The 2010 FIFA World Cup high-frequency data economics: Effects on international tourism and awareness for South Africa

Abstract

The 2010 FIFA World Cup in South Africa was undoubtedly a great experience for both soccer fans and their South African hosts, and focused unprecedented and favourable media attention on South Africa. Despite the tournament's manifest success, however, its short-term impact on international tourism to South Africa, in the form of immediate positive effects on the economy, has turned out to be much smaller than expected or even as reported during the tournament – as this paper shows, using high-frequency daily data on tourism. This sobering outcome may be attributable to self-defeating expectation effects and this paper is a warning against overly optimistic economic impact studies which could undermine the short-term benefits of major sporting events. The paper also investigates the awareness effects of sport mega-events, and potential long-term development effects, by using data from electronic social networks.

Keywords:

• sport mega-event
• sport economics
• tourism
• self-defeating expectations
• Google
• Facebook
• social networks
• 2010 FIFA World Cup

JEL codes:

• L83
• R53
• R58

1. Introduction

The 2010 FIFA World Cup was clearly successful as a tournament and as an advertisement for South Africa's capacity to host a global event. Expectations were high prior to the tournament regarding the home team's performance and the tournament's magnitude as a major event with the potential to benefit the South African economy.

The home team's strong performance, better than could reasonably have been expected, was the source of much local pride. Further, the tournament was an organisational and logistical success, contributed to the cultural appreciation of the game of football, and was an advertisement for South Africa's capacity to host a global event despite the sometimes pessimistic forecasts in the international media. An otherwise critical international press did not contradict the ‘summa cum laude’ grade assigned to the World Cup by FIFA president Sepp Blatter (Reeves, 2010).

An economic analysis of the 2010 FIFA World Cup should be conducted at the micro or managerial level of the Local Organising Committee (LOC) and of FIFA. The financial outcomes of the LOC are not yet known. FIFA, which earned US$3.4 billion in total commercial revenues, provided the LOC with US$423 million, an amount that is considered to have been sufficient to ensure that the tournament was within budget (Pedroncelli, 2010). The costs of the stadiums and the transport infrastructure were almost entirely publicly financed, but these costs should not be attributed solely to the 2010 FIFA World Cup, since these investments have long-term benefits. It is too early to assess the ultimate impact of these public investments on the long-term growth path of the economy.

On a micro-level, there are always winners, losers and those who are unaffected by a major event of this kind. On the one hand there are reports of increased demand for vuvuzelas, travel in luxury coaches, tour operators and tourist attractions close to the stadiums (personal communication, Mariette du Toit-Helmbold, CEO, Cape Town Tourism, 25 June 2010). On the other, South African manufacturers seem to have been largely unaffected by the event (personal communication, Denny Thaver, Coface, SA's trade credit protection company, 2 July 2010).

But there was also controversy: Many South Africans experienced the tournament preparations as disruptive (Tolsi, 2010). FIFA's prescriptive approach and the LOC's plans were also controversial. Cornelissen (2009), for example, raises the concern that the plans would entrench existing spatial inequality in the country, and Darkly and Horn (2009) point out that owners of guesthouses were discouraged by the rigidity of the FIFA and LOC requirements for accreditation.

To appreciate the total economic impact of a major sporting event, it is necessary to move beyond data on individual enterprises and sectors to investigate the effect of the tournament using meso and macro economic data. These data have the advantage of aggregating possible increases in the incomes of individuals with the losses of others, thereby presenting a general economic picture.

The paper starts with a discussion of the net tourist increase to South Africa during the FIFA World Cup, as this is the major source of short-term economic benefit. We used high-frequency data to estimate the likely tourist flow. After this follows a discussion of possible reasons for the disappointing number of actual tourist arrivals after the optimistic estimates. But there have been other short-term benefits, including a feel-good effect and longer-term gains in the form of image and improved international perception of the economy. Evidence for these effects is discussed in the final section of the paper.

