North Dakota’s economy started to boom about eight years ago due to advances in hydraulic fracturing, or fracking, and by 2014 it had the fastest growing economy in the nation. The city of Williston, in the middle of the state’s Bakken oil patch, grew by an astounding 67% from 2010 to 2014 according to the U.S. Census, and by even more according to city sewage data compiled by Reuters.

Though some are predicting a new boom as oil prices rise, for the time being lower oil prices have weakened North Dakota’s economy and the residents of Williston—many of whom moved there to take advantage of it—are trying to figure out what to do next. Do they stay and wait for things to get better? Or do they move on to greener pastures?

The city’s officials also face tough decisions. During the boom the city government, loaded with cash and facing increased demand for city services, built a $57 million dollar high school and other public facilities. The city’s liabilities quadruped from 2008 to 2014.

Now tax revenue is falling fast; the city’s sales tax receipts fell 47% from March 2015 to March 2016. Moody’s also downgraded Williston’s general obligation bonds to junk status, which increases the cost of future borrowing. If the population continues to decline, the city will be left with new infrastructure and buildings that have to be maintained by a smaller tax base.

While many in Williston and the surrounding area remain optimistic about the future, in many ways what is occurring there now is a time-lapsed version of what happened in many Midwestern and Northeastern cities in the 20th century. One can get a sense of the similarity by looking at a graph of city populations.

The left panel in the figure below from the Reuters article shows Williston’s population from 2010 to 2015: under 20,000 in 2010, rising to over 30,000 by 2014, and then falling, with the whole thing looking like an inverted U.

We can see the same shape in the populations of other declining cities, as shown below.

This figure depicts the populations of Scranton, PA; Cleveland, OH; Dayton, OH; Pittsburgh, PA; Buffalo, NY; and Newark, NJ from 1900 to 2010. There is nothing unique about this particular group of cities and dozens of others could have been chosen that have similar population paths. All of them have the same inverted U shape because they all have a similar story.

First, each had an economic boom due to location, the discovery of a new way to use a resource or the invention of a new product: In Williston it was fracking; in Dayton it was the cash register, airplane and manufacturing in general; in Pittsburgh it was steel.

The boom in tradeable goods creates a demand for non-tradeables – haircuts, bars, banks, restaurants, hotels, retail stores, legal services and salons to name a few. Housing prices rise and new housing is built in response to the higher prices. The city’s population grows.

City officials flush with cash invest in infrastructure, create a social safety net, construct public parks and engage in all-around city beautification. Public sector unions are able to negotiate higher wage increases and lavish pension benefits since everyone is confident the money will be there in the future.

Though not without its problems – increased congestion, residents’ anxiety about how growth will affect the character of their neighborhoods, construction inconveniences – the economic and population boom is the good part. After the boom there is usually a 10 to 20 year period of population stability.

Then it happens – the bust. Entrepreneurs in other cities and even other countries invent new products that supplant the old ones. In the U.S., changing institutions in the American south, technological advances like air conditioning, and rising incomes have caused people to migrate to areas with better climates. Competition from abroad weakened the domestic auto industry and with it cities like Detroit and Dayton. Silicon Valley emerged in California and people moved west. As southern and western cities grew and became more cosmopolitan, Cleveland and Buffalo looked less and less attractive leading to additional out-migration. As shown in the figure above, this out-migration has slowed since the 1980s but has yet to completely stop.

The result is a shrinking tax base with too many financial commitments. Bonds issued in better years need to be paid off and underused infrastructure still needs to be maintained. But with less people, either taxes need to increase or some level of deterioration must be accepted. Many of the above cities tried a combination of the two, but higher taxes and urban blight tend to push people away.

In under a decade Williston has gone through an economic cycle that took nearly a century in many other cities. And it’s not the first time; similar booms and busts occurred in Williston in the 1950s and 80s. Where it goes from here is not clear but it will ultimately be up to the people of Williston who remain.

Lessons

The biggest lesson from Williston’s and the other cities’ experiences is that city governments need to stay simple and flexible. They can’t provide everything to everybody. Cities in general lack the revenue generating mechanisms to provide a social safety net. Taxes and policies that redistribute income at the city-level cause people to leave since it is relatively easy to move to a lower tax jurisdiction.

To the extent that redistributive policies are sensible, it is easier for countries, and to a lesser extent states, to implement them since larger jurisdictions make it costlier to avoid such policies. Only a couple of cities – New York and San Francisco come to mind – appear to be able to redistribute wealth at any scale without losing their wealthiest residents, but they are the exception. Billionaire David Tepper’s recent decision to leave New Jersey and its effect on the state’s budget highlights how redistributive tax policy can backfire if it induces wealthy residents to leave.

Moreover, pension payments must be paid in the future and it’s hard to predict how large the future population will be that has to provide these payments. In the past, declining cities had bigger populations with bigger government workforces, but unfunded pension benefits must be paid by today’s smaller tax bases. Scranton is an example of the problems this can cause. Defined contribution plans alleviate this uncertainty and prevent over-promising since the money is paid as the benefits accrue.

Infrastructure should be built slowly and kept simple – function over form. Issuing bonds is often necessary but should be done prudently. Infrastructure levies for specific spending should be offered to voters; if they want it they will vote for it.

Finally, local rules and regulations often stifle entrepreneurship and economic growth. Specialization and innovation are the comparative advantages of a city, and those that don’t exploit these advantages will decline. Too many rules and regulations discourage entrepreneurs from creating new products and services and can drive out innovative companies, as Austin officials recently discovered when they tried to impose additional rules on ridesharing company Uber. Dynamic cities often take their dynamism for granted and end up crushing it under the weight of local government.

On top of that, the enforcement of complex rules, regulations, and ordinances requires a large government workforce that must be paid and often shrinks too slowly when populations decline, contributing to the aforementioned pension problem.

Cities are always rising and falling in America; sometimes over a century, sometimes in just a decade. City officials and residents must recognize that the good times are unlikely to last forever. Doing so will encourage them to act prudently when it comes to economic policy and expenditures, ultimately enabling America’s cities to age gracefully.