diff --git a/causal-inference-for-the-brave-and-true/08-Instrumental-Variables.ipynb b/causal-inference-for-the-brave-and-true/08-Instrumental-Variables.ipynb
index ba054f2..a2904b8 100644
--- a/causal-inference-for-the-brave-and-true/08-Instrumental-Variables.ipynb
+++ b/causal-inference-for-the-brave-and-true/08-Instrumental-Variables.ipynb
@@ -776,7 +776,7 @@
"\n",
"## Instrumental Variables by Hand\n",
"\n",
- "Having both our reduced form and our 1st stage, we can now scale the effect of the first stage by the reduced form. Since the first stage coefficient was something like 0.1, this will multiply the effect of the reduced form coefficient by almost 10. This will give us our unbiased IV estimate of the average causal effect:\n",
+ "Having estimated both the reduced form and the 1st stage, we can divide the reduced-form effect by the 1st-stage coefficient. Since the 1st-stage coefficient was about 0.1, this scales the reduced-form coefficient by about 10. This gives us the IV estimate of the average causal effect:\n",
"\n",
"$\n",
"\\mathrm{ATE}_{IV} = \\dfrac{\\text{Reduced Form}}{\\text{1st Stage}} \n",
@@ -957,23 +957,23 @@
"cell_type": "markdown",
"metadata": {},
"source": [
- "The return on education is estimated to be lower with OLS than with 2SLS. This suggests that OVB might not be as strong as we first though. Also, notice the confidence intervals. 2SLS has a much wider CI than the OLS estimate. Let's explore this further\n",
+ "The return on education is estimated to be lower with OLS than with 2SLS. This suggests that OVB might not be as strong as we first thought. Also, notice the confidence intervals. 2SLS has a much wider CI than the OLS estimate. Let's explore this further\n",
"\n",
"## Weakness of Instruments\n",
"\n",
"\n",
"\n",
- "When dealing with IV, we need to remember we are estimating the ATE indirectly. Our estimates depend on both the first stage and the second stage. If the impact of the treatment on the outcome is indeed strong, the second stage will also be strong. However, it doesn't matter how strong the second stage is if we have a weak first stage. A weak first stage means that the instrument has only a very small correlation with the treatment. Therefore, we can't learn much about the treatment from the instrument.\n",
+ "When dealing with IV, we need to remember we are estimating the ATE indirectly. Our estimates depend on both the first stage and the second stage. If the impact of the treatment on the outcome is indeed strong, the second stage will also be strong. However, it doesn't matter how strong the second stage is if we have a weak first stage. A weak first stage means that the instrument adds little predictive power for the treatment after accounting for the included controls. Therefore, we can't learn much about the treatment from the instrument.\n",
"\n",
- "The formulas for the IV standard errors are a bit complex and not so intuitive, so we will try something else to grasp this problem. We will simulate data where we have a treatment T with effect 2.0 on the outcome Y, an unobserved confounder U and an additional control X. We will also simulate multiple instruments with different strengths on the 1st stage.\n",
+ "The formulas for the IV standard errors are a bit complex and not so intuitive, so we will use a controlled simulation to grasp this problem. We draw an instrument $Z$, an unobserved confounder $U$, a control $X$, and two error terms once. We then vary only the first-stage coefficient $\\pi$ across 50 separate data-generating processes while keeping those draws fixed. In every design, $Z$ is independent of $U$ and the error terms, and it affects $Y$ only through $T$. Instrument exogeneity and the exclusion restriction therefore hold by construction, while the true causal effect of $T$ on $Y$ remains 2.0.\n",
"\n",
"$$\n",
"\\begin{align}\n",
- "X \\sim & N(0, 2^2)\\\\\n",
- "U \\sim & N(0, 2^2)\\\\\n",
- "T \\sim & N(1+0.5U, 5^2)\\\\\n",
- "Y \\sim & N(2+ X - 0.5U + 2T, 5^2)\\\\\n",
- "Z \\sim & N(T, \\sigma^2) \\text{ for }\\sigma^2 \\text{ in 0.1 to 100}\n",
+ "X, U, Z \\stackrel{ind}{\\sim} & N(0, 2^2)\\\\\n",
+ "\\epsilon_T, \\epsilon_Y \\stackrel{ind}{\\sim} & N(0, 5^2)\\\\\n",
+ "T_{\\pi} = & 1 + 0.5U + \\pi Z + \\epsilon_T\\\\\n",
+ "Y_{\\pi} = & 2 + X - 0.5U + 2T_{\\pi} + \\epsilon_Y\\\\\n",
+ "\\pi \\in & [0.01, 2]\n",
"\\end{align}\n",
"$$"
]
@@ -1004,68 +1004,65 @@
" \n",
"
\n",
"
\n",
+ "
X
\n",
"
U
\n",
+ "
Z
\n",
"
T
\n",
- "
...
