BHI FaxSheet: NJ 1994 and 1995 Cuts Led to 25,000 New Jobs

In 1994 and again in 1995, New Jersey cut its graduated personal income tax rates, which ranged from 2% - 7% in 1993, to 1.9% - 6.65% in 1994 and further to 1.7% - 5.68% by 1995. In 1997 the effective average state tax on personal income is 2.71%, but would have been 3.08% without the tax reductions.

According to an econometric analysis performed by the Beacon Hill Institute, the tax reductions have resulted in the creation of 25,000 additional jobs and $2.1 billion in new capital stock. These increases are over and above any changes that may have occurred as a result of the upswing in the business cycle.

Year	Cumulative Change in Payroll	Cumulative Change in Jobs	Cumulative Change in Capital Stock	Cumulative "Static" Tax Revenue Effect	Cumulative "Dynamic" Tax Revenue Effect	Cumulative Net Tax Revenue Effect
1994	$383 million	11,667	$0.9 billion	-$194 million	$10 million	-$183 million
1995	$814 million	23,997	$1.9 billion	-$469 million	$21 million	-$448 million
1996	$861 million	24,500	$2.0 billion	-$499 million	$23 million	-$476 million
1997	$929 million	25,017	$2.1 billion	-$539 million	$25 million	-$514 million

The tax cuts have boosted the annual payroll in New Jersey by $930 million and raised employment by 0.7%, helping bring the state’s unemployment rate down from 7.1% in January 1994, when it was well above the national average, to 5.2% in August 1997 (very close to the national average). With the exception of April of this year (when it was also 5.2%), this is the lowest rate in seven years.

The tax reductions produced a net revenue loss of about $514 million in 1997. This is the net result of two effects: (1) the “static” revenue loss of $539 million (the loss computed on the assumption that the tax cuts had no effect on the New Jersey economy) and (2) the “dynamic” revenue gain of $25 million (the revenue gained as a result of the increased production and personal income brought about by the tax cuts).

Despite the job growth New Jersey has enjoyed, critics of the tax cuts say that the new jobs being created have mostly been low-wage service sector jobs and that high-paying industrial jobs have decreased. It is true that most of the increase in jobs has been in the service sector, in line with the long-run trend in the structure of New Jersey’s economy. But it is not correct that the new jobs pay low wages, given that the average annual wage in the state rose from $32,815 in 1994 to $37,115 by 1997, significantly faster than the inflation rate. The BHI tax model confirms this: The relevant structural equation shows that a change in the state personal income tax rate does not alter the wage rate, implying that the jobs induced by a tax cut must pay at least as well as the average job does today.

______________________________________________________________________________

Two steps are needed in order to measure the effect of the tax cuts on the variables of interest – these variables being the number of jobs (L), the wage rate (w), the capital stock (K) and state tax revenue (TR). First we must establish baseline values for the variables, projecting them out through the end of 1997. Then we use the Beacon Hill Institute’s State Tax Analysis Modeling Program (STAMP) to estimate and project employment, wages and the capital stock in the absence of the tax cuts. These projections may then be compared with the baseline values in order to isolate the effects of the tax cuts.

The baseline projections are shown in the table below, and were constructed as follows. Information on the number of jobs is available for 1994 and 1995, and information on the number of workers is available through June 1997. We suppose that the growth rate of jobs is the same as the growth rate of the number of workers for 1996 (1.65%) and assume that the number of jobs in 1997 grows at the same rate as the number of workers did for the first half of 1997 relative to the first half of 1996 (1.67%).

The value of the total payroll is known for 1994-1996, and we assume that it will continue to grow (in nominal dollars) at the same speed as it did between the first quarter of 1996 and the first quarter of 1997, which is by 7.43% per year. We also assume that the capital stock grows at the same speed as the total payroll, so that the capital to labor ratio does not change. Since information on the capital stock in New Jersey is available for 1994, we use the historical growth rates of payroll to generate capital stock values for 1995 and 1996, and the projected growth of payroll to arrive at the capital stock for 1997. The average wage is simply the total value of payroll divided by the number of jobs.

A fuller description of the tax analysis model is given in An Application of the BHI Tax Analysis Model to the Massachusetts Mutual Fund Industry (Beacon Hill Institute, October 1996); the basic model has been modified and re-estimated so that it is applicable to New Jersey. The model is designed to allow the analyst to trace the effects of changes in state tax rates on employment, wages and the stock of capital. It begins with a series of structural equations that aim to capture the behavior of firms and households, and then re-arranges these equations to arrive at a set of reduced form equations which are both theoretically consistent and may be estimated econometrically. The equations are estimated using data from 1970-1994. In simplified form, the two that are relevant here are the labor equation and the capital equation.

