[Script Info] Title: [Events] Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text Dialogue: 0,0:00:00.40,0:00:05.70,Default,,0000,0000,0000,,The path from cause to effect\Nis dark and dangerous, Dialogue: 0,0:00:05.70,0:00:08.90,Default,,0000,0000,0000,,but the weapons\Nof econometrics are strong, Dialogue: 0,0:00:09.70,0:00:13.70,Default,,0000,0000,0000,,wield differences-in-differences\Nwhen witnessing parallel trends. Dialogue: 0,0:00:14.45,0:00:17.00,Default,,0000,0000,0000,,♪ [music] ♪ Dialogue: 0,0:00:20.10,0:00:21.42,Default,,0000,0000,0000,,Masters of metrics Dialogue: 0,0:00:21.42,0:00:24.80,Default,,0000,0000,0000,,look for convincing\Nceteris paribus comparisons. Dialogue: 0,0:00:25.10,0:00:29.42,Default,,0000,0000,0000,,The ideal comparison contrasts\Ntreatment and control groups Dialogue: 0,0:00:29.42,0:00:30.60,Default,,0000,0000,0000,,that look similar. Dialogue: 0,0:00:30.60,0:00:34.70,Default,,0000,0000,0000,,But sometimes this sort\Nof comparability is elusive. Dialogue: 0,0:00:34.70,0:00:36.80,Default,,0000,0000,0000,,When treatment and control groups Dialogue: 0,0:00:36.80,0:00:40.10,Default,,0000,0000,0000,,evolve similarly\Nin the absence of treatment, Dialogue: 0,0:00:40.10,0:00:42.31,Default,,0000,0000,0000,,even if from different\Nstarting points, Dialogue: 0,0:00:42.31,0:00:44.90,Default,,0000,0000,0000,,there's hope for causal inference. Dialogue: 0,0:00:45.60,0:00:48.40,Default,,0000,0000,0000,,The weapon that exploits\Nparallel evolution, Dialogue: 0,0:00:48.66,0:00:50.89,Default,,0000,0000,0000,,masters say parallel trends, Dialogue: 0,0:00:50.89,0:00:53.36,Default,,0000,0000,0000,,is called differences-in-differences... Dialogue: 0,0:00:53.36,0:00:54.30,Default,,0000,0000,0000,,(voice whispering)\NDifferences-in-differences Dialogue: 0,0:00:54.40,0:00:56.90,Default,,0000,0000,0000,,- ...or DD for short.\N- Alright. Nice. Dialogue: 0,0:00:56.90,0:00:59.99,Default,,0000,0000,0000,,Let's see how DD\Ncan help us understand Dialogue: 0,0:00:59.99,0:01:03.06,Default,,0000,0000,0000,,one of the most important\Neconomic events Dialogue: 0,0:01:03.06,0:01:04.37,Default,,0000,0000,0000,,in US history. Dialogue: 0,0:01:05.30,0:01:08.30,Default,,0000,0000,0000,,Look back with me now\Nat the Great Depression-- Dialogue: 0,0:01:08.80,0:01:12.20,Default,,0000,0000,0000,,the worst economic catastrophe,\Nour country has ever known. Dialogue: 0,0:01:13.10,0:01:16.20,Default,,0000,0000,0000,,Unemployment hit 25% in 1933-- Dialogue: 0,0:01:16.60,0:01:19.51,Default,,0000,0000,0000,,a level not seen before or since. Dialogue: 0,0:01:19.51,0:01:22.10,Default,,0000,0000,0000,,Millions lost their homes\Nor their land. Dialogue: 0,0:01:22.60,0:01:24.74,Default,,0000,0000,0000,,Suicide spiked, and hungry families Dialogue: 0,0:01:24.74,0:01:26.77,Default,,0000,0000,0000,,relied on soup kitchens\Nand bread lines Dialogue: 0,0:01:26.77,0:01:28.16,Default,,0000,0000,0000,,to keep from starving. Dialogue: 0,0:01:29.40,0:01:34.00,Default,,0000,0000,0000,,Economists argue fiercely over\Nthe causes of the Great Depression. Dialogue: 0,0:01:34.00,0:01:36.98,Default,,0000,0000,0000,,Most agree, however,\Nthat a key piece of the puzzle Dialogue: 0,0:01:36.98,0:01:39.62,Default,,0000,0000,0000,,is an