Why Does Sovereign Risk Differ for Domestic and Foreign Investors? Evidence from Scandinavia, Daniel Waldenström - PDF

Description
IFN Working Paper No. 677, 2006 Why Does Sovereign Risk Differ for Domestic and Foreign Investors? Evidence from Scandinavia, Daniel Waldenström Research Institute of Industrial Economics P.O.

Please download to get full document.

View again

of 20
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information
Category:

Internet & Technology

Publish on:

Views: 16 | Pages: 20

Extension: PDF | Download: 0

Share
Transcript
IFN Working Paper No. 677, 2006 Why Does Sovereign Risk Differ for Domestic and Foreign Investors? Evidence from Scandinavia, Daniel Waldenström Research Institute of Industrial Economics P.O. Box SE Stockholm, Sweden Why Does Sovereign Risk Differ for Domestic and Foreign Investors? Evidence from Scandinavia, * Daniel Waldenström November 20, 2006 Abstract Recent theoretical models suggest that the costs governments face when defaulting on their domestic and external debt may differ considerably. This paper examines if this proposed cost difference is reflected in sovereign risk spreads across domestic and foreign markets. Specifically, I analyze market yields on Danish government debt in both Denmark and Sweden during , i.e., a period full of political shocks as well as a wartime segmentation of Scandinavian capital markets. By linking the exogenous wartime shocks to changes in the costs of defaulting on domestic and external sovereign debt, it is found that these costs explain a significant part of the variation in the sovereign risk spread across markets. The result is robust to a multitude of tests and the inclusion of additional yield spread influences such as differences in macroeconomic fluctuations, portfolio allocation opportunities, local risk aversion and microstructure institutions. JEL classification: F34, G15, G18, N20, N24, N44. Keywords: Sovereign risk, Investor heterogeneity, Domestic debt, External debt, Market segmentation, Political economy, Cliometrics. * I would like to thank Timothy Guinnane, Mariassunta Giannetti, Magnus Henrekson, Naomi Lamoreaux, Andreas Madestam, Kim Oosterlinck, Jean-Laurent Rosenthal, Jonas Vlachos, Eugene White and seminar participants at UCLA, Stockholm Institute for Financial Research, Stockholm School of Economics and Solvay Business School for comments and suggestions. Financial support from Jan Wallander s and Tom Hedelius Research Foundation is acknowledged. Research Institute of Industrial Economics, P.O. Box 55665, SE Stockholm, Ph , Fax , 1 1 Introduction Domestically issued domestic-currency government debt has become a major source of public finance for many emerging market economies over the last years (Reinhart et al., 2003, Table 14). Despite this development, standard models of sovereign debt only focus on external government debt, issued in foreign currency to a foreign market, when assessing the costs of default, e.g., deteriorated credit reputation on international capital markets (Eaton and Gersovitz, 1982; Kletzer and Wright, 2000), risk of triggering various forms of direct sanctions (Bulow and Rogoff, 1989a,b) and curbed trade flows (Rose, 2005). In a more recent sovereign debt default literature, however, researchers argue that the costs of domestic defaults are both quantitatively as important as the costs of external debt defaults and different in their nature. Specifically, Drazen (1998) and Di Gioacchino et al. (2005) characterize domestic creditors as belonging to the constituency upon which the government relies for its political support. Unlike external debt holders, domestic debt holders are thereby able to credibly threaten to punish sovereigns in case of a domestic default by refusing re-election. Similarly, Gelpern and Setser (2004) argue that local elites represent another group of large domestic government bondholders and they may be expected to exert all their influence to prevent the government from repudiating its domestic debt, possibly pushing it to instead repudiate its foreign loans. 1 Taken together, these models can be seen as complementing the traditional models cited above, their main message being that if the groups holding domestic and external debt, and their means to punish a defaulting sovereign, differ the expected costs of domestic and external defaults will most likely also differ. Past empirical studies of sovereign defaults have primarily focused on countries external government debt, some examining the explanatory power of the standard models (e.g., Obstfeld and Taylor, 2003; Rose, 2005; Mitchener and Weidenmier, 2005) while others have searched for additional institutional or political factors to either explain changes in market-based risk assessment of external debt (e.g., Mauro et al., 2003; Hilscher and Nosbusch, 2004) or the incidence of actual external debt defaults (Kohlscheen, 2003; Bordo 1 Furthermore, Reinhart et al. (2003) argue that another cost of a domestic default is the possibly resulting turbulence on domestic banking markets, which could hurt government finances (Reinhart et al., 2003). Roubini (2001) emphasizes moral and equity reasons for discriminating between domestic and external defaults. 