Thursday, February 26, 2009

Why Brooklyn arena tax revenues likely would be lower than projected (and why the IBO should take another look)

New York City's Independent Budget Office (IBO) may not be ready to recalculate a cost-benefit analysis for the planned Atlantic Yards arena, but there's surely a reason to do so, because one key statistic has likely changed, and one key assumption was likely wrong from the start.

As I describe below, that could mean a 23.2% decline in expected new spending, and a significant--if not quite as high--decline in sales tax revenues. If so, there'd be even more evidence that the arena would represent a loss to the city rather than, as previously analyzed by the IBO, a "modest fiscal surplus."

And it would be another reason to tilt the balance between public and private benefit from the project--the question at the heart of the AY eminent domain case argued Monday--a bit toward the latter.

My estimates are hardly foolproof. But any future study should offer a range of potential outcomes, rather than rely on past efforts at precision, which are inherently flawed.

And a decline in revenues would raise further questions about whether the project deserves federal stimulus funds, despite the developer's lobbying.

Questionable assumptions

Let me explain. When Smith College sports economist Andrew Zimbalist, Forest City Ratner's paid consultant, calculated expected tax revenues from the planned Atlantic Yards arena, one key element was what Zimbalist called the "recapture of tax revenues presently generated in New Jersey," current home of the Nets basketball team.

That estimate depends on multiple variables.

1) Do the arena attendance figures remain accurate? Likely no.

2) Did Zimbalist used the appropriate formula to calculate how many fans from New Jersey attended games in New Jersey? Again, likely no.

3) Did Zimbalist accurately estimate how many of those New Jersey fans would go to Brooklyn and thus represent new spending in New York? Perhaps yes at the time, but likely no at this point.

Keep in mind that, in terms of economic gain for the city (as opposed to gain for the developer), the key is not whether the Brooklyn arena would fill up, but how many fans would be come from out of state and deliver new revenues. (Spending by New Yorkers would not be a boost, since it mostly would be diverted from spending on other local events.) And if that New Jersey number declines, then estimates for new New York City tax revenues go down, as well.

Based on a dodge

First, an aside. When paid by Forest City Ratner, Zimbalist in his 5/1/04 report (Estimated Fiscal Impact of the Atlantic Yards Project on the New York City and New York State Treasuries) praised Atlantic Yards as different from most sports facility projects because it involved moving a team across state lines, not building a new arena for a team already in place. That means the AY arena would generate new local and state taxes, and New Jersey would lose out.

Unmentioned was that the burden fell on federal taxpayers to subsidize tax-exempt bonds for the new facility. Zimbalist, when not paid by FCR, wrote in a book published a year earlier that "there appears to be no rationale whatsoever for the federal government to subsidize the financial tug-of-war among the cities to host ball clubs."

He conveniently forgot that criticism in his report for FCR, which was not peer-reviewed but was taken seriously by the New York Times, among others.

Attendance numbers game

The IBO and Zimbalist used average attendance figures of 14,765, based on the first 32 home games at the Continental Airlines Arena, now Izod Center, in two seasons, 2002-03 and 2003-04. (The Nets actually drew more, averaging 15,184 in 2002-03 and 14,952 in 2003-04, according to ESPN.)

This year, announced attendance figures, though slightly higher, have been more suspect. For example, when the average was reported at 16,098, Julian Garcia of the Daily News wrote December 11, that "seems suspiciously high to those who regularly attend games at the Meadowlands."

Al Iannazzone of the Record, writing the next day, suggested several reasons for the dropoff, including the economy and the lingering promise that the Nets would move to Brooklyn. (Now the regularly-shifting official Brooklyn opening date is 2011, though many doubt that and I think 2012 is a more likely best-case scenario.)

Real attendance figures

At the halfway point of this season, the number was 15,153. (It's since dipped to 14,739, which is nearly the number Zimbalist used.)

However, if attendance (at the halfway point) is overstated by 10%, or 1515, then the gate count would be 13,638. If the overstatement is 20%, or 3031, then the average gate count would be 12,122.

I suggested a 25% fudge factor, which would turn the 15,153 average into 11,365 attendees. But I'll use the 20% discount in some calculations below.

Sales tax revenues are applied to ticket sales, concessions, and novelties. New revenues diminish if those paying for tickets (or getting them for free) don't attend, since there'd be no spending on concessions or novelties. They also diminish if tickets are given away or significantly discounted.

That suggests that current attendance, even if reported as slightly higher than the average in Zimbalist's report, produces lower tax revenues because so many tickets are essentially distributed free.

Zimbalist's numbers

So, how did Zimbalist calculate new New York tax revenues? He assumed that 30% of current New Jersey fans of the Nets would also attend games in Brooklyn.

There are three flaws in his assumptions:
1) the current total attendance
2) the calculation of current Nets fans from New Jersey
3) the calculation of how many of those fans would go to Brooklyn.

