National Economic Review
National Institute of Economic and Industry Research
No. 68 October 2013
The National Economic Review is published four times each year under the auspices of the Institute’s Academic Board.
The Review contains articles on economic and social issues relevant to Australia. While the Institute endeavours to provide reliable forecasts and believes material published in the Review is accurate it will not be liable for any claim by any party acting on such information.
Editor: Kylie Moreland
National Institute of Economic and Industry Research
This journal is subject to copyright. Apart from such purposes as study, research, criticism or review as provided by the Copyright Act no part may be reproduced without the consent in writing of the relevant Institute.
Mercantilist and equilibrium fallacies in regional economics
Dr Ian Manning, Deputy Executive Director, NIEIR
The present paper discusses the divergent conclusions of two studies: that of Abelson, ‘Evaluating Major Events and Avoiding the Mercantilist Fallacy’, published in Economic Papers, and an economic impact evaluation completed in 2005 by the National Institute of Economic and Industry Research. In this case, the conflicting results are not due to major disagreements on theory: NIEIR agrees with most of the theoretical statements in Abelson’s article. Here, disagreement arises from the application of theory, or the relevance of assumptions
In the March 2011 edition of Economic Papers Peter Abelson accuses unspecified economists of committing the ‘mercantilist fallacy’ in their evaluation of major events. He shadow-boxes the culprits without naming them or examining any of their reports in detail, but the shadow which haunts his discussion of the 2005 Formula 1 Grand Prix in Victoria is easily identified as the economic impact evaluation completed in 2005 by the National Institute of Economic and Industry Research (NIEIR). Abelson is not to be blamed for shadow-boxing since the Australian Grand Prix Corporation, which holds copyright to the NIEIR report, has chosen not to publish anything but the bottom-line result, a gross benefit to the Victorian economy of approximately $175 million. This benefit contrasts strongly with the result of a study by Applied Economics (2007) (i.e. Abelson under his consulting hat) that estimates that the same event imposed net costs on Victoria of $6.7 million. En route to their final conclusions, NIEIR and Abelson agree that visitor-generated expenditure in Victoria was of the order of $60 million. Abelson whittles this down to a benefit of $9.4 million in ‘local production surpluses’, whereas NIEIR expands it to an increase of over $100 million in gross state product. (The total published by the Grand Prix Corporation includes a number of additional costs and benefits which are not relevant to Abelson’s accusation. For completeness, they are discussed towards the end of this article.)
How can economists come to such divergent conclusions? In this case, it is not due to major disagreements on theory: NIEIR agrees with most of the theoretical statements in Abelson’s article. The disagreement concerns the application of theory, or, to put it another way, the relevance of assumptions. The assumptions NIEIR applies to event assessment are described by Manning (2012). The present article reports a detailed comparison with the neoclassical approach to the same topic.
The importance of assumptions
Abelson (2011, p. 48) does not define the mercantilist fallacy precisely: only as analysis, which, in his view, ‘typically exaggerate(s) the benefits of export income’. Benefit exaggeration arises when the analyst fails to realise that ‘export income is valuable only in so far as it has more value than the consumption foregone’ (Abelson, 2011, p. 52). Along these lines, Abelson (2011, p. 48) summarises his argument against subsidies to events that generate export income as follows:
“Any visiting consumer wants a service in return for their expenditure and the provision of this service almost always requires the use of resources that could be employed in other activities. Consequently, an external injection of funds guarantees neither net employment generation nor a welfare-enhancing economic project.”
The mechanism proposed for this crowding out is price equilibration of demand and supply assuming upward-sloping supply curves based on rising marginal costs in all relevant markets. An extreme case of this mechanism is provided by general equilibrium models, in which rising marginal costs are necessary to generate results, but the assumption of rising marginal costs has also frequently, and unthinkingly, been transferred to economic impact analysis. Such transfer is only appropriate where there is empirical evidence that the postulated price movements actually take place; it generates false results where the relevant supply curves are horizontal or declining. The mercantilist fallacy, accordingly, has a counterpart, the equilibrium fallacy, which occurs when the market conditions that generate the mercantilist fallacy are wrongly assumed to apply.
