This chapter provides an overview of the Swedish non-financial corporate bond market with respect to size, risk profile and issuer characteristics. It explores the nature of the investors in the market and describes some of the most common covenants used by Swedish corporate issuers. The chapter also includes a description of secondary market activity and a case study of the COVID‑19 pandemic’s effect on the Swedish non-financial corporate bond market. The last section offers an overview of the real estate sector’s use of corporate bond markets.
The Swedish Corporate Bond Market and Bondholder Rights
![](/adobe/dynamicmedia/deliver/dm-aid--85fa64e0-2401-482b-b39a-d3c47d23c0a3/c443352d-en.jpg?quality=80&preferwebp=true)
1. The Swedish corporate bond market landscape
Abstract
1.1. Market size
The Swedish corporate bond market for non-financial companies is relatively young and has changed significantly over the past two decades. Between 2000 and 2008, it did not grow substantially and was to a large extent the territory of large, established companies within a limited number of industries. However, in the years after the 2008 financial crisis, notably as access to bank lending diminished, the market began growing in earnest. Between 2000 and 2008 issuance by non-financial companies averaged USD 6.0 billion annually, a number which more than doubled to USD 12.3 billion in the period between 2009‑21 (Figure 1.1, Panel A). As a comparison, the average annual amounts of total equity (IPOs and SPOs) issued by non-financial companies during these periods were USD 6.6 billion and USD 8.3 billion, respectively. In other words, average annual bond issuance has surpassed annual average equity issuance. This is notable given the prevalence of equity financing in Sweden, which had the highest equity market capitalisation to GDP ratio and the highest net stock market listing in the EU in the period from 2010 to 2018 (Oxera, 2020[1]). However, the size difference between corporate bond and equity markets remains much smaller in Sweden than in larger markets. For example, while average annual bond issuance has been 1.5 times larger than total average annual equity issuance in Sweden since 2009, the equivalent figure in the United States is 5.5 times. Nevertheless, the fact that bond issuances has overtaken equity in size is significant and illustrates the pace at which the Swedish market is growing. At the end of 2021, the total amount of outstanding non-financial corporate bonds was USD 77 billion, more than twice the amount in 2008 (Figure 1.1, Panel B).
Figure 1.1. The size of the Swedish non-financial corporate bond market
![](/adobe/dynamicmedia/deliver/dm-aid--b369525a-7753-40fc-8ac0-71969cd77553/image3.png?quality=80&preferwebp=true)
Source: OECD Capital Market Series dataset, Thomson Reuters Datastream, see Annex for details.
Debt securities, notably corporate bonds, have also increased as a share of companies’ total debt financing, meaning market-based debt financing has increased faster than bank loans. After having remained remarkably stable between 9% and 11% from 2000 to 2012, the share of debt securities in total debt financing had grown to 16% in 2020, the vast majority of which is made up by long-term securities (Figure 1.2, Panel A). This figure is slightly higher than in the Euro Area, but lower than other parts of the world where market-based financing is more developed (see further discussion under Chapter 3). Sweden remains a largely bank-dependent economy, with loans representing 85% of non-financial companies’ aggregate debt financing. Forty-seven percent of total debt financing is made up by inter-company loans, with the remaining share split between short-term loans (21%) and long-term loans (17%) (Figure 1.2, Panel B).1
Figure 1.2. Swedish non-financial companies’ use of debt securities and bank loans
![](/adobe/dynamicmedia/deliver/dm-aid--7ea70e35-694f-4412-8d9f-2661d61c316b/image4.png?quality=80&preferwebp=true)
Note: Panel A shows the share of debt securities (long- and short-term) in total debt financing (sum of total loans and total debt securities) for non-financial companies. Securities with original maturities below one year are classified as short-term.
Source: Swedish Financial Accounts from Statistics Sweden (SCB).
1.2. Risk profile of the corporate bond market
Looking at the risk profile of the Swedish corporate bond market reveals several notable trends. Firstly, as seen in Panel A of Figure 1.3, the credit quality of Swedish bond issuances has been decreasing, both since 2000 but more notably since 2009 when the market began growing substantially. For bonds issued in 2021, the average value‑weighted rating was just slightly above BBB-, the lowest investment grade rating. It has only ever been lower in 2004 and 2016. This structural decline in credit quality follows a similar trend globally, as is clear from the figure. To an extent, this is driven by investors trading off credit protection for higher yields in the general low-yield environment that has prevailed following the extensive expansionary monetary policies conducted by a number of central banks across the world since the 2008 financial crisis, and more recently in response the euro and COVID‑19 crises. Notably, the Swedish Riksbank was the first central bank to bring its main repurchase rate into negative territory in early 2015.