The labor equation shows how changes in state tax rates (and other variables) affect the number of jobs in New Jersey. The estimated equation is of the form

where ln(L) is the natural log of the number of jobs, and t is the average marginal state tax rate on earned income (i.e. the marginal tax rate faced by New Jersey taxpayers, on average). The coefficient 0.0142 is statistically significant, and measures the effect of a change in the tax rate (t) on ln(L); in other words ln(L)/t = -0.0142.

The capital equation is similar, except that the dependent variable is capital rather than labor. Specifically the estimated equation is of the form

where ln(K) is the change in the natural log of the value of the capital stock from one year to the next. The coefficient -0.0173 is statistically significant.

In order to estimate the impact of the income tax cuts, we first have to compare the average marginal tax rate under the tax cuts, t_c, with the rate under the hypothetical case of no tax cuts, i.e., under the 1993 tax systems, t_h.

The latest data available for New Jersey tax returns by income group are for 1994. Using the 1994 data, we first calculate t_c,94. We compute the t that would have prevailed if there were no tax cut in 1994, t_h,94, by applying the 1993 tax schedule to the 1994 taxable income. The results are: t_c,94 = 3.490% and t_h,94 = 3.713%. Therefore, the 1994 tax cut reduces the t by 0.223 percentage points. Since state tax return data are not available for 1995, we use the 1995 data on gross income by income group obtained from the federal income tax returns. The estimated t_c,95 is 3.376% and t_h,95 is 3.831%, resulting in a reduction of 0.455 percentage points. Using a similar method, the reductions of t for 1996 and 1997 are estimated to be 0.457 percentage points and 0.459 percentage points respectively.

We project the employment that would not have been created in the absence of a tax cut. In 1994, t would have been 3.713% rather than 3.490%, so t = 0.223. Thus from equation (1) we have

The baseline value of ln(L) [=15.1212 = ln(3,690,289)] now falls by 0.00317 to 15.118; taking the antilog gives the number of jobs with the tax cut, which is 3,678,622, or an decrease of 11,667 under the baseline case.

An essentially similar procedure is applied for the three subsequent years. The detailed results are shown in the table below, and indicate that by 1997, an estimated 25,017 jobs have been created which would not have existed in the absence of the cuts in the state tax on earned income.

The effects of the tax cuts on the stock of capital are estimated in the same manner as for employment. Thus from equation (2) we have for 1994

The baseline value of ln(K) [=12.313 = ln($222,674 million)] now falls by 0.00398 to 12.309; taking the antilog gives a new capital stock, after the tax cut, of $221,790 million, or an increase of $884 million over the baseline case. Then, by subtracting this growth rate from the actual growth rate, we obtain the growth rate of K caused by factors other than tax cuts, which is then used to project the capital stock under no tax cut for 1995. This same approach is used for estimating 1996 and 1997 levels.

The estimated reduced form equations of the tax analysis model show that the state tax rate on unearned income does not have a statistically significant effect on the wage rate in New Jersey. Wages are thus assumed to follow the baseline projections both with, and without, the tax cut. The total value of the payroll, in the presence of the tax cuts, is calculated by multiplying these average wage rates by the number of people employed.

What effect did the cuts in the tax rate have on state tax revenue? First there is a “static” revenue loss, which is measured as the reduction in the average tax rate (ATR) times the tax base, which we take to be the baseline payroll. For instance, the static revenue loss in 1994 was $194 million (=0.160%*$121,098 million). But this overstates the true revenue loss, because there is also a “dynamic” revenue effect: The tax cuts lead to an increase in the number of jobs and hence the total payroll, and therefore to some offsetting increase in revenue. For instance, in 1994 the dynamic revenue gain was $10 million (=2.70%*$383 million).

According to BHI estimations, as of 1997, the number of new jobs created by the tax cuts was 25,017, raising employment by 0.7% above the level it would otherwise have achieved. The capital stock was $2.1 billion (almost 0.8%) higher than it would otherwise have been. Annual state tax revenue was $514 million lower than it would otherwise have been; this total consists of a static revenue loss of $539 million, offset by higher revenue of $25 million generated by taxes on the additional payroll resulting from the tax cuts.

Appendix Table: Effect of New Jersey Tax Cuts on Employment, Capital Stock and Tax Revenue

Notes: Slight rounding errors may occur. a: AMTR, Average Marginal Tax Rate, is the marginal state tax rate on earned income faced, on average, by taxpayers in New Jersey. b: ATR, Average Tax Rate, is total state tax revenue on earned income divided by earned income (payroll).

New Jersey's 1994 and 1995 Personal Income Tax Cuts led to 25,000 new jobs and $2.1 billion in new capital