1. TRYGVE HAAVELMO AND THE EMERGENCE OF CAUSAL CALCULUS Judea Pearl University of California Los Angeles (www.cs.ucla.edu/~judea)

2. Haavelmo’s major breakthroughs The evolution of causal calculus What causal calculus can do for us? Will Haavelmo’s legacy survive? OUTLINE

3. 1.An economic model is a set of hypothetical experiments, qualitatively encoded in a system of equations. 2.An economic model is capable of answering policy intervention questions, with no further assistance from the modeller. 3.There is a formal way of taking an arbitrary model, combining it with data, and derive valid answers to policy questions. HAAVELMO’S THREE INSIGHTS

4. “We have in mind some actual experiment, or some design of an experiment, which we could at least imagine arranging." (Haavelmo, 1944) His example: What information did the modeller intend a to carry in Eq. 1 and what information would a provide if we are able to estimate its value. WHAT IS AN ECONOMIC MODEL? Perhaps a = g[P(x,y,z...)] ? There is no such g [ ].

5. TRADITIONAL STATISTICAL INFERENCE PARADIGM Data Inference Q(P) (Aspects of P ) P Joint Distribution e.g., Infer whether customers who bought product A would also buy product B. Q = P(B | A)

6. Data Inference Q(M) (Aspects of M ) Data Generating Model M – Invariant strategy (mechanism, recipe, law, protocol) by which Nature assigns values to variables in the analysis. Joint Distribution THE STRUCTURAL MODEL PARADIGM M “A painful de-crowning of a beloved oracle!”

7. “I am speaking, of course, about the equation: {y = a + bx +  }. What does it mean? The only meaning I have ever determined for such an equation is that it is a shorthand way of describing the conditional distribution of {y} given {x}.” (Holland, 1995, p. 54) CLINGING DESPARATELY TO AN OLD ORACLE Fifty-two years later (1995): “Beneath all multiple equation models, there is only a set of conditional distributions.” (Richard Berk, 2004, p. 196)

8. “For example, in the model: until the relationship between z t and u t is specified, the meaning of  is uncertain since E [ z t u t ] could be either zero or non-zero on the information provided.” (Hendry, 1995) ECONOMISTS CLINGING TO THE OLD ORACLE Fifty-two years later (1995): “The joint density is the basis: SEMs are merely an interpretation of that.” (1998, personal communication)

9. “A state implements tough new penalties on drunk drivers: What is the effect on highway fatalities?... [This effect] is an unknown characteristic of the population joint distribution of X and Y.” (Stock and Watson, 2007, Introduction to Econometrics, Chapter 4, pp. 111) ECONOMISTS CLINGING TO THE OLD ORACLE Sixty-four years later (2007): Sixty-eight years later (2011): “An econometric model specifies the statistical relationship that is believed to hold between the various economic quantities pertaining to particular economic phenomenal under study.” (Wikipedia – “Econometric Models,” 2011)

10. “to predict consumption,... under the Government policy,... we may use the `theoretical’ equations obtained.” (1943, p.12) WHAT CAN OUR NEW ORACLE DO? Haavelmo: Provide policy advice. Haavelmo: “this is only natural, because now the Government is, in fact, performing `experiments' of the type we had in mind when constructing each of the two equations.'' (1943, p.12) “How can anyone predict outcomes of experiments that where never performed?'' Skeptics:

11. “Assume that the Government decides, through public spending, taxation, etc., to keep income, r t, at a given level,... the only change in the system being that, instead of we now have where g i is Government expenditure, so adjusted as to keep r constant, whatever be u and v,...'' (1943, p. 12) HOW CAN OUR NEW ORACLE DO ALL THAT? (THE ALGORITHMIZATION OF INTERVENTION) Haavelmo “surgery”: Modify an equation by adding an adjustable force g i, while keeping all other equations in tact.

12. Haavelmo's surgery (1943) Add adjustable force Strotz and Wold surgery (1960). “Wipe out” the equation and replace it with Graphical surgery (Spirtes et al., 1993; Pearl, 1993). Wipe out incoming arrows to r do -calculus (Pearl, 1994) Structural counterfactuals (Balke and Pearl, 1995) Y r (u) = Y(u) in the model Unification with Neyman-Rubin Y x ( u ) and Lewis (1973) THE EVOLUTION OF CAUSAL CALCULUS u v ry

13. In particular: ALGORITHMIZATION OF COUNTERFACTUALS Semantics: The counterfactual Y x (u) = y, which reads “ Y would be y (in situation u ), had X been x ” is said to be true, if Y = y in the x -mutilated model, under exogenous conditions U = u. Joint probabilities of counterfactuals:

14. AXIOMS OF CAUSAL COUNTERFACTUALS 1.Definiteness 2.Uniqueness 3.Effectiveness 4.Composition 5.Reversibility Y would be y, had X been x (in state U = u )

15. CAUSAL MODEL (M A ) A - CAUSAL ASSUMPTIONS Q Queries of interest Q(P) - Identified estimands Causal inference T(M A ) - Testable implications Statistical inference Goodness of fit Model testingProvisional claims A* - Logical implications of A Q - Estimates of Q(P) CAUSAL MODEL (M A ) Data ( D ) THE LOGIC OF CAUSAL ANALYSIS

