Thursday, April 14, 2011
The Economic Effects of the Ryan Plan: Assuming the Answer?
Last week the House Budget Committee, chaired by Congressman Paul Ryan (R, WI) issued a Budget Resolution for fiscal year 2012 along with a Republican blueprint for addressing the nation’s long-term fiscal imbalance.[1] The “Ryan plan” would squeeze discretionary spending and health care entitlements very hard, lower marginal tax rates, and expand the tax base.
· We agree that addressing the nation’s long-term federal fiscal imbalance is critically important, and that doing so might head off an eventual fiscal crisis that could threaten our standard of living over the long haul.
· The Committee’s report included a simulation analysis showing the economy strengthening immediately as a result of the fiscal contraction; that is, a negative short-run fiscal multiplier.[2]
· We don’t believe this finding, which was generated by manipulating an econometric model that would not otherwise have produced the result.
· That analysis implied other questionable results — some of them probably unintended — including over $1 trillion of net new borrowing from abroad over the coming decade and the construction of several million unoccupied houses.
· We consider the analysis both flawed and contrived, and are concerned it will create the false impression among some legislators that implementation of the Budget Resolution would entail no short-run macroeconomic pain.
In this Macro Focus we outline some of our very strong reservations about the macroeconomic analysis underlying the Committee’s report. In a subsequent Focus we will present our own estimates of the macroeconomic impacts of this kind of deficit reduction.
Reasonable Results?
At the request of the House Budget Committee and using the Global Insights (GI) model of the US economy, analysts at the Center for Data Analysis at the Heritage Foundation simulated the results of implementing the Budget Resolution on top of a baseline reflecting CBO’s Alternative Fiscal Scenario.[3] A subset of the results released by Heritage is summarized in the table on page 9. Despite meaningful cuts in federal outlays, there is an immediate rise in real GDP which, over the next decade, grows to 2.2% of the baseline value. By 2021 the federal debt is reduced by nearly $10 trillion, and the yield on 10-year Treasury notes is down 84 basis points. This sounds pretty good, but how seriously should we take these results and the analysis that gave rise to them?
Believable Baseline?
To start, consider that the baseline shows the economy recovering cyclically to full employment by 2016, with real GDP growing at a trend-like rate of 2.4% thereafter, inflation averaging slightly more than 2.5%, and the yield on 10-year Treasury debt holding steady at 5.4%. While this seems reasonable, under CBO’s Alternative Fiscal Scenario the ratio of debt to GDP explodes. From 2016, when the economy first nears full employment, through 2021, the baseline shows federal debt held by the public rising 55% — or $9 trillion — all with no change whatsoever in long-term interest rates after 2017.
If the analysis shows that trimming $10 trillion of debt by 2021 reduces long-term rates by 84 basis points, then isn’t it the case that in the baseline the $9 trillion expansion of debt after 2016 should keep pushing rates further above 5.4% by 2021? If not, then what is preventing rates from rising?
To be clear, the model apparently was lined up using adjustment factors to produce the Alternative Fiscal Scenario under CBO intermediate-term economic assumptions.[4] These typically assume a return to full employment with constant inflation and average real interest rates no matter what fiscal assumptions are used. In that sense, the CBO baseline itself is not internally consistent. So, a legitimate question is: does the Budget Committee forecast that, under current policy, interest rates will actually stabilize over the second half of the coming decade?
Peek-a-boo
There were actually two sets of results. The first showed real GDP immediately rising by $33.7 billion in 2012 (or 0.2%) relative to the baseline, with total employment rising 831 thousand (or 0.6%) and the civilian unemployment rate falling a stunning 2 percentage points, a decline that persisted for a decade. (This path for the unemployment rate is labeled “First Result” in the table.) The decline in the unemployment rate was greeted — quite correctly, in our view — with widespread incredulity. Shortly thereafter, the initial results were withdrawn and replaced with a second set of results that made no mention of the unemployment rate, but not before we printed a hardcopy! (This is labeled “Second Result” in the table.)
Okun Overkill…?
