Six-Year Income Tax Revenue Forecast FY

Transcription

1 Six-Year Income Tax Revenue Forecast FY Prepared for the Prepared by the Economics Center February

2 TABLE OF CONTENTS EXECUTIVE SUMMARY... i INTRODUCTION... 1 Tax Revenue Trends... 1 AGGREGATE FORECAST... 5 COMPONENTS OF INCOME TAX COLLECTIONS... 9 Payroll Withholding... 9 Business Income Individual Non-Withholding POLICY CONSIDERATIONS Tax Credits to Employers Job Creation and Property Investment Reimbursement Policy and Process Change Implications for Net Income Tax Revenues and Future Forecasts Ohio House Bill ECONOMIC CONDITIONS Regional Employment Characteristics National and Regional Economic Outlook SUMMARY BIBLIOGRAPHY APPENDIX: METHOD AND APPROACH Autoregressive Integrated Moving Average (ARIMA) Forecast

3 Vector Error Correction Model (VECM) Forecast Comparison with Prior Forecast Industry Groupings

4 EXECUTIVE SUMMARY To assist the City of Cincinnati in planning, the Economics Center has completed an updated six-year forecast of net income tax revenues and components. Estimates for the net income tax forecast for the six-year period of FY FY 2022 appear in Table A. Aggregate net income tax revenues are projected to grow at 2.3 percent annually on average from FY 2017 through FY 2022, slightly lower than the average annual rate from 2003 to While any individual year may deviate from this annual average, realized net revenue growth may be slightly lower or higher, on average the year-over-year growth is anticipated to approximate this annual average compound rate. Table A: Net Income Tax Revenue Forecast Fiscal Year Income Tax Net Growth Net Operating US Inflation Real Growth rate Revenues ($M) Rate Income ($M) Forecast 2017 $ % -1.1% $ % 2018 $ % 2.2% $ % 2019 $ % 0.7% $ % 2020 $ % 0.0% $ % 2021 $ % -0.9% $ % 2022 $ % -2.4% $ % Average 2.3% * -0.1% * 2.4% * Source: Economics Center model results and calculations. US inflation projection calculated from NECC forecast. *compound average annual rate Baseline forecast estimates are an average constructed from two statistical models used to produce the forecast. The 95-percent confidence intervals estimated for the baseline forecast produce upper and lower bound net revenue amounts. The bounds provide the range within which baseline revenues are estimated to fall with 95-percent probability. The range increases in later years, indicative of the additional uncertainty in the forecast for longer time horizons. Table B: Forecast Upper and Lower Bounds, FY 2017-FY 2022 Fiscal Year Total ($M) Lower Bound Upper Bound 2017 $345.0 $ $332.2 $ $340.8 $ $348.0 $ $352.5 $ $353.3 $484.0 Source: Economics Center model results and calculations. i

5 Business income tax revenues exhibit the greatest volatility among the sources of income tax revenues. Business Income tax revenues are anticipated to grow 2.9 percent annually on average during the forecast period. Individual non-withholding revenues exhibit considerable seasonal patterns, with April manifesting as the predominant month when these revenues are received. The average annual growth rate for these revenues is anticipated to be 0.9 percent. An internal policy change, converting Property Investment Reimbursement Agreements (PIRAs) to Job Creation Tax Credits (JCTC) lowers future forecasted net revenue values by shifting the funding of these incentives from an appropriated departmental expenditure to a tax expenditure deducted directly from income tax payroll withholding. o o o o These economic incentives add to the employment tax base through the portion of payroll taxes for retained and new jobs that accrue to the City. Net income tax revenues have displayed historical growth in the presence of these credits and incentives, about 3 percent on average annually. As a tax expenditure, the incentive payments reduce net income tax revenues compared to the baseline total estimated. Although the overall City budget is unaffected as departmental expenditures experience a commensurate reduction. Baseline forecast estimates for FY 2018 FY 2022 were adjusted to reflect these refundable JCTC credits. Growth in anticipated payments refunded out of net tax revenues yields lower growth in net revenues, and a slower growth rate than in the absence of refunding these payments directly out of income tax revenues. This dynamic is illustrated in Figure A. ii

6 Millions ($) $450.0 $430.0 $410.0 $390.0 $370.0 $350.0 $330.0 $310.0 $290.0 $270.0 $250.0 Figure A: Baseline and Adjusted (Final) Net Tax Revenue Forecasts Fiscal Year (*forecast) PIRA Conversion Adjusted (Final) Total Baseline Total Source: Economics Center model results and calculations Increased use of economic development incentives introduces greater volatility into future net tax revenues, for several reasons such as: o o o contracts are negotiated on an individual basis and as such do not exhibit a consistent or predictable pattern; refunds occur with approximately a two-year lag following job creation, thus payroll withholdings will increase in the first year but a refund must be anticipated in a future fiscal year; employers do not always request the negotiated refund as actual job creation may differ from anticipated at the time of the agreement. Consistent with existing national and regional forecasts, the current revenue forecast projects growth but lower than the previous forecast from a year ago. Analysis of employment and industry data illustrate the diverse commercial base of the local economy. Education and Health Services, Professional and Business Services, Trade, Transportation and Utilities, Financial Activities and Manufacturing are industry sectors of note. Consistent with national projections, local Manufacturing employment is expected to decline. Education and Health Services and Financial Activities employment are both expected to grow nationally and locally. Data on local employment projects declines in Trade, Transportation, and Utilities employment as well as Business and Professional Services in contrast with national projections. iii

