July  2020, 7(3): 239-252. doi: 10.3934/jdg.2020017

Network ANOVA random effects models for node attributes

1. 

Instituto Interdisciplinario de Economía Política de Buenos Aires, (IIEP-BAIRES-UBA)-CONICET, Facultad de Ciencias Económicas, Universidad de Buenos Aires, Av. Córdoba 2122 2do piso, C1120AAQ, Ciudad Autónoma de Buenos Aires, Argentina

2. 

Banco Central de la República Argentina and Universidad Nacional de La Plata

* Corresponding author: Gabriel Montes-Rojas

We are grateful to the Guest Editors, Viktoriya Semeshenko, Gabriel Brida and Andrea Roventini, and to two anonymous referees for helpful comments and suggestions. Matías Pardini and Emilio Sáenz Guillén provided invaluable research assistance. We also thank the Gerencia Principal de Estadísticas Económicas of Banco Central de la República Argentina for the data used in the empirical application.

Received  February 2020 Revised  April 2020 Published  July 2020

This paper develops a subgraph network random effects error components structure for network data to perform analysis of variance. In particular, it proposes a model for evaluating the network interdependence of nodes attributes allowing for edge and triangle specific components. The latter serve as a basal model for modeling more general network effects. Consistent estimators of the variance components and Lagrange Multiplier specification tests for evaluating the appropriate model of random components in networks structures is proposed. Monte Carlo simulations show that the tests have good performance in finite samples. The proposed tests is applied to the unsecured (Call) interbank market network in Argentina.

Citation: Gabriel Montes-Rojas, Pedro Elosegui. Network ANOVA random effects models for node attributes. Journal of Dynamics & Games, 2020, 7 (3) : 239-252. doi: 10.3934/jdg.2020017
References:
[1]

D. AcemogluA. Ozdaglar and A. Tahbaz-Salehi, Systemic risk and stability in financial networks, American Economic Review, 105 (2015), 564-608.  doi: 10.3386/w18727.  Google Scholar

[2]

P. AngeliniA. Nobili and C. Picillo, The interbank market after August 2007: What has changed and why?, Journal of Money, Credit and Banking, 43 (2011), 923-958.   Google Scholar

[3]

L. AnselinA. BeraR. Florax and M. Yoon, Simple diagnostic tests for spatial dependence, Regional Science and Urban Economics, 26 (1996), 77-104.  doi: 10.1016/0166-0462(95)02111-6.  Google Scholar

[4]

L. BargigliG. di IasioL. InfanteF. Lillo and F. Pierobon, The multiplex structure of interbank networks, Quantitative Finance, 15 (2015), 673-691.  doi: 10.1080/14697688.2014.968356.  Google Scholar

[5]

S. Battiston, M. Puliga, R. Kaushik, P. Tasca and G. Caldarelli, Debtrank: Too central to fail? Financial networks, the fed and systemic risk, Scientific Reports, 2. doi: 10.1038/srep00541.  Google Scholar

[6]

M. L. Bech, J. T. E. Chapman and R. J. Garratt, Which bank is the central bank?, Journal of Monetary Economics, 57. doi: 10.1016/j.jmoneco.2010.01.002.  Google Scholar

[7]

A. Bera and Y. Bilias, Rao's score, Neyman's $c(\alpha)$ and Silvey's lm tests: An essay on historical developments and some new results, Journal of Statistical Planning and Inference, 97 (2001), 9-44.  doi: 10.1016/S0378-3758(00)00343-8.  Google Scholar

[8]

A. BeraG. Montes-Rojas and W. Sosa-Escudero, General specification testing with locally misspecified models, Econometric Theory, 26 (2010), 1838-1845.  doi: 10.1017/S0266466609990818.  Google Scholar

[9]

A. BeraG. Montes-Rojas and W. Sosa-Escudero, A new robust and most powerful test in the presence of local misspecification, Communications in Statistics - Theory and Methods, 46 (2017), 8187-8198.  doi: 10.1080/03610926.2016.1177077.  Google Scholar

[10]

