Page 79 - Contributed Paper Session (CPS) - Volume 4
P. 79

CPS2129 Matilde Bini et al.
            Response variable is the Interest Coverage from 2008 to 2017, measured on a
            continuous  scale;  it  represents  a  short-term  risk  indicator.  Time-varying
            covariate is the Leverage which is the indicator of financial distress and it varies
            across  firms  and  time.  Time-invariant  covariates  are  Sector  of  Economic
            Activity (Food and Tobacco, Textile and Leather, Wood, Publishing and Paper,
            Refining,  Chemistry  and  Rubber,  Metallurgy  and  Steel  industry,  Electric
            machines  and  Mechanical,  Means  of  transport  and  Other  industries  and
            maintenance as reference category); they vary across firms but not in time.
            Distal outcome is the ROE 2017, which is an outcome of firms’ performance
            that is predicted from the growth of Interest Coverage. Here following the
            path  diagram  of  the  LGCM  with  ten  repeated  measures  of  the  outcome
            variable (Interest coverage: Int_cov), a time-varying covariate (Leverage: Lev),
            the time invariant covariates (Sector of economic activity) and a distal outcome
            (ROE 2017):


















            5.  Result
                Here below we show the results from the analysis performed by fitting a
            LGCM. A set of alternative unconditional LGCMs with correlated measurement
            errors  between  adjacent  time  point  were  first  estimated  (linear,  quadratic,
            Piecewise linear with 2 knots and latent basis), in order to identify the more
            suitable  functional form  for  the  individual  latent  trajectories.  Based  on the
            model goodness of fit the quadratic form LGCM was preferred (RMSEA=0.028,
            CFI=0.980, TLI=0.976). The corresponding model parameters were estimated
            by  using  the  Full  Information  Maximum  Likelihood  (FIML;  Arbuckle,  1996)
            method with robust standard errors:














                                                                68 | I S I   W S C   2 0 1 9
   74   75   76   77   78   79   80   81   82   83   84