Akee is the Chair of American Indian studies program at UCLA. The other authors are mostly academics at Duke in various health departments. Authors study the effect of an exogenous increase in household income due to government transfer independent of household characteristics on long-term outcomes of children in those households. Study finds increases in educational attainment and lower likelihood to commit crime as a teenager.
Researchers use data from the Great Smoky Mountains Study of Youth, a longitudinal study of child mental health in rural North Carolina. Both American Indian and non-Indian children were sampled. Halfway through the survey, a casino opened on the Eastern Cherokee reservation, a portion of profits of which were distributed every 6 months on a per-capita basis to all adult tribal members regardless of employment, income, or other household characteristics.
Poorer treated households do not use the additional income to substitute away from full-time employment, so it not that they are spending more time with their children. Children and parents of treated households report higher quality of parent-child interaction.
Some questions I have: study sees significant decreases in likelihood to commit crime between ages of 16-17, but not for ages 18-19, nor ages 20-21. Also, they see no change in having ever commited a moderate/violent crime by age 21. Does this suggest the 16-17 effect is noise? Do we care if people are less likely to commit a crime between 16-17 if they're just as likely to commit a crime between 18-19?
Related numbers: Today we spend roughly $12,000 per pupil per year (Census). In North Carolina, that number is $8,800. (Michigan is $12,000, and New York is $20,000!) So does that mean we're increasing expenditure by 50% for 10 years to get 1 additional year of educational attainment? Presumably there are other benefits to the direct cash transfer. What should we be willing to spend for a marginal year of attainment?
During the 20th century, marriage was the biggest predictor of whether an American woman participated in the labor market. Upon marriage, a woman left the labor force permanently. (Interestingly, this effect is much less true amongst African-American women. Goldin speculates that work was seen as socially undesirable among Whites, but, due to slavery's legacy, not among African-Americans.) She spends an enormous amount of time assembling Census data and then survey data to correct Census data more/less to back these claims regarding women's participation and experience rates in the market.
More notes to come...
Phil Knight does a world-tour, builds an iconic brand.
What struck me most is how much he and his companions believe in their mission to sell running shoes. There's a funny story early in the book about Knight getting waylaid in Hawaii by the nice weather and pretty ladies, and to make ends meet, he takes a job as a door-to-door encyclopedia salesperson, but never makes a sale. With Nike, he believes so much in running and the shoes they sell that he's able to sell to Japanese shoe makers, factories, sales reps, track stars, his family -- the world, really. He doesn't go outright and say it, but it's alluded to many times, that he believes it was his destiny to sell shoes, to be a prophet/messiah for running.
The vast majority of the narrative is a story about lack of capital. Banks won't lend to Nike, because they're too leveraged, taking all profit and plowing into further growth. In a world of more developed capital markets, ie, venture, Phil Knight would have slept a lot better, Nike would have grown faster, and this memoir would be 80% shorter.
Nike has been innovating the running shoe since its first days as a shoe manufacturer. Nike's recent break of the 2-hour marathon barrier reflects this tradition. After I finished the memoir, I bought a new set of Nike running gear.
Steckel, Richard H. and White, William, Engines of Growth: Farm Tractors and Twentieth-Century U.S. Economic Welfare (March 2012). NBER Working Paper No. w17879, Available at SSRN: https://ssrn.com/abstract=2014575
Per capita number of people working on farms has dropped enormously. In 1910, 13 million people worked on farms. In 2012, 6 million people worked on farms. The population in 1910 was 92 million. In 2012, it was 310 million. This represents a 6 fold decrease in required labor inputs from 1 in 7 people to fewer than 1 in 50 people.
Low agricultural productivity growth encourage developmental economists to recommend transferring resources out of agriculture into higher-growth industries as a necessary condition for countries to grow. The low-growth in agriculture was true in total dollar or unit terms (total dollar value of annual agricultural products, total bushels produced), but not in rate terms (bushels per labor hour).
In 1910, farmland made up half of total land area in the continental US (879 million acres; Arizona and New Mexico were not admitted as states yet). 1/3 of this, 310 million acres, were used for growing crops.
Prior to 1910, most crops were grown, cultivated, and harvested using animal power or manual labor. Steam and gasoline engines had limited impact on farming due to their massive size and weight.
1910 census values horses and mules on the farm at $2.6 billion, or about 6% of all farm property, and implenents and machinery valued at $1.27 billion. Capital intensity can be calculated at $12.52 per acre, staying constant for the first half of the century.
Advancements in tractor technology: power lift by Deere in 1927; rubber tires, replacing steel wheels, in 1932; and development of diesel engines in the mid-1930s.
Tractor development stops in 1934, changing little in design for 30 years. Despite the depression, tractor sales increase rapidly until 1960, when increase in percentage of farms iwth tractors stops (outside of the cotton South).