Our purpose is not to fault Abelson’s claim that event assessments may suffer from mercantilist fallacies nor is it to deny that the eighteenth-century Corn Laws that so displeased Adam Smith were an example of the mercantilist fallacy in action. Again, we do not enter into the welter of arguments, such as those surrounding the concept of export-led growth and the circumstances in which additional exports may or may not add to the consumption possibilities of the exporting region. The case under discussion is whether subventions to the organisers of public events can have welfare-increasing results via the generation of additional export income, while admitting that whether they will or not depends on the circumstances of the place and time. In the example considered by Abelson the fallacy lies in assuming that Victoria, at the time of the analysed event, was experiencing full employment in the sense that additions to the demand for factors of production due to increased exports raised their prices and so ensured that increases in export production crowded out other production. This is an empirical, not a theoretical matter, concerning market conditions in and around Albert Park, Melbourne, Victoria in March 2005.
A digression: Labour mobility
As Abelson himself reports, in the extreme context of general equilibrium theory rising labour supply curves do not necessarily condemn public support of events as expressions of the mercantilist fallacy. In an imaginative assessment of the 2005 Grand Prix, the same event as assessed by Abelson and NIEIR, the Allen Consulting Group (ACG) (2007) drew on the Monash Multi-Regional Forecasting model (MMRF), the doyen of Australian computable general equilibrium (CGE) models. The MMRF assumes that economies operate at equilibrium full employment with rising marginal costs.
The ACG observed that the event transferred visitor expenditure from the visitors’ states and countries of origin to Victoria. The increase in expenditure in Victoria would be matched by decreases in the states of visitor origin, creating demand for labour in Victoria and excess supply elsewhere. Given the assumed hypersensitivity of labour markets to changes in demand, equilibrium would be maintained by transfer of labour from the states of visitor origin to Victoria, increasing Victorian gross state product by around $60 million at the expense of the other states. If this reasoning is correct, an injection of funds by Victoria does precisely what eighteenth-century mercantilists wanted to do: it strengthens Victoria at the expense of other states and expresses mercantilist truth rather than fallacy. However, the exercise is pointless because most of the ‘benefit’ goes to people who have lost their jobs in other states and very few individuals actually benefit. Abelson is critical of this analysis on the grounds that it omits relocation costs and depends on an incredible degree of sensitivity of interstate migration to tiny variations in the demand for labour. He argues that it is ‘far more likely’ that a one-off event ‘would simply increase part-time labour for a few days or weeks’, a suggestion that is difficult to reconcile with the dependence of his own results on rising marginal labour costs (Abelson, 2011, p. 57).
Data on interstate migration support Abelson’s criticisms of ACG. Because employed persons come with partners and dependents, we assume that every additional employment position created in 2005 required the interstate migration of two people. The following equation was estimated:
The equation is nonlinear. The higher the employment to population ratio, the greater the proportionality and the greater the net immigration rate, a form which allows the ACG labour mobility assumption to be tested. When the data for the March quarter 2005 is plugged into the equation and the employment level is increased by 1,000, the resulting increase in net interstate migration into Victoria is 43 persons, which is considerably lower than what is implied by ACG. We accordingly condemn ACG’s application of MMRF to event assessment as an example of the equilibrium fallacy.
Short-run crowding out via the labour market
Abelson (2011, p. 58) gives various reasons why the ACG assessment of the 2005 Grand Prix using CGE modelling went wrong: mainly that CGE models are ‘not primarily designed for the task of event assessment’ and are ‘not well-suited to estimate the micro intra-industry impacts of small and temporary events’. He proposes cost–benefit analysis as the appropriate methodology, whereas NIEIR was briefed to undertake an economic impact analysis concentrating on effects on gross state product and employment. This difference in aim accounts for part of the difference in the results, mainly because Abelson deducts an estimate of the value of leisure foregone as employment increases and NIEIR does not. A serious problem with the value of leisure foregone is that it cannot be directly observed and is likely to be low, or even negative, when previously unemployed persons gain work. However, the difference between cost–benefit and economic impact analysis does not explain the wide divergence of numerical results, which arise because Abelson did not follow up his own suggestion that labour supply for small and temporary events is likely to come from part-time workers and, instead, based his cost–benefit analysis on the assumption that relevant markets are subject to rising marginal costs. This is an empirical question, and failure to check the empirical position exposes him, like ACG, to the equilibrium fallacy.
