october 2014

Forecasting the uncertain future

by Todd Graham, principal forecaster, Metropolitan Council

Todd GrahamWhere and when are population and economic growth expected? And how much growth? Just about everyone has an opinion or an ambitious vision. At Metropolitan Council, the planning agency for the seven-county Twin Cities metro area, our forecast team works to bring analytic expertise to this discussion: What do we really expect based on market analysis, regional economics, and demography? Our approach to providing answers involves use of forecast models.

Models are an attempt to represent real-world systems in a simplified way, but also in a way that allows us to understand and experiment with the big picture. Real-world systems are reduced in model-space to representations of key components, behaviors, and interactions between moving parts.

Our potential for insight from models depends on how well the models are reasoned. Ideally, forecasters understand the systems at issue, apply appropriate methods, articulate appropriate mathematical representations in model-space, use reliable data, and make judicious assumptions. Longer timespans, added detail, and the multiplicity of systems represented can add complexity and multiply forecast uncertainties.

A great illustration of the challenges of long-range forecasting is the array of forecasts of the Twin Cities metro’s future population. Census 2010 counted 2,850,000 people in the metro. Since then, the metro’s population has grown to 2,951,000 in 2013. Depending who you ask, the recent population growth rate could slow dramatically, could accelerate, or could remain about the same.

There are at least five published forecasts of the Twin Cities metro’s population:

  • Minnesota State Demographer projects 2040 metro population of 3,342,000.
  • HCF/Profamy, a demography consultant, projects metro population of 3,604,000 in 2040.
  • Metropolitan Council projects metro population of 3,674,000 in 2040.
  • REMI Inc., a vendor of regional economic models, projects metro population of 3,906,000 in 2040.
  • Woods & Poole Economics projects metro population of 3,950,000 in 2040.

These substantially varying expectations mostly derive from model methodology or assumptions.

At the low end of the range, Minnesota’s State Demographer, charged with informing state agencies, assumes that future migration flows will resemble 2000-2010 performance: Minnesota employment did not grow during that recent decade; migration was negligible; thus the State Demographer infers that long-range, future migration will be negligible*. At the high end of the range, economic consultancies like Woods & Poole typically start with industry-centered, regional macroeconomic models. Such models presume that the fastest-growing regions will be those with a diverse industry mix, good business conditions and cost advantages, good access to production inputs, and good access to markets and workforce. The key difficulty for such models is including in the model framework, and properly weighting, the demographic and other noneconomic factors that might limit workforce growth.

Most metropolitan planning agencies in North America, including Metropolitan Council, practice integrated modeling of regional economies and regional population. As the economy grows (or contracts), workforce demand grows (or contracts). In an integrated model, economic growth, workforce growth, and population growth are forecasted simultaneously, not in isolation or with historically-based assumptions. Economic opportunity and wage growth draw in migrants from the rest of the nation and the world. There may also be noneconomic or quality of place dynamics that promote (or detract from) a region’s migration. Forecasters attempt to balance these competing dynamics in model-space.

All forecasts are premised on assumptions. At Metropolitan Council, we assume we know the important trends (aging of the baby boomers), or else we assume steady-state conditions (e.g. Minnesota’s cost-of-living remain slightly better than the national average). Still other aspects of the future are more difficult; there is a long list of uncertainties.  

In summary, the future has not happened yet. Trends may change. Assumptions will fail. But in the meantime, we structure modeled futures with the best evidence and techniques available, and considerate of the limits of our knowing. 

Todd Graham is a principal forecaster for Metropolitan Council, where he provides project leadership in development monitoring, land use economics, demography, and economic analysis. Recent projects include preparation of the Council’s 2040 socioeconomic and land use forecasts. From 1997 to 2004, Graham was a regional analyst and regional analysis manager for the Department of Employment and Economic Development.

*Editor's note: The State Demographic Center bases its assumptions about Minnesota’s future migration trends on the U.S. Census Bureau’s population projections for the U.S. to 2060. Assuming that Minnesota will continue to grow as its respective share of the nation, the State Demographic Center projects that the number of migrants (on net) gained each year will grow from about 5,000 in 2015, to 12,000 by the year 2040. As a point of reference, between 2001 and 2009 period, Minnesota gained an average of 6,000 migrants each year (net).

Find more about the State Demographic Center projections here <http://mn.gov/demography/projections>

Find more details about the Metropolitan Council’s forecasts here <http://www.metrocouncil.org/Data-and-Maps/Data/Census-Forecasts-Estimates.aspx> .

On Minnesota Compass, we include Twin City Region population forecasts by both the State Demographic Center and Metropolitan Council.

Opinions expressed in Minnesota Compass guest columns do not necessarily reflect the views of Minnesota Compass. Compass welcomes a range of views about issues pertaining to quality of life in Minnesota.

Minnesota Compass

Minnesota Compass
www.mncompass.org
Led by Wilder Research

© 2017. All rights reserved.