STAMP dialogs

Contents

Models for time-series data
Dialogs for Unobserved Components Modelling

Dialogs for Unobserved Components Modelling

Dialogs for model formulation and estimation:
Formulate
Select components
Estimate
Options
Progress
Regression coefficients
Select interventions
Edit and fix parameter values
Dialogs for model evaluation:
Test Menu
Further Output
Component graphics
Weight functions
Residual graphics
Auxiliary Residual graphics
Prediction graphics
Forecasting
Store in Database

Formulate - STAMP unobserved components module

Use this dialog for single equation dynamic model formulation: to formulate a new model, or reformulate an existing model.

Database
Mark all the variables you wish to include in the new model or add to the existing model, using the Tab or Arrow keys and selecting by spacebar or by left clicking the mouse. After you have pressed << (or double-clicked if you are using a mouse), the database variables are added to the model with the default lag length.

The variable at the top of the list will by default become the endogenous (Y) variable. To select a different dependent variable, see below.

Lags
At the top you can choose how the lag length is set with which variables are added to the model:
Selection
This list box shows the current model.
The variable at the top of the list will by default become the endogenous (Y) variable.
To select a dependent variable which is listed further down:
  1. mark the current dependent variable and right-click to clear its status;
  2. mark the new variable, and right-click to change to Y: endogenous.
You can also assign a new status to a variable by first selecting the variable and then selecting a status from the bar at the bottom of the Selection box. Confirm by clicking the Set button. If you have marked variables in the model, you can delete them, or assign a status to them.
<<
Adds the currently selected database or special variables to the model.
>>
Deletes the currently selected variables from the model.
Clear>>
Deletes the whole model, so that you can start from scratch.

Previous models
Use this to recall a previously estimated model.
OK
Press OK to move to the Select components or Estimation.

Select components - STAMP unobserved components module

This dialog is for choosing the components of the model.

Basic components
Select whether or not you wish to include a Level, Slope, Seasonal (this option is not available for annual data) and Irregular component in the model. For Level, Slope and Seasonal you can choose the component to be Fixed or Stochastic You can select and deselect the options by left clicking the box on the right of the option or you can use the Tab or Arrow keys of your keyboard and select the option by pressing spacebar.
Cycle(s)
You have the option to include up to three cycles in the model. The three cylces have different default frequencies. You can also choose to include Autoregressive components by selecting the AR(1) or AR(2) option.
Options
There are three additional options, which can be selected by marking the box on the right of the option. After you have pressed OK in the Select components dialog, for every option you selected a dialog will be shown.
  • Set regression coefficients
  • Select interventions
  • Edit and fix parameter values
  • Regression coefficients - STAMP unobserved components module

    For every explanatory variable you can choose the type of regression coefficient by selecting the variable (using Tab or Arrow keys and pressing spacebar, or by clicking) and then selecting the type of coefficient from the list that will be shown on the right of the variable, by using Tab or Arrow keys and pressing Enter or by clicking. Press OK when finished. This will take you to the next option or, if you did not select any other option, to Estimation

    Select interventions - STAMP unobserved components module

    When this dialog opens, it will show a list of interventions. By default the list will consist of one irregular intervention. You can change the type of intervention by clicking irregular and selecting a type from the list that will be shown. You also have the option to include interventions in Level and Slope.
    One cell to the right you can select the period in which the intervention takes place. If you wish to include the intervention in the model, mark the box under Select.

    Add
    Adds a new intervention to the current list. This intervention will be irregular and unselected by deafult. You can modify new interventions in the same way as described above.
    <
    Moves one cell to the left in the list.
    >
    Moves one cell to the left in the list.
    Del
    Deletes the selected intervention from the list.
    Save As...
    Allows you to save the current list of interventions. You can choose to save the list in Matrix, Excel, OxMetrics or Comma seperated form.
    Load...
    Allows you to load a list of interventions saved earlier.
    Clear
    Deletes the whole list, so that you can start from scratch.
    OK
    Press OK to move to the next option or, if you did not select any other option, to Estimation

    Edit and fix parameter values - STAMP unobserved components module

    If you wish to fix any of the parameters of the components used in the model, simply mark the box under Fix of the component you wish to fix the parameter for. You can then set the fixed value by selecting the cell under Value which currently contains a default value and then by changing this into the desired value.

    <
    Moves one cell to the left in the list.
    >
    Moves one cell to the left in the list.
    Save As...
    Allows you to save the current values of the parameters. You can choose to save the list in Matrix, Excel, OxMetrics or Comma seperated form.
    Load...
    Allows you to load a list parameter values saved earlier.
    Clear
    Deletes the whole list, so that you can start from scratch.
    OK
    Press OK to move to Estimation

    Estimate - STAMP unobserved components module

    Select the sample period and an estimation method.