A. Bera and M. Yoon, Specification testing with locally misspecified alternatives, Econometric Theory, 9 (1993), 649-658.  doi: 10.1017/S0266466600008021.  Google Scholar

[11]

G. G. BoothU. G. Gurun and H. Zhang, Financial networks and trading in bond markets, Journal of Financial Markets, 18 (2014), 126-157.  doi: 10.1016/j.finmar.2013.08.001.  Google Scholar

[12] A. G. Chandrasekhar, Econometrics of network formation, in The Oxford Handbook of the Economics of Networks, Oxford University Press, Oxford, 2016.   Google Scholar
[13]

A. de Paula, Econometrics of network models, CEMMAP Working Paper CWP52/15, (2017).  Google Scholar

[14]

S. de Raco and V. Semeshenko, Labor mobility and industrial space in Argentina, Journal of Dynamics & Games, 6 (2019), 107-118.  doi: 10.3934/jdg.2019008.  Google Scholar

[15]

B. S. Graham, Network data, NBER Working Paper 26577, (2019). Google Scholar

[16]

H. H. Kelejian and I. R. Prucha, A generalized moments estimator for the autoregressive parameter in a spatial model, International Economic Review, 40 (1999), 509-533.  doi: 10.1111/1468-2354.00027.  Google Scholar

[17]

H. H. Kelejian and I. R. Prucha, Specification and estimation of spatial autoregressive models with autoregressive and heteroskedastic disturbances, Journal of Econometrics, 157 (2010), 53-67.  doi: 10.1016/j.jeconom.2009.10.025.  Google Scholar

[18]

Y. Kumagai, Social networks and global transactions, Journal of Dynamics & Games, 6 (2019), 211-219.  doi: 10.3934/jdg.  Google Scholar

[19]

S. LangfieldZ. Liu and T. Ota, Mapping in the UK interbank system, Journal of Banking & Finance, 45 (2014), 288-303.   Google Scholar

[20]

J.-L. Molina-BorboaS. Martinez-JaramilloF. Lopez-Gallo and M. van der Leij, A multiplex network analysis of the mexican banking system: Link persistence, overlap and waiting times, Journal of Network Theory in Finance, 1 (2015), 99-138.  doi: 10.21314/JNTF.2015.006.  Google Scholar

[21]

S. PolednaJ. L. Molina-BorboaS. Martínez-JaramilloM. van der Leij and S. Thurner, The multi-layer network nature of systemic risk and its implications for the costs of financial crises, Journal of Financial Stability, 20 (2015), 70-81.  doi: 10.1016/j.jfs.2015.08.001.  Google Scholar

[22]

A. TemizsoyG. Iori and G. Montes-Rojas, The role of bank relationship in the interbank market, Journal of Economic Dynamics & Control, 59 (2015), 118-141.  doi: 10.1016/j.jedc.2015.07.008.  Google Scholar

[23]

A. TemizsoyG. Iori and G. Montes-Rojas, Network centrality and funding rates in the e-mid interbank market, Journal of Financial Stability, 33 (2017), 346-365.  doi: 10.1016/j.jfs.2016.11.003.  Google Scholar

[24]

C. Upper, Simulation methods to assess the danger of contagion in interbank markets, Journal of Financial Stability, 7 (2011), 111-125.  doi: 10.1016/j.jfs.2010.12.001.  Google Scholar

show all references

References:
[1]

D. AcemogluA. Ozdaglar and A. Tahbaz-Salehi, Systemic risk and stability in financial networks, American Economic Review, 105 (2015), 564-608.  doi: 10.3386/w18727.  Google Scholar

[2]

P. AngeliniA. Nobili and C. Picillo, The interbank market after August 2007: What has changed and why?, Journal of Money, Credit and Banking, 43 (2011), 923-958.   Google Scholar

[3]

L. AnselinA. BeraR. Florax and M. Yoon, Simple diagnostic tests for spatial dependence, Regional Science and Urban Economics, 26 (1996), 77-104.  doi: 10.1016/0166-0462(95)02111-6.  Google Scholar

