Measures of fiscal risk in oil-exporting countries

Abstract

The recent relatively high levels of global oil prices have led to a significant improvement in the public finances of several oil-exporting countries. However, despite the increase in fiscal buffers, medium-term risks remain high. Fiscal vulnerabilities have increased as a consequence of the substantial spending packages that have been implemented in recent years. This has raised break-even prices – that is, the price levels that ensure that fiscal accounts are in balance at a given level of spending – in these countries. This study analyses such risks and develops measures of fiscal risk stemming from oil price fluctuations. An empirical application to oil-exporting countries from the Middle East and North Africa region is included. Additionally, it is worth noting that countries with large net assets and proven oil reserves are much less vulnerable to fiscal risk than is indicated by standard measures based on break-even prices.

Keywords:

• Fiscal risk
• stochastic simulations
• oil prices
• volatility
• break-even prices

JEL Classification:

• C53
• G17
• H60
• Q38

Acknowledgements

Without any implication, the authors wish to thank Alberto Behar, Paul Cashin, Serhan Cevik, Raphael Espinoza, Davide Furceri, Joong Shik Kang, Maxym Kryshko, Aiko Mineshima, Tigran Poghosyan, Christine Richmond, Pedro Rodriguez and two anonymous referees for constructive discussions and helpful comments. The views expressed in this paper are those of the authors and do not necessarily represent those of the IMF or IMF policy.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. These buffers reflect the significant increase in foreign assets and noticeable reduction in public debt.

2. For more details, see, for instance, Caballero and Krishnamurthy (2004), Gavin and Perotti (1997), Ilzetzki and Végh (2008), Kaminsky, Reinhart, and Végh (2005), Riascos and Végh (2003), and Talvi and Végh (2005), among others.

3. See, for example, Medina (2010) and Kaminsky (2010).

4. This includes the six Gulf Cooperation Council countries (Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and the United Arab Emirates) as well as Algeria, Iran, Iraq, Libya, and Yemen.

5. Alternative oil prices, such as West Texas Intermediate and Fateh Dubai, could be easily analyzed in the same way if break-even prices for those series were available. Note, however, that these oil prices tend to exhibit a similar pattern in terms of historic volatility, and thus the results would not vary considerably from those of Brent prices presented here.

6. The U.S. CPI series was rebased so that the index is set to 100 in 2011. The analysis was also performed using the U.S. PPI series as a deflator and, essentially, similar results were obtained (available upon request).

7. These break-even prices are estimated by IMF country teams, covering each of the respective countries, and are the projected prices (for 2012–2017) as of July 2012.

8. For instance, it is used to model stock prices in the seminal Black-Scholes option pricing model. See Black and Scholes (1973) for details.

9. Intuitively, period of high volatility in high-frequency series would generate relatively large swings in oil prices, both upwards and downwards, partially offsetting each other on average (or in cumulative terms) over a period of time (say, one year). Thus, the resulting cumulative variation in oil prices over these periods is not likely to differ significantly from the cumulative variation in oil prices over periods of low volatility.

10. Note that the estimation of the geometric Brownian motion model only depends on the historical properties of the real Brent price series and is independent on the assessment of breakeven prices.

11. Note that we are using annual data, thus our risk measure relates to the probability of the annual average of crude oil prices in a given year falling below the break-even price.

12. The full set of results, for every year of the period 2011–17, are presented in Tables 3 and 4 in Appendix I.

13. Note that we are dealing here in terms of annual averages (and thus these results do not refer to the probability of Brent oil prices falling below the break-even price on any given day, or during a portion of the year).

14. Note that in those countries, that probability was equal to 100% in 2011 (see Table 3). The average price of Brent in 2011 was US$111.32 per barrel.

15. The correlation between measures I and II for the year 2012 is very high at 0.93.

16. Consistently, some countries (e.g. Kuwait, Qatar) appear to be less vulnerable to oil price shocks, while others (e.g. Algeria, Yemen) seem to be more vulnerable to such shocks. These vulnerabilities stem from higher break-even prices, which in turn are affected by a variety of factors such as spending pressures, non-oil revenues, and oil production.

17. The correlation between the estimated break-even prices in 2011 and 2017 is positive and equal to 0.54.

18. In our sample, Oman and Saudi Arabia illustrate this point well. Although Oman appears to be just slightly more vulnerable than Saudi Arabia on our two risk measures, Saudi Arabia has significantly larger net assets, and its proven oil reserves are almost 50 times larger than those of Oman. Hence, in reality, one would expect Saudi Arabia to be much less vulnerable (when compared with Oman) than what our risk measures based on break-even prices would suggest.

19. The Jarque–Bera statistic at 1.559 is assessed against a χ² distribution with two degrees of freedom. The resulting p-Value for this test is.449, implying the failure to reject the null at that significance level. See Jarque and Bera (1980) for details on this mis-specification test.

20. The resulting test statistic (TR², where T is the number of observations) is equal to.101, which is assessed against a χ² distribution with one degree of freedom, with implied p-value of.751. For more details on the properties of the ARCH test, see Engle (1982).