RDP 2024-10: How Do Global Shocks Affect Australia? 2. Related Literature

This paper sits within the broader literature on spillovers of foreign shocks to small open economies. This literature has generally found that economic and financial spillovers tend to significantly affect small open economies, and that the United States is an important source of global shocks for these countries. This finding is robust across different methodologies: VAR-related methods (Georgiadis 2015; Dedola, Rivolta and Stracca 2017; Greenwood-Ninmo, Nguyen and Shin 2021), factor models (Ha et al 2020), and panel methods (Iacoviello and Navarro 2019). Some structural models also corroborate this finding (Croitorov et al 2020), though standard dynamic stochastic general equilibrium models can have difficulty matching reduced form evidence (Justiniano and Preston 2010; Georgiadis and Jančoková 2020). These spillovers are especially large to emerging markets, the external financing of which is affected more by US monetary policy than for advanced economies (Tillmann 2016; Anaya, Hachula and Offermanns 2017).

Prior literature has generally found more modest effects of global shocks on Australian economic variables than for other small open economies, although these estimates vary substantially across studies. Brischetto and Voss (1999) estimate a structural vector autoregression (SVAR) for Australia with oil prices and US interest rates in the foreign block, and find that shocks to these foreign variables explain around 20 per cent of variation in Australian GDP and interest rates, and around 10 per cent of variation in the CPI and exchange rate. Dungey (2001), also estimating an SVAR but with a larger foreign block including financial variables, finds that foreign shocks explain a higher proportion of variation in Australian GDP – around 40 per cent after two years – but suggests that this might be an overestimate of the contribution of foreign shocks relative to the results of a historical decomposition. Nimark (2007) and Liu (2008) both estimate small open economy structural models for Australia (with the latter also constructing a sign-restricted SVAR with a foreign block), and find that foreign shocks explain a significant proportion of variation in Australian variables. Specifically, Nimark (2007) estimates that foreign shocks explain around half of variation in nominal GDP and the cash rate, and around 30 per cent in trimmed mean CPI inflation, while Liu (2008) finds that foreign shocks explain more than half of the variation in the output gap.

Post-GFC Australian literature has tended to focus more on particular foreign shocks, especially terms of trade shocks, but has also generally continued to find that foreign shocks explain a large proportion of variation in Australian variables. Jääskelä and Smith (2011) find that foreign shocks explain around 30 per cent of variation in GDP and export prices, but 70 per cent of variation in the exchange rate. Manalo, Perera and Rees (2014) find that foreign shocks explain around 60 per cent of variation in the exchange rate after 10 quarters. There is also some more recent literature showing that US monetary policy shocks have substantial effects on Australia's stock markets (Ha 2021). One reason we might expect estimates of the effects of foreign shocks on the Australian economy to have changed over time, however, is that the Chinese economy has gradually become more important to Australia (Andrews and Kohler 2005; Guttmann et al 2019). This may make models which specify a foreign block as only the United States or a group of advanced economies (as in all of the preceding studies) less appropriate now, compared to previous decades.

The broader literature posits four main channels through which shocks spill over between countries: a trade channel, an exchange rate channel, an asset price channel, and the international banking channel. The trade channel operates as economic and financial developments in one country affect demand for another country's exports (Baxter and Kouparitsas 2005). The exchange rate channel hypothesises that if a country has flexible exchange rates, then the country's exchange rate changes with developments in foreign countries. These changes in exchange rates then affect economic activity domestically through ‘expenditure switching’, as the relative price of domestic goods and foreign goods (final and intermediate) change (Frenkel and Razin 1987). The asset price channel operates via changes in foreign equity (or other asset) prices, which may affect domestic business investment as listed companies' stock valuations or valuations of direct foreign investment and domestic households' wealth change (Pilbeam 1992).

