The Cherokee Land Lottery: What Happens When You Randomize Wealth

Back in August I wrote a post on Give Directly, a charity that simply gives money to poor people, and it sparked one of the better debates on the blog. CashRebel recently wrote a post on the subject as well, and he specifically asks whether such a program would work here in the United States. Would simply providing wealth to American familes have long term positive results? Would the lives of those families improve?

As interesting as these questions are, they are a bit academic. It makes for good debate, but how do you get data? Because even if you could give life-changing sums of wealth to people here in this country, it would take decades to see the long term impacts of that charity. Maybe generations. Wouldn't it be great if someone had performed a natural experiment about randomized wealth a long time ago, here, in this country? Well, thanks to an idea of Georgia politicians in the 1830s, an article by economic professors Hoyt Bleakley at the University of Chicago and Joseph Ferrie at Northwestern, and the excellent podcast from Freakonomics, we get to learn of a quirky and sad bit of American history. (And if you have thirty minutes, I definitely recommend listening to the podcast.) Their work reveals what happened when a huge sum of wealth was randomly distributed in the Cherokee Land Lottery.

The Natural Experiment:
Because of a previous scandal in Georgia, in the 1830's the state's politicians were under extreme pressure from citizens. To quell public pressure, politicians decided to distribute the land that was tragically taken from the Cherokee people (and led to the the Trail of Tears), via a series of lotteries. If you were a white male who had lived in Georgia for three years or more and were over 18, you were eligible to buy one ticket for 12.5 cents: one "draw" per man maximum. (Some other groups, such as widows, orphans, or veterans, received two draws.) Winners of the lottery would receive 160 acres of land, worth approximately $700 at the time (equivalent to 900 days' wages for an unskilled laborer). This amount of wealth was roughly equal to the median wealth of residents of the time. Winning meant that the even the poorest Georgians would immediately have the wealth of an average citizen: a huge jump up the socioeconomic ladder. The payout was so good that there was near universal participation: the economists estimate that between 97.2% and 99.5% of eligible residents participated in the lottery. And the lottery was no real long-shot: roughly 19%, one in five, participants won. With a twelve cent ticket giving a 19% chance of winning $700 of land (worth nearly three years of wages), residents had every incentive to participate.

This sort of organic randomization, a natural experiment, basically never happens in real life. And it happens to address a great question for economists and policy makers as well: is the lack of wealth a barrier to investing in your family's human capital (e.g. - sending them to school, improving their future prospects, etc.)? Finally, because the first US Census was held relatively soon afterwards in 1850, the lottery winners and their children could be tracked. Researchers could use the data from this experiment, along with census data, to see the long term impact from this sudden impact of wealth on families. The Cherokee Land Lottery, while a tragedy for the Native Americans and a blemish on our history, was a windfall for these economists.

The Results:
So, what happened? Did the families benefit over the years? Did their children have greater wealth, income, school attendance, or literacy? It turns out, no. Amazingly, there was no long term benefit in terms of human capital, wealth, or income. From the paper:

"Although winners had slightly more children than non-winners, they did not send them to school more. Sons of winners have no better adult outcomes (wealth, income, literacy) than the sons of non-winners, and winners’ grandchildren do not have higher literacy or school attendance than non-winners’ grandchildren."

Depressing, right? The families who won this once-in-a-lifetime windfall of wealth, did not see any long term benefit. This is a period where Americans were actually sending their children to school. And the common wisdom is that poverty is an impediment to that: the poor don't have access to education (or equivalent education) because they don't have the means. Additionally, people who were literate did make more money: so there was an incentive to send children to school. But, in this setting, a shock of wealth did not have any snowball effect. The lottery winners' children were no better off than non-winners'.

The Conclusions:
A first objection is that giving land is not the same thing as giving cash. Give Directly provides cash, not acres of land. But it should be noted that back in the 1830's, the typical investment outlets for our capital (stocks, mutual funds, REITs, etc.) did not exist yet. The measures of wealth at the time in rural Georgia were land and slaves. So providing land seems to be a reasonable equivalent of providing wealth.

As this experiment didn't result in any long term benefits for the recipients of sudden wealth, what lessons can we take from it? The authors find that results are in line with what we'd expect today. They note that we can see the same general result with lottery winners: there are typically not long term benefits from a sudden injection of wealth. Within a matter of years, the money is gone and the families (and often the recipients themselves) are no better for having won (and are often worse off). (Though it should be noted that modern lotteries have a negative expected value, while the Cherokee Land Lottery had a very positive expected value.)

The study suggests that the disadvantages for children in poor households don't seem to simply be the direct result of a lack of capital. This authors argue that money, in particular, just being given to the fathers and mothers in poor households, isn't likely to be what the poor need. Something is needed, but a sudden influx of cash alone is not it. Based on this evidence, the simple act of giving money to poor people isn't likely to be the answer. There are likely cultural and socioeconomic issues in play that are far too complex to be solved by simply throwing money at them. The poor might indeed need capital, but there needs to be something else (likely a lot of somethings) paired with it. (As readers of the blog will recognize, this is what Mrs. Done by Forty argued all along. As usual, the lesson seems to be that I should not disagree with my wife.)

In closing, Hoyt Bleakley describes the potential takeaways for our current society. From the podcast:

"When you select a particular group of the population, and you either give them money or you cajole them to get a little more schooling by bribing them, with a cash transfer or cell phone minutes or what have you, you have to ask whether there's some other set of characteristics that they have that makes it hard for them to take advantage of those opportunities. And maybe there's an intervention that helps them better manage those other characteristics; that makes it such that that's less of a disadvantage for them. Whereas giving them something, you say, 'Well this was great for me; it will be great for you,'...that's perhaps not the right approach."

I don't know about you, but I find the whole thing to be interesting, but kind of a downer. Sorry for the depressing post, today. But what do you think of the study? Do you believe that the authors are wrong, and that an influx of cash via charities like Give Directly would have long term positive impacts?


*Photo, possibly of a Cherokee land rush claim, is from rcstanley at Flickr Creative Commons.