\n",
- "
Z_48
\n",
- "
Z_49
\n",
+ "
Y
\n",
"
\n",
" \n",
"
\n",
"
\n",
"
0
\n",
- "
2.696148
\n",
- "
8.056988
\n",
- "
...
\n",
- "
-117.798705
\n",
- "
-13.485292
\n",
+ "
-1.424781
\n",
+ "
-0.925639
\n",
+ "
1.893819
\n",
+ "
10.736423
\n",
+ "
15.070533
\n",
"
\n",
"
\n",
"
1
\n",
- "
2.570240
\n",
- "
0.245067
\n",
- "
...
\n",
- "
-209.727577
\n",
- "
-70.792948
\n",
+ "
1.507533
\n",
+ "
-0.884565
\n",
+ "
0.135373
\n",
+ "
-6.031675
\n",
+ "
-12.226334
\n",
"
\n",
"
\n",
"
2
\n",
- "
0.664741
\n",
- "
5.597510
\n",
- "
...
\n",
- "
60.562232
\n",
- "
47.619414
\n",
+ "
-0.089006
\n",
+ "
2.085831
\n",
+ "
3.987078
\n",
+ "
-5.987409
\n",
+ "
-6.216387
\n",
"
\n",
"
\n",
"
3
\n",
- "
1.037725
\n",
- "
0.493532
\n",
- "
...
\n",
- "
78.136513
\n",
- "
-108.322304
\n",
+ "
0.903625
\n",
+ "
-2.024321
\n",
+ "
-1.919581
\n",
+ "
-7.135441
\n",
+ "
-5.916757
\n",
"
\n",
"
\n",
"
4
\n",
- "
-2.590591
\n",
- "
-6.263014
\n",
- "
...
\n",
- "
78.776566
\n",
- "
-80.547214
\n",
+ "
2.690203
\n",
+ "
2.419974
\n",
+ "
1.831828
\n",
+ "
15.334038
\n",
+ "
42.761057
\n",
"
\n",
" \n",
"\n",
- "
5 rows × 53 columns
\n",
""
],
"text/plain": [
- " U T ... Z_48 Z_49\n",
- "0 2.696148 8.056988 ... -117.798705 -13.485292\n",
- "1 2.570240 0.245067 ... -209.727577 -70.792948\n",
- "2 0.664741 5.597510 ... 60.562232 47.619414\n",
- "3 1.037725 0.493532 ... 78.136513 -108.322304\n",
- "4 -2.590591 -6.263014 ... 78.776566 -80.547214\n",
- "\n",
- "[5 rows x 53 columns]"
+ " X U Z T Y\n",
+ "0 -1.424781 -0.925639 1.893819 10.736423 15.070533\n",
+ "1 1.507533 -0.884565 0.135373 -6.031675 -12.226334\n",
+ "2 -0.089006 2.085831 3.987078 -5.987409 -6.216387\n",
+ "3 0.903625 -2.024321 -1.919581 -7.135441 -5.916757\n",
+ "4 2.690203 2.419974 1.831828 15.334038 42.761057"
]
},
"execution_count": 17,
@@ -1074,17 +1071,22 @@
}
],
"source": [
- "np.random.seed(12)\n",
+ "np.random.seed(13)\n",
"n = 10000\n",
- "X = np.random.normal(0, 2, n) # observable variable\n",
- "U = np.random.normal(0, 2, n) # unobservable (omitted) variable\n",
- "T = np.random.normal(1 + 0.5*U, 5, n) # treatment\n",
- "Y = np.random.normal(2 + X - 0.5*U + 2*T, 5, n) # outcome\n",
+ "X = np.random.normal(0, 2, n) # observable variable\n",
+ "U = np.random.normal(0, 2, n) # unobservable (omitted) variable\n",
+ "Z = np.random.normal(0, 2, n) # instrument\n",
+ "treatment_noise = np.random.normal(0, 5, n)\n",
+ "outcome_noise = np.random.normal(0, 5, n)\n",
"\n",
- "stddevs = np.linspace(0.1, 100, 50)\n",
- "Zs = {f\"Z_{z}\": np.random.normal(T, s, n) for z, s in enumerate(stddevs)} # instruments with decreasing \\mathrm{Cov}(Z, T)\n",
+ "strengths = np.geomspace(0.01, 2, 50)\n",
"\n",
- "sim_data = pd.DataFrame(dict(U=U, T=T, Y=Y)).assign(**Zs)\n",
+ "def make_data(strength):\n",
+ " T = 1 + 0.5*U + strength*Z + treatment_noise\n",
+ " Y = 2 + X - 0.5*U + 2*T + outcome_noise\n",
+ " return pd.DataFrame(dict(X=X, U=U, Z=Z, T=T, Y=Y))\n",
+ "\n",
+ "sim_data = make_data(strengths[-1])\n",
"\n",
"sim_data.head()"
]
@@ -1093,7 +1095,7 @@
"cell_type": "markdown",
"metadata": {},
"source": [
- "Just to double check, we can see that the correlation between Z and T is indeed decreasing."
