1

French Foreign investment in the late 19th Century;

A Modern Portfolio Theory Analysis1

David Le Bris2

University of Orléans and Paris Sorbonne

Amir Rezaee

University of Orléans and EDHEC Business School

Abstract:

In this paper we analyze the reasons of French capital export to Russia during the late 19th century. Using the Modern Portfolio Theory we show that the French investors had understood the underlying notions of international portfolio diversification. So in order to maximize their portfolio’s performance they choose to invest also in foreign securities and this almost 50 years before the formal development of optimal portfolio theory by Markowitz. Our results show that although between the foreign securities the German debt would get a significantly higher mean-variance optimized portfolio performance but probably because of the hostile atmosphere toward Germany in France during this period this security was excluded from the French portfolios.

JEL classification: N23, O16, G11.

Keywords: Capital Export, Modern Portfolio Theory, International Diversification, Optimal Portfolio.

Introduction

The proportion of foreign securities quoted on the Paris Stock Exchange was

constantly rising in the late 19th and early 20th century. Although at that time almost

all the investments at the French Bourse were made by the individual investors, their

high amount of savings permitted them to invest also in foreign countries. Thanks to

this high level of savings estimated between 2 and 2.5 billion francs per year just

1 The authors thank Georges Gallais-Hamonno, Pierre-Cyrille Hautcoeur, Angelo Riva, Jean-Christophe Meyfredi, Frédéric Herbin, Octave Jokung, Iliya Komarev, Peter Daly, Cyrille Piatecki, Jean-Paul Pollin, Christian Rietsch, Kim Oosterlinck, Catherine D’Hondtn, Lionel Porte as well as the participants of the June 2009 LEO seminar at University of Orléans for helpful comments. 2 Contact e-mails: david.lebris@gmail.com & amir.rezaee@edhec.edu

2

before the Franco-Prussian war in 1870, the tribute demanded by Germans has been

subscribed and paid during just two following years of war. As shown in figure 1 the 5

billion war tribute subscription gets the foreign securities market down to the level of

the last 1860’s. Nevertheless, the 70s is the start of continuous development of foreign

values market for a 45 years period. The financial crisis in 1882 due to the crash of

one of the major French Banks “Union Générale” is the only event that reduces at

least the steady development of the foreign securities market for a while. As a result,

just before WWI, the French investment in foreign stock and bonds in both the Paris

official (Parquet) and Curb (Coulisse) market reached 45 Billion francs. This sum is

composed roughly half of the securities quoted on the Paris market and made

France the second largest capital exporting country in the world after Britain.

5

10

15

20

25

30

35

40

45

1866 1868 1870 1872 1874 1876 1878 1880 1882 1884 1886 1888 1890 1892 1894 1896 1898 1900 1902 1904 1906 1908 1910 1912 1914

Billi

on Fr

ancs

Figure 1. Par value of foreign securities quoted on Paris markets (Parquet and Coulisse)

Source: Catin (1927)

It is important to not that during this time the French provincial Stock Exchanges

like Lyon and Lille were also quoting major foreign companies from neighboring

countries such as Switzerland and Belgium. In this paper we focus only on capital

quoted at the Paris Bourse.

3

A glance at table 1 demonstrates that the main part of French capital was

invested in foreign Government debts. Thought during this time period, the issues of

foreign companies rise, but the government bonds percentage remains

unchanged3.

Million francs

Austrian 1641.1Grand Duchy of Baden 31Bavarian 60Belgian 484.4Danubian Principalities 46.6Egyptian 560.3Spanish 2712American 3032.8Haitian 30Honduras 103.7Hungarian 150Italian 1892.7Mexican 570Peruvian 298Portuguese 502Prussian 8.7Vatican 402.4Russian 2196.7Switzerland 7.2Tunisian 76Turkish 2225.5

17031.1Bank 665.4Railway 7284.1Mining 149.7Miscellaneous 831.9Total Private securities 8931.1ALL 25962.2Source : Catin(1927)

Total Government securites

Table 1. Nominal value of foreign securities quoted on Paris in 1870 (Parquet and Coulisse)

Among all the foreign securities quoted on the Paris Stock Exchange in the late

19th century, the Russian securities dominated others. Catin (1927) evaluated the

nominal value of Russian securities 13.2 billion Francs composing a quarter of all

foreign securities quoted on Paris in 1914.

Although Neymarck (1911) shows that the Russian debts and equities have been

popular between French investors since the early 70s but some researchers consider 3 In this paper we borrow constantly the data given by Catin simply because in his thorough PhD thesis he make a vast analysis of the previous works( Neymarck, Rafflovich, Leroy-Beaulieu,…) on foreign securities listed on Paris market and propose a unique bird eye view of the subject.

4

some political and diplomatic engagements the origin of the preference of investors

for Russian securities. Figure 2 depicts the nominal value of Russian government and

municipal bonds admitted to the Paris market per year (Parquet and Coulisse). Even

if this data set covers the conversion of older issues, the total amount of issues

compared to the other foreign bonds remains frankly high.

