. And in China, where several currencies, many with limited acceptance, circulated, any analysis of supra-local capital markets must also consider domestic currency markets. Looking at currency trading raises an often overlooked set of questions concerning relationships between state and economy in early twentieth-century China.

This paper examines capital markets in one province—Shandong—between the Boxer Uprising and World War II (1900–1937). Since some but not all parts of the province experienced rapid growth in these years, we must first explain why, as interest rates declined in more prosperous areas, funds did not flow to the more "backward" regions, in which credit commanded a higher price. It is not new to note that the links between rapidly growing regions and other parts of China were weak. Much of the literature on pre-World War II China argues that the rapid growth of those areas that were stimulated by the international economy did little for the rest of the country.^[1] But it is hard to prove that money made in China's "advanced" regions was rarely invested in the hinterland. Moreover, a number of scholars using aggregate national data now argue that there was an increasingly national

[1] For reasons of brevity, a review of the literature on various controversies mentioned here has been omitted. These references are available in my Ph.D. dissertation, "The Making of a Hinterland: State, Society, and Economy in Inland North China, 1900–1937" (Yale University, 1988), chap. 1, nn. 1, 2, 5, 7, and 94.

capital market, in which coastal funds were invested in the hinterlands. This position rests on three points. First, investors seek high returns, and institutional and technological changes were making information about far-flung opportunities cheaper. Second, regional differences in interest rates were narrowing. Third, without such investment, the unbalanced trade between coastal cities and the countryside (Shanghai, for instance, sold far more than it bought) should have led to money shortages and deflation in the countryside, causing a divergence between urban and rural prices; no such divergence has been observed.

Local credit markets are also controversial. Since North China had relatively little tenancy, discussions of economic exploitation there have centered on credit and marketing. Although numerous anecdotes show money-lenders abusing local monopolies, some of the aggregate data suggest that the cheaper funds available in Shanghai, Tianjin, and overseas made their way, through intermediaries, to enough outlets to create competitive rural credit markets.

Finally, recent research has highlighted both the development of national markets in prewar China and surprising continuities between Chinese state making before and after 1949. However, relationships between these two trends have received little attention. A provincial case study cannot resolve all these issues—especially since we still lack information from below the county level^[2] —but it can help us see how credit markets worked, and highlight important relationships between state making and market making.

I began by borrowing a technique used by two economic historians to estimate interest rates for medieval Europe. Using the estimates derived for each of Shandong's 12 prefectures, I did simple tests for market integration and found a surprising pattern. Not only did the province's money market not show the integration economists would likely predict where transport and information costs are low; the lines across which prices suddenly jumped did not correspond to the physiographic regions at whose boundaries most historians would expect trade networks to break down. Eventually, an investigation of traditional archival and published sources showed that this peculiar market segmentation could be traced to deliberate manipulation by certain local elites and local governments. To be more precise, the evidence suggests the following.

1. The province included three large and distinct regional capital markets (see Map 10.1). Interest rates in the poorest market (Southwest Shandong) were about 1.5 percent per month higher than in the most prosperous region (the North Coast) and approximately 0.6 percent higher than in the middle, or Heartland.

[2] See, e.g., "Shandong Gaomi Wei xian zhi nongcun jiehuo," Gong shang banyuekan 6, no. 4 (Feb. 15, 1934): 49.

Map 10.1.
Approximate Map of Shandong's Regional Capital Markets, 1911–1937.

2. The interest-rate differentials changed little between 1900–1911 and the 1930s.

3. These lasting differences were largely due to the ability of politically powerful people to restrict the flow of money in and out of their own county and to manipulate local silver-copper exchange rates while limiting others' access to the cheaper credit and silver available in more prosperous areas.

4. The artificial barriers kept the return on any hoardable or loanable asset (such as grain or money) higher in poor areas than elsewhere. Moreover, politically connected people who could cross the barriers made much greater currency-trading profits than those available either in purely local commerce or in commodities governed by national and international markets.

5. Divergence in commodity prices between the North Coast and the other regions was limited because the Heartland exported enough cash crops to avoid a silver drain and the Southwest, with fewer exports, bought few modern goods and exported laborers, who brought silver home. When Japan seized Manchuria in 1931 and curtailed migratory labor, the Southwest experienced a local deflation beyond that caused by the world depression.

6. While county governments most often obstructed currency and capital

markets, provincial officials were also caught between the advantages and disadvantages they expected from Shandong's integration into a larger economy.

7. The truncated money markets in Shandong, during a period in which many other markets became more integrated, show how politics shaped and limited market participation. These limits also handicapped governments trying to move resources, affecting state making as well as market making.

Shandong's Regional Capital Markets

We have numerous interest rates from particular times and places, but most cannot be used in systematic comparisons because they omit crucial information, such as who was charged this rate or what security was put up.^[3] However, we do have useful county-by-county figures for 1933.^[4] For 1900–1911 we can estimate interest rates using monthly grain prices and patterns of post-harvest price increases in each prefecture of Shandong, as Donald N. McCloskey and John Nash did for medieval England.^[5] Over the long haul, holding grain a few months before selling it will be no more or less profitable than selling the grain immediately and holding money; otherwise, more people would turn to the more profitable activity, until it ceased being so. Thus, the average rise in grain prices after the harvest should approximate the interest rate plus other storage costs (including safeguards against theft, and provisions for losses from rotting or rats).

This method involves looking at pairs of prices a month or a few months apart. For technical reasons, the most useful quotations are the high prices from each prefecture, recorded in the prefectural capitals.^[6] Ignoring pairs that end in or pass through harvest months—when prices reflected the value of the new grain rather than the accrued value of whatever little grain remained from the previous year—leaves 50 to 60 pairs for each of six grains in the 11 prefectures with reliable data. In ten, the results for the different grains are quite close, and so more trustworthy.^[7] We do not know how much

[3] E.g., Shandong zheng su shichaji , 1934, and Shina shobetsu zenshi (hereafter SSZ ) have many rates but no supporting details; the same is true of numerous magazine articles, gazetteers, memoirs, and so forth.

[4] Zhongguo Shiyebu, Guoji Maoyiju, Zhongguo shiye zhi: Shandong sheng (hereafter ZSZS ; Chinese industrial handbook: Shandong Province; Nanjing, 1934), pt. 5, pp. 91–97; for a discussion of these data, see Pomeranz, "Making of a Hinterland," app. B.

[5] Donald N. McCloskey and John Nash, "Corn at Interest: The Extent and Cost of Grain Storage in Medieval England," American Economic Review 74, no. 1 (Mar. 1984): 174–87; this method is discussed on pp. 178–83.

[6] The papers in Part One of this volume provide detailed explanations of the Qing system of grain price reporting.

[7] The prefectures that do not conform to this pattern are discussed in Pomeranz, "Making of a Hinterland," chap. 1, n. 12, and app. A.

of the average price increase represents interest, but other storage costs were similar across prefectures.^[8] Thus regional differences should reflect relative interest rates. The results are shown in Table 10.1.