+ "The preview above uses the strongest first stage, $\\pi=2$. As a quick validity check, the sample correlation between $Z$ and $U$ is close to zero. It is not exactly zero because of sampling variation; the variables are drawn independently in the data-generating process."
]
},
{
@@ -1102,48 +1104,134 @@
"metadata": {},
"outputs": [
{
- "data": {
- "text/plain": [
- "Z_0 0.999807\n",
- "Z_1 0.919713\n",
- "Z_2 0.773434\n",
- "Z_3 0.634614\n",
- "Z_4 0.523719\n",
- "Name: T, dtype: float64"
- ]
- },
- "execution_count": 18,
- "metadata": {},
- "output_type": "execute_result"
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "|Corr(Z, T)| at pi=2: 0.6157215054852875\n",
+ "|Corr(Z, U)|: 0.007343391806201979\n"
+ ]
}
],
"source": [
- "corr = (sim_data.corr()[\"T\"]\n",
- " [lambda d: d.index.str.startswith(\"Z\")])\n",
- "\n",
- "corr.head()"
+ "print(\"|Corr(Z, T)| at pi=2:\", abs(sim_data.corr().loc[\"Z\", \"T\"]))\n",
+ "print(\"|Corr(Z, U)|:\", abs(sim_data.corr().loc[\"Z\", \"U\"]))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
- "Now, we will run one IV model per instrument we have and collect both the ATE estimate and the standard error."
+ "Now, we will run one just-identified IV model for each value of $\\pi$. We collect the ATE estimate, its conventional standard error, and the homoskedastic first-stage $F$ statistic. This is a sequence of one-instrument designs, not a many-instrument or overidentified design."
]
},
{
"cell_type": "code",
"execution_count": 19,
"metadata": {},
- "outputs": [],
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "
\n",
+ "\n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
\n",
+ "
strength
\n",
+ "
first_stage_f
\n",
+ "
ate
\n",
+ "
se
\n",
+ "
\n",
+ " \n",
+ " \n",
+ "
\n",
+ "
0
\n",
+ "
0.010000
\n",
+ "
0.118719
\n",
+ "
-0.417392
\n",
+ "
7.499425
\n",
+ "
\n",
+ "
\n",
+ "
10
\n",
+ "
0.029485
\n",
+ "
1.224024
\n",
+ "
1.247142
\n",
+ "
1.119127
\n",
+ "
\n",
+ "
\n",
+ "
20
\n",
+ "
0.086935
\n",
+ "
11.239293
\n",
+ "
1.751550
\n",
+ "
0.308300
\n",
+ "
\n",
+ "
\n",
+ "
30
\n",
+ "
0.256327
\n",
+ "
99.505175
\n",
+ "
1.916500
\n",
+ "
0.101579
\n",
+ "
\n",
+ "
\n",
+ "
40
\n",
+ "
0.755774
\n",
+ "
870.378805
\n",
+ "
1.971767
\n",
+ "
0.034316
\n",
+ "
\n",
+ "
\n",
+ "
49
\n",
+ "
2.000000
\n",
+ "
6107.090355
\n",
+ "
1.989342
\n",
+ "
0.012960
\n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
"
+ ],
+ "text/plain": [
+ " strength first_stage_f ate se\n",
+ "0 0.010000 0.118719 -0.417392 7.499425\n",
+ "10 0.029485 1.224024 1.247142 1.119127\n",
+ "20 0.086935 11.239293 1.751550 0.308300\n",
+ "30 0.256327 99.505175 1.916500 0.101579\n",
+ "40 0.755774 870.378805 1.971767 0.034316\n",
+ "49 2.000000 6107.090355 1.989342 0.012960"
+ ]
+ },
+ "execution_count": 19,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
"source": [
- "se = []\n",
- "ate = []\n",
- "for z in range(len(Zs)):\n",
- " formula = f'Y ~ 1 + X + [T ~ Z_{z}]'\n",
- " iv = IV2SLS.from_formula(formula, sim_data).fit()\n",
- " se.append(iv.std_errors[\"T\"])\n",
- " ate.append(iv.params[\"T\"])"
+ "results = []\n",
+ "for strength in strengths:\n",
+ " data = make_data(strength)\n",
+ " iv = IV2SLS.from_formula(\"Y ~ 1 + X + [T ~ Z]\", data).fit(cov_type=\"unadjusted\")\n",
+ " results.append({\n",
+ " \"strength\": strength,\n",
+ " \"first_stage_f\": iv.first_stage.diagnostics.loc[\"T\", \"f.stat\"],\n",
+ " \"ate\": iv.params[\"T\"],\n",
+ " \"se\": iv.std_errors[\"T\"],\n",
+ " })\n",
+ "\n",
+ "plot_data = pd.DataFrame(results).sort_values(\"first_stage_f\")\n",
+ "plot_data.iloc[[0, 10, 20, 30, 40, 49]]"
]
},
{
@@ -1153,9 +1241,9 @@
"outputs": [
{
"data": {
- "image/png": "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\n",
+ "image/png": "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",
"text/plain": [
- ""
+ ""
]
},
"metadata": {},
@@ -1163,12 +1251,11 @@
}
],
"source": [
- "plot_data = pd.DataFrame(dict(se=se, ate=ate, corr=corr)).sort_values(by=\"corr\")\n",
- "\n",