The relevant markets can be identified from the expenditure surveys carried out as part of the assessment of events. The industries that gain noticeable additional demand during an event comprise accommodation, restaurants, bars and air transport. Other visitor expenditures on motor transport, entertainment and shopping, although substantial, are but small percentage additions to resident demand and, hence, are unlikely to strain regular provision. Instead, production is likely to increase through the utilisation of normal excess capacity: shops and petrol stations are a little busier and entertainment venues sell additional seats. Similarly, even though it is likely to experience a noticeable increase in demand, air transport has sophisticated means of matching demand to capacity, including complex pricing, and is likely to have the capacity to meet the surge in traffic caused by an event. Notice, however, that these observations imply a range of capacity over which sales can vary without causing any changes in wage rates or other factor prices.
The case is different in the hospitality industries (accommodation, restaurants and bars) because the additional demand is likely to be large enough to require additional labour input, at least within a small area, such as the environs of Albert Park. When the hospitality industry requires additional labour, the full employment assumption that underlies rising marginal costs gives it only one source, people already working elsewhere: ACG assumes interstate and Abelson assumes in other industries located near the event. It is also assumed that labour will be attracted to hospitality by the offer of slightly higher wages. Relaxing the full employment assumption allows us to list a number of additional sources of labour, beginning with the offer of additional time to existing employees (the hospitality industry in Melbourne in 2005 had numerous part-time workers who might be persuaded to work additional hours and also had the option of offering overtime). The industry could also offer work to people otherwise unemployed or underemployed.
The conventional measure of spare capacity in labour markets is the unemployment rate estimated quarterly by the ABS Labour Force Survey. This rate was defined in the era of male full employment in full-time jobs and is now highly unsatisfactory as an indicator of excess capacity, both because of the increase in part-time employment and (more importantly) the increasing capacity of the Commonwealth Government to move social security recipients between payments that require job search and those that do not (NIEIR, 2011, pp. 7– 10 and 59).
In the March quarter of 2005 Victoria experienced official (ABS) unemployment rates of 5.2 per cent for males and 5.7 per cent for females. In addition, ABS surveys confirmed the presence of part-time workers who wanted to work longer hours, the presence of discouraged workers and an overall ratio of hours worked to the working-age population well short of ratios common elsewhere in the OECD. From these sources it can be estimated that in March 2005 approximately 700,000 people were available in Victoria to undertake and, in most cases, could adequately provide, the generally low and semi-skilled services required to support an event. Even the most optimistic estimates of the employment opportunities created by the Grand Prix represent less than 1 per cent of this available labour.
Despite the slack in the labour market in Melbourne in March 2005, it could be argued that the hospitality industries were competing with other industries for skills in demand. For non-hospitality industries to be adversely affected, the competition must be inter-industry. Competition for skills within the hospitality industry does not reduce factor supply to other industries, although it might lead to an increase in hospitality prices and, hence, possibly choke off some of the increase in demand. In 2005 the scarce skills that the hospitality industry needed to recruit were, by and large, industry-specific (e.g. skilled chefs and suave waiters) so that an increase in demand for skilled labour was not likely to spill over into withdrawal of staff from other industries.
The question of capital capacity and accommodation prices
If it is conceded that suitable underemployed and unemployed labour was available at the time and in the place where an event occurred, the labour requirements of the event would have been met without withdrawal of labour from other industries and, therefore, without any mercantilist effects via labour supply. However, labour is not the only input to the holding of an event and it might be asked whether mercantilist effects can arise from limited capital capacity, which for events effectively narrows down to the stock of hotel rooms, bars and restaurants. The short-run answer to this question is no. Accommodation capacity is of crucial relevance to the ex-ante assessment of events because it can limit the number of visitors and, hence, the export earnings from the event, but is not relevant to short-run ex-post assessments such as the three under discussion because these are based on actual visitor numbers and expenditures. In these cases the main possible capacity effect is distributional: price increases in local hospitality venues, which transfer income from local consumers to hospitality providers. However, there is little evidence of price inflation due to a spike in demand for accommodation and eating out in Victoria in March 2005.