epidemic of bank failures. Dialogue: 0,0:01:39.80,0:01:41.90,Default,,0000,0000,0000,,This was before deposit insurance. Dialogue: 0,0:01:42.10,0:01:46.49,Default,,0000,0000,0000,,So if your bank went bankrupt,\Nyour savings disappeared with it, Dialogue: 0,0:01:52.60,0:01:56.10,Default,,0000,0000,0000,,Faced with a banking crisis,\Nthe central bank has a choice: Dialogue: 0,0:01:56.40,0:01:58.52,Default,,0000,0000,0000,,lend freely to troubled banks Dialogue: 0,0:01:58.52,0:02:01.10,Default,,0000,0000,0000,,or stand aside and refuse to lend. Dialogue: 0,0:02:01.50,0:02:05.44,Default,,0000,0000,0000,,Lending freely to banks in trouble\Nis called easy money. Dialogue: 0,0:02:05.44,0:02:08.10,Default,,0000,0000,0000,,Refusing to lend is called tight money. Dialogue: 0,0:02:10.20,0:02:12.87,Default,,0000,0000,0000,,Monetarist masters Milton Friedman\Nand Anna Schwartz Dialogue: 0,0:02:12.87,0:02:14.61,Default,,0000,0000,0000,,famously called\Nthe Great Depression Dialogue: 0,0:02:14.61,0:02:16.35,Default,,0000,0000,0000,,the "Great Contraction," Dialogue: 0,0:02:16.80,0:02:18.37,Default,,0000,0000,0000,,accusing the Federal Reserve Dialogue: 0,0:02:18.37,0:02:21.20,Default,,0000,0000,0000,,of inflicting a misguided policy\Nof tight money Dialogue: 0,0:02:21.20,0:02:24.00,Default,,0000,0000,0000,,on the nation's teetering\Nfinancial institutions. Dialogue: 0,0:02:24.40,0:02:25.87,Default,,0000,0000,0000,,They argued that easy money Dialogue: 0,0:02:25.87,0:02:27.98,Default,,0000,0000,0000,,would have kept\Nmany banks in business, Dialogue: 0,0:02:27.98,0:02:29.70,Default,,0000,0000,0000,,shortening the Great Depression, Dialogue: 0,0:02:30.40,0:02:32.24,Default,,0000,0000,0000,,But others disagree! Dialogue: 0,0:02:32.24,0:02:33.77,Default,,0000,0000,0000,,If banks are insolvent Dialogue: 0,0:02:33.77,0:02:35.95,Default,,0000,0000,0000,,because of unwise\Nlending decisions, Dialogue: 0,0:02:35.95,0:02:38.90,Default,,0000,0000,0000,,then bailouts just encourage\Nmore foolishness. Dialogue: 0,0:02:39.60,0:02:41.30,Default,,0000,0000,0000,,Economists call this problem, Dialogue: 0,0:02:41.50,0:02:46.10,Default,,0000,0000,0000,,moral hazard the debate over bailouts\Nin. Moral hazard continues. Today Dialogue: 0,0:02:46.50,0:02:47.90,Default,,0000,0000,0000,,should Financial Behemoth. Dialogue: 0,0:02:47.90,0:02:51.50,Default,,0000,0000,0000,,Lehman Brothers, have been allowed to\Nfail on the eve of the Great Recession Dialogue: 0,0:02:52.00,0:02:56.60,Default,,0000,0000,0000,,in an Ideal World. We'd answer this\Nquestion by applying different fed policies Dialogue: 0,0:02:56.60,0:02:58.40,Default,,0000,0000,0000,,to randomly selected regions, Dialogue: 0,0:02:59.00,0:03:01.50,Default,,0000,0000,0000,,but we can still learn a\Nlot by using differences Dialogue: 0,0:03:01.50,0:03:06.30,Default,,0000,0000,0000,,and differences to compare Trends across\Nareas with different monetary policies. Dialogue: 0,0:03:10.90,0:03:15.70,Default,,0000,0000,0000,,How's that even possible? Don't the same\Nfed policies. Apply to all banks in the US? Dialogue: 0,0:03:15.80,0:03:17.40,Default,,0000,0000,0000,,Yeah, good question. Dialogue: 0,0:03:17.70,0:03:22.50,Default,,0000,0000,0000,,The Federal Reserve System is divided\Ninto