The interest rates fall into three regional groupings; the regions are demarcated on the map. The North Coast area, which consisted of four prefectures, included the treaty ports of Yantai and Weihaiwei and had roughly 6.8 million people. With unusually strong ties to booming Manchuria, right across the Bohai Gulf, and access to the central and South China coastal regions, Shandong's North Coast had a privileged position, which gave it relatively easy access to hard currency (to be discussed later) and particularly low interest rates. Holding grain there earned less than 0.5 percent per month. At the other extreme is the second region: the Southwest, the prefectures of Yanzhou, Caozhou, and Jining, with about 6.4 million people.^[9] Returns there averaged 2 percent per month.

In the remaining prefectures, ranging from Yizhou (1.2 percent) to Dongchang (1.7 percent), returns averaged 1.4 percent. This largest region, the Heartland, with 21 million people, presents the least uniform picture but still appears to have been a meaningful capital market. Moreover, these prefectures are clearly separate from the other two groups, and traded with each other much more than with those areas. They were far more involved in outside trade than the Southwest was, and more a part of the North China regional economy than the North Coast, which was instead closely tied to Manchuria.

Rates derived from 1903–11 grain prices in neighboring prefectures of Zhili Province echo this pattern.^[10] Coastal Zunhua, closely tied to Manchuria, had rates like those in Shandong's North Coast (under 0.8 percent); Daming, adjoining Southwest Shandong, had rates like those in the Southwest (over 2 percent); three prefectures bordering the Shandong Heartland resembled that zone (rates were slightly over 1 percent).

Measures of dispersion for the Shandong results confirm the regional pattern. Within each individual prefecture except Laizhou, the returns from storing different grains are closely bunched around the prefectural mean for all grains. Entry and exit from local markets, storage activities, and switching between particular assets must have been fairly easy; as we shall see, this was not true of interregional money markets. To the extent that these grains are a random sample of all hoardable assets, the results suggest that in any

[8] The kinds of storage facilities in different parts of the province were the same, and inland areas were drier. Banditry was more serious in the Southwest than elsewhere, but in the hundreds of crime reports I have reviewed, stored grain was almost never a target (grain on the road to market sometimes was). The rat population may have been unevenly distributed, but I know of no reason to think so.

[9] There are a number of problems with the Jining data, which are discussed in Pomeranz, "Making of a Hinterland," chap. 1, n. 15, and app. A.

[10] See ibid., chap. 1, n. 17, and app. A.

given prefecture, there is a two-thirds probability that the average monthly return on any asset over these years would be within 0.2 percent of the mean return for all assets in that prefecture.^[11]

On the other hand, specifying a grain without knowing what prefecture it was stored in has little predictive value. Thus, the concept of a functioning market in savings within each prefecture has predictive power, while the idea of a provincewide market in the storage of any particular grain—much less all grains—does not. The rates of return for different grains within each of the three regions outlined here are also tightly bunched around the regional mean for all grains, though not quite as tightly bunched as within individual prefectures. There are differences in rates of return between different prefectures in the same region, but they are much smaller than the changes that we find if we cross regional boundaries.

These divisions changed very little during the 1900–1937 period. A 1933 provincial survey presents the "highest" and "ordinary" interest rates charged by individuals, pawnshops, cooperatives, and "stores" (shang dian )—many of which also acted as local banks—in each county. The ordinary rates quoted for stores are probably the more reliable and more important. North Coast store rates averaged 1.95 percent; Southwest ones, 3.5 percent; and Heartland rates, 2.5 percent. In loans from individuals, the Southwest-Heartland difference is much smaller; however, it was probably easier to get an accurate picture of rates charged by stores than by the huge number of individual lenders.^[12]

While the pattern of interest rates in 1933 resembles that derived for 1900–1911, the rates themselves are higher. This need not mean that real interest rates rose. The 1900–1911 figures represent what one earned simply by transferring grain into the future, without it ever leaving one's courtyard. Consequently, these rates should reflect only minimal compensation for risk. Nor would they reflect differences in bargaining power between lenders and possibly desperate borrowers. Thus, these rates should resemble those on the absolutely safest short-term money loans. Our limited evidence confirms this: the safest loans in Laizhou (North Coast) cost 0.4 percent per month, loans from trade associations to their member firms cost 0.6 percent in Jinan (Heartland), and a loan in the Southwest—the lowest rate I have found for the region—cost 1.5 percent.^[13] The 1933 rates should include larger risk premiums and, after the 1910s and 1920s, may also include anticipated inflation. These higher premiums would make the rates on all the loans in the 1933 data higher than those inferred from the 1900–1911 data. What matters

[11] See ibid., app. A, for complete figures.

[12] See ibid., app. B, for a discussion of these data and a breakdown of the results.

[13] SSZ , 4:1125, 1156; Shandong quansheng caizheng shuomingshu , p. 28.

for our purposes, however, is that the same pattern of regional differences persists, suggesting that similar nonmarket forces operated in both periods.

Other Signs of Regional Distinctness

The northeasternmost part of Shandong was largely cut off from the rest of the province by mountains, making trade in bulky products difficult and making the area less a part of Shandong's economy than Manchuria's. For instance, the North Coast produced only half the grain it needed; the rest came across the Bohai Gulf from Manchuria. People crossed the gulf too; until World War I, this area accounted for almost all Shandong settlers in Manchuria, while Yantai, the North Coast's largest port, handled most of Shandong's trade with Manchuria. Moreover, huge numbers of people from this area worked in Manchuria either seasonally or for several years.^[14]

However, the North Coast region outlined by our statistics extends westward beyond this extreme northeastern part of the peninsula. It includes all the prefectures along the Bohai Gulf, including areas easily reached from central and western Shandong. This broader definition of the North Coast is suggested by logic as well as statistics: if access to the North Coast provided superior opportunities, people on the western part of that coast would try to join in, whether or not they were cut off from the Heartland. The question is why more of the Heartland did not become involved.

The broader definition of the North Coast is also reflected in a set of rates charged in 1911 for money transfers from Huangxian, a major banking center.^[15] The ten listed destinations stretch throughout our expanded North Coast. The furthest, Xincheng, was 250 kilometers from Huangxian, but only 16 perfectly flat kilometers from the important Heartland trading center of Zhoucun and 120 flat kilometers from Jinan, Shandong's capital and largest city. However, the only destination outside the North Coast listed was foreign-controlled Qingdao. That a network with such modest costs—less than 0.5 percent of the shipment for the most expensive destination—would suddenly stop short of important markets strongly suggests that the boundaries of this trading area were set by political rather than strictly geographic factors.

Various analysts have also seen a distinct southwestern zone only loosely linked to the rest of Shandong. Its boundaries, however, are disputed, and are based on social differences—number of gentry, incidence of violence, etc.—rather than physical ones.^[16] No insuperable geographic barriers sepa-

[14] Thomas Gottschang, "Migration from North China to Manchuria, 1891–1942: An Economic History" (Ph.D. diss., University of Michigan, 1982), pp. 77–81.

[15] SSZ , pp. 1145–47.

