age-varying distributions for shocks

[sbenthall]

Consider the Adjust shock in risky asset models.

HARK/HARK/ConsumptionSaving/ConsRiskyAssetModel.py

Lines 184 to 216 in 00ce66a

	def get_Adjust(self):
	"""
	Sets the attribute Adjust as a boolean array of size AgentCount, indicating
	whether each agent is able to adjust their risky portfolio share this period.
	Uses the attribute AdjustPrb to draw from a Bernoulli distribution.
	Parameters
	----------
	None
	Returns
	-------
	None
	"""
	if not ("AdjustPrb" in self.time_vary):
	self.shocks["Adjust"] = Bernoulli(
	self.AdjustPrb, seed=self.RNG.randint(0, 2 ** 31 - 1)
	).draw(self.AgentCount)
	else:
	Adjust = np.zeros(self.AgentCount) # Initialize shock array
	for t in range(self.T_cycle):
	these = t == self.t_cycle
	N = np.sum(these)
	if N > 0:
	AdjustPrb = self.AdjustPrb[t - 1]
	Adjust[these] = Bernoulli(
	AdjustPrb, seed=self.RNG.randint(0, 2 ** 31 - 1)
	).draw(N)
	self.shocks["Adjust"] = Adjust

This is some complicated case-handling code that could be extracted into an age-varying distribution class and put into distribution.py.

Related to #1003 and #890

[sbenthall]

I'll need the test case from #832 to complete the work on this.

Otherwise, it's quite easy.

[sbenthall]

fixed by #1018