Economic perspectives

Transcription

1 Economic perspectives Has there been an economic ivien from evolution? 1 Jo Armstrong, Richar Harris, John McLaren an John Moffat 1. Introuction It is now over twelve years since the restoration of Scotlan s parliament, after a hiatus of almost three centuries. Sufficient time has therefore elapse that it is possible to provie some evience on whether Scotlan s economy has inee performe better uner evolution. Thus we look at prouctivity, GVA per hea, employment, an R&D to see if there has been any relative improvement post Having one this, two of the channels through which evolution may affect these variables will be iscusse: the composition of expeniture an policy innovation 2. This is particularly timely given that the UK an Scottish parliaments are currently consiering proposals which will give further fiscal powers to the Scottish parliament, an the Scottish government is planning to hol a referenum on full inepenence in the autumn of What happene? In consiering Scotlan s post-evolution, there is a nee to consier what is most likely to bring long-run (sustainable) economic growth to the nation. Accoring to Krugman (1997), in the etermination of living stanars, prouctivity isn t everything but in the long run, it is almost everything. Similarly, Baumol (1984) states that it can be sai without exaggeration that in the long run probably nothing is as important for economic welfare as the rate of prouctivity growth. Figure 1 shows our emphasis on the central role of prouctivity in etermining living stanars an ientifies innovation an efficiency alongsie human capital as the eterminants of prouctivity. Opinions expresse in economic perspectives are those of the authors an not necessarily those of the Fraser of Allaner Institute Figure 2 shows workplace prouctivity, measure as GDP per hour worke, in the ifferent nations of the UK (an the G7 excluing the UK) relative to prouctivity in the US since Scotlan s prouctivity in 2010 was 80.4% of the US level. This is own slightly from a figure of 80.8% of the US level in Throughout the perio, Scotlan s prouctivity has been higher than in Northern Irelan an Wales (notably the Welsh position has eteriorate over time) but lower than in Englan. There is no obvious positive stepchange in prouctivity performance in the evolve nations since 1999.

2 USA = 100 Figure 1: Drivers of growth Source: CPPR (2008) Figure 2: GDP per hour worke, UK an G7 countries relative to USA, Englan Scotlan Wales Northern Irelan UK G7 exc. UK Source: ONS Labour Prouctivity

3 % eviation from UK GVA p.h. Figure 3: Relative (healine) GVA per hea, UK regions, Wales Scotlan N. Irelan Greater SE rest of Englan Source: Calculations base on ONS Regional Accounts Turning to a wier measure of economic well-being, Figure 3 shows relative GVA per hea of population, relative to the UK average, for the four nations of the UK (with Englan ivie into the Greater South East 3 an the rest of Englan). Scotlan has been close to the UK average since During that perio, the Greater South East has improve its GVA per hea significantly while the rest of Englan an Wales has seen significant relative falls in their GVA per hea. Since 1999, Scotlan has manage to raise its GVA per hea, relative to the UK average, so it has now almost reache parity with the UK average. However, as Figure 3 shows, there is still a large gap between Scotlan an the Greater South East, although improvement has been seen relative to the rest of Englan, Wales an Northern Irelan. Following the approach of Harris an Trainor (1999), whether there has been any greater convergence or ivergence post-evolution can be teste econometrically using the following equation: D j - t * ev + e ( ysc - yc ) t = ( ysc - yc ) t 1 + m + g 1t + g 2 (1) where y SC an y C are GVA per capita in Scotlan an a comparator, respectively; μ is an intercept; t is a time tren; t t an ev t is a ummy variable that takes the value of one from 2002 onwars (i.e. the perio post-evolution) 4. The epenent variable is therefore measuring the change in the gap between GDP per capita for Scotlan an a comparator region (3 ifferent comparators are use below). The parameter 1 measures whether the gap between Scotlan an the comparator region is trening upwars or ownwars over time 5 an 2 inicates whether this tren (if it exists) has accelerate or ecelerate since evolution. Reflecting what is seen in Figure 3, Table 1 shows that when the comparator use is the UK (either excluing the Greater SE or just excluing the Continental Shelf), there is evience of a small, but significant, acceleration in the rate of convergence since evolution (when the latter is measure post-2001) 6. When the comparator is Greater SE, the evolution time tren is not significantly ifferent from zero. However, the inability to reject the null (H0: ) that the lag of the gap in GVA per capita between Scotlan an the comparator region is not ifferent from zero suggests that there is no equilibrium relationship between the two series (so the above results regaring the evolution tren nee to be interprete with caution) 7. A further measure of Scotlan s absolute an relative improvement in welfare/growth is the level of employment.

