Definition
Applied General Equilibrium (AGE) Analysis:
- Based on neoclassical theories of firm and household behavior
- Use economy-wide simulation model
- Time frame is long enough to achieve equilibrium
- Usually comparative static analysis (static), but also have dynamic versions.
Comparing with AGE analysis, the limitations of partial equilibrium (PE) analysis
- does not acknowledge finite resource endowment (subsidy pulls resource away from other sectors)
- does not consider the effect of subsidy / tax (who pays subsidy / get tax revenue)
- does not capture income effect endogenously ( no link between consumer's income and its expenditure)
- cannot apply consistency check: Walras's Law
PE example: subsidy
Consider an example: subsidy on food production
From PE aspect, we only consider supply and demand change of this good (food) (single market), assuming all possible equilibrium on the single market can be produced and consumed.
This PE example is equivalent to a supply-demand figure and we add a subsidy for producer
In functional form, it would be:
PS = PM*TO: producer's price = market price * tax ratio (> 1 because it is a subsidy), producer earns more.
QO = S(PS): production is the supply function of producer's price
QD= D(PM): consumption is the demand function of consumer's price
QO = QD: this is the market clear condition holds at the equilibrium
If we write them in percentage term:

Understanding percentage form model: we start with an equilibrium condition. When exogenous variable shocks (here is to), the linearized model tells the percentage change of endogenous variables in response.
Where η are price elasticity: % change of quantity with respect to % change of price
Note: variables in lower case are percentage change. They are not percentage change between variables but are changes before and after we alter the tax / subsidy.
Note: according to linearization of equation, if P and Q has linear relationship with no intercept: for example, PS = βQO, then the elasticity = 1.
PS = βQO
ps = qo
Note: although after shock, PM and PS do not equal their original value, but we know the change of tax rate to, which is exogenous.
Using the market clear condition: QO = QD so qo = qd, solve it as a function of parameters (η) and exogenous shock (to), we know the impact of tax on market price under new equilibrium is:
to > 0 (subsidy), ηs > 0, ηd < 0

(subsidy goes to the agent with less elasticity)
Note: this is the PE model: we focus on a certain goods' supply and demand, and solve it when price and quantity both match.
More detailed explanation of this example is available here.
Multiple commodities
If there are n commodities, the percentage change in market price would be:

Where:
: n matrix of supply elasticities
: n matrix of demand elasticities
Limitation of PE model
- It ignores the impact of shock on other goods. For example, subsidy in agriculture would boost input use in agriculture, since the total endowment of inputs in an economy is given, other sector would have less input to use, but this effect is not considered in PE model.
- It ignores who afford the subsidy, and the opportunity cost of subsidy (would have been spent for other purpose)
- It does not have an explicit budget constraint for households, no link between source and use of income. Especially when the shock is large enough to change expenditure pattern
- It does not have a definitive check on the model. We cannot use Walras's law to check the consistency of model.
Or to summarize:
- producer: not consider input competition with other sectors (infinite input supply with given exogenous price)
- consumer: not consider budget competition with other commodities
- source and effect of subsidy
- no Walras's law
GE Example: subsidy
To solve the impact of shock with GE aspect, we need to develop a GE model (a simple one here).
Closed economy
Four agents
- representative food producer
- representative nonfood producer
- representative household
- fiscal entity
Two inputs
- farmland (land), only used in food production
- capital / labor aggregate, used in food and nonfood production
Two goods
- food
- nonfood
This figure shows the flow of material in the economy (flow of monetary value is opposite)
For household, we have the utility function with two goods (food and nonfood)
Each goods have their share of consumption (CONSHR)
For food producer, we have the production function with two inputs (land and labor)
CONSHR: Share of consumption on goods
FACTSHR: factor share of factor i spent on production for good j, divided by all goods j
FACTSHR(labor, food) + FACTSHR (labor, nonfood) = 1
QFE(labor, food) = FACTSHR(labor, food) * QO(labor)
SVA: SVA(labor, food) + SVA(land, food) = 1, so I think SVA is the share of value added (share of each factor over total factor input)
To summarize:
- CONSHR: for the consumer, separating consumption on different goods
- FACTSHR: for an input, separating its inputs supplied to different producers: separate single inputs to sectors
- SVA: for a producer, separating its use of different types of input: share of inputs used in a single sector
Note: So far, we haven't set production and utility function. All equations are based on the fractional relationship of variables (forever hold)
Income formation
Before tax or subsidy:
Household income y from two primary factors
means endowment. The consumer have no incentive to set aside inputs.
Y is spent on goods:
CONSHR(food) Y is spent on food,
CONSHR(nonfood) Y is spent on nonfood.
After tax or subsidy
Y = PM(land)* QO(land) + PM(labor)*QO(labor) + net taxes
Producer
Assume constant return to scale
Input use by sector
QFE(factor, good) = FACTSHR(factor, good) * QO(factor): the quantity of factor spent on certain goods' production
Based on "Hertel, T. W. (1990). General equilibrium analysis of US agriculture: What does it contribute?. The Journal of Agricultural Economics Research, 42(3), 3-9:"
For food sector, cost share for labor and land are
and
, they are substitutable in food production with elasticity of substitution
(Here
is represented as a non-negative value)
Then the food sector's partial equilibrium supply response is
Note: this
is the same as supply elasticity
in partial equilibrium model.
Tax's impact on market price
Let the price of land to be numeraire, so
, pm is the percentage change of market price (price for consumer)
[recall: to is subsidy for food production ps = pm + to]
So we have a vector of pm changes in the economy:
x is a shift factor, which is a function of all type model shares and elasticities
This is the special version of 4 good model solved by Keller in chapter 10.7
Note: although formulas are written in multiple lines, they are not vectors, but just one line formula.
Revise this equation
Set land price to be numeraire, and omit supply = demand for land.
Special cases of the model
If
= 0 (demand is inelastic) [I suppose
is the substitution elasticity between consumed goods)
or
becomes infinity (supply is perfectly elastic)
Then we have: x = 0
pm(food) = -to
pm(nonfood) = 0
pm(labor) = 0
Then all price changes goes to consumer, who is inelastic (same conclusion as we get from PE model)
If
= 0 (no substitution between land and labor)
or
= 0 (no labor is used in food sector, or only land is used in food sector)
Then we have:

This is based on
, 
Since x > 0, in this case all price go down relative to land price. Thus landowner's income rises (the income from land can now afford more expenditure since other goods' price decrease).
Implication: benefit of subsidy are shifted back to landowners.
Compare PE and GE analysis of subsidy
If we focus on the impact of subsidy on food price
And assume the food sector is small (
is close to 0, I think it means the impact on food sector has little impact on the competition of labor with nonfood sector)
In that case, PE analysis is good enough.
PE model for pm(food): 
From GE model, we have:
Why we need GE analysis
- Theoretical consistency:
- CGE model is not a black box
- We can use Walras's law as a definitive computational check
- Accounting consistency
- Taken the impact of limited resource into consideration, including
- commodity and factor market clearing condition
- private and public household budget constraints
- balance of payment conditions
- Taken the impact of limited resource into consideration, including
- Inter-industry effects
- To have an exhaustive model to account for all effects outside of the target sector, since other sector may also related with the target sector
- Welfare analysis
- PE model focus on commodity price, consumer and producer surplus
- GE model focus on policy change on factors, household, and ultimately people.
When should we avoid using GE
- When only interested in sectoral effects
- When need to introduce lots of complexity in the industry
- When economic wide data is not available
- When time is a constraint
When should use GE
- When research question across multiple sectors, for example:
- trade liberalization
- tax reform
- growth and technological change
- When inter-industry linkages are important