2. International tourism: The origin of short-term income and employment effects

2.1 Optimistic expectations and their impact on tournament preparations

The additional impact on local income and employment of a major sporting event results only from expenditures by non-residents visiting the event region. Expenditures by domestic residents on the event do not typically contribute to the impact of the event beyond the immediate duration of the tournament; even if individuals purchase televisions and other items used in relation to the event, the event usually affects the timing of the consumption decisions rather than the overall level of consumption and savings. Any increased consumption by residents during the tournament is typically counterbalanced by reduced demand in other months and other sectors. Nor did the public investment in stadiums and infrastructure for the occasion of the tournament necessarily imply an expenditure boost: with limited public finances, any additional spending on stadiums will be compensated for by reduced spending on other public purposes, if not immediately, then over time as the government's intertemporal budget constraint comes into play.

This explains the emphasis on international tourism in ex ante impact studies of the effect of the 2010 FIFA World Cup. The event raised high expectations: 230 000 foreign tourists were predicted to stay for an average of 15 days (Grant Thornton, 2004), and this number even increased to 380 000 overseas visitors (Grant Thornton, 2008). To estimate the impact of the tournament it is important to distinguish between ‘international tourists’ and ‘overseas tourists’ since differences are to be expected between the spending patterns of tourists from other countries in Africa and those from Europe, Asia and the Americas.

The optimistic forecasts with respect to tourist inflows were well received on the basis of the perceived effects of previous large sporting events hosted in South Africa in the low tourist season months, including the rugby and cricket world cups (Spronk & Fourie, 2010). By contrast, early warnings (e.g. Du Plessis & Maennig, 2007; Allmers & Maennig, 2009) from World Cup host nations such as France (1998) and Germany (2006) and from European Cup organisers that the net increase in the number of tourists that accompanies such events is typically small or even insignificant received no attention in the optimistic ex ante impact studies (e.g. Grant Thornton, 2004, 2008). The optimistic expectations were reflected in the considerable private sector investment in the tourism sector in the years leading up to the tournament. Table 1 compares the number of hotel beds available in the various categories of hotels in 2007 and 2010.

Table 1:  The supply of hotel beds in South Africa

Level	2007	2010	Av. growth p.a.
5 star	8 013	10 295	8.7
4 star	12 585	21 049	18.7
3 star	23 714	26 698	4
2 star	3 559	4 185	5.5
1 star	3 156	3 645	4.9
Total	51 027	65 872	8.9
Source: Pam Golding Tourism & Hospitality Consulting (Pam Golding, 2010).

While there was considerable expansion over this period, especially at the luxury end of the market, it would not be appropriate to attribute this solely or even largely to the approaching 2010 FIFA World Cup. Instead, rising demand had long necessitated investment in extra capacity. Nevertheless, the approaching 2010 FIFA World Cup affected the timing of these projects, advancing their completion dates in anticipation of the tournament. Some of this investment was of doubtful long-term viability, and there was a risk that some of the new hotels would fail when room occupancy and rates per room dropped after the tournament (Cokayne, 2010).

The available hotel beds listed in Table 1 exclude the many guesthouses in the South African tourism sector. Regrettably, there is no formal estimate of the number of these establishments and their available beds. However, the Department of Tourism reported that prior to the tournament approximately one third of the accommodation pre-booked through FIFA's MATCH system was non-hotel accommodation (DoT, 2010). The role of this system was a major concern for the owners of guesthouses. Even before the tournament, Darkey and Horn (2009) reported survey evidence of these owners' misgivings about the manner in which the FIFA and the LOC arranged accommodation, and 60% of the long-term bed-and-breakfast proprietors expected the tournament to have a negative impact on their business.

2.2 Divergent estimates of tourist arrivals

Official data on tourist arrivals for the months covered by the tournament have been released by Statistics South Africa. The data are shown in Table 2 for June and July 2009, and for the same months in 2010, covering the tournament, which started on the 11 June 2010 and lasted until the final on 11 July 2010.