0

1

2

3

4

5

6

7

8

0

1 000

2 000

3 000

4 000

5 000

6 000

1871

1872

1873

1874

1875

1876

1877

1878

1879

1880

1881

1882

1883

1884

1885

1886

1887

1888

1889

1890

1891

1892

1893

1894

1895

1896

1897

1898

1899

1900

1901

1902

1903

1904

1905

1906

1907

1908

1909

1910

1911

1912

1913

Figure 2. Nominal value of Russian government and municipal bonds admitted to the Paris market per year (Parquet and Coulisse)

Nominal Value (Million Francs)

Number

Source: Catin (1927)

In this paper we test a classic hypothesis proffered by historians by using modern

portfolio theory (MPT). i. e. we examine the hypothesis of geographical diversification

to find out if investment in Russian securities could be justified by the MPT propositions

and not only by diplomatic reasons. The consistency of investment decision of French

investors with a rational decision-making frame work could explain the extent to

which the French investors improved their portfolios by investing in the international

securities.

While the theoretical demonstration of international diversification improving

portfolio performances is proposed officially by Markowitz (1952) in the early 1950s, its

premise was developed by some financier of the late 19th century. In an article

published in Le Rentier on January 7, 1878, Alfred Neymarck the founder of one of

the most influential financial journals of that time writes that “many of the investors

either by calculation or as a precaution desire not to develop their portfolios only

5

with domestic securities”4(authors’ translation). He repeats this idea of diversification,

and in his guide to investing (Que doit-on faire de son argent?)5 He advises investors

to diversify their portfolio by investing in geographically different countries.

Goetzmann and Ukhov (2006) give other examples of British financiers who

recommend geographical diversifications to investors. They also test the extent to

which the MPT explains investment of vast sum by the British in the overseas. They find

that diversification played an important role in the decision of British investors to

allocate a significant fraction of their portfolio to overseas securities.

In the French case Parent and Rault (2004), using an economic methodology,

investigate the influences affecting French assets abroad prior to 1914. They show

that the destination of French financial flows was consistent with rational economic

behavior. Le Bris (2009) finds that by WWI compared to higher foreign return the main

incentive that pushed investors to buy foreign assets was the low correlation

between domestic and foreign returns. Given these results, one can ask what is the

reason behind the high concentration of investments on Russian securities? Why

were Russian securities chosen by the French investor? Can this be explained by a

higher Russian securities rate of return? Could the mean-variance optimization

explain this phenomenon?

Utilizing original data sets of French securities mixed up with various foreign debt

data series, we run some optimal portfolios for a typical French investor and we

tested the rationality of his/her choices for the period 1870-1913. Our results suggest

that Russian debt provided satisfactory diversification for French investors’ portfolio.

This is demonstrated by the performance of the portfolio shown by the Sharpe ratio

and it is significantly high when Russian bond is added in. Nevertheless, the test

applied to other foreign assets indicates that the choice of Russian bond was not the

best one. German bonds propose even higher performances rather than Russian

ones. We suggest that this phenomenon could be explained by the fact that mean-

variance optimization does not take to account all factors influencing the

investment choices. Nevertheless, the basics of investor’s behavior were consistent

with the application of modern portfolio theory.

4 Borrowed from Catin(1927). 5 See Neymarck Neymarck, A., 1913. Que doit-on faire de son argent? (Marchal et Godde, Paris).

6

Data Taken from some recently created indices on 19th century French equity prices,

the French stock returns set ,historical CAC40 index, is borrowed from Le Bris and

Hautcoeur (2008) and the French Corporate bond returns comes from the Index

constructed by Rezaee (Juin 2008). The French bond returns is calculated on French

3% consol series. Since no data set is available concerning foreign securities quoted

on the French market during the 19th century, our foreign data set comes from the

British market.6 The Student t-test is not statistically significant for the sets of French

and English bond markets, as a result; our data set from London is also representative

of price movements in Paris. Nevertheless in order to ensure that the London and

Paris stock exchanges were cointegrated during our study period ,the johansen

cointegration tests have applied in two sets of French 3% consol prices quoted on

Paris and London. Johansen Maximum Likelihood (ML) procedure is a multivariate

test that is based on VAR approach to cointegration. There are two test statistics

produced by the Johansen ML procedure namely, Trace test and Maximal

Eigenvalue test. Here we use both of them to determine the number of cointegrating

vector present. The test relies on the relationship between the rank of a matrix and its

eigenvalues or characteristic roots. Given the results (table 2), for both tests the test

statistic exceeds its critical value (5%) when the null is r=0 we can conclude that at

least one cointegrating vector is present. Neverthless for more than one

cointegrating vector the test statistic is less than the critical value so we conclude

that only a single cointegrating vector is present between two variables. Our results in

table 2 depicting cointegration between Paris and London multiquoted securities

confirm previous observations that show a strong financial integration between the

Paris and London stock markets during the Gold standard Era especially after the

invention of the telegraph7. Hence, the Russian bond returns series is computed on

the monthly spread compared to the UK Consoles published in Investor’s Monthly

Manual and mentioned in Ferguson (2001) . These authors give also the other foreign

bonds’ spread such as German, Italian, Argentinean and Spanish ones and this

6 Details of data sets on Appendix 2 7 For integration between Paris and London markets see also Homer and Sylla Homer, S., and R. Sylla, 1991. A History of Interest Rates (Rutgers University Press, London).

