Constructing the Yield Curve for Sri Lankas Government Bond Market

Show simple item record

dc.contributor.author RATHNASINGHA, D.L.P. M.
dc.contributor.author DAYARATHNE, K.P.N.S.
dc.date.accessioned 2021-09-14T06:03:42Z
dc.date.available 2021-09-14T06:03:42Z
dc.date.issued 2021
dc.identifier.citation RATHNASINGHA, D., & DAYARATHNE, K. (2021). Constructing the Yield Curve for Sri Lankas Government Bond Market. International Journal of Business and Economic Affairs, 6, 56-69. en_US
dc.identifier.uri http://archive.cmb.ac.lk:8080/xmlui/handle/70130/5938
dc.description.abstract The primary objective of the study was to construct the most applicable model that can be used for tting yield curves for Sri Lanka Government Bonds. It is important to adhere to a scienti c method of deciding the yields for respective securities in line with the yield curve than ad-hock and personal judgments for bond Investors, traders, and monetary authorities. Nelson-Siegel model used Ordinary Least Squares after xing the Lambda parameters to make the model linear and exible enough to handle. Primary examination of the Nelson-Siegel parsimonious approach on Sri Lanka Government Securities yields t the yields having an average R2 of 91.4%. But R2 was not consistent throughout the entire study period i.e., 2010 to 2018 as R2 fell below 80% in some of the periods. Hence, the di erent discounting function was used to enhance the accuracy and predictability of the model applying Ordinary Least Squares technique. Two models were derived for the nal test and evaluation and found one model with two exponential decay functions and tenor as another predictor. This model exhibits over 97% R2 even after xed for two Lambdas. Further, the out-of-sample predictability of the model surpassed the other models with an average of higher R2 values and less Mean Residual Sum of Squared values. Clustering algorithms (listed under Weka) were used to examine the di erent patterns of the yield curve during the study period. The same model was re-constructed for identi ed clusters xing separate Lambda but the comparison of separate lambda for each cluster and overall lambda did not indicate signi cant di erences. The nal model comprises the traditional Nelson-Siegel model, but more towards the Dobbie-Wilkie Model, which accounted for two exponential dece functions and a time variable. In Sri Lanka, very few attempts have been made to understand the behavior of the yield curve. That resulted to have controversial and unorthodox practices to judge the optimal yields for respective government bonds, especially for the longer tenor bonds beyond 5 years. Monetary authorities are geared to develop and extend the Sri Lankan Government securities Bond yield curve beyond the 10 years (tenor) which will encourage foreign investors to invest in Sri Lanka bonds. Countries with active longer tenor yield curves have been bene ted from FDI. en_US
dc.language.iso en en_US
dc.subject Yield curve en_US
dc.subject Nelson-Siegel model en_US
dc.subject Government bond market en_US
dc.subject Sri Lanka en_US
dc.title Constructing the Yield Curve for Sri Lankas Government Bond Market en_US
dc.type Article en_US


Files in this item

This item appears in the following Collection(s)

Show simple item record

Search DSpace


Advanced Search

Browse

My Account