Table 2:  Official tourist data for June and July 2009 and 2010: Number of visitors and their countries of origin

Region	June 2009	July 2009	June 2010	July 2010	Benchmark month*	Tournament month*
Europe	72 144	91 904	123 702	98 113	80 707	117 282
North America	25 502	28 951	50 902	32 068	27 274	45 314
Central & South America	4 605	5 699	47 188	11 731	5 092	35 621
Australasia	7 813	9 464	18 450	8 911	8 567	15 462
Asia	13 999	15 704	30 914	26 209	14 905	29 909
Middle East	2 794	5 124	6 189	6 010	3 681	6 259
Africa	376 280	461 630	421 074	494 815	414 658	456 296
Unspecified	2 294	4 938	22 892	2 557	3 282	16 169
Total	505 431	623 414	721 311	680 414	558 165	722 311
Total ex-Africa and unspecified	126 857	156 846	277 345	183 042	140 226	249 847
Increment						109 621
*The benchmark and tournament months were constructed to match the number of days the tournament lasted in June and July 2010, assuming that tourists would have arrived no later than 10 June 2010 and counting arrivals until 11 July 2010.
Source: StatsSA (2010).

The data in Table 2 are intended to reflect only tourist arrivals. However, the ability to assess the purpose of a visit is imperfect and especially in the case of arrivals from mainland Africa these data include very substantial numbers of migrant workers in addition to tourists. For that reason StatsSA published a separate monthly total excluding arrivals from Africa (as well as ‘unspecified’ arrivals) – a convention we have repeated in Table 2. Since the data in Table 2 are shown per calendar month, the final two columns create a ‘tournament month’ from the 2009 and 2010 data, by adding the number that arrived in June starting on the 10th to the number that arrived in July until the 11th. According to StatsSA's data, 109 621 more ‘overseas’ visitors arrived in South Africa during the tournament month than during the comparable days of 2009.

A number of alternative and notably divergent estimates of tourist arrivals are available (Lapper & Blitz, 2010). High forecasts were in the order of 350 000, such as the Grant Thornton prediction of 373 000 made shortly before the tournament (Baumann, 2010; Business Day, 2010). Some estimates even exceeded 500 000, such as the estimate reportedly made by the LOC midway through the tournament (Naidu & Piliso, 2010). Estimates at the lower end were close to 200 000 (Du Plessis & Venter, 2010). On a regional level, a survey conducted by Cape Town Tourism indicated that occupancy levels in the Cape Town Metropole for the first two weeks of the tournament averaged 34% in the first week of the 2010 FIFA World Cup and 40% in the second week, with positive effects for accommodation in the neighbourhood of Green Point stadium. However, establishments that were further afield were disappointed in the number of reservations (personal communication, Mariette du Toit-Helmbold, 25 June 2010).

2.3 High-frequency data on tourist arrivals

This paper compares such numbers with evidence from two sources. We use data first on actual arrivals at the international airports and second on occupancy rates in the hotels of the major cities. To begin the analysis, we note that the number of international plane landings did not increase substantially between June 2009 and June 2010. Daily data on plane landings at the three South African international airports in Johannesburg (where the vast majority of international flights arrive), Cape Town and Durban were collected from the ACSA (Airports Company South Africa) website www.acsa.co.za/. Baumann et al. (2009) used similar data in an earlier paper.

The average number of international arrivals per day for all three airports on the dates of 10 June (the day before the opening ceremony of the 2010 FIFA World Cup and the first day with a visible increase in arrivals) and 11 July (the day of 2010 FIFA World Cup final) increased to an average of 133 from an average of 125 international arrivals in the time period from 22 April (first normal flight day after Easter) to 9 June: an average addition of eight plane arrivals per day (+6%).

To calculate the transport capacities, the specific types of planes landing in South Africa were taken from the ACSA daily observations and combined with the ‘typical’ and ‘maximal’ seat capacities, taken from http://flug.airego.de/ (Aircraft Encyclopaedia – www.flugzeuginfo.net/).

Figure 1 shows an increase in the daily seat capacities during the 2010 FIFA World Cup period of some 11%, compared to the time period from 22 April 2010 to 9 June 2010, before the Cup. In this April–June period, the daily capacity was at about 29 980 seats, whereas it increased to some 33 350 seats (+3370) per day in the period from 10 June 2010 (the day before the opening ceremony and the first day showing a significant increase) to 11 July (the day of the final). Assuming that the available capacity was fully utilised, this implies a maximum increase in capacity of some 107 800 seats.