7 INTRODUCTION To assist the City of Cincinnati in planning and projections, the Economics Center estimated a six-year forecast of net income tax revenues and major components for the period of fiscal year (FY) 2017 to FY In 2013 the City of Cincinnati adjusted its accounting schedule from calendar year to a July-June fiscal year. All estimates presented in this report are for a July-June fiscal year, including historical data. To prepare the forecast, the Economics Center analyzed historical tax revenue trends for the City of Cincinnati, as well as information on tax incentive payments. This information was combined with data on national economic trends to develop the forecast projections. The Economics Center also examined data on regional industry employment and wages for the City. This analysis identified significant industry sectors influencing the local economy and illustrates the benefits to the local economy from its relatively diverse commercial base. While the city s economy is subject to national trends affecting its most significant industries, the diverse commercial base allows declines in one sector to be offset by growth in another. Tax Revenue Trends For FY 2016, the City of Cincinnati received nearly $371 million in net income tax revenues from all sources. FY 2016 exhibited a 4 percent nominal increase in revenues over the prior year, a 3.2 percent real increase after adjusting for inflation. The pattern of net income tax revenues exhibits varying growth rates, with years of high growth clustered together generally followed by a few years of slower growth. The real growth rate also illustrates volatility in net income tax revenues overtime. Historical revenue trends are displayed in Table 1. 1

8 Table1: Net income Tax Revenues, FY 2004-FY 2016 * Fiscal Year Total Net Revenues ($M) Growth Rate Real Growth Rate 2004 $ % -1.8% 2005 $ % 3.7% 2006 $ % 2.3% 2007 $ % 1.8% 2008 $ % 0.7% 2009 $ % -4.9% 2010 $ % -4.0% 2011 $ % 3.8% 2012 $ % -1.6% 2013 $ % 2.2% 2014 $ % 0.6% 2015 $ % 4.4% 2016 $ % 3.2% Source: Economics Center calculations *In the second half of FY2016 a change in the collections process accelerated the receipt of $7.2 million in withholding revenues. To better reflect actual annual dynamics and account for this idiosyncratic process change, these collections have been omitted from the analysis. The long run average growth rate of net revenues is approximately 3 percent and has displayed little change since While historically the real growth rate of net revenues has been closer to zero on average, the post-recession local economy has displayed some gains. Recent growth in real net revenues has increased the long run average rate to about 2 percent overall. Figure 1 displays the nominal and real growth rates of net income tax revenues. 2

9 Figure 1: Nominal and Real Growth Rates of Net Income Tax Revenues, FY 2004-FY 2016 Percent Change 8.0% 6.0% 4.0% 2.0% 0.0% -2.0% -4.0% -6.0% Fiscal Year Growth Rate Linear (Growth Rate) Source: Economics Center model results and calculations Real Growth Rate Linear (Real Growth Rate) Historically, payroll withholding is the largest component of income tax collections, generating about 80 percent of net revenues. Individual non-withholding revenues have consistently accounted for approximately 5 percent of total net income tax revenues. Business income tax revenues have displayed the greatest volatility since FY 2004, accounting for as little as 8 percent of total net revenues to almost 16 percent. Generally, increases in the business earnings tax revenues coincide with declines in payroll withholdings. Shares of the components of income tax revenues are displayed in Table 2. 3

10 Table 2: Components of Net Income Tax Revenues, FY 2004-FY 2016 *,** Fiscal Year Individual Non- Payroll Business Withholding Withholding % 9.1% 85.9% % 8.3% 86.8% % 11.3% 83.0% % 12.4% 82.4% % 14.2% 80.1% % 15.7% 79.4% % 13.1% 81.9% % 11.3% 83.6% % 13.2% 81.4% % 11.9% 82.9% % 12.2% 82.6% % 13.5% 81.3% % 14.8% 80.0% % 13.3% 81.1% Source: Economics Center calculations *Individual non-withholding and Payroll Withholding net revenues are adjusted for employee refunds. Prior series deducted these refunds from the Individual component; however, these refunds are made from Payroll Withholding. ** Components may not sum to 100% due to rounding. Analysis of monthly receipts illustrates consistent seasonal patterns for some of the components of income tax revenues. Figure 2 displays the monthly seasonal factors, a comparison of average monthly to annual collections. The seasonal factor indicates the extent to which net collections within a specific month tend to deviate from a simple average of annual net collections. 1 Seasonal variation depends on when filings and revenues are received and when refund disbursements occur. Not surprisingly, individual non-withholding revenues display the strongest seasonal fluctuations with the majority of revenues realized in April. Business income tax revenues exhibits a weak quarterly pattern. Payroll withholding displays the greatest stability across months, with slight peaks occurring in January, April, and October. Seasonal patterns affect monthly cash flows, and may impact calculation of annual totals if receipts are accelerated or delayed due to process changes for example. 1 To calculate seasonal factors the average monthly collections over the period (e.g. average of April totals) were divided by the average annual collection for the period. 4

