The unintended consequence of English Votes for English Laws.

Armstrong, Angus ; Ebell, Monique

The Scotland Bill 2015-16 would make the Scottish government one of the most powerful sub-central governments in the OECD in terms of its control over spending and taxation. The UK government has also announced plans to introduce 'English Votes for English Laws' (EVEL), where the support of a majority of English MPs would be necessary to pass legislation deemed to impact on England only. The objective of this paper is to examine the potential for spillovers to arise in monetary unions of asymmetric nations where fiscal policy choices are taken locally. We extend a model of Chari and Kehoe (2008) to show the sub-optimal consequences of devolved fiscal policy in a moneteary union with a dominant member state. Because England is so much larger than the other constituent nations of the UK, its fiscal policy choices will have a commensurately stronger impact on UK monetary policy. As a result, UK monetary policy might be inappropriate for the smaller nations, calling into question the economic efficiency of EVEL. This is a general result which arises from the asymmetry of nations rather than specific UK funding arrangements or behavioural responses.

Keywords: devolution; English votes for English laws; fiscal and monetary policy; coordination; spillovers

JEL Classifications: E42; E58; E61; E63; H77

Introduction

Immediately after the Scottish referendum last year, Prime Minister Cameron set up the Smith Commission. Its remit was to reach an agreement between the five main political parties in Scotland to devolve further spending and taxation powers. The Scottish government already controls 60 per cent of total public spending and 10 per cent of revenues raised in Scotland. (1) Under the Smith Commission's recommendations, included in the Scotland Bill 2015-16, the Scottish government would control 65 per cent of spending and 45 per cent of revenues. By these metrics, Scotland would be among the most powerful sub-central governments in the OECD. (2)

At the same time, the Prime Minister announced that "just as Scotland will vote separately in a Scottish parliament on their issues of tax, spending and welfare, so too England as well as Wales and Northern Ireland should be able to vote on these issues". (3) The issue of MPs in England voting on tax, welfare and spending matters which only affect England is referred to as English Votes for English Laws (EVEL). A UK-wide survey in 2014 showed that 64 per cent of English respondents broadly support such changes. (4) On the face of it, EVEL seems to be a common sense proposal, consistent with the idea of representation and taxation being indivisible.

The objective of this paper is to examine the impact of EVEL on welfare through its influence on monetary policy. We extend a model of Chari and Kehoe (2008) to show that because England is so much larger than the other constituent nations of the UK, its fiscal policy choices will have a commensurately stronger impact on UK monetary policy. As a result, UK monetary policy might be inappropriate for the smaller nations, a form of spillover. Kneebone (1994) recounts an example of this type of spillover in Canada, where the province of Ontario makes up 38.5 per cent of Canada's population and 36.5 per cent of its GDP. (5)

"An example of the conflict between provincial governments is the commonly heard complaint that a loose fiscal policy by Ontario during the mid1980s caused the Bank of Canada to impose a more restrictive monetary policy than would otherwise have been necessary. Since by necessity any monetary policy is imposed on all members of a currency union, Ontario's fiscal decisions are thought to have had a substantial negative impact on other provinces". (6)

It is important to note that the spillovers we show in the paper are not the UK-specific spillovers transmitted through the block grant or behavioural changes induced by tax policies in a neighbouring country. The spillovers we show are general to asymmetric nations which conduct local fiscal policy in a monetary union. We suggest that these spillovers are, in fact, quite common and potentially very significant. A recent example is the European Central Bank's ill-fated tightening of monetary policy in 2011. While this may have been appropriate for a very large country like Germany growing at over 3 per cent at the time, it was wholly inappropriate for smaller countries in the monetary union facing depression conditions, such as Greece which contracted by 9 per cent.

EVEL

The newly elected UK government has tabled a change in Standing Orders (parliamentary procedures) to introduce EVEL. Under the new proposals the Speaker would decide whether a bill (or parts of a bill) refers only to England (or England and Wales). If deemed to be the case then the bill would follow a new procedure whereby English (or English and Welsh) MPs would form a Grand Committee with the power to refer the bill back to the House. The proposals fell short of denying MPs from all UK constituencies from a vote on all legislation, despite English MPs not having reciprocal rights on devolved legislation. However, the Grand Committee would have the capacity to block or veto legislation which impacts only on England (or England and Wales). Most importantly, since income tax in Scotland will be set by the Scottish Parliament, income tax in England would be deemed as an English issue. And since income tax must be included in the Finance Act each year, this effectively hands the Grand Committee an annual opportunity to veto the wishes of the UK government on English income tax. (7) The UK government would no longer have full control over income tax; the most visible and highest yielding form of taxation in the Union.