To see how amiss the first result seemed to us suppose that our first inclination, which was to consider the reported decline in the unemployment rate as cyclical, had been correct. A typical “Okun’s law” relationship would show a 0.5 percentage-point decline in the unemployment rate for each 1 percentage-point decrease in the output gap. The results showed GDP increasing by 0.2 percent the first year, so a reasonable expectation is that the unemployment rate would fall by about 0.1 percentage point, not twenty times that![5]
…Or NAIRU Nonsense?
An alternative interpretation of the sudden and persistent drop in the unemployment rate was that it reflected an immediate decline of 2 percentage points or so in the Non-Accelerating-Inflation Rate of Unemployment (NAIRU, which was not reported) that quickly became reflected in the observed unemployment rate as well.
This did seem consistent with the result that inflation remained at the baseline path over the whole decade despite the much lower average unemployment rate initially simulated under the Budget Resolution. Furthermore, there is an argument that, by raising the relative reward for work, cutting transfers and/or the marginal tax rate on wages could lower the NAIRU by reducing search times in the labor market. However, it just seemed incredible to us that this (or any) fiscal plan would immediately reduce the NAIRU to a nonsensical value.
Labor Force Fiasco?
Whatever the explanation of the initially reported decline in the unemployment rate — and there’s more on that just below — the results were confusing. The simulation showed total nonfarm payroll employment rising by 831 thousand in 2012, with the unemployment rate falling 2 percentage points. One way to have produced this configuration of results is for the civilian labor force to have fallen sharply. For example, the MA forecast for the labor force in 2012 is 156 million, of which 2% is 3.1 million. Since the simulation showed payroll employment rising only 831 thousand, the drop in the unemployment rate could imply a decline in the labor force of 2.3 million, or an immediate 0.9 percentage-point decline in the labor force participation rate. This doesn’t square with the discussion of participation rates in the report.
“Labor Participation Rates. Taxes on labor affect labor-market incentives. Aggregate labor elasticity is a measure of the response of aggregate hours to changes in the after-tax wage. These are larger than estimated micro-labor elasticities because they involve not only the intensive margin (more or fewer hours), but also, and even more so, the extensive margin (expanding the labor force (MA italics)).”
But if the participation rate didn’t decline, and even rose, then it must be that household employment initially increased 2.3 million (or more) relative to total nonfarm establishment employment.
Ready Response!
Either way, something went awry here. In response, Heritage pulled the initial tabulations and replaced them with new ones that did not show or discuss the unemployment rate but left all the other results unchanged. Then, in a separate update, Heritage published a new path for the unemployment rate averaging 1.5 percentage points higher than the original path, along with the following explanation[6]:
“In further response to the Chairman’s letter of February 28, 2011 requesting technical advice and assistance, we have given additional scrutiny to calculations concerning the unemployment rate under the Chairman’s proposed budget plan. As a result of that examination, we are making an adjustment to one variable — the full-employment unemployment rate, which is one component of the equation for the overall unemployment rate.”
In other words, it was the case — nonsensical, in our view — that the NAIRU fell sharply in the first set of results.[7] We do know that in the GI model the NAIRU is an exogenous variable, so this is something that Heritage assumed. Why, and why the revised path of the NAIRU?[8]
In any event, the update goes on to say “while the adjustment has an impact on the unemployment rate in the model, the overall results elsewhere in the model do not change significantly.”[9] Of course it can’t be the case that no other variables changed. In particular, for a given population, if the unemployment rate rises for whatever reason, either household employment must be lower and/or the labor force participation rate must be higher. Is it really the case that in a general equilibrium model such large changes in labor force participation, household employment, and the NAIRU have no other significant impacts? We doubt it.[10] To us, these machinations call into serious question this part of the exercise.
Multiplier Mischief
The simulation shows real federal non-defense purchases down by $37.4 billion in 2012, but real GDP up by $33.7 billion, so the short-run “fiscal multiplier” is negative.[11] As noted above, that analysis was prepared using the GI model of the US economy. We are not intimately familiar with this model but have the impression it is a structural macro model in which near-term movements in GDP are governed by aggregate demand while long-term trends in output are determined by the labor force, the capital stock, and total factor productivity. Obviously we can’t object to this paradigm, since we rely on it, too.
However, precisely because we are so familiar with the characteristics of such systems, we doubt that the GI model, used as intended, shows a negative short-run fiscal multiplier. Indeed, GI’s own discussion of its model makes clear the system does, in fact, have a positive short-run fiscal multiplier.[12] This made us wonder how and on what grounds analysts at Heritage manipulated the system to produce the results reported.