Figure 1: Seat capacity of international flights to South African international airports, 1 April to 2 August 2010

This modest estimate is consistent with those made by Cape Town Tourism indicating that the number of international arrivals in Cape Town was 44% higher in June 2010 than in June 2009 (personal communication, Mariette du Toit-Helmbold, 25 June 2010). First, the base period is modest: the number of international tourists visiting South Africa in 2009 was sharply down from previous years. Total quarterly arrivals from the USA, Great Britain and Germany, which make up approximately 45% of total air arrivals, reached more than 250 000 visitors during the high season from 2004 until the summers of 2007 and 2008 before dropping (during the first two quarters of 2009) to levels not seen since 2003. Second, Cape Town Tourism's estimate of international arrivals seems to include tourists who initially landed in Johannesburg.

We now turn to the second method of estimating additional tourism from the 2010 FIFA World Cup, that using hotel occupancy rates. Table 3 shows the hotel occupancy rates for the three major centres of Cape Town, Durban and Gauteng for each month of June since 2007 and includes a tournament (2010) and benchmark month (2009) calculation, following the method used in Table 2. Figure 2 shows the same data, but in terms of the daily frequency for June and July 2010.

Table 3:  Hotel occupancy rates in the major cities and regions

	Cape Town	Durban	Gauteng
June 2006	52.5	79.6	68.2
June 2007	55	74.2	73.1
June 2008	48.1	74.6	72.5
June 2009	49.5	67.3	54.9
June 2010	51.7	61.8	77.3
Benchmark month*	45.7	70.9	53.9
Tournament month*	64.1	60	85
*The benchmark and tournament months were calculated as in Table 2.
Data source: STR Global.

Figure 2: Hotel occupancy rates (as a percentage) in the major cities for June/July 2010

The data in Figure 2 should be read alongside the information on capacity expansion in Table 1, which indicates that growth especially occurred in high-end accommodation. Figure 2 shows the extent to which 2010 FIFA World Cup visitors used Gauteng as their base for the tournament because of the number of stadiums in close proximity and the easy access by road and air to more distant stadiums. The tournament month is a useful abstraction for the purposes of identifying those tourists who arrived for the FIFA World Cup. The daily hotel occupancy data for Gauteng showed an increase over the comparable day in 2009 only from 10 June onwards. In Cape Town the difference only emerges from 11 June onwards. The Gauteng hotels were fully booked on the day of the finals, but within two days had returned to the level of 2009. As mentioned above, the data on seat capacity for international flights support the same abstraction.

Combining the occupancy rates for the tournament and benchmark months with the number of available hotel rooms in the major cities gives a figure of 11 684 additional hotel rooms occupied on an average day during the tournament (and an additional 5842 occupied guest house rooms). To complete the calculation we need an assumption about the average duration of stay for international visitors. We base our assumption on survey data from the 2006 FIFA World Cup in Germany, showing that European visitors stayed on average six days while other overseas international guests stayed 11.4 days on average. For the South African tournament we assume that African visitors stayed on average for six days, other overseas visitors 11.4 days, and that South African visitors who stayed in hotels did so for two nights on average. Finally, we assume that 80% of the rooms were occupied by two persons and the remainder by single occupants. With these data and assumptions the estimated number of additional overseas arrivals is 65 000, with another 26 000 African arrivals and 72 000 South African tourists.

2.4 Explanations for the lower than expected turnout

A number of factors are likely to have played a role in the disappointingly small tourism impact of the 2010 FIFA World Cup. The first is the international recession. Our data on arrivals compare the 2010 FIFA World Cup period to the days before and after the event, which were affected by the crisis in the same manner. Without the crisis, the numbers of tourists both before, during and after the event could have been substantially larger. The effect of the recession may have been accentuated by South Africa's geographic location, being a long-haul (and therefore higher cost) destination for Europeans, Asians and Americans.

