Counting Jobs Created or Saved by the "Stimulus" Bill, Part 5:
On the first anniversary
17 Feb 2010 in Regulatory Economics
Yesterday the White House released the Administration's first annual report on the "stimulus bill" (the American Recovery and Reinvestment Act of 2009, or "ARRA"). The Administration and its critics are sparring over how many jobs the bill "created or saved."
Where do the numbers come from?
We've blogged several times now on the estimation and reporting of the employment effects of ARRA. Here's quick refresher, which will help to keep the first year report in perspective:
- Part 1, November 20, 2009: We discussed the scandal that had erupted over errors in the data reported on Recovery.Gov, the Administration's portal for disseminating results. Errors ranged from assigning codes for nonexistent congressional districts to reporting as jobs "saved" instances in which recipients used ARRA funds to give raises.
Objectivity and integrity are fundamental information quality principles. Estimates of jobs "saved" or "created" that are not objective are inherently untrustworthy, which undermines public respect for and confidence in the program.
Recovery.Gov is operated by the Recovery Accountability and Transparency Board, a 12-member group of agency inspectors general. We reported that the Chairman Earl Devaney had testified that the Board had no responsibility for ensuring the accuracy of information it reported.
That responsibility had been delegated to funding agencies, and through them to recipients, who were directed to report the numbers of jobs "created" or "saved." The Office of Management and Budget published guidance recipients were supposed to use to make these determinations. OMB's "guidance" didn't provide any guidance, however. It said "the relevant party conducting a data quality review required by this Guidance (i.e., recipients, sub-recipients, Federal agencies) must use its discretion in determining the optimal method for detecting and correcting material omissions or significant reporting errors." In other words, anything goes.
Meanwhile, Recovery.Gov commits wholesale violations of the federal Information Quality Act. The Board publicizes figures for jobs saved or created that have no analytic merit. The Board has no procedures for receiving error correction petitions from the public and no procedure for reviewing and responding to those petitions.
- Part 2, November 24, 2009: The Board is reported as having no plans to remedy errors in the data reported or summarized on Recovery.Gov.
Some recipients want to correct reporting errors. We discussed why the data input scheme for Recovery.org is designed to impair error correction. The Board's defense is that recipients are responsible for the accuracy of their submissions, and funding agencies are responsible for reviewing them. Thus, the Board believes that it has no responsibility for "ensuring and maximizing" the quality of data it disseminates, as the Information Quality Act requires. The White House, which also is subject to the Information Quality Act, promotes he Board's figures without caveat.
- Part 3, January 23, 2010: White House Council of Economic Advisors chairman Christina Romer claims that ARRA "has created or saved between 1.7 million and 2 million jobs." We looked closely at the methodology used to produce this estimate. We learned that it is not actually an estimate at all, but rather a calculation derived from a certain assumptions concerning the employment effects of government spending. The greater the spending, the higher these figures would be.
In short, Romer's figures are modeling projections based on her January 2009 model. They are not estimates of what has actually happened under ARRA.
- Part 4, February 1, 2009: The Administration issued a quarterly report on ARRA that made similar claims. What had changed is the OMB changed its guidance concerning how to report jobs "created" or "saved."
OMB abandoned this effort entirely. Its new guidance directs recipients to report the number of jobs funded by ARRA.
WHAT DOES THE FIRST ANNIVERSARY REPORT SAY?
Growth in Gross Domestic Product (GDP)
The Report (PDF) cites in Figure 1 (p. 1) quarterly rates of growth in gross domestic product (GDP) for 2008:Q4 through 2009:Q4. The reported rate of GDP growth in the fourth quarter is 5.7%. The implied reference for these data is a CEA report dated January 13, 2010. However, the CEA report says "[t]he only official GDP data that we have under the program are for the second and third quarters of 2009" (p. 6).
The actual reference for these data is not reported, but it can be ascertained to be the Commerce Department's Bureau of Economic Analysis. On January 29, BEA reported an "advance" estimate of 5.7% for GDP growth in the 4th quarter of 2009. BEA attributed this change as follows:
The increase in real GDP in the fourth quarter primarily reflected positive contributions from
private inventory investment, exports, and personal consumption expenditures (PCE). Imports, which are a subtraction in the calculation of GDP, increased.
The acceleration in real GDP in the fourth quarter primarily reflected an acceleration in private inventory investment, a deceleration in imports, and an upturn in nonresidential fixed investment that were partly offset by decelerations in federal government spending and in PCE.
Motor vehicle output added 0.61 percentage point to the fourth-quarter change in real GDP after adding 1.45 percentage points to the third-quarter change. Final sales of computers subtracted 0.03 percentage point from the fourth-quarter change in real GDP after subtracting 0.08 percentage point from the third-quarter change.
