Econ 440: Lecture 8 PDF

Econ 440: Lecture 8 Prof. Ragan Petrie Texas A&M University September 19, 2024 Decline in drug overdose deaths Source: NPR, Sept 18, 2024 Good news, but why? ▶ Possible explanations ▶ Demand: Decline in use of fentanyl, methamphetamines, xylazine and other synthetic chemica...

Econ 440: Lecture 8 Prof. Ragan Petrie Texas A&M University September 19, 2024 Decline in drug overdose deaths Source: NPR, Sept 18, 2024 Good news, but why? ▶ Possible explanations ▶ Demand: Decline in use of fentanyl, methamphetamines, xylazine and other synthetic chemicals ▶ Supply: law enforcement efforts at border; gangs mixing xylazine with fentanyl which delays onset of withdrawal ▶ Use adaptation: drug users more likely to carry overdose-reversal medication (naloxone) ▶ Decline in drug user population: most have died off already ▶ Important public health question and limited budget to do something ▶ Confounding explanations, may not be causal ▶ How could you use experimental methods to learn what might be most cost-effective way to decrease drug overdose deaths? Time Preferences Time Preferences  Time Preferences The Standard Model: Exponential Discounting The Standard Model: Exponential Discounting  Time Preferences The Standard Model: Exponential Discounting Motivation ▶ Which would you rather have? ▶ $100 today OR $90 one year from now ▶ $100 today OR $100 one year from now ▶ $100 today OR $110 one year from now ▶ $100 today OR $150 one year from now ▶ If you value money today more than the same amount in the future, then we say you are impatient  Time Preferences The Standard Model: Exponential Discounting Consumption Over Time ▶ Consider a stream of consumption (or wealth or income) over T time periods, starting with period 1: c = (c1 , c2 , c3 ,... , cT ) ▶ ct is the consumption in period t ▶ Example, T = 3 periods: (c1 , c2 , c3 ) = ($5, $10, $0) ▶ Utility is a function of the entire stream of income: U(c) = f (c1 , c2 , c3 ,... , cT ) ▶ If impatient, then would prefer to have an extra dollar today rather than tomorrow, implying: ∂U ∂U > ∂ct ∂ct+1 or more generally: ∂U ∂ct+1 ∂U δu(2) ▶ Treatment 2: choose early option if δ 4 u(1) > δ 5 u(2) ▶ Reduces to u(1) > δu(2), same as Treatment 1 ▶ Thus we expect same percentage subjects choosing early option in both treatments ▶ What actually happened? ▶ Treatment 1 (immediate): 60% choose early option ▶ Treatment 2 (delay): 30% choose early option