1. The original discussion on the bank lending channel by Bernanke and Blinder (1988) also raises the possibility of a link between the perceived riskiness of loans and banks' supply of loans, but this is not done in the context of changes in monetary policy.
2. Garcia-Appendini, Gatti, and Nocera (2017) analyze how banks' funding policies using covered bonds can affect the willingness of banks to lend and change the overall asset risk of intermediaries.
3. In a default, lenders may have to make fresh exposures to stressed projects to enable them to make good on accumulating interest payments.
4. Equal sharing is assumed based on the literature; see, for instance, Esty (2001). The author assumes equal commitment on the loan underwriting when estimating the fee distribution among lead and colead arrangers. Gatti et al. (2013) use the same assumption where no data on the composition of the syndicate are provided.
5. In virtually all instances, the composition of the syndicate is the same across tranches for a project. In cases where banks differ across tranches (10 projects), syndicate size is defined as the number of unique banks providing financing across all tranches.
6. The multilateral development banks in the sample are the Asian Development Bank and the International Finance Corporation; the bilateral development banks are China Development Bank, Development Bank of Japan, Netherlands Development Finance Company, KfW IPEX-Bank GmbH, and Korea Development Bank.
7. Because longer-tenor bank loans will be more difficult to come by once Basel III capital standards take effect, the role of debt capital markets will gain further prominence.
8. Following Acemoglu et al. (2003), volatility of GDP growth is measured as the standard deviation of lagged values of GDP growth rates. Volatility of the inflation rate follows the same logic. Volatility of exchange rates is calculated as the standard deviation of the first difference of logarithms of the local currency unit versus the dollar exchange rate (Clark et al. 2004). The volatility estimates use data of the 20 lagged quarters together with the current quarter.
9. Clustering standard errors at the syndicate level implies that correlation of the error term or terms within a syndicate is controlled, which is the level at which the variables or interests are observed. Thus, the possibility that projects (error terms of the study's projects) are not independently and identically distributed within the same syndicate because of omitted variables at the syndicate level is effectively accounted for.
10. To deal with the problem of limited variation in the governance indicators in the country variables, two sets of regressions are conducted: one without country variables (time-varying, albeit with limited variation, and time-invariant) and one with country variables. Wherever possible, country variables at the quarterly frequency are used to gain more variation over time.
11. Another clear trend emerging in the debt capital market for project finance transactions is the development of various forms of partnerships between traditional bank lenders and institutional investors, where banks provide know-how in screening and monitoring projects, and play a delegated monitoring role on behalf of institutional investors joining a partnership (OECD 2014).