7

between 1870 and 1913. Using these spreads and adding UK Consol’s return, we

have computed their monthly rates of returns8.

λtrace 5% critical valueH0: r=0 Ha: r=>1 293.76 15.49H0: r=<1 Ha: r=2 1.50 3.84

λmax 5% critical valueH0: r=0 Ha: r=1 292.26 14.26H0: r=1 Ha: r=2 1.50 3.84

Table 2. Johansen ML tests

* The Trace test is a joint test, the null hypothesis is that the number of cointegrating vectors is less than or equal to r, against a general alternative hypothesis that there are more than r cointegrating vectors. The Maximal Eigenvalue test conducts separate tests on each eigenvalue. The null hypothesis is that there are r cointegrating vectors present against the alternative that there are r+1 present.

a. Trace tests

b. Maximum Eigenvalue tests

Here we focalize only on the foreign government bonds listed on the Paris stock

exchange (Parquet). The reason is that according to Martin (1919) just before WWI

the percentage of foreign government and municipal debts was twice the rate of

private issues. Hence, the creation of optimal portfolio based on government debt

should be closer to the historical reality. It is important to mention that at that time

the foreign bonds were quoted in local currencies but with always a kind of

guarantee of parity with one of the main gold currency (French franc or sterling). As

the gold standard leads to stability of prices and exchange rate among these main

currencies, we only used nominal data sets.

Table 2 depicts some individual characteristics of the assets considered in this

study as well as the correlation between them. Surprising evidence is the rather low

coefficient of correlation between international bonds compared to average

coefficient observed between present day markets (0.77)9. A look at this table

dissipates the idea that French investment in Russia was motivated simply by higher

rates of return. Table 2 shows that Spanish, Argentinean and US municipal bonds

offered better returns than Russian bonds. Nevertheless they were not as popular as

the Russian ones.

8.For robustness and quality test of this data set see Le Bris Le Bris, David, 2009, Why did French Savers buy Foreign Asset befor 1914? Decomposition of the International Diversification Benefit, Eurohistock April 2009 (Madrid (University Carlos III)). 9 See Hautcoeur & Le BrisLe Bris, David, and Pierre-Cyrille Hautcoeur, 2008. A Challenge to Triumphant Optimists? A New Index for the Paris Stock-Exchange (1854-2007) (SSRN).

8

French French French corp. Russian German Spanish Argentine Italian UK US US mun.stocks bonds bonds bonds bonds bonds bonds bonds bonds bonds bonds

Return 5,40% 4,35% 4,20% 6,33% 4,62% 9,40% 7,95% 7,31% 2,81% 4,55% 6,17%Standard deviation 6,92% 5,58% 3,42% 6,88% 2,38% 17,04% 12,23% 6,32% 3,41% 4,00% 9,91%

French stocks 1French bonds 0,57 1French corporate bonds 0,41 0,76 1Russian bonds 0,14 0,11 0,10 1German bonds -0,09 0,04 -0,05 0,29 1Spanish bonds -0,06 -0,15 0,01 -0,16 0,14 1Argentine bonds -0,06 -0,09 -0,14 0,04 0,15 0,34 1Italian bonds 0,26 0,17 0,24 0,23 0,44 0,21 0,04 1UK bonds 0,18 0,18 0,33 -0,08 -0,48 -0,12 0,03 -0,02 1US bonds 0,34 0,33 0,42 0,08 0,01 0,09 0,06 0,21 0,28 1US municipal bonds 0,13 0,14 0,01 0,02 -0,08 -0,09 0,01 -0,04 0,11 -0,07 1

Correlation coefficient

Table 3. Assets characteristics

Optimal portfolio Modern Portfolio Theory

A portfolio is defined by allocating fractions of initial wealth to individual assets10.

The fractions (or weights) must sum to 1; but some of these weights may be negative

if short selling is allowed. The return of a portfolio is the weighted sum of the returns of

its individual assets, with the weights being those that define the portfolio. The

expected return of the portfolio is, likewise, equal to the weight average of the

expected returns of the individual assets. The variance of the portfolio is determined

by a more complicated formula: σ2= ∑ni , j=1 ωi ωj σij , where the ωi ‘s are the weights

and the σij ‘s are the covariances.

From a given collection of n risky assets, there results a set of possible portfolios

made from all possible weights of the n individual assets. If the mean and standard

deviation of these portfolios are plotted on a diagram with vertical axis “the mean

(r)” and horizontal axis “the standard deviation (σ)”, the region so obtained is called

the feasible region. Two alternative feasible regions are defined: one allowing short

selling and one not allowing shorting.

It can be argued that, investors who measure the value of a portfolio in terms of its

mean and its standard deviation, who are risk averse, and who have the

nonsatiation property will select portfolios on the upper left-hand portion of the

feasible region namely, the efficient frontier (figure 3).