    Choose the estimation sample
    Enter the sample period you wish to use for the estimation. The maximum sample is given one line up.
    The default is the whole sample.
    Choose the estimation method
    Select the method of estimation you want to use.
    OK
    Pressing OK starts the estimation, unless there still is something missing or wrong in the dialog.

    Progress

    The Progress dialog is used to review the progress made to date in the model reduction, when using the general-to-specific modelling strategy.

    To offer a default sequence, STAMP decides that model A could be nested in model B if the following conditions hold:

    STAMP does not check if the same variables are involved, because transformations could hide this. As a consequence STAMP does not always get the correct nesting sequence, and it is the user's responsability to ensure nesting.

    E.g. DCONS = α + βDINC is nested in: CONS = a + b1 CONS1 + b2 INC + b3 INC1 through the restrictions b1 = 1 and b3 = -b2.

    There are two options on the dialog to select a nesting sequence:

    Mark Specific to General
    Marks more general models, finding a nesting sequence with strictly increasing log-likelihood.
    Mark General to Specific
    Marks all specific models that have a lower log-likelihood.
    The default selection is found by first setting the most recent model as specific, and then setting the general model that was found as the general model.

    Additional dialog items are:

    <
    To move a model up in the modelling sequence.
    >
    To move a model down in the modelling sequence.
    Del
    Tp permanently delete a model from the modelling sequence.
    OK
    Prints the progress report, consisting of:
    1. number of observations, paramaters, and log-likelihood.
    2. Information criteria: reported are the Schwarz Criterion (SC), the Hannan-Quinn (HQ) Criterion, and the Akaike criterion (AIC).
    3. F or Chi-squared tests of each reduction.

    Options (all)

    Controls maximization settings, and what is automatically printed after estimation (in addition to the normal estimation report). Model options referes to settings which are changed infrequently, and are persistent between runs of STAMP.

    Maximization Settings

    Maximum number of iterations:
    Note that it is possible that the maximum number of iterations is reached before convergence. The maximum number of iterations also equals the maximum number of switches in cointegration.

    Write results every:
    By default no iteration progress is displayed in the results window. It is possible to write intermediate information to the Results window for a more permanent record. A zero (the default) will write nothing, a 1 every iteration, a 2 every other iteration, etc.

    Write in compact form:
    Writes one line per printed iteration (see Write results every).

    Convergence tolerance:
    Change the convergence tolerance levels (the smaller, the longer the estimation will take to converge). See under numerical optimization for an explanation of convergence decisions.

    Default:
    Resets the default maximization settings.

    Test Menu

    This dialog box gives access to a selection of diagnostic testing procedures. Mark the tests you want to be executed, then press OK.

  • More written output
  • Components graphics
  • Weight functions
  • Residuals graphics
  • Auxiliary Residual graphics
  • Prediction graphics
  • Forecasting
  • Store in Database
  • More written output

    Components graphics

    Plots Y with
    Gives you the option to plot the dependent variable Y togehter with the Trend. The options to plot Y togehter with the Trend and other components are only available if those components are included in the model.
    Select components
    Mark the boxes of the components of which you wish to see a plot. Only components that are included in the model can be selected.
    Select type of plot
    You can choose to make individual plots of the components or to make a plot of the composite signal. The latter can also be plotted together with Y or cross-plotted against Y.
    Further plots
    Gives you the option to plot Detrended Y, Seasonally adjested Y and individual seasonals plots.
    Prediction, filtering and smoothing
    You can choose whether you want to use Predictive filtering, Filtering or Smoothing to estimate the components. Smoothing is set as default. If you wish to use more than one method, you can choose to plot the different estimates together or separately.
    Further options
    Components estimates can be plotted together with confidence intervals.
    Anti-log analysis can be used (this is useful when the data are in logs).
    You can choose which part of the sample will be used for the components estimates and which part of the estimates will be plotted.
    Default is the whole sample for both options.
    The selected components can be stored in the OxMetrics database. OxMetrics will prompt for a variable name.

    Weight functions

    Residuals graphics

    One step ahead prediction residuals (standardized)
    Plots the standardized residuals of the one step ahead predictions.
    Diagnostic plots
    Mark the boxes of the desired plots. You can choose a Correlogram and specify the number of Lags, a Spectrum and specify the range, a Histogram and a QQ-plot.
    Cumulative plots
    Gives you the option to plot the cumulative sum of the residuals, the cumulative sum of squares of the residuals and the cumulative sum scaled so that it can be compared with the critical values of the t-test.
    Write
    Mark the Diagnostics box if you wish to view the Goodness-of-fit based on the residuals and Serial correlation statistics for the residuals. These will be shown in OxMetrics/Text/Results
    Store
    Mark the Residuals box if you wish to store the residuals in the OxMetrics database.

