Innovation and Gender during Industrialization

Midway: The case of patenting activity in France 1791-1913

Youssouf Merouani

2023-11-27

A survey.

Merouani, Y. & Perrin, F. (2022). “Gender and the Long-Run Development Process: A Survey of the Literature.” European Review of Economic History, vol. 26, no. 4, pp. 612–641.

There has been a dispelling of myths about women

  • About how women worked in the past
  • About how much they earned

About how much they contributed to the economic prosperity of developed and developing economies.

many divisive debates, and unanswered questions

How does this project fit into the broader discussion?

Why?

  • Technology and science key in the industrialization process.
  • Still a research gap regarding women’s role in the process.

What?

  • A long-run view of innovation in France during its industrialization.

How?

  • The whole patenting activity during the century (1791-1913)
  • Geocoded industry data
  • Individual level data

How does this project fit into the broader discussion?

Why?

  • Technology and science key in the industrialization process.
  • Still a research gap regarding women’s role in the process.

What?

  • A long-run view of innovation in France during it’s industrialization.

How?

  • The whole patenting activity during the century (1791-1913)
  • Geocoded industry data
  • Individual level data

Data.

The first papers, use exclusively patent data.

Data collection and refinement

Data extension

Origins.

Women Inventors. On the Origins of the Gender Patenting Gap. (Merouani, Y. & Perrin, F.)

Women-linked Patents, 1791-1900

Women-linked Patents, 1791-1900

Gender Patenting Gap, 1791-1900

Step 1

\[ Pr(Y_i = 1) = \Phi\left(\beta_0 + \beta_1 \cdot \text{Year1844}_i + \beta_2 \cdot \text{PatentLength}_i + \beta_3 \cdot \text{TeamSize}_i + \sum_{j=1}^{J} \beta_{4j} \cdot \text{Sector}_{ji} + \sum_{k=1}^{K} \beta_{5k} \cdot \text{Education}_{ki} + \varepsilon_i\right) \]

\(Y_i = 1\) indicates women-linked patent

  • Financial resources
  • Team size
  • Female oriented sectors
  • Human capital

Step 2

\[ \log\left(\frac{P(Y_i = j)}{P(Y_i = K)}\right) = \beta_{0j} + \beta_{1j} \cdot \text{Year1844}_i + \beta_{2j} \cdot \text{PatentLength}_i + \beta_{3j} \cdot \text{TeamSize}_i + \sum_{l=1}^{L} \beta_{4jl} \cdot \text{Sector}_{li} \]

\(Y_i\) representing the marital status

  • Differencess across marital status

Findings

  • Increase of women patentees over time
  • Fixed Gender patenting gap by 1840’s
  • Positive effects of decreased financial constrains and larger teams
  • higher likelihood in female-oriented sectors
    • (+) Textile, Clothing, Industrial Arts, Office Articles, Medicine/Hygiene, and Paris articles
    • (-) Agriculture, Railways, Weaponry, and Mining
  • Both upper-tail human capital and “learn by doing”
  • Patenting across marital status
  • But differences in sectoral behavior depending on marital status
  • All variables have a small effect

Networks

Innovation networks: Dynamics of invention during the French industrialization. (Merouani, Y.)

Anatomy of a Network Graph

To understand a complex system, we first need to know how its components interact with each other.

  • Nodes: The individuals and the firms. Inventors and patent agents.
  • Edges: The direct interactions between the nodes. Collaborations and representation.

Inventor-only network

  • \(N\): 370,953 nodes
  • \(E\): 60,813 edges
  • \(\langle k \rangle\): 0.328 (Average Degree)
  • \(D\): \(8.84 \times 10^{-7}\) (Density)
  • Connected Components: 318,059
  • \(S\): 29 (Size of Largest Component)

Inventor-only network

  • \(N\): 370,953 nodes
  • \(E\): 60,813 edges
  • \(\langle k \rangle\): 0.328 (Average Degree)
  • \(D\): \(8.84 \times 10^{-7}\) (Density)
  • Connected Components: 318,059
  • \(S\): 29 (Size of Largest Component)

Full network

  • \(N\): 380,353 nodes
  • \(E\): 283,839 edges
  • \(\langle k \rangle\): 1.495 (Average Degree)
  • \(D\): \(39.27 \times 10^{-7}\) (Density)
  • Connected Components: 141,871
  • \(S\): 202,860 (Size of Largest Component)

Who are the central innovators?