Points on the efficient frontier can be characterized by an optimization problem

originally formulated by Markowitz (1952). This problem seeks the portfolio weights

that minimize variance for a given value of mean return. Mathematically, this is a

problem with a quadratic objective and two linear constraints. If shorting is allowed

(so that the weights may be negative as well as positive), the optimal weights can

10 A main part of this section is borrowed from very pedagogical book of Luenberger, David G., 1998. Investment Science (Oxford University Press, New York).

9

be found by solving a system of n+2 linear equations and n+2 unknowns. Otherwise if

shorting is not allowed, the Markowitz problem can be solved by special quadratic

programming packages.

An important property of the Markowitz problem, when shorting is allowed, is that

if two solutions are known, then any weighted combination of these two solutions is

also a solution. This leads to the fundamental two-fund theorem: investors seeking

efficient portfolios need only invest in two master efficient funds.

Usually it is appropriate to assume that, in addition to n risky assets, there is

available a risk-free asset with fixed rate of return rf . The inclusion of such an asset

greatly simplifies the shape of the feasible region, transforming the upper boundary

into a straight line. This line is the efficient frontier. The straight line frontier touches the

original feasible region (the region defied by the risky assets only) at a single point F,

namely optimal portfolio. This leads to the important one-fund theorem: investors

seeking efficient portfolios need only invest in one master fund (portfolio) of risky

assets and in the risk-free asset. Different investors may prefer different combinations

of these two.

The optimal portfolio F can be found by solving a system of n linear equations and

n unknowns. When the solution to this system is normalized so that its components

sum to 1, the resulting components are the weights of the risky assets in the master

fund (optimal portfolio).

The risk-return trade-off of the a portfolio could be found by the Sharpe ratio

computed as

Sp=(rp-rf)/σp

Where rp is the expected return on a given portfolio, rf is the risk-free interest rate,

and σp is the standard deviation of the return on the portfolio. Portfolios with higher

Sharpe ratios offer more attractive risk-return trade-off.

Evidence from optimization of Domestic and all foreign assets

To evaluate the role of the foreign assets, we first construct minimum variance

portfolios with only domestic assets. We then include foreign assets in the investment

opportunity set and construct again the minimum variance portfolios. We follow this

10

procedure to quantify the improvement in the risk-return trade-off due to the

inclusion of foreign investments in a classic short-sale constrained optimal portfolio.

Although short-selling has been practiced in Paris Bourse since the 19th century, but

considering in this study the portfolio of a typical French investor with long-term

investment horizon, it seems less likely that they appeal to the short selling strategy.

So, henceforward, we construct only short sale constrained portfolios.

Table 3 depicts the considerable improvement of investor’s portfolio when it

passes from a pure domestic to internationally diversified portfolio. The Sharpe ratio

of diversified portfolios is substantially higher than the domestic one. And even in the

case of the optimal portfolio, it reaches 0.98 points. The results obtained by pure

domestic as well as international diversification show that in order to get higher

performances, an investor has to exclude government bonds of its portfolio. These

results are not consistent with previous historical studies11 that show at least a quarter

of French portfolio was composed of government bonds. The same result is obtained

by Goetzmann and Ukhov (2006) considering UK railway bonds. e.g., despite their

popularity among British investors, UK railway bonds are excluded from optimal

portfolios. Hence, in order to approximate our results with historical facts we construct

new portfolios with a minimum of 25% of government bonds, even though this

constraint decreases the performance of our portfolios and it is no more an

“optimal” portfolio. Figure 3 represents both the constrained and non-constrained

efficient frontiers and this lower performance appears on the dashed curve.

11 See Bourguignon and Levy-Leboyer Bourguignon, F., and M. Levy-Leboyer, 1984, An Econometric Model of France during the 19th Century, European Economic Review 25, 107-141.and Rezaee Rezaee, Amir, Juin 2008, L'éfficience du marché obligataire français au 19eme siècle, Document de Recherche du LEO (Université d'Orléans).

11

French stocks 5% 1% 39% 4% 0% 0% 42% 0%(2.4) (1.0) (8.9) (2.4)

French bonds 0% 0% 0% 0% 25% 25% 25% 25%(0) (0) (2.2) (0)

French corporate bonds 95% 6% 61% 8% 75% 0% 33% 0%(2.4) (2.1) (9.1) (4.6)

Russian bonds 0% 10% 0% 14%(0.1) (3.2)

German bonds 57% 38% 47% 15%(1.7) (5.7)

Spanish bonds 2% 5% 2% 8%(0.2) (0.94)

Argentine bonds 0% 4% 0% 6%(0) (1.9)

Italian bonds 0% 14% 0% 22%(0) (3.7)

UK bonds 33% 0% 25% 0%(1.4) (1.1)

US bonds 0% 8% 0% 1%(1.2) (4.1)

US municipal bonds 2% 9% 0% 9%(0.5) (1.9)

Standard deviation 3,41% 1,37% 4,04% 2,48% 3,74% 1,87% 4,61% 3,20%(0.12) (0.05) (0.36) (0.20)