    Auxiliary residuals graphics

    Prediction graphics

    Select type of predictions
    Choose between One-step ahead and Multi-step ahead predictions.
    Choose the post-sample size.
    Plot predictions and Y
    Predictions can be plotted together with Y or with standard errors. They can also be scaled by a factor of choice or be cross-plotted against Y.
    Plot residuals
    The residuals of the predictions made can be plotted. These can also be plotted together with standard errors or scaled by some factor. Standardized residuals can also be plotted, as can the cumulative sum of the standardized residuals.
    Write
    Mark the Prediction tests box if you wish to view the Prediction analysis the post-sample predictions and the Post-sample predictive tests, which are the Chi^2 test and the Cusum t-test.
    These will be shown in OxMetrics/Text/Results

    Forecasting

    Store in database

    Allows you to save any of the listed items in the OxMetrics database. Note that forecasts must be generated using Test/Forecast before they can be stored.

    OxMetrics will prompt for a variable name.

    Lags

    When creating lags, STAMP appends the lag length as extra characters in a name, preceded by an underscore. E.g. CONS_1 is CONS one period lagged.

    Lagging a variable leads to the loss of observations, but seasonals can be lagged up to the frequency without loss. STAMP handles variables in models through lag polynomials.

    Sample periods are automatically adjusted when lags are created.

    STAMP stores the lag information, and uses it to recognize lagged variables for Dynamic Analysis. Lags created this way are not physically created, and do not consume any memory. However, when you compute a lag using the calculator, a new variable will be created in the database, which will NOT be treated as a lagged version of that variable, but as any other variable.

    Dynamic model formulation

    A dynamic equation is specified as an autoregressive-distributed lag model:

    B0(L) yt = c + B1(L) x1,t + B2(L) x2,t + ... + Bk(L) xk,t + et,   t = 1,...,T.      (1)

    In (1), the lag polynomials are defined by:

    Bi (L) = Σnij=mi bi,j Lj    with 0 ≤ mini,    i = 1,...,k.

    `Solving' (1) yields:

    yt = Σki=1 Hi (L) xit,   where   Hi (L)=Bi (L) / B0(L).

    Zero is a legitimate order for a lag polynomial. Thus, static or dynamic models are equally easily specified.

    A model in STAMP is formulated by:

    1. Which variables are involved;
    2. The orders of the lag polynomials;
    3. The status of variables (only when it is not legitimate to treat all regressors as valid conditioning variables, and you wish to use Instrumental Variables).

    The following information is needed to estimate an equation:

    1. The model formulation;
    2. The sample period;
    3. Optionally, the number of static forecasts to be withheld for testing parameter constancy;
    4. The method of estimation;
    5. Optionally, the number of observations to be used to initialize the recursive estimation (when available).

    The available single-equation estimators are (see Volume I):

    Single-equation estimation output is discussed in Volume I. Models may be revised interactively after formulation and after estimation. Afterwards, the estimated model can be analyzed.

    STAMP facilitates a general-to-specific modelling strategy.

    Ordinary Least Squares (OLS)

    Ordinary Least Squares is the standard textbook method. OLS is valid if the data model is congruent.

    Congruency

    The requirements for congruency are:

    1. Homoscedastic innovation errors;
    2. Weakly exogenous regressors;
    3. Constant parameters;
    4. Theory consistency;
    5. Data admissibility;
    6. Encompassing rival models.

    STAMP provides tests of most of the aspects of model congruency.

    Instrumental Variables (IV)

    A structural representation is parsimonious with parameters but has regressors which are correlated with the error term. IV requires that the reduced form is a congruent data model. The Instrumental variables are the reduced form regressors. Instrumental Variables include two stage least squares (2SLS) as a special case.

    STAMP needs to know the status of the variables in the model:
    1. At least one endogenous variable on the right-hand side;
    2. At least as many instruments as endogenous rhs variables.

    STAMP computes:
    1. The estimate of all the reduced form equations;
    2. The estimate of the structural form equation;
    3. Tests of the over-identifying restrictions.

    Autoregressive least squares (RALS)

    Autoregressive least squares requires that the restricted dynamic model is data congruent, where the restrictions correspond to COMFAC constraints selected (since an autoregressive error is a more parsimonious representation). Various orders of autoregression can be selected, and a grid is estimable for single orders.

    Multiple optima to the likelihood function commonly occur in the COMFAC class, thus case 5. is recommended. Direct fitting of 4. may not find the optimum. .

    RALS numerical optimization

    The log-likelihood function f(θ) for RALS is a sum of squares of non-linear terms.

    Let the regression and the autoregressive error parameters be β and ρ. Then f(β, ρ) is non-linear but is linear in β given ρ and conversely.

    The Gauss-Newton method exploits this fact. It is a reliable choice, but need not find global optima. Like Newton-Raphson, Gauss-Newton uses analytical first and second derivatives. Hendry (1976) reviews alternative methods.

    The autoregressive error can be written as

    ut = Σri=s ρi ut-i + εt   with   εt ~ IN(0, σ2).