[4]

L. BargigliG. di IasioL. InfanteF. Lillo and F. Pierobon, The multiplex structure of interbank networks, Quantitative Finance, 15 (2015), 673-691.  doi: 10.1080/14697688.2014.968356.  Google Scholar

[5]

S. Battiston, M. Puliga, R. Kaushik, P. Tasca and G. Caldarelli, Debtrank: Too central to fail? Financial networks, the fed and systemic risk, Scientific Reports, 2. doi: 10.1038/srep00541.  Google Scholar

[6]

M. L. Bech, J. T. E. Chapman and R. J. Garratt, Which bank is the central bank?, Journal of Monetary Economics, 57. doi: 10.1016/j.jmoneco.2010.01.002.  Google Scholar

[7]

A. Bera and Y. Bilias, Rao's score, Neyman's $c(\alpha)$ and Silvey's lm tests: An essay on historical developments and some new results, Journal of Statistical Planning and Inference, 97 (2001), 9-44.  doi: 10.1016/S0378-3758(00)00343-8.  Google Scholar

[8]

A. BeraG. Montes-Rojas and W. Sosa-Escudero, General specification testing with locally misspecified models, Econometric Theory, 26 (2010), 1838-1845.  doi: 10.1017/S0266466609990818.  Google Scholar

[9]

A. BeraG. Montes-Rojas and W. Sosa-Escudero, A new robust and most powerful test in the presence of local misspecification, Communications in Statistics - Theory and Methods, 46 (2017), 8187-8198.  doi: 10.1080/03610926.2016.1177077.  Google Scholar

[10]

A. Bera and M. Yoon, Specification testing with locally misspecified alternatives, Econometric Theory, 9 (1993), 649-658.  doi: 10.1017/S0266466600008021.  Google Scholar

[11]

G. G. BoothU. G. Gurun and H. Zhang, Financial networks and trading in bond markets, Journal of Financial Markets, 18 (2014), 126-157.  doi: 10.1016/j.finmar.2013.08.001.  Google Scholar

[12] A. G. Chandrasekhar, Econometrics of network formation, in The Oxford Handbook of the Economics of Networks, Oxford University Press, Oxford, 2016.   Google Scholar
[13]

A. de Paula, Econometrics of network models, CEMMAP Working Paper CWP52/15, (2017).  Google Scholar

[14]

S. de Raco and V. Semeshenko, Labor mobility and industrial space in Argentina, Journal of Dynamics & Games, 6 (2019), 107-118.  doi: 10.3934/jdg.2019008.  Google Scholar

[15]

B. S. Graham, Network data, NBER Working Paper 26577, (2019). Google Scholar

[16]

H. H. Kelejian and I. R. Prucha, A generalized moments estimator for the autoregressive parameter in a spatial model, International Economic Review, 40 (1999), 509-533.  doi: 10.1111/1468-2354.00027.  Google Scholar

[17]

H. H. Kelejian and I. R. Prucha, Specification and estimation of spatial autoregressive models with autoregressive and heteroskedastic disturbances, Journal of Econometrics, 157 (2010), 53-67.  doi: 10.1016/j.jeconom.2009.10.025.  Google Scholar

[18]

Y. Kumagai, Social networks and global transactions, Journal of Dynamics & Games, 6 (2019), 211-219.  doi: 10.3934/jdg.  Google Scholar

[19]

S. LangfieldZ. Liu and T. Ota, Mapping in the UK interbank system, Journal of Banking & Finance, 45 (2014), 288-303.   Google Scholar

[20]

J.-L. Molina-BorboaS. Martinez-JaramilloF. Lopez-Gallo and M. van der Leij, A multiplex network analysis of the mexican banking system: Link persistence, overlap and waiting times, Journal of Network Theory in Finance, 1 (2015), 99-138.  doi: 10.21314/JNTF.2015.006.  Google Scholar

[21]

S. PolednaJ. L. Molina-BorboaS. Martínez-JaramilloM. van der Leij and S. Thurner, The multi-layer network nature of systemic risk and its implications for the costs of financial crises, Journal of Financial Stability, 20 (2015), 70-81.  doi: 10.1016/j.jfs.2015.08.001.  Google Scholar

