SANDAG Can Still Deliver on TransNet Promises

Questions have been raised by Voice of San Diego about the San Diego Association of Governments’ ability to continue improving our region’s transportation system, as well as the agency’s conduct in sharing information with the public about its efforts.

Before answering those specific questions, a little background is important. SANDAG – which is overseen by a board of elected officials from all of the region’s 18 cities and the county – administers our region’s voter-approved TransNet half-cent sales tax for transportation.

Using this local source of revenue since 1988, SANDAG and its partner agencies have successfully attracted state and federal matching funds to complete more than 650 projects. And in the first eight years of the 40-year TransNet Extension program, the agency has finished or started construction on 40 percent of the projects promised to voters.

Nevertheless, as we continue to grow, there are increasing needs in the region. Recognizing this, the SANDAG board placed Measure A on the November ballot seeking an additional half-cent sales tax to fund transit, open space, highway, bike/pedestrian and local infrastructure improvements. Fifty-eight percent of voters supported the measure, but it fell short of the two-thirds majority needed to pass.

Starting a few weeks before the election, Voice raised a series of questions related to the TransNet program and about how much revenue Measure A would likely generate. Here, fundamentally, are the three primary questions and my responses to them:

Has SANDAG overestimated the revenue it is likely to receive from the existing TransNet program?

Maybe, but we won’t know for sure until 2048.

SANDAG uses a complex computer model to forecast population, jobs and housing growth. This model looks out over long periods of time and is tied to state Department of Finance estimates. The model then uses a multi-step process to forecast taxable retail sales in the future, factoring a large number of inputs into a series of equations. Small changes in any of the inputs may have little impact in the short run, but can potentially multiply into bigger effects as the model projects out over decades.

The taxable retail sales number generated by this model is one element that is used later in a budgeting process to estimate future TransNet revenue.

SANDAG staff recognized more than a year ago that the model’s taxable retail sales estimates appeared aggressive, but it was not immediately clear why. Nor was it clear what, if any, impact that might have on the TransNet revenue estimates. Staff started the initial work of updating the model. They planned to address the taxable retail sales issue as part of a larger effort to refresh the model, which is budgeted this fiscal year.

VOSD raised questions about the TransNet revenue forecast in late October as part of its election coverage of Measure A. Rather than wait, SANDAG staff dug into the issue, focusing their efforts on the taxable retail sales number that came from the computer model. After a month of work, staff concluded that the model overestimated taxable retail sales. Those overestimates will be corrected as part of the update to the model.

In the meantime, an independent methodology using a consensus of three national forecasts was developed so that it could be used in budgeting for the TransNet program while the SANDAG model is being updated. This more conservative forecast estimated that TransNet will bring in approximately $17.3 billion between now and 2048, roughly $3.3 billion less than previously forecast (in year-of-collection dollars).

This new forecast is lower than the previous one used in SANDAG’s budgeting for the TransNet program, but – by the very nature of forecasting – it is difficult to conclude that it will turn out to be more accurate in the long run. Fundamentally, the new forecast results in a 3.9 percent average annual growth rate in TransNet revenue, a half a percent lower than the previous forecast. If sales tax revenue grows annually at just one-tenth of 1 percent faster than the new forecast predicts, the region would collect an additional $250 million in revenue over the remainder of the program.

To add perspective, if we looked independently at each of the three national forecasts used to create the consensus forecast, TransNet revenue estimates would range from $16.1 billion to $20.1 billion.

Can SANDAG still complete the projects it promised to voters back in 2004 when they extended TransNet for another 40 years?

Yes, if SANDAG continues to attract federal and state matching funds at a similar rate as it has in the past.

The agency’s ability to complete projects depends on much more than just TransNet, which is a single revenue source in a very large capital improvement program. Project costs, the availability of matching funds and the state of the economy also are enormous factors.

The increasing price tags for the projects may be the biggest challenge. As planners and engineers have looked more closely at the projects, estimated costs have increased. For example, State Route 78 will cost more to widen than initially expected because we now know it will be necessary to replace the bridges that pass over the highway. As a result of these recent updates to cost estimates, the expected total to complete the projects included in the TransNet extension have increased by about $8.4 billion (from $19.4 billion to $27.8 billion).

Overall, SANDAG needs to attract approximately $3 in outside funds for every $1 generated locally by TransNet to complete the program over the coming 30 years. That ratio is similar to the track record SANDAG has with the TransNet program over the past 30 years.

To be exact, the ratio needed is 3.4-to-1. That is, it will be if all the projections we’re relying on – population, inflation, costs, revenue estimates, etc. – over the next three decades turn out to be correct. With changes to any of these factors – such as the increased federal investment in infrastructure being discussed by the incoming administration – that ratio could change significantly.

Had SANDAG concluded before the election that there were problems with its model that could have resulted in the $18 billion Measure A revenue forecast being overestimated?

No – before the election, SANDAG technical staff had not discovered how, or if, the agency’s computer model could have caused an overestimation of the Measure A revenue forecast.

SANDAG staff prepared a revenue forecast for Measure A more than a year before the election, estimating that it would bring in $18 billion over 40 years (in constant 2015 dollars).

After the revenue forecast was prepared, staff recognized in late 2015 that taxable retail sales forecasts produced by the agency’s computer model seemed aggressive. Staff decided to address the issue as part of a planned update to the model.

Over the past five years, SANDAG has been working on a program to update its computer modeling capabilities. The next model slated to be updated – the Demographic and Economic Forecasting Model, used to help forecast population, jobs, housing and economic changes – was generating the estimate of taxable retail sales. SANDAG staff did some initial work to start updating this model. And, going forward, the agency planned to continue to work on a full update, including the source code, data, and econometric equations that drive the model. As part of that larger process, staff planned to address the model’s taxable retail sales estimates.

A few weeks before the election, Voice of San Diego raised the question of whether the sales tax revenue forecasts for the existing TransNet program and for the proposed Measure A sales tax were too high. SANDAG technical staff dug into the agency’s very complex computer model sooner than planned. After a month of intensive investigation, it was discovered that the growth rates of economic output for certain employment sectors used in the model were aggregated incorrectly from the source data, causing overestimations of taxable retail sales. Those taxable retail sales figures were used in revenue forecasting for TransNet and Measure A. By this time, the general election was over.

In the wake of the election, SANDAG board members asked for a status report on the existing TransNet extension program. In December, staff presented an update, including what has been accomplished to date, as well as updated cost estimates for the projects yet to be completed, a review of the many ups and downs in sales tax revenue and a more conservative sales tax revenue forecast based on the consensus of three national forecasts.

The bottom line question: What are the next steps, and can SANDAG deliver the projects in the TransNet extension ordinance?

SANDAG intends to put into place additional quality control of its forecasting process. Even with an updated model and additional quality assurance, the agency will continue to reassess its revenue forecast every year and make adjustments for both the short-term and long-term financial outlook.

It is through this disciplined approach that SANDAG has been able to deliver projects year after year. I remain confident that many opportunities will present themselves over the remaining 32 years in the TransNet program to bring in matching funds. SANDAG has a proven track record of being able to take advantage of those opportunities and deliver on its promises to voters in the San Diego region.

San Diego County Supervisor Ron Roberts is chairman of the SANDAG Board of Directors.