{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Linear Regression: Rolling and Recursive\n",
    "\n",
    "**Functions**\n",
    "\n",
    "`sm.OLS`, `plt.title`, `plt.legend`, `plt.subplots`, `plt.plot`\n",
    "\n",
    "### Exercise 35\n",
    "For the same portfolios in the previous exercise, compute rolling $\\beta$s\n",
    "using 60 consecutive observations."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-09-28T12:33:53.629390Z",
     "iopub.status.busy": "2023-09-28T12:33:53.629390Z",
     "iopub.status.idle": "2023-09-28T12:33:56.100580Z",
     "shell.execute_reply": "2023-09-28T12:33:56.099572Z"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Exercise 36\n",
    "\n",
    "For each of the four $\\beta$s, produce a plot containing four series: \n",
    "\n",
    "* A line corresponding to the constant $\\beta$ (full sample) \n",
    "* The $\\beta$ estimated on the rolling sample \n",
    "* The constant $\\beta$ plus $1.96 \\times$ the variance of a 60-observation estimate of $\\beta$. \n",
    "  The 60-month covariance can be estimated using a full sample VCV and rescaling it by T/60 \n",
    "  where T is the length of the full sample used to estimate the VCV. Alternatively, the VCV\n",
    "  could be estimated by first estimating the 60-month VCV for each sub-sample and then averaging\n",
    "  these.\n",
    "* The constant $\\beta$ minus $1.96 \\times$ the variance of a 60-observation estimate of $\\beta$. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-09-28T12:33:59.010487Z",
     "iopub.status.busy": "2023-09-28T12:33:59.010487Z",
     "iopub.status.idle": "2023-09-28T12:34:06.378621Z",
     "shell.execute_reply": "2023-09-28T12:34:06.378621Z"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Exercise 37\n",
    "\n",
    "Do the factor exposures appear constant?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-09-28T12:34:06.382426Z",
     "iopub.status.busy": "2023-09-28T12:34:06.382426Z",
     "iopub.status.idle": "2023-09-28T12:34:06.387574Z",
     "shell.execute_reply": "2023-09-28T12:34:06.387574Z"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Exercise 38\n",
    "\n",
    "What happens if only the market is used as a factor (repeat the exercise excluding SMB and HML)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-09-28T12:34:06.390581Z",
     "iopub.status.busy": "2023-09-28T12:34:06.389582Z",
     "iopub.status.idle": "2023-09-28T12:34:09.306277Z",
     "shell.execute_reply": "2023-09-28T12:34:09.306277Z"
    }
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.5"
  },
  "pycharm": {
   "stem_cell": {
    "cell_type": "raw",
    "metadata": {
     "collapsed": false
    },
    "source": []
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}