16. Model 2 (Linear version)Model 1 ILLUSTRATING THE WORKING OF CAUSAL CALCULUS W1W1 Z2Z2 W2W2 Z1Z1 X Z3Z3 W3W3 Y U ’ s are mutually independent

17. Model 1 WHAT ARE THE TESTABLE IMPLICATIONS OF THE MODEL? W1W1 Z2Z2 W2W2 Z1Z1 X Z3Z3 W3W3 Y Missing edges: Z 1 – Z 2, Z 1 – Y, Z 2 – X... Separating sets: {0}, {X, Z 2, Z 3 }, {Z 1, Z 3 }... Zero regression coefficients: r {Z 1 Z 2 } = 0 and r {YZ 1 XZ 2 Z 3 } = 0 These imply *all* misspecification tests Testable implications (FINITE!): Z 1 Z 2 Z 1 Y | {X 1,Z 2,Z 3 }, Z 2 X | {Z 1, Z 3 }.

18. IMPLIED MISSPECIFICATION TESTS W1W1 Z2Z2 W2W2 Z1Z1 X Z3Z3 W3W3 Y Question 4: If we regress Z 1 on all other variables in the model, which regression coefficient will be zero? Answer: All but these three Question 5: If we regress Z 1 on Z 3 and W 1, which regression coefficient will change if we add Y as a regressor? Answer: The coefficient of Z 3 will change and the coefficient of W 1 will remain invariant. Non-invariance may not be misspecification.

19. CAUSAL CALCULUS AS AN ORACLE FOR INTERVENTIONS W1W1 Z2Z2 W2W2 Z1Z1 X Z3Z3 W3W3 Y Suppose we wish to estimate the average causal effect of X on Y Which subsets of variables need to be adjusted to obtain an unbiased estimate of ACE ? Is there a single variable that, if measured, would allow an unbiased estimate of ACE ?

20. Front Door EFFECT OF WARM-UP ON INJURY (After Shrier & Platt, 2008) No, no! Watch out! Warm-up Exercises ( X ) Injury ( Y ) ???

21. THE MACHINERY OF CAUSAL CALCULUS Rule 1: Ignoring observations P(y | do{x}, z, w) = P(y | do{x}, w) Rule 2: Action/observation exchange P(y | do{x}, do{z}, w) = P(y | do{x},z,w) Rule 3: Ignoring actions P(y | do{x}, do{z}, w) = P(y | do{x}, w)

22. DERIVATION IN CAUSAL CALCULUS Smoking Tar Cancer P (c | do{s}) =  t P (c | do{s}, t) P (t | do{s}) =  s   t P (c | do{t}, s) P (s | do{t}) P(t |s) =  t P (c | do{s}, do{t}) P (t | do{s}) =  t P (c | do{s}, do{t}) P (t | s) =  t P (c | do{t}) P (t | s) =  s  t P (c | t, s) P (s) P(t |s) =  s   t P (c | t, s) P (s | do{t}) P(t |s) Probability Axioms Rule 2 Rule 3 Rule 2 Genotype (Unobserved)

23. WHAT ELSE CAN CAUSAL CALCULUS DO FO US? W1W1 Z2Z2 W2W2 Z1Z1 X Z3Z3 W3W3 Y Equivalent models Identifying Counterfactual queries (ETT, PC) Mediation Causes of Effects External validity Finding instruments Is there an instrumental variable for the Z 3  Y relationship? Answer: No Can we turn Z 1 into an IV? Answer: Yes, condition on W 1.

24. WILL HAAVELMO'S LEGACY SURVIVE? Two attempted hijackings 1.The regressionism assault (1975 - 2000) Economic model is a parameterized family of joint distributions.

25. THE CAUSAL RENAISSANCE: VOCABULARY IN ECONOMICS From Hoover (2004) “Lost Causes”

26. WILL HAAVELMO'S LEGACY SURVIVE? Two attempted hijackings 1.The regressionism assault (1975 - 2000) Economic model is a parameterized family of joint distributions. 2.The “instrumentalism” assault (1995 - ) Economic analysis is a potential clinical trial with missing data (potential outcomes) and one treatment. Instruments are needed, no structure. 3.The structuralist defense Finding an instrument requires structural assumptions. The "experimentalists" choice of language deprives them from making those assumptions transparent.

27. FORMULATING ASSUMPTIONS THREE LANGUAGES 2. Counterfactuals: 1. English: Smoking ( X ), Cancer ( Y ), Tar ( Z ), Genotypes ( U ) X Y Z U ZXY 3. Structural:

28. THE STRUCTURAL-COUNTERFACTUAL SYMBIOSIS 1.Express theoretical assumptions in structural language. 2.Express queries in counterfactual language. 3.Translate (1) into (2) for algebraic analysis, Or (2) into (1) for graphical analysis. 4.Use either graphical or algebraic machinery to answer the query in (2).

29. Formal basis for causal and counterfactual inference (complete) Unification of the graphical, potential-outcome and structural equation approaches Friendly and formal solutions to century-old problems and confusions. CONCLUSIONS

30. He is wise who bases causal inference on an explicit causal structure that is defensible on scientific grounds. After Haavelmo, 1943 CONCLUSIONS

31. Thank you