Return 4,26% 4,12% 4,67% 5,68% 4,24% 4,21% 4,74% 6,11%(0.16) (0.05) (0.20) (0.22)

Sharpe ratio 0,30 0,63 0,35 0,98 0,26 0,51 0,32 0,89(0.04) (0.04) (0.04) (0.04)

Port

folio

Wei

ghts

Short-sale constrainend Optimization-Minimum Risk Short-sale constrained Optimization

Bootstrap standard errors are in the parentheses( 5000 draws for each result)

Short-sale and French Bonds weight > 25 % constrained optimization -

Minimum risk

Short-sale and French Bonds weight > 25 % constrained optimization

Optimal Portfolio

Table 4. Optimal portfolio of French differnt investment securities and foreign debt

As the results show, due to the high correlation between three domestic securities

when we impose important weight to one of them in an internationally diversified

portfolio, the left two other securities are automatically abandoned in favor of lowly

correlated foreign bonds.

These quite surprising results in table 3 concern the international diversified

portfolios. As mentioned earlier, considering important amount of French investments

in Russia, we expect quite a high slice of portfolios allocated to Russian debt,

although here the weight of Russian bonds (14%) comes after the Italian (22%) and

German (15%) bonds. Strangely the exact allocation of French portfolio to the Russian debt

was 10%.(in 1900’s) This result is consistent with the Russian proportion in optimal short sales

constrained portfolio(Table 3 column 2)12. This result is all the more interesting when we

look at the weights of each foreign bond in minimum risk portfolio. In the late

portfolio preferred by investors with high risk aversion, the two Russian and Italian 12 See Des Essars Des Essars, P., 1897, Les dépots de titres à la Banque de France, Journal de la société de statistiques de Paris 38. and Valeurs Mobilières Des Valeurs Mobilières, Office National, 1914. Annuaire 1913-1914 (Office National des Valeurs Mobilères (Ancienne Association Nationale des Porteurs Français de Valeurs Etrangères), Paris).

12

bonds are abandoned in favor of German bonds. The bottom line of our results is

that the best foreign debt that contributes to decrease the performance of risk

averse as well as high return oriented investors’ portfolio is German debt. Hence,

German bonds should be considered as the predominant security between all

foreign assets.

Bootstrapping

Before going further into our mean-variance portfolio optimization we present our

robustness test. The sensitivity of mean-variance optimization to the small differences

in expected returns has been known to researchers in Finance13. As the reliability of

our deductions is dependent on the robustness of the calculated weights, we

applied the bootstrapping test to examine the robustness of the portfolio

characteristics. Hence, from the return of data series, we draw a new sample by

replacing and we estimate the vector of expected returns as well as the variance-

covariance matrix and given these two parameters, we compute optimal portfolio

characteristics (Weights, Sharpe ratio and etc.). We repeat this procedure N time

and using this N observation we construct standard error of all characteristics of our

portfolios.

2,00%

3,00%

4,00%

5,00%

6,00%

7,00%

8,00%

9,00%

10,00%

0,00% 2,00% 4,00% 6,00% 8,00% 10,00% 12,00% 14,00% 16,00% 18,00%

Retu

rn

Standard deviation

Figure 3. French only and French and Foreign Efficient Frontiers

French Corporate Bonds

OptimalPortfolios

Risk free rate

French Stocks

French Bonds

13 Jorion Jorion, P., 1985, International portfolio diversification with estimation risk, Journal of Business 58, 259-278.

13

Evidence from optimizing of domestic and foreign bonds portfolio: selecting 4 out

of 11 assets

As we show in the previous section, using a classical optimization for the portfolio

leads to results which are not consistent with the historical facts. In this section we test

other hypotheses to evaluate the optimal portfolio of a French investor in the late

19th century.

Daumard (1973) research’s on the French inheritance shows that an important

proportion of French investors were not really wealthy investors. Consequently, it is

difficult to imagine that they detain a well-stocked and highly diversified portfolio. It

was also probably expensive to hold many lines of foreign bonds since the cost of

transactions as well as informational costs were considerably higher at the time.

Additionally Barber and Odean (2000) show that today, the average American

investors’ portfolio is composed of just 5 assets.

Considering the previous studies and in order to simulate the most probable

French investment portfolio, we test the optimal portfolio for domestic assets plus one

foreign bond. Therefore, the influence of each foreign asset is clearly isolated. In

addition, this procedure allows us to compare the effect of each country’s debt on

French diversified portfolio.

Table 4 indicates the characteristics of minimum risk as well as optimal portfolios

composed by group of French domestic plus one foreign bond. Regarding the

Sharpe rations as ranking benchmark, portfolios composed of Domestic assets and

Russian bond are relegated to the forth place. Here the portfolio composed of

German bonds tacks the first place far from the other portfolios. It is followed by

Italian and Argentinean portfolios and in the fourth stage by the Russian portfolio.

14

Nevertheless, the bootstrap tests do not indicate a significant difference between

the Sharpe ratios of the previous two portfolios.

