In which a group of graying eternal amateurs discuss their passions, interests and obsessions, among them: movies, art, politics, evolutionary biology, taxes, writing, computers, these kids these days, and lousy educations.

E-Mail Donald
Demographer, recovering sociologist, and arts buff

E-Mail Fenster
College administrator and arts buff

E-Mail Francis
Architectural historian and arts buff

E-Mail Friedrich
Entrepreneur and arts buff
E-Mail Michael
Media flunky and arts buff

We assume it's OK to quote emailers by name.

Try Advanced Search

  1. Seattle Squeeze: New Urban Living
  2. Checking In
  3. Ben Aronson's Representational Abstractions
  4. Rock is ... Forever?
  5. We Need the Arts: A Sob Story
  6. Form Following (Commercial) Function
  7. Two Humorous Items from the Financial Crisis
  8. Ken Auster of the Kute Kaptions
  9. What Might Representational Painters Paint?
  10. In The Times ...

Sasha Castel
AC Douglas
Out of Lascaux
The Ambler
Modern Art Notes
Cranky Professor
Mike Snider on Poetry
Silliman on Poetry
Felix Salmon
Polly Frost
Polly and Ray's Forum
Stumbling Tongue
Brian's Culture Blog
Banana Oil
Scourge of Modernism
Visible Darkness
Thomas Hobbs
Blog Lodge
Leibman Theory
Goliard Dream
Third Level Digression
Here Inside
My Stupid Dog
W.J. Duquette

Politics, Education, and Economics Blogs
Andrew Sullivan
The Corner at National Review
Steve Sailer
Joanne Jacobs
Natalie Solent
A Libertarian Parent in the Countryside
Rational Parenting
Colby Cosh
View from the Right
Pejman Pundit
God of the Machine
One Good Turn
Liberty Log
Daily Pundit
Catallaxy Files
Greatest Jeneration
Glenn Frazier
Jane Galt
Jim Miller
Limbic Nutrition
Innocents Abroad
Chicago Boyz
James Lileks
Cybrarian at Large
Hello Bloggy!
Setting the World to Rights
Travelling Shoes

Redwood Dragon
The Invisible Hand
Daze Reader
Lynn Sislo
The Fat Guy
Jon Walz


Our Last 50 Referrers

« Toby's Movie | Main | It's Hank's Fault »

September 29, 2008

Hits and Misses: New York Forecasts

Donald Pittenger writes:

Dear Blowhards --

In a recent post I mentioned that in my dark past, I created population forecasts of counties in New York state. This admission might have been a mistake, because Benjamin Hemric and Michael B quickly appeared in Comments asking for more information.

Since my back is now against the wall, here goes:

In the mid-1960s Nelson Rockefeller was governor, a man who believed in big, bold government projects -- a trait not unheard of in other occupants of the governor's mansion in Albany. Among his other accomplishments were the transformation of a collection of teachers colleges and other schools into the State University of New York (SUNY) system and the building of the Albany Mall (which I wrote about here).

Another initiative was the establishment of an agency named the Office of Planning Coordination (OPC). Perhaps not having read (or having forgotten about) Friedrich Hayek, Rockefeller and his advisers thought that planning was a Good, Rational Activity -- which it is in an ideal world. Elsewhere, such an agency would be yet another collection of over-educated bureaucrats who would pass their time awaiting retirement by writing studies whose destiny would be the oblivion of a state archives file drawer.

Unfortunately, OPC was saddled with a slight problem: it was intended to hold actual power. In fact, the concept was that OPC would be co-equal with the budget agency, something almost unheard of. Poor Rockefeller: he didn't realize that no one was co-equal with Budget. So, on its creation, OPC faced a mortal enemy. Other enemies quickly emerged amongst other agencies that resented being dictated to as well as county and local governments long leery of the actions of higher-level governments. This soon translated into lack of support by legislators.

Nevertheless, all went seemingly well. I was hired during the summer of 1970 to create regional and county population projections, a task formerly performed by the Cornell Aeronautical Laboratory in Buffalo.

All was jolly, state government cruising along, taxpayer money being spent and spent and spent. Until after the November election where Rockefeller had earned his fourth term. Then he jammed on the spending brakes. Cuts had to be made. Employees would have to be let go. This was something new to New York State employees. In the end, two agencies were targeted, one being OPC. OPC didn't disappear, but it lost about half of its staff (if I remember right); actually many dismissed people popped up in other agencies, so I question how much economizing actually happened.

Even though I hadn't completed my six-month trial period, I was retained (to the displeasure of others who thought their jobs were secure). That was because the Office of Planning Services (OPS), the renamed, shrunken agency felt population forecasts were a key product.

