Congressional Budget Office paper contradicts Moody's model on effectiveness of different stimulus measures

The January 2008 CBO white paper Options for Responding to Shortterm Economic Weakness was written by five economists and attempted to examine the effects of different kinds of government stimulus measures (chiefly, tax cuts vs spending in terms of GDP impact). Unlike the current Moody's model being bandied around by stimulus supporters (and most highly touted by Dr. Mark Zandi), the CBO report does not attempt to place a direct dollar multiplier figure on different types of stimulus, instead measuring them in terms of Large, Medium, and Small impact.

Nonetheless, the similarity of the categories used with those of the Moody's model allows for a rough comparison.

So let's take a look at how different stimulus strategies shake out in the two sets of analyses.

The CBO rates the cost-effectiveness of various stimulus strategies thus [in no particular order between the items in each category]:

Large impact
Lump Sum Tax Rebates
Tax Withholding Holiday
Extending or Expanding Unemployment Benefits
Temporarily extending Food Stamp benefits

Medium impact
Extending the AMT patch
Incentives for new investment
General aid to State governments

Small impact
Across the board tax rate cuts
Deferring or eliminating tax rate increases
Cut in corporate tax rates
Extending operating and loss carryback provisions
Investing in public works

Here are the bang for the buck multipliers according to Moody's, in descending order of effectiveness:

1.73 Temporary increase in Food Stamps
1.64 Expanding Unemployment Benefits
1.59 Increased infrastructure spending
1.36 General aid to State governments
1.29 Payroll tax holiday
1.26 Refundable lump sum tax rebate
1.03 Across the Board Tax cut
1.02 Non-refundable lump sum tax rebate
0.48 Extend AMT patch
0.37 Make dividend and capital gains cuts permanent
0.30 Cut in corporate tax rate
0.29 Make Bush tax cuts permanent
0.27 Accelerated depreciation

There are obvious significant points of agreement: increasing food stamps, unemployment benefits, payroll tax holidays, and refundable lump sum tax rebates score well in both models.

I have highlighted the two major points of disagreement. Moody's indicates that increased infrastructure spending and general aid to State governments are both high-multiplier strategies, but the CBO estimates that infrastructure/public works spending only has a Small impact and general aid to States only has a Medium impact.

The difference between a high-multiplier and Medium impact for State government may not be that significant because the CBO does not provide a comparative multiplier.

However, that difference in ranking infrastructure/public works spending is critical to any stimulus strategy.

Here's what Mark Zandi actually has to say about that robust multiplier for infrastructure spending, which is a lot more candid and honest than the people are being who simply reprint his numbers and march on talking about shovel-ready projects [my emphasis, obvious]:

Increased infrastructure spending is also a particularly effective way to stimulate the economy included in the House stimulus plan. The plan includes $160 billion in such spending over two years, with $90 billion in more traditional infrastructure such as highway construction, public transit and waterways; and $70 billion for a variety of energy, science and healthcare projects. The boost to GDP from every dollar spent on public infrastructure is large—an estimated $1.59—and there is little doubt that the nation has underinvested in infrastructure for some time, to the increasing detriment of the nation's long-term growth prospects.

The argument against including infrastructure spending as a part of any fiscal stimulus plan is that it takes substantial time for the funds to flow into the broader economy. Infrastructure projects can take years from planning to completion. Moreover, even if the funds are used to finance only projects that are well along in their planning—so-called shovel-ready projects—it is difficult to know just when projectswill get under way and when the money will be spent. These are reasonable concerns in most recessions, but the economy's current problems appear likely to continue for some time. It is also reasonable to be worried that this spending will be used on pork-barrel projects chosen not for political rather than economic reasons. To address this worry, policymakers plan to put in place tight controls to monitor the spending.

Notice that last phrase in bold, which suggests that the only reasons Zandi is doubling down on slow-impact infrastructure bang because the recession is going to be a long one.

Now contrast this to the CBO analysis--also quoted in full and with my emphasis added:

In addition to stimulating firms’ investment in plant and equipment, the government can invest in capital itself as a means of boosting demand. Federal, state, and local governments are responsible for large swaths of the economy’s capital stock, which includes ports, bridges, and roads. Those responsibilities also include various forms of reconstruction, such as in areas badly damaged by natural disasters. Proposals also exist for large-scale government investment in new technologies, such as new-generation power plants, facilities that produce alternative fuels, and automobiles that use alternative fuels.

Conceptually, spending on these kinds of projects seems to offer an appealing way to counteract an economic downturn and has the potential to enhance long-term economic growth. Because these projects are capital projects, their timing can be flexible. When demand is not sufficient to fully employ productive resources in the economy, a backlog of such projects is available that can employ workers and use capital. If those resources were indeed not being used fully, the social cost of the projects could be substantially reduced.

Practically speaking, however, public works involve long start-up lags. Large-scale construction projects of any type require years of planning and preparation. Even those that are “on the shelf” generally cannot be undertaken quickly enough to provide timely stimulus to the economy. For major infrastructure projects supported by the federal government, such as highway construction and activities of the Army Corps of Engineers, initial outlays usually total less than 25 percent of the funding provided in a given year. For large projects, the initial rate of spending can be significantly lower than 25 percent.

Some of the candidates for public works, such as grant-funded initiatives to develop alternative energy sources, are totally impractical for countercyclical policy, regardless of whatever other merits they may have. In general, many if not most of these projects could end up making the economic situation worse because they would stimulate the economy at the time that expansion was already well under way.

This comparison makes the point that I have been drumming home here, here, here, and here:

Over-reliance on a single Keynesian analysis, based on a proprietary model that is disputed in at least one major sector by CBO analysis, to justify hundreds of billions in stimulus spending is dangerous.

And the people who spend their time pasting and re-pasting Zandi's bang for the buck chart without the slightest understanding that it was not sent down from Mt Sinai carved in stone, are either ideologically blind, intellectually lazy, or simply don't know what the hell they are talking about.