11 Figure 2: Monthly Net Revenue Patterns Seasonal Factor Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Individual Non-Withholding Payroll Withholding Business Net Income Source: Economics Center calculations AGGREGATE FORECAST To forecast net income tax revenues for the six-year horizon, FY 2017 through FY 2022, the Economics Center analyzed historical patterns for total net revenues and its components. Consistent with the prior forecast, two estimation methodologies were used an autoregressive integrated moving average (ARIMA) model and a vector error correction model (VECM). While the ARIMA model utilizes only the historical data on revenues, the VECM model also incorporates information regarding dynamics of the national economy. Also consistent with the prior forecast, the initial estimate which forecasts the second half of FY 2017 is from the ARIMA model. The remaining forecast years are a consensus estimate of both the VECM and the ARIMA models. 2 The ARIMA and VECM estimates form the baseline net tax revenue forecast. The Economics Center then accounted for potential refunds out of payroll withholdings due to the conversion of the Property Investment Reimbursement Agreements (PIRAs) to a Job Creation Tax Credit (JCTC) for the years FY FY The PIRAconversion adjusted estimates constitute the final forecast. Total net revenue estimates, along with upper and lower bounds constructed for 95-percent confidence intervals, and the forecast growth rates are presented in Table 3. Forecasted net revenues for the current fiscal year are estimated at $371 million, growth of 0.1 percent over the previous year. 2 Two refinements from the prior forecast to improve the estimates include accounting for seasonal fluctuations within the VECM model and an average estimate for all years after the initial fiscal year. See the appendix for more detail. 3 PIRA conversion is discussed in greater detail in a subsequent section. 5

12 After accounting for potential future PIRA-converted JCTC refunds, annual projected revenues are forecasted to increase modestly, reaching about $416 million in FY While the actual growth rate varies year to year, the average annual expected nominal growth rate is 2.3 percent. Table3: Forecasted Net Income Tax Revenues and Growth Rates, FY 2017-FY 2022 Total ($M) Forecast Fiscal Year Forecast Lower Bound Upper Bound Real Growth US Growth Rate Inflation Rate 2017 $371.2 $345.0 $ % -1.1% 1.3% 2018 $387.7 $332.2 $ % 2.2% 2.2% 2019 $400.0 $340.8 $ % 0.7% 2.5% 2020 $409.6 $348.0 $ % 0.0% 2.4% 2021 $415.8 $352.5 $ % -0.9% 2.5% 2022 $415.9 $353.3 $ % -2.4% 2.5% Average 2.3% * -0.1% * 2.4% * Source: Economics Center model results and calculations. US inflation projection calculated from NECC forecast. *compound annual average rate 6

13 Figure 3 displays the forecast value along with the upper and lower bounds. Net income tax revenues are forecasted to increase at a slightly lower rate than historically observed. The range for the forecast values (upper and lower bounds) are $332 to $454 million in FY 2018 and $353 to $484 million in FY The upper and lower bounds are formed from the 95-percent confidence intervals, calculated from the baseline forecast. Figure 3: Net Revenue Forecast with Upper and Lower Bounds, ($M) $500.0 $400.0 Millions ($) $300.0 $200.0 $100.0 $- Fiscal Year (*forecast) Source: Economics Center model results and calculations Upper Bound Forecast Lower Bound 7

14 Nominal average annual growth over the forecast period is 2.3 percent, slower than the historical rate of change of 2.9 percent due to the anticipated PIRA-converted JCTC refunds. As noted in Figure 4, the anticipated future growth rate follows a pattern similar to past revenues with slower growth forecasted for later periods. Figure 4: Net Revenue Forecast, Total ($M) and Growth Rate $ % Millions ($) $400.0 $300.0 $200.0 $ % 4.0% 2.0% 0.0% -2.0% Percent Change $ % Fiscal Year (*forecast) Total Net Revenues ($M) Source: Economics Center model results and calculations Growth Rate Approximately 74 percent of net income tax revenues accrue to the City s Operating Fund (1.55 percentage points), with the remainder of revenues distributed among Infrastructure (0.10 percentage points), Public Transit (0.3 percentage points), and Permanent Improvements (0.15 percentage points) funds. The forecasted estimates for net income tax revenues result in approximately $286 million for the operating fund in FY 2018, with a lower and upper bound of $245 to $335 million respectively. Table 4: Forecasted Net Operating Revenues ($M), FY 2017-FY 2022 Fiscal Year Forecast Lower Bound Upper Bound 2017 $274.0 $254.6 $ $286.2 $245.2 $ $295.3 $251.6 $ $302.3 $256.9 $ $306.9 $260.2 $ $307.0 $260.8 $357.2 Source: Economics Center model results and calculations 8

