Collaboration for the Bioeconomy

Philipp Jonas Kreutzer
philipp_jonas.kreutzer@ekh.lu.se

Lund University

Josef Taalbi

Lund University

2025-09-11

Collaboration Is Expected to be Crucial For Bioeconomy Transition

We test central hypotheses of its importance for the bioeconomy transition with new panel data of collaboration behind commercialized innovation in Sweden 1970–2021

Node’s Neighbors

Figure 1: Predictions of Subsequent Innovation By Neighbor Predictors, Poisson Model Including Year FE and Clustered Standard Errors

Node’s Brokerage Positions

Figure 2: Predictions of Subsequent Innovation By Brokerage Predictors, Poisson Model Including Year FE and Clustered Standard Errors

Node’s Knowledge Access and Cognitive Proximity

Figure 3: Predictions of Subsequent Innovation By Knowledge Predictors, Poisson Model Including Year FE and Clustered Standard Errors

Model Specification

\[\text{Innovations}_{it} \sim \text{Poisson}(\lambda_{it})\] \[\log(\lambda_{it}) = \alpha + \boldsymbol{X}_{it-1} \boldsymbol{\beta} + \text{Age}_{it-1} + \text{Year}_t + \varepsilon_{it}\]

\[ \boldsymbol{X}_{it-1} = \begin{bmatrix} \color{#d47a17}{\text{Neighbors: Direct & Indirect Ties}} \\ \color{#9370DB}{\text{Brokerage: 2-step Betweenness}} \\ \color{#73B4A0}{\text{Knowledge: Access & Proximity}} \end{bmatrix} \]

\[\text{All variables interacted with bioeconomy firm indicator}\]

Robustness: Specifications without year effects and variable groups tested

Data From SWINNO Database

Significant Swedish Innovations

~ 5000 unique innovations

Literature-Based Innovation Output method

from 15 independent trade journals (Sjöö et al. 2014)

Example Page From a Source Article

Collaboration Network

Figure 4: Plot of Swedish Innovation Collaboration

Take Aways

  1. Collaboration matters for innovation output, but does not matter more for bioeconomy firms.
  2. Many direct ties, but not indirect ties, were associated with higher innovation output.
  3. Bioeconomy firms benefitted less from access to knowledge, and were too cognitively similar to neighbors.

Contact Us!

philipp_jonas.kreutzer@ekh.lu.se

swedishinnovationdata.se

@swinnoproject

References

Everett, Martin G., and Stephen P. Borgatti. 2020. “Unpacking Burt’s Constraint Measure.” Social Networks 62 (July): 50–57. https://doi.org/10.1016/j.socnet.2020.02.001.
Sjöö, Karolin, Josef Taalbi, Astrid Kander, and Jonas Ljungberg. 2014. “A Database of Swedish Innovations, 1970-2007.” Lund Papers in Economic History General Issues (133): 77.
Weidenman, Per. n.d. “The Serrano Database for Analysis and Register-Based Statistics.” Accessed February 7, 2025.

Appendix

Model Specification (Full)

\[\text{#innovation}_{it} \sim \text{Poisson}(\lambda_{it})\]

\[\log(\lambda_{it}) = \alpha + \boldsymbol{X}_{it-1} \boldsymbol{\beta} + \text{Age}_{it-1} + \text{Year}_t + \varepsilon_{it}\]

\[ \boldsymbol{X}_{it-1} = \begin{bmatrix} \color{#9370DB}{\text{Centrality}} \\ \color{#d47a17}{\text{Neighbors}} \\ \color{#73B4A0}{\text{Knowledge}} \end{bmatrix} \\ = \begin{bmatrix} \color{#9370DB}{\text{2-step Betweenness}} \\ \color{#d47a17}{\text{Direct Ties}, \text{Indirect Ties}} \\ \color{#73B4A0}{\text{Accessible Knowledge}, \text{Cognitive Proximity}, (\text{Cognitive Proximity})^2} \end{bmatrix} \]

\[\text{Each } X_j \times (1 + \text{is bioeconomy firm})\]

Cognitive Proximity

\[ J(A, B) = \frac{|A \cap B|}{|A \cup B|},\]

where, \(A\) and \(B\) represent the knowledge bases of node \(i\) and its neighbor \(j\), respectively.

Burt’s Constraint versus 2-Step Betweenness (Everett and Borgatti 2020)

Figure 5: Brokerage Comparison

Defining the Bioeconomy

Table 1: Key Sectors Used in Query
SNI Code Description
02 Forestry and related services
20 Wood and wood product manufacturing except furniture
21 Pulp, paper and paper product manufacturing
36 Furniture manufacturing; other manufacturing
Table 2: Swedish Keywords used in Query: WHERE description LIKE %keyword% OR ...
Swedish English
virke timber
cellulos cellulose
lignin lignin
spån chip
bark bark
levulinsyra levulinic acid
furfural furfural
svarttjära black tar
svartlut black liquor
växtbas plant-based
ved wood
trä timber
skog forest
biobränsle biofuel
biologiskt biological
nedbrytbar biodegradable
papper paper
karton carton
lyocell lyocell

Patent Propensity

(a) Patent Propensity of Bioeconomy Innovations by Sector
(b)
Figure 6

Network Descriptives

Table 3: Descriptive Statistics of Sweden’s Innovation Collaboration Network
Total Network Only Bioeconomy Excluding Bioeconomy
Edges 2,202 294 1,152
Nodes 1,489 244 1,245
Average Degree 2.96 2.41 1.85
SD Degree 6.14 2.08 2.11
Min Degree 0 1 0
Max Degree 111 11 28