The scholarly literature lists other potential reasons for major sporting events having only a limited impact on tourism (e.g. Coates & Humphreys, 2000; Baade & Matheson, 2004; Du Plessis & Maennig, 2009; Allmers & Maennig, 2009): One important reason is the crowding-out of ‘normal’ tourists due to the noise, traffic jams and other disturbances. In the South African case, there was considerable crowding-out in the hunting sector, a market whose peak season is typically June (The Star, 2010), because overseas hunters found that flights to South Africa were now more expensive and some were fully booked, and that transport facilities were now catering for the tournament and were thus very expensive to divert to the game lodges.

Two other reasons may apply in the special case of South Africa. First, the international perception of the country as a dangerous location may have had a higher than expected negative effect on the expected increase in tourist arrivals. Second, price crowding-out is sometimes mentioned in the scholarly literature but rarely supported with data. In the case of South Africa, a special form of price crowding-out may have been at work: self-defeating expectations, i.e. the adverse impact on actual arrivals of exaggerated price rises in the tourist sector, where these price rises were based on the ex ante anticipation of massive tourist arrivals.

Figures 3a and 3b show the seven-day moving averages (to adjust for systematic weekly fluctuations) of prices for flights from Frankfurt, London, Madrid, Paris and Rome to Johannesburg on fixed days during, before and after the 2010 FIFA World Cup. These airports are the international hubs of the nations that were expected to have the most influential soccer teams in the tournament.

Figure 3: (a) Prices for Frankfurt–Johannesburg flight connections for flight dates 9 May, 9 June and 9 July. Seven-day moving average of the cheapest prices between 15 December, 2009 and 3 August, 2010. (b) Prices for London–Johannesburg flight connections for flight dates 9 May, 9 June, 9 July and 9 August. Seven-day moving average of the cheapest prices between 15 December, 2009 and 3 August, 2010

The cheapest prices of such flights were collected on a daily basis beginning on 15 December 2009 via the publicly available internet portal www.billigflieger.de/, a German online booking site. In the case of the flight connection between Frankfurt and Johannesburg on 9 May 2010 (a date well before the 2010 FIFA World Cup) the supply price between 15 December 2009 and 8 May 2010 varied by approximately €350 for most of the observation period and by approximately €480 in the period from the middle of March 2010 to the beginning of April 2010, with an average variation of €387.2. For the same flight on 9 June 2010, prices were over €900 during the period between end of January and the end of March. From that time forwards, there was a downward correction in flight prices, possibly due to a combination of an intervention by South Africa's Competition Commission and the airlines' realisation that they would not be able to reach sufficient bookings for their flights at the inflated fares. However, even after this correction prices for flights were consistently at least 50% higher than in the non-World Cup periods.

Flight prices for the other connections between the major European airports and Johannesburg had similar developments to the one describe for Frankfurt. We restrict ourselves to providing an additional Figure 3b for London.1 The price crowding-out effect indicated by these figures may have been biased downwards because our statistics exclusively use the cheapest flights and do not control for the number of seats available for this price.

An additional factor that could have constrained demand was the relatively high local tourist sector prices (in dollar terms). To illustrate this factor, Figure 4 shows the Econex/Portfolio real price index in rand and dollar terms. Rand appreciation over recent years also led to higher dollar prices for international tourists.

Figure 4: Rand and US dollar tourist sector price index

Further, higher than normal hotel prices may have affected the number of tourists visiting for the 2010 FIFA World Cup. As with the 2006 FIFA World Cup, there is clear evidence that the inflexibility of the tournament schedule improved the pricing power (in local currency) of hotels more than is usual for the season, and that improvement is reflected in the average room rates (in rand, adjusted for inflation) reported in Table 4.

Table 4:  Average room rates per day – rand constant 2006 prices

	Cape Town	Durban	Gauteng
June 2006	649	549	544
June 2007	724	626	670
June 2008	805	733	810
June 2009	909	832	959
June 2010	1983	1429	2182
Benchmark month*	1275	827	1461
Tournament month*	2670	1254	3075
*The benchmark and tournament months were calculated as in Table 2.
Data source: STR Global.

