Private Payments Systems in Historical Perspective:
The Banco Central System of Mexico

Patrice Robitaille, Board of Governors of the Federal Reserve System

NOTE: The summary which appeared in the October 1997 issue of The Newsletter of The Cliometric Society contained graphics and formatting which cannot be accurately reproduced in this version.

The early 1900s was a particularly interesting period in Mexican banking. Mexican banks were far less regulated than they are today, and many banks had the power issue their own currencies (called bank notes) that were redeemable for specie on demand. The Mexican banking system was considered a stable one by Conant (1911), an influential foreign observer, and others1. However, the system came under stress in 1907 and went through a period of turmoil through the Mexican revolution, prompting a series of reforms that culminated in the founding of the Bank of Mexico (Mexico's modern-day central bank) in 1925. This paper focuses on a private payment system whose problems and collapse in 1913 played a major role in galvanizing banking reforms.
In 1899, a private bank called Banco Central Mexicano became the Mexico City correspondent for state banks of issue. State banks, unlike the two large "city" banks (so-called because their headquarters were in Mexico City), were not permitted to branch beyond specific states or regions. Banco Central redeemed state bank notes for specie at par and acted as state banks' agent for the clearing of other payments2. In 1908, however, a banking crisis prompted the Mexican government to provide a large amount of financial assistance to Banco Central and other large banks in an effort to keep all bank notes redeemable for specie at par and to protect other bank creditors from losses. Banking and other economic problems continued, fueling social unrest and political turmoil3. In late 1913, the Huerta government was confronted with a civil war and foreign sources of capital had dried up, prompting it to suspend bank note convertibility. By 1916, when the forces of Venustiano Carranza had established enough control over the country to form a new central government, bank notes could be exchanged for specie at only a small fraction of their face value.
This study is motivated by issues that have arisen as the result of the growth of private payment systems in recent years. As noted by Calomiris and Kahn (1996), the increasing complexity of private payment systems has prompted regulators to reassess their role and to place increasing importance on private mechanisms for managing risk. How well private payment systems can control risk-taking and cope in times of stress has been an open question, prompting some researchers to turn to historical experiences. Calomiris and Kahn look at the Suffolk Bank System, which operated in New England before the U.S. Civil War, and argue convincingly that it worked remarkably well in constraining members' behavior (see also Mullineaux 1987).

The first part of this paper looks at the origins of the Banco Central System (BCS). The last two decades of the 1800s was a time of political stability; previously, Mexico had suffered through decades of political turmoil and military conflicts internally and against foreign powers. Major improvements in transportation and communications in the 1880s and 1890s and the suppression of tariffs at the state level (alcabadas) in 1896 spurred growth in interregional trade. A new bank law in 1897 relaxed entry restrictions for state banks of issue. These developments led to an increase in the number of state bank notes that traveled greater distances. State bank notes traded at discounts when they were far enough from the issuing bank, including in Mexico City, the country's leading financial and commercial center. This phenomenon was seen by contemporary observers as undesirable; notably, Casasus (1908) and Gurza (1905) believed that bank note discounts undermined monetary stability.
The BCS was a mechanism through which the state banks could extract rent from the city banks. The city banks' notes circulated at par throughout the country. This reflected in part their special privileges that increased the demand for, and helped support the value of, their notes. In 1897, the city banks held 85% of the assets of the banks of issue. Banco Nacional de Mexico, the larger of the two, served as the fiscal agent for the federal government, and its notes were the only ones that were acceptable for federal payments - and acceptance was required to be at par. Banco Nacional had also lobbied unsuccessfully for several years to get the federal government to enforce an 1884 law that essentially gave the bank a monopoly over the power to issue currency. Even through Banco Nacional dropped its opposition to the entry of new state banks in 1897, the bank's informal role as a broker for state bank notes generated complaints on more than one occasion (see Barrera Lavalle 1909 p.208).
It was widely believed by contemporaries that fraud by state banks had contributed heavily to the Mexican banking system's problems after 1907. Cerda (1997) observes that adverse economic and political conditions deserve much of the blame for the banking system's problems, but maintains that mismanagement by the state banks played an important role. Direct evidence of fraud is hard to come by, and any explanation for systematic fraud would have to focus on the working of the BCS. For the scheme to be viable (that is, to keep Banco Central from being flooded with requests for redemptions of state bank notes), Banco Central had to have constrained the behavior of its members.
In two respects, Banco Central seems to have been at a disadvantage relative to the Suffolk Bank in terms of its ability to discipline bank behavior. Unlike the Suffolk Bank System, which was restricted to New England, the BCS was on a national scale. The greater distance between Banco Central and some of its banks and made it harder for the Mexico City bank to monitor other banks. Also, Banco Central did not necessarily require payment in specie when it returned state bank notes for specie. According to Conant (op. cit.), it allowed member banks to borrow and required repayment only once every six months. The Suffolk Bank, in contrast, at first required payment in specie. Later, participating banks were allowed to borrow temporarily, but were required to pay off their balances every week (Calomiris and Kahn op. cit.).
Could the greater potential for "excessive" risk-taking under the BCS system have been exacerbated by its different capital structure? State banks were required to hold stock in the bank. If Banco Central operated as a cooperative, did some members exploit the arrangement for their own profit at the expense of other members? Banco Central appears to have been structured more like a private bank than a cooperative. Banco Central's shares were also held outside investors through private placements and public offerings. In 1900, Banco Central suspended the par redemption of the notes of Banco Yucateco after the state bank's management informed its correspondent that it could no longer redeem its notes for specie. Suspension was a powerful mechanism for disciplining bank behavior, but Banco Yucateco's management did not see matters this way. Banco Yucateco blamed foreign shareholders of Banco Central for the suspension, and lobbied other state banks to purchase the foreign investors' shares. One-half of Banco Central's shares were in the hands of Bleichroeder and Deutchebank, two major German investment houses of the late nineteenth century, and J.P. Morgan. Although the share purchase (some time in 1901) initially left the state banks collectively with a two-thirds stake in the bank's stock and in control of the management, this stake appears to have been diluted by subsequent share offerings that were taken up by outside investors.
Furthermore, the evidence suggests that Banco Central continued to operate like a profit-maximizing bank after 1901. Although the bank's liabilities relative to its specie holdings increased substantially after 1901, subsequent capital increases enabled it to maintain a fairly constant ratio of liabilities to capital (see Table 1)4. On the other hand, especially after 1901, state banks tended to have higher ratios of notes to specie and higher ratios of specie to market capital than did than did the city banks, Banco Nacional and Banco de Londres (see Tables 2 and 3). Total liabilities of banks were also looked at (results not shown), even though Banco Nacional's role as a federal treasury agent makes it hard to interpret the results. (The balance sheets of banks of issue did not distinguish among non-bank note liabilities in any meaningful way.) If one ignores Banco Nacional (whose risk was the highest by this measure), state banks tended to have higher ratios of liabilities to specie and to market capital than did Banco de Londres. The empirical work has focused on the period up to 1904, in an attempt to separate the effects of the BCS from the effects of the change in the monetary regime in 1905. That year, Mexico abandoned a silver standard for bimetallism.