- "plt.scatter(plot_data[\"corr\"], plot_data[\"se\"])\n",
- "plt.xlabel(\"Corr(Z, T)\")\n",
+ "plt.scatter(plot_data[\"first_stage_f\"], plot_data[\"se\"])\n",
+ "plt.xscale(\"log\")\n",
+ "plt.xlabel(\"First-Stage F Statistic (log scale)\")\n",
"plt.ylabel(\"IV Standard Error\");\n",
- "plt.title(\"Variance of the IV Estimates by 1st Stage Strength\");"
+ "plt.title(\"Standard Errors by 1st Stage Strength\");"
]
},
{
@@ -1178,9 +1265,9 @@
"outputs": [
{
"data": {
- "image/png": "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\n",
+ "image/png": "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",
"text/plain": [
- ""
+ ""
]
},
"metadata": {},
@@ -1188,44 +1275,47 @@
}
],
"source": [
- "plt.scatter(plot_data[\"corr\"], plot_data[\"ate\"])\n",
- "plt.fill_between(plot_data[\"corr\"],\n",
+ "plt.scatter(plot_data[\"first_stage_f\"], plot_data[\"ate\"])\n",
+ "plt.fill_between(plot_data[\"first_stage_f\"],\n",
" plot_data[\"ate\"]+1.96*plot_data[\"se\"],\n",
- " plot_data[\"ate\"]-1.96*plot_data[\"se\"], alpha=.5)\n",
- "plt.xlabel(\"Corr(Z, T)\")\n",
+ " plot_data[\"ate\"]-1.96*plot_data[\"se\"], alpha=.5,\n",
+ " label=\"Conventional 95% interval\")\n",
+ "plt.axhline(2, color=\"black\", ls=\"dashed\", label=\"True ATE\")\n",
+ "plt.xscale(\"log\")\n",
+ "plt.xlabel(\"First-Stage F Statistic (log scale)\")\n",
"plt.ylabel(\"$\\hat{ATE}$\");\n",
- "plt.title(\"IV ATE Estimates by 1st Stage Strength\");"
+ "plt.title(\"IV Estimates and Conventional 95% Intervals\")\n",
+ "plt.legend();"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
- "As we can see in the plots above, estimates vary wildly when the correlation between T and Z is weak. This is because the SE also increases a lot when the correlation is low.\n",
- "\n",
- "Another thing to notice is that **2SLS is biased**! Even with high correlation, the parameter estimate still does not reach the true ATE of 2.0. Actually, 2.0 is not even in the 95% CI! 2SLS is only consistent, which means that it approaches the true parameter value if the sample size is big enough. However, we can't know how big is big enough. We can only stick to some rules of thumb to understand how this bias behaves:\n",
+ "As the first-stage $F$ statistic approaches zero, the IV estimate becomes extremely imprecise and sensitive to sampling noise. In contrast, estimates based on the strongest first stages are close to the true ATE of 2.0. Since $Z$ is drawn independently of $U$ and has no direct path to $Y$ in every simulated design, this deterioration comes from weak identification rather than a violation of instrument validity.\n",
"\n",
- "1. 2SLS is biased towards OLS. This means that if OLS has a negative/positive bias, 2SLS will also have it. The advantage of 2SLS is that it is at least consistent, where OLS is not, in the case of omitted variables. In the example above, our unobserved U impacts negatively the outcome but its positively correlated with the treatment, which will result in a negative bias. That is why we are seeing the ATE estimate below the true value (negative bias).\n",
+ "The shaded regions are conventional normal-approximation 95% intervals. Those intervals are not reliable under weak identification, so they should not be interpreted as valid weak-IV inference. Weak-identification-robust procedures, such as Anderson-Rubin tests and confidence sets, are designed for this setting. Under standard fixed-instrument asymptotics, instrument validity and a nonzero first stage make 2SLS consistent as the sample size grows. In finite samples, however, weak first stages can produce unstable estimates, severe distortions in conventional confidence intervals, and substantial bias.\n",
"\n",
- "2. The bias will increase with the number of instruments we add. If we add too many instruments, 2SLS becomes more and more like OLS.\n",