[16] Joseph Esherick, Origins of the Boxer Uprising (Berkeley and Los Angeles, 1987), pp. 12–14; Sun Jingzhi, Huabei de jingji dili (Beijing, 1958), pp. 132–34; Shandong minzhong jiaoyu yuekan 5, no. 4 (May 1934), pp. 85–88.

rate the Southwest, however defined, from the Heartland.^[17] After 1912, Jining and Yanzhou were also linked to Jinan by rail.^[18]

Different Market, Same Map. Regional Currency Markets

The political underpinnings of market boundaries become clear when we turn to the currency market. Like all of China, Shandong used numerous silver, copper, brass, and paper currencies during this period, with no fixed exchange rates between them.^[19] In most of China, silver and silver-denominated paper were increasingly dominant,^[20] but in Shandong copper remained important; consequently, so did the copper-silver market.

According to a 1933 study, copper was used for agricultural sales in five of 12 Shandong places surveyed and was the only currency farmers received in four locations. Farmers made at least some purchases with copper in eight of 13 locations. People often received loans in copper but, except in one place, had to repay in silver.^[21] Prices for large items were generally quoted in silver, and taxes were always set in silver.^[22]

In a unified provincial capital market, silver-copper exchange rates in different counties should have converged or at least followed parallel paths. In fact, however, exchange rates varied wildly from county to county, and trends were often far from parallel. Shanghai rates may be considered "national" prices, though Tianjin, Beijing, and Hankou rates did not always exactly track Shanghai's.^[23] Prices in Yantai and other North Coast ports closely tracked Shanghai rates,^[24] confirming this area's membership in a national, or perhaps littoral, economy. However, inland areas are another story.

Even within the Heartland, there were significant differences between some nearby counties, such as Jinan and Taian. In the early 1920s rates in Linqing and Qingping, two adjacent Heartland counties, diverged by as

[17] For a brief survey of Shandong geography in English, see Esherick, Boxers , pp. 11–20; for greater detail, see Hou Renzhi, ed., Xu tianxia junguo libingshu, Shandong zhi bu (Beijing, 1940), pp. 137–216.

[18] Zhongguo Jiaotongbu, Jiaotong Shi Weiyuanhui, Jiaotong shi (1937), pp. 3531–35.

[19] Li Guijin, "Qingmo bizhi gaige ji qi shibai yuanyin de qiantan," in Jingji shi , 1984, no. 2, pp. 129–30, has a brief summary of the late Qing situation; by 1919, the North-China Herald (hereafter NCH ), Oct. 18, 1919: 172, had counted 115 new kinds of coinage since 1912 and untold varieties of paper.

[20] Leonard Hsu, Silver and Prices in China (Shanghai, 1935), pp. 68, 75.

[21] Ibid., pp. 68, 75.

[22] Ibid., p. 75; SSZ , pp. 260–61.

[23] See, e.g., Tongji yuebao , Dec. 1929: 25; Dec. 1930: 52.

[24] Hsu, Silver and Prices , p. 85; Chintao Shubigun Minseibu Tetsudobu, Tohoku Santo (Bokkai Santo engan shoko Iken Chifu kan toshi) chosa hokokusho (Chosa Shiryo no. 17; Qingdao, 1919) pp. 54 (Yangjiagou), 148 (Ye County), 252 (Longkou), 286 (Huang County), 317 (Penglai).

much as 33 percent. However, after 1926 the difference hovered between 0 percent and 7 percent, only slightly above the cost of shipping copper coins between the two places.^[25] Except for in Jinan, rates within the Heartland rarely differed by more than 20 percent, and sharply opposed trends do not last long.

Moreover, until 1927, trends in the Heartland (again, excepting Jinan) were similar to those in Shanghai and Yantai. After that, however, littoral rates leveled off, while those in the Shandong Heartland kept climbing. By 1933–34 silver cost about 75 percent more copper in Qingping and Linqing (the only Heartland rates available for those years) than in Shanghai.^[26] In most ways, North China was less chaotic after the spring of 1928. However, since it was governments and their allies that kept the currency market divided, it is logical that with increased political stability, littoral and Heartland rates would diverge.

Evidence is scant, but it appears that the Shandong government's growing strength after 1927 allowed it to reclaim some North Coast areas, causing rates there to diverge from Shanghai's and resemble the Heartland's. A July, 1928 quotation from Longkou, a small North Coast port, is very close to rates in Jinan, Jiaozhou, Linqing, and Qingping but 27 percent above those in Shanghai.^[27] However, Yantai, the largest North Coast port, escaped government control and stayed part of the littoral economy. Even three years after Japan seized Manchuria, illegal silver shipments from Yantai to Manchuria affected the local money supply enough to force the government to print more notes; this arbitrage kept the difference in silver-copper rates between Yantai and Dalian down to 8 percent,^[28] far less than the North Coast—Heartland difference. Provincial state making expanded the Heartland zone, but there were limits.

The most dramatic differences, however, are between Heartland and Southwest rates. My most complete set of exchange-rate data is for Jining, the commercial center of the Southwest;^[29] those data resemble nothing for the North Coast or Heartland. Jining silver prices were ordinary in 1904 but

[25] Sources and calculations are in Pomeranz, "Making of a Hinterland," chap. 1, n. 40, and app. C.

[26] Xuxiu Qingping xianzhi (1936), jingji : 25a–b; Linqing xianzhi (1934), jingji : pp. 22b–24a. The Shanghai data were created using the annual index numbers for copper and silver yuan exchange rates in Shanghai from Tongji yuebao , Dec. 1930, p. 51, and prices for selected dates in Hsu, Silver and Prices , p. 85 (for the dates at which the two series both have data, the rates coincide perfectly).

[27] Quotation from Chinese Economic Bulletin 13, no. 3 (July 21, 1928), in Records of Former German and Japanese Embassies and Consulates, 1890–1945 (hereafter Records ), reel 5, p. 4618834.

[28] Quotation from Qingdao Times , Dec. 7, 1934, in Records , reel 5, p. 4619093.

[29] The Jining data may be found in Hai Shan, pp. 48–78. Sources for the other conversion rates appear in Pomeranz, "Making of a Hinterland," app. C.

then soared. By 1909, Jining silver commanded 40 percent more copper than in Jinan and 60 percent more than in Shanghai. By 1914, it was roughly double the price in both Taian (Heartland) and Shanghai; by 1921, almost triple; in 1929, almost six times Linqing and Qingping rates and about eight times the Shanghai rate; and in 1933, roughly double Qingping and triple Shanghai rates.^[30]

Transport costs cannot explain these differences, since Jining was only 100 kilometers by rail from Heartland Taian. Nor can information problems. Currency markets in late Qing Jinan, Wei County, and Yantai generated publicly available daily prices; the same was true in early Republican Jining.^[31] By this time, other commercial information regularly reached even fairly remote counties by telegraph. All the pieces were in place for lucrative arbitrage.