Distribution of centrality scores

How does connectivity evolve?

Average degree evolution for the two networks

Average Degree Per IPC Sector. Size of the circles represent the relative share of connections within that sector compared to others.

Findings

  • “Democratic” dispersed process of connectivity between innovators.
  • In the full network, key players (patent agents) form hubs - taking most connections.
  • Central innovators have important cross-industry patents.
  • Central patent agents have strong industry ties & key know-how.
  • Women had the highest collaboration intensity in non-female oriented sectors.
  • Very small difference in collaboration intensity for men across sectors

Appendix: Origins paper

Appendix: Networks paper

Degree distribution of the Inventor-only network

Degree distribution of the full network

Average degree evolution for male and female inventors

Career Length and Unique Connections by Starting Decade of the Patent Agents. Circle size indicates the total number of patents divided by the number of agents for that decade

Record linking using the Fellegi-Sunter model

\[ W(i, j) = \sum_{k \in A} \log \left( \frac{m_k}{u_k} \right) + \sum_{k \in D} \log \left( \frac{1 - m_k}{1 - u_k} \right) \]

Where:

  • \(A\) is the set of fields that agree between record \(i\) and record \(j\),
  • \(D\) is the set of fields that disagree between record \(i\) and record \(j\),
  • \(m_k\) is the probability that the two records agree on field \(k\) given that they are a match,
  • \(u_k\) is the probability that the two records agree on field \(k\) given that they are not a match.

Fields I use:

  • Surname
  • Prenames
  • Marital status & Maiden name
  • HISCO and Sector
  • Application year
  • Specialization signature

Results of record linking

  • 389 thousand patents
  • 449,539 observations of inventors, collapsed into 370,954 unique individuals and firms
    • 355,638 nodes are unique individuals (350,940 male, 4 698 female)
    • 15,316 nodes are unique firms (11,948 firms, 3,368 family firms)
  • 216,115 observation of patent agents, collapsed into 8,896 unique agents

Fingerprinting Inventors’ Specialization

Step 1: Data Preparation

  • Cleaning Patent Descriptions
  • Remove common words, punctuation, entity names
  • Extract core words (lemmas)

Step 2: TF-IDF Analysis

  • Term Frequency-Inverse Document Frequency
  • TF: Frequency of a word in a patent
  • IDF: Uniqueness of a word across all patents

Step 3: Identifying Key Terms

  • Select words with high TF-IDF scores
  • Reflect unique aspects of an inventor’s work

Step 4: Creating the Fingerprint

  • Compile a list of significant words
  • Represents the inventor’s specialization

Collaboration and productivity

Top 10 with most unique connections
Fullname Patents Start End Years Unweighted Degree Sector
HENRY Henry 225 1847 1902 56 23 Railways
CHERADAME Antoine Leopold 23 1857 1875 19 20 Machinery
DIXON John 12 1825 1830 6 19 Textiles
ROUEN Pierre Isidore 19 1829 1846 18 16 Lightning
DARSONVAL Jacques Arsene 60 1879 1900 22 15 Precision Instrument
CHAMPONNOIS Hugues 28 1848 1881 34 15 Chemistry
POIRRIER Francois Alcide 11 1874 1880 7 15 Textiles
POTEZ Hyacinthe Aine 20 1861 1887 27 14 Railways
LAVILLE Jean Baptiste 7 1862 1868 7 14 Clothing
BERTHAUD Fils 9 1887 1895 9 13 Textiles

