Why Is Western Mass Losing Population? Sixty Years, Town by Town

The premise of the question is wrong, and it is worth saying so before anything else. Western Massachusetts is not losing population and has not lost population over any long horizon you care to measure. In 1960 these 101 towns held 729,581 people. At the last census they held 828,188, which is 13.5% more.

Only 11 of the 101 towns have fewer people now than in 1960. 57 of them recorded their highest ever population at one of the last two censuses. The region as a whole is at or near its peak.

So why does everybody believe the opposite? Because the places that lost people are the places everybody has heard of, and the places that gained are places most people could not find on a map.

The Mill Cities Took The Entire Loss

Five towns account for the story people are telling, and four of them are cities:

• Monroe: 210 in 1960, 118 in 2020, a change of -43.8%
• North Adams: 19,905 in 1960, 12,961 in 2020, a change of -34.9%
• Adams: 12,391 in 1960, 8,166 in 2020, a change of -34.1%
• Holyoke: 52,689 in 1960, 38,238 in 2020, a change of -27.4%
• Pittsfield: 57,879 in 1960, 43,927 in 2020, a change of -24.1%

Between them, the cities on that list are down roughly 39,572 people from their 1960 counts. That is a serious, visible, generational decline, and it happened in the places with the newspapers, the hospitals and the reputations. If your impression of this region comes from driving through North Adams or Holyoke, you are looking directly at it.

The cause is not mysterious. These are cities built around specific industries at specific addresses, textiles, paper, machine tools, electrical equipment. The industries left between the 1950s and the 1990s and the housing stock stayed exactly where it was. That is the origin of the single most useful fact about this region's property market: there is more housing in those cities than there is demand for it, and prices have reflected that for fifty years.

Meanwhile The Hill Towns Filled Up

Now the other half, which nobody writes about:

• Shutesbury: 265 in 1960, 1,717 in 2020, a change of +547.9%
• Mount Washington: 34 in 1960, 160 in 2020, a change of +370.6%
• Tolland: 101 in 1960, 471 in 2020, a change of +366.3%
• Holland: 561 in 1960, 2,603 in 2020, a change of +364.0%
• Peru: 197 in 1960, 814 in 2020, a change of +313.2%

Several of these towns have four, five or six times the population they had in 1960. They are still small in absolute terms, which is exactly why the growth is invisible: a town going from 200 people to 1,100 is a transformation locally and a rounding error regionally.

What drove it was the car, then the mortgage, then the internet, in that order. Land in the hill towns was cheap because it was inaccessible, and each of those three things made it less inaccessible. The last of them is still working: a job that can be done from a spare room changes the calculation for a town forty minutes from anything, which is why what the broadband record actually shows is now a property question rather than a convenience one.

What This Means If You Are Buying

Three things follow from the pattern, and all of them are practical.

1. Cheap housing in the cities is a supply story, not a warning. Springfield, Holyoke, North Adams and Pittsfield have housing stock built for many more people than live there now. That is why the same money buys so much more there, and it is a structural condition rather than a signal that something is about to go wrong. What it does mean is that price appreciation there has historically been slower, because supply is not the constraint.
2. Hill town growth has not been matched by building. Population up several hundred percent, permits scarce, and the result is a market where very little is for sale at any given moment. What actually gets built here now covers the permit side of that, and it is thinner than the growth would lead you to expect.
3. Check the peak column before you read a town's trajectory. A town that peaked in 1960 and one that peaked at the last census are in genuinely different situations, and both can have the same population today.

Every Town, Seven Censuses

Sortable and filterable. The peak column is the quickest read: it tells you whether a town's high-water mark is in living memory or a long way behind it.

The Honest Caveat

Sixty years is a long horizon and it can hide a recent turn. Growth across the whole period is compatible with a decline in the last ten years, and for some towns here that is exactly what the row shows. If you are interested in what is happening now rather than since 1960, which towns are actually growing today is the shorter-horizon version, and who is arriving and what they earn covers the flows behind it.

And a count is a count. It says nothing about ages, and a town holding its population while growing older is a different proposition from one holding it by attracting families. Both look identical in the table above.