+ "A related but distinct result concerns overidentified designs. With many weak instruments, finite-sample 2SLS bias can be substantial and often points towards OLS; adding more weak instruments can aggravate that problem. The controlled experiment above does not demonstrate this many-instrument result: it deliberately fits one instrument at a time so that only first-stage relevance changes.\n",
"\n",
"Besides knowing how this bias behaves, a final piece of advice is to avoid some **common mistakes when doing IV**:\n",
"\n",
"1. Doing IV by hand. As we've seen, IV by hand will result in wrong standard errors, even if the parameter estimates are right. The SE won't be completely off. Still, why do it if you can use software and get the right SE?\n",
"\n",
- "2. Using anything other than OLS on the 1st stage. Lots of Data Scientist encounter IV and think they can do better. For example, they see a dummy treatment and think about replacing the 1st stage by a logistic regression, after all, they are predicting a dummy variable, right?. The problem is that this is plain wrong. The consistency of IV relies on a property that only OLS can give, which is the orthogonality of the residuals, so anything different than OLS on the 1st stage will yield something biased. (OBS: there are some modern techniques that use Machine Learning for IV, but their results have been, at best, questionable).\n",
+ "2. Naively replacing the OLS projection in textbook 2SLS. With a binary treatment, it can be tempting to fit a logistic regression or a black-box predictor in the first stage and then insert its fitted values into the usual second stage. That plug-in procedure is not ordinary 2SLS: it changes the estimator, and the conventional 2SLS standard errors no longer follow automatically. The linear first-stage projection supplies the sample moment and orthogonality structure used by textbook 2SLS, even when the endogenous treatment is binary. This does not mean OLS is the only possible IV method. Nonlinear IV, GMM, control-function, and modern machine-learning IV estimators can be valid under their own moment conditions, regularity assumptions, cross-fitting schemes, and inference theory. The warning is to use a method designed for that setting rather than swapping a predictive model into the standard 2SLS recipe.\n",
"\n",
"## Key Ideas\n",
"\n",
"We've taken some time here to understand how we can work around omitted variable bias if we have an instrument variable. An instrument is a variable that is correlated with the treatment (has a first stage), but only affects the outcome through the treatment (exclusion restriction). We saw an example of an instrument with quarter of birth to estimate the effect of education on income.\n",
"\n",
- "We then delve into the mechanics of estimating the causal effect with IV, namely, using 2SLS. We've also learned that IV is no silver bullet. It can be quite troublesome when we have a weak first stage. Also, although consistent, 2SLS is still a biased method to estimate the causal effect. \n",
+ "We then delve into the mechanics of estimating the causal effect with IV, namely, using 2SLS. We've also learned that IV is no silver bullet. It can be quite troublesome when we have a weak first stage: even with valid instruments, 2SLS can be unstable and substantially biased in finite samples. \n",
"\n",
"## References\n",
"\n",
"I like to think of this entire book as a tribute to Joshua Angrist, Alberto Abadie and Christopher Walters for their amazing Econometrics class. Most of the ideas here are taken from their classes at the American Economic Association. Watching them is what is keeping me sane during this tough year of 2020.\n",
"* [Cross-Section Econometrics](https://www.aeaweb.org/conference/cont-ed/2017-webcasts)\n",
"* [Mastering Mostly Harmless Econometrics](https://www.aeaweb.org/conference/cont-ed/2020-webcasts)\n",
+ "* [Instrumental Variables Regression with Weak Instruments](https://www.nber.org/papers/t0151)\n",
"\n",
"I'll also like to reference the amazing books from Angrist. They have shown me that Econometrics, or 'Metrics as they call it, is not only extremely useful but also profoundly fun.\n",
"\n",