Although the hotel prices reported above were the actual prices, the decision to visit South Africa was made weeks and months ahead of time based on prices quoted at that time. We recorded the lowest quoted prices for hotel rooms collected on a daily basis since 30 April 2010 via the publicly available internet portal www.hrs.de (the data from this site are similar to those obtainable from www.hotels.com). Figure 5 shows that the lowest prices for accommodation in Johannesburg for 9 May and 9 August (both dates near to but outside of the 2010 FIFA World Cup period) were more or less constantly between €45 and €50 per night over the entire data period. By contrast, during the World Cup period it was not possible to book accommodation for less than €120 per night at any point after late April. Advertisements for the cheapest accommodation increased to approximately €700 for the night of 25 June if booked one month in advance. Accommodation suppliers then obviously realised that they would not find sufficient bookings at such a high price and reduced supply prices, but subsequently it was still not possible to book accommodation for 25 June for less than approximately €260. Similar observations are available for the other 2010 FIFA World Cup cities. The same applies to the price quotations for car rentals.2

Figure 5: Prices for hotel accommodation in Johannesburg for the dates 9 May, 12 June, 25 June, 11 July and 9 August. Seven-day moving average of the cheapest prices between 30 April, 2010 and 3 August, 2010, in euros

Air carriers, car rental agents and hoteliers enjoyed an enhanced price environment during the 2010 FIFA World Cup, with rates doubling and tripling. Restaurants are also likely to have profited handsomely, though we have no data to confirm this.

The negligible increase in the capacities of air carriers in combination with the sharp price increases depicts an inelastic supply and high windfall profits for the sector. It remains unclear, however, whether the modest supply reaction by airlines was due to (i) internal information that, on the basis of their pricing strategies, no additional supply would be needed or (ii) that the inherent logistical costs of changing the flight schedules would have been too high to increase supply (personal communication, Joachim Hunold, CEO, Air Berlin, 29 July 2010).

The economy of a host nation may well experience increased income from such price effects, that is, higher prices may be accompanied by higher salaries. But slack labour market conditions made it unlikely that labour would benefit greatly. Indeed, non-agricultural wages in South Africa increased on average by only 1.4% during the second quarter of 2010, compared with an average quarterly increase of 3.4% over the preceding four quarters, using seasonally adjusted data from the South African Reserve Bank's quarterly bulletin.

In the case of airfares, most of the profit (with the exception of South African Airways flights) clearly went abroad. In the case of the hotel industry, no data on the capital shares of foreign investors are available, but at least in the case of the international hotel chains a high proportion of the profit did not stay in South Africa (Matheson, 2009).

The modest and sector specific economic benefits mentioned above are especially frustrating given the urgent need for poverty alleviation in a society with deep inequalities and widespread underdevelopment (see Van der Berg, 2010, for a description of these socioeconomic challenges). Against this background, winning the bid to host the tournament raised expectations of a significant impact on poverty and inequality. So optimistic were these expectations that Czeglédy (2009) identifies indications of millenarianism.

Pillay and Bass (2009), among others, also cautioned against raising expectations and articulating the expected benefits of the tournament exclusively in pro-poor terms. Cornelissen (2009) added the concern that the legacy of the tournament would reinforce existing spatial inequalities. But these words of caution did not resonate widely, and the result was frustration with the ultimately modest impact on poverty and inequality in South Africa.

In summary, according to our data on additional international plane landings, we do not see any evidence for a net increase in World Cup related overseas tourism beyond approximately 90 000 to 118 000 persons. This range is consistent with StatsSA's estimated increment of 109 600 for the tournament month shown in Table 2. Even in 2010 FIFA World Cup cities, occupancy rates were far from 100% during the event period, implying that the South African economy could not make full use of its resources.

Additionally, even this modest positive effect in June and July does not prove that the tournament had a positive tourism effect because ‘time-switcher’ effects have to be considered. In the cases of the 2004 European Cup in Portugal and the 2006 FIFA World Cup in Germany, increased numbers of international tourist in the event months of June and July were partly or fully offset by lower numbers in other months. It is conceivable that tourists who would otherwise have travelled to the host nations during May or August transferred their stay in a utility-maximising way to the FIFA World Cup months (Allmers & Maennig, 2009).

Our results confirm previous studies on the effects of sport mega-events which, almost without exception, find that the ex post measurable sport events effects are significantly below the effects claimed ex ante by their organisers and sponsors, and that these events generate scarcely any significant tourism, income or employment impact (e.g. Baumann et al., 2009, and the literature cited therein).