[22]

A. TemizsoyG. Iori and G. Montes-Rojas, The role of bank relationship in the interbank market, Journal of Economic Dynamics & Control, 59 (2015), 118-141.  doi: 10.1016/j.jedc.2015.07.008.  Google Scholar

[23]

A. TemizsoyG. Iori and G. Montes-Rojas, Network centrality and funding rates in the e-mid interbank market, Journal of Financial Stability, 33 (2017), 346-365.  doi: 10.1016/j.jfs.2016.11.003.  Google Scholar

[24]

C. Upper, Simulation methods to assess the danger of contagion in interbank markets, Journal of Financial Stability, 7 (2011), 111-125.  doi: 10.1016/j.jfs.2010.12.001.  Google Scholar

Figure 1.  LM tests for edge effects, $ \sigma_\mu^2=0 $, Erdös-Rényi random graph

Notes: Monte carlo experiments based on 2000 replications. Solid line: $ LM_{\mu} $. Dashed line: $ LM_{\mu\delta} $. Dotted line: $ LM_{\mu(\delta)} $

Figure 2.  LM tests for edge effects, $ \sigma_\mu^2=0 $, Queen spatial structure

Notes: Monte carlo experiments based on 2000 replications. Solid line: $ LM_{\mu} $. Dashed line: $ LM_{\mu\delta} $. Dotted line: $ LM_{\mu(\delta)} $.

Figure 3.  LM tests for triangle effects, $ \sigma_\delta^2=0 $, Erdös-Rényi random graph

Notes: Monte carlo experiments based on 2000 replications. Solid line: $ LM_{\delta} $. Dashed line: $ LM_{\mu\delta} $. Dotted line: $ LM_{\delta(\mu)} $. Dash-dot line: $ LM_{\delta-\mu} $

Figure 4.  LM tests for triangle effcts, $ \sigma_\delta^2=0 $, Queen spatial structure

Notes: Monte carlo experiments based on 2000 replications. Solid line: $ LM_{\delta} $. Dashed line: $ LM_{\mu\delta} $. Dotted line: $ LM_{\delta(\mu)} $. Dash-dot line: $ LM_{\delta-\mu} $

Figure 5.  Subgraph joint tests for edge and triangle effects and spatial Moran's Ⅰ LM test

Note: P-values in log-scale. Dashed line is the 10% critical value and dotted line to the 5% critical values. Horizontal axis corresponds to joint test for edge and triangle effects (LMµ, δ). Vertical axis corresponds to Moran's Ⅰ LM tests for spatial error.

Figure 6.  Subgraph tests for edge and triangle effects

Note: P-values in log-scale. Dashed line is the 10% critical value and dotted line to the 5% critical values. Horizontal axis corresponds to tests for edge effects (LMµ). Vertical axis corresponds to tests for triangle effects (LMδ).

Figure 7.  Robust subgraph tests for edge and triangle effects

Note: P-values in log-scale. Dashed line is the 10% critical value and dotted line to the 5% critical values. Horizontal axis corresponds to tests for edge effects robust to triangle effects (LMµ(δ)). Vertical axis corresponds to tests for triangle effects robust to edge effects (LMδ(µ)).