For evidence on this point, we can turn to the implicit deflator of the Victorian consumption of accommodation, cafes and restaurants, divided by the overall Victorian implicit consumption deflator. An upward trend would be expected, due to the effect on the overall Victorian implicit consumption deflator of rapid growth in labour productivity in goods industries compared to service industries, the China effect on goods prices and the hedonic price adjustment for electronic equipment. Accommodation prices did, indeed, drift upwards in the March quarter 2005, but only modestly at 0.7 per cent per annum, much less than the increase in health and education prices. These data give no evidence of capacity constraints in the Victorian tourism industry around March 2005.
These data series also show that real hospitality price growth tends to fall below trend when output growth is high and vice versa. In the late 1990s, when output growth exceeded 10 per cent per annum, real tourism sector prices fell, indicating that productivity growth in the Victorian tourism sector is positively related to output growth and, therefore, that the sector is subject to increasing returns to scale. For every 1-per cent increase in output growth, productivity growth (output per employee) has increased by approximately 0.6 per cent. Therefore, the industry supply curve may be downward sloping, rather than horizontal as argued so far.
Multipliers in the short run
With capital capacity not a concern and underemployed labour available, the short-run impact of an event turns into a simple exercise in macroeconomics. Due to the increase in demand, average fixed cost will be less than in a no-event base case, although average variable cost may be greater due to overtime working and/or the costs of induction of recently-unemployed people. With these effects offsetting each other, production will rise to meet demand at reasonably similar average cost. Indeed, the econometric result is that demand will be met at falling average cost due to increases in productivity.
Additional export sales add to regional demand and employ otherwise underemployed capital and labour. Not only does the additional demand fail to crowd out alternative production, it generates a Keynesian multiplier, which, in traditional fashion, peters out as savings and imports increase. NIEIR’s final estimate of the benefits of visitor expenditure at the Grand Prix thus increases from under $60 million to over $100 million. The underlying assumption is that capacity limitations are not triggered in any relevant market, including those benefiting from multiplier effects. (We note that, true to its brief, NIEIR is here conducting an economic impact analysis rather than a cost–benefit analysis in which alternative means of generating multiplier effects might be relevant.)
This is not a final estimate of the impact of the event. Various other positive items include expenditures by overseas-based media and competitors, expenditures by Victorians who run down their savings rates and import substitution effects (local motor racing enthusiasts receive local satisfaction rather than heading off interstate or overseas to attend substitute events). Negative items include the import content of the event itself and the ‘tourism repulsion effect’, which, in addition to potential visitors repulsed by the lack of discounts for accommodation during the event, includes local residents who temporarily abandon Melbourne rather than put up with the crowds and noise. These effects and their multipliers account for the final NIEIR estimate of a gross benefit of $175 million. There is room to argue over the details, but the main short-run conclusion is clear. Doing away with the equilibrium fallacy not only yields a much more positive assessment of the benefits of injecting funds into events that generate export income in a less than fully employed economy, it means that the mercantilist fallacy did not apply to events held in Melbourne in March 2005.
No claim is made that this is a universal result. To take an extreme example, the labour shortages in the Pilbara make the argument inapplicable to an event held in Karratha in 2011. In these circumstances, diversion of labour to event-staging would almost certainly curtail work in the construction sector.
The long run
The estimate for Victoria in 2005 is a short-run result and does not preclude the possibility that the mercantilist fallacy might apply in the long run, either through long-run labour market effects or through long-run capital effects. However, long-run effects are unlikely to apply to events that are strictly one-off as distinct from one of a series. Since the Grand Prix was one of a series, we will take this as background to the long-run discussion.
A long-run labour-market effect would arise if, for example, a series of events encourages workers to acquire hospitality-industry skills rather than train to alleviate skills shortages in non-subsidised industries. This argument requires a limited supply of potential trainees and limited but flexible capacity in training institutions. The argument remains possible, but it is hard to argue that it applies currently. Hospitality training institutions are specialised (rather than flexible) and have lately been very busy training aspiring immigrants.