15 COMPONENTS OF INCOME TAX COLLECTIONS Income tax for the City consists of an equal 2.1 percent of income generated from three sources payroll withholding, business profit or income, and individual non-withholding. Each of these series displays slightly different dynamics. As payroll withholding is the largest component of income tax revenues its patterns tend to dominate the patterns of the total; however, the dynamics of all components influence total revenues. As noted previously, payroll withholdings tend to move inversely with business income. Analysis of each income tax revenue source illustrates that Payroll and Individual Non-Withholdings have the strongest seasonal components. These sources of income tax revenue are more sensitive to changes in national employment, while business income tends to be more sensitive to national gross domestic product (GDP). To produce the forecast, net payroll withholdings and total revenues were jointly forecasted in the VECM, and individually analyzed using the ARIMA model. Both series were seasonally adjusted. A residual (the sum of net business income and individual non-withholding revenues) was calculated from the forecasts for payroll withholdings and total revenues. Historical shares of these components were then analyzed to estimate future business income and individual non-withholding net revenues. To account for the conversion of Property Investment Reimbursement Agreements (PIRAs) to Job Creation Tax Credits (JCTC), an estimate of future JCTC refunds has been deducted from forecasted payroll withholding net revenues. Discussed in greater detail in a subsequent section, the internal policy change results in refunds being issued from payroll withholdings whereas formerly these incentives were paid as an expenditure out of a departmental budget for the City. Payroll Withholding Forecast estimates for payroll withholdings exhibit moderate growth through most of the forecast period. Payroll withholdings in the city have grown at an average annual rate of 2.4 percent historically. Future revenues are forecast to grow at a slightly lower rate of 2.3 percent annually from FY 2017 to FY The slower growth rate for payroll withholding net revenues are due largely to the conversion of PIRAs to JCTCs. Table 5: Payroll Withholding Net Revenues Fiscal Year Net Revenues ($M) Share of Total Growth Rate 2017 $ % 1.0% 2018 $ % 0.2% 2019 $ % 3.0% 2020 $ % 1.8% 2021 $ % 1.1% 2022 $ % 5.5% Average 2.3% * Source: Economics Center model results and calculations *compound average annual rate 9

16 Figure 5: Payroll Withholding Net Revenues $360.0 $340.0 $320.0 Millions ($) $300.0 $280.0 $260.0 $240.0 $220.0 $200.0 Fiscal Year (*forecast) Source: Economics Center model results and calculations Business Income Business income tax revenues exhibit less consistency annually than payroll withholding revenues. These revenues have grown at an average annual rate of about 5.9 percent between 2003 and 2016, although with considerable variability. The projected annual growth rate through the forecast period displays similar volatility as observed historically, although with a lower growth rate in the forecast period. A compound average annual growth rate of approximately 2.9 percent is projected for the period FY 2017 through FY Table 6: Business Income Net Revenues Fiscal Year Net Revenues ($M) Share of Total Growth Rate 2017 $ % -2.9% 2018 $ % 24.4% 2019 $ % 4.4% 2020 $ % 5.1% 2021 $ % 3.6% 2022 $ % -18.4% Average 2.9% * Source: Economics Center model results and calculations *compound average annual rate 10

17 $78.0 Figure 6: Business Income Net Revenues $68.0 $58.0 Millions ($) $48.0 $38.0 $28.0 $18.0 $8.0 Fiscal Year (*forecast) Source: Economics Center model results and calculations Individual Non-Withholding Similar to business income net revenues individual non-withholding exhibits considerable variability annually. The post-recession period has seen slower growth than prior. From 2003 to 2007, individual non-withholding revenues grew at an average annual rate of 7.8 percent. In the recovery period from 2011 to 2016, the average annual growth rate was about half, or 4 percent. The projected compound average annual growth rate for this component is 0.9 percent per year for the forecast period. In other words, on average the annual growth rate of individual non-withholding revenues is anticipated to be just under 1 percent; however, the actual growth rate may be higher or lower in any given year. Table 7: Individual Non-Withholding Net Revenues Fiscal Year Net Revenues ($M) Share of Total Growth Rate 2017 $ % -4.8% 2018 $ % 21.9% 2019 $ % 2.4% 2020 $ % 3.1% 2021 $ % 1.7% 2022 $ % -20.0% Average 0.9% * Source: Economics Center model results and calculations *compound average annual rate 11

18 Millions ($) $28.0 $26.0 $24.0 $22.0 $20.0 $18.0 $16.0 $14.0 $12.0 $10.0 $8.0 Figure 7: Individual Non-Withholding Net Revenues Fiscal Year (*forecast) Source: Economics Center model results and calculations POLICY CONSIDERATIONS Tax Credits to Employers Job Creation and Property Investment Reimbursement Policy and Process Change Recently the Property Investment Reimbursement Agreements (PIRA) have been converted to Job Creation Tax (JCTC). As a result of this change, refunds associated with the JCTC are drawn from payroll withholding whereas PIRAs previously were issued as an expenditure in a departmental budget. The City may offer a JCTC to a company as an incentive to expand in or move to Cincinnati, based on certain eligibility criteria. The company then commits to maintain a level of retained and new employment for an agreed upon term. The length of the JCTC and value are negotiated on a per case basis. The value of the JCTC is based on a percentage of payroll withholding taxes from the new employment created. This value is then credited back to the company. In practice the JCTC is refunded out of the City s payroll withholding revenues but in principle functions as a reduction in a business earnings tax obligation. 4, Credit.aspx