Another notable development is the prevalence of BBB rated bonds (again, the lowest investment grade category) in the composition of investment grade issuance. In 2021, BBB rated bonds made up almost 76% of total issuance, up from an average of less than 11% from 2000‑08 (Figure 1.3, Panel B). By far the largest corresponding decrease has taken place within the A grade category. This increase in the weight of the lowest credit quality category within investment grade issuance is also in line with global trends, although the Swedish share of BBB rated bonds is significantly higher than the global figure.
Importantly, a substantial share of Swedish bond issuances do not have credit ratings. Of the total number of issues between 2000 and 2021, 46% were unrated. Of the rated bonds, 83% were investment grade (Figure 1.3, Panel C). It bears mentioning that when looking at amounts rather than number of bonds, the share of unrated bonds is significantly lower at 19%. This is expected, since larger issuances typically target a broader, often international investor base that require the bonds they invest in to have credit ratings. According to the Swedish Riksbank (2014[2]), in 2014 about two‑thirds of the unrated bonds issued in Sweden were considered investment grade by the banks. However, this practice of banks supplying so‑called “shadow ratings” has since been discontinued, after ESMA ruled it constituted a breach of the Credit Rating Agencies Regulation (CRAR) and consequently fined five major Nordic banks (ESMA, 2018[3]). Having a credit rating from one of the main rating agencies is often a precondition for accessing capital from many types of institutional investors, who use credit ratings as an aggregate risk management tool for large portfolios and do not necessarily follow business models that allow for detailed due diligence of individual bonds. Certain institutional investors, notably pension funds, are also constrained by regulation to holding bonds over a certain rating.
One reason for the sizeable share of unrated bonds in Sweden is likely the fact that obtaining credit ratings from the large, international rating agencies carries a significant cost that may be unaffordable to smaller issuers. For example, S&P Global Ratings has disclosed that credit ratings for most transactions involve a fee of up to 7.1 basis points of the total transaction value, with a floor of USD 110 000, effectively meaning that any issue below USD 155 million will carry a cost above 7.1 basis points.2 For a bond issue of USD 60 million – the median size in Sweden in 2021 – the cost would be 18.3 basis points. In addition, all else equal, a larger company size is associated with a higher rating, possibly further discouraging smaller issuers from obtaining ratings. Scale (e.g. total sales) is one of the five factors Moody’s uses to assign its ratings (OECD, 2021[4]).
Recognising the existence of these dynamics favouring larger issuers as well as the need for an active market for research on credit worthiness of smaller companies, some countries have implemented systems where ratings are provided domestically at reduced cost. This is done with the understanding that smaller size issues are typically not intended for large, international investors who would require a rating from (at least) one of the established agencies, but that it is beneficial to have easily accessible information that can help investors gauge default risks. For example, in France the Banque de France provides a form of credit score for individual firms for a fee through the FIBEN system (OECD, 2020[6]).
Figure 1.3. Credit ratings of Swedish non-financial issuances
![](/adobe/dynamicmedia/deliver/dm-aid--c406042e-d59b-4a0b-8286-6e67d2cc3230/image5.png?quality=80&preferwebp=true)
Note: In Panel C, issuance is by number of bonds.
Source: OECD Capital Market Series dataset, Thomson Reuters Datastream, see Annex for details.
At the end of 2021, the non-investment grade (“high yield”) segment of the Swedish non-financial bond market made up 12.8% of total outstanding amounts, including unrated bonds (in terms of the number of issuers, 4.1% were classified as non-investment grade in 2021). This share has increased somewhat as the market has grown. The lowest share was recorded in 2010 at 5.3%, after three years without non‑investment grade issuance in 2006, 2008 and 2009. It peaked at 16.5% in 2017 (Figure 1.4, Panel A). This is slightly below the share of outstanding non-investment grade bonds globally, which is 18.3%. One possible explanation as to why non-investment grade markets are less prevalent in certain jurisdictions relates to their insolvency systems. If insolvency/bankruptcy regimes are inefficient, financially distressed firms will resort to out‑of‑court restructuring processes, where banks typically have greater bargaining power due to, among other reasons, their size and the fact that they are concentrated lenders. This, in turn, leads bond investors to require higher credit spreads for companies that are more likely to find themselves in financial distress (i.e. non-investment grade firms) by more than would be the case under an efficiently functioning bankruptcy regime, hampering the development of a market for such bonds (Becker and Josephson, 2016[7]). Indeed, looking at a number of indicators developed by the OECD to measure the efficiency of an insolvency regime, Sweden ranks 28th among 33 countries, notably because of the system’s high personal costs of failure to entrepreneurs and a lack of prevention and streamlining (Adalet McGowan and Andrews, 2018[8]).