For now, though, let's look only at the second issue--how to estimate the number of current New Jersey fans.

Part of that calculation was simple. Of current season-ticket holders, 67.9% live in New Jersey, with "the large majority of the remaining holders live in New York, with a small proportion living in Connecticut and even smaller share in Pennsylvania," according to his report.

What about the rest of the fans? Zimbalist could have simply extrapolated from the season-ticket numbers to approximate the home states of those attending on a game-by-game basis, then adjusting the New Jersey percentage upward to account for the fact that those making game-day attendance decisions likely live closer to the arena.

Instead, his formula lowered the number of Nets fans from New Jersey. Lacking Nets data on the home state of fans who are not season-ticket holders, he instead used data regarding fans who attend New York Jets games at the Meadowlands: 51% from New Jersey, 44.7% from New York and 4.3% from Connecticut.

So, using those proportions for the balance of Nets fans, his blended average of New Jersey attendance at the Meadowlands arena is 60.5%, down from the 67.9% of season-ticket holders.

Basketball vs. football

In their May 2004 critique of Zimbalist's report, Jung Kim and Gustav Peebles demolished his assumptions regarding attendance:
Any knowledgeable local sports fan knows that the Jets, due to their glory years at Shea Stadium and the historical territorial split between Jets and Giants fans, have a large chunk of their fan base in Queens and Long Island. To boot, we all know that football fans must travel to New Jersey to see the game live, which is hardly true for basketball.

Also, I'd add, the far more infrequent football schedule and weekend game days make it more likely fans would decide to make a relatively long trip.

Kim and Peebles noted that the decrease in New Jersey attendance from 67.9% (season-ticket holders) to 60.5% has significant consequences for Zimbalist’s calculations. They wrote:
If, as we assume, there are more NJ-based fans who have to make the above decision, then this places a greater burden on NY-based fans in terms of fan attendance. In revenue terms, a greater share will be diverted from existing sales taxable expenditure in New York. Following Dr. Zimbalist’s other assumptions, we find that 55.9% of spending at the Atlantic Yards arena would be new to the NY economy, compared to his figure of 59.4%.

Looking at the numbers

I think the attendance percentage would end up even lower, but first, let's look at the calculations under both sets of assumptions, as noted in the chart below.

Even if there were a larger base in New Jersey, leading to slightly more fans coming from New Jersey, under the Kim/Peebles calculations, the smaller number of existing New York fans means there'd have to be a larger number of additional New York-based fans to fill up the Brooklyn arena.

Remember, the new New York fans--the largest segment of attendees--would mostly be re-channeling entertainment dollars, so the larger that number, the lower the new revenues. Only a fraction of the revenue from current New York-based fans would represent new spending.

(Click to enlarge; note that the percentages are very slightly different from the text above.)


20% lower attendance

What happens if we apply the above assumptions to a 20% lower attendance? The need for additional new fans from New York goes up and, consequently, the amount of new spending would go down, to 52.4% under the Zimbalist assumptions and 49.5% under the more credible Kim/Peebles assumptions.

(The chart below and subsequent charts are my adaptations of an original Kim/Peebles spreadsheet.)


If fewer New Jersey fans come to Brooklyn

But we shouldn't stop there.

Why should we assume, as did Zimbalist, that 30% of New Jersey fans of the Nets would also attend games in Brooklyn? Though Kim/Peebles accepted that assumption, at this point, the long goodbye may have soured a larger number of New Jersey fans. That means a smaller percentage of the New Jersey fan base would come to Brooklyn.

Let's adjust that assumption to 20% of current New Jersey fans. If applied to the 2004 attendance figures, the percentage of new spending goes down, from 59.5% to 55.3% under the Zimbalist assumptions, or from 56% to 51.3% under the Kim/Peebles calculations.


But we're not done yet. Let's apply the assumption to a reduction in current attendance of 20%. That means fewer New Jersey-based fans would come to Brooklyn, and to fill the arena the team would have to draw even more New York-based fans--not 8708 under the original Zimbalist calculations but perhaps 11,672--who currently don't go to Nets games.


New spending down 23.2%?

What's the implication of these numbers? It would bring new spending down from 59.5%, under the original Zimbalist calculations, to 45.7%, under my calculations, which assume:
--a lower attendance base
--a larger number of New Jersey-based fans
--a smaller percentage of those fans coming to Brooklyn
--a larger number of new fans who already live in New York.

The reduction from 59.5% to 45.7% represents a decline of 23.2%, a figure that should be applied to expected sales tax revenues.

New revenue calculations

What does that number mean in practice?

Zimbalist started us along the way;
When I alter the assumption that 30 percent of current Nets attendees from New Jersey also attend games at the Atlantic Yards arena, the following results obtain. When the share is lowered to 25 percent, new sales tax revenues fall from $6.43 million in 2008 to $6.26 million, or a decrease of 2.6 percent. When the assumed share is raised to 35, the sales tax revenues grow to $6.62 million in 2008, or an increase of 2.9 percent.