The results just reported do not provide clear evidence that there was any systemic tendency to engage in fraud or "excessive" risk-taking. There was considerable variation in risk levels. Also, no attempt is made to control for the riskiness of the economic environment faced by each of the state banks. It could be that a higher level of riskiness reflected higher underlying risk.
An attempt was made to address the question more systematically with panel estimations. The observations were from December of each year. Four measures of bank risk for which data are available over the 1897 to 1904 period were used as dependent variables: notes relative to specie, total liabilities relative to specie, notes relative to book capital, and total liabilities relative to book capital. These measures were regressed on bank dummies, time dummies, and a dummy variable that was equal to one if an observation represented a state bank after 1900 and zero otherwise. The results weakly support the hypothesis that even after controlling for year and bank specific factors, state banks tended to have higher levels of risk. Although the coefficient on the interactive dummy was always positive, most of the time it was not significantly different from zero. Another dummy variable was added that equaled one if an observation was a city bank after 1900 and zero otherwise. Its coefficient was always negative, suggesting that city banks tended to have reduced risk after 1900, but was never significantly different from zero. The lack of statistical significance probably reflects the tests' low power; only two banks are alternatives to the state banks (with varied in number from 7 in 1897 to a peak of about 25).
Although these results still provide no conclusive evidence of systematic behavioral problems, additional information that suggests that Banco Yucateco appears to have taken risks that seemed unjustified by its own investment opportunities. Banco Yucateco's case is an important one. Banco Yucateco had become the largest of the state banks by 1901. Moreover, the institution of a limited state bank mutual aid scheme in 1904 potentially allowed a bank's problems to be transmitted to other state banks. Although the 1897 bank law greatly discouraged states from having more than one state bank, there were a couple of exceptions to the rules, one of these being in the state of Yucatan. Banco Yucateco's measures of bank risk were nearly always higher than those of the other Yucateco bank. Either in July or November 1903 (accounts differ), Banco Nacional began a note redemption scheme with the two Yucateco banks under which each bank would accept each others' notes at par. While the structure of the arrangement is not entirely clear, April 1904, Banco Yucateco's management notified Banco Nacional that it could not redeem its notes and asked for a loan. Banco Nacional's management postponed its decision.
Paralleling this development, in March of 1904, state banks broadened the original agreement with Banco Central to include the mutual aid scheme. Each state bank also agreed to accept at par notes of any other state bank. This, according to Rosenzweig (1965 p. 828), was intended to "avoid excessive concentration of exchange at Banco Central and reduce to a more manageable level the balances of each bank with Banco Central". What was not appreciated was that the move only transferred some of the burden of managing risk from Banco Central to each state bank. The following May, El Economista Mexicano reported that although notes of Banco Yucateco were being accepted by the other Yucateco state bank, "notes are not exchanged regularly and balances build up".
Around 1904, as Cerda (op. cit.) notes, Banco Nacional became more active in providing loans to state banks. A sizable loan had been provided to Banco Oriental in 1902. Banco Nacional extended a loan to Banco Central in mid-1904 and renewed at the following December at the request of Joaquim Casasus. Casasus, a major architect of the 1897 bank law with a distinguished career in public service, was also heavily involved in the affairs of Banco Central. The monetary reform of 1905 brought the perception of greater monetary stability in Mexico and generated inflows of capital from abroad, a good portion of which took the form of direct foreign investment in many Mexican banks, including Banco Yucateco. Thus, we conclude that government involvement in banking provided some support to the state bank note par redemption scheme even before a large amount of money was injected into the banking system by the government in 1908.