While the outstanding share of non-investment grade bonds has been increasing since 2010, this only partially explains the downward trend in the rating index shown in Panel A of Figure 1.3. This indicates that the main driver of lower ratings is a change in the composition of investment grade issuance, as shown in Panel B of Figure 1.3. Notably, the value‑weighted average rating in the non-investment grade category has not been declining (Figure 1.4, Panel B). If anything, it has shown a slight upward trend, although with relatively significant swings over time.
Figure 1.4. The Swedish non-investment grade corporate bond market
![](/adobe/dynamicmedia/deliver/dm-aid--3efcf958-134c-46ee-af3a-7054b594f097/image6.png?quality=80&preferwebp=true)
Note: Shaded areas in Panel B represent years with no non-investment grade issuance.
Source: OECD Capital Market Series dataset, Thomson Reuters Datastream, see Annex for details.
An important aspect of the risk level of a bond market is the maturity and repayment profile of the outstanding stock of bonds. Aside from a spike in 2015, the value‑weighted average maturity of Swedish non-financial corporate bonds has been fluctuating between approximately 6 to 10 years since 2000, with no discernible structural trend either upwards or downwards.3 Since 2018, however, maturities have lengthened somewhat, reaching 8.4 years in 2021. This is slightly lower than the global figure of 9.4 years. Since 2010, Swedish investment grade bonds have on average had 2.7 years longer maturities than non‑investment grade bonds (excluding 2015), a difference which is smaller than the global one during the same period (4.2 years) (Figure 1.5, Panel A). Since the ratio between the average maturity of investment grade and non-investment grade bonds in Sweden is similar to the global ratio, it follows that the divergence is driven by generally shorter average maturities in Sweden.
When companies have large amounts of debt coming due within a short time period, they may be exposed to refinancing risk, especially in case of general financial distress and tighter credit conditions. This can pose a risk to a market as a whole if a large share of total outstanding bonds is coming due, amplifying existing shocks and affecting the real economy. Panel B of Figure 1.5 illustrates the debt coming due within the next three years and its share of the total outstanding debt. While the amount due in the next three years has increased substantially after the 2008 financial crisis, more than doubling since 2009, it has not increased significantly as a share of total outstanding debt. In 2021, the share of debt coming due in the next three years (38%) was very similar to that ten years earlier (37%). This is in line with the (relatively) constant maturities shown in Panel A. It is worth noting that the share was at its lowest in 2009 (at 20%), when crisis-induced risk aversion likely drove investors towards safer issuers who were able to issue at longer maturities. The Swedish non-investment grade market was effectively frozen in both 2008 and 2009, with zero issuance in both years.
Figure 1.5. Maturity and repayment profile of Swedish non-financial issuances
![](/adobe/dynamicmedia/deliver/dm-aid--450915ac-2d83-4204-a806-a9ee617f18fc/image7.png?quality=80&preferwebp=true)
Note: Includes unrated bonds.
Source: OECD Capital Market Series dataset, Thomson Reuters Datastream, see Annex for details.
Between 2000 and 2021, bonds denominated in Euros made up 57% of total Swedish non-financial issuance. Along with the Swedish Krona (SEK) at 31%, these two currencies made up 88% of total amounts. The majority of the remainder was issued in US dollars. However, the currency composition of Swedish bonds has changed substantially in the last decade (Figure 1.6, Panel A). Up until 2010, bond issues in SEK were minimal (with the exception of 2007 and 2008) and the market was dominated by euro denominated bonds.4 However, as the market grew and became accessible to a larger number of companies around 2010, the share of domestic currency bonds increased sharply. Between 2010 and 2021, the share of SEK denominated bonds in total issuance averaged 42%, compared to 4% in the period from 2000 to 2009. In 2021, the share was 56% (Figure 1.6, Panel B). Panel C shows the average and median size of bonds issued in the two different currencies, clearly illustrating how bonds (and issuers) issued to the international markets (proxied by euro denominated issuance) and the domestic ones (in SEK) differ in character. The average (median) size of Swedish corporate bonds issued in Euros between 2000 and 2021 was USD 491 (409) million, more than five times the amount in SEK, which was USD 93 (64) million. With this in mind, the increasing share of SEK issuance shown in Panel B is indicative of the growing accessibility of the Swedish bond market for smaller domestic companies.