Table 1.  Empirical size
N LMµ LMδ LMµ,δ LMµ(δ) LMδ(µ) LMδµ
Erdös-Rényi random graph
Size 1%
100 0.009 0.016 0.0145 0.009 0.0165 0.0115
225 0.012 0.0115 0.015 0.013 0.012 0.009
400 0.013 0.012 0.0085 0.0095 0.0075 0.007
Size 5%
100 0.043 0.05 0.0465 0.042 0.052 0.041
225 0.052 0.0485 0.0495 0.052 0.0495 0.041
400 0.047 0.0475 0.049 0.046 0.046 0.0435
Size 10%
100 0.082 0.0885 0.0855 0.089 0.092 0.0765
225 0.1045 0.092 0.102 0.098 0.0995 0.0875
400 0.089 0.087 0.093 0.0965 0.099 0.0915
Spatial queen structure
Size 1%
100 0.0115 0.0105 0.0105 0.01 0.011 0.0115
225 0.0075 0.0065 0.012 0.0145 0.0135 0.014
400 0.0085 0.0085 0.0095 0.012 0.011 0.011
Size 5%
100 0.0475 0.0515 0.047 0.048 0.044 0.046
225 0.045 0.039 0.0565 0.0595 0.052 0.0525
400 0.046 0.0465 0.049 0.0535 0.049 0.0505
Size 10%
100 0.0965 0.0975 0.0955 0.094 0.09 0.097
225 0.0965 0.09 0.1 0.1085 0.1115 0.1125
400 0.0935 0.0965 0.098 0.096 0.0995 0.1015
Notes: Monte carlo experiments based on 2000 replications.
N LMµ LMδ LMµ,δ LMµ(δ) LMδ(µ) LMδµ
Erdös-Rényi random graph
Size 1%
100 0.009 0.016 0.0145 0.009 0.0165 0.0115
225 0.012 0.0115 0.015 0.013 0.012 0.009
400 0.013 0.012 0.0085 0.0095 0.0075 0.007
Size 5%
100 0.043 0.05 0.0465 0.042 0.052 0.041
225 0.052 0.0485 0.0495 0.052 0.0495 0.041
400 0.047 0.0475 0.049 0.046 0.046 0.0435
Size 10%
100 0.082 0.0885 0.0855 0.089 0.092 0.0765
225 0.1045 0.092 0.102 0.098 0.0995 0.0875
400 0.089 0.087 0.093 0.0965 0.099 0.0915
Spatial queen structure
Size 1%
100 0.0115 0.0105 0.0105 0.01 0.011 0.0115
225 0.0075 0.0065 0.012 0.0145 0.0135 0.014
400 0.0085 0.0085 0.0095 0.012 0.011 0.011
Size 5%
100 0.0475 0.0515 0.047 0.048 0.044 0.046
225 0.045 0.039 0.0565 0.0595 0.052 0.0525
400 0.046 0.0465 0.049 0.0535 0.049 0.0505
Size 10%
100 0.0965 0.0975 0.0955 0.094 0.09 0.097
225 0.0965 0.09 0.1 0.1085 0.1115 0.1125
400 0.0935 0.0965 0.098 0.096 0.0995 0.1015
Notes: Monte carlo experiments based on 2000 replications.
[1]

Mahendra Piraveenan, Mikhail Prokopenko, Albert Y. Zomaya. On congruity of nodes and assortative information content in complex networks. Networks & Heterogeneous Media, 2012, 7 (3) : 441-461. doi: 10.3934/nhm.2012.7.441

[2]

W. Patrick Hooper, Richard Evan Schwartz. Billiards in nearly isosceles triangles. Journal of Modern Dynamics, 2009, 3 (2) : 159-231. doi: 10.3934/jmd.2009.3.159

[3]

W. Patrick Hooper, Richard Evan Schwartz. Erratum: Billiards in nearly isosceles triangles. Journal of Modern Dynamics, 2014, 8 (1) : 133-137. doi: 10.3934/jmd.2014.8.133

[4]

Alexander Blokh. Necessary conditions for the existence of wandering triangles for cubic laminations. Discrete & Continuous Dynamical Systems - A, 2005, 13 (1) : 13-34. doi: 10.3934/dcds.2005.13.13

[5]

Robert W. Ghrist. Flows on $S^3$ supporting all links as orbits. Electronic Research Announcements, 1995, 1: 91-97.

