1function oo_=disp_moments(y,var_list,M_,options_,oo_)
2% function disp_moments(y,var_list,M_,options_,oo_)
3% Displays moments of simulated variables
4% INPUTS
5%   y                   [double]       nvar*nperiods vector of simulated variables.
6%   var_list            [char]         nvar character array with names of variables.
7%   M_                  [structure]    Dynare's model structure
8%   oo_                 [structure]    Dynare's results structure
9%   options_            [structure]    Dynare's options structure
10%
11% OUTPUTS
12%   oo_                 [structure]    Dynare's results structure,
13
14% Copyright (C) 2001-2019 Dynare Team
15%
16% This file is part of Dynare.
17%
18% Dynare is free software: you can redistribute it and/or modify
19% it under the terms of the GNU General Public License as published by
20% the Free Software Foundation, either version 3 of the License, or
21% (at your option) any later version.
22%
23% Dynare is distributed in the hope that it will be useful,
24% but WITHOUT ANY WARRANTY; without even the implied warranty of
25% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
26% GNU General Public License for more details.
27%
28% You should have received a copy of the GNU General Public License
29% along with Dynare.  If not, see <http://www.gnu.org/licenses/>.
30
31warning_old_state = warning;
32warning off
33
34if isempty(var_list)
35    var_list = M_.endo_names(1:M_.orig_endo_nbr);
36end
37
38nvar = length(var_list);
39ivar=zeros(nvar,1);
40for i=1:nvar
41    i_tmp = strmatch(var_list{i}, M_.endo_names, 'exact');
42    if isempty(i_tmp)
43        error ('One of the variable specified does not exist') ;
44    else
45        ivar(i) = i_tmp;
46    end
47end
48
49y = y(ivar,options_.drop+1:end)';
50
51ME_present=0;
52if ~all(M_.H==0)
53    if (isoctave && octave_ver_less_than('6')) || (~isoctave && matlab_ver_less_than('8.1'))
54        [observable_pos_requested_vars, index_subset, index_observables] = intersect_stable(ivar, options_.varobs_id);
55    else
56        [observable_pos_requested_vars, index_subset, index_observables] = intersect(ivar, options_.varobs_id, 'stable');
57    end
58    if ~isempty(observable_pos_requested_vars)
59        ME_present=1;
60        i_ME = setdiff([1:size(M_.H,1)],find(diag(M_.H) == 0)); % find ME with 0 variance
61        chol_S = chol(M_.H(i_ME,i_ME)); %decompose rest
62        shock_mat=zeros(options_.periods,size(M_.H,1)); %initialize
63        shock_mat(:,i_ME)=randn(length(i_ME),options_.periods)'*chol_S;
64        y_ME = y(:,index_subset)+shock_mat(options_.drop+1:end,index_observables);
65        y_ME_only = shock_mat(options_.drop+1:end,index_observables);
66        m_ME = mean(y_ME);
67        y_ME=get_filtered_time_series(y_ME,m_ME,options_);
68        y_ME_only_filtered=get_filtered_time_series(y_ME_only,mean(y_ME_only),options_);
69        s2_ME = mean(y_ME.*y_ME);
70        s_ME = sqrt(s2_ME);
71        zero_variance_ME_var_index=index_subset(abs(s_ME')<options_.zero_moments_tolerance);
72    end
73end
74
75
76m = mean(y);
77
78% filter series
79y=get_filtered_time_series(y,m,options_);
80
81s2 = mean(y.*y);
82s = sqrt(s2);
83oo_.mean = transpose(m);
84oo_.var = y'*y/size(y,1);
85oo_.skewness = (mean(y.^3)./s2.^1.5)';
86oo_.kurtosis = (mean(y.^4)./(s2.*s2)-3)';
87
88zero_variance_var_index=find(abs(s)<options_.zero_moments_tolerance);
89oo_.skewness(zero_variance_var_index)=NaN;
90oo_.kurtosis(zero_variance_var_index)=NaN;
91s(zero_variance_var_index)=0;
92s2(zero_variance_var_index)=0;
93oo_.var(zero_variance_var_index,:)=0;
94oo_.var(:,zero_variance_var_index)=0;
95
96
97labels = M_.endo_names(ivar);
98labels_TeX = M_.endo_names_tex(ivar);
99
100if ~options_.nomoments
101    z = [ m' s' s2' oo_.skewness oo_.kurtosis ];
102    title='MOMENTS OF SIMULATED VARIABLES';
103    title=add_filter_subtitle(title, options_);
104    headers = {'VARIABLE'; 'MEAN'; 'STD. DEV.'; 'VARIANCE'; 'SKEWNESS'; 'KURTOSIS'};
105    dyntable(options_, title, headers, labels, z, cellofchararraymaxlength(labels)+2, 16, 6);