Collaboration and productivity

Most collaborative and productive innovators
Fullname Patents Start End Years Degree (Weighted) Sector
HENRY Henry* 224 1846 1901 55 123 Railways
SCHAFFER Bernhard 64 1861 1881 20 113 Railways
BUDENBERG Christian Friedrich 64 1861 1881 20 113 Railways
LECOINTE Jules 55 1858 1873 15 105 Chemistry
LECOINTE Eugene 46 1863 1873 10 103 Chemistry
VILLETTE Auguste 46 1863 1873 10 103 Chemistry
MARGUERITTE Louis 113 1850 1885 35 84 Chemistry
HALSKE Johann Georg 49 1855 1883 28 80 Precision Instruments
SIEMENS Werner 49 1855 1883 28 80 Precision Instruments
EDISON Thomas Alva 89 1872 1899 27 75 Precision Instruments


Collaboration and productivity

Most collaborative and productive innovators
Fullname Patents Start End Years Degree (Weighted) Sector
HENRY Henry* 224 1846 1901 55 123 Railways
SCHAFFER Bernhard 64 1861 1881 20 113 Railways
BUDENBERG Christian Friedrich 64 1861 1881 20 113 Railways
LECOINTE Jules 55 1858 1873 15 105 Chemistry
LECOINTE Eugene 46 1863 1873 10 103 Chemistry
VILLETTE Auguste 46 1863 1873 10 103 Chemistry
MARGUERITTE Louis 113 1850 1885 35 84 Chemistry
HALSKE Johann Georg 49 1855 1883 28 80 Precision Instruments
SIEMENS Werner 49 1855 1883 28 80 Precision Instruments
EDISON Thomas Alva 89 1872 1899 27 75 Precision Instruments


All the top most collaborative individuals are male. They are concentrated in the Railways, Chemistry and Precision instrument sectors.

Collaboration and productivity

Most collaborative and productive innovators
Fullname Patents Start End Years Degree (Weighted) Sector
HENRY Henry* 224 1846 1901 55 123 Railways
SCHAFFER Bernhard 64 1861 1881 20 113 Railways
BUDENBERG Christian Friedrich 64 1861 1881 20 113 Railways
LECOINTE Jules 55 1858 1873 15 105 Chemistry
LECOINTE Eugene 46 1863 1873 10 103 Chemistry
VILLETTE Auguste 46 1863 1873 10 103 Chemistry
MARGUERITTE Louis 113 1850 1885 35 84 Chemistry
HALSKE Johann Georg 49 1855 1883 28 80 Precision Instruments
SIEMENS Werner 49 1855 1883 28 80 Precision Instruments
EDISON Thomas Alva 89 1872 1899 27 75 Precision Instruments


The most collaborative individuals are also the most productive.

Women with highest number of unique connections

Top 10 women with most unique connections}
Fullname Maiden Name Patents Start End Years Unweighted Degree Sector
CARROT Catherine FOUGEDOIRE 2 1887 1887 2 12 Railways
FANGET Mariette FOUGEDOIRE 2 1887 1887 2 12 Railways
VEYRAND VINCENT Marie FOUGEDOIRE 2 1887 1887 2 12 Railways
FOUGEDOIRE Catherine 2 1887 1887 2 12 Railways
FOUGEDOIRE Marie BELLAT 2 1887 1887 2 12 Railways
BELLET Adele Louise MODINI 5 1866 1869 4 10 Railways
LEVIANDIER Pauline 16 1882 1887 6 7 Chemistry
HANTZ Amandine 15 1890 1899 10 7 Lightning
DE CHALEON Victoire Zoe HAUDINET 7 1870 1871 2 7 Machinery
JOSSE Amelie Alphonsine Constance CONTANT 5 1867 1872 6 7 Office/Education Articles

Most collaborative individual innovators (Women)

Top 10 most collaborative women innovators
Fullname Maiden Name Patents Start End Years Degree (Weighted) Sector
HANTZ Amandine 15 1890 1899 10 41 Lightning
HANTZ Adrienne 13 1890 1899 10 35 Lightning
MERLE Emilie 16 1877 1899 23 28 Clothing
MERLE Appolonie 16 1877 1899 23 28 Clothing
AUROY Julie Augustine LOISELEUR DESLONGCHAMPS 21 1875 1885 11 19 Textiles
LEVIANDIER Pauline 16 1882 1887 6 19 Chemistry
JANIOT Alexandrine Louise 20 1889 1899 11 18 Railways
ROHART Marie Leontine RUFIN 14 1882 1889 8 18 Chemistry
DE CHALEON Victoire Zoe HAUDINET 7 1870 1871 2 18 Machinery
MISSIRE Marie Louise MAGRIN 19 1885 1902 18 16 Lightning