In the parliamentary debate, some critics pointed out that decisions by the Grand Committee would have consequences for other nations of the UK. This would contradict Oates' (1972) famous Decentralisation Theorem: local provision is only at least as efficient as central provision of goods and services if there are no spillovers or externalities to other regions. The Smith Commission proposed that this be resolved by a 'no detriment' principle whereby the UK and Scottish governments would not be financially worse off due to the devolution of additional powers and that post-devolution decisions by the governments would only impact on their own economies. The UK government's Command Paper (2015) proposed that post-devolution spillovers could be resolved through abatement or adjustments to the Barnett Formula. However, this is fraught with conceptual, implementation and measurement problems (see Armstrong, 2014, for an early discussion).

An important conceptual issue is that economic and monetary unions enable risk sharing, which raises welfare for all members (a positive sum game). Governor Carney (2015) has stressed that the common feature of enduring unions is fiscal risk sharing. Since greater devolution curbs risk sharing resulting in an overall loss, it is difficult to see how compensation consistent with 'no detriment' is possible.

Critics of EVEL point out that there are direct spillovers through the block grant transfer to devolved governments. Changes to departmental expenditure limits which affect only England (such as the Department of Health) will alter the grant to Scotland via the Barnett formula. While this is correct, it will become less of an issue as the block grant is reduced as more tax powers are devolved. Critics also point out that a change in taxation one side of the border is likely to induce a behavioural change on the other, thereby also creating a (very difficult to measure) spillover.

A substantive issue raised by Cuthbert (2015) is that devolving tax and spending powers to Scotland requires symmetrical ring fencing of tax and spending powers for the rest of the UK. This violates the practice of not hypothecating taxes. Suppose that English MPs wished to raise income tax to fund higher spending on defence (a reserved or non-devolved spending for the entire UK). There would need to be an offsetting transfer from Scotland to England. With a smaller tax base this could have a greater economic impact. The solution would be changes in devolved taxes to cover devolved spending. Alternatively, if English MPs wished to retire some of the national debt using income tax revenue this would also benefit Scotland and so would require another transfer. This is hypothecation through the 'back door'.

In this paper we examine a more general form of spillover. We ask whether decentralised fiscal policy in a monetary union of nations of very different size can lead to sub-optimal monetary policy for the junior partner. Note that this is a general result and not specific to the case of the UK. This calls into question whether EVEL is economically efficient. It draws out the importance of the asymmetry of size between members of a monetary union when fiscal policy is no longer coordinated. While we show in the following section that none of the large OECD federal nations have anything like the degree of asymmetry as in the UK, there are clear cases where this issue has emerged in practice, such as the case of Ontario cited above.

The precise welfare consequences are extremely difficult to measure, however, and it is difficult to see how to quantify any adverse effects on Scotland as a basis for compensation. We present a formal model to show the sub-optimal consequences of devolved fiscal policy in a monetary union with a dominant member state.

An asymmetric union

Under the Scotland Bill 2015-16, the Scottish government's control over spending and taxation will make it one of the most powerful sub-central governments of any advanced federation. Yet, crucially, the UK is not a federation. Even after devolving these powers to Scotland the UK will have one of the most centralised governments in the world (see Travers in this volume). This matters because in federal countries the nation is divided into powerful states, either many or few but of fairly equal size. Table 1 shows that federal nations typically have many more states, which are more similar in size (e.g. Australia and Germany). In cases where there are many states (e.g. the US) then it is unlikely that one can have a large impact on national macroeconomic policy. In cases where there are broadly similar sized states then no federal state has anything close to a majority of the population.