Investment Incentives?
The incremental strength of GDP is primarily in the components of gross private domestic investment. How can we make sense out of this? There is an argument that, in today’s environment, the sudden adoption of, or an expectation of a serious fiscal fix that was not already priced into markets could lower interest rates enough to stimulate the interest sensitive components of aggregate demand, leading to an immediate rise in output. This can’t be what’s going on here. In 2012, long-term rates are lower by a measly 1 basis point and, by 2014, by only 9 basis points.
The report does argue that the proposed reduction in the corporate tax rate from 35% to 25% would lower the cost of capital and stimulate investment.[13] However, it is not clear to us what should happen to the overall cost of business capital in the simulation. In standard neoclassical theory, the effect of a reduction in the corporate tax rate on the user cost of capital is ambiguous in sign, depending on the interaction of the tax rate with other provisions of the tax code. In addition, tax reform may take away some or all of the current deductions for interest, depreciation, and property taxes, tending to raise the cost of capital.[14] In our own calculations, cutting the top corporate income tax from 35% to 25% does reduce the user cost of capital at any given rate of interest by a very slight amount. However, if this is combined with the elimination of just the interest deduction, the result is a very substantial increase in user cost.[15]
Crowding Out Credibility
So, as we parsed the simulation results, we couldn’t see what was stimulating aggregate demand at unchanged interest rates and in the face of large cuts in government consumption and transfer payments…until we read this:
“Economic studies repeatedly find that government debt crowds out private investment, although the degree to which it does so can be debated. The structure of the model does not allow for this direct feedback between government spending and private investment variables. Therefore, the add factors on private investment variables were also adjusted to reflect percentage changes in publicly held debt (MA italics).”
Let’s talk this through. The model undoubtedly has the standard identities equating saving and investment, so if government saving rises during a fiscal contraction, there must be a corresponding combination of changes in gross private domestic investment, personal saving, and foreign capital inflows. The usual mechanism by which this re-allocation occurs is through a decline in the interest rate which affects each term in the identity differently depending on the parameters of the underlying utility and production functions. But unless saving is infinitely elastic with respect to the interest rate, and/or domestic investment is insensitive to the cost of capital, some investment will get “crowded in” during a fiscal contraction through a decline in the interest rate…the standard stuff. Heritage, however, did something additional. It just adjusted up investment demand at every level of the interest rate, as if federal debt appeared directly in these equations.
This is purely ad hoc. Is such an investment equation routinely derived from microeconomic principles? Not that we’re aware of. Are there any genres of models — large scale structural macro models, overlapping generation computable general equilibrium models, dynamic stochastic general equilibrium models — that actually show this direct effect? Not that we’re aware of.[16] What would investment look like in the baseline, which shows exploding federal debt, if the baseline reflected this direct effect? Why is it that only federal debt should appear in the investment equations as an agent for “crowding in” business investment. Why doesn’t deficit reduction directly crowd in debt-financed consumer spending, like vehicle sales?
And what is the empirical evidence for this direct effect? The analysis cites two fairly well-known papers — one by Thomas Laubach and the other by Eric Engen and Glenn Hubbard.[17] However, both of these papers examined the relationship between deficits, debt, and interest rates — the conventional channel though which crowding in occurs and the one that already is included in the model. Neither paper says anything about a direct effect (that is, for any given interest rate) of government debt on investment. Indeed, neither paper even discusses investment equations.
In sum, we have never seen an investment equation specified this way and, in our judgment, adjusting up investment demand in this manner is tantamount to assuming the answer. If Heritage wanted to show more crowding in, it should have argued for a bigger drop in interest rates or more interest-sensitive investment, responses over which there is legitimate empirical debate. These kinds of adjustments would not have reversed the sign of the short-run fiscal multiplier in the manner that simply adjusting up investment spending did.
Hilarious Housing?