However, our results go a step further. They make it clear that the overly optimistic forecasts themselves may have contributed to the smaller than expected tourism effect. The over-confident expectations of the tourism industry, as reflected in its pricing behaviour, were self-defeating because they dampened the enthusiasm of potential visitors and thus also the potential increase in the number who visited South Africa for the Cup.

3. Other short-term economic effects and potential long-term effects

A feel-good effect might be the most evident positive outcome of a sport mega-event. Nevertheless, only a few studies attempt to evaluate this phenomenon of benefiting from an event without active attendance at the stadium, and they do so using the concept of willingness to pay. Before the 2006 FIFA World Cup, only one out of five Germans surveyed revealed a ‘willingness to pay’ (WTP) greater than zero for the 2006 FIFA World Cup to take place in Germany (Heyne et al., 2010). On average, the WTP was €4.26 per person which, with 82 million inhabitants, corresponds to approximately €351.5 million. After the 2006 FIFA World Cup, the proportion with a positive WTP had increased to 42.6% and the average was €10.0, amounting to €830.8 million for the whole country. It follows that this feel-good effect was one of the largest effects of the World Cup.

While the feel-good effect refers most obviously to the impact on locals, there is a related potential impact on foreigners due to an improved international perception of the host country (Du Plessis & Maennig, 2007). There is at least preliminary evidence for such awareness effects arising from South Africa's 2010 tournament. Figure 6 uses the number of Google hits as a proxy for awareness and plots the index of the number of Google hits for the search words ‘South Africa’, ‘Germany’, ‘Namibia’ and ‘World Cup 2010’ for each day from 19 April 2010 to 3 August 2010. The data are a visual representation of the rising international awareness of South Africa, showing an increase of approximately 60% during the 2010 FIFA World Cup period compared with April 2010. The same measure of awareness for Germany, possibly due to its national team performance, increased by up to approximately 40%, whereas for Namibia it did not increase by more than 20% for any day in the observation period. The biggest increase in awareness was for the 2010 FIFA World Cup tournament itself, which might suggest that only part of the awareness of major events is directed towards the host country.

Figure 6: Index of the number of Google hits from 19 April 2010 to 3 August 2010

Google hits are a supply-side indicator, however, because Google records the number of webpages relevant to a given keyword. From a media economics perspective, the demand side is also interesting.

As a proxy, we counted the number of members of the Facebook groups ‘South Africa 2010’, ‘My South Africa’, ‘Namibia’ and ‘Super Bowl’. Figure 7 plots the index of the number of members of these Facebook groups for each day from 19 April 2010 to 3 August 2010 and shows that the number of members of ‘My South Africa’ increased by approximately 20%, well above the near-to-zero increase of the Facebook group ‘Super Bowl’. It must be mentioned, however, that the number of members of the group ‘Namibia’ increased by nearly the same amount. In addition, the number of members in the group ‘South Africa 2010’ increased by more than 170%, again relativising the awareness effect for the host country.

Figure 7: Index of the number of members in Facebook groups from 19 April 2010 to 3 August 2010

Besides awareness, another positive effect on the host country's sporting environment may apply in the long run: the positive effect of the new stadiums on spectator demand. In Germany, the ‘novelty effect’ of all stadium projects since 1963 was equivalent to a rise in spectator numbers of approximately 10% per match (Feddersen et al., 2006). In addition, the average revenue per ticket increased due to the expansion of the area for VIP and business seating, so the overall ticket proceeds may also increase. These increased receipts improve the ability of a club to acquire top players in the international market, which, in the medium term, leads to increased national and international competitiveness. The general impact of stadiums on urban development is studied by Ahlfeldt and Maennig (2010b), and for the measurement of the external effects of stadiums see Tu (2005) and Ahlfeldt & Maennig (2009, 2010a). For a discussion of the potential effects of South African stadiums, see Du Plessis & Maennig (2009), Maennig & Du Plessis (2009) and Maennig & Schwarthoff (2010, 2011).