4 =1 (smoothe tren-cycle series) Table 1: Estimate parameters from moel of convergence Comparator: UK excluing Greater South EastUK excluing Continental Shelf Greater South East Constant ** (0.0047) (0.0045) (0.0140) (Scotlan GVA Comparator GVA)t (0.0851) (0.0718) (0.0780) Trent (0.0002) Devolutiont Tren ** (0.0001) ** (0.0002) ** (0.0002) *** (0.0003) (0.0003) No. of Observations Note: Stanar errors in parenthesis. **/*** significant at 5/1% level base on stanar t-test. Figure 4: Employment levels (1999q4=1), employe an self-employe, NI (nonphd) NI (PHD) LON (non-phd) LON (PHD) WAL (non-phd) WAL (PHD) SCO (non-phd) SCO (PHD) reng (non-phd) reng (PHD) Source: Labour Force Survey

5 region R&D per unit of GVA minus UK figure (%) =1 (smoothe tren-cycle series) Figure 5: UK Employment levels (1999q4=1), employe an self-employe, (Public aministration, efence, health an eucation - PHD - sector versus non-phd sector) Lonon rest of Englan Wales 1.12 Scotlan N Irelan Source: Labour Force Survey Figure 6: R&D spening per unit of GVA relative to UK figure, North East Yorks-Humbersie West Milans Wales Scotlan Northern Irelan In 2010, the Eastern region was 234% above the UK average; South East was 75.1% above the UK average Source: Business Enterprise R&D an Regional Accounts

6 Figure 4 shows the tren in employment across the ifferent nations of the UK (with Englan split into Lonon an the rest of Englan) since the last quarter of All nations have higher employment than at the beginning of the perio with Northern Irelan achieving the most remarkable rise in employment. Scotlan also performe relatively well until However, the reboun in employment after the recession has been far smaller in Scotlan than in other parts of the UK which means that, over the perio as a whole, Scotlan has only performe better than the rest of Englan. Figure 5 shows that, when the employment figures in Figure 4 are isaggregate into two broa sectors (public aministration, efence, health an eucation labelle PHD from now on versus all other inustries) 8, then, in all parts of the UK, most of the rise in employment is attributable to the PHD sector. In Scotlan, Wales an Lonon, aggregate employment growth woul have been non-existent or negative, ha this sector not expane. This is clearly a worrying fining given that such employment growth in the PHD sector is more likely to be stagnant or negative in the near future, because of the current UK government s commitment to cuts in public expeniture. Figure 6 shows R&D expeniture per unit of GVA relative to the UK average for selecte UK regions 9. This is important as R&D is a key eterminant of prouctivity (see, e.g. Harris & Moffat, 2011). Among the selecte regions, Scotlan s R&D performance has improve slightly since 1995 but has remaine relatively poor. Of the selecte regions, only Yorkshire & Humbersie an Wales ha lower R&D expeniture per unit of GVA in To the extent that R&D is a leaing inicator of future performance, this is a worrying fining. Note: these R&D figures are base on information covering the most important R&D spening firms in the UK; thus they are likely to unerestimate R&D spening by smaller firms. Table 2 presents nationally representative ata from a ifferent source, showing Scotlan (an to a lesser extent Wales) to have the smallest proportion of firms innovating an/or conucting R&D 10. One argument for evolution is that it allows bugets to be spent in accorance with local preferences (cf. the literature on fiscal feeralism, especially Tiebout, 1956; Oates, 1972). One way of testing this hypothesis is to look at whether there have been changes in the composition of expeniture since evolution. This can be one with ata from the ONS Public Expeniture Statistical Analyses an using the following moel: spening it ev i it time t 3 1 ev it time t (2) where spening it measures the proportion of expeniture going to a given area of expeniture in region i at time t, evit is a ummy variable that takes the value of one from 1999 onwars in evolve region ; time t is a time tren that it shows how expeniture has grown in the non-evolve regions of the UK (i.e. the regions of Englan). The ev coefficient on it timet is of greatest interest as it shows whether the percentage point increase (ecrease) in expeniture has been faster (slower) in evolve region an therefore provies a measure of the egree of policy heterogeneity 11. However, it must be acknowlege, that this metho will not necessarily capture policy heterogeneity because ifferences in policy o not necessarily require changes in expeniture (the same amount can be spent, but spent on a ifferent mix of unerlying services capture by the aggregate figures). Furthermore, given that a large proportion of spening is on wages, which will increase or ecrease at the same rate across the UK, looking for variation in expeniture totals may be a very strict test of policy heterogeneity. The results from estimating equation (2) by OLS regression 12, for those areas in which the majority of expeniture is uner the control of the Scottish government, are given in Table 3. Taking health as an example, the coefficients can be interprete as follows (taking each in turn): at the start of the perio, on average 18.3% of ientifiable expeniture went towars health across the English regions; there was no significant ifference in the amount of ientifiable expeniture going towars health in Scotlan at the start of the perio; health expeniture in Englan has growth by 0.4 percentage points per year since 1999; health expeniture in Scotlan has grown by 0.1 percentage points less (i.e. 0.3 percentage points) than in Englan over the perio. This latter we can take as evience of policy heterogeneity in Scotlan. It is reasonable to expect that expeniture on enterprise an economic evelopment, agriculture, forestry an fisheries, transport an eucation an training will have the most irect impact on economic performance (through potential increases in prouctivity). Expeniture on eucation an training, an transport, has been growing by 0.1 fewer percentage points, an by 0.2 more percentage points, respectively, in Scotlan compare to Englan. The ifference is positive but not statistically significant in enterprise an economic evelopment an agriculture, forestry an fisheries. In sum, therefore, there is no clear inication of expeniture moving towars those areas that are likely to improve the performance of the Scottish economy in the future However, there has been little evience of such economic policy innovation. Most recent economic policy ocuments (see Northern Irelan Executive, 2011; Scottish Government, 2007; Welsh Assembly Government, 2010) from the evolve nations focus on the same rivers of growth such as R&D, training an investment an employ the same type of methos to encourage them (base on comparable analysis unertaken at HM Treasury after 1997.