The first set of OPS forecasts appeared in 1972, about a year after necessary benchmark data from the 1970 census were released. At the time, New York was more of a major state than it is now. Its Empire State nickname was not a case of bragging because for many decades it was the most populous state and the home to many large and important businesses. But the trends were against it. Population has been tending to shift from the north and east to the south and west for a long time, becoming especially noticeable for the last 50 years. As long ago as 1962 California snatched away its claim to being the state with the greatest number of inhabitants.

So we knew that New York was slipping, but were unsure if that slippage would be gradual or sudden. Being government bureaucrats, we hedged in the former direction but, in the end, the second condition was the reality. Another reason for caution was the decimated state of the agency. It had barely survived one assault and more trouble could be expected; we didn't want to be in the wave-making business. Even so, OPS disappeared not long after I left it at the end of 1974.

I should add that I was new at the population forecasting game. I had completed my doctoral exams at Dear Old Penn and was chipping away at my demography dissertation, so it could be said that I had had a reasonable amount of training. Except that I hadn't been trained to do state, regional and county forecasts. The only projection training I had received had to do with national populations -- a slightly different technical animal. In fact, I'm not sure if there was any academic training available in the field of sub-national population forecasting back in the 1960s. My life raft was the methodology from the Cornell Aeronautical Laboratory work, which I was able to expand upon. A few years later my self-training had reached the point where I was able to write a book on the subject, but that's of no concern for the present tale.

My training led me to approach forecasting in terms of the traditional components of demographic change: fertility, mortality and migration. For each of the state's 62 counties, I worked out sets of future rates for each component (actually, mortality rates were the same for every county at any time interval, but in principal I could have introduced variations). Today, I would proceed differently, specifying population control totals first and having the computer system work out appropriate age-sex detail later. Such totals would be the result of analyses of historical share changes (state to the nation, regions to the nation, counties to the nation, regions and counties to the state, etc. for various time periods) and then using judgment to select the most appropriate resulting projections and do some fine-tuning.

The 1972 set of forecasts was a collaborative work. Beside myself, my boss and his boss were involved in settling on numbers that seemed reasonable at the time. We also conferred with regional planning agencies and other interested parties. For these reasons, I can't be totally praised or blamed for the results, though my contribution was important.

The projections went forward from 1970 to 2000 in five-year intervals. Now that the impossibly far-in-the-future 2000 date is eight-year-old history, I'll focus on how accurate we were for that point. Please bear in mind that 30 years is a long time for almost any kind of forecast. Also be aware that forecast accuracy almost always decays in proportion to the distance from the benchmark or start point.

We predicted the the state would have 22.65 million people by 2000; the 2000 census reported only 18.98 million, so we were 3.68 million people and more than 19 percent high. We released a second set of projections in 1974, reducing the 2000 forecast to 21.33 million -- closer, but still too optimistic. I'll deal with the 1972 numbers here.

I created a spreadsheet to compare the forecast with the census data. I'm not sure how to present it as a whole, so will present selected results to give you a sense of how things went. And I'll focus on the county projections which were the heart of the effort.

In case you aren't familiar with New York state, click here for a counties map -- and if you click on a county, you'll get the Census Bureau's latest estimates on various characteristics as well as some 2000 census data. Since the map has boundaries only, here are some orientation hints: Buffalo is in Erie County, Rochester in Monroe County, Syracuse in Onondaga County, Utica in Oneida County and Binghamton is in Broome County.

Given that we over-projected the state, it stands to reason that most of the counties suffered a similar fate, given that they would be dragged in the same direction. Even so, nine counties -- about 15 percent of the total -- were off by five percent or less. They were: Allegheny, Cattaraugus, Essex, Greene, Kings (Brooklyn), Lewis, Saratoga, Tompkins and Yates. The most accurate forecast was for Saratoga County, a suburb of Albany. For 2000 we forecast 199,653 population and the census turned up 200,635, a difference of 982 or about a half-percent error.

To balance this, we were grossly off for many other counties, the worst being Dutchess where we were 84 percent high. Other counties where we were 25 percent or more high were Broome, Chemung, Genesee, Livingston, Nassau, Oneida, Onondaga, Orange, Oswego, Putnam, Rensselaer, Schenectady, Schuyler, Seneca, Suffolk, Tioga, Ulster, Wayne and Westchester. Errors of 50 percent and more occurred for Dutchess, Orange, Putnam and Suffolk counties.

The pattern is that we over-forecast suburban and exurban counties. I recall having had the hypothesis that one area ripe for growth was the mid Hudson Valley, which is why Dutchess, Orange and Putnam counties are so far in error.

I mentioned that the high state total meant that many counties were pulled by it in the same direction. As a simple way of correcting for that factor, I computed county shares of the statewide population in 2000 for the forecast and the census. Fifteen, or nearly a quarter of the counties had projected shares within 5 percent of census shares. These counties were: Broome, Chemung, Chenango, Clinton, Columbia, Delaware, Hamilton, Niagara, Ontario, Orleans, Richmond (Staten Island), St. Lawrence, Sullivan, Warren and Wyoming. Of these, Chenango, Warren and Wyoming had shares that were essentially without error. Seven counties' shares were off by 25 percent or more: Dutchess, Franklin, Jefferson, Orange, Rockland, Suffolk and Tioga.