It could also be argued that subsidised serial events adversely affect production in other industries because enhanced visitor demand induces capital investment in the hospitality industry, which reduces investment and output in other industries. There is no need to quarrel with the first step in this argument. Serial events are, indeed, likely to induce investment in capacity. Because hotel occupancy in Melbourne peaks in March, the month in which the Grand Prix is held, the event adds to peak hotel occupancy. A simple investment model for accommodation is that long-run room supply adjusts to expected growth in demand, as indicated by the trend rate of growth adjusted to the extent that current occupancy rates are more or less optimal in the peak month. The March room occupancy rate in Melbourne has oscillated around 75 per cent. When the actual room occupancy rate moves above this, investment in hotel rooms drives the rate back below the 75-per cent mark. Given that around 45 per cent of net additional and enhanced-duration visitors to major events stay in hotels, it is estimated that the 2005 Grand Prix increased demand by approximately 70,000 visitor nights, or (at an average room occupancy of 1.5) approximately 47,000 room nights. To maintain the 75-per cent peak-month average room occupancy rate, this translates into 2,000 rooms, which in March 2005 represented 8.8 per cent of available rooms. This simple model explains the increase in room supply in Melbourne Statistical Division from 29,900 in March 1996 to 37,100 in March 2006, with the Grand Prix accounting for a little over a quarter of the increase. The question is whether this had ill effects on other industries.
Investment crowd out
If a series of subsidised events induces investment in additional accommodation, does this crowd out other investment? In general equilibrium models, equilibrium of aggregate investment and aggregate savings is ensured by movements in the appropriate price: interest rates. Investible funds are pooled by the finance sector, which distributes them disinterestedly across projects to the point where prospective marginal rates of return are equalised to the interest rate. Under this account of investment, a change in expectations which increase investment in any one industry will reduce investment in others, so redistributing capacity and generating a mercantilist long-run result.
However, this theory is far from universally accepted. A major challenge has been mounted to the information requirements of the model, the argument being that a disinterested, rational distribution of investment funds is impossible in a world of uncertainty and risk (Kornai, 1971). In this world, investment tends to be determined more by immediate past industry profitability and the current climate of business opinion than it is by dispassionate contemplation of comparative opportunity, and a positive outlook in hospitality is as likely by contagion to generate investment in other industries as it is to crowd such investment out. A second challenge has concentrated on the assumption that savings are pooled and allocated by a disinterested financial system. The flow of funds accounts show that the financial system is currently marginal to the finance of Australian industrial investment, as distinct from household investment in housing. Instead, businesses reinvest their own savings: particularly depreciation allowances, but also retained profits. If this is the case, increased profitability in the hospitality sector leading to increased investment does not affect investment prospects in other industries unless there is limited physical capacity in the construction industry, which did not appear to be the case in Victoria in the mid-2000s.
Multipliers in the long run
The question as to whether an expanded hospitality sector reduces long-run production in other industries is partly a question about full employment. Returning to Albert Park, the simplest argument is that the lack of full employment in March 2005 was temporary and cyclical, in which case the 2005 event might be diagnosed as an appropriate Keynesian stimulus with long-run mercantilist fallacy effects to the extent that it encouraged investment, which increased production in the hospitality sector at the expense of production in other industries once full employment returned. This is a practical question, and the practical answer is that Victoria has underemployed labour, as it has had since the end of full employment in the 1970s. In the presence of this chronic underemployment it is reasonable to assume that subventions to events that generate export incomes will continue to provide jobs for otherwise underemployed workers, including via multiplier effects (Brain, 1986). It is true that in some parts of Australia severe shortages have developed for construction and related skills, but this is not the case for hospitality-related skills in Melbourne.