In the simulation, the component of GDP that initially increases most, both in absolute and in percentage terms, is residential investment. This is really hard to fathom. There’s no change in pre-tax interest rates to speak of, hence the after-tax mortgage rate presumably rises with the decline in marginal tax rates even as the proposed tax reform curtails some or all of the mortgage interest deduction. It’s hard to imagine the financing cost of housing not rising, at least initially. True, the simulation shows the number of households increasing 75 thousand in the first year, but the simulation also shows residential investment jumping $89 billion, or 21%. Given the size of this increase, we can suspect that residential investment was also adjusted up directly. In any event, at today’s real value of approximately $210 thousand per newly completed housing unit[18], the extra investment is enough to build roughly 425 thousand units in 2012, of which 350 thousand would be empty at a time when we estimate there already is an excess of more than a million units that could be absorbed by new household formation.
And this imbalance only worsens through time. Cumulating the increases in residential investment by perpetual inventory using a 1.5% annual depreciation rate, we calculate that by 2021 the real residential housing stock is up $1.023 trillion. At a decadal average of roughly $234 thousand per newly completed housing unit, this translates into an additional 4.4 million units. By the end of 2021 households are up only 379,000, so that 4 million unoccupied housing units have been built, creating an overhang as large as we estimate developed at the peak of the recent housing boom.[19]
Foreign Foibles
The usual story one tells about deficit reduction is that as government or government-financed consumption shrinks relative to GDP, domestic interest rates fall, putting downward pressure on the dollar, which in turn encourages higher net exports and hence smaller reliance on borrowing from abroad.
In this simulation, however, real net exports immediately decline by $82 billion relative to the baseline and eventually by $183 billion. Through 2021 the cumulative decline in net exports is $1.090 trillion in real terms, and — we’re guessing — perhaps $1.2 trillion in nominal terms. The current account balance (which is not reported) likely deteriorates more given the compounding of net factor payments to the rest of the world. Hence, this simulation might imply that one decade into the Republican plan the US has $1.3 trillion or so more net indebtedness to the rest of the world than under the baseline.[20]
Without more details we can’t be sure what’s going on here, but it looks as if this is at least partially the result of directly adjusting up domestic demand. With not much decline in interest rates, there’s probably little decline in the dollar (which is not reported) to stimulate net exports. But since domestic demand is adjusted up directly, and the income elasticity of imports likely is high, more imports get sucked in immediately, implying an increase in foreign capital inflows. Without these direct adjustments, GDP likely would show an initial decline and interest rates would fall further and faster. Net exports would rise both because domestic demand would be weaker and because imports would become relatively more expensive as the dollar weakened.
Perhaps it could be argued that this extra foreign indebtedness at least is financing private investment. Still, we have to ask: is all this extra debt to foreigners a sensible result, one that really is intended? It is our impression that reducing US indebtedness to the rest of the world is supposed to be a major benefit of, and hence motivation for, long-term deficit reduction.
Other Oddities
There are other things wrong with this simulation. Despite additional capital accumulation over the next decade, real household net worth is essentially unchanged after ten years, and real net worth per household is actually lower than in the baseline. GDP per employee (e.g., average productivity) is actually 0.4% lower through 2016 than in the baseline despite the extra capital formation. We could go on, but won’t. Suffice it to say we see a lot of internal inconsistencies in the analysis.
Concluding Comments
Political winds and economic realities portend a fiscal contraction. Understanding the macroeconomic implications of both the amount and composition of any fiscal adjustment is critical to designing a good strategy for addressing the nation’s secular budget imbalance. We agree with most other economists that there is a long-run gain to deficit reduction, and that it matters how the deficit is reduced. Furthermore, we understand that there are defensible models that show less short-run pain from a fiscal contraction than does ours. The debate over the modeling of these fiscal effects is one we will eagerly join.
In our opinion, however, the macroeconomic analysis released in conjunction with the House Budget Resolution is not relevant to the coming discussion. We believe that the main result — that aggressive deficit reduction immediately raises GDP at unchanged interest rates — was generated by manipulating a model that would not otherwise produce this result, and that the basis for this manipulation is not supported either theoretically or empirically. Other features of the results — while perhaps unintended — seem highly problematic to us and seriously undermine the credibility of the overall conclusions.
This is indeed unfortunate, since it might encourage some legislators to believe that slicing federal debt dramatically can produce long-run gain without short-run pain. After all, if the short-run fiscal multiplier really is negative, why not reduce the deficit (or even debt) to zero overnight? Unfortunately, we expect budget hawks to advance this argument during the coming debate over whether to raise the debt ceiling. That would be an especially bad time to make the mistake of assuming the answer.