Figure 1.6. Currency composition of Swedish non-financial bond issuance
![](/adobe/dynamicmedia/deliver/dm-aid--4c5c57eb-7312-4adf-bd15-ad0c6d9ff8af/image8.png?quality=80&preferwebp=true)
Note: Panels A and B both show shares of the total amount issued.
Source: OECD Capital Market Series dataset, Thomson Reuters Datastream, see Annex for details.
Data on the exchanges used for listings are consistent with this development. Whereas international exchanges are dominant when looking at total issuance amounts between 2000 to 2021, local exchanges is the single largest category when looking at number of issues. Specifically, Luxembourg is the most important foreign exchange, listing 39% of all bonds issued over the last two decades by amount, followed by London at 20%. Local exchanges represent 17%. A relatively sizeable share of bonds are classified as unlisted (11%) or over-the‑counter (6%). Looking instead at the number of issues, 38% are listed on local exchanges (Figure 1.7, Panel A). It should be noted that there are certain differences between investment grade and high-yield bonds with respect to the most common exchanges. For example, investment grade bonds are listed on the Luxembourg exchange to much greater extent than non-investment grade ones (by amount from 2000 to 2021, 29% versus 3% respectively), and non-investment grade bonds are more commonly unlisted than investment grade ones (15% versus 8% by amount from 2000 to 2021).
Also in line with the growing prevalence of domestic currency denominated bonds shown above, Panel B of Figure 1.7 shows how the share of local exchange issuance has increased over time. Similar to the development in SEK denominated issuance, the share of locally listed bonds was effectively zero up until 2011, after which it has grown substantially, reaching 44% in 2021. It bears mentioning that while most Swedish corporate bonds are listed on an exchange, all trading takes place OTC. The local exchange does not have a mechanism for secondary market trading in corporate bonds. Refer to Section 3.3 for an overview of alternative models.
Figure 1.7. Issuance by exchange
![](/adobe/dynamicmedia/deliver/dm-aid--bbf7054d-0724-4f92-8147-fa10c318da48/image9.png?quality=80&preferwebp=true)
Note: Exchanges included in the Local category are: Stockholm/Nasdaq Nordic and Nordic MTF.
Source: OECD Capital Market Series dataset, Thomson Reuters Datastream, see Annex for details.
Looking at exchanges and currency denomination in conjunction shows more clearly the types of bonds that are issued in different markets. Figure 1.8 shows the currency composition of issuance by exchange, which differs significantly. On local exchanges, 90% of all bonds issued (by amount) between 2000 and 2021 were denominated in SEK, with Euro issuances making up the clear majority of the remaining 10%. The domestic currency share is also substantial for bonds classified as OTC, at 83%. Contrarily, on foreign exchanges euro denominated bonds represented as much as 78% of issuance, and the Swedish krona no more than 10%. Interestingly, for bonds without listings the US dollar is the single largest currency, representing 44% of issuance.5
Figure 1.8. Currency distribution of issuance by exchange, 2000‑21
![](/adobe/dynamicmedia/deliver/dm-aid--c18a4a3f-7b7e-48e5-b3a9-80d80d9340cf/image10.png?quality=80&preferwebp=true)
Note: Exchanges included in the Local category are: Stockholm/Nasdaq Nordic and Nordic MTF.
Source: OECD Capital Market Series dataset, Thomson Reuters Datastream, see Annex for details.
The share of floating-rate bonds in total amounts issued in Sweden has increased significantly in the past decade, from 18% in 2010 to 53% in 2021 (Figure 1.9). Issuing floating rate bonds can help increase investor demand, and thus liquidity, for a security since it offers holders positive exposure to higher interest rates if broader market conditions change, thus limiting to an extent the risk of being locked into lower than currently prevailing rates. However, from an issuer perspective it also increases the exposure to rate hikes which would lead to increased debt servicing costs. This can also have financial stability implications for a market more broadly if the share of floating rate debt is high. If cost levels become unsustainable for a large enough number of companies, this may in turn lead to widespread defaults and further pressure on capital markets.