Economies of scale
Even if subventions to a series of events causes a transfer of resources into investment in the hospitality industries, this is not necessarily to be regretted, because as we noted in connection with capacity limits and price effects, the hospitality industries are subject to increasing returns to scale. More formally, productivity growth for the Victorian tourist industry has been positively related to output growth. This has been tested using quarterly data for the 1-digit accommodation and restaurant ANZSIC industry in Victoria between June 1994 and June 2006. The estimated equation is:
The coefficient for (VAT/VATt–4) is strongly positive, which suggests that the hospitality industry exhibits strong economies of scale. The short-run explanation is that once the hotel rooms are in place, the restaurant tables laid out and the core staff employed, any additional demand can be met with low marginal costs. The long-run explanation may be due to economies of scale at the level of the individual hotel or to Marshallian external economies at the local industry level (Marshall, 1920, ch IX). Either way, increased productivity is likely to compensate for any mercantilist effects of a transfer of investment resources into the hospitality industry, always assuming that industries that suffer from the transfer of investment are not equally subject to economies of scale.
The additional rooms are available for the remainder of the year. Because of economies of scale, accommodation prices can be cut, helping to fill the additional rooms and so maintain the stable relationship between the peak room occupancy rate and the average annual room occupancy rate. This is an average 8 percentage point difference. Assuming that average occupancy is maintained, for the other 11 months of the year there are an additional 470,000 occupied room nights or 700,000 visitor nights. If the visitors come from outside Victoria and, on average, spend $100 a night in Victoria, total additional exports amount to $70 million. As explained above, these will generate multiplier effects. However, it should be acknowledged that other subsidised events also contribute to the attraction of visitors. Because of the uncertainties, NIEIR did not include this effect in its estimate of the benefits of the Formula One Grand Prix.
The point at issue is whether ‘external injections of funds’ to support events that generate export income add to the welfare of consumers in the jurisdiction injecting the funds. The claim that there is little if any addition to welfare is based on short-run equilibrium assumptions: essentially that any addition to demand in the hospitality industries must withdraw labour from production elsewhere, either interstate (ACG, 2007) or in other industries locally (Abelson, 2011). In a rhetorical flourish, Abelson accuses any analyst who argues to the contrary of committing the mercantilist fallacy. While admitting that this assumption may be approximately true in some places and at some times, we have asserted that it is not generally true and, in particular, was not true in Melbourne in March 2005. Economic assessment of an event in Melbourne at that time based on the assumption of rising marginal costs is, therefore, an example of the equilibrium fallacy.
Although the mercantilist fallacy did not apply to hospitality events in Melbourne in March 2005, an argument can be formulated that a series of such events would lead to expansion of the hospitality industry at the expense of other industries. The expansion of hospitality is, indeed, likely but additions to investment in one industry do not necessarily crowd out investments in other industries and long-run additions to capacity do not crowd out long-run production in other industries if there is chronic underemployment. Finally, given that the hospitality industry is subject to economies of scale, the effect on other industries may be less important than the cost-reducing effects in hospitality itself.
Theories that depend on equilibrium under rising marginal costs can be very attractive intellectually but it is a mistake to draw conclusions from them in circumstances where their underlying assumptions are not met. The equilibrium fallacy can be very seductive.
Abelson, P. (2011), ‘Evaluating Major Events and Avoiding the Mercantilist Fallacy’, Economic Papers, vol. 30, pp. 48–59.
ACG (Allen Consulting Group) (2007), ‘Commissioned Study B. Computable General Equilibrium Analysis’ (of Formula 1 Grand Prix). In: Victorian Auditor-General (2007), State Investment in Major Events, Victorian Printer, Melbourne.
Applied Economics (2007), ‘Commissioned Study A. Cost–Benefit Analysis’ (of Formula 1 Grand Prix). In: Victorian Auditor-General (2007), State Investment in Major Events, Victorian Printer, Melbourne.
Brain, P. J. (1986), The Microeconomic Structure of the Australian Economy, Longman Cheshire, Melbourne.
Kornai, J. (1971), Anti-equilibrium: On Economic Systems Theory and the Tasks of Research, North Holland, Amsterdam.
Manning, I. (2012), ‘The Economic Impact of Public Events’, National Economic Review, no. 67, pp. 19–28.
Marshall, A (1920), Principles of Economics (8th edition), Oxford, Oxford University Press.
NIEIR (2011), State of the Regions Report 2011–12, Australian Local Government Association, Canberra.