19 Implications for Net Income Tax Revenues and Future Forecasts To the extent that credits offered do not equal the total value of a company s payroll withholding, these credits contribute to net income tax revenues for the City through maintaining or expanding the City s employment base. On average, an examination of 48 PIRAs and JCTCs negotiated since 2000 indicate that agreements have rebated, or will refund, about 57 percent of the value of payroll withholdings to the qualifying companies. In other words, the City receives about 43 percent, on average, of gross income tax revenues associated with new and/or retained employment commitments under the JCTC agreements. Assuming this employment and the associated income tax revenue would never have occurred within the City but for the JCTC program, in this way such incentives contribute to the tax revenue base. Historically net income tax revenues have grown on average 3 percent annually in the presence of these incentive agreements. As tax credits are negotiated on a case by case basis, it is difficult to model their potential future impact. Data on JCTCs and PIRAs indicate an irregular pattern overtime, with some years containing only a single agreement particularly prior to (calendar year) Following 2009, incentive agreements are more common, with up to seven agreements occurring in a year. Rates and terms of agreements also vary widely from year to year, as would be expected with contracts tailored to specific employer characteristics. Particularly if high-growth segments of a company or local employment receive JCTCs, the forecast may overstate net revenues based on expected job growth. 6 Due to the nature of the JCTCs being negotiated on a case-by-case basis, this program does not follow a standard or predictable process, as they become more frequent greater uncertainty is introduced into the tax revenue forecast. In the absence of refunding these incentives directly out of income tax revenues, net revenues accruing to the City can be expected to increase as employment increases by some proportion of jobs and wages. As the value of the JCTC is directly tied to the value of payroll withholding, net revenues realized are not as closely tied to employment within the City. Further complicating future projections, there is approximately a two-year lag from the time the credit is earned (when jobs are created) and when the refund is issued. The revenue will be received the first fiscal year when employment is created, but refunded in a future fiscal year. Also, not all employers file for the job credits or refunds even if eligible, or file for amounts different than the negotiated maximum depending on actual job creation numbers. In other words, now that these refunds will be issued out of income tax revenues directly, the future patterns of net revenues are more difficult to predict. Substantively, the change does not necessarily alter the overall City budget but converts what had been an explicit expenditure appropriation instead to a tax expenditure or revenue reduction. Thus, the change impacts the total amount of net revenues accruing to the City and the timing. Additionally, future income tax revenue forecasts will be subject to greater volatility and uncertainty. To account for the potential impact of the PIRA to JCTC conversion, estimates of expected JCTC credits 6 White (2017) 13

20 have been deducted from the baseline forecast for annual withholding net revenues. Credits to be refunded during the forecast period are estimated based on data from the City. Table 8 displays the estimated future refunds used to adjust the net revenue forecast. The City estimates between $5.7 and almost $9 million in refunds committed through existing contracts for the FY 2019 through FY Based on these estimates, expected credits exhibit a 24 percent average annual growth rate. Estimates for FY 2021 and FY 2022 assume that future contracts and credits will grow at the same 24 percent compound rate. If more or larger agreements are negotiated, then the amount of future payments will be greater further lowering future net income tax revenues. If fewer or smaller contracts are added as existing contracts end, then future refunds will be lower, thus increasing net income tax revenues. Table 8: Estimated Annual Value of Future Refunded Credits * Fiscal Year Estimated Refund 2018 $ 5,760, $ 6,987, $ 8,865, $ 10,999, $ 13,645,950 Source: City of Cincinnati (FY FY 2020), and Economics Center calculations (FY 2021 FY 2022). *PIRA conversions are already accounted for in FY 2017 data, thus no estimate or adjustment is made for FY As illustrated in Figure 8, estimated converted PIRA payments will lower net income tax revenues as they constitute an anticipated increase in future refunds, for both payroll withholding and total net revenues. If future incentive payments grow, then both the level and the growth rate of net income tax revenues will be lower than without these refunds. The baseline forecast illustrates anticipated net revenues before accounting for these anticipated incentive refunds. As noted, the gap between the baseline and the PIRA conversion adjusted (final) forecasts grows over time indicative of a slower growth rate in forecasted net income tax revenues under the PIRA conversion. As the refunds constitute a larger share of payroll withholdings (relative to total net revenues), the incentive refunds have a proportionally greater impact on the forecasted trend of realized net payroll withholding tax revenues. 14

21 Figure 8: Baseline and Adjusted (Final) Net Income Tax Revenue Forecasts $450.0 $400.0 Millions ($) $350.0 $300.0 $250.0 $200.0 Fiscal Year (*forecast) PIRA Conversion Adjusted (Final) Total Baseline Total PIRA Conversion Adjusted (Final) Withholding Baseline Withholding Source: Economics Center model results and calculations Ohio House Bill 5 Effective January 1, 2016 House Bill 5 (HB 5) modified certain aspects on all components of income tax revenues. 7 In particularly, HB 5: o o o impacts individual non-withholding revenues through limiting deductions of employee business expenses and extending tax exemption for non-resident income earned; changes how business calculate net operating losses (NOL) for deductions, as well as provides more flexibility in allocation methods and filing; requires businesses remitting employee withholdings to do so twice a year if withholdings exceed $12,000 annually. Interest and penalties have also changed, and penalties are capped for late filings. Calculations of deductions and exemptions and changes to allocation methods may impact the amount refunds and net revenues received by the City. Changes to required timing of filings may impact the timing of revenue receipt. With greater flexibility and options available to income tax filers, more volatility may be introduced into net revenues complicating future forecasts