The Politics of Economic Geography: Public Finance

There was, however, a major obstacle to currency movements: numerous county governments banned the export of more than small amounts of certain currencies (usually silver, sometimes copper).^[32] This obstructed both arbitrage and cross-county lending: if it is hard to get hard currency out of a county, it is not attractive to send money in. In setting up these barriers, county governments increased the profits of certain local money shops and, more important, their own revenues. Although peasants were often paid in copper, taxes were set in silver. In the 1800s, the exchange rate used for tax payments varied widely, in many places exceeding 5,000 copper jing qian per silver tael (a 1,800-to-1 rate in the units used here), but by the 1890s the market rate in cities along the seacoast was less than half that. In 1896 the Shandong governor tried to standardize the official rate at 4,800 to 1, still roughly double the Tianjin rate.^[33] If county governments or their allies could enforce high rates at tax-collection time and then dispose of their copper at market rates, they would double their real revenues. As market rates shifted in favor of silver after 1905, county governments' arbitrage profits declined, but they remained significant.

High exchange rates were imposed on all direct taxes but did not enhance

[30] Ibid., app. C and chart, pp. 52–55.

[31] Shandong quansheng caizheng shuomingshu , p. 28; Chintao Shubigun Minseibu Tetsudobu, pp. 462–65; NCH , Nov. 22, 1907: 455; Chintao Minseibu Tetsudobu, Chosa Shiryo , p. 88.

[32] See, for instance, NCH , Nov. 10, 1905: 304; Nov. 15, 1907: 394.

[33] The situation in the late nineteenth century is summarized in Esherick, Boxers , pp. 170–72; a post-1900 primary source that gives the government's view is Sun Baoqi's memorial of 9/25/Xuantong 1 (1909) in archives of the Huiyi Zhengwuchu, file #573, document 4802, First Historical Archives, Beijing.

national or provincial receipts. Counties collected all taxes and forwarded a fixed amount of silver to the provincial capital.^[34] Thus, the profits from artificial exchange rates were retained locally. In fact, the Shandong governor complained that since Yellow River control demanded large copper expenditures, the province suffered from receiving a fixed silver income;^[35] he also argued that locally imposed conversion rates led to tax resistance and so prevented the levying of necessary new taxes.^[36] In 1901 the province decreed that taxes should be computed with the market exchange rate,^[37] but counties continued to set their own market rates and to enforce them by restricting currency flows.

Moreover, these restrictions often led to currency shortages in inland counties, which in turn created very profitable opportunities for issuing local coins or paper. While the printing of worthless paper by various "national" governments (especially during the warlord period, 1916–28) is well known, county reports of the Shandong Office to Encourage Industry reveal a constant struggle to limit the currencies issued by local governments and merchants. These local currencies, typically denominated in copper and often issued with no financial backing, were especially common in western Shandong, most of all in the Southwest region.^[38] Nearer the coast, copper coins were sometimes scarce, but a variety of foreign silver and paper rushed in to fill any vacuum.^[39] It is not surprising that Jining, the Southwest trading center with the sky-high silver-copper exchange rate, had one of the worst local-currency problems, which became a major target of political insurgents.^[40]

However, officials' efforts to quarantine their domains from market forces were not entirely cynical. The writings of economic officials in early twentieth-century Shandong have an almost mercantilist tone. Growth is desired, but primarily because it will help preserve China's independence, not

[34] See, e.g., NCH , Nov. 15, 1907: 394. John Schrecker, Imperialism and Chinese Nationalism (Cambridge, 1971), pp. 213–14, says that when Germany adopted market-rate conversions in Qingdao, local tax collectors, but not the government, lost.

[35] Reprinted in Wu Tongju, Zaixu xingshui jinjian (Taibei, 1966), 10:3748.

[36] Jining zhilizhou xuzhi (1929), 4:14a–b; NCH , Mar. 18, 1904: 575.

[37] En xianzhi (1909), 4:4a; doc. 4802 (memorial of 9/25/Xuantong 1), file 573, files of Huiyi Zhengwuchu, First Historical Archives, Beijing; Shandong zazhi , no. 62 (6/10/Xuantong 2): 11a–b.

[38] See, e.g., Shandong quanye huikan , no. 4 (Jan. 1922), ge xian shiye zhuangkuang : 17 (on Shan County); no. 7 (May 1922), ge xian quanye baogao : 32–33 (Heze); no. 11 (Feb. 1923), ge xian chengji baogao : 20 (Heze); no. 13, gongdu : 41 (Tangyi), 45 (Linqing).

[39] See, e.g., Shandong zazhi , no. 78 (11/20/Xuantong 2): 14b–15a; NCH , Jan. 30, 1909: 250.

[40] See Tang Chengtao, "Huoguo yangmin de 'li Ji qianpiao,'" Jining shi shiliao , 1983, no. 1, pp. 89–91; Wu Guogui, "Yutang xinghuo—Jining Shi Yutang jiangyuan gongyun gai," Shandong gongyun shi ziliao , no. 16 (Feb. 10, 1985): 13.

because it will promote individual welfare.^[41] Import substitution and export promotion projects were conceived as ways of escaping debt, reclaiming foreign concessions, or achieving other "self-strengthening" goals that would not be best served by full integration into the coastal economy.^[42] This stress on national political goals in economic policy fit perfectly in the intellectual climate of early twentieth-century China, where even advocates of Western liberalism stressed liberating individual energies more as a stratagem in China's struggle for autonomy than as an end in itself.^[43] The mixture of self-serving and sincere motives for impeding currency flows is exemplified in the provincial government's attempt to stop the export of copper coins from Shandong.^[44]

Even in the nineteenth century, copper coins had sometimes been worth more as metal than as coins, leading both Chinese and foreigners to buy coins and melt them down. During World War I, the world price of copper (as metal) climbed rapidly. Japanese merchants in Shandong aggressively bought up copper coins for use at home. After the Japanese seized Qingdao in 1914, they had little trouble evading provincial restrictions, and this trade spread along the Qingdao-Jinan railway. Between 1915 and 1919, according to the U.S. consul in Jinan, Shandong "was almost denuded of copper currency," with over 22 million yuan worth of copper exported. Coin exports from Shandong in these years were about 12 percent of the annual capacity of all of China's copper mints, or the equivalent of the leakage from all of China during the 1899 currency panic. Strong demand for copper coins raised their value further above the levels set by many local authorities, making restrictions even more important to county finance.

Provincial and national authorities were worried both about the symbolism of this trade—Japan used some of this copper to make bullet casings—and the control it gave Japan over the money supply along the railway. Complaints about this trade disappeared in the early 1920s but became common again after 1928.

Though provincial governments suffered from restrictions on currency

[41] See, e.g., Shandong zazhi, no. 88 (4/30/Xuantong 3): 7b–8a; no. 89 (5/15/Xuantong 3): 8a–9b; no. 90 (5/30/Xuantong 3): 7a–8b; Nong shang gongbao 1, no. 5 (Dec. 15, 1914), zhengshi : 14; Lin Maoquan (head of the Shandong Office to Encourage Industry in the 1920s), Wenji, pp. 2a–b, 71a.

