BEARX-tutorials/XConditionalForecastsPanel.m at master · OGResearch/BEARX-tutorials · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
%% Conditional forecasts
%%
% * Prepare a reduced-form model for experiments with zero restrictions
% * Prepare a table with conditions
% * Prepare a table with a "simulation plan"
% * Run and report a conditional report using all shocks vs seletected shocks


clear
close all
rehash path

addpath ../sandbox
addpath ../bear

%% Convenience functions
%
% The |extremesFunc| function compresses any number of samples (draws from the
% posterior) into two numbers - the minimum and the maximum.


percentiles = [10, 50, 90];
prctilesFunc = @(x) prctile(x, percentiles, 2);
extremesFunc = @(x) [min(x, [], 2), max(x, [], 2)];

defaultColors = get(0, "defaultAxesColorOrder");

%% Prepare data and a reduced-form model
%
% * Same as in introCommonTasks

inputTbx = tablex.fromCsv("panel_data.csv");

estimStart = datex.q(1972,1);
estimEnd = datex.q(2014,4);
estimSpan = datex.span(estimStart, estimEnd);

meta = model.Meta( ...
    endogenous=["YER", "HICSA", "STN"], ...
    units=["US", "EA", "UK"], ...
    exogenous="Oil", ...
    order=4, ...
    intercept=true, ...
    estimationSpan=estimSpan, ...
    ...
    identificationHorizon=12, ...
    shockConcepts=["DEM", "SUP", "POL"] ...
);

dataH = model.DataHolder(meta, inputTbx);

estimatorR = estimator.NormalWishartPanel(meta);

modelR = model.ReducedForm( ...
    meta=meta ...
    , dataHolder=dataH ...
    , estimator=estimatorR ...
    , stabilityThreshold=Inf ...
);

%% Indentify a SVAR using Cholesky with reordering
%
% * Use Cholesky as if the endogenous variables were ordered in a different
% way than in meta
% * If a certain trailing portion of the order follows the meta order, you can
% omit that part


identChol = identifier.Cholesky(order=["YER", "HICSA", "STN"]);

% Equivalent to
% identChol = identifier.Cholesky(order=["DOM_CPI"]);

modelS0 = model.Structural(reducedForm=modelR, identifier=identChol);

modelS0

modelS0.initialize();
info0 = modelS0.presample(100);

modelS0.Presampled{1}.D
modelS0.Presampled{2}.D

respTbx0 = modelS0.simulateResponses();
respTbx0 = tablex.apply(respTbx0, extremesFunc);
respTbx0 = tablex.flatten(respTbx0);

respTbx0


%% Create forecast assumptions


fcastStart = datex.shift(estimEnd, 1);
fcastEnd = datex.shift(estimEnd, 12);
fcastSpan = datex.span(fcastStart, fcastEnd);
initStart = datex.shift(fcastStart, -modelS0.Meta.Order);

[dataTbx, planTbx] = tablex.forConditional(modelS0, fcastSpan);
dataTbx
planTbx

dataTbx{datex("2015-Q4"), "US_YER"} = -1.5;
dataTbx{datex("2016-Q4"), "US_HICSA"} = 5.5;
%dataTbx{datex("2016-Q3"), "STN"} = 5.5;

dataTbx{:, "Oil"} = inputTbx{end, "Oil"};


%% Run across-the-board vs selective conditions forecasts


planTbx{datex("2015-Q4"), "US_YER"} = "DEM POL";
planTbx{datex("2016-Q4"), "US_HICSA"} = "DEM SUP";
%planTbx{datex("2016-Q3"), "DOM_CPI"} = "SUP";


histContTbx = modelS0.calculateContributions();

rng(0);

[cfcastTbx1, cfcastContTbx1] = modelS0.conditionalForecast( ...
    fcastSpan, ...
    conditions=dataTbx, ...
    plan=[], ...
    contributions=true ...
);

cfcastPrctilesTbx1 = tablex.apply(cfcastTbx1, prctilesFunc);


rng(0);

[cfcastTbx2, cfcastContTbx2] = modelS0.conditionalForecast( ...
    fcastSpan, ...
    conditions=dataTbx, ...
    plan=planTbx, ...
    contributions=true ...
);

cfcastPrctilesTbx2 = tablex.apply(cfcastTbx2, prctilesFunc);


%% Visualize conditional forecasts


plotSettings = { ...
   {"color"}, {defaultColors(2,:); defaultColors(1,:); defaultColors(2,:)},  ...
   {"lineStyle"}, {":";"-";":"}, ...
};

ch = visual.Chartpack( ...
    span=datex.span(initStart, fcastEnd), ...
    namesToPlot=[modelS0.Meta.EndogenousNames, modelS0.Meta.ShockNames], ...
    plotSettings=plotSettings ...
);

ch.Captions = "Across-the-board conditional forecast";
ch.plot(cfcastPrctilesTbx1);

ch.Captions = "Selective conditional forecast";
ch.plot(cfcastPrctilesTbx2);