Most productive individual innovators (Women)

Top 10 most productive women innovators
Fullname Maiden Name Patents Start End Years Degree (Weighted) Sector
REDIER Marie Josephine Herminie MICHELLE 31 1867 1885 19 3 Precision Instruments
RALU Denise CORNILLAC 24 1882 1886 5 15 Chemistry
PROPHETE Flore Felicite 23 1853 1874 22 11 Chemistry
AUROY Julie Augustine LOISELEUR DESLONGCHAMPS 21 1875 1885 11 19 Textiles
JANIOT Alexandrine Louise 20 1889 1899 11 18 Railways
MISSIRE Marie Louise MAGRIN 19 1885 1902 18 16 Lightning
PLANCHE Louise COSTE 19 1889 1905 17 16 Office/Education Articles
CAUZIQUE Marie Anne Pelagie Veronique LEROUX 18 1861 1887 27 4 Office/Education Articles
MATHIAN Claudine HILLAIRE 17 1887 1901 15 14 Lightning
ARBOUCAU Eugenie FOURQUET 17 1882 1902 21 1 Machinery

Firms with the most unique connections

Firms with highest number of unique connections
Fullname Patents Start End Years Unw. Degree Sector
GAUDET et compagnie 17 1852 1859 8 27 Mining & Metallurgy
COMPAGNIE PARISIENNE DECLAIRAGE ET DE CHAUFFAGE PAR LE GAZ 79 1862 1901 40 26 Lightning
LIZARS et compagnie 23 1844 1871 28 23 Lightning
TESTE pere et fils 20 1871 1885 15 20 Clothing
MONOT pere et fils 14 1877 1886 10 16 Ceramics
GENESTE HERSCHER ET COMPAGNIE 92 1877 1900 24 16 Lightning
COMPAGNIE DE FIVES LILLE 144 1869 1902 34 15 Railways
SOCIETE P. LEBOEUF ET GUION 27 1893 1900 8 15 Lightning
HUGGINS ET COMPAGNIE 7 1878 1880 3 15 Agriculture
CAIL et compagnie 40 1855 1882 28 14 Chemistry

Most collaborative inventor firms

Top 10 most collaborative innovating firms
# Fullname Patents Start End Years Degree * Weight Sector
1 FRIED. BAYER & COMPANY 317 1883 1902 19 311 Chemistry
2 COMPAGNIE PARISIENNE DE COULEURS D’ANILINE 279 1881 1900 19 275 Chemistry
3 BADISCHE ANILIN und SODA FABRIK 230 1878 1906 28 171 Chemistry
4 COMPAGNIE DE FIVES LILLE 143 1868 1901 33 144 Railways
5 JAPY freres 171 1845 1901 56 131 Precision instruments
6 AKTIENGESELLSCHAFT FUR ANILINFABRIKATION 125 1889 1901 12 125 Chemistry
7 SIEMENS & HALSKE 116 1881 1904 23 118 Railways
8 GENESTE HERSCHER ET COMPAGNIE 91 1876 1899 23 95 Lightning
9 MANUFACTURE LYONNAISE DE MATIERES COLORANTES 89 1889 1900 11 88 Chemistry
10 S. DES MATIERES COLORANTES ET PRODUITS CHIMIQUES DE SAINT DENIS 83 1882 1901 19 81 Chemistry

Most collaborative inventor firms

Top 10 most collaborative innovating firms
# Fullname Patents Start End Years Degree * Weight Sector
1 FRIED. BAYER & COMPANY 317 1883 1902 19 311 Chemistry
2 COMPAGNIE PARISIENNE DE COULEURS D’ANILINE 279 1881 1900 19 275 Chemistry
3 BADISCHE ANILIN und SODA FABRIK 230 1878 1906 28 171 Chemistry
4 COMPAGNIE DE FIVES LILLE 143 1868 1901 33 144 Railways
5 JAPY freres 171 1845 1901 56 131 Precision instruments
6 AKTIENGESELLSCHAFT FUR ANILINFABRIKATION 125 1889 1901 12 125 Chemistry
7 SIEMENS & HALSKE 116 1881 1904 23 118 Railways
8 GENESTE HERSCHER ET COMPAGNIE 91 1876 1899 23 95 Lightning
9 MANUFACTURE LYONNAISE DE MATIERES COLORANTES 89 1889 1900 11 88 Chemistry
10 S. DES MATIERES COLORANTES ET PRODUITS CHIMIQUES DE SAINT DENIS 83 1882 1901 19 81 Chemistry