twelve districts each headed by Dialogue: 0,0:03:22.50,0:03:27.30,Default,,0000,0000,0000,,a regional bank. Today fed policy\Nis set at the national level. Dialogue: 0,0:03:27.40,0:03:32.20,Default,,0000,0000,0000,,But in the 1930s Regional, feds\Ncould do pretty much as they liked, Dialogue: 0,0:03:33.20,0:03:35.40,Default,,0000,0000,0000,,and here's what's so awesome about that Dialogue: 0,0:03:35.50,0:03:40.60,Default,,0000,0000,0000,,in 1930. The Atlanta fed running the\N6th District, followed an easy money. Dialogue: 0,0:03:40.70,0:03:45.40,Default,,0000,0000,0000,,Policy sending wheelbarrows of cash\Nto rescue insolvent institutions, Dialogue: 0,0:03:45.90,0:03:50.60,Default,,0000,0000,0000,,the st. Louis fed running the 8th\NDistrict, followed a Tight money policy. Dialogue: 0,0:03:51.10,0:03:53.90,Default,,0000,0000,0000,,Let Veil The foolish\Nthey said in st. Louis Dialogue: 0,0:03:54.30,0:03:59.50,Default,,0000,0000,0000,,and so a natural experiment in\Nmonetary policy was born even better. Dialogue: 0,0:03:59.60,0:04:01.90,Default,,0000,0000,0000,,This is a within State experiment, Dialogue: 0,0:04:02.00,0:04:05.20,Default,,0000,0000,0000,,the border between the sixth and\Nthe eighth districts run smack Dialogue: 0,0:04:05.20,0:04:06.70,Default,,0000,0000,0000,,through the middle of Mississippi. Dialogue: 0,0:04:07.30,0:04:10.50,Default,,0000,0000,0000,,So northern, Mississippi had tied\Nmoney while Southern Mississippi. Dialogue: 0,0:04:10.70,0:04:15.20,Default,,0000,0000,0000,,Had easy money, but under the same state\Nlaws in banking regulations, in both. Dialogue: 0,0:04:15.80,0:04:19.70,Default,,0000,0000,0000,,The treatment group is the\Ndistrict. 6 part of Mississippi, Dialogue: 0,0:04:20.10,0:04:23.10,Default,,0000,0000,0000,,which had access to Easy\NMoney during the crisis. Dialogue: 0,0:04:23.80,0:04:27.60,Default,,0000,0000,0000,,The control group is the\Ndistrict 8 part of Mississippi, Dialogue: 0,0:04:27.80,0:04:30.10,Default,,0000,0000,0000,,which had Tight money during the crisis. Dialogue: 0,0:04:31.30,0:04:35.40,Default,,0000,0000,0000,,The key year in our natural\Nexperiment was 1930, Dialogue: 0,0:04:35.80,0:04:41.90,Default,,0000,0000,0000,,Caldwell and Company a massive Financial\NEmpire. In the South Came Crashing. Down. Dialogue: 0,0:04:42.70,0:04:46.00,Default,,0000,0000,0000,,Banking is a business built\Non confidence and Trust Dialogue: 0,0:04:46.50,0:04:52.90,Default,,0000,0000,0000,,the Caldwell meltdown caused a panic that\Nled to a widespread Bank Run all at once. Dialogue: 0,0:04:53.00,0:04:56.90,Default,,0000,0000,0000,,Depositors wanted their money\Nback, causing Banks to go bankrupt Dialogue: 0,0:04:57.00,0:04:58.20,Default,,0000,0000,0000,,and shut their doors. Dialogue: 0,0:05:01.00,0:05:04.40,Default,,0000,0000,0000,,We'll use differences in\Ndifferences to measure the effect Dialogue: 0,0:05:04.40,0:05:09.10,Default,,0000,0000,0000,,of contrasting monetary policies\Nin response to the Caldwell crisis. Dialogue: 0,0:05:12.50,0:05:16.30,Default,,0000,0000,0000,,This figure plots. The number\Nof banks in Mississippi by year Dialogue: 0,0:05:16.40,0:05:18.80,Default,,0000,0000,0000,,for the eighth and sixth districts. Dialogue: 