Comment on CEPR Policy Insight 91 section 4.2.1

Sorry for the title which is pure click bait.

I would like to discuss a reform of the Stability and Growth Pact proposed by Agnès Bénassy-Quéré Markus Brunnermeier, Henrik Enderlein, Emmanuel Farhi, Marcel Fratzscher, Clemens Fuest, Pierre-Olivier Gourinchas, Philippe Martin, Jean Pisani-Ferry, Hélène Rey, Isabel Schnabel, Nicolas Véron, Beatrice Weder di Mauro, and Jeromin Zettelmeyer in CEPR Policy Insight 91 “Reconciling risk sharing with marketdiscipline: A constructive approach to euro area reform” (pdf warning). This is just one section of the proposal (4.2.1) and it is the one related to some of my own research.

Vox EU has a good executive summary of the proposal. I quote the relevant passage (I will add details from the long detailed proposal especially on points which were not clear to me from the summary)

Martin Sandbu likes the proposal.

Second, replacing the current system of fiscal rules focused on the ‘structural deficit’ by a simple expenditure rule guided by a long-term debt reduction target. The present rules both lack flexibility in bad times and teeth in good times. They are also complex and hard to enforce, exposing the European Commission to criticism from both sides. They should be replaced by the principle that government expenditure must not grow faster than long-term nominal output, and should grow at a slower pace in countries that need to reduce their debt-to-GDP ratios. A rule of this type is both less error-prone than the present rules and more effective in stabilising economic cycles, since cyclical changes in revenue do not need to be offset by changes in expenditure.

Monitoring compliance with the fiscal rule should be devolved to independent national fiscal watchdogs, supervised by an independent euro area-level institution, as elaborated below. Governments that violate the rule would be required to finance excess spending using junior (‘accountability’) bonds whose maturity would be automatically extended in the event of an ESM programme (the status of the existing debt stock would remain unaffected). The real-time market pressure associated with the need to issue such bonds would be far more credible than the present threats of fines, which have never been enforced. And the cost at which these junior sovereign bonds are issued will depend on the credibility of government policies to tackle fiscal problems in the future.

To summarize more, the proposal is to focus more on expenditure and less on deficits and to enforce the rules through the junior status of bonds issued above the allowed amount. The second aspect of the proposal is more interesting and I have an (even) more favorable view of it, but I will focus on the first for most of this post.

The starting point is that the current adjustments used to calculate structural budget balance are not transparent, are controversial and according to the authors (and others including me) are just incorrect so the rules as applied force pro-cyclical fiscal policy. I have two problems with the proposal “that government expenditure must not grow faster than long-term nominal output” as stated in the executive summary. First it appears to be simply assumed that changes in government revenue are all cyclical — That Ronaldo Reagano will not be elected head of any European government. Second it is assumed that calculation of “long-term nominal output” is transparent, uncontroversial and won’t force spending cuts during recessions.

Some of my distress is relieved when I read the actual proposal.

4 Nominal expenditures are calculated net of interest payments, of unemployment spending (except when these are due to discretionary changes to unemployment benefits), and of the estimated impact of any new discretionary revenue measures (changes in tax rates and tax bases). The first two adjustments allow for more counter-cyclicality, while excluding the effect of expenditure-increasing structural measures. The last adjustment is meant to preclude the manipulation of tax rules (for example, tax cuts ahead of an election) that are not compensated by offsetting expenditure measures.

The nominal spending ceiling thus becomes a new formula for structural budget balance. The cyclical adjustments are remove changes due to changes in interest payments, unemployment spending (except when these are due to discretionary changes …) and tax revenue changes (except for those due to changes in the tax code).

The proposed system is at least as opaque as the SGP, the non-transparent part of the current process is the estimation of potential output. This must still be estimated in order to forecast normal nominal GDP growth (which, as they explain, is real potential GDP growth + 2%).

The actual application of the proposal would require and agreed definition of “unemployment spending”. Is this unemployment insurance or unemployment insurance plus unemployment assistance ? How does one correct for a change whose impact will increase as the population ages ? Netting out the effect of discretionary “changes” requires a baseline from which to estimate the change — an increase in duration of benefits can have a tiny effect the year it is introduced and a large effect later — is the effect on current spending of a 3 year old reform an increase “due to discretionary changes”.

To net out the effects of discretionary changes in unemployment benefits, one would need to estimate the effect of such changes, would increasing the replacement ratio to 110% have an effect on unemployment duration ? How about “new discretionary tax measures”, what if Arthur Lafferbaud were to assert that reduced tax rates caused increased revenue, so the correction for the effects of changes in the tax code would be to allow more spending if taxes were cut ? The effects of unemployment benefits and taxes on expenditures and revenue are exceedingly controversial. The controversies would have to be resolved for the plan to be implemented.

I think it is more practical to admit that the proposal is a new estimate of structural deficits and subtract a guess of the effect of the cycle on revenues and unemployment spending. I guess one might give fiscal authorities the right to appeal that there is no plausible discretionary change which could have caused a decline in revenue or an increase in unemployment spending, but I don’t like a rule where the key clauses are in parentheses and begin with the word “except”.

I think the real difference between the proposal and the current SGP rules is hidden in the phrase “grow … long-term nominal output (that is, the sum of potential output growth and expected inflation)”. This implies that the potential output is estimated as part of forecasts of output in the distant future. In the current approach, completely different time series models of potential output are used for medium term forecasts and for cyclical corrections. The potential output used for cyclical corrections is markedly pro-cylical which implies small cyclical corrections and a demand for procyclical fiscal policy. I think the problem would be resolved if it were required that “potential output” be given one definition and models which imply a high probability that the natural rate of unemployment will be negative in the not too distant future are ruled out (they are currently used to attempt to dictate fiscal policy).

On the other hand, it might be better diplomacy to propose an entirely new approach which is really mainly new because it doesn’t allow such models than to say that the old calculations are just incorrect and the application of the SGP has been nonsensical.

OK now the more interesting part. I do like the proposal for accountability bonds. It is enforceable — owners of regular bonds have standing to sue if a state attempts to even partially default on regular bonds while also paying something to owners of accountability bonds. My only concern is that I am pretty sure that the new rule would amount to a pretty much absolute ceiling on Italian public spending. I’m not buying and accountability BTPs from the Italian Treasury and I doubt many other people would be eager to buy them. That is, my problem with an actually effective enforcement mechanism is that, given the effects of fiscal rules so far, I don’t want them to be enforced effectively.