Relax assumption on government consumption good production

[jdebacker]

We currently assume that goods/services making up government consumption expenditures are produced from just one industry (namely, industry $M$).

However, given data from an input/output table, it would be trivial to find demand for output from any sector $m$ that results from gov't consumption expenditures.

Analogous to how we model differentiated consumption goods, we can have a unit of gov't expenditures be made up of outputs from different sectors using a fixed-coefficient matrix. E.g., demand for industry $m$ output from gov't consumption will be given as:

$$ G_{m,t} = \pi_{gov, m} G_{t} $$

where $\pi_{gov, m}$ is the fraction of a unit of gov't spending that comes from industry $m$.

The price of a unit of gov't consumption will then be:

$$ p_{gov,t} = \sum_{m=1}^M \pi_{gov, m} p_{m} $$

Government spending in the government's budget constraint, currently $G_{t}$ will be replaced by $p_{gov,t}G_{t}$ since this change will mean that gov't expenditures are not only on the numeraire good (which is the current assumption in OG-Core).