The majority are chemistry companies specializing in colouring agents. Many connected to Germany. Some of the largest pharmaceutical companies today.

Most productive inventor firms

Top 10 most productive innovating firms
Fullname Patents Start End Years Degree Sector
FRIED. BAYER & COMPANY 317 1883 1902 19 311 14. Chemistry
COMPAGNIE FRANCAISE POUR LEXPLOITATION DES PROCEDES THOMSON HOUSTON 291 1894 1900 6 1 Railways
COMPAGNIE PARISIENNE DE COULEURS DANILINE 279 1881 1900 19 275 Chemistry
BADISCHE ANILIN & SODA FABRIK 230 1878 1906 28 171 Chemistry
FARCOT et fils 175 1860 1879 19 77 Railways
JAPY freres 171 1845 1901 56 131 Precision instruments
COMPAGNIE DE FIVES LILLE 143 1868 1901 33 144 Railways
AKTIENGESELLSCHAFT FUR ANILINFABRIKATION 125 1889 1901 12 125 Chemistry
SIEMENS & HALSKE 116 1881 1904 23 118 Railways
SOCIETE MIGNON ET ROUART 94 1862 1885 23 37 Chemistry

Most productive inventor firms

Top 10 most productive innovating firms
Fullname Patents Start End Years Degree Sector
FRIED. BAYER & COMPANY 317 1883 1902 19 311 14. Chemistry
COMPAGNIE FRANCAISE POUR LEXPLOITATION DES PROCEDES THOMSON HOUSTON 291 1894 1900 6 1 Railways
COMPAGNIE PARISIENNE DE COULEURS DANILINE 279 1881 1900 19 275 Chemistry
BADISCHE ANILIN & SODA FABRIK 230 1878 1906 28 171 Chemistry
FARCOT et fils 175 1860 1879 19 77 Railways
JAPY freres 171 1845 1901 56 131 Precision instruments
COMPAGNIE DE FIVES LILLE 143 1868 1901 33 144 Railways
AKTIENGESELLSCHAFT FUR ANILINFABRIKATION 125 1889 1901 12 125 Chemistry
SIEMENS & HALSKE 116 1881 1904 23 118 Railways
SOCIETE MIGNON ET ROUART 94 1862 1885 23 37 Chemistry


Only a few of the top 10 most productive firms are not in the top ten most collaborative.

Regression Analysis of Factors Influencing Collaboration
Y = Node degree Model 1 Model 2 Model 3 Model 4
Patent agents 1.07*** 1.05*** 1.05*** 1.05***
(0.01) (0.01) (0.01) (0.01)
No. patents 0.03*** 0.03*** 0.03***
(0.00) (0.00) (0.00)
Female -0.01 -0.00
const 0.29*** 0.27*** 0.27*** 0.20***
Sector Controls No No No Yes
R-squared Adj. 0.46 0.46 0.46 0.46
No. observations 370953 370953 370953 370953

Regression Analysis of Factors Influencing Women’s Collaboration
Y = Node degree Model 1 Model 2 Model 3 Model 4
Patent agents 1.06*** 1.02*** 1.02*** 1.02***
(0.02) (0.02) (0.02) (0.02)
No. patents 0.06*** 0.06*** 0.06***
(0.02) (0.02) (0.02)
Married -2.20 -2.03
(1.66) (1.78)
const 0.29*** 0.23*** 2.43 2.21
% (0.01) (0.02) (1.66) (1.78)
Sector Controls No No No Yes
R-squared Adj. 0.44 0.44 0.45 0.46
No. observations 4813 4813 4813 4813