7 In oing so, they ten to follow the UK strategy ocuments (HM Treasury, 2000, 2001) Another argument for evolution is that it encourages policy innovation by creating inter-jurisictional competition. However, there has been little evience of such economic policy innovation. Most recent economic policy ocuments (see Northern Irelan Executive, 2011; Scottish Government, 2007; Welsh Assembly Government, 2010) from the evolve nations focus on the same rivers of growth such as R&D, training an investment an employ the same type of methos to encourage them (base on Table 2: Percentage of establishments proucing a prouct innovation or unertaking R&D, Prouct innovation Blue-sky innovation a R&D South East Eastern Englan East Milans South West West Milans UK Yorks-Humbersie North East Lonon Wales North West Scotlan a Introuction of a new prouct that is new to the inustry (not just the firm) Source: weighte ata from Community Innovation Surveys, Table 3: Estimates of parameters from Equation (2), Depenent variable - % of ientifiable expeniture going to: Constant Devolution Scotlan Devolution Wales Devolution NI Time Tren Time Tren Devolution Scotlan Time Tren Devolution Wales Time Tren Devolution NI General Public Orer Public Services & Safety Enterprise & Economic Development Agriculture, Fisheries & Forestry Transport Housing & Community Amenities Health Recreation, Culture & Religion Eucation & Training 0.015*** 0.061*** 0.015*** 0.020*** 0.033*** 0.012*** 0.183*** 0.020*** 0.160*** (0.000) (0.002) (0.001) (0.001) (0.002) (0.001) (0.001) (0.000) (0.001) 0.008*** *** *** (0.001) (0.007) (0.004) (0.004) (0.007) (0.005) (0.005) (0.002) (0.004) 0.006*** *** ** 0.007*** ** (0.001) (0.007) (0.004) (0.004) (0.007) (0.005) (0.005) (0.002) (0.004) 0.008*** 0.048*** 0.015*** 0.013*** * 0.029*** *** *** (0.001) (0.007) (0.004) (0.004) (0.007) (0.005) (0.005) (0.002) (0.004) *** 0.000* 0.001*** 0.004*** *** 0.001*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) ** ** ** (0.000) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.000) (0.001) 0.001*** (0.000) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.000) (0.001) *** ** *** *** (0.000) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.000) (0.001) Observations Note: stanar errors in parenthesis..