2 and Oosterlinck, 2005). In Van Rijckeghem and Weder (2004), however, the default likelihoods of both domestic and external debt are investigated. They find a variation in the degree of influence from a collection of political institutional variables (e.g., democracy vs. dictatorship, parliamentary vs. presidential electoral systems), but primarily that they only bind in conjunction with severe macroeconomic stability (e.g., low levels of inflation). Tomz (2004) studies how individual voter preferences correspond to the expected distributional effects of a domestic debt default, finding that they are well in line with the political default cost -channel proposed by the model of Drazen (1998). This paper examines how changes in the expected costs of defaulting on domestic and external debt influence the spread in sovereign risk between domestic and foreign markets. To my knowledge, this is the first time such an analysis is conducted. The study rests on a unique historical episode, World War II, when a series of exogenous political shocks both shifted the relative cost of domestic and external defaults and abruptly segmented international capital markets. While the exogeneity of the cost shocks is important for the interpretation of the measured effects, capital market segmentation is necessary for being able to contrast the different theories on external and domestic default costs, as it ensures a stable linkage between creditor nationality and debt type with no arbitrage trading or debt buybacks by borrowing governments. The empirical analysis uses newly assembled market yields on Danish government bonds traded on Danish and Swedish markets during to calculate a Danish sovereign risk spread over debt types. This is done by separating out all non-risk related yield influences on the spread, such as market differences in macroeconomic fluctuations, portfolio diversification opportunities, risk aversion and market microstructure regulations. The resulting sovereign risk spread is regressed on the exogenous variable containing the wartime shocks that, arguably, significantly influenced the costs of defaulting on the domestic and external debt that the Danish government expected to bear. 2 The paper proceeds as follows. Section 2 describes the institutional settings in the Danish and Swedish secondary bond markets, including details on pricing, trading and regulatory changes in the study period. Section 3 presents the basic empirical methodology and Sec- 2 There were trading halts on the Copenhagen Stock Exchange in September 1939 and April-May 1940, but Danish financial newspapers reported the bond quotes on the curb market which were regarded as representative (see section 2). Moreover, the analysis in section 6 shows that the impact on government yields of these circuit breakers was marginal. 3 tion 4 discusses the data. In Section 5, the main results along with a robustness analysis are presented. Section 6 concludes. 2 Institutional setting This section outlines the institutional framework of the Danish and Swedish secondary government bond markets in the late 1930 s and 1940 s. The main messages are the following. First, these markets basically functioned freely throughout this period, despite the wartime regulations that were particularly experienced by the Copenhagen Stock Exchange. Second, the amount of trading in Danish government bonds in both markets was sufficiently large to assert that there were sufficient levels of market liquidity in these loans. Third, the degree of Danish-Swedish capital market integration varied considerably during the period. During the interwar period, both markets were fully integrated with each other, but the outbreak of World War II and capital controls imposed by both countries shortly thereafter abruptly put a halt to this and instead completely segmented the markets. This implies that during the war, only domestic (Danish) investors were allowed to trade Danish domestic debt in Denmark and only foreign (Swedish) investors were allowed to trade Danish external debt in Sweden. Table 1 presents details on the size and activity of the Danish and Swedish bond markets as well as some specific regulatory changes concerning the pricing of bonds. The Copenhagen market was considerably larger than that in Stockholm, listing more bond loans and reporting an overall larger traded volume. In particular, a large trading volume is important for there to be enough market liquidity and, in turn, roughly efficient prices (Silber, 1991; Longstaff, 2005). Judging from the statistics of the exchange-based turnover and estimates of the trading in the Swedish over the counter (OTC) market, which is a traditionally important venue for bond trading, the amounts traded were significant in both countries and, in particular, in the individual government bonds analyzed in this study. Most market microstructure regulations were the same in Copenhagen and Stockholm during most of the period. As shown in Table 1, a major exception was the practice of circuit breakers, mainly in Copenhagen. These included two trading halts, the first between September 1 and 10, 1939 and the second between April 9 and May 22, More importantly, price limits were imposed on bonds after the outbreak of the war. Between September 1 and 22, 1939, the prices of bonds (and stocks) were prohibited to move two percent- 4 age points under their prewar (August 30) closing price. Between September 23, 1939 and February 23, 1946, the price limits were relaxed to restrict daily price falls beyond one percentage point. 