porate bonds are endogenously determined.7 As a result, empirical estimations of
eqs.(1) and (2) are carried out by means of Three-Stage Least Squares (3SLS).8
3 Results
The empirical estimates of the system of eqs.(1) and (2) are reported in Ta-
ble 1. The two regressions are initially estimated without any lagged dependent
variables. Then, in order to account for lagged dependent variables which might
become significant, the two regressions are supplemented with lags of order one,
two, and so forth. Following this procedure it can be shown that eq.(1) includes
lagged dependent variables up to the fourth lag, while eq.(2) includes lagged de-
pendent variables up to the fifth lag.9
The variable that exerts the strongest influence on primary placements of shares
is the returns on S&P500 which are significant at the 1% level. This result is
in line with the literature.10 The coefficient associated with volatility is posi-
tive, but not statistically significant. Consequently, the volatility of stock market
prices does not affect primary placements of shares. This surprising result can
be explained recalling that primary placements include not only IPOs of privately
owned firms, but they also include the issuance of convertible bonds and capital in-
7To save space these results are not reported.
8The model has been supplemented with dummy variables to account for monthly seasonality,
the Stock Market Crash of October 1987 and the collapse of LTCM fund of November 1998.
9When lagged dependent variables of order higher than four and five are included, empirical
results show that these terms are not statistically significant. Thus, the identification process of the
two regressions implies kS = 4 and kB = 5.
10See, for instance, Lowry (2003) and Welch (2004).