Forecasting is a tricky game. Complete accuracy is rare and probably accidental. You know that you are more than likely to be overly influenced by recent events when you establish your assumptions regarding the future. And there are all those unknown unknowns that will occur to ruin your hard work.



posted by Donald at September 29, 2008


But didn't you yourself earlier claim you had to say the decline would be gradual rather than sudden to avoid losing your job?

Nothing wrong with that, but don't beat yourself up for being wrong afterward. It's like complimenting the boss's tie and then lamenting afterward your lack of taste. You do what you have to to survive, same as people have been since the dawn of time.

Posted by: SFG on September 30, 2008 10:01 AM

Everyone in NY has known forever that the size of the state government is the problem. Taxes are not high, they're disastrously high. Doesn't matter. The socialists want what they want. The progressive agenda must go forward, the peasantry be damned.
One thing I can guarantee: not even this Wall Street collapse will result in smaller government. They'll just ratchet up property taxes one more time, thereby driving out even more of what's left of the tax base.
Solution? There is no solution. Lefties are quite content to level us all down to misery. As long as THEY continue to rule.

Posted by: ricpic on September 30, 2008 11:00 AM

SFG -- I said nothing that drastic in the comment to which you're probably referring. We were all worried about not making waves. On the other hand, we didn't have strong evidence because we lacked decent current estimates in those days. The Census Bureau wasn't yet into that game and another NYS agency was doing the job using an incredibly naive method: subtracting deaths from births, then adding the result plus the previous decade's net migration, prorated to the census benchmark population. Such a method makes it impossible to glean current trends.

On reflection, what I said in the comment about only letting losers lose mostly had to do with a set of township numbers that we cobbled together at the time of the 1974 revision.

The proverbial bottom line is that none of us knows the future. We knew NYS was losing steam (and had been for years), weren't sure if its pace would change. So we took the cautious path, something we likely would have done if the agency's position had been strong.

Posted by: Donald Pittenger on September 30, 2008 11:02 AM

Thanks for this. Fascinating in many ways, among them as a reminder of what New York State was once like: big, burly, bossy, and about to crumble under its own weight. You've got me curious about one thing, well, one among many: How necessary is it for something like a state govt to have forecasts? If the reality is that 1) everyone knows a decline has set in, and 2) nobody can really predict the future ... Is it really necessary to go through with the kinds of exercises you've described here? Is it an exericse that bureaucracy demands?

Posted by: Michael Blowhard on September 30, 2008 6:01 PM

Michael -- As best I remember, the NYS forecasts were what I term "general purpose" numbers that were "official" (scary concept, that) for government planning activities. (Private parties and firms were free to use 'em too.) So if some project was trotted out in Broome County, say, the numbers would be looked at as a check on whether or not it might be needed.

At the more practical level, long-range numbers can be used for infrastructure planning purposes. For instance, what diameter sewer pipe to lay -- pick something too small, you might end up digging it out and replacing it with something more adequate, which is expensive; pick something too large and that will increase your costs unnecessarily, though less so. But as I said, any long-range (5 or 10+ years) projection can be pretty iffy, especially for places where trends aren't stable.

In Washington State, we came up with statewide age-sex projections each fall that were mostly intended for short-term (2-6 year) budget planning, especially for stuff like caseloads. Large area + short horizon means the numbers were less risky, though seldom exact.

Posted by: Donald Pittenger on September 30, 2008 6:32 PM

Thanks, Donald, for a retrospective look at the statistics and a behind the scenes look at the agency.

One of the reasons I asked was that I had in mind that only "good" news was expected from the agency by its patron (the government) and that the patron wouldn't be happy with the messenger once he heard what the agency really had to say. Also I was wondering if officials of various counties (or from the state government itself) might have lobbied the agency for a rosier picture than the data merited.

What made me think of the above were some news stories in the late 1980s, which if I remember correctly, reported a "brain drain" in NYS and the governor's response denying a brain drain. And then there are all those stories about lawsuits regarding census bureau statistics.

But if I understand you correctly, the politics was more of an inter-agency type.

- - - - - -

By the way, it's interesting how demographic projections seem to get interpreted as hard facts by politicians, the media and the general public -- at least in NYC. A year or two ago a projection came out that NYC would grow by another one (?) million people or so. While I'm not saying this won't happen, it's amazing how politicians, the media and the general public seem to talk about this as though it is an absolute fact and that projections are never, ever off, and that a beneficial population increase will not have to be "earned" with a good economy, etc.

Posted by: Benjamin Hemric on September 30, 2008 10:40 PM

Post a comment

Email Address:



Remember your info?