3 The Stockholm market also practiced price limits during this period, but only at the beginning of the war (September December 1939) and with much wider bands of the allowed price changes (beyond five, and often more, percentage points). Hence, the Stockholm limits were much less restrictive than those in Copenhagen and from historical sources, they only seem to have been binding on a few occasions. 4 The degree of capital market integration between Denmark and Sweden varied dramatically in the study period. From being fully integrated during the interwar era, capital controls imposed by the governments in both countries effectively stopped almost all crossborder capital flows between these two and all other Western countries. 5 This development is depicted in Figure 1. That the wartime segmentation of Danish and Swedish capital markets was not overridden by some third-country market is also indicated in Figure 1 by the instantaneous drops in the capital flows to and from both Great Britain and the United States. The primary market supply of Danish domestic and external government debt may influence secondary market yields. According to standard asset pricing theory, an increased supply increases the government s debt service load on fiscal inflows which, all else equal, increases the sovereign risk. An increased supply also pushes the secondary market prices of government bonds down or, equivalently, the yields on government bonds up. In other words, increasing the relative stock of domestic to external debt hence decreases the sovereign yield spread, if the latter is beforehand perceived as more risky than the former. 6 Historically, the Danish government issued four new domestic and no new foreign loans during Accordingly, the isolated primary market effects on Danish sovereign 3 On all these events, see the Danish financial weekly Finanstidende Sep. 6, 1939, p. 1020; Sep. 27, 1939, p and Feb. 27, 1946, p Algott (1963, pp. 182ff) argues that the stock exchange board generally accepted daily price drops well beyond the official limit of five percentage points, especially for foreign-issued bonds. The period when the limits are said to have been the most restrictive is December 1939 and then mainly Finnish government bonds were concerned. 5 Although the capital controls were not imposed until January 1940 (see Table 1), the drop in recorded cross-border flows to and from Sweden (Figure 1) was observed already in the last quarter of During that period, the flows were only 1.8 percent (outflows) and 3.8 percent (inflows) of the levels in the fourth quarter of The empirical literature contains ample evidence on a positive relation between debt to GDP and sovereign yield spreads (see, e.g., Hilscher and Nosbusch, 2004). 7 Statistics Denmark (1969), pp yields should have been positive (i.e., yield-increasing) in the case of domestic debt and unchanged in the case of external debt. Politically motivated interventions in government bond markets represent another potential source of distortions of observed market returns. The most obvious case would be the government commanding the Danish central bank Nationalbanken to purchase government debt, whenever its price was falling more than what is wanted. Taking stock with the available statistical and anecdotal evidence, however, such non-market interference was infrequent and, in any case, not effective in the long run. The Danish bonds held by Nationalbanken only increased marginally between June and December 1939, and even decreased during the most critical period, December 1939 and June 1940 (Svendsen, 1968, p. 16). Furthermore, the monetary policy issues addressed by Nationalbanken itself in the early war years rather concerned how to prevent interest rates from falling too much in the light of the abundant liquidity levels in the Danish economy. 8 In other words, the Danish central bank worked to raise, not reduce, market interest rates during the war period. 3 Estimation methodology The estimation approach is based on relating differences in sovereign risk between domestic and foreign markets with changes in the relative cost of defaulting on domestic versus external debt. For this purpose, one needs to separate out all influences on observed sovereign yield spreads, both nominal and real, that are not associated with default risk, e.g., macroeconomic fluctuations (expected depreciation and inflation differential), portfolio allocation (market interest rate and stock market return differentials) and differences in institutional market microstructure constraints. 