The case of the UK is quite different. Table 1 shows that there are only four constituent nations to the UK and England is 85 per cent of the total UK population. Moreover, there is no national government for the largest member of the union and, if there were, England would be much more dominant within the UK than any of the other federal states in Australia, Canada, Germany, Switzerland or the United States. The next largest federal state would be Ontario, which has 39 per cent of Canada's population. Whereas none of the major federal countries is dominated by a single state, England would clearly be dominant in the UK. We leave it to political scientists to suggest whether a federal model could ever work with one nation so dominant in terms of size. And even if all constituent nations of the UK had fully devolved fiscal powers (assuming that this might be possible), then England would still be the dominant economic nation of the UK simply by virtue of its size.

This paper asks: what is the consequence of this asymmetry of size for macroeconomic policy in a political and monetary union? When federal states, regions or even countries within a monetary union are symmetric--identical in every relevant way--then setting an optimal monetary policy for the union is the same as setting an optimal policy for each member state. Similar countries require similar policies. The similarity can come from features of a country which are difficult to alter, such as their relative sizes. The similarity can also be due to homogeneous preferences and similar economic conditions, leading to similar fiscal and other economic policies as well as monetary policy.

When two countries are asymmetric (or dissimilar), then setting monetary policy requires trading off the optimal policy in one country against the optimal policy for the other. A compromise must be reached. Under current arrangements, the Bank of England's mandate is to set monetary policy optimally for the entire UK. Implicitly, this requires placing weights on the appropriate monetary policy for England and Scotland in proportion to their sizes or economic importance. Scotland is roughly onetenth the size of England in terms of both population and GDP. As a result, monetary policy can be expected to be set much more closely in line with the needs of England than those of Scotland. And as the Bank of England's latest Inflation Report makes clear, the Monetary Policy Committee's projections (on which they base policy) are "conditioned on the tax and spending plans outlined in the March 2015 Budget". (8)

England's dominant size means that English taxation and spending decisions will have a stronger influence on UK-wide monetary policy. The policy may not be consistent with the fiscal policy chosen separately by the Scottish government. Given that the UK government has committed to deliver the Smith Commission recommendations, and the fixed asymmetry of nations, we are firmly in a second best world. In our view, the influence of English decisions on the Scottish economy justifies a Scottish vote on English laws. One might ask whether the reverse is also true: should English parliamentarians also retain a vote on Scottish fiscal policy? Elere, we argue that the case for English votes for Scottish laws is much weaker.

Monetary union with asymmetric nations

Chari and Kehoe (2008) analyse a model of monetary union among symmetric countries with fiscal and other non-monetary policy decisions taken on a regional level. They find that optimal monetary and non-monetary policies are independent of one another. We extend Chari and Kehoe (2008)'s analysis to the case of asymmetric regions in a monetary union which differ along some key dimensions, and show that now the fiscal and nonmonetary policy decisions of one region do indeed have an effect on the other region, via their impact on unionwide monetary policy.

The welfare of the ith region is expressed by the function [V.sup.i]([[tau].sub.i], [pi]) which depends on its choice of fiscal policy [[tau].sub.i], and on the common monetary policy [pi]. There are two ways that states or nations may differ, coinciding with Oates' (1972) two main reasons that regions choose to have devolved fiscal powers in the first place. First, preferences might differ across regions, so that each region would have a distinct welfare function. The welfare function [V.sup.i]([[tau].sub.i], [pi]) can either vary across regions because they have different parameterisations (differing degrees of risk aversion, for example) or because they take different functional forms (perhaps reflecting constant absolute risk aversion rather than constant relative risk aversion preferences in some regions, for example) or both.

Second, preferences might be identical, but economic conditions (such as the current or expected business cycle shocks, financial asset or property price shocks or the benefits from certain kinds of policies) might differ across regions. For example, the Scottish economy is probably a net gainer from higher oil prices (the gains from higher oil revenues outweigh the higher consumption costs) while England's economy loses from the same shock.

Both of these sources of heterogeneity across regions would give rise to different optimal non-monetary policy choices [[tau].sub.i] [not equal to] [[tau].sub.i] for i [not equal to] j. In addition, England and Scotland are countries of very different size. Both by population and by GDP, England is about 10 times as large as Scotland. This size gap will turn out to be important, because it determines both how monetary policy is set and who loses most under EVEL.

There are N regions indexed i = 1, ... N which are in a monetary union. Monetary policy is set by a single central bank, whose objective is to maximise welfare across the monetary union. In the case of the UK, the four regions are the constituent nations of England, Scotland, Wales and Northern Ireland (or Scotland and the rest of the UK), and monetary policy is set by the Bank of England for the entire UK. Elere we focus on only two regions: England and Scotland.