    Numerical optimization

    Numerical optimization is used to maximize the likelihood log L(θ) as an unconstrained non-linear function of θ.

    STAMP maximization algorithms are based on a Newton scheme:

    θk+1 = θk + skQk-1 qk,

    with

    STAMP and STAMP use the quasi-Newton method developed by Broyden, Fletcher, Goldfarb, Shanno (BFGS) to update K = Q-1 directly, estimating the first derivatives numerically.

    Owing to numerical problems, it is possible (especially close to the maximum) that the calculated θ does not yield a higher likelihood. Then an s in [0,1] yielding a higher function value is determined by a line search. Theoretically, since the direction is upward, such an s should exist; however, numerically it might be impossible to find one.

    The convergence decision is based on two tests:
    1. based on likelihood elasticities (dlogLik/dlog|θ|) (scale invariant):
    | qk,j θk,j | ≤ eps   for all j when θk,j not zero,
    | qk,j | ≤ eps for all j when θk,j = 0.

    2. based on the one-step-ahead relative change in the parameter values (assuming step length 1) (scale variant, but relative change is infinite if any θ = 0)
    | θk+1,j - θk,j | ≤ 10 * eps * | θk,j |   for all j when θk,j not zero,
    | θk+1,j - θk,j | ≤ 10 * eps for all j when θk,j = 0.

    The status of the iterative process is given by the following messages:

    1. No convergence!
    2. Aborted: no convergence!
    3. Function evaluation failed: no convergence!
    4. Maximum number of iterations reached: no convergence!
    5. Failed to improve in line search: no convergence!
      s has become too small.
      Test 1 was passed, using eps2.
    6. Failed to improve in line search: weak convergence.
      s has become too small.
      Test 1 was passed, using eps2.
    7. Strong convergence
      Both tests were passed, using eps1.

    The chosen default values are: eps1 = 1E-4, eps2 = 5E-3.

    You can:

    1. Set the initial values of the parameters to zero or the previous values;
    2. Set the maximum number of iterations;
    3. Write iteration output;
    4. Change the convergence tolerances eps1 and eps2;
    5. Care must be exercised with this: the defaults are `fine-tuned': some selections merely show the vital role of sensible choices!
    6. Choose the maximization algorithm;
    7. Plot a grid of the log-likelihood.
      The `fineness', number of points and centre can be user-selected. Up to 16 grids can be plotted simultaneously. A grid may reveal potential multiple optima.

    Options 1., 5. and 6 are mainly for teaching optimization.

    NOTE: estimation can only continue after convergence.

    Modelling strategy

    STAMP has two modes of operation: general-to-specific and unordered.

    General-to-specific

    1. Begin with the dynamic model formulation;
    2. Check its data coherence and cointegration;
    3. Transform to a set of variables with low intercorrelations but interpretable parameters;
    4. Delete unwanted regressors to obtain a parsimonious model;
    5. Check the validity of the model by thorough testing.

    STAMP monitors the progress of the sequential reduction from the general to the specific and will provide the associated F-tests, Schwarz and σ values.

    Unordered Search

    Nothing commends unordered searches:
    1. No control is offered over the significance level of testing;
    2. A `later' reject outcome invalidates all earlier ones;
    3. Until a model adequately characterizes the data, standard tests are invalid

    Dynamic analysis

    After estimation, unrestricted general models like (1) in the Dynamic Model Formulation are analysed:

    B0(L) yt = c + B1(L) x1,t + B2(L) x2,t + ... + Bk(L) xk,t + et,   t = 1,...,T.     (1)

    where

    Bi (L) = bi,0 + bi,1 L + bi,2 L2 + ... + bi,n Ln.

    Static long-run solution
    If the roots of B(L) lie outside the unit circle we can rewrite (1) as (forgetting about c and e):
    yt = Σki=1 Hi (L) xi,t,   where   Hi (L)=Bi (L) / B0(L).     (2)

    If E[x] has remained at a constant level x for long enough, y will reach its long-run solution:

    E[y] = Σki=1 Hi (1) E[xi],   where   Hi (1)=Bi (1) / B0(1).     (3)

    (reported with asymptotic standard errors).

    Static Forecasting

    STAMP allows you to retain observations to compute forecast statistics. For OLS/RLS/RALS these are comprehensive 1-step ahead forecasts.

    For IV/RIV, since there are endogenous regressor variables, the only interesting issue is that of parameter constancy, and the only output is the forecast Chi˛ test.

    Dynamic forecasts can be made from single equation models as well as from simultaneous equations system. STAMP will compute analytical standard errors of dynamic forecasts, and can take parameter uncertainty into account.

    Correlations

    The correlation matrix of selected variables is a symmetric matrix, with the diagonal equal to one. Each cell records the simple correlations between the two relevant variables.

