Program

Detailed program of the workshop

Main topic areas

• What's (new) in Iris
• Data management
• Empirical time series modeling
• Structural modeling
• Reporting

---

A. What's (new) in Iris?

Block A1 [Quick overview]: Overview of most common Iris functionality, and its hierarchy

Data management

• +databank package
• Dater objects, +dater package
• Series objects
• +x13 package
• NamedMatrix objects

Structural modeling

• Model objects
• Plan objects
• Explanatory objects

Time series modeling

• VAR and SVAR objects
• Dynafit objects
• Armani objects

Reporting functionality

• Chartpack object
• +rephrase package

Shrinkage estimation utilities

• Posterior object
• SystemProperty, SystemPrior, SystemPriorWrapper objects
• +distribution package
• +dummy package

Block A2: Fundamental changes in Iris and the reasons behind

• (Almost) everything is an object or a namespace (package)
• New front-end objects to replace older functionality, and the reasons behind
• New front-end namespaces (packages)
• Static constructors

New object	to replace...
Series	tseries
Model	model
Plan	plan
Explanatory	rpteq
Dater	dates functions

New package	to replace
+databank	db... functions
+rephrase	+report
+x13	x13 functions
+dater	dates functions

• Why new implementation of time series?
• Why new implementation of models and simulation plans?
• Why a new reporting package?

Block A3 [Quick overview]: New in-house algorithms

• Kalman filter
• Nonlinear solver in steady state
• Nonlinear solver in dynamic simulations: stacked-time vs period-by-period
• Backward looking models with no steady state
• Time frames in simulation plans
• Posterior simulator
• System priors

---

B. Data management

Block B1: The concept of databanks, dates and time series in Iris

• Matlab structs as databanks
• Matlab timeseries vs Iris Series
• Matlab datetime vs Iris Dater and +dater

Block B2: Importing, downloading, exporting data

• API for Fred, IMF, ECB
• Importing/exporting data to/from CSV and Excel files

Block B3: Batch processing of data

• Databank functions apply, copy, filterFields, batch, merge
• Naming convetions based on prefixes, suffixes
• Using Explanatory objects for preprocessing and postprocessing

Block B4: Time series manipulation, UV filtering, X13

• Univariate and multivariate time series, concatenation
• Basic interpolation/extrapolation
• Clipping, rebasing, redating time series
• Differencing and cumulating time series
• HP filter, local-level filter, Butterworth filter, cutoff frequency/periodicity
• ARMA reconstructors/deconstructors
• New implementation X31 interface

---

C. [Lower priority than D] Empirical time series modeling

Block C1: Vector autoregressions, dynamic factor models

• Specifying VARs, DFMs, VARs with exogenous variables
• Plain vanilla estimation of VARs and DFMs
• Filtering and simulating VARs

Block C2: Shrinkage estimators for VARs

• Constructing and adding prior dummy observations: why and how
• Litterman priors, asymptotic mean priors, sum of coefficients priors
• Resampling from VARs

Block C3: Mixed-frequency models and conditional forecasting

• Designing VARs with mixed-frequency data
• Forecasting with ragged edge data
• Conditioning in VAR simulations

Block C4: Time-domain and frequency-domain properties

• Autocovariance and autocorrelation functions
• Power spectrum and spectral density functions
• Bootstrapping and resampling VARs

---

D. [Higher priority than C] Structural modeling

Block D1: Using Iris preparer to build more complex models

• Creating multi-area and/or multi-sector models
• Mutliple model source files
• Preparser commands !if, !for, <...>
• Dynamic and steady versions of equations !!

Block D2: Nonlinear models: steady state

• Nonlinear models with growth, log status of individual variables !log-variables
• Block sequential analyzer of steady state, blazer
• Advanced options in steady state calculations: fix, fixLevel, fixGrowth, exogenize, endogenize
• Steady state databank

Block D3: Nonlinear models: dynamic simulations

• First order solution and its use in nonlinear solution methods
• Solution method options firstOrder, stacked, period
• Terminal condition and its influence
• Occasionally binding inequalities
• Time frames in nonlinear simulations, anticipated vs unanticipated events, frame databanks

Block D4: Bayesian estimation with system priors

• Estimation with priors on individual parameters vs system properties
• Example of system properties: sacrifice ratio, delayed policy reaction, trend-gap frequency response functions
• Implementation of SystemProperty, SystemPrior, SystemPriorWrapper objects
• Posterior mode and posterior simulator with system priors

---

E. Reporting

Block E1: On-screen plotting

• Chartpack object

Block E2: HTML reporting

• +rephrase package
• Overview of the endless possibilities
• Interactive features in charts, tables, pagers
• User CSS styling