New ‘iconic’ stadiums such as those in Cape Town, Durban, Port Elizabeth and Johannesburg may also play a role in long-term urban development. There is no standard definition of ‘iconic buildings’, but there are plenty of existing examples, such as the Sydney Opera House, which is inseparable from the worldwide image of that city, and the Guggenheim Museum in Bilbao, which has converted that previously neglected city into one of the five most visited cities in Spain. It is currently open to debate whether stadiums in general and the South African stadiums in particular fulfil the requirements for such positive external effects.

4. Conclusion

The central theme of this paper is the claim that the economic impact of international tourist arrivals during a FIFA World Cup such as the 2010 tournament in South Africa will be considerably smaller, at least in the short term, than was widely expected prior to the event. This phenomenon is not unique to South Africa but is the general result of most ex post analyses of FIFA World Cups. The focus of this paper was on accurately estimating tourist arrivals during the 2010 FIFA World Cup because the immediate economic impact must derive from these arrivals. We found a modest number of arrivals – a net increase in tourists of some 90 000 to 108 000 persons, which was much lower than optimistic expectations – which would have constrained the impact on local income and employment. This result is consistent with Du Plessis and Venter's (2010) estimated tourist arrivals, which they used to calculate that the likely short-term impact of the tournament was 0.1% of GDP.

It seems plausible that higher supply prices, in anticipation of very high demand during the event, contributed to the modest number via a crowding-out effect. Prices for flights to South Africa during the 2010 FIFA World Cup period were three times higher than normal for bookings made between the end of January and the end of March 2010. Although these prices dropped closer to the tournament, they remained at least 50% higher than normal. A similar observation can be made for hotel prices and for price quotations for car rentals. Partly because of South Africa's elastic labour supply and the near-zero capacity effects of the 2010 FIFA World Cup, the price increases translated less into higher salaries than into higher profits for capital owners and foreign investors.

Overly optimistic forecasts by ex ante impact studies of approximately 400 000 overseas visitors may have induced this pricing behaviour that damped down the potential increase in the number of tourists. This is not an argument against the appropriate use of ex ante impact studies. On the contrary, ex ante quantitative forecasts or measurements play an important role in sport event economics as in other economic areas. But it is a warning that ex ante economic impact studies with overly optimistic claims may induce self-defeating expectations. At present, it is not possible to determine the likelihood of positive medium- or long-term effects of the 2010 FIFA World Cup; on the basis of empirical studies of comparable sporting events, it appears that such effects cannot be guaranteed (e.g. Hotchkiss et al., 2003; Jasmand & Maennig, 2007; Hagn & Maennig, 2008).

Although the ‘core’ criteria of economic performance (such as income and employment) are usually the focus of impact studies, modern economics recognises other effects as well, such as awareness and image effects. Our study indicated a rising awareness of South Africa that may well be one of the largest short-term economic effects of the tournament. There is widespread agreement that the 2010 FIFA World Cup was successfully hosted by friendly, relaxed and committed South Africans. The country takes pride in having passed such an important test under the scrutiny of the world (Schifferle, 2010). The positive media coverage and visual documentation of an incident-free 2010 FIFA World Cup has given South Africa a chance to reintroduce itself to the rest of the world. The depressing stories about crime, racial tension, HIV and power shortages have been displaced by a host of positive new images. This could translate into increased tourism and increased trade and attract foreign direct investors.

A channel by which the tournament could raise the long-term growth of the South African economy is through the impact on local self-confidence. This new self-confidence is manifested in sporting affairs, as the revived plans to host the Olympic Games may indicate. Beyond this, a poll conducted among 400 South Africans shows that 85% of South Africans now have more hope for a positive future for all South Africans than they did before the tournament (Rasmussen, 2010). While the authors have little doubt about the potential for long-term positive economic effects, it is too early to find supporting statistical evidence. If our expectations also prove to be overly optimistic, then South African Minister of Sport Mankenkhesi Stofile may have been correct when he said in 2007 that ‘The memory of that tournament will be a lasting legacy’; in other words, we will have to look to less tangible benefits of such events rather than hope for substantial economic enrichment.

Notes

¹However, European data are available from the authors on request.

²Details available from the authors on request.