    The mean:

    m = T-1 ΣTi=1 xi,

    and standard deviation:

    s = (T-1)-1 ΣTi=1 (xi - m)2

    of the variables are also given.

    NOTE that the standard deviation here is based on 1/(T-1).

    Data density and histogram

    Histograms are a way of looking at the sample distributions of statistics. Then, on the basis of the original data, density functions may be interpolated to give a clearer picture of the implied distributional shape: similarly, cumulative distribution functions may be constructed (and compared on-screen to a Cumulative Normal Density).

    Non-parametric density estimation

    Given observations:

    (x1 ... xT)

    from some unknown probability density function f(X), about which little may be known a priori. To estimate that density without imposing too many assumptions about its properties, a non-parametric approach is used in STAMP based on a kernel estimator. The kernel K used is the Normal or Gaussian kernel. Research suggests that the density estimate is little affected by the choice of kernel, but is largely governed by the choice of window width, h.

    Owing to the importance of the window width h in estimating the density, the non-parametric density estimation menu offers control over the choice of window width, h = CσTP. By default, P = -0.2 and C = 1.06 in STAMP. For normal densities this choice will minimize the Integrated Mean Square Error.

    For more information see: Silverman B.W. (1986). Density Estimation for Statistics and Data Analysis, London: Chapman and Hall.

    Correlogram (ACF, PACF)

    The correlogram or autocorrelation function (ACF) of a variable, or of the residuals of an estimated model, plots the series of correlation coefficients { rj } between xt and xt-j.

    The length s of the ACF is chosen by the user, leading to a figure which shows (r1, r2, ..., rs) plotted against (1,2,..., s).

    A related statistic is the Portmanteau (also called Box-Pierce or Q-statistic):

    T Σsj=1 rj2.

    The partial autocorrelation coefficients correct the autocorrelation for the effects of previous lags. So the first partial autocorrelation coefficient equals the first normal autocorrelation coefficient.

    Spectrum

    A stationary series can be decomposed in cyclical components with different frequencies and amplitudes. The spectral density gives a graphical representation of this. It is symmetric around 0, and only graphed for [0,π] (the horizontal axis in the STAMP graphs is scaled by π, and given as [0,1]).

    The spectral density consists of a weighted sum of the autocorrelations, using the Parzen window as the weighting function. The truncation parameter m can be set (the larger m, the less smooth the spectrum becomes, but the lower the bias).

    A white-noise series has a flat spectrum.

    Diagnostic testing

    Test types

    Many tests report a Chi^2 and an F form. In the summary, only the F-test is reported, which is expected to have better small-sample properties.

    F-tests are usually reported as

    F(num,denom) = Value [Probability] /*/**

    For example

    F(1, 155) = 5.0088 [0.0266] *

    where the test statistic has an F distribution with 1 degree of freedom in the numerator, and 155 in the denominator. The observed value is 5.0088, and the probability of getting a value of 5.0088 or larger under this distribution is .0266. This is less than 5% but more than 1%, hence the star.

    Significant outcomes at a 1% level are shown by two stars: **.

    Chi^2 tests are also reported with probabilities, as e.g.:

    Normality Chi^2(2)= 2.1867 [0.3351]

    The 5% Chi^2 critical values with 2 degrees of freedom is 5.99, so here normality is not rejected (alternatively, Prob(Chi^2 ³ 2.1867) = 0.3351, which is more than 5%).

    Auxiliary regression tests

    Many diagnostic tests are done through an auxiliary regression.
    In this case two forms of the test are reported:
    1. TR^2 which has a Chi^2(r) distribution for r restrictions;
    2. (T-k-r)R^2/r(1-R^2), which has an F(r,T-k-r) distribution.
    The F-form may be better behaved in small samples.

    Autoregressive Conditional Heteroscedasticity (ARCH)
    Checks whether the residuals have an ARCH structure:
    E[ ut2 | ut-1 , ..., ut-r ] = Σri=s αi ut-i2,

    with [0 &leq; s &leq; r &leq; 12] and e ~ IID(0, τ2). An F-statistic and the αs are reported. The null hypothesis is no ARCH, which would be rejected if the test statistic is too high. This test is done by regressing the squared residuals on a constant and lagged squared residuals (now some observations are lost at the beginning of the sample).

    Normality
    The Normality test checks whether the variable at hand (either a database variable or the residuals), here called u, are normally distributed as:
    ut ~ IN(0,1)   with E[ut3] = 0,   and E[ut4] = 3σ2.

    A Chi^2 test is reported (with 2 degrees of freedom), and the output includes all moments up to the fourth. The null hypothesis is normality, which will be rejected at the 5% level, if a test statistic of more than 5.99 is observed.

    Full report includes:
    mean:

    m = T-1 ΣTi=1 xi;

    moments:

    mj = T-1 ΣTi=1 (xi - m)j;

    (reported as m21/2);
    skewness:

    m3 / m23/2;

    excess kurtosis:

    m4 / m22  -  3.

