# Code for "Reserve Speculative Attacks", by

# Manuel Amador, Javier Bianchi, Luigi Bocola, and Fabrizio Perri

# published at

# The Journal of Economics Dynamics and Control,

# November 2016, Volume 72, Pages 125-137

#

# Python 3.5

#

# Main file

import modelclass as gs

import time

from matplotlib import pyplot as plt

import seaborn as sns

if __name__ == "__main__":

# Money demand parameters

rbar = None # Using log-log money demand

psi = 415 # elasticity parameter

zlb = - 100 # No ZLB

# Foreign interest rate process

ihigh = gs.annual_to_monthly(1.5 / 100) # high interest rate

ilow = gs.annual_to_monthly(0.0 / 100) # low interest rate

interest_transition = {

0: (1 - 1.7/100, 1.7/100),

1: (1.0/100, 1 - 1.0/100)

} # transition matrix for the interest rate

lam = 0.4 / 100 # probability of abandonment (lambda)

ebar = 0.7 # exchange rate if abandonment occurs

epeg = 1.0 # level of exchange rate peg

gamma = 3.5 / 100 # probability of money demand increasing by g

g_val = 0.505 # growth rate of money demand ( g )

# initial level of money demand

b0 = psi * ((1 + ihigh) * (lam * ebar / epeg + (1 - lam)) - 1)

N = 24 # number of states

par = {

'psi': psi,

'rbar': rbar,

'zlb': zlb,

'b0': b0,

'N': N,

'nw': 0.2, # net worth of the central bank

'pibar': 1.6, # constraint on central bank losses

'epeg': epeg,

'ebar': ebar,

'gamma': gamma,

'lam': lam,

'g_val': g_val,

'c_states': (0, 1), # (low perm, low transitory, high transitory)

'c_transition': interest_transition,

'c_istar_values': (ilow, ihigh)

}

sns.set_style("whitegrid") # optional seabird settings

t1 = time.clock()

model = gs.Model(**par)

print("time : {} seconds".format(time.clock() - t1))

gs.do_plots(model, save_to_file=False, file_name='benchmark_figure.pdf')

robustness = {}

labels = {}

markers = ['^--', 'o-', 's--']

robustness['ebar'] = []

labels['ebar'] = []

for ebar_val, m in zip([0.65, 0.7], markers):

par2 = par.copy()

par2['ebar'] = ebar_val

robustness['ebar'].append(gs.Model(**par2))

lab = {}

lab['label'] = '$1 - \\bar S = {:.2}$'.format(1 - ebar_val)

lab['marker'] = m

labels['ebar'].append(lab)

robustness['lambda'] = []

labels['lambda'] = []

for lambda_values, m in zip([0.6, 0.4], markers):

par2 = par.copy()

par2['lam'] = lambda_values / 100

robustness['lambda'].append(gs.Model(**par2))

lab = {}

lab['label'] = '$\\lambda = ' + str(lambda_values) + '$'

lab['marker'] = m

labels['lambda'].append(lab)

robustness['pibar'] = []

labels['pibar'] = []

for pibar_val, m in zip([1, 1.6], markers):

par2 = par.copy()

par2['pibar'] = pibar_val

robustness['pibar'].append(gs.Model(**par2))

lab = {}

lab['label'] = '$\\bar \\Pi = ' + str(pibar_val) + '$'

lab['marker'] = m

labels['pibar'].append(lab)

robustness['psi'] = []

labels['psi'] = []

for psi, m in zip(

[212, 415],

markers):

par2 = par.copy()

par2['psi'] = psi

par2['b0'] = (

psi * ((1 + par2['c_istar_values'][1]) *

(par2['lam'] * par2['ebar'] / par2['epeg'] +

(1 - par2['lam'])) -

1))

robustness['psi'].append(gs.Model(**par2))

lab = {}

lab['label'] = '$\\psi = {:.2}$'.format(psi / 12 / 100 - 0.01)

lab['marker'] = m

labels['psi'].append(lab)

fs = 15

b_range = list(range(4))

plt.figure(figsize=(10, 8))

for i, key in enumerate(['ebar', 'lambda', 'pibar', 'psi']):

plt.subplot(221 + i)

model_list = robustness[key]

temp = labels[key]

e_min = model_list[0].epeg

c = 1 # plot only the high interest rate

for m, lab in zip(model_list, temp):

e_list = [m.e_rate['top'][(1, b, c)] for b in b_range]

plt.plot(e_list, lab['marker'], label=lab['label'])

e_min = min(e_min, min(e_list))

plt.ylim([.85, 1.01])

plt.title([

'A. Size of Appreciation Shock',

'B. Probability of Appreciation Shock',

'C. Tightness of Loss Constraint',

'D. Elasticity of Money Demand'][i], fontsize=fs)

plt.xticks(b_range, [b + 1 for b in b_range])

if i > 1:

plt.xlabel('b')

if i in (0, 2):

plt.ylabel('Exchange Rate')

plt.legend(loc=3, handlelength=4)

plt.subplots_adjust(wspace=0.3)

# plt.savefig('robustness_E.pdf')

plt.figure(figsize=(10, 8))

for i, key in enumerate(['ebar', 'lambda', 'pibar', 'psi']):

plt.subplot(221 + i)

model_list = robustness[key]

temp = labels[key]

c = 1 # plot only the high interest rate

for m, lab in zip(model_list, temp):

plt.plot([m.reserves[(1, b, c)]

for b in b_range], lab['marker'], label=lab['label'])

plt.legend(loc=4, handlelength=4)

plt.title([

'A. Size of Appreciation Shock',

'B. Probability of Appreciation Shock',

'C. Tightness of Loss Constraint',

'D. Elasticity of Money Demand'][i], fontsize=fs)

if i > 1:

plt.xlabel('b')

if i in (0, 2):

plt.ylabel('Reserves')

plt.xticks(b_range, [b + 1 for b in b_range])

plt.ylim([0, 12])

plt.subplots_adjust(wspace=0.25)

# plt.savefig('robustness_F.pdf')

robustness['psi'][0].e_rate['top']

plt.show()