0,0:05:19.20,0:05:24.10,Default,,0000,0000,0000,,Let's start in 1929 a year\Nbefore the Caldwell crash. Dialogue: 0,0:05:24.20,0:05:30.40,Default,,0000,0000,0000,,There are 169 banks open in the\Neighth and 141 banks open in the Dialogue: 0,0:05:31.30,0:05:32.50,Default,,0000,0000,0000,,Over the next year, Dialogue: 0,0:05:32.70,0:05:37.00,Default,,0000,0000,0000,,we see a similar handful of\Nbanks fail. In both districts, Dialogue: 0,0:05:37.40,0:05:41.70,Default,,0000,0000,0000,,the change in the number of banks\Nin operation. Is remarkably similar. Dialogue: 0,0:05:42.10,0:05:44.60,Default,,0000,0000,0000,,That's what parallel Trends look like Dialogue: 0,0:05:45.50,0:05:48.80,Default,,0000,0000,0000,,in November. 1930, Caldwell, crashes. Dialogue: 0,0:05:48.90,0:05:50.40,Default,,0000,0000,0000,,And the Panic begins Dialogue: 0,0:05:51.20,0:05:55.30,Default,,0000,0000,0000,,Banks. Fail frequently in the 8th\NDistrict, which had Tight money, Dialogue: 0,0:05:55.80,0:06:00.10,Default,,0000,0000,0000,,but the decline is slower in the\N6th District which had easy money. Dialogue: 0,0:06:00.80,0:06:03.00,Default,,0000,0000,0000,,The diverging Trends in this period, Dialogue: 0,0:06:03.00,0:06:06.30,Default,,0000,0000,0000,,might be attributable to\Neasy versus Tight money Dialogue: 0,0:06:06.60,0:06:11.80,Default,,0000,0000,0000,,in July, of 1931, the 8th\NDistrict, abandons type money. Dialogue: 0,0:06:11.90,0:06:14.20,Default,,0000,0000,0000,,So now both districts are easy. Dialogue: 0,0:06:14.70,0:06:16.90,Default,,0000,0000,0000,,Parallel Trends are restored Dialogue: 0,0:06:17.30,0:06:22.20,Default,,0000,0000,0000,,in a counterfactual world where the sixth\Ndistrict follows a Tight money policy. Dialogue: 0,0:06:22.50,0:06:23.80,Default,,0000,0000,0000,,What might have happened Dialogue: 0,0:06:24.30,0:06:28.40,Default,,0000,0000,0000,,if we extrapolate the trend of\Nthe 8th District to the sixth, Dialogue: 0,0:06:28.50,0:06:29.90,Default,,0000,0000,0000,,it would look like this. Dialogue: 0,0:06:30.40,0:06:30.70,Default,,0000,0000,0000,,So Dialogue: 0,0:06:30.80,0:06:35.90,Default,,0000,0000,0000,,The treatment effective easy money is\Nhow much the 6th District deviated From Dialogue: 0,0:06:35.90,0:06:38.90,Default,,0000,0000,0000,,the Path implied by\Nthe 8th District trend. Dialogue: 0,0:06:41.10,0:06:44.00,Default,,0000,0000,0000,,How many banks did the\NEasy Money treatment save Dialogue: 0,0:06:44.00,0:06:49.80,Default,,0000,0000,0000,,this table reports data for the treatment\Ngroup District Six in the first row Dialogue: 0,0:06:49.80,0:06:54.00,Default,,0000,0000,0000,,and data for the control group\NDistrict 8 in the second row. Dialogue: 0,0:06:54.00,0:06:57.40,Default,,0000,0000,0000,,The First Column shows the\Nnumber of banks in business, Dialogue: 0,0:06:57.40,0:07:03.40,Default,,0000,0000,0000,,before The Crisis began in 1930.