    The reported test statistic has a small-sample correction. Also reported is the asymptotic form of the test (skewness2 *T/6 + excess_kurtosis2 *T/24), which requires large samples for the asymptotic Chi2(2) distribution to hold.

    NOTE that the standard deviation here is based on 1/T.

    Tests for linear restrictions

    If we write the model as

    y = + u, where y is (T x 1), β is (k x 1) and X is (T x k),

    then linear restrictions can be expressed in vector form as:

    = r, where R is a (p x k) matrix, and r a (p x 1) vector.

    E.g. the two restrictions: α = 1 and β = -γ in

    CONS = b + α CONS1 + β INC + γ INC1

    can be expressed as:

    | 0 1 0 0 | R = | |, r' = [0 1]. | 0 0 1 1 |

    STAMP allows you to test general linear restrictions by specifying R and r, in the form of a (p x k+1) matrix [R : r]. Simple linear restrictions of the form α =... = δ = 0 can be done by selecting the relevant variables.

    The null-hypothesis Ho: = r is rejected if we observe a significant test statistic.

    Two tests of linear restrictions are routinely reported in STAMP:
    1. Ho: b = 0, where the test-statistic is the t-ratio of b.
    2. Ho: α = ... = δ = 0 (all coefficients apart from the constant are zero).
    Shown as the F-statistic which follows R^2 (and can be derived from it).

    Tests for general restrictions

    Given the estimated coefficients θ, and their variance-covariance matrix V[θ], we can test for (non-) linear restrictions of the form:

    f(θ) = 0;

    The null hypothesis Ho: f(θ) = 0 will be tested against Ha: f(θ) ≠ 0 through a Wald test:

    w = f(θ) ' (JV[θ] J')-1 f(θ)

    where J is the Jacobian of the transformation:

    J = ∂ f(θ)/∂q'.

    The statistic w evaluated at θ has a Chi^2(r) distribution, where r is the number of restrictions (i.e. equations in f(θ)). The null hypothesis is rejected if we observe a significant test statistic.

    E.g. the two restrictions implied by the long-run solution of:

    CONS = b + α CONS1 + β INC + γ INC1 + δ INFLAT

    are expressed as

    (β + γ) / (1 - α) = 0;
    δ / (1 - α) = 0;

    which has to be fed into STAMP as (coefficient numbering starts at 0!):

    (&1 + &2) / (1 - &0) = 0; &3 / (1 - &0) = 0;

    Common-Factor Test

    The COMFAC test evaluates error-autocorrelation claims by checking if the model's lag polynomials have factors in common. If so, the model's lags can be simplified with an autoregressive error; if not, the model cannot be re-expressed with an autoregressive error. Chi^2 tests of each possible common factor and of sequences are shown.

    The COMFAC test option is only feasible for unrestricted dynamic models (which have a closed lag system), which are not estimated by Autoregressive Least Squares.

    The algorithm was developed and written by Denis Sargan and Juri Sylvestrowicz.

    We have recently discovered that the COMFAC test outcome may change if ordering of the variables in the model is changed (but only if there are at least several lag polynomials of the same length). This is due to testing different formulations of the restrictions in the Wald test (i.e. computing determinants of different submatrices).

    Omitted variables

    This tests if some variables should be added to the model, which can be any variables in the database matching the present sample.

    If the estimated model is

    y = + u,

    then the omitted variables test, tests for γ = 0 in

    y = + +v,

    The Lagrange Multiplier F-test is reported, and the null hypothesis is rejected when its value is significant.

    This test is not available for Autoregressive Least Squares or non-linear models.

    Encompassing tests

    Encompassing evaluates against rival models to see if they embody specific information excluded from the model under test.

    Encompassing tests are only available for single equation models estimated by OLS or IV.

    Four tests are calculated:

    1. The Cox non-nested hypotheses test (Cox, 1961)
    This tests whether the adjusted likelihoods of two rival models are compatible. It is equivalent to checking variance encompassing. This test is invalid for IV estimation, and omitted in that case.
    2. The Ericsson Instrumental Variables test (Ericsson, 1983)
    This is an IV equivalent to the Cox test.
    3. The Sargan restricted/unrestricted reduced form test (Sargan, 1964)
    This checks if the restricted reduced form of a structural model encompasses the unrestricted reduced form including exogenous regressors from rival models.
    4. The joint model F-test
    checks if each model parsimoniously encompasses the linear nesting model.

    Invariance

    The F-test is invariant to variables in common between the rival models. The Cox and the Ericsson tests are not invariant: their values change with the choice of overlapping variables.

    Consult e.g. Ericsson (1983) or Hendry and Richard (1987) for details.

    Status of variables

    STAMP checks for valid choices of variables:
    1. Endogenous variables are matched;
    2. Instruments in Model 1 are treated as exogenous in Model 2 even if you denote them as endogenous;
    3. The models must be non-nested.