If we assume that only 20 percent of Nets fans from New Jersey come to Brooklyn, the projected 2008 sales tax revenues fall to $6.08 million.

Looking more closely

Now things get a little more complicated. First, remember that Zimbalist's calculation applied to likely inaccurate assumptions regarding the New Jersey fan base, so it should be recalculated, with the projected revenues lowered (which I'll do below).

Under Zimbalist's original calculation, a reduction from $6.43 million to $6.08 million represents a cut of 5.44%. However, the charts above suggest a reduction from 59.5% of new spending to 55.3% of new spending--a 7.05% drop. So the numbers are obviously inexact.

Consider that the final chart shows new spending at 45.7%. That implies a 23.2% decline from the 59.5% figure in Zimbalist's report. However, if we extrapolate from the difference between the numbers in the above paragraph, the reduction in revenue may be slightly less than 23.2%, perhaps 20%,

Bottom line

But the bottom line is that an arena that was producing a modest surplus for the city under Zimbalist's calculations may well--after additional city contributions and more plausible assumptions regarding the home state of new fans--represent a loss.

The IBO, in its September 2005 Fiscal Brief on Atlantic Yards, calculated a net fiscal impact of $28.5 million for the city. That may already be a loss, given that the city's contribution of $100 million (Debt Service) has more than doubled.

More than 60% of the new tax revenues from arena operations would come from sales tax revenues, according to the IBO. That's nearly $70 million. Subtract 20% and that's another $14 million reduction in the net fiscal impact.

The role of the IBO

Am I sure of these calculations? No, but neither was Zimbalist--who at least, in footnotes, offered alternate numbers--nor was the IBO, which accepted Zimbalist's assumptions rather than provide multiple scenarios. The IBO used attendance figures of 14,765.

It also "used Forest City Ratner Companies' assumption that about half of those attending Nets games at the Atlantic Yards arena will be from the ranks of those attending now, based on data on recent attendance levels at Nets games. Assuming that 20 percent of the spending by the other half of the fans attending the Nets games will be new to New York City, IBO estimates that nearly 60 percent of Nets’ games sales revenue will be new to the city’s economy."

That's a rounding off of the 59.4% figure. But what if it were 45.7%, as noted in the above chart?

When the IBO does a recalculation, and it should, it should produce a range of results, given that differing assumptions can result in very different predictions.

New spending beyond current fans

To further explain Zimbalist's calculations--incorporated by Kim/Peebles and the IBO--note that they include an adjustment for new revenues from new fans:
The money they spend at the new Brooklyn arena will be largely recirculated within the New York economy, and for the most part will not represent new revenues.
However, some of these expenditures will be new either to the New York City or the New York State economy or both. The sources of this new money are the following. Some people from out of state (principally from New Jersey and Connecticut) will be new Nets fans. They will be attracted either to the new Frank Gehry-designed arena, to new players on the team or to the team itself. Second, other attendees will attend Nets games as an add-on to their leisure expenditures. Primarily, these individuals will be from upper income brackets who do not need to reduce other leisure-time expenditures in order to be able to afford Nets games. Third, others may attend Nets games and reduce out-of-town leisure spending. Fourth, some corporations may purchase premium seating and catering services as an add-on to their entertainment budgets. Fifth, some of the spending at the Atlantic Yards arena will come from fans in Nassau County, Suffolk County, or Westchester County who did not attend games at CAA. Together these three counties have a population of 3.74 million. When these fans spend money at the new Atlantic Yards arena on tickets, concessions, or novelties, it will bring new sales tax revenue to New York City (though not to New York State.)

Another reason for doubt

Remember, Zimbalist's report had another huge fundamental flaw, especially regarding revenues from non-basketball events in Brooklyn. He wrote:
Many of the numbers used in this report concerning Nets attendance, ticket prices, construction costs and other items come from projections done by or for FCRC. I have discussed these estimates with FCRC and they seem reasonable to me. FCRC projects that the arena will not host an NHL team and that it will host 224 events during the year (assuming the eventual closing of CAA, no new arena in Newark, no NHL and no minor league hockey events at the Atlantic Yards arena.)


While the CAA (Continental Airlines Arena, now Izod Center) may or may not close, a new arena in Newark opened in 2007. As Kim and Peebles pointed out in their report, his assumption was inherently flawed because it gave no place for the New Jersey Devils to play.

The Empire State Development Corporation, in responses to a comment about the closing of the New Jersey arena in the Atlantic Yards Final Environmental Impact Statement, simply punted, claiming that estimates of events at the Brooklyn arena--though the same number as in Zimbalist's report--came from a different source.

But when the IBO does its recalculation, it should factor in the presence of a competing arena in New Jersey.

No comments:

Post a Comment