\NThe second column shows 1931. Dialogue: 0,0:07:03.40,0:07:08.90,Default,,0000,0000,0000,,This is the key period when each district\Nhad differing monetary policies during Dialogue: 0,0:07:08.90,0:07:10.90,Default,,0000,0000,0000,,the crisis the rightmost. Dialogue: 0,0:07:11.20,0:07:11.60,Default,,0000,0000,0000,,Column Dialogue: 0,0:07:11.70,0:07:18.00,Default,,0000,0000,0000,,reports changes within the district\Ndistrict, 6 lost 14 Banks while District Dialogue: 0,0:07:18.10,0:07:23.80,Default,,0000,0000,0000,,8 lost 33, the mathematical formula\Nfor the treatment effect is simple. Dialogue: 0,0:07:24.40,0:07:28.70,Default,,0000,0000,0000,,We subtract the change in banks\Nin operation, in the 8th District Dialogue: 0,0:07:28.90,0:07:32.20,Default,,0000,0000,0000,,from the change in banks\Nin operation in the sixth. Dialogue: 0,0:07:32.80,0:07:35.70,Default,,0000,0000,0000,,Hence. The name\Ndifferences in differences, Dialogue: 0,0:07:37.00,0:07:40.90,Default,,0000,0000,0000,,negative 14, minus negative, 33 equals 9. Dialogue: 0,0:07:41.10,0:07:41.60,Default,,0000,0000,0000,,19. Dialogue: 0,0:07:42.50,0:07:46.80,Default,,0000,0000,0000,,We estimate that 19 Banks\Nwere saved by easy money Dialogue: 0,0:07:47.30,0:07:51.70,Default,,0000,0000,0000,,in practice tables and figures like\Nthose shown here are the beginning Dialogue: 0,0:07:51.80,0:07:54.30,Default,,0000,0000,0000,,rather than the end of a DD analysis, Dialogue: 0,0:07:55.40,0:08:00.00,Default,,0000,0000,0000,,the problem of how to gauge the\Nstatistical, significance of DD, estimates Dialogue: 0,0:08:00.10,0:08:02.30,Default,,0000,0000,0000,,turns out to be exceedingly, tricky Dialogue: 0,0:08:02.40,0:08:05.60,Default,,0000,0000,0000,,and a regression is typically\Npart of the solution, Dialogue: 0,0:08:09.20,0:08:10.70,Default,,0000,0000,0000,,the key assumption behind a Dialogue: 0,0:08:11.10,0:08:17.70,Default,,0000,0000,0000,,DD analysis is that of parallel Trends\Nrecall, the principle of ceteris paribus, Dialogue: 0,0:08:17.90,0:08:21.70,Default,,0000,0000,0000,,our ideal comparison would have\Nthe two districts experienced, Dialogue: 0,0:08:21.70,0:08:27.80,Default,,0000,0000,0000,,an identical business environment, except\Nfor one factor easy or tight money. Dialogue: 0,0:08:29.20,0:08:33.50,Default,,0000,0000,0000,,Both districts would have identical\Ntypes of customers who would go bankrupt Dialogue: 0,0:08:33.50,0:08:35.30,Default,,0000,0000,0000,,at exactly the same rate. Dialogue: 0,0:08:35.70,0:08:38.60,Default,,0000,0000,0000,,The skill of their employees\Nwould be equal and so on Dialogue: 0,0:08:39.20,0:08:43.40,Default,,0000,0000,0000,,perfect ceteris, paribus comparisons\Nwould allow us to clearly see Dialogue: 0,0:08:43.60,0:08:48.70,Default,,0000,0000,0000,,the causal effect of different fed\Npolicies in this case, that's not possible. Dialogue: 0,0:08:49.10,0:08:53.60,Default,,0000,0000,0000,,But the idea of parallel Trends\Nis based on a similar concept. Dialogue: 0,0:08:53.60,0:08:57.90,Default,,0000,0000,0000,,If we see that the two regions\Nexperienced similar Trends in the number Dialogue: 0,0:08:57.90,0:08:58.80,Default,,0000,0000,0000,,of banks over. Dialogue: 0,0:08:59.00,0:09:04.00,Default,,0000,0000,0000,,Time in the absence of treatment, we\Ncan assume they are good comparisons. Dialogue: 0,0:09:04.40,0:09:09.80,Default,,0000,0000,0000,,We see that the two districts move in\Nparallel both before the crisis. And after, Dialogue: 0,0:09:10.10,0:09:12.40,Default,,0000,0000,0000,,when they have the same fed policy, Dialogue: 0,0:09:13.10,0:09:18.30,Default,,0000,0000,0000,,the only