    Output

    The output is summarized in an encompassing table:
    1. The type of test statistic;
    2. The value of each outcome;
    3. The degrees of freedom of each test;
    4. The null that Model 1 is valid is on the left;
    5. The null that Model 2 is valid is on the right.

    If the left-side tests are insignificant, Model 1 encompasses Model 2.
    If the left-side tests are significant, Model 1 fails to encompass Model 2.
    Similarly for the rightside tests with models 1 and 2 interchanged.

    Model 1 encompasses Model 2 implies Model 1 also parsimoniously encompasses the linear nesting model. If not, Model 2 contains specific data information not captured by Model 1.

    The algorithm incorporated in STAMP was written by Neil Ericsson.

    Identities

    Identities are exact (linear) relations between variables, as in the components of GNP adding up to the total by definition. In STAMP, identities are created by marking identity endogenous variables as such during dynamic system formulation.

    Identities are ignored during system estimation/analysis. They come in at the model formulation level, where the identity is specified just like other equations. However, there is no need to specify the coefficients of the identity equation, as STAMP automatically derives these by estimating the equation (which must have an R^2 of at least 0.99).

    Unrestricted variables

    Variables can be classified as unrestricted during dynamic system formulation. Such variables will be partialled out, prior to estimation, and their coefficients will be reconstructed afterwards. Although unrestricted variables do not affect the basic estimation, there are important differences:

    Following estimation:
    the R^2 measures and corresponding F-test are relative to the unrestricted variables.
    In recursive estimation:
    the coefficient of unrestricted variables are fixed at the full sample values.
    In cointegration analysis:
    unrestricted variables are partialled out together with the short-run dynamics, whereas restricted variables (other then lags of the endogenous variables) are restricted to lie in the cointegrating space.
    In simultaneous equations estimation:
    unrestricted variables are partialled out prior to estimation. FIML estimation of the smaller model could improve convergence properties of the non-linear estimation process.

    Dynamic System and Model Formulation

    The simultaneous equations modelling process in STAMP starts by focusing on the System, often called the unrestricted reduced form (URF), which can be written as:

    (1) yt = π0 + πi yt-i + πj zt-j + vt,  vt ~ IN(0,Ω)   i = 1,...,m,  j = m+1,...,m+r.

    where yt, zt are respectively (n x 1) and (q x 1) vectors of observations at time t, t = 1,...,T, on the endogenous and non-modelled variables. A more compact way of writing the system is:

    (2) yt = Πwt + vt

    where w contains z, lags of z and lags of y, and Π is (n x k).

    A vector autoregression (VAR) arises when there are no z's (but there could be a constant, seasonals or trend). An example of a 2-equation system is:

    CONS = β0 + β1 CONS1 + β2 INC1 + β3 CONS2 + β4 INC2 + β5 INFL,
    INC = β6 + β7 CONS1 + β8 INC1 + β9 CONS2 + β10 INC2 + β11 INFL.

    This system would be a VAR when β5 = β11 = 0.

    Non-modelled variables can be classified as unrestricted. Variables defined by identities are also allowed.

    To obtain a structural dynamic model, premultiply the system (2) by a non-singular matrix B, which yields:

    (3) Byt = BΠwt + Bvt.

    We shall write this as:

    (4) Byt + Cwt = ut,   t = 1,...,T;  ut ~ IN(0,σ),

    or succinctly:

    Axt = ut

    The restricted reduced form (RRF) corresponding to this model is (note that the Π of (5) is a restricted version of that in (3)):

    (5) yt = Πwt + vt,   with  Π = -inv(B)C.

    Identification of the model, through within equation restrictions on A, is required for estimation. Some equations of the model could be identities. An example of a model with the previous system as unrestricted reduced form is:

    CONS = β0 + β1 CONS1 + β2 INC + β3 INFL,
    INC = β4 + β5 INC1.

    The philosophy behind STAMP is first to develop a congruent system. If the system displays symptoms of mis-specification, there is little point in imposing further restrictions on it. From a congruent system a model is derived.

    A system in STAMP is formulated by:

    1. which variables yt, zt are involved;
    2. the orders of the lag polynomials;
    3. classification of the ys in endogenous variables and identity endogenous variables;
    4. any non-modelled variable may be classified as unrestricted. Such variables will be partialled out, prior to estimation, and their coefficients will be reconstructed afterwards.

    A model in STAMP is formulated by:

    1. which variables enter each equation, including identities;
    2. coefficients of identity equations need not be specified, as STAMP automatically derives these by estimating the equation (requires an R^2 of at least 0.99);
    3. constraints, if the model is going to be estimated by CFIML or RCFIML.

    When a model has been formulated, it can be estimated and evaluated, a detailed description of estimators and tests is in Volume II. STAMP facilitates a general-to-simple modelling strategy.

