Data for: Are syndicated loans really cheaper?

This database contains information on large corporate and middle market commercial loans filed with the Securities and Exchange Commission, or obtained through other reliable public sources. The Compustat database complements the former database by providing information on borrowers. The fusion of the two databases uses the link table provided by Roberts and Chava (2008). Each contract listed in DealScan, referred to as a deal, consists of one or several facilities or tranches. DealScan provides a unique identification number for each deal allowing the identification of all the tranches belonging to the same deal. Different tranches in a deal can show heterogeneous characteristics in terms of interest rate spreads, amount, currency, maturity, default probability, among others. In particular, not all members of a syndicated loan participate in every tranche of the deal. Our initial information was composed of deals originated in the US market for the period 1986-2013 (inclusive) for which borrower accounting information is available in Compustat. In our empirical analysis, several measures related to the syndicate structure and its previous relationships with the borrower require using the information of the four years prior to each deal active date. This implies that deals from 1986-1989 are only used for this purpose, and therefore the final sample only considers loans issued between 1990 and 2013. Additionally, some loans are excluded from the 1990-2013 period. First, loans to borrowers that are government entities, banks, or financial institutions, and/or regulated borrowers, such as transportation and public utilities (industries identified as SIC 91-99, 60-67 and 40-49). Second, deals where at least one tranche is denominated in currencies other than the US dollar, or for which some basic information is missing, such as the facility amount or tranche interest spread. For the sake of homogeneity, deals with a base rate other than LIBOR were also excluded. After this cleaning process, the final sample consists of 32,102 tranches corresponding to 21,034 deals to 5,206 borrowers over the period 1990-2013. However, of those 32,102 potential observations, a smaller number have complete information for all the variables that we will use in each particular econometric model.本数据库收录了向美国证券交易委员会（Securities and Exchange Commission, SEC）提交备案的大型企业与中端市场商业贷款信息，同时包含通过其他可靠公开渠道获取的同类贷款数据。康普斯塔特数据库（Compustat）通过提供借款方相关信息，对本数据库形成补充。两个数据库的匹配融合，通过Roberts与Chava（2008）提供的关联表实现。 DealScan数据库（DealScan）中收录的每一份贷款合约（亦称“交易”，deal），均包含一个或多个授信额度或分档（tranche）。DealScan为每一笔交易分配唯一标识码，可用于识别归属同一交易的所有分档。同一交易内的不同分档，在利率利差、授信规模、计价货币、到期期限、违约概率等维度上，均可能呈现异质性特征。尤为值得注意的是，银团贷款的全体参与方并非都会参与该交易下的每一分档授信。 本研究初始数据集涵盖1986年至2013年（含首尾两年）美国市场内发放的交易，且其借款方的会计信息可在康普斯塔特数据库中获取。在实证分析环节，多项与银团结构、银团与借款方既往合作关系相关的指标，需要使用每笔交易生效日前四年的相关信息。这意味着1986年至1989年的交易仅用于前述指标的计算，因此最终样本仅纳入1990年至2013年间发放的贷款。 此外，1990年至2013年区间内的部分贷款也被剔除出样本：其一，借款方为政府实体、银行或其他金融机构的贷款，以及受监管的借款方（如交通运输与公共事业行业，对应标准行业分类码（Standard Industrial Classification, SIC）91-99、60-67与40-49）相关贷款均被剔除；其二，存在至少一个分档以非美元货币计价的交易，或缺失部分基础信息（如授信额度、分档利率利差）的交易均被剔除。为保证样本同质性，基准费率不为伦敦同业拆借利率（London Interbank Offered Rate, LIBOR）的交易也被一并剔除。 经过上述清洗流程后，1990年至2013年间的最终样本共包含32102个分档，对应21034笔交易，覆盖5206家借款方。但在这32102条潜在观测样本中，仅有更少一部分具备后续各计量模型所需的全部变量完整信息。