Federal spending during the 4th quarter was substantial:
Real federal government consumption expenditures and gross investment increased 0.1 percent in the fourth quarter, compared with an increase of 8.0 percent in the third. National defense decreased 3.5 percent, in contrast to an increase of 8.4 percent. Nondefense increased 8.1 percent, compared with an increase of 7.0 percent. Real state and local government consumption expenditures and gross investment decreased 0.3 percent, compared with a decrease of 0.6 percent.
It's important to remember that GDP captures all private and public spending. Thus, every dollar of additional federal spending adds a dollar to GDP.
In a technical note, BEA attempts to inform the public about how ARRA shows up in the National Income and Product Accounts, and explain why it is so difficult to pinpoint its effects:
American Recovery and Reinvestment Act of 2009
BEA's national accounts include the effects of the federal outlays and tax cuts included in the American Recovery and Reinvestment Act of 2009. Because most of the outlays and tax reductions from ARRA during the last four quarters were in the form of grants to state and local governments, tax reductions for individual and business taxpayers, and social benefits to households, their effects on GDP show up indirectly through the effects on GDP components such as consumer spending, residential investment, and state and local government spending. Thus, BEA’s accounts do not directly identify the portion of GDP expenditures that is funded by ARRA. It is possible, however, to identify the effects of ARRA on components of income.
During the fourth quarter, ARRA provisions lowered personal taxes about $85 billion (annual rate) and raised government social benefits to persons about $69 billion, thus raising disposable personal income about $154 billion. Corporate taxes were also lowered about $20 billion. ARRA funded about $80 billion in current grants to state and local governments (such as Medicaid and education grants), about $25 billion in capital transfers to state and local governments and home buyers, and about $3 billion in subsidies for housing and energy. Further information, including estimates of the effects of ARRA for the first three quarters of 2009, is available at http://www.bea.gov/recovery/index.htm.
BEA also provides tables showing its best estimates of ARRA's effects on government sector transactions and how ARRA funds are allocated through the National Income and Product Accounts. More than 75% of the total consisted of transfer payments (pie chart).
Jobs Created or Saved
Employment effects are more interesting because increasing employment was the stated purpose of ARRA.
The Report cites in Figure 4 (p. 3) the same figures previously published by the CEA, which were derived by calculation from its pre-ARRA model. Unlike the GDP data mentioned above, the Report clearly identifies the source.
The Report gives a map (Figure 5, p. 4) showing state-by-state employment gains. These gains are proportional to population. As reported by CEA (pp. 20-21):
Of course, simply because their populations are larger, we estimate that larger states have seen larger jobs impacts. Similarly, because their employment is more cyclically sensitive, industrial states are estimated to have had larger employment effects relative to their populations. Finally, both because of their industrial composition and because state fiscal relief and aid to individuals directly impacted have been larger in states hit harder by the recession, we estimate that states with higher unemployment rates at the time of passage have seen larger employment effects of the ARRA relative to their populations.
In short, in the CEA model employment effects are essentially proportional to government spending, which under ARRA has a high correlation with state-level population.
Where Did the Money Go?
The Report includes a helpful pie chart (Figure 9) showing the top 10 federal programs in which spending has been supplemented by ARRA. We've reproduced the figure below:
The total amount of additional spending on these 10 programs was $182.6 billion. Subsidies for health care ($54.3 b) or food stamps ($8.5 b), pension payments ($13.9 b), and student loans ($8.7 b) have value, but not for saving or creating jobs.
Payments to States ($36.9 b) and K-12 education ($10.0 b + $8.7 b) substituted for State contributions to these joint federal/State programs. They prevented layoffs of State and local government employees (if budgets otherwise would have been cut) or increases in State and local taxes (if service levels were kept constant).
When these two groups are added they total 77% of spending -- essentially the same percentage the BEA had allocated to transfer payments.
Only two of the 10 program areas experienced spending increases that might have resulted in private sector job growth: highway infrastructure investments ($24 b) and energy efficiency ($14.1 b). In both cases, a substantial fraction of these amounts had to go toward materials (e.g., concrete, steel, machinery). Supplying these materials has only indirect employment effects, and only if production had to be ramped up instead of being provided through underutilized capital or inventory.
For these reasons, the employment effects claimed in the Report (2 million jobs as of 4th quarter 2009) seem highly unlikely to have actually materialized. The Labor Department's Bureau of Labor Statistics reports that 8.4 million jobs have been lost since the recession began in December 2007:
Following steep job losses toward the end of 2008 and in the first quarter of 2009, job losses began to moderate. Over the past 3 months, nonfarm payroll employment has shown little net change.