Domesticonly Russia Germany Spain Argentina Italy UK US US mun.

French stocks 39% 20% 12% 35% 24% 12% 39% 25% 25%(8.9) (5.6) (2.2) (7.5) (5.3) (5.4) (8.7) (7.1) (6.8)

French bonds 0% 0% 0% 0% 0% 0% 0% 0% 0%(2.2) (0.93) (0) (5.5) (0.69) (3.1) (2.5) (0.9) (0.06)

French corporate bonds 61% 35% 16% 44% 51% 19% 61% 28% 51%(9.1) (9.0) (3.7) (10.1) (7.1) (9.3) (9.0) (11.3) (8.9)

Russian bonds 44%(7.6)

German bonds 72%(3.3)

Spanish bonds 20%(3.1)

Argentine bonds 25%(4.2)

Italian bonds 68%(8.4)

UK bonds 0%(0)

US bonds 47%(12.0)

US municipal bonds 24%(6.5)

Standard deviation 4,04% 4,01% 1,99% 4,73% 3,94% 4,85% 4,04% 3,54% 3,92%(0.36) (0.32) (0.07) (0.43) (0.32) (0.39) (0.34) (0.23) (0.30)

Return 4,67% 5,39% 4,64% 5,68% 5,43% 6,47% 4,67% 4,66% 4,97%(0.20) (0.26) (0.08) (0.28) (0.24) (0.31) (0.20) (0.17) (0.24)

Sharpe ratio 0,35 0,53 0,70 0,51 0,55 0,66 0,35 0,40 0,44(0.04) (0.04) (0.04) (0.04) (0.04) (0.04) (0.05) (0.04) (0.05)

Domesticonly Russia Germany Spain Argentina Italy UK US US mun.

French stocks 5% 2% 3% 6% 4% 2% 2% 0% 2%(2.4) (1.9) (1.4) (2.3) (2.2) (1.9) (1.8) (1.2) (2.0)

French bonds 0% 0% 0% 0% 0% 0% 0% 0% 0%(0) (0) (0) (0) (0) (0) (0.1) (0) (0)

French corporate bonds 95% 81% 30% 91% 86% 83% 48% 63% 87%(2.4) (2.3) (2.8) (2.3) (2.3) (2.7) (3.8) (3.6) (2.2)

Russian bonds 17%(2.1)

German bonds 66%(2.2)

Spanish bonds 4%(0.9)

Argentine bonds 10%(1.3)

Italian bonds 16%(2.6)

UK bonds 50%(3.6)

US bonds 37%(3.7)

US municipal bonds 10%(1.1)

Standard deviation 3,41% 3,18% 1,90% 3,35% 3,14% 3,26% 2,78% 3,09% 3,24%(0.12) (0.12) (0.05) (0.11) (0.11) (0.10) (0.07) (0.10) (0.11)

Return 4,26% 4,59% 4,52% 4,47% 4,63% 4,71% 3,53% 4,33% 4,43%(0.16) (0.15) (0.08) (0.15) (0.15) (0.15) (0.13) (0.14) (0.16)

Sharpe ratio 0,30 0,42 0,67 0,36 0,44 0,45 0,10 0,35 0,36(0.04) (0.04) (0.04) (0.04) (0.04) (0.04) (0.04) (0.04) (0.04)

Table 5. Characteristics of optimal as well as minimum risk portfolios composed by group of French domestic securities and one foreign debt

b) Short sale constrained optimization-Minimum risk

Bootstrap standard errors are in the parentheses (5000 draws for each result)

a) Short sale constrained optimizationGeographic diversfication

Geographic diversfication

15

Here again the 25 % weight constraint on French government bonds is imposed for the following portfolios. Thus, the results are more likely to be close to historical reality.

16

Domesticonly Russia Germany Spain Argentina Italy UK US US mun.

French stocks 42% 18% 6% 34% 23% 4% 42% 23% 26%

French bonds 25% 25% 25% 25% 25% 25% 25% 25% 25%

French corporate bonds 33% 5% 0% 17% 23% 0% 33% 0% 22%

Russian bonds 52%

German bonds 69%

Spanish bonds 24%

Argentine bonds 29%

Italian bonds 71%

UK bonds 0%

US Bonds 52%

US municipal bonds 27%

Standard deviation 4,61% 4,60% 2,35% 5,28% 4,55% 5,05% 4,61% 3,93% 4,57%

Return 4,74% 5,55% 4,60% 5,89% 5,61% 6,50% 4,74% 4,69% 5,08%

Sharpe ratio 0,32 0,50 0,57 0,50 0,52 0,64 0,32 0,37 0,40

Domesticonly Russia Germany Spain Argentina Italy UK US US mun.