We contrast two regimes: English votes for English laws (EVEL) and policy coordination (PC). Under EVEL, each region sets its own fiscal policy independently and non-cooperatively. England would set its fiscal policies (and other non-monetary and non-fiscal policies such as structural reforms) independently and without Scottish input, and vice versa. In a parliamentary system, only English MPs would vote on English policies and only in the interests of the English electorate. Equally, only Scottish MPs would vote on Scottish policies and only take the interests of the Scottish electorate into account. (9)

Under policy coordination (PC), in contrast, fiscal policies are set jointly by the regions. In a parliamentary system, English and Scottish MPs would vote on English fiscal policy. They would also all join together to vote on Scottish fiscal policy. Note that this does not necessarily imply that fiscal policy is uniform across diverse regions: if England and Scotland differ along important dimensions, then fiscal policy choices will reflect those differences. For example, the tax rates which would maximise joint welfare might be higher in one region than in the other. PC does mean that fiscal policy is chosen in such a way as to maximise the joint utility of all regions, taking into account the impact of each region's policies on the union's monetary policy. Because PC maximises joint utility, it defines a first-best benchmark for the case of asymmetric regions.

In both regimes, the central bank is assumed to choose its monetary policy after the regional governments have set their non-monetary policies. This is a realistic assumption, as implementing a fiscal policy requires forward planning, while the central bank target rate can be changed at very short notice. It is consistent with how the Monetary Policy Committee responds to fiscal policy today.

The central bank chooses monetary policy n to solve

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)

where [[alpha].sub.i], is the weight on region i in the central bank's utility function. We assume that [[alpha].sub.i] is a population weight, proportional to region z's share of total population in the monetary union. (10)

The central bank's optimality condition is:

[N.summation over (i=1)] [[alpha].sub.i][V.sup.i.sub.[pi]]([[tau].sub.i], [pi]) = 0 (2)

where [V.sup.i.sub.[pi]]([[tau].sub.i], [pi]) represents the impact on region i's welfare of a marginal change in monetary policy. (11) Generally, monetary policy will be tailored to suit larger regions more closely, because they have a larger weight in the monetary authority's objective.

If all regions were symmetric, so that they could all be expected to set identical non-monetary policies, then the optimal monetary policy would also be identical for all regions. (12) When countries are asymmetric, however, it will generally not be the case that a single choice of monetary policy will turn out to be optimal for each of the diverse regions. Now, the central bank will face trade-offs in setting monetary policy among regions, and its optimal policy will be sub-optimal for both regions. (13) For example, if the optimal monetary policy for the heterogeneous regions i and j were such that[[pi].sup.*.sub.i] > [[pi].sup.*.sub.i], (14) then the central bank's optimal policy would be [[pi].sup.*.sub.i] > [[pi].sup.*] > [[pi].sup.*.sub.i], with the common monetary policy variable set too high for j but too low for i. Also, the size weighting would imply that the monetary policy variable would be set to be more in line with the optimal monetary policy for the larger country.

In either case, optimal monetary policy, defined implicitly by equation (2), gives an optimal monetary policy for each set of regional non-monetary policies:

[pi] = [pi]([[tau].sub.1],... [[tau].sub.N]| [[alpha].sub.1],... [[alpha].sub.N]) (3)

Thus, the optimal monetary policy becomes a function of all regional fiscal policy choices, weighted by their sizes. In this way, the fiscal policy choices which each region makes will have an impact on the monetary policy and hence the welfare of all other regions in the monetary union. Also, the larger regions' fiscal policy will have a greater impact, as larger regions' welfare has a higher weight in the central bank's objective function.