22 ECONOMIC CONDITIONS Regional Employment Characteristics Employment and wage data for the city illustrate key industries which are main drivers of earnings, and thus income tax revenues, locally. National conditions for these industries are likely to affect local performance. Employment data for major industries indicate (calendar year) 2015 may represent a peak year in terms of recent growth rates, experiencing larger than average employment and wage gains. According to employer data from the U.S. Bureau of Labor Statistics, total employment within the city has grown 5.3 percent from Q to Q Examination of first quarter employment of 2013 to 2016 illustrates major employment groups relevant to the city s economy. The largest employment group is Education and Health Services which comprised more than one-quarter of total employment in quarter one of Professional and Business Services accounted for the next largest segment of city employment at 19 percent. Trade, Transportation, and Utilities completes the top three major sectors and accounted for 14 percent of employment within the city. 9 8 For the analysis city employers were identified by applying three filters. First employer data was filtered for establishments located within zip codes contained within the City wholly or in part. Once zip codes were identified, observations were screened a second time to include locations indicating Cincinnati or a neighborhood within the City (e.g. Clifton or Walnut Hills). Finally, once city-based employers were identified, observations with multiple establishments reporting employment and wage data across all locations were excluded. 9 See the appendix for descriptions of the industry groupings. 16

23 Table 9: City Employment by Major Industry Group Sector 2013 Q Q Q Q1 Professional and Business Services 51,950 51,365 53,243 51,913 Trade, Transportation, and Utilities 37,787 35,821 38,455 37,725 Education and Health Services 74,054 69,239 74,551 74,608 Total for All Industries 258, , , ,220 Source: Economics Center calculations from ES-202 data In addition to accounting for almost 60 percent of all employment, these industries also generated 60 percent of total wages and earnings in the first quarter of Professional and Business Services contributed the largest share to local wages at nearly 27 percent. Educational and Health Services accounted for just under one-quarter of wages while Trade, Transportation, and Utilities contributed approximately 10 percent. Table 10: City Total Wages by Major Industry Group ($M) Sector 2013 Q Q Q Q1 Professional and Business Services $1,035 $1,060 $1,129 $1,091 Trade, Transportation, and Utilities $412 $381 $453 $429 Education and Health Services $905 $864 $956 $933 Total for All Industries $3,732 $3,740 $4,089 $4,094 Source: Economics Center calculations from ES-202 data Recent employment and wage data illustrate 2015 as experiencing an above average growth rate. In general, employment has grown about 1.7 percent per year in the city, while wages have increased by approximately 3.1 percent annually between 2013 and However, 2015 displays 5.3 percent annual growth in employment over the same quarter in 2014, while total wages displayed 9.3 percent growth. First quarter data from 2016 indicate slower growth in both employment and wages. Analysis of employment and wage data also illustrates two additional industry groupings significant for the local economy Financial Activities and Manufacturing. Financial Activities is not among the largest industries, accounting for about 9 percent of total city employment, but has experienced an annual growth rate above the all-industry average, 4.5 percent, between the first quarters of 2013 and While this industry contains about one-third of the employment compared to the city s largest sector (Education and Health Services), Financial Activities is associated with 14 percent of total earnings, nearly half that generated by Education and Health Services. In other words, although this industry is relatively small, these jobs are high wage and have exhibited above average growth. Additionally, Manufacturing is significant in that both employment and wages have been growing from 2013 to Within the period examined, Manufacturing employment experienced an above average growth rate of 5.4 percent. Similar with other major industry trends for the city, much of the growth in manufacturing employment occurred in 2014 and 2015 with the rate slowing substantially in Wages have also grown substantially in Manufacturing over the period, about 11.7 percent, substantially faster than the all-industry average for the city. 17

24 Table 11: City Employment and Total Wages for Financial Activities and Manufacturing Sector 2013 Q Q Q Q1 Financial Activities Employment 21,810 22,013 22,410 24,909 Total Wages ($M) $515 $531 $554 $589 Manufacturing Employment 15,776 16,898 18,239 18,476 Total Wages ($M) $263 $284 $373 $366 Source: Economics Center calculations from ES-202 data Data from Economic Modeling Specialists, Inc. (EMSI) provides employment projections for local industries. 10 The data indicate continued growth in Education and Health Services and Financial Activities employment locally. Manufacturing and Trade, Transportation and Utilities employment are expected to see the greatest declines 8.2 and 6.7 percent respectively. Professional and Business Services employment is also expected to decline locally, by one-percent. Table 12: City Employment Projections, * Sector Percent Change Professional and Business Services 52,998 52, % Trade, Transportation, and Utilities 29,051 27, % Education and Health Services 61,507 65, % Financial Activities 21,171 22, % Manufacturing 18,805 17, % Total for All Industries 261, , % Source: Economics Center calculations from EMSI Q estimates including QCEW, non-qcew, and Self- Employment estimates. *Employment estimates are for the calendar year. National and Regional Economic Outlook National economic dynamics influence the performance of the local economy. National projections for the key sectors previously discussed anticipate continued slowed growth in Manufacturing and accelerated growth in Education and Health Services. Growth in Manufacturing real GDP nationally is anticipated to be strong through 2024, followed by Financial Activities. As illustrated in Figures 9 and 10, the two highest value producing industries nationally are associated with the slowest employment growth, or declines, with future projections. 10 For data from EMSI city employment is defined as employment contained within one of the 26 zip codes contained within the city wholly or in part. EMSI estimates for small geographies are based on a proprietary methodology and algorithm calculated from standard publicly available sources. 18