Amount of land used to grow field crops stayed roughly the same. In 1954, 329 million acres were used (6% increase from 1909). Large decreases in corn and cotton acreage were balanced by 18 million acres of soybeans and 16 million acres of grain and sorghum. By 1954, tractors were used to plow and prepare nearly all this land.
Farm labor force of 1954 was significantly smaller than that of 1910. Half of the nation's farm operators held paying jobs away from the farm. More than 1 million of these operators worked more than 200 days per year off-farm. The hired work force worked fewer days per year, mostly during harvest time, as tractors had replaced the demand for more steady labor. Hired laborers are estimated to have increased by 10% from 2.8 million to 3 million.
Further reading: Johnson and Gustafson estimated that mechanization was responsible for one-third of the increase in corn yields in the 20th century.
Einav, Farronato, and Levin are all economists interested in industrial organization. They are at Stanford, Harvard Business School, and Stanford Graduate School of Business, respectively.
This paper is interested in peer-to-peer internet marketplaces: (1) design of internet marketplaces, (2) the economics of peer production, and (3) regulatory issues that were emerging in 2015.
Market Design of P2P marketplaces
Transactions rely on spot transactions over long-term contracts/employment relationships.
The decline in auctions in e-commerce extends to P2P markets. Many marketplaces started with auction pricing but have moved to some type of set pricing. Prosper (a lending marketplace) replaced its auction mechanism (interest rate) with a centralized pricing algorithm. Study shows risk is priced about the same, but the funding process has been greatly simplified.
The pricing unit matters greatly. Airbnb prices by night, not by number of guests per unit, as the former is easier to observe, reducing monitoring costs. TaskRabbit cleaning priced by hour not by job, and online ad marketplaces (Google, Facebook) moving to click/conversion pricing over impression pricing, are examples of aligning incentives between marketplaces and market participants.
Feedback and review systems are extremely pervasive and 98% of feedback is positive, but seems to weed out the worst actors.
A central challenge with P2P marketplaces is presenting options to buyers. P2P marketplaces are characterize by high degrees of heterogenity, influencing decision by marketplace to de/centralize the matching process. Transport is more homogenous so Uber centralizes the process. Accomodation is more heterogeneous, so Airbnb decentralizes the process to facilitate individual choice.
Economics of Peer Production
The authors create a theoretical model of professional/dedicated sellers and casual/peer sellers and derive intuitive results from the model. This is less interesting.
Volatile market demand favors the peer sellers. High capacity costs and frequent low-end marginal costs favor peer sellers. A study of TaskRabbit worker price sensitivity showed it to be very elastic: a 10% increase in wage rate led to workers applying for 30% more jobs.
P2P markets work when a new technology reduces the marginal cost of advertising (finding customers). A critical mass of supply is necessary, as the marginal advertising cost is reduced for each market participant with each additional supply unit.
Regulatory Issues
Two possible reasons for regulation: (1) to protect consumers (in response to market failure), (2) to protect incumbents.
There isn't a global reputation system for contractor workers. This was pitched as a business idea recently
Hall is now Uber's Chief Economist and Alan Krueger was a Princeton labor economist who served as chair of the Council of Economic Advisors in the Obama administration.
Through direct access to Uber's observed ride data and driver surveys, Hall and Krueger determine Uber drivers greatly prefer the flexibility Uber offers.
Uber driver growth has been driven by people seeking flexible work arrangements
When surveyed, 3 out of the top 4 reasons for joining Uber were related to having control of work scheduling.
UberX growth was exponential, whereas UberBlack growth was linear. Proportion of drivers driving 15 or fewer hours a week trended up, whereas drivers driving between 16 to 35 hours and 35+ hours trended down. UberBlack drivers are more likely
When surveyed, 74% (5%) of drivers said Uber made their life better (worse).
When asked to state preference between work descriptions that corresponded to independent contractor versus employee relationships, 79% of drivers said they would prefer an independent contractor relationship.
Female drivers were more likely to state a preference for flexibility than male drivers were. This reminds me of findings of Claudia Goldin.
Uber is largely supplemental income
80% of drivers report they were working full- or part-time before starting on Uber. 10% of drivers were unemployed immediately before joining Uber. For the overall economy, 25% of new hires come from unemployment.
The number one factor in driver growth was meeting unmet demand
Number of Uber drivers in a city is a linear function of (1) city's population, (2) how long Uber has been operating in the city, and (3) log number of existing taxi licenses. All factors are positive, so (3)'s implication is that the the more licenses there are, the more unmet demand there was (or, the more artifically constrained supply was). Factors that did not determine driver growth include (1) population density, (2) population GDP, (3) taxi earnings, (4) regional unemployment, (5) cost of a 5-mile trip in the city, and (5) cars as a proportion of poulation.