English votes for English laws

Under EVEL, the Scottish government sets its own fiscal policy and the Grand Committee can veto UK government taxes in order to maximise regional welfare. Each regional body, aware of the central bank's reaction function, chooses its optimal fiscal policy [[tau].sub.i], to maximise its own welfare:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (4)

subject to (3). The regional government's optimal fiscal policy under EVEL satisfies:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (5)

Clearly, the Scottish government and Grand Committee only takes the impact of monetary policy on itself into account (second term), neglecting the monetary spillovers onto other regions of its fiscal policy. If, for example, one large region were to choose a low income tax policy (or veto higher income taxes), and the central bank were to tighten monetary policy to accommodate it, then all regions would feel the effects of tighter monetary policy, whether it was appropriate or not. As in the Ontario example from the mid-1980s, if the central bank were to tighten monetary policy to discipline a large region, then all regions would suffer the effects. Under EVEL, the large region would neglect the damage it is doing to other regions.

Policy coordination

Under PC, regional governments set their fiscal policies jointly, in order to maximise joint welfare across the monetary union. Setting fiscal policies jointly in a monetary union with heterogeneous regions will generally not result in uniform fiscal policies across the monetary union: instead, each region's fiscal policy will be chosen taking both regional conditions and the impact on the whole union via monetary policy spillovers into account.

Formally, optimal fiscal policy for region i solves

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (6)

subject to (3). The regional government's optimal fiscal policy under PC satisfies:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (7)

Under policy coordination, each government's regional policy is chosen taking its effects on inflation across the entire monetary union into account (second term of (7)). That is, policy coordination internalises the monetary spillovers onto other regions.

Mapping EVEL and PC onto the UK

Next we explain how to map EVEL and PC onto the UK political landscape. Under EVEL, Scottish MPs would vote on Scottish fiscal policy (in the Scottish parliament), and English MPs would vote (through its veto powers) on English only fiscal policy. English MPs would neglect the impact of their fiscal policy decisions on the monetary policy faced by Scotland. At the same time, Scottish MPs would neglect the impact of their decisions on the inflation faced by England.

Policy coordination, however, involves fiscal policy being set jointly by England and Scotland. One way to interpret policy coordination is that it mimics current UK practice with regard to England: government policy is set according to the preferences of government MPs from the entire UK. Government policy for England is assumed to be formed as a weighted average of the policies preferred by Scottish and English MPs in the government. If, as is currently the case, only one Scottish MP were in the ruling party/coalition, then the weight on Scottish preferences in (6) would be accordingly small. In the long run, taking the [[alpha].sub.i]'s to be population weights is the same as assuming that, in the long run, Scotland's MPs are in government in proportion to Scotland's share of population--overweighted in Labour but underweighted in Conservative governments in recent history.

Another way of putting this is that EVEL might have little short-term impact on parliamentary voting outcomes, as the current government has a majority that does not rely on Scottish MPs. Even without EVEL, policy can be expected to reflect the needs of English constituents and monetary policy will be adjusted accordingly. In the longer run, however, it is perfectly plausible that future governments might depend on Scottish MPs. In fact a realistic scenario coming into the recent election was a Labour and SNP coalition government or working agreement. In this case, the policies under EVEL and PC might be expected to differ. With monetary policy more responsive to fiscal policy set by the English Grand Committee (presumably by a Conservative majority), this would have been to the detriment of Scotland.

Do no EVEL

There are two main points to be made about EVEL. First, EVEL is sub-optimal. This can be seen by comparing the optimality conditions for EVEL (5) and for PC (7). The two conditions coincide for England and Scotland whenever [patrial derivative][V.sup.E] ([[tau].sub.E], [pi])/[patrial derivative][pi] = [patrial derivative][V.sup.S] ([[tau].sub.S], [pi]) [patrial derivative][pi], that is, whenever Scotland and England are perfectly symmetric, with identical preferences and policies. If Scotland and England are asymmetric either due to preferences or economic conditions, then the optimality conditions for EVEL and PC do not coincide, and EVEL does not achieve the joint welfare optimum.

Second, Scotland stands to lose more by EVEL than England. The reason is that the monetary policy chosen by the central bank will be much more strongly influenced by economic conditions in England (equation 2) than by those in Scotland. (15) As a result, the impact of English fiscal policy decisions on overall monetary policy--and hence on Scotland--will be much larger than vice versa. (16) In our view this gives a rationale for allowing Scottish MPs to continue to vote on English fiscal policy, even when the issues at hand are only directly relevant under English jurisdiction. Indeed, even identifying English policy which would not have a direct spillover to Scotland is a challenge. That said, it is true that it would also be better if English MPs retained a vote on Scottish fiscal policy. However, because Scotland's impact on the central bank's monetary policy is so small, the losses to England from having a Scottish parliament with sovereignty over Scottish fiscal policy are also very small.