French stocks 0% 0% 0% 0% 0% 0% 0% 0% 0%

French bonds 25% 25% 25% 25% 25% 25% 25% 25% 25%

French corporate bonds 75% 57% 6% 70% 64% 58% 20% 36% 66%

Russian bonds 18%

German bonds 69%

Spanish bonds 5%

Argentine bonds 11%

Italian bonds 17%

UK bonds 55%

US bonds 39%

US municipal bonds 9%

Standard deviation 3,74% 3,50% 2,30% 3,63% 3,47% 3,57% 3,03% 3,40% 3,61%

Return 4,24% 4,61% 4,53% 4,50% 4,64% 4,77% 3,47% 4,37% 4,42%

Sharpe ratio 0,26 0,39 0,56 0,34 0,40 0,43 0,07 0,33 0,32

b) Short-sale and French bonds weight > 25 % constrained optimization - Minimum risk

Bootstrap standard errors are in the parentheses (5000 draws for each result)

Table 6. Characteristics of optimal as well as minimum risk portfolios composed by group of French domestic securities and one foreign debt (French bonds weight > 25 % constrained)

Geographic diversficationa) Short-sale and French bonds weight > 25 % constrained optimization

Geographic diversfication

17

According to table 5, although the German bond remains highly preferred,

nevertheless, the other weights are shaded. Here there is not a significant difference

between the Sharpe ratio of the Russian, Italian and Argentinean portfolios.

Therefore, we cannot find the best second performance , the three following bonds

would be chose to create the best performance after the German one. (Figures 4

and 5)

3,00%

3,50%

4,00%

4,50%

5,00%

5,50%

6,00%

6,50%

0,00% 1,00% 2,00% 3,00% 4,00% 5,00% 6,00% 7,00%

Figure 4. Group of French Securities and Russian Debt Efficient Frontiers

French Coporate Bonds

French Bonds

Russian Bonds

International efficient frontier

OptimalPortfolios

Risk free rate

French Stocks

OptimalPortfolios

Domestic efficient freontier

Domestic efficient frontier with a constraint of 25 % of Fench Bonds

International efficient frontier with a constraint of 25 % of Fench Bonds

18

3,00%

3,50%

4,00%

4,50%

5,00%

5,50%

0,00% 1,00% 2,00% 3,00% 4,00% 5,00% 6,00% 7,00%

Figure 5.Group of French Securities and German Debt Efficient Frontiers

French Coporate Bonds

French Bonds

French Stocks

German Bonds

International efficient frontierInternational efficient frontier with a constraint of 25 % of Fench Bonds

OptimalPortfolios

Risk free rate

OptimalPortfolios

Domestic efficient frontier

Domestic efficient frontier with a constraint of 25 % of Fench Bonds

Interpretations: Limits of mean-variance optimization

While mean-variance optimization is the best method to construct an optimal

portfolio but as it lies on the theory of rational beliefs, it cannot explain some

investment decisions influenced by factors like feelings, conservatism, anchoring and

etc14. As Markowitz tried to simplify the reality in his theory, he ignored other

behavioral elements that can play a role in the composition of portfolio. And some

are known as puzzles in finance literature (i.e. Equity Risk Prime Puzzle, Home bias

Puzzle,…)

Considering our previous results, we can deduct that the German bonds would

maximize the performance and propose a better portfolio diversification compared

to other foreign bonds. Nevertheless, the German debt composes few portion of

French portfolio at the time. The previous tables clearly show that the notion of

diversification was known to the investors but their portfolio stock picking was

influenced by the national atmosphere of the late of 19th century. At that time

France still suffers from the humiliating defeat of Franco-Prussians war of 1870. In

addition the Prussian victory was also the origin of the German Reich. As mentioned

14 For a complete list see Kahneman and Tversky Kahneman, D., and A. Tversky, 2000. Choices, Values and Frames (Cambridg University Press, London).

19

previously, in order to liberate some territories from the German army occupation,

France had to pay a huge tribute of 5 billion francs; this amount represents about 25

% of French GDP of the time. France was also forced to cede the Alsace-Lorraine

region to freshly created German Reich. During the period that separates this defeat

to the WWI, which covers our time period, France was engaged in a sultry feeling of

revenge. As a consequence, it is difficult to imagine that French investors chose

German debt in order to diversify their portfolio.

After removing the German debt, among the possible security choices, Russian

debt could offer the next best diversification mainly because in one hand, the

portfolios composed by Russian bond offers the second highest performance

(considering Bootstrap Standard error) amount the other foreign bonds and in the

other hand considering its poorly developed economy for which large amount of

foreign capitals is needed, only Russia could float a huge amount of financial

equities. Obviously it was not exactly the case for Italy and Argentina. Hence, in

order to obtain an optimal performance in a portfolio composed by French assets

and one foreign debt a French investor has to invest at least 18 % of his wealth in

Russian bonds. This level of allocation is the highest after the German one and could

justify the important investment of France on Russian debt.

Conclusion

In this paper we tried to find out why in the late 19th century France allocated a

large part of its portfolio to Russian debt. Our results show that the choice of Russian

bonds was consistent with Modern portfolio Theory. The results depict also that all of

the historical investment decisions could not be explained by the MPT especially

concerning the German debt. That is, this debt could offer similar even better

performance than Russian ones but some factors such as national feelings prevented

the French investor to buy them.