[1] House Committee on the Budget, The Path to Prosperity: Restoring America’s Promise; Fiscal Year 2012 Budget Resolution (April 5, 2011).
[2] “Economic Analysis of the House Budget Resolution by the Center for Data Analysis at the Heritage Foundation” (http://budget.house.gov/UploadedFiles/heritageanalysis452011.pdf)
[3] CBO’s Alternative Fiscal Scenario is essentially an extension of current policy as opposed to current law. For example, in this scenario the current tax code and current entitlement programs are extended indefinitely. The result is exploding federal deficits and debt. For a description of this baseline, see the Congressional Budget Office, The Long-Term Budget Outlook (June 2010).
[4] The report does discuss diverging from CBO economic assumptions after 2020. Essentially, the baseline for the analysis extends the “steady state” economic assumptions that characterize the last few years of CBO’s intermediate projections.
[5] Of course, if the decline in the unemployment rate were cyclical, we’d expect immediately to see tighter monetary policy. Indeed, according to our preferred rules for monetary policy, the Federal Open Market Committee, faced with the macroeconomic configuration shown in the first set of results and believing the decline in the unemployment rate to be cyclical, would eventually raise interest rates by nearly 500 basis points relative to the baseline. This would totally overwhelm the modest decline in rates shown in the simulation itself.
[6] The Heritage Foundation, Heritage Updates Budget Unemployment Estimate (April 6, 2011) at http://blog.heritage.org/2011/04/06/heritage-updates-ryan-budget-estimate/
[7] In a recent discussion with GI, we’ve learned that such a large, ill-advised decline in the assumed NAIRU does, in fact, produce a temporary large gap between household and establishment employment.
[8] The first set of results seems to imply a reduction in the NAIRU of about 2 percentage points, while the second set of results seems to imply a reduction of about 0.6 percentage point.
[9] “Do not change significantly” seems an understatement. In fact, the other numbers reported do not change at all, in some instances to three decimal places.
[10] In fact, it is our understanding that in the GI model a reduction in NAIRU lowers productivity at full employment. However, from the first to the second set of results there is no change at all in the estimated impact of the Budget Resolution on GDP per employee.
[11] The summary table does not report detailed assumptions for defense purchases. However, the text mentions cumulative cuts in defense spending that are even larger than for non-defense purchases. Hence the first-year cut in total federal purchases likely is well larger than the $37.4 billion for non-defense purchases alone.
[12] IHS Global Insight, Fiscal Stimulus and the US Economic Outlook (January 26, 2009) at http://www.ihs.com/products/global-insight/industry-economic-report.aspx?ID=106593556&pu=1&rd=globalinsight_com#
[13] The analysis does not describe these calculations in detail or show results for the cost of capital.
[14] The Budget Resolution argues strongly for base-broadening, but does not offer any detail on which tax expenditures would be curtailed or eliminated.
[15] For example, in the case of business equipment other than computer hardware and software, we estimate that cutting the corporate tax rate from 35% to 25% would lower the user cost of capital by 0.5%, but that combining the lower tax rate with the complete elimination of the interest deduction would raise the user cost by 18%. These calculations were done using depreciation schedules that will pertain after the current stimulus provisions expire, but even these have accelerated provisions that could be curtailed.
[16] GI has confirmed with us that this direct crowding out effect is not in the GI model.
[17] See Thomas Laubach, “New Evidence on the Interest Rate Effects of Budget Deficits and Debt,” Board of Governors of the Federal Reserve System (May 2003); and Eric M. Engen and R. Glenn Hubbard, “Federal Government Debt and Interest Rates”, in NBER Macroeconomics Annual 2004, Volume 19, Mark Gertler and Kenneth Rogoff, eds. (Cambridge, MA: MIT Press, April 2005).
[18] This is a weighted average of $252 thousand for completed single-family units and $88 thousand for completed multi-family units.
[19] It would have been much easier to do this checking if housing starts or the unit housing stock were reported. Of course, that would also have made it easier to discover this bizarreness!
[20] This would be easy to check if the current account deficit were reported.
No comments:
Post a Comment