The task of the Smith Commission was to devolve substantial fiscal powers to Scotland to shape its economic future. A direct consequence is that there will be less pooling of fiscal resources and therefore less risk sharing. As a result, economic cycles between the nations of the UK are likely to diverge compared with today. When the next large economic shock arrives, the one-size monetary policy is less likely to be appropriate for Scotland because it will be set more to suit England. Even in federal nations with a less dominant state (province), such as Ontario in Canada, there is evidence of spillovers in monetary policy. This paper shows that the current proposal for EVEL, is likely to be the wrong answer to the West Lothian Question. (17)

NOTES

(1) This includes the powers under the Scotland Act (2012) which came into effect this year. See Armstrong and Ebell (2014).

(2) Two more powerful sub-central governments are the Basque Country and Navarra which practically have full fiscal autonomy in Spain. In terms of per capita output they are the richest of Spain's 17 autonomous communities (Basque Country per capital income is one third above the national average) and account for 6 per cent of the nation's total population.

(3) David Cameron, quoted on the BBC 19th September, 2014 http://www.bbc.co.uk/news/uk-politics-29271765.

(4) Jeffrey et at. (2014), available at http://sites.cardiff.ac.uk/wgc/ files/2014/10/Taking-England-Seriously_The-New-EnglishPolitics.pdf.

(5) Ontario Ministry of Finance, Fact Sheet June 2015, available at: http://www.fin.gov.on.ca/en/economy/ecupdates/factsheet.html

(6) Kneebone (1994), p. 154.

(7) The point that income tax must be legislated each year is made by Gallagher: http://www.centreonconstitutionalchange.ac.uk/ blog/problems-evel-i-taxation. Keating points out that between 1964 and 1966 and for a few months in 1974 the government had not won a majority of seats in England. http://www. centreonconstitutionalchange.ac.uk/blog/evel-and-democraticreform.

(8) Bank of England (2015), p. 19.

(9) For brevity, in what follows we neglect Northern Ireland and Wales, focusing the discussion instead on the relationship between England and Scotland.

(10) Using GDP weights gives very similar results for Scotland, but reduces the weights on Northern Ireland and Wales relative to England.

(11) We assume that [V.sup.i]([[tau].sub.i], [pi]) has a unique optimum in [pi], that is, there is a unique monetary policy [[pi].sup.*.sub.i] such that [V.sup.i.sub.[pi]]([[tau].sub.i], [[pi].sup.*.sub.i]) = 0.

(12) If all regions were symmetric, then there would be a common optimal monetary policy [[pi].sup.*.sub.i] = [[pi].sup.*.sub.j] = [pi] such that [V.sup.i.sub.[pi]]([[tau].sub.j], [pi]) = [V.sup.j.sub.[pi]]([[tau].sub.j], [pi]) = 0. Equation (2) would reduce to [V.sup.i.sub.[pi]]([[tau].sub.j], [pi]) = 0 for all i.

(13) With asymmetric regions, there is no common optimal monetary policy, and [[pi].sup.*.sub.i] [not equal to] [[pi].sup.*.sub.j]. It will generally be the case that the (second-best) optimal monetary policy [pi] will involve trade-offs between the regions, so that if [V.sup.i.sub.[pi]]([[tau].sub.j], [pi]) > 0 then [V.sup.j.sub.[pi]]([[tau].sub.j], [pi]) < 0, and the common monetary policy is sub-optimal for both regions: [[pi].sup.*.sub.i] [not equal to] [[pi].sup.*.sub.j] [not equal to] [pi].

(14) That is, if the [[pi].sup.*.sub.i], satisfying [V.sup.i.sub.[pi]]([[tau].sub.i], [[pi].sup.*.sub.j]) = 0 and the [[pi].sup.*.sub.j] satisfying [V.sup.j.sub.[pi]]([[tau].sub.j], [[pi].sup.*.sub.j]) = 0 were such that [[pi].sup.*.sub.i] > [[pi].sup.*.sub.j].

(15) This is because the weight on English welfare in the central bank's objective function is about ten times larger than the weight on Scottish welfare.