25 The groupings most significant to the local economy in terms of employment and wages Educational and Health Services and Professional and Business Services are expected to see greater growth in employment nationally, but account for less economic output as measured by real GDP. Assuming the city economy follows national trends, gains in Education and Health Services and Financial Activities sectors, along with gains to Manufacturing output, may offset any losses from Manufacturing employment. Employment and output in Trade, Transportation and Utilities and Professional and Business services is anticipated to grow nationally in contrast to local employment data. Thousands of Jobs 30,000 25,000 20,000 15,000 10,000 5,000 Figure 9: National Employment by Industry Grouping 0 Manufacturing Financial Activities Professional and Business Services Education and Health Services Trade, Transportation, and Utilities Source: U.S. Bureau of Labor Statistics, Employment Projections 19

26 Figure 10: National GDP by Industry Grouping (Chained 2009$) 7,000 6,000 5,000 Billions ($) 4,000 3,000 2,000 1,000 0 Manufacturing Financial Activities Professional & Business Services Source: U.S. Bureau of Labor Statistics, Employment Projections Educational and Health Services Trade, Transportation, and Utilities Two key components explicit for the local forecast include national gross domestic product (GDP) and national nonfarm employment. As noted in the most recent release from the National Economic Estimating Conference (NECC), national indicators are still projecting growth in coming quarters, although at a slower pace for some than originally forecasted in December of 2015, particularly towards the end of the forecast period. For example, while employment and output projections are still strong, forecasts are slightly lower and slower than estimated last year. 11 Figures 11 and 12 illustrate changes in short run national economy projections between forecasts developed in December 2015 and November National Economic Estimating Conference. 20

27 Billions ($) Figure 11: National Gross Domestic Product Short Run Forecasts 21, , , , , , ,000.0 November 2016 Forecast December 2015 Forecast Source: National Economic Estimating Conference Figure 12: National Total Nonfarm Payroll Employment Short Run Forecasts Millions November 2016 Forecast December 2015 Forecast Source: National Economic Estimating Conference 21

28 These slowing, but still positive, economic forecasts are consistent with recent trends as well as other existing state and national forecasts. In general, the weaker outlook is due to lower consumer spending and slower growth in the housing market than anticipated. Higher inflation and lower consumer spending contribute to moderate growth projections for the nation. 12 In January of 2017, the Ohio Tax Conference also released its Economic Outlook, citing positive but moderating economic growth. Inflation and inflation expectations began rising. According to the forecast, Ohio s economy has generally followed the nation in terms of performance. Recent trends illustrate that the state has experienced slower job growth than the U.S. as a whole. 13 For the state and nation modest growth is anticipated for the near term. The forecast anticipates that Ohio s economy will continue to lag the nation and that slower job growth will also mean slower GDP growth. 14 The forecast also notes anticipated slow growth in domestic manufacturing. Consistent with the Ohio Tax Conference forecast, the Cleveland Federal Reserve Bank in its most recent Beige Book entry 15 also notes that declining manufacturing employment in the region has offset gains in other industries recently, despite economic growth overall. Additionally, the overview notes that within Freight Transportation payrolls and hiring remain flat. These dynamics may affect the outlook and future performance for Manufacturing and Trade, Transportation, and Utilities employment in the city. SUMMARY The current six-year forecast for net income tax revenues for the City displays continued growth through FY Consistent with regional and national forecasts which anticipate slowed growth relative to a year ago, net income tax revenue projections for the City are slightly lower than those produced in Revenue projections estimate net income tax revenue of $371 million for FY 2018, increasing to $416 million by FY Much of the growth is driven by Payroll Withholding, the largest source of income tax revenue. Payroll Withholding net revenues are forecasted to be $304 million in FY 2018, growing to $340 million by FY Individual Non-Withholding Revenues and Business income revenues are also projected to grow, but exhibit more volatility than payroll withholding. The conversion of PIRAs to JCTCs alters the manner in which these incentives enter into the City budget. As PIRAs, incentive payments were treated as an explicit expenditure in a departmental budget. With the conversion to JCTCs, incentive payments constitute refunds out of net income tax revenues. As such, net revenues will be lower relative to the baseline forecast without making these payments. Increased use of economic development incentives introduces greater volatility into the City budget, complicating future forecasts. Legislative changes that provide income tax filers with 12 NECC (2016) 13 Berson (2016). 14 Berson (2016) 15 Fourth Federal Reserve District (November 2016) 22