106    if options_.TeX
107        dyn_latex_table(M_, options_, title, 'sim_moments', headers, labels_TeX, z, cellofchararraymaxlength(labels)+2, 16, 6);
108    end
109end
110
111if ~options_.nocorr
112    corr = (y'*y/size(y,1))./(s'*s);
113    corr(zero_variance_var_index,:)=NaN;
114    corr(:,zero_variance_var_index)=NaN;
115    if options_.contemporaneous_correlation
116        oo_.contemporaneous_correlation = corr;
117    end
118    if ~options_.noprint
119        title = 'CORRELATION OF SIMULATED VARIABLES';
120        title=add_filter_subtitle(title,options_);
121        headers = vertcat('VARIABLE', M_.endo_names(ivar));
122        dyntable(options_, title, headers, labels, corr, cellofchararraymaxlength(labels)+2, 8, 4);
123        if options_.TeX
124            headers = vertcat('VARIABLE', M_.endo_names_tex(ivar));
125            lh = cellofchararraymaxlength(labels)+2;
126            dyn_latex_table(M_, options_, title, 'sim_corr_matrix', headers, labels_TeX, corr, lh, 8,4);
127        end
128    end
129end
130
131if ~options_.noprint && length(options_.conditional_variance_decomposition)
132    fprintf('\nSTOCH_SIMUL: conditional_variance_decomposition requires theoretical moments, i.e. periods=0.\n')
133end
134
135ar = options_.ar;
136if ar > 0
137    autocorr = [];
138    for i=1:ar
139        oo_.autocorr{i} = y(ar+1:end,:)'*y(ar+1-i:end-i,:)./((size(y,1)-ar)*std(y(ar+1:end,:))'*std(y(ar+1-i:end-i,:)));
140        oo_.autocorr{i}(zero_variance_var_index,:)=NaN;
141        oo_.autocorr{i}(:,zero_variance_var_index)=NaN;
142        autocorr = [ autocorr diag(oo_.autocorr{i}) ];
143    end
144    if ~options_.noprint
145        title = 'AUTOCORRELATION OF SIMULATED VARIABLES';
146        title=add_filter_subtitle(title,options_);
147        headers = vertcat('VARIABLE', cellstr(int2str([1:ar]')));
148        dyntable(options_, title, headers, labels, autocorr, cellofchararraymaxlength(labels)+2, 8, 4);
149        if options_.TeX
150            headers = vertcat('VARIABLE', cellstr(int2str([1:ar]')));
151            lh = cellofchararraymaxlength(labels)+2;
152            dyn_latex_table(M_, options_, title, 'sim_autocorr_matrix', headers, labels_TeX, autocorr, cellofchararraymaxlength(labels_TeX)+2, 8, 4);
153        end
154    end
155
156end
157
158
159if ~options_.nodecomposition
160    if M_.exo_nbr == 1
161        oo_.variance_decomposition = 100*ones(nvar,1);
162    else
163        oo_.variance_decomposition=zeros(nvar,M_.exo_nbr);
164        %get starting values
165        if isempty(M_.endo_histval)
166            y0 = oo_.dr.ys;
167        else
168            if options_.loglinear
169                y0 = log_variable(1:M_.endo_nbr,M_.endo_histval,M_);
170            else
171                y0 = M_.endo_histval;
172            end
173        end
174        %back out shock matrix used for generating y
175        i_exo_var = setdiff([1:M_.exo_nbr],find(diag(M_.Sigma_e) == 0)); % find shocks with 0 variance
176        chol_S = chol(M_.Sigma_e(i_exo_var,i_exo_var)); %decompose rest
177        shock_mat=zeros(options_.periods,M_.exo_nbr); %initialize
178        shock_mat(:,i_exo_var)=oo_.exo_simul(:,i_exo_var)/chol_S; %invert construction of oo_.exo_simul from simult.m
179
180        for shock_iter=1:length(i_exo_var)
181            temp_shock_mat=zeros(size(shock_mat));
182            temp_shock_mat(:,i_exo_var(shock_iter))=shock_mat(:,i_exo_var(shock_iter));
183            temp_shock_mat(:,i_exo_var) = temp_shock_mat(:,i_exo_var)*chol_S;
184            y_sim_one_shock = simult_(M_,options_,y0,oo_.dr,temp_shock_mat,options_.order);
185            y_sim_one_shock=y_sim_one_shock(ivar,1+options_.drop+1:end)';
186            y_sim_one_shock=get_filtered_time_series(y_sim_one_shock,mean(y_sim_one_shock),options_);
187            oo_.variance_decomposition(:,i_exo_var(shock_iter))=var(y_sim_one_shock)./s2*100;
188        end
189        oo_.variance_decomposition(zero_variance_var_index,:)=NaN;
190        if ME_present
191            oo_.variance_decomposition_ME=oo_.variance_decomposition(index_subset,:)...