20

Appendix 1: Russian diversification tests

Domesticonly monthly hypothetical 1 annual 2

French stocks 39% 20% 32% 10%French bonds 0% 0% 0% 0%French corporate bonds 61% 35% 51% 48%Russian bonds 44%Russian bonds hypothetical return 16%Russian bonds annual data 42%Standard deviation 4,04% 4,01% 3,70% 3,63%Return 4,67% 5,39% 4,61% 5,22%Sharpe ratio 0,35 0,53 0,37 0,54

b) Short-sale and French Bonds weight > 25 % constrained optimizationDomestic

only monthly hypothetical annualFrench stocks 42% 18% 34% 4%French bonds 25% 25% 25% 25%French corporate bonds 33% 5% 24% 18%Russian bonds 52%Russian bonds hypothetical 17%Russian bonds annual data 53%Standard deviation 4,61% 4,60% 4,23% 4,46%Return 4,74% 5,55% 4,67% 5,42%Sharpe ratio 0,32 0,50 0,34 0,491 In order to test the effect of correlation between french and Russian debt on the optimal portfolio thereturns on russian debt and the corresponding domestic debt are set to be same. Yet it is still optimal to include russian debt in the portfolio and take advantage of the diversification opportunity that investing in russia offers. 2 The monthly returns series tacken from annual returns are used to compare the importance of frequency of the observations. the results show that the frequency of observations has no influance on the final portfolio performance.

Russian diversification

Russian diversification

a) Short sale constrained optimization

Appendix 2: Data Sources

US Municipal bonds:

1870-1913: monthly data, Macaulay (1938) available online:

http://www.nber.org/databases/macrohistory/contents/chapter13.html

French stocks: 1854-1988: approximated Cac 40, monthly data, Le Bris (2008)

French corporate bonds: 1838-1914: Rezaee’s all market index(Juin 2008)

French treasury long term rate:

1854-1913: Rente 3 %, monthly data,

21

French Bills: 1854-1913: Banque de France‘s taux d’escompte, annual data interpolated, INSEE’s Statistical Yearbooks

1864-1913: NBER and Banque de France,

UK Consols: 1854-1913: monthly data, Kloveland (1994)

UK Bills: 1854-1913: open market rate of discount, annual average interpolated, Homer & Sylla (1998)

German, Argentinean, Russian, Spanish, Italian spread with UK Consols:

1870-1913: Investor’s Monthly Manual, monthly data, Fergusson & Batley (2001)

References BIBLIOGRAPHIE:

Barber, B. M., and T. Odean, 2000, Trading Is Hazardous to Your Wealth: The Common Stock Investment Performance of Individual Investors, Journal of Finance 55.

Bourguignon, F., and M. Levy-Leboyer, 1984, An Econometric Model of France during the 19th Century, European Economic Review 25, 107-141.

Catin, Roger, 1927, Le portefeuille étranger de la France entre 1870 et 1914, (Faculté de droit de l'université de Paris).

Daumard, A., 1973. Les fortunes françaises au XIXème siècle (Mouton, Paris). Des Essars, P., 1897, Les dépots de titres à la Banque de France, Journal de la société

de statistiques de Paris 38. Des Valeurs Mobilières, Office National, 1914. Annuaire 1913-1914 (Office National

des Valeurs Mobilères (Ancienne Association Nationale des Porteurs Français de Valeurs Etrangères), Paris).

Ferguson, N., and R. Batley, 2001, Event Risk and the International Bond Marlet in the Era of Classical Gold Standard, (Oxford University).

Goetzmann, W., and Andrey Ukhov, 2006, British Investment Overseas 1870-1913: A Modern Portfolio Theory Approach, Review of Finance 10, 261-300.

Homer, S., and R. Sylla, 1991. A History of Interest Rates (Rutgers University Press, London).

Jorion, P., 1985, International portfolio diversification with estimation risk, Journal of Business 58, 259-278.

Kahneman, D., and A. Tversky, 2000. Choices, Values and Frames (Cambridg University Press, London).

Le Bris, David, 2009, Why did French Savers buy Foreign Asset befor 1914? Decomposition of the International Diversification Benefit, Eurohistock April 2009 (Madrid (University Carlos III)).

Le Bris, David, and Pierre-Cyrille Hautcoeur, 2008. A Challenge to Triumphant Optimists? A New Index for the Paris Stock-Exchange (1854-2007) (SSRN).

Luenberger, David G., 1998. Investment Science (Oxford University Press, New York). Markowitz, H., 1952, Portfolio Selection, Journal of Finance 7, 77-91. Martin, G., 1919. Les problèmes du Credit en France (Payot, Paris).

22

Neymarck, A., 1911. Finances Contemporaines ,Les budgets(1872-1903) (Alcan F., Paris).

Neymarck, A., 1913. Que doit-on faire de son argent? (Marchal et Godde, Paris). Parent, A., and Ch. Rault, 2004, The Influances Affecting French Assets Abroad prior to

1914, Journal of Economic History 64. Rezaee, Amir, Juin 2008, L'éfficience du marché obligataire français au 19eme

siècle, Document de Recherche du LEO (Université d'Orléans).