    Cointegration Analysis

    The vector autoregression can be written in equilibrium-correction form as:

    Δyt=( π1+π2-In) yt-1-π2Δyt-1+Φqt+vt,

    or, writing P0=π1+π2-I, and δ1=-π:

    Δyt=P0yt-1+δ1Δyt-1+Φqt+vt.
    (eq:1.1)

    Equation (eq:1.1) shows that the matrix P0 determines how the level of the process y enters the system: for example, when P0=0, the dynamic evolution does not depend on the levels of any of the variables. This indicates the importance of the rank of P0 in the analysis. P0=∑πi-In is the matrix of long-run responses. The statistical hypothesis of cointegration is:

    H(p):    rank( P0) ≤p.

    Under this hypothesis, P0 can be written as the product of two matrices:

    P0=αβ',

    where α and β have dimension n×p, and vary freely. As suggested by Søren Johansen, such a restriction can be analyzed by maximum likelihood methods.

    So, although vt~INn[0,Ω], and hence is stationary, the n variables in yt need not all be stationary. The rank p of P0 determines how many linear combinations of variables are I(0). If p=n, all variables in yt are I(0), whereas p=0 implies that Δyt is I(0). For 0<p<n there are p cointegrating relations β'yt which are I(0). At this stage, we are not discussing I(2)-ness, other than assuming it is not present.

    The approach in STAMP to determining cointegration rank, and the associated cointegrating vectors, is based on the Johansen procedure.

    Estimator Generating Equation

    All model estimation methods in STAMP are derived from the Estimator Generating Equation (EGE).

    We require the reduced form to be a congruent data model, for which the structural specification is a more parsimonious representation. The structural model is:

    BY' + CW' = U',

    or using A = (B : C):

    AX' = U',

    with the restricted reduced form (RRF)

    Y'= ΠW' + V'

    (so Π = -inv(B)C). Writing Q' = (Π' : I), we have that AQ = 0, and can write the restricted reduced form as:

    X'= QW' + V'.

    The structural model involves regressors which are correlated with the error term. Instruments (reduced form regressors) are used in place of structural form regressors to estimate the unknown coefficients in A, denoted θ. The general estimation formulation is based on the EGE.

    The available estimation methods are described in Volume II. 1SLS applies OLS to each equation, imposing a diagonal errorr varance matrix. This estimator is not consistent for a simultaneous system, but is offered for systems that are large relative to the data available, where its MSE properties may be the best that can be achieved.

    Dynamic Forecasting

    STAMP allows you to retain observations to compute forecasts and forecast statistics. Both 1-step ahead (static, ex-post) and h-step ahead (dynamic, ex-ante) forecasts are available. The 1-step forecasts are computed automatically after system and model estimation if observations are reserved. Three 1-step test statistics are offered:

    1. Using Ω: This test ignores parameter uncertainty and intercorrelations between forecast errors, thus taking only innovation uncertainty into account.
    2. using V[e]: This test takes parameter uncertainty into account, but ignores intercorrelations between forecast errors.
    3. using V[E] (only for the system): This statistic takes both parameter uncertainty and intercorrelations between forecast errors into account, making it a better calibrated test statistic.

    Dynamic forecasts are available separately, up to the end of the database sample period (observations are required for all exogenous variables, but not for endogenous variables and their lags). Dynamic forecasts can be with or without 95% error bars, but only the innovation uncertainty is allowed for in the computed error variances. Two types of forecasts are available for graphing:

    1. Dynamic forecasts
      Select this to graph the dynamic forecasts (the sequence of 1, 2, 3,...,h-step forecasts).
    2. h-step forecasts
      Up to h forecasts, the graphs will be identical to the dynamic forecasts. Thereafter values of the endogenous variables which go more than h periods back will use actual values.

    The database sample can be extended with ease if longer-horizon forecasts are desired.

    Matrix File

    Seceral formats are available to load and save matrices:

    An example of a matrix file is: +---------+ ¦ 2 3 ¦ &lt;-- dimensions, a 2 by 3 matrix ¦//comment¦ &lt;-- a line of comment ¦ 1 0 0 ¦ &lt;-- first row of the matrix ¦ 0 1 .5 ¦ &lt;-- second row of the matrix +---------+

    Closed lag system

    With a closed lag system is meant that there are no gaps in the lag polynomials. So a closed system is e.g.:

    CONS = b + α CONS1 + β INC + γ INC1

    however, without INC (i.e. β = 0), it wouldn't be closed. You could then replace INC lagged by INC1 = lag(INC, 1), and close the lag system (because STAMP will not know that INC1 is a lagged variable; STAMP only recognizes lags when they are created within the model formulation dialog).

    Missing values

    The data sample for analysis is automatically selected to not include any missing values within the sample. In cross-section regression, any observation with missing values is automatically omitted from the analysis, so in-sample observations with missing values are simply skipped.


    This file last changed .