[42] See, for instance, Lin Maoquan, Wenji, pp. 1a–7b; another, slightly different, example is the great interest by Shandong merchants and politicians in improving the harbor and otherwise building up Yantai as an alternative to foreign-controlled Qingdao and the waning of that interest when Qingdao was returned to China in 1922; see, e.g., NCH, Mar. 4, 1922: 576, 585.

[43] See, e.g., Benjamin Schwartz, In Search of Wealth and Power: Yen Fu and the West (Cambridge, 1964); Joseph Levenson, Liang Chi-chao and the Mind of Modern China (Cambridge, 1953).

[44] The following account is condensed from Pomeranz, "Making of a Hinterland," pp. 61–64.

movement between counties, they supported restrictions that kept coins from leaving the province. Provincial officials wanted a unified provincial currency and economy but were unwilling to let this happen through integration into international markets; even while they were redoubling efforts to prevent currency exports, they were (unsuccesfully) ordering all counties to abide by exchange rates for notes that would be wired from Jinan.

We have few details about how currency restrictions were enforced, but persistent exchange rate differentials suggest significant successes. In 1933 counties very close to each other, such as Shouguang (North Coast) and Linju (Heartland), and Heze and Cao (both Southwest), still quoted very different exchange rates.^[45] There are also accounts of people being searched for currency at train stations and other checkpoints.^[46] County governments could not fully police their borders, but policing rail and water routes greatly reduced currency arbitrage. (Postal remittances were irregular, and could be limited to firms adhering to local exchange rates.)^[47] Rate differentials between Jining and the nearest Heartland points far exceeded even the costs of moving copper by wheelbarrow.

The Politics of Economic Geography: Private Gain

Soldiers, however, were often exempt from inspection, and at a higher level, the politically connected could profit handsomely by evading restrictions.^[48] A hint of how currency moved across counties may be gathered from the other service offered by the Huangxian currency-moving company mentioned above: convoying opium, which required either extensive official contacts or private armed forces.^[49]

One revealing arbitrage story involves less privileged participants, from Fan County in far southwestern Shandong. In 1917, North China suffered huge floods. A North-China Herald reporter sent to the southwest Shandong—Henan—Zhili border area, near the origin of the flooding, found Japanese merchants paying 0.155 yuan per jin for copper coins worth about 0.167 yuan back on the coast.^[50] This trade had previously been carried on elsewhere,

[45] Shouguang xianzhi, 11:12a–b; Shandong zheng su shichaji, pp. 362, 371, 832.

[46] NCH, Nov. 10, 1905: 304; Nov. 15, 1907: 394.

[47] Clifrton O. Carey, "Narrative Account of Experiences in China," p. 8, in packet labeled "Letters, Sept.–Dec., 1919," Clifton O'Neill Carey Papers, Bentley Historical Library, University of Michigan; NCH, Sept. 27, 1907: 720.

[48] On soldiers, see NCH, July 29, 1916: 189. On politicians, see NCH, Dec. 22, 1917: 719, and Mar. 30, 1918: 752, which have examples drawn from Kaifeng, Henan, near Southwest Shandong.

[49] SSZ, 4:1145.

[50] NCH, Oct. 27, 1917: 216; the weight-to-value conversion ratio derived from the figures here matches figures from the American consul in Jinan quoted in Frederic Lee, Currency, Banking, and Finance in China (Washington, D.C., 1926), p. 32.

but high water had now allowed it to move further upstream, where silver fetched an even higher price. The Japanese were making their fourth trip to the area,

allowing just enough interval for the natives to take their [silver] dollars upstream or inland and bring all the available [copper] cash to this shipping point. . . . Last year, said the locquacious [village] headman [who was supervising the exchange], this trade did not come so far upstream and the officials prevented the shipment of cash down to the buyers, but this year the buyers had come themselves and "bought an open road" [bribed the officals] so that all was easy, and the impoverished countryside, under water for 3 months, was recovering, thanks to the cash trade.^[51]

Several points emerge from this story. First, more people knew about this trade than could ordinarily participate; the people in this village were normally excluded. Second, who participated was determined by official prohibitions and, conversely, by bribes. Third, even after getting by the formal government, the merchants dealt, not with individual peasants, but with a village head who had organized "his" people into teams. Fourth, the biggest profits were made on the inland side of the transaction. If the copper the Japanese bought for 0.155 yuan was indeed worth 0.167 yuan on the coast, they cleared less than 8 percent between Fan County and the coast, minus their costs. The villagers who took the coins further inland (on foot) received 1,400 copper cash for silver coins that cost them roughly 1,000,^[52] a 40 percent profit before accounting for their costs (largely a matter of their time). Local restrictions on currency flows created hefty profits for a favored minority of local traders, not for outside merchants.

Fragmented Markets and Distribution

Weakly integrated currency and capital markets redistributed wealth from some hinterland actors to others. High interest rates benefited hinterland creditors and hurt hinterland debtors. Distorted silver-copper rates redistributed wealth from private citizens to local governments. High exchange rates hurt those who were paid in copper but bought at least some things in silver (or at silver-denominated prices), and benefited people who were paid in silver. Finally, restrictions on currency movement benefited those who had the connections to circumvent those restrictions. However, specifying the members of these groups and how much they gained or lost is very difficult.

High interest rates aided not only banks and pawnshops but also certain stores, which probably provided more credit than official financial institu-

[51] NCH, Oct. 27, 1917:216, emphasis added.

[52] Ibid.

tions did.^[53] Buyers of cash crops were important credit sources in the Heartland and the North Coast,^[54] but not in the Southwest. In Jining, the shops that doubled as credit institutions were mostly those that sold modern or "foreign" goods: factory yarn, cloth, kerosene, matches, and cigarettes.^[55] Since they often got credit from their suppliers, who were located where money was cheaper, they were major beneficiaries of the high local interest rates.^[56] With their access to cheaper funds from outside the region, these "foreign goods" stores enjoyed a particularly large spread between their own cost of funds and what they charged their customers.

We have very little information on the cost of funds for these privileged firms, but to the extent that they could get outside loans from suppliers or other contacts, they stood to make very large profits. We do know that Jining's largest charity, whose directors included the city's leading gentry, merchants, and officials, received from 8 percent to 12 percent per year for its deposits at the city's leading pawnshops and merchant firms.^[57] These rates were likely influenced by self-dealing, since the charity's directors were also leaders of the firms with which they deposited money; thus these rates cannot be taken as indicative of ordinary deposit rates. However, if well-connected firms were paying 1 percent per month to the charity endowment because that was even approximately what other funds would have cost them, then the spread that they enjoyed with local consumer rates at over 2 percent per month was indeed substantial.

A 1940 survey argues that yarn and foreign goods merchants had become Jining's mercantile elite and notes a sharp conflict between them and the older grain, leather, and wool merchants^[58] —exporters of local products who lacked access to the importers' relatively cheap credit and were in turn surprisingly unimportant as lenders to local farmers. Unfortunately the survey does not describe this conflict. However, it does say that many importers

[53] Stores had long been a major source of credit throughout China. In early twentieth-century Shandong, they became still more important, partly because of discriminatory levies on pawnshops; in fact, some pawnshops became stores with a lending sideline in order to escape these charges. See Zhang Yufa, Zhongguo xiandaihua de quyu yanjiu: Shandong sheng, 1860–1916 (Taibei, 1982), p. 583, and ZSZS, pt. 10:28–31, on the small number of pawnshops remaining in the 1930s.