(16) In the appendix we formalise this argument. We show that as long as the curvature properties of the two countries' welfare functions are the same, then the central bank's monetary policy wll be influenced (much) more by a marginal change in English fiscal policy than by an equivalent change in Scottish fiscal policy. That is, the slope of the monetary policy function [pi]([[tau].sub.S] [[tau].sub.E]) is steeper in the English fiscal policy variable [[tau].sub.E] than in [[tau].sub.S]. A change in English fiscal policy will have a bigger impact on monetary policy than an equivalent change in Scottish fiscal policy.

(17) The West Lothian Question was put by Tam Dalyall MP for West Lothian. How could he be able to vote on policies in Blackburn in England, but after devolution English MPs cannot vote on policies in West Lothian?

REFERENCES

Armstrong, A. (2014), speech for 'Policy Scotland', Glasgow University, http://niesr.ac.uk/sites/default/files/publications/ Policy per cent20Scotland2.pdf.

Armstrong, A. and Ebell, M. (2014), 'Commentary: Devolution within the UK', National Institute Economic Review, 230, November.

Bank of England (2015), Inflation Report, May, Bank of England.

Carney, M. (2015), Speech, 'Fortune favours the bold', http://www. bankofengland.co.uk/publications/Documents/speeches/2015/ speech794.pdf.

Chari, V.V. and Kehoe, P. (2008), 'Time inconsistency and freeriding in a monetary union 'Journal of Money, Credit and Banking, 40(7), pp. 1329-55.

Cuthbert,J.R. (2015), 'Scotland's fiscal framework: Finance Committee call for evidence on Smith Commission implementation', The Jimmy Reid Foundation.

HM Government (2015), Scotland in the United Kingdom: Towards an Enduring Settlement, Cm8990.

Jeffrey, C., Wyn Jones, R., Henderson, A. and Scully, R. (2014), Taking England Seriously--The New English Politics, The Future of England Survey 2014.

Kneebone, R.D. (1994), 'Deficits and debt in Canada: some lessons from recent history', Canadian Public Policy, 20(2), pp. 152-64.

Oates, W.E. (1972), Fiscal Federalism, New York, Harcourt Brace Jovanovich.

Smith Commission (2014), Report of the Smith Commission for Further Devolution of Powers to the Scottish Parliament, The Smith Commission.

APPENDIX

To find the slope of the monetary policy function (3) [pi]([[tau].sub.1], ... [[tau].sub.N |[[alpha].sub.1], ... [[alpha].sub.N]) in region i's fiscal policy [[tau].sub.i] we apply the implicit function theorem to the central bank's optimality condition (2).

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

In the case of two regions, England and Scotland, we have that

[absolute value of ([partial derivative] [pi]/[partial derivative][[tau].sub.E])] > [absolute value of ([partial derivative] [pi]/[partial derivative][[tau].sub.S])]

whenever

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

Assuming that the policy weight on England will be ten times as large as that on Scotland, so that [[alpha].sub.E] = 10[[alpha].sub.S], yields that

[absolute value of ([partial derivative] [pi]/[partial derivative][[tau].sub.E])] > [absolute value of ([partial derivative] [pi]/[partial derivative][[tau].sub.S])]

whenever:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

Clearly, if the cross-derivatives are of similar magnitudes, then [absolute value of ([partial derivative] [pi]/[partial derivative][[tau].sub.E])] > [absolute value of ([partial derivative] [pi]/[partial derivative][[tau].sub.S])]. If the cross derivatives are equal, then the monetary policy reaction to a marginal change in English fiscal policy will be ten times as large as the monetary policy reaction to a marginal change in Scottish fiscal policy.

Angus Armstrong and Monique Ebell *

* National Institute of Economic and Social Research and Centre for Macroeconomics. E-mail: a.armstrong@niesr.ac.uk or m.ebell@niesr.ac.uk. The authors are supported by the Economic and Social Research Council through its Centre on Constitutional Change.

Table 1. Size distribution of federal states

              No. of   CV of population    Largest state
              states    across states     total population

Australia       8            0.95               32%
Canada          13           1.50               39%
Germany         16           0.95               22%
Switzerland     26           1.06               18%
US              51           1.13               12%
Memo: UK        4            1.57               84%

Source: National Statistics.

Note: CV is coefficient of variation.