Manero (op. cit.) asserted that the Banco Central System never worked very well despite the expectation that Banco Central, being two-thirds owned by its members, would have acted in their interests and limited bad behavior. Our results suggest that Banco Central acted in ways that would have constrained bank behavior. Government involvement in supporting par redemption may have thus been partly responsible for the problems that Mexican banks experienced after 1907, but future research should focus on the role of other factors.

1 Conant's work for the U.S. National Monetary Commission was largely based on an article published in The Bankers Magazine in November 1908.
2 There is far less information on Banco Central's role in clearing checks, which were becoming increasingly popular by the turn of the century. Federal bank inspectors often noted the discount rates of bank drafts ("giros") on distant localities. This seems to suggest that the Banco Central System was not intended to and did not support par check clearing.
3 1911, Porfirio Diaz, who had been in power for nearly 30 years, was forced into exile, and in 1913, President Francisco Madero was assassinated.
4 Since specie reserves included holdings of other banks' notes before 1901, the increase thereafter is exaggerated to some degree.

Brothers, Dwight S., and Leopoldo Solis, 1966, Mexican Financial Development, Austin, University of Texas Press.
Calomiris, Charles, and Charles Kahn, "The Efficiency of Self-Regulated Payments Systems: Learning from the Suffolk System", Journal of Money, Credit, and Banking, November 1996, Part 2, pp. 766-797.
Casasus, Joaquin D., Las Reformas a la Ley de Instituciones de Credito, Mexico City, 1908; From articles published in El Tiempo.
Cerda, Luis, Historia Financiera del Banco Nacional de Mexico, Mexico City: Fomento Cultural Banamex, A.C., 1997.
Conant, Charles, "The Banking System of Mexico", in Banking in Belgium and Mexico, Washington, D.C.: National Monetary Commission Vol. XVI, 1911.
Gurza, Jaime, Nuestros Bancos de Emision, Imprenta Central, Mexico, 1905.
Manero, Antonio, La Revolucion Bancaria en Mexico, Mexico City, Talleres Graficos de la Nacion, 1957.
McCaleb, Walter Flavius, Present and Past Banking in Mexico, New York, Harper and Brothers, 1920.
Mullineaux, Donald, "Competitive Monies and the Suffolk Bank System: A Contractual Perspective", Southern Economic Journal, April 1987, pp. 884-898.
Rosenzweig, F., "Moneda y Bancos", pp.789-885 in Cosio Villegas, Daniel, Historia Moderna de Mexico, Ed. Hermes 1965, Vol VII, Tome II.

Table 1 Total liabilities of Banco Central relative to specie and to capital (December of each year)

	Liabilities/specie*	Liabilities/capital **
1899 2.7 1.6
1900 3.5 0.9
1901 5.8 1.3
1902 14.5 2.4
1903 15.0 1.7
1904 15.0 1.7
1905 21.9 1.2
1906 33.1 1.5
1907 42.9 1.6

* specie includes holdings of other bank notes prior to 1901.
** capital is defined as book capital, which is paid in capital plus reserves and provisions.

Table 2 Notes in Circulation Relative to Specie Holdings

State Banks			Banco Nacional 	Banco de Londres
Year No. Obs. Median Mean Std. Dev.
1897 7 1.55 1.51 0.34 1.11 1.67
1898 13 1.59 1.45 0.42 0.82 1.89
1899 13 1.49 1.49 0.31 0.90 1.70
1900 15 1.74 1.70 0.34 0.93 1.59
1901 17 1.69 1.74 0.37 1.33 1.04
1902 21 1.67 1.64 0.44 1.62 1.82
1903 26 1.73 1.74 0.30 0.96 1.60
1904 26 1.71 1.76 0.22 0.62 1.20

Table 3 Notes in Circulation Relative to Market Capital
State Banks			Banco Nacional	Banco de Londres
Year No. Obs. Median Mean Std. Dev.
1900 15 0.89 1.08 0.77 0.58 0.76
1901 16 0.82 1.07 0.61 0.47 0.60
1902 21 0.84 1.02 0.51 0.46 0.64
1903 22 0.84 0.97 0.45 0.35 0.69
1904 26 0.95 1.01 0.55 0.31 0.60