The effects of geopolitical risk and economic policy uncertainty on dry bulk shipping freight rates

ABSTRACT

We examine the effects of geopolitical risk (GPR) and economic policy uncertainty (EPU) on shipping freight rates using a Bayesian VAR model. A positive shock to global GPR has an immediate positive, but gradually diminishing, effect on dry bulk shipping freight rates. This effect is driven by global rather than country-specific GPR shocks. Positive shocks to EPU indices for the U.S., Brazil, and China trigger a negative response of dry bulk shipping freight rates that builds gradually over several months. Historical cumulative effects of both GPR and EPU shocks on freight rates can be large and of different signs during different subperiods. Our results are important for both shipowners and charterers when fixing chartering strategies and prioritizing investments in newbuilding or second-hand vessels.

KEYWORDS:

• Geopolitical risk
• economic policy uncertainty
• shipping freight rates
• bayesian VAR

JEL CLASSIFICATION:

• C1
• D80
• R40

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ In addition, the riskiness of operating in the shipping industry may be due to other factors, such as segmentation of shipping markets (Kavussanos 1996; Tsouknidis 2016) or poor corporate governance mechanisms (Andreou, Louca, and Panayides 2014). For a comprehensive review of related shipping finance literature, see Alexandridis et al. (2018).

² In several instances throughout this paper, we refer to earnings and freight rates interchangeably, in order to make the analysis and discussion easier to follow.

³ For an analysis of the interplay between supply, demand, and freight rates, see Stopford (2009).

⁴ A study related to the shipping industry that uses the GPR index is Kotcharin and Maneenop (2020). They document that shipping firms increase their cash reserves after a GPR shock, possibly to protect against cash flow risks.

⁵ Caldara and Iacoviello (2019) provide further detailed information about the construction of the GPR index and its subcomponents.

⁶ Baker, Bloom, and Davis (2016) provide further details about the construction of the EPU indices.

⁷ GPR indices are available at: https://matteoiacoviello.com/gpr.htm (last accessed April 2020); EPU indices are available at: https://www.policyuncertainty.com (last accessed April 2020). As a robustness test, we replaced EPU-US with the global EPU index in our estimations, which, however, is available only from 1997 onward. Our results remain qualitatively the same due to this change, and are available from the authors upon request.

⁸ The sample period begins at the earliest possible date when taking into account the availability of all variables.

⁹ Dry bulk cargo vessels are categorized by capacity as follows: capesize 100,000+ deadweight tonnage (dwt), panamax 60,000–100,000 dwt, handymax 40,000–60,000 dwt, and handysize 10,000–40,000 dwt.

¹⁰ The dry bulk sector involves the transportation of homogeneous dry and wet bulk commodities – typically raw materials such as crude oil, iron ore, grains, coking and thermal coal, bauxite, and alumina, etc. – on non-scheduled routes on a one ship-one cargo basis. In 2019, dry bulk vessels carried more than 60% of the world’s seaborne trade measured in ton-miles. Capesize vessels are used primarily in the trade of iron ore and coal commodities (UNCTAD 2019). UNCTAD (2019) also cites the top importers of iron ore as China, Japan, South Korea, and Taiwan; the top exporters of coking coal as Australia, Canada, and the U.S.; and the top exporters of steam coal as Australia, Canada, Colombia, and Indonesia (World Coal Association). Therefore, the shipping trade routes among these countries are the major routes in the seaborne coal and iron ore trades. Similarly, panamax vessels have a large concentration of trade in the shipping route of the U.S. (the top exporter of grains) to Asian countries, such as China, India, and Japan, who are the top importers of grains. Since the GPR index is not available for Australia, we examine Brazil, the U.S., and China, which represent a sizable part of the shipping trade conducted by capesize and panamax vessels.

¹¹ According to Clarksons (2013) (‘Sources & Methods for the Shipping Intelligence Weekly’), daily net freight rates (earnings) for each route are computed as the net of total revenue minus the bunker costs based on prices at representative regional bunker ports, minus the port costs after currency adjustments and total commissions, divided by the number of voyage days. Details of these calculations and their constituent parameters and assumptions are in Annexes 1-4 of Clarksons (2013). For bulkers, average earnings for each ship type are averages of the voyage earnings for selected routes. The constituent routes are in Annexe 4 (b) of Clarksons (2013).

¹² In a robustness test, we also estimate the model in a standard VAR framework. The results are qualitatively similar and available from the authors upon request.

¹³ Our results remain qualitatively the same if vessel earnings are not deflated or de-seasonalized.

¹⁴ The Minnesota prior was originally developed by Litterman (1986) at the University of Minnesota and the Federal Reserve Bank of Minneapolis, and imposes a random walk representation for all variables. This seems to be a reasonable assumption for the prior for most macroeconomic variables, except those characterized by substantial mean reversion (Auer 2019). However, shipping freight rates exhibit substantial mean reversion (Kyriakou et al. 2018). Moreover, the Minnesota prior imposes a fixed and diagonal covariance matrix of the residuals, which rules out possible correlations among residuals of different variables. Kadiyala and Karlsson (1997) and Robertson and Tallman (1999) suggest that a normal Wishart prior retains the principles of the Minnesota prior, but relaxes the assumptions on the covariance matrix structure of the residuals.

¹⁵ In robustness tests, our estimates remain almost identical when we change the ordering of the variables in the BVAR model, i.e., the EPU index first and the GPR index second.

¹⁶ Each observation of a variable does not generally coincide with its unconditional mean. This is because, in each period, the structural shocks realize and push all variables away from their equilibrium values.