time the district's behave differently\Nis when the FED policy is different. Dialogue: 0,0:09:19.40,0:09:26.70,Default,,0000,0000,0000,,In view of this fed policy is a likely\Ncause of diverging Trends from 1930 to 1931. Dialogue: 0,0:09:27.80,0:09:31.80,Default,,0000,0000,0000,,But we should also check for other\Nchanges unique to northern, Mississippi. Dialogue: 0,0:09:32.30,0:09:33.20,Default,,0000,0000,0000,,What do you mean? Dialogue: 0,0:09:33.50,0:09:39.10,Default,,0000,0000,0000,,Imagine that bad tornadoes? Hit Northern,\Nbut not Southern. Mississippi in 1930, Dialogue: 0,0:09:39.60,0:09:44.30,Default,,0000,0000,0000,,these tornadoes devastate Farms\Ncausing Farmers to default on loans, Dialogue: 0,0:09:44.60,0:09:46.80,Default,,0000,0000,0000,,which drives their Banks out of business. Dialogue: 0,0:09:47.40,0:09:52.20,Default,,0000,0000,0000,,Then the sixth and eighth districts\Nwould differ in not one, but two ways Dialogue: 0,0:09:52.70,0:09:57.20,Default,,0000,0000,0000,,fed policy, and whether and\Nwe'd have trouble identifying Dialogue: 0,0:09:57.60,0:10:02.50,Default,,0000,0000,0000,,Policy as the causal Factor behind\Nincreased bank failures in the eighth. Dialogue: 0,0:10:07.20,0:10:10.20,Default,,0000,0000,0000,,DD credibility lives or\Ndies with the claim that the Dialogue: 0,0:10:10.20,0:10:13.60,Default,,0000,0000,0000,,only reason northern\NMississippi was special in 1930 Dialogue: 0,0:10:13.90,0:10:16.00,Default,,0000,0000,0000,,is differing. Regional fed policy. Dialogue: 0,0:10:16.60,0:10:22.10,Default,,0000,0000,0000,,We're in DD heaven with strong Visual\NEvidence of parallel Trend in general. Dialogue: 0,0:10:22.20,0:10:24.20,Default,,0000,0000,0000,,The first step in evaluating whether Dialogue: 0,0:10:24.20,0:10:30.20,Default,,0000,0000,0000,,to use DD is usually this type of\Nvisual confirmation of parallel Trends Dialogue: 0,0:10:30.20,0:10:31.70,Default,,0000,0000,0000,,outside of the period. Dialogue: 0,0:10:31.70,0:10:36.90,Default,,0000,0000,0000,,When we expect to see a treatment\Neffect, the treatment in our example, Dialogue: 0,0:10:37.10,0:10:39.70,Default,,0000,0000,0000,,Easy money in the face of bank failures Dialogue: 0,0:10:40.50,0:10:45.00,Default,,0000,0000,0000,,metrics Masters, use DD to\Nexplore effects of many policies Dialogue: 0,0:10:45.80,0:10:47.90,Default,,0000,0000,0000,,like the minimum, legal drinking age, Dialogue: 0,0:10:48.50,0:10:52.20,Default,,0000,0000,0000,,and environmental changes\Nlike access to clean water. Dialogue: 0,0:10:52.90,0:10:54.20,Default,,0000,0000,0000,,In our next video. Dialogue: 0,0:10:54.50,0:10:59.20,Default,,0000,0000,0000,,We'll see an example of how regression\Nis used to implement a DD approach. Dialogue: 0,0:11:00.80,0:11:06.90,Default,,0000,0000,0000,,Are you a teacher click to explore ways to\Nuse these videos in class if your learner Dialogue: 0,0:11:07.00,0:11:11.20,Default,,0000,0000,0000,,Make sure this video sticks by taking\Na few quick practice questions, Dialogue: 0,0:11:11.60,0:11:14.20,Default,,0000,0000,0000,,or if you're ready.\NClick for the next video. Dialogue: 0,0:11:14.60,0:11:20.00,Default,,0000,0000,0000,,You can also check out Mr. Use website for\Nmore courses, teacher resources and more.