Reading the Independent Congressional Research Service Reports on the Stimulus So You Don’t Have To

[digg-reddit-me]According to Wikipedia, “the Congressional Research Service (CRS) is the public policy research arm of the United States Congress. As a legislative branch agency within the Library of Congress, CRS works exclusively and directly for Members of Congress, their Committees and staff on a confidential, nonpartisan basis. CRS reports are highly regarded as in-depth, accurate, objective, and timely, but as a matter of policy they are not made directly available to members of the public.”

Which makes the reports it has prepared on various stimulus measures extremely interesting – as their reports strive to give the consensus expert view of the issues involved, without partisan affiliation. I’ve prepared some highlights from the non-partisan (and confidental) CRS reports on the effectiveness of tax cuts and other stimulus measures (available thanks to Wikileaks…H/t Marc.)

CRS – RS21126 – Tax Cuts and Economic Stimulus: How Effective Are the Alternatives (PDF)

The summary:

While temporary individual tax cuts likely have smaller effects than permanent ones, temporary cuts contingent on spending (such as temporary investment subsidies or a sales tax holiday) are likely more effective than permanent cuts. (Sales tax holidays may, however, be very difficult to implement in a timely fashion).

The report offers a quick explanation of why tax cuts aren’t ideal short term stimulus on page 1:

A tax cut that is saved will have no short term stimulative economic effect (or long term one, if the cut is financed by a deficit, since increased private saving would be offset by decreased government saving). Thus, in general, tax cuts received by individuals will not be successful as a short run stimulus if they lead to additional saving, and tax cuts received by firms will not be successful unless they lead to spending on investment (or lead quickly to spending on consumption by shareholders).

The problem in this instance is that private virtue – saving money – undermines the public good stimulus seeks to achieve – stimulating the economy – which can only be achieved by spending money. The report discusses corporate tax cuts on page 5:

General corporate rate cuts are less likely to be effective than investment subsidies because they have a smaller “bang-for-the-buck.” because much of their cost is a windfall that only affects cash flow and not the return to new investment. Since even temporary investment subsidies do not appear to have worked effectively, a corporate rate cut would be expected to have a small effect.

CRS – Report R4104 – Economic Stimulus: Issues and Policies (PDF) (It appears the Wikileaks link is down, so try this.)

The summary:

Economists generally agree that spending proposals are somewhat more stimulative than tax cuts since part of a tax cut may be saved by the recipients. The most important determinant of the effect on the economy is the stimulus’ size. [my emphasis]

The report describes the cause of our current panic/crisis as the ripple effect of the sudden collapse of Lehman Brothers, the near miss at Merrill Lynch, and the government rescue of AIG – all following upon the failure of other large institutions over the months before hand – from Bear Stearns to Fannie Mae and Freddie Mac.

These actions eroded market confidence further, resulting in a sudden spike of the commercial paper rate spread from just under 90 basis points to 280 basis points, a spike that in times past might have been called a panic. If financial market confidence is not restored and private market spreads remain elevated, the broader economy could slow more due to difficulties in financing consumer durables, business investment, college education, and other big ticket items.

The report doesn’t get into telling the story of how the world economy almost came to an end at 2pm on September 18. One of the problems the report explains with stimulus is that we can only design it to be effective based on our economic forecasts. As the report explains on page 8, economic forecasting has its problems:

Economic forecasts are notoriously inaccurate due to the highly complex and changing nature of the economy, so it is difficult to accurately assess how deep the downturn will be, and how much fiscal stimulus would be an appropriate response.

But the report nevertheless ventures a guess, on page 2, based not on any particular forecaster, but only the consensus among them:

Forecasters now predict that GDP will continue to contract until the second half of 2009 and the rate of decline will accelerate.  If correct, this recession would be the longest in the period since World War II.

The report discusses three elements by which to judge the stimulus – how fast it works; how effective the stimulus is per dollar paid; and the size. Predicting the right size for the stimulus must primarily be based on the economic forecasts – which the report notes are “notoriously inaccurate.” But the first two measures should be maximized no matter what the forecast is. The report describes why effectiveness is important in a section called “Bang for the Buck” on page 8:

If the goal of stimulus is to maximize the boost to total spending while minimizing the increase in the budget deficit (in order to minimize the deleterious effects of “crowding out”), then maximum bang for the buck would be desirable. The primary way to achieve the most bang for the buck is by choosing policies that result in spending, not saving. Direct government spending on goods and services would therefore lead to the most bang for the buck since none of it would be saved…

One non-spending measures the report analyzes is seen to have a good “bang for the buck” but to take too long to act:

Investment incentives are attractive, if they work, because increasing investment does not trade off short term stimulus benefits for a reduction in capital formation, as do provisions stimulating consumption. Nevertheless, most evidence does not suggest these provisions work very well to induce short-term spending. This lack of effectiveness may occur because of planning lags or because stimulus is generally provided during economic slowdowns when excess capacity may already exist.

On page 10, the report cites Mark Zandi of Moody’s Economy.com, and an advisor to John McCain, and includes a chart of his estimates of the multiplier effect of the various policy proposals. The report qualifies it’s endorsement of Zandi’s work, saying, that there is significant disagreement about fundamental matters among economists, but that:

Qualitatively, most economists would likely agree with the general thrust of  [Zandi’s] estimates, however—spending provisions have higher multipliers becausetax cuts are partially saved, and some types of tax cuts are more likely to be saved by theirrecipients than others.

I’ve graphed the values Zandi provides to give a visual measure of the different in the types of stimulus:

[Click on the image for a larger version of the chart.]

As you can see, the spending measures have far greater “bang for the buck” values. The report acknowledges that Zandi’s and most other economic models might understate the stimulus from tax cuts – but the alternate explanation for the data given by defenders of tax cuts is that they take longer to have an effect:

there is a behavioral lag, since time elapses before the recipient of a transfer or tax cut increases their spending. For example, the initial reaction to the receipt of rebate checks was a large spike in the personal saving rate… It is unclear how to target recipients that would spend most quickly, although presumably liquidity-constrained households (i.e., those with limited access to credit) would spend more quickly than others. In this regard, the advantage to direct government spending is that there is no analogous lag.

These reports seems to provide a good deal of information mainly missing from the public debates – as Republicans talked about tax cuts, tax cuts, tax cuts as the best type of stimulus. Whether the stimulus bill we end up with works or not, these reports at least help explain the assumptions underlying it.