[54] See, e.g., Quanye huikan, no. 12 (March 1923), gongdu: 30 (on Northwest Shandong in general).

[55] Kokuritsu Pekin Daigaku, Santo sainei kenjo o chushin toseru nosan butsu ryutsu ni kansuru ichi kosatsu (Beijing, 1942), hereafter Santo sainei, pp. 94–100.

[56] On credit (in money, not just goods) from modern goods producers to their agents in the Jining area, see the letter of the British-American Tobacco agent Frank H. Canaday to A. Bassett of May 20, 1925, in the F. H. Canaday Papers, Harvard-Yenching Library, vol. 17: 92, 96. Credit was provided even though the company's Jining agent was "the richest man in the city" (vol. 15, letters of Aug. 20 and Aug. 26, 1923).

[57] Jining zhilizhou xuzhi, 5:10b–11a, 13b, 21b.

[58] Santo sainei, pp. 12–13.

were also local officials, while grain traders were not, and that importers, unlike exporters, got funds from Jining's two modern bank branches.^[59]

These Jining importers and a few other influential firms thus appear to exemplify the politically connected oligopolists with access to cheap credit that we are looking for. This hypothesis, however, should not be confused with the claims made in some literature that such traders were parasitic "compradors," as opposed to a productive "national bourgeoisie." The foreign goods they sold, though imports to the area, were often made in Qingdao, Tianjin, or Jinan factories. Perhaps the thing these importers offered that was most genuinely foreign in origin was credit, which came to them (directly or indirectly) from suppliers and modern banks based in the treaty ports. However, the problem was not that they brought in coastal funds but that they acted as a cartel, offering less credit at higher prices than would have been available were cross-county transactions unrestrained. And, as with the currency arbitrage described above, it was the local conduits for outside credit that profited from there being few such conduits, not would-be or actual outside lenders. American complaints that hinterland retail agents passed along currency losses—but not gains—to their foreign suppliers/creditors also suggest that merchants like the Jining importers were hardly subservient to treaty port interests.^[60]

Most farmers, laborers, and merchants outside the privileged circle were at least sometime borrowers—and therefore losers from the restrictions on capital mobility. However, some better-off members of these groups were also at least occasional lenders, or even net creditors, who gained from high rates. Being such did not require cash lending: farmers who did not need to sell immediately at harvesttime could hoard their crop and earn the area's high interest rate.^[61] In general, though, we know very little about who could and could not afford to do this.

For government overcharges, taxpayers were the obvious losers. Local governments and money changers gained, but we do not know how much or how the spoils were divided. Some officials tried to force local money shops to accept conversion rates that would have given government all the profit, but they often failed.^[62]

Cash-crop farming, which gave peasants a way to obtain silver for taxes, may have enabled some of them to escape this gouging. Several Southwest Shandong counties, however, would not let their citizens escape, even if they had silver on hand: they required that taxes be computed in silver but paid in

[59] Ibid., pp. 10, 31.

[60] Lee, Currency, p. 56.

[61] Santo sainei, pp. 86, 92.

[62] This would appear to be the case described in NCH, Nov. 15, 1907: 393–94, where the government itself had trouble obtaining silver. See also Schrecker, Imperialism, p. 213; Shandong zazhi, no. 62 (6/10/Xuantong 2): 11a–11b, and no. 77 (10/1/Xuantong 1): 14b.

copper.^[63] Evidence about how currency restrictions affected people as consumers, workers, and so forth is even more fragmentary, though we know that at least some cash-crop farmers were paid in silver, and so gained, while virtually all casual laborers were paid in copper, and thereby lost.^[64]

Fragmented Markets and Regional Economic Performance

The most basic effect of Shandong's highly imperfect capital market is clear. It kept capital scarcer in the Heartland and especially in the Southwest than it would have been otherwise; thus, certain projects in those regions were never carried out, even though they would have paid as well as projects that were carried out in the North Coast.

The precise effects on the Heartland of limited access to outside capital are harder to assess than those on the Southwest. The Heartland had its own access to the international economy through Qingdao. That port, however, was always dominated by one power (first Germany, then Japan) that tried to monopolize its trade, and credit and silver were never as cheap there as in Yantai, Weihaiwei, or Shanghai.^[65] The northwestern Heartland was also connected to Tianjin. The Heartland sent some people to Manchuria but received fewer remittances than the North Coast; those remittances seem to have pushed interest rates to particularly low levels in the North Coast and in Zunhua, Zhili.

Significant capital infusions did reach the Heartland. Improved cotton varieties, for instance, had an enormous impact, especially on northwestern Shandong; their introduction was financed by Qingdao and Tianjin mills and by a provincial government spurred by fear of the growing influence wielded by Japanese-owned mills.^[66] Once the new crop was established, the mills often advanced credit against it.^[67] Little is known about the rates on these loans; at times these buyer/suppliers established local monopolies that allowed them to become quite abusive.^[68] In general, however, the influx of foreign money meant that farmers got cheaper credit and did better than farmers growing for more local markets.

The growth of cash crops spared the Heartland the silver shortages common in the Southwest. As of 1933, most Heartland counties exported more

[63] NCH, Nov. 15, 1907: 393–94.

[64] See, e.g., SSZ, 4:234, 250, 260; Records, reel 37, pp. 4660213–19, 4660230; Wu Guogui, "Yutang xinghuo," p. 13.

[65] On Qingdao rates, see ZSZS, pt. 5, p. 90; Jiao-Ao xianzhi, shihuo: 85–88.

[66] For a general account, see Pomeranz, "Making of a Hinterland," chap. 2.

[67] Minami Manshu Tetsudo Kabushiki Kaisha, Tenshin Jimusho Chosaka, Santo no mensaku (Tianjin, 1942), pp. 47–54; Kanbe Masao, ed., Toa keizai kenkyu (Tokyo, 1942), 1:118–20.

[68] See, e.g., Shandong shiye gongbao, no. 3 (Sept. 1931), xunling: 27.

than they imported, which left some silver for taxes and for repaying whatever cross-county lending did occur. There is no reason to expect deflation, marked price divergence from the North Coast, or some of the peculiarities that did appear in the Southwest. Still, 30 years of persistent interest rate differences between the Heartland and the North Coast remind us that, with barriers to currency flows and conversion, the Heartland did not gain all it could have from the development of the coastal and Manchurian economies.

Having no outside loans except those financing exports may have suited Heartland officials, whose economic goals stressed the balance of payments and "independence." It did not, however, maximize growth. Capital infusions were limited by the extent to which the Heartland specialized in exports—and specialization depended on a still-backward transport network.