9 Specifically, the variables of interest are: Macroeconomic fluctuations: I use a set of parity relationships from international economics to define the influence from expected exchange rate depreciation (on nominal yield spreads) and expected inflation rate differences (on real spreads). The uncovered interest rate parity (UIP) states that nominal yields should be equal across markets, once expected depreciation is controlled for. The real interest rate parity (RIP) states that real yields 8 In fact, new treasuries and long-term government bond loans were issued and higher cash reserve ratios for the banking system were imposed (Johansen, 1986, pp ; Hoffmeyer, 1968, pp , ). 9 All identical yield determinants across debt types and markets, such as the term premium in this paper, cancel out entirely. 6 should equalize across markets (at least in the long run ). 10 While the concepts are admittedly stylized, they are widely used in studies of market integration and interest rate differentials as benchmarks for understanding the role of macroeconomic factors on these issues (see, e.g., Jackson and Lothian, 1993; Lothian, 2001; Lothian and Wu, 2005). Portfolio allocation factors: The cost of holding fixed-income securities depends on the fluctuations in market interest rates (the interest rate risk) and the return on other investments in the market (Cuthbertson, 1996, ch. 9). These influences are accounted for by including market differentials for the market interest rate and stock market return. Institutional differences: Changes in market microstructure constraints, e.g., price limits and trading halts (Charemza and Majerowska, 2000) and illiquidity (Silber, 1991; Longstaff, 2005), can have sizeable effects on recorded asset returns. Section 2 showed that most bond market institutions (e.g., taxes on cash flows, commission fees and market liquidity) were either the same in Denmark and Sweden or did not change during the study period, implying that they enter as constants in the empirical estimations. As for the circuit breakers in Copenhagen, a separate analysis in Section 5.1 shows that they had no lasting effect on sovereign spreads. Costs of defaulting on domestic and external government debt: The introduction discussed two literatures on sovereign debt default, one exclusively focusing on the external default costs of deteriorated international credit reputation (e.g., Eaton and Gersovitz, 1982; Kletzer and Wright, 2000), direct sanctions (Bulow and Rogoff, 1989a,b) or curbed trade flows (Rose, 2005), and another that specified domestic default costs, mainly in the form of political punishments by deprived domestic creditors also being voter groups (Drazen, 1998; Di Gioacchino et al., 2005) or local special interests (Gelpern and Setser, 2004). In order to analyze whether these costs really influenced the market-assessed sovereign risk spread between Danish external and domestic debt during World War II, I use the fact that they shifted significantly, to different degrees, following some of the major political wartime shocks hitting Denmark and the Danish government. Specifically, by applying the theoretical sovereign debt models to the political and economic history of Denmark I deduct 10 The RIP result draws on a combination of UIP and the relative purchasing power parity (PPP), which states that expected depreciation should correspond to the differences in expected inflation. 7 the shift in the relative cost of defaulting on domestic versus external debt following each of these severe shocks. 11 Table 2 displays the result of this theory- and history-based classification. During the initial pre-war period, Denmark faced high default costs on all debt types. External default costs were high, since Denmark had a default-free credit history and hence, potentially high reputational costs in terms of more expensive future borrowing. Domestic costs were also high, since bondholding was widespread among the Danish people and, accordingly, the creditors were likely to be a large part of the government s constituency. 12 The first severe political shock came with the outbreak of World War II, when external default costs were reduced for two reasons. First, the reputational costs of a default are likely to be smaller if the default is driven by exogenous fiscal shocks, as during wars, and not purely by the will of sovereigns. Second, historians report that the Danish government disliked the fact that in early 1940, the Swedish government refused to promise to support of Denmark in case of an attack (Lidegaard, 2005, p. 150). The domestic default costs remained high, however, since the economic difficulties caused by the war made the Danish people more inclined to check that the government did not try to inflate away public debt, e.g., by printing extra money. 13 The German occupation on April 9, 1940 profoundly changed the Danish political situation. Although the Danish government remained in charge of most fiscal and political issues, the Germans taxed the country heavily and took over the residual control
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks