Exercise 2.1
Consumption Innovations
Problem
In the economy with AR(1) endowment shocks studied in section 2.2, we found that \(E_{t}c_{t+1}=c_{t}\), which means that \(c_{t+1} = c_{t}+\mu_{t+1}\), where \(\mu_{t+1}\) is a white noise process that is unforecastable given information available in t. Derive the innovation \(\mu_{t+1}\) as a function of r, \(\rho\), and \(\epsilon_{t+1}\).
Answer
Using equation (2.18), we have:
\[ c_{t+1} = \frac{r}{1 + r - \rho} y_{t+1} - r d_t \tag{1} \]
\[ c_t = \frac{r}{1 + r - \rho} y_t - r d_{t-1} \tag{2} \]
Subtracting (2) from (1):
\[ c_{t+1} - c_t = \frac{r}{1 + r - \rho} (y_{t+1} - y_t) - r (d_t - d_{t-1}) \tag{3} \]
Now plug in the AR(1) process for \(y_t\):
\[ y_{t+1} = \rho y_t + \epsilon_{t+1} \]
So:
\[ y_{t+1} - y_t = (\rho - 1) y_t + \epsilon_{t+1} \]
Then:
\[ c_{t+1} - c_t = \frac{r}{1 + r - \rho} ((\rho - 1) y_t + \epsilon_{t+1}) - r(d_t - d_{t-1}) \tag{4} \]
Now use the law of motion for external debt (equation 2.21):
\[ d_t - d_{t-1} = \frac{1 - \rho}{1 + r - \rho} y_t \]
Substitute into the equation:
\[ c_{t+1} - c_t = \frac{r}{1 + r - \rho} ((\rho - 1) y_t + \epsilon_{t+1}) - r \cdot \frac{1 - \rho}{1 + r - \rho} y_t \tag{5} \]
The terms involving $ y_t $ cancel:
\[ c_{t+1} - c_t = \frac{r}{1 + r - \rho} \epsilon_{t+1} \tag{6} \]
The innovation term is:
\[ \mu_{t+1} = c_{t+1} - c_t = \frac{r}{1 + r - \rho} \epsilon_{t+1} \]