The Southwest was hurt much more, though. Even here, some outside capital flowed in, but only in very special circumstances.^[69] We would not expect outsiders to have made retail loans to Southwest Shandong peasants for weddings, funerals, and the like; even without currency restrictions, information and collection problems would have discouraged remote lenders. What is striking is the absence of outside credit for production and marketing in the Southwest, even of crops exported to areas with cheaper money.

In the Heartland, and in nearby northern Jiangsu (just south of Shandong), trade in cash crops was financed by merchant buyers in a chain stretching back to the coastal ports and world economy.^[70] However, the 1940 survey of Jining and nearby villages shows that credit for trade in Southwest Shandong came from below, not above: farmers generally borrowed from other village farmers.^[71] Cash crops were brought to market by poor farmers, who borrowed from richer peasants in the village to finance their trip; they usually paid the actual producers only after returning from market.^[72] The description of the market in Jining emphasizes how eager these xiao banzi (petty traders) were to get back to their villages;^[73] time pressure probably weakened their bargaining power, and was probably partially due to pressures from producers who needed cash quickly to pay taxes and debts. These xiao banzi rarely borrowed from Jining grain shops, and did not—indeed, usually could not—give advances against crops.^[74] Thus, even if a few received credit from above, it was not passed on.

[69] See Pomeranz, "Making of a Hinterland," p. 80 and chap. 1 n. 129.

[70] See, e.g., "The Peanut Trade of Tsingtao," Chinese Economic Bulletin 11, no. 348 (Oct. 22, 1927), pp. 213–14. On northern Jiangsu, see Chen Bozhuang, Xiaomai ji mianfen (Shanghai, 1936), p. 52.

[71] Santo sainei , pp. 13, 106. This pattern is confirmed by the figures in John L. Buck, Land Utilization in China: Statistical Volume (1937; New York, 1982), p. 404.

[72] Santo sainei , p. 40.

[73] Ibid., pp. 33–34, 39–40.

[74] Ibid., pp. 39–40.

Other potential sources also failed to relieve Southwest farmers' dependence on local credit. In 1939 most grain and most bean and nut oil firms in Jining were very small, were less than five years old, and did no business with Jining's two modern banks.^[75] "Guest merchants" who came to Jining from Taian, Jinan, and elsewhere got money from modern bank branches, but only after the harvest, when they were about to purchase the crop;^[76] thus they could not have passed these funds along as credit to finance production or lower levels of marketing. Southwest peasants and exporters had no chance to finance improvements with credit cheaper than that generated locally, as cash-crop growers elsewhere did. This was true even though many in the Southwest produced for ultimate buyers—Jinan flour mills and foreign buyers of peanut, cotton seed, and bean oils^[77] —who could borrow more cheaply.

With even exporters cut off from cheap funds, Southwest Shandong suffered frequent, serious silver shortages. The area responded in part by making do with very few modern-sector goods. As of 1933, the Southwest imported far fewer outside goods than any other part of the province: only 35 percent as much per capita, for instance, as a sample of 18 equally remote northwestern Heartland counties.^[78] Despite this austerity, however, the Southwest as a whole ran a merchandise trade deficit.^[79]

With a trade deficit, taxes to pay, and little outside credit, the Southwest exported labor, particularly to Manchuria. Though Southwest Shandong was more remote from Manchuria than any part of either Shandong or Hebei (the home provinces of 95 percent of the migrants to Manchuria), it sent tens of thousands of people per year there.^[80] Three years after Japan's seizure of Manchuria, remittances lost by 16 Southwest counties were estimated at 4 million yuan per year,^[81] while Manchurian remittances to all 215 Shandong and Hebei counties were about 30 million yuan .^[82] Jining alone may have lost 3 million yuan per year.^[83]

Before 1931, Manchurian remittances probably covered Southwest Shandong's trade deficits, at least before we add taxes.^[84] Thus, currency problems did not cause marked deflation or price divergence in the Southwest before

[75] Ibid., pp. 13, 26–27, 68–69.

[76] Ibid., p. 68.

[77] Ibid., pp. 47–48, 65–72.

[78] ZSZS , pt. 10:151–201.

[79] Ibid., pt. 10:165–89, gives the county-by-county figures.

[80] For estimates based on figures for 1933, see Toa Kenkyusho, Santo Koshogun chitai no chiiki chosa , Report no. 14, Class C, no. 158-D, pp. 127–29, based on figures for 1933.

[81] "Luxi ge xian nongcun jingji xianzhuang," Nongcun jingji 1, no. 11 (Sept. 1, 1934), p. 75.

[82] Gottschang, "Migration," pp. 137–38.

[83] "Luxi jingji xianzhuang," p. 76.

[84] ZSZS , pt. 10:156–201, has the trade figures; see also "Luxi jingji xianzhuang."

1931; the absence of outside capital is reflected instead in high interest rates, austerity, and massive temporary and permanent emigration. Even though rural elites in the Southwest sometimes prevented export growth,^[85] local governments and their allies generally met their silver obligations to higher levels of government while retaining enough control to manipulate exchange and interest rates to their benefit—and to the detriment of the population as a whole.

After 1931 this balance was disrupted. Between 1930 and 1934 the Shanghai price of wheat—Southwest Shandong's major cash crop—declined 40 percent.^[86] In Southwest Shandong, however, things were much worse. Of 13 Southwest counties that reported price trends in 1934, one reported a 50 percent decline in both wheat and land prices; four reported 60 percent declines; one a 75 percent decline; and seven 80 percent declines.^[87] This alone does not demonstrate price divergence from other areas, but since wheat was the Southwest's biggest product and biggest export, it is suggestive. So is the price of silver money (in copper, which might serve as a proxy for all goods sold for copper) in Jining. After retreating from the dizzying heights of the mid to late 1920s, that price stabilized from 1930 to 1933, but then took off again, rising faster in Jining over the next two years than anywhere else. At the least, this shows that three years after Manchuria was closed and a silver drain began, the high rates being offered for credit and hard currency were not drawing outside funds to the Southwest. By contrast, the more open North Coast counties also had merchandise trade deficits, and traditionally received even larger remittances from Manchuria, but avoided a hard currency shortage even after Japan's takeover of Manchuria dealt a severe shock to trade and migration across the Bohai Gulf.^[88]

Fragmented Markets and State Making

This summary of government influence on the capital market and economy leads back to the effects of economic phenomena on the state. Shandong officials announced goals of decreasing imports and increasing exports to accumulate silver. Far less attention was paid to commodities that circulated locally.^[89] Meanwhile, both the province and many county governments tried

[85] See Pomeranz, "Making of a Hinterland," chap. 2, for an extended example: the unwillingness of Southwestern village leaders to cooperate with the dissemination of new cotton varieties.

[86] Wu Chengming, "Woguo banzhimindi banfengjian guonei shichang," Lishi yanjiu , no. 168 (April, 1984), p. 115.

[87] "Luxi jingji xianzhuang."

[88] ZSZS , pt. 10:189–201.

[89] See, e.g., the discussion of development plans in Shandong quanye huikan , no. 10 (Jan. 1923), lunshuo : 1–3.

to make their jurisdictions economically meaningful units, in which they could control matters such as exchange rates and coinage free from broader market forces. Such freedom generally enhanced the particular government's short-term revenue. It also benefited merchants tied to local governments and those who could move money despite restrictions.

