___ ____ ____ ____ ____(R) /__ / ____/ / ____/ ___/ / /___/ / /___/ Statistics/Data Analysis Title dsimih use -- load dynamic simulation results generated by dsimih create into Stata data set memory Syntax dsimih use [filename] [, noci se level(#) clear] options description ------------------------------------------------------------------------- noci do not create variables for upper and lower confidence bounds se create variables that contain standard errors of dsimih stats setypes(selist) create variables for standard errors and confidence intervals based on standard errors of type(s) selist only; default: all setypes level(#) set confidence level clear specifies that it is okay to replace the data in memory, even though the current data have not been saved to disk number(#) access results number # ; only allowed if the using modifier refers to a .ster file ------------------------------------------------------------------------- Description dsimih use retrieves dynamic simulation results generated by dsimih create into Stata data set memory. Results can be pulled from e(), from a .ster file, or from a dsimih file. Abbreviations, definitions, notation, syntax elements This help entry uses terminology defined in svarih and dsimih. Options noci prevents dsimih use from creating variables that contain lower and upper confidence bounds at level #. By default, these variables are created based on asymptotic and for (parametric and residual) bootstrapped standard errors, which makes a total of six variables per stat. If e(dsimih) does not contain numbers for a particular type of standard error, the confidence interval variables are still created by dsimih use, with all values set to missing. se causes dsimih use to create variables that contain standard errors for SIRFs and SFEVDs. Series are created for asymptotic and for (parametric and residual) bootstrapped standard errors, which makes a total of three variables created for each dsimih stat. If e(dsimih) does not contain numbers for a particular type of standard error, the standard error variables are still created by dsimih use, with all values set to missing. setypes(selist) specifies the standard errors types based on which standard error variables and/or confidence band variables are created. selist may contain individual tokens asymptotic, bs and bsp. They stand for asymptotic standard errors, standard errors from a residual bootstrap, and standard errors from a parametric bootstrap based on draws from the normal distribution, respectively. Specifying multiple standard error types as in setypes(asymp bsp) is allowed. The default is to create variables for all setypes. level(#) specifies the confidence level, as a percentage, for confidence intervals, when they are reported. The default is level(95) or as set by set level. clear specifies that it is okay to replace the data in memory, even though the current data have not been saved to disk. Remarks dsimih use retrieves dynamic simulation results generated by dsimih create into Stata data set memory. It accesses DS results in e() by default but is also capable of accessing DS results stored in files using the optional filename. For details on how the latter works, see dsimih etodta. It is important to note that dsimih use, by default, does not load the results data as-is. When loading dsimih data, dsimih use uses the variables that contain information on the different types of standard errors to construct variables that contain upper and lower confidence bounds for the dsimih stats 'sirf' and 'sfevd'. Then, by default, dsimih use drops all standard error variables. Moreover, since variables 'impulse' and 'response' are stored under a numeric encoding in e(), dsimih use converts those variables to appropriate string variables. This is not necessary if you load a dsimih file since it already contains variables 'impulse' and 'response' as string variables. It is imperative to note that, due to the default loading actions just mentioned, the data set that you see when typing browse after . dsimih use is different from what is actually stored in e() or in filename. If you want to see the exact contents of e() or of filename, type . dsimih use [filename] , se noci Examples Excecuting the following statements will change current e()-results. Bootstrap replication numbers are set to values that are inappropriate for analysis but appropriate for quick execution of example statements. Generate example estimates (see svarih examples): . webuse lutkepohl2 . svarih examples bfa_constr , ereplace . dsimih describe , modelstats cmdline . dsimih create . version 11.2: set seed 123456 . dsimih create , step(12) bs fromb reps(10) Pull DS results from e() into data set memory: . dsimih use , clear . describe Customize on setype and confidence level: . dsimih use , setype(asy bs) level(68) . describe To get the data that is actually stored, use options se and noci. . dsimih use , se noci . describe Data can also be pulled from saved .ster files. . drop _all . estimates save dsimih_examplefile , replace . dsimih use dsimih_examplefile.ster , setype(bs) . describe It is also possible to retrieve data from dsimih files created by dsimih etodta. . drop _all . dsimih etodta dsimih_examplefile2.dsimih , replace These are the variables that are actually stored in the file: . describe using dsimih_examplefile2.dsimih These are the variables in memory when the data is loaded without any options . dsimih use dsimih_examplefile2.dsimih . describe . erase dsimih_examplefile.ster . erase dsimih_examplefile2.dsimih Author Daniel C. Schneider, Goethe University Frankfurt, dan_schneider@outlook.com Also see Help: [TS] irf, dsimih, dsimih create, dsimih table, dsimih graph, dsimih describe, dsimih drop, dsimih etodta