Using Ordinal Logistic Regression To Predict Terms of Payment in Shipbuilding Industry
Ante Rozga¹, Ante Luetic²
¹Faculty of Economics, University of Split, Split, Croatia; ²Brodosplit, Split, Croatia

We have analysed delay of payments towards suppliers in shipyard “Brodosplit“ in Split, Croatia. Delay of payments was ordered into four groups: 0-60 days, 61-90 days, 91-120 days and over 120 days. Several predictors were proposed at the beginning of the study. Those variables are different suppliers regarding the sort of materials they offer. The first one was “ML” which represents the most expensive material whose potential suppliers are the shipyard and the ship-owner. Variable “A” is composed of direct materials which need further analysis of the offers. Variable “N” is standard material which is specified by the list of materials whose term of delivery is within 60 days. Variable “ZM” is also standard material which is specified by special demand whose term of delivery is over 60 days. Variable “L” is the list of materials whose assortment and quantity is limited by catalogue at the annual level.

Delay of payment was treated as dependent ordered variable and we used ordinal logistic regression. Predictors were dichotomous whose modalities were “yes” or “no” which depends if supplier is on the list or not. Reference category for dependent variable was last one: terms of payments over 120 days as the longest term. Applying ordinal regression we have found two variables to be statistically significant: ZM (p = 0,002) with the estimate of regression coefficient of -3.524 and L (p = 0,075) with the estimate of regression coefficient of -1.051.

All statistics in ordinal regression were sufficient and significant and this model could be regarded as good for prediction.

Hosmer, David W. and Lemeshow, Stanley (1989). Applied Logistic Regression. John Wiley & Sons, New York.

Bender, R. & Benner, A. (2000). Calculating Ordinal Regression Models in SAS and S-Plus. Biom. Journal 42, 677-699.

Chu, W., & Ghahramani, Z. (2005). Gaussian Processes for Ordinal Regression. Journal of Machine Learning Research, 6, 1019-1041.

McCullagh, P. and Nelder (1989). Regression Models for Ordinal Data (with Discussion), Journal of the Royal Statistical Society – B 42(2), 109-142.

O'Connell, A. A. (2006). Logistic Regression Models for Ordinal Response Variables. Thousand Oaks: SAGE.

SPSS, Inc. (2002). Ordinal Regression Analysis. SPSS Advanced Models 10.0, Chicago, Il.

Biography: Ante Rozga, Ph.D. is professor of statistics at University of Split, Croatia. He graduated statistics at University of Zagreb. After graduation he continued his studies at postgraduate level at University of Zagreb and The London School of Economics and Political Science. At LSE he specialised time series analysis and attended other courses relevant for his university career. He made his Ph.D. in statistics.He was director of Faculty of Economics. He was visiting professor at the University of Venice, University of Padova, The London School of Economics and Faculty of Economics and Politics, Cambridge. He was active participant in many conferences and congresses related to statistics and education process. He took part in many research projects in Croatia and abroad.Professor Rozga took active part in the project of financing higher education, sponsored by Unesco.