Many officials did seek greater economic integration. In some areas not discussed here, such as road building, important advances were made.^[90] County Offices to Encourage Industry of the 1920s never tried to retire all local currency, but they often did try to centralize issuing authority, usually in the county chamber of commerce.^[91] In the 1930s the provincial government tried to eliminate local paper currency, though with limited success.^[92] However, these efforts aimed at concentrating control at that particular level of government; control was not to be allowed to pass to a higher government unit or broader market. Provincial efforts to standardize currency rates within Shandong while banning currency exports are a good example.

Government policies often hindered economic growth in Shandong, particularly in the Southwest. But how well did government do for itself? This is a complicated question, but some general outlines are clear. While the provincial government could not control currency trading in the rich areas near Yantai and along the Jinan-Qingdao railway, it was fairly successful elsewhere, thus protecting its own influence and revenue. However, its control was never secure.

Meanwhile, many counties maintained independent monetary policies even into the 1930s. The provincial government's greatest success against local-currency trading was probably in Zouping, where a team of outsiders was brought in to take over the entire county government; until then, numerous orders to suppress local money and adopt standard exchange rates had been ignored.^[93] The 1935 national currency reform came too shortly before the outbreak of war to judge its success in Shandong, but in 1937 much of the province had very limited ties to national or international currency and capital markets, and the Southwest was still only loosely tied to the rest of the province's money markets. Unwilling to accept integration into larger markets, provincial leaders were also unable to impose full integration onto their own system.

[90] See county reports in Shandong zheng su shichaji .

[91] See, e.g., Shandong quanye huikan , no. 4 (Jan. 1922), ge xian shiye zhuangkuang : 17 (on Shan county); no. 7 (May 1922), ge xian quanye baogao : 32–33 (Heze); no. 11 (Feb. 1923), ge xian chengji baogao : 20 (Heze); no. 13, gongdu : 41 (Tangyi), 45 (Linqing).

[92] Luo Ziwei, "Zouping sichao mianguan," Xiangcun jianshe xunkan 4, no. 29, pp. 26–31; Shandong caizheng gongbao 6, no. 3 (Dec. 1934), mingling : 41; Shandong shengxianzheng jianshe shiyanqu gongbao , no. 14 (Nov. 27, 1935): 6–7; no. 23 (Dec, 15, 1935): 3 (page numbers supplied; original is unpaginated).

[93] Luo Ziwei, "Zouping," pp. 26–31.

This failure probably did not cost them much. Immediate provincial revenue increases from faster growth would have been small. The Southwest, poor and loosely tied to Jinan, was unlikely to yield more revenue. However, provincial officials could not ignore the Southwest. Two of the area's biggest problems—the Yellow River, which entered Shandong through the Southwest, and unusually serious banditry—endangered the rest of the province. Officials dealing with these problems were severely handicapped by the Southwest's economic isolation. Localities that regularly rebuilt earthen dikes balked at building cheaper stone dikes, which would have required paying hard currency to distant quarries, rather than making local expenditures for dirt, food, and labor.^[94] Subcounty leaders were supposed to maintain sections of the dikes, either from local resources or a small, fixed subsidy; they received no specifications. Given very high interest rates, they often found that building and rebuilding minimal barriers every two to three years was a cheaper way to meet their formal obligations than making one large outlay for dikes that could have lasted 20 years or more. However, less permanent dikes meant more floods.^[95]

After 1900, coastal areas in Shandong and elsewhere constructed fewer and fewer of their riverworks with locally levied labor and in-kind assessments (of stalks, earth, etc.) and used more cash taxes and public borrowing to buy specialized materials and hire contractors—including, sometimes, foreign contractors who brought new technology and cheaper credit. In eastern Shandong, the result was better and cheaper flood control. In the Southwest, however, reliance on local in-kind contributions increased, in part because the province forced localities to pay more of these costs than in the 1800s and because local governments preferred not to spend hard currency. In 1934 provincial officials who had just supervised Southwest Shandong's largest riverworks project in 50 years—all done with local materials and in-kind levies—explained that trying to do the project with money would have caused "panic."^[96]

Flood control would have increased output enough for taxes to easily pay for the work; thus, it was a perfect project for local governments to borrow for. Local governments often borrowed money to finance flood control projects near the coast, but not in Southwest Shandong. Local riverworks officials complained that they could not arrange even the short-term credit needed for routine maintenance.^[97]

[94] This discussion of riverworks is condensed from Pomeranz, "Making of a Hinterland," chaps. 4 and 5; citations below are examples from much longer lists. On the cost effectiveness of stone dikes, see Shandong hewu tekan , no. 2 (Jan. 1930), zhuanjian : 3–5, 10–15.

[95] See Pomeranz, "Making of a Hinterland," 223–35, and apps. G and H.

[96] Shandong Sheng Jiansheting, Junzhi Wan Fu Zhu Shui He zhi (Jinan 1934), p. 119.

[97] Shandong hewu tekan , no. 8 (Jan. 1936), gongdu : 89; Hewu jibao , no. 1 (Apr. 1919): 117; Shandong jianshe yuekan 2, no. 3 (Mar. 1932), baogao : 45.

The failures of river control in western Shandong were not primarily due to currency problems; many other factors were involved. Nonetheless, currency policies that sacrificed economic integration to local revenues and profits made it hard to move the public's resources through time and space to get the most out of them, just as they limited most private actors' ability to do this.

In sum, one critical part of state making in early twentieth-century Shandong—the search for greater revenue—interfered with market integration. This clash contrasts sharply with the pattern in early modern Europe, where state and market making usually reinforced each other, and market integration made the extraction and movement of state revenue easier.^[98] Shandong's poorly integrated markets, in turn, helped frustrate the development of an administration that could move resources to achieve things that localities could not.

Yet these problems cannot be blamed entirely on the governments involved. Government in early twentieth-century China did not consume a particularly large part of the country's income,^[99] and further economic growth surely required that some governments find enough revenue to provide increased amounts of various public goods (social stability, flood control, etc.). Some local officials may have favored the methods discussed here because they were profitable for themselves and their allies, but they probably also felt these methods were less disruptive than many alternatives and may have considered them important to defending Chinese authority against foreign intrusions. Finally, the foreign intrusions that helped provoke these measures (and which also sometimes disrupted currency flows, as in 1931) were inseparable from the stimulus that made economic links to the coast so useful. Early twentieth-century Shandong saw both state making and market making, but under these circumstances, they clashed rather than meshed.

[98] See, e.g., the discussion in Charles Tilly, "Reflections on the History of State-Making," in Tilly, ed., The Formation of National States in Western Europe (Princeton, 1975), pp. 17, 30–31, 52–57.

[99] Most estimates place government spending at 3–5 percent of gross domestic product. See, e.g., Philip C. C. Huang, The Peasant Economy and Social Change in North China (Stanford, 1985), pp. 280–84.