8 comparable analysis unertaken at HM Treasury after In oing so, they ten to follow the UK strategy ocuments (HM Treasury, 2000, 2001). 3. Conclusion Our review of economic inicators has faile to provie any strong evience of a significant impact following the (re)- creation of the Scottish parliament on Scotlan s relative economic performance. While there has been some post evolution impact in terms of improve GVA per hea relative to the UK, as shown in Figure 3 an Table 1, the remainer of the evience fails to highlight why this might have occurre. For Wales an Northern Irelan there is even less to suggest evolution has resulte in any economic ivien. However, it ought to be acknowlege that our approach can be criticise on the grouns that it may be unsuitable for ientifying a causal impact of evolution. A better approach, particularly in relation to prouctivity, woul be one to estimate the impact of evolution at a micro-level, as that woul allow us to control for many of the eterminants of firm prouctivity (see, for example, the approach use by Harris an Moffat, 2011). The etaile work neee to unertake this approach is something we plan to o in the near future. References Baumol, W. (1984), On Prouctivity Growth in the Long Run, Atlantic Economic Journal, 12, 3, CPPR (2008) CPPR Analysis of Scottish Government Targets. (available from Harris, R. & Moffat, J. (2011), Plant-level Determinants of Total Factor Prouctivity in Great Britain, , SERC Discussion Papers 0064, Spatial Economics Research Centre, LSE. Harris, R. & Trainor, M. (1999), Manufacturing Inustries in the UK: Was There Convergence uring the Perio? Scottish Journal of Political Economy, 46, 5, HM Treasury (2000), Prouctivity in the UK: The Evience an the Government s Approach, Available from: HM Treasury (2001), Prouctivity in the UK: 3 The Regional Dimension, Available from: Krugman, P. (1997), The Age of Diminishe Expectations, Cambrige, MA: MIT Press. Northern Irelan Executive (2011), Economic Strategy, Available from: Oates, W.E. (1972) Fiscal Feeralism, New York: HBJ. Scottish Government (2007), The Government Economic Strategy, Available from: 2.pf Tiebout, C. (1956), A Pure Theory of Local Expenitures, Journal of Political Economy, 64, Welsh Assembly Government (2010), Economic Renewal: A New Direction, Available from: n.pf Ennotes 1 This paper is base on the presentation the Urban an Regional Economic Stuy Group on 11th January, We wish to thank the participants for comments, with the usual isclaimer that only we are responsible for the final views expresse here. 2 Ientifying a causal impact of evolution on ifferent inicators of economic performance is ifficult. This is because of the problems inherent in estimating what woul have happene to Scotlan s economy, ha the Scottish parliament not been create (the counterfactual). As a result, in this paper, we generally rely on comparisons of Scotlan s performance with that of other regions an with its performance prior to evolution. Both have shortcomings as measures of what woul have happene in the absence of a Scottish parliament because of ifferences in others factors across time, an across regions, that will affect performance. 3 Specifically, Lonon, the South East an the East of Englan. 4 We starte with ev t having the value of one from 1999 onwars, but the results were not significant for Scotlan. However, as any policy changes will take time to fe through to change outputs, then using a later start ate for the ummy seems reasonable. 5 A negative (positive) sign inicates that the gap is getting larger (smaller). 6 Results (not shown here) for Wales an Northern Irelan never show any evience of convergence or ivergence, even when (to give evolution a fairer chance of working) we have experimente by setting ev t to later years. 7 Note, the t-values obtaine from the analysis must be compare to the Dickey-Fuller istribution, an not the Stuent s t-istribution. 8 Note, the first broa sector (public aministration, efence, health an eucation) mostly comprises

9 employment in the public sector (some 77% of total employment in 2010.q3 was in the public sector base on ata from the ONS series Public Sector Employment Statistics ). Thus overall, most jobs epen irectly on public sector spening. 9 R&D spening in Northern Irelan rose significantly in Part of the reason seems to be a significant increase in spening by the aerospace inustry in the Province. 10 Similar ata for is available for Northern Irelan (but was not available here). 11 It may be thought unlikely that any significant policy heterogeneity will emerge immeiately after the creation of the evolve boies. To allow for a elaye impact of evolution, we experimente by changing the efinition of evit to being a ummy variable that takes the value of one from 2000, 2001 an so on onwars in evolve region. Using 2000 to 2003, there was little impact on the results for Scotlan but using 2004 onwars a larger ev number of the coefficients on the it timet became statistically significant although their magnitue remaine small. This implies that it took a lengthy perio of time for the Scottish Parliament to begin to eviate from UK spening priorities. 12 This metho is not strictly applicable in the current situation because the epenent variable is boune between 0 an 1. However, it has the avantage of proviing results that are easy to interpret.

10