192                .*repmat((s2(index_subset)./s2_ME)',1,length(i_exo_var));
193            oo_.variance_decomposition_ME(:,end+1)=var(y_ME_only_filtered)./s2_ME*100;
194            oo_.variance_decomposition_ME(ismember(observable_pos_requested_vars,intersect(zero_variance_ME_var_index,zero_variance_var_index)),:)=NaN;
195            oo_.variance_decomposition_ME(ismember(observable_pos_requested_vars,setdiff(zero_variance_var_index,zero_variance_ME_var_index)),1:end-1)=0;
196            oo_.variance_decomposition_ME(ismember(observable_pos_requested_vars,setdiff(zero_variance_var_index,zero_variance_ME_var_index)),end)=1;
197        end
198        if ~options_.noprint %options_.nomoments == 0
199            skipline()
200            title='VARIANCE DECOMPOSITION SIMULATING ONE SHOCK AT A TIME (in percent)';
201            title=add_filter_subtitle(title,options_);
202            headers = M_.exo_names;
203            headers(M_.exo_names_orig_ord) = headers;
204            headers = vertcat(' ', headers);
205            lh = cellofchararraymaxlength(M_.endo_names(ivar))+2;
206            dyntable(options_, title, vertcat(headers, 'Tot. lin. contr.'), ...
207                     M_.endo_names(ivar), [oo_.variance_decomposition sum(oo_.variance_decomposition,2)], lh, 8, 2);
208            if ME_present
209                headers_ME = vertcat(headers, 'ME');
210                dyntable(options_, [title,' WITH MEASUREMENT ERROR'], vertcat(headers_ME, 'Tot. lin. contr.'), M_.endo_names(ivar(index_subset)), ...
211                         [oo_.variance_decomposition_ME sum(oo_.variance_decomposition_ME, 2)], lh, 8, 2);
212            end
213            if options_.TeX
214                headers = M_.exo_names_tex;
215                headers = vertcat(' ', headers);
216                labels = M_.endo_names_tex(ivar);
217                lh = cellofchararraymaxlength(labels)+2;
218                dyn_latex_table(M_, options_, title, 'sim_var_decomp', vertcat(headers, 'Tot. lin. contr.'), ...
219                                labels_TeX, [oo_.variance_decomposition sum(oo_.variance_decomposition, 2)], lh, 8, 2);
220                if ME_present
221                    headers_ME = vertcat(headers, 'ME');
222                    dyn_latex_table(M_, options_, [title, ' WITH MEASUREMENT ERROR'], 'sim_var_decomp_ME', ...
223                                    vertcat(headers_ME, 'Tot. lin. contr.'), ...
224                                    labels_TeX(index_subset), ...
225                                    [oo_.variance_decomposition_ME sum(oo_.variance_decomposition_ME, 2)], lh, 8, 2);
226                end
227            end
228
229            if options_.order == 1
230                fprintf('Note: numbers do not add up to 100 due to non-zero correlation of simulated shocks in small samples\n\n')
231            else
232                fprintf('Note: numbers do not add up to 100 due to i) non-zero correlation of simulated shocks in small samples and ii) nonlinearity\n\n')
233            end
234        end
235
236    end
237end
238
239warning(warning_old_state);
240end
241
242function y = get_filtered_time_series(y, m, options_)
243
244if options_.hp_filter && ~options_.one_sided_hp_filter  && ~options_.bandpass.indicator
245    [hptrend,y] = sample_hp_filter(y,options_.hp_filter);
246elseif ~options_.hp_filter && options_.one_sided_hp_filter && ~options_.bandpass.indicator
247    [hptrend,y] = one_sided_hp_filter(y,options_.one_sided_hp_filter);
248elseif ~options_.hp_filter && ~options_.one_sided_hp_filter && options_.bandpass.indicator
249    data_temp=dseries(y,'0q1');
250    data_temp=baxter_king_filter(data_temp,options_.bandpass.passband(1),options_.bandpass.passband(2),options_.bandpass.K);
251    y=data_temp.data;
252elseif ~options_.hp_filter && ~options_.one_sided_hp_filter  && ~options_.bandpass.indicator
253    y = bsxfun(@minus, y, m);
254else
255    error('disp_moments:: You cannot use more than one filter at the same time')
256end
257
258end
259