[6]

Hans G. Kaper, Bixiang Wang, Shouhong Wang. Determining nodes for the Ginzburg-Landau equations of superconductivity. Discrete & Continuous Dynamical Systems - A, 1998, 4 (2) : 205-224. doi: 10.3934/dcds.1998.4.205

[7]

Andrey Yu. Verisokin, Darya V. Verveyko, Eugene B. Postnikov, Anastasia I. Lavrova. Wavelet analysis of phase clusters in a distributed biochemical system. Conference Publications, 2011, 2011 (Special) : 1404-1412. doi: 10.3934/proc.2011.2011.1404

[8]

Iliya D. Iliev, Chengzhi Li, Jiang Yu. Bifurcations of limit cycles in a reversible quadratic system with a center, a saddle and two nodes. Communications on Pure & Applied Analysis, 2010, 9 (3) : 583-610. doi: 10.3934/cpaa.2010.9.583

[9]

Jianfeng Jia, Xuewei Liu, Yixin Zhang, Zhe Li, Yanjie Xu, Jiaqi Yan. Rumor propagation controlling based on finding important nodes in complex network. Journal of Industrial & Management Optimization, 2020, 16 (5) : 2521-2529. doi: 10.3934/jimo.2019067

[10]

Theodore Kolokolnikov, Juncheng Wei. Hexagonal spike clusters for some PDE's in 2D. Discrete & Continuous Dynamical Systems - B, 2020  doi: 10.3934/dcdsb.2020039

[11]

Sergey V. Dmitriev, Asiya A. Nazarova, Anatoliy I. Pshenichnyuk, Albert M. Iskandarov. Dynamics of edge dislocation clusters interacting with running acoustic waves. Discrete & Continuous Dynamical Systems - S, 2011, 4 (5) : 1079-1094. doi: 10.3934/dcdss.2011.4.1079

[12]

Patrick Flynn, Quinton Neville, Arnd Scheel. Self-organized clusters in diffusive run-and-tumble processes. Discrete & Continuous Dynamical Systems - S, 2020, 13 (4) : 1187-1208. doi: 10.3934/dcdss.2020069

[13]

Martin Wechselberger, Warren Weckesser. Homoclinic clusters and chaos associated with a folded node in a stellate cell model. Discrete & Continuous Dynamical Systems - S, 2009, 2 (4) : 829-850. doi: 10.3934/dcdss.2009.2.829

[14]

Donghi Lee, Makoto Sakuma. Simple loops on 2-bridge spheres in Heckoid orbifolds for 2-bridge links. Electronic Research Announcements, 2012, 19: 97-111. doi: 10.3934/era.2012.19.97

[15]

Marco Tosato, Jianhong Wu. An application of PART to the Football Manager data for players clusters analyses to inform club team formation. Big Data & Information Analytics, 2018  doi: 10.3934/bdia.2018002

[16]

K. Habibul Kabir, Masahiro Sasabe, Tetsuya Takine. Optimal visiting order of isolated clusters in DTNs to minimize the total mean delivery delay of bundles. Numerical Algebra, Control & Optimization, 2011, 1 (4) : 563-576. doi: 10.3934/naco.2011.1.563

[17]

Hassan Allouba. Brownian-time Brownian motion SIEs on $\mathbb{R}_{+}$ × $\mathbb{R}^d$: Ultra regular direct and lattice-limits solutions and fourth order SPDEs links. Discrete & Continuous Dynamical Systems - A, 2013, 33 (2) : 413-463. doi: 10.3934/dcds.2013.33.413

[18]

Mapundi K. Banda, Michael Herty, Axel Klar. Gas flow in pipeline networks. Networks & Heterogeneous Media, 2006, 1 (1) : 41-56. doi: 10.3934/nhm.2006.1.41

[19]

Radu C. Cascaval, Ciro D'Apice, Maria Pia D'Arienzo, Rosanna Manzo. Flow optimization in vascular networks. Mathematical Biosciences & Engineering, 2017, 14 (3) : 607-624. doi: 10.3934/mbe.2017035

[20]

Juan Manuel Pastor, Javier García-Algarra, José M. Iriondo, José J. Ramasco, Javier Galeano. Dragging in mutualistic networks. Networks & Heterogeneous Media, 2015, 10 (1) : 37-52. doi: 10.3934/nhm.2015.10.37

 Impact Factor: 

Article outline

Figures and Tables

[Back to Top]