The channel we examine most closely in this paper, however, is the international banking channel. Since the GFC, this channel, and particularly financial spillovers via global banks have received particular attention. According to this channel, any shocks that alter the cost of funding for global banks can spill over to other economies as these banks adjust their leverage positions. Miranda-Agrippino and Rey (2020) show that there exists a ‘global financial cycle’ – that is, one common factor that drives significant variation across global asset prices – and that shocks originating from the United States (in particular monetary policy shocks) are central to the global financial cycle and its effects on other economies via (de)leveraging of global banks. This can be explained by the central role of the United States in the international financial system (Gourinchas, Rey and Sauzet 2019). The US dollar is not the only dominant currency for invoicing global trade but the United States is also the primary global provider of (US dollar-denominated) safe assets, with these assets trading in deep, liquid and open financial markets. Due to this important role for the US dollar, shocks originating from the United States can therefore alter the cost of funding for global banks.

Along with Miranda-Agrippino and Rey (2020), several other papers also emphasise this channel (Avdjiev, McCauley and Shin 2015; Cecchetti, Mancini-Griffoli and Narita 2017; Avdjiev et al 2019). Buch et al (2019) identify the transmission of foreign monetary policy spillovers to domestic lending in a cross-country panel VAR via within-country heterogeneity in banks' international exposures, and find significant spillovers from US monetary policy. Kearns, Schrimpf and Xia (2023) find that there is significant interest rate co-movement across countries, and that a country's financial openness is the most significant factor in driving stronger interest rate spillovers from monetary policy shocks. However, little evidence exists on how relevant this channel is for Australia.

Global banks play a lesser role in the Australian financial system than in those of other economies, which raises the question to what extent the findings from Miranda-Agrippino and Rey (2020) apply to Australia. Better understanding of the channels through which global spillovers affect Australia is consequential for policymakers, because different channels would suggest a different optimal allocation of monitoring capabilities. Particularly, a more important banking channel would suggest policymakers should pay greater attention to the interaction between domestic financial stability and macroprudential policy and the dynamics of foreign shocks.

In contrast to some other economies, the Australian financial system has historically been dominated by authorised deposit-taking institutions, in particular the ‘Big 4’ banks. On the asset side, these banks are domestically focused, and are primarily involved in traditional banking services (deposit taking and lending), with securities trading on their own accounts being only a small proportion of net income compared to global banks (Donovan and Gorajek 2011; Davis 2011). On the liabilities side, although the majority of major banks' bond funding comes from offshore bonds (making up slightly less than 15 per cent of total funding), most of these liabilities are hedged back to Australian dollars (Atkin and Harris 2023; Johnson 2022). Most of banks' funding overall is in turn either linked directly or indirectly (via hedging) to the domestic bank bill swap rate (Fitzpatrick, Shaw and Suthakar 2022; Brassil 2022). As a consequence, policymakers tend to believe that foreign monetary policy shocks are not directly reflected in Australian financial conditions via banks (Kent 2019). This situation likely differs from that of other advanced economies with more internationalised banking systems, such as those of European countries (Schnabel and Seckinger 2019). As a result, other channels such as the trade channel, the exchange rate channel or the asset price channel may be more important for Australia. We discuss these in the next section.

Given this lesser role of global banks in the Australian economy compared with other advanced economies, this paper therefore also seeks to shed light on how much global shocks matter for Australia and whether the Australian banking system is indeed a significant source of transmission for global shocks.

Finally, this paper builds on the literature on FAVAR models. FAVAR models originated as a way of summarising information in a domestic economic context that policymakers may respond to, without explicitly including all relevant variables, by assuming that a small number of factors drive common variation across a large number of variables (Bernanke, Boivin and Eliasz 2005). FAVAR models have been used to assess economic and financial spillovers to several other small open economies: the United Kingdom (Mumtaz and Surico 2009), New Zealand (Karagedikli and Thorsrud 2010), Canada (Vasishtha and Maier 2013), Norway (Aastveit, Bjørnland and Thorsrud 2016) and Switzerland (Bäurle, Gubler and Känzig 2017). We largely follow this open economy FAVAR literature in constructing our model, though we also conduct some novel exercises in Section 4 to test alternative factor selections.