29 greater flexibility may also impact the amount and timing of net revenues for the City, and be another source of volatility. The local economy is largely characterized by employment in Education and Health Services and Professional and Business Services. Employment in these industries is expected to grow nationally over the forecast period. Output for these industries is also anticipated to grow, but the value of output is low compared to other sectors such as Manufacturing. Manufacturing and Financial Activities are also significant industry sectors for the local economy. These high output industries are expected to experience slower employment growth, or declines in the case of Manufacturing, nationally. Due to a relatively diverse commercial base locally, growth in some sectors may offset declines in another. 23

30 BIBLIOGRAPHY Berson, D.W. Economic Outlook for 2016 and beyond for Ohio & the U.S. Presented at the Ohio Tax Conference. January City of Cincinnati. Hamilton County. Single Audit for the Year Ended June 30, Economics Center. Five Year Income Tax Forecast. Prepared for the City of Cincinnati Florida Office of Economic and Demographic Research. National Economic Estimating Conference. November Florida Office of Economic and Demographic Research. National Economic Estimating Conference. July 2016 Florida Office of Economic and Demographic Research. National Economic Estimating Conference. December Federal Reserve Bank of Cleveland. Fourth District Summary of Commentary on Current Economic Conditions (Beige Book). November White, D. "The Economy's Expanding. So Why Aren't Tax Revenues?" Governing. January Creation-Tax-Credit.aspx

31 APPENDIX: METHOD AND APPROACH Consistent with prior forecasts the Economics Center employed two specifications to develop the current forecast. Technical aspects of these specifications are provided in greater detail below. Analyses were conducted on the net total revenues and net payroll withholding revenues. Monthly net revenue estimates from January 2002 through December 2016 are the primary data for analysis. Potential seasonality within the monthly data was taken into account in each approach. For the Vector Error Correction Model (VECM) forecast, data were seasonally adjusted manually and forecasted estimated then re-adjusted for seasonal affects to produce a series consistent with historical variations. The impact of macroeconomic dynamics associated with the national economy was included in the VECM in the form of nominal national GDP and total nonfarm payroll employment. Historical data obtained were monthly, however forecasted values derived from the National Economic Estimating Conference (NECC) were quarterly. To integrate the forecast data into the series, each was transformed using a 3-period moving average (MA(3)). Each of these series was then differenced (once for GDP and twice for employment) to account for non-stationarity in the levels. Statistical analysis of monthly net income tax revenues indicated the series did not exhibit non-stationarity. Thus, the analysis considered these series in levels. Autoregressive Integrated Moving Average (ARIMA) Forecast An ARIMA model is a regression model that utilizes a moving average of the dependent variable along with other components to predict future values of the independent variable. The Economics Center used the X-13 ARIMA procedure developed by US Census Bureau. This particular model allows a user not only to run an ARIMA specification, but to do so using a seasonal adjustment technique created by the Census Bureau. The model includes indicators of the seasons, which pick up on effects such as a large portion of non-withholding filings arriving during April each year. Ultimately, the following model was used: The dependent variable net income tax collections was tested for non-stationarity. 16 Statistical tests indicated the monthly series do not exhibit unit roots, thus the analysis was conducted on the levels of the variables. Vector Error Correction Model (VECM) Forecast A Vector Autoregression (VAR) is a statistical model that attempts to capture interdependencies of time series variables. The contemporaneous change in each variable is determined using a 16 Regular growth in the value of the variable over time. 25

32 mathematical equation that incorporates lags (previous values) of the variable of interest as well as lags of other variables in the model. Model diagnostics (Johansen test) indicated that variables within the data set were cointegrated of the first order. 17 VAR models do not account for cointegration, thus the model specification was revised to include an error correction term. A Vector Error Correction Model (VECM) is a variation of a VAR model that includes an error correction term to account for cointegration between the variables. The error correction term ( ) allows a statistical model to take into account how one variable may influence another over time. In this regression, GDP and Cincinnati income tax revenues are intricately related. This implies that a rise in GDP may cause a simultaneous rise in Cincinnati income tac revenues. The error term will account for this by incorporating a lag of the residual values, or by subtracting the estimated error due to cointegration from the previous period. Because the regression was performed with monthly data, a period of one year s worth of previous values (12 lags) of the dependent variable was used. Intuitively, one year of previous data on tax collections should be enough to capture any anomalies not accounted for in explanatory variables. Ultimately, the following regression formula was used: Where and are vectors of the variables discussed previously. Comparison with Prior Forecast In 2016, the Economics Center conducted a five-year income tax revenue forecast for the City. The current forecast estimates are comparable to the prior. 18 While the prior and current forecasts utilize the same basic approach and methodology, a few adjustments were made to the current forecast. As monthly net revenues were used for the model estimation in the current round, seasonal factors were incorporated into the VECM forecast. Both total revenues and payroll withholding were modeled using the ARIMA and VECM specifications for the current forecast. 19 Modeling both series permitted imputation of the residual net revenues the sum of individual non-withholding and business income. Forecasts for these remaining two components of income tax revenues were then 17 Multiple variables share underlying trends. 18 City budget estimates and projections have been adjusted to reflect slower growth than forecasted in the prior year. FY 2017 estimates in the current forecast are consistent with existing City projections. 19 Individual Non-Withholding and Business Income net revenue series did not permit the same level of statistical modeling due to non-normality in the error structures. 26