Figure 1.9. Split between fixed and floating rate bonds over time
![](/adobe/dynamicmedia/deliver/dm-aid--0cf70b56-8364-4cff-ac1f-4ea317b49b38/image11.png?quality=80&preferwebp=true)
Note: The figure shows the shares of fixed and floating rate bonds by amount issued.
Source: OECD Capital Market Series dataset, Thomson Reuters Datastream, see Annex A for details.
1.3. Issuer characteristics
In addition to changes to the market at the macro level, when trying to understand how the market has developed over time it is also useful to look at changes at the issuer level. Figure 1.10 below provides an overview of the composition of issuers on the Swedish bond market during the past 20 years. Panel A classifies issuers into three groups depending on their previous experience with bond markets: first-time issuers are issuers that have never issued a bond before; returning issuers has previously issued bonds, but more than five years ago; and active issuers have issued at least one bond within the last five years. As for many of the data shown thus far, there is a clear change in trend after the 2008 financial crisis, and in particular after 2010. Between 2000 and 2010, the average share of active issuers was 63% with first time issuers representing 23%. The average number of issuers annually was only eight. In sharp contrast, the share of first-time issuers averaged 37% between 2011 and 2021, with an average of 39 issuers annually. In 2020, the number of issuers dropped sharply and the market was dominated by active issuers, indicating that in crisis times the Swedish bond market was not a fully resilient source of funding for many companies (see further discussion under Section 1.8.1 below). However, 2021 was the year on record with the largest issuer count, with 73 companies issuing bonds.
Panel B shows the corresponding change in the distribution of issue sizes as well as the median issue size over time. Similarly, up until 2009 the market was clearly dominated by large companies. Issues in excess of USD 500 million represented as much as a third of issues between 2000 and 2010. In 2007 it was as high as 62%. The two smaller categories below USD 100 million – likely the most attainable for smaller companies – made up no more than 9% on average during this period. The dominance of large companies was particularly evident during the 2008 financial crisis when only very large and creditworthy Swedish companies had access to this type of market-based financing. This led to a remarkably high median issue size, which peaked at USD 630 million in 2009 – higher than recorded in either Europe or the United States during the same period (OECD, 2021[4]). This dropped sharply after 2010, reaching USD 60 million in 2021, a decrease of more than 90% compared to roughly a decade earlier. In the same year, issues below USD 100 million made up 68% of total issuance while the largest category (above USD 500 million) only represented 5%. These developments clearly demonstrate an increased accessibility and use of corporate bond markets among smaller Swedish companies.
Figure 1.10. Characteristics of issuers/issues on the Swedish non-financial corporate bond market
![](/adobe/dynamicmedia/deliver/dm-aid--28b9b073-f309-437e-91e0-60acbd17c73f/image12.png?quality=80&preferwebp=true)
Note: A company is defined as a first-time issuer if its bond issue in a given year is its first issue since the start of our series (January 1980). A “returning issuer” is a company which made its last bond issue more than five years ago. If the company issued bonds in at least one of the past five years, it is defined as an “active issuer”.
Source: OECD Capital Market Series dataset, Thomson Reuters Datastream, see Annex for details.
The increase in the number of issuers and general broadening of the market is also reflected in the concentration of issuers. Specifically, Figure 1.11 below shows the share of the top three, five and ten largest issuers in total issuance over time. The share of the ten largest issuers has fallen sharply over the past decade, from 76% in 2012 to 51% in 2021, with the three largest issuers’ share falling from 35% to 25% in the same period. Notably, in 2020, as the COVID‑19 crisis hit, market concentration increased to 73%, similar to the levels seen back in 2012. This follows from the sharp decrease in the number of issuers during the first year of the COVID‑19 crisis shown in Figure 1.10 above. Within the investment grade category of the Swedish bond market, in 2020 only one of 132 issues came from a new issuer (Nordic Trustee, 2020[9]).
Figure 1.11. Top issuers’ share in total issuance
![](/adobe/dynamicmedia/deliver/dm-aid--5789902d-b5a3-4e97-b8d9-1de3dbd96823/image13.png?quality=80&preferwebp=true)
Note: Graph starts from 2011 since the average number of annual issuers prior to that was no more than 8. In the period from 2011 to 2021, the annual average was 39, reaching 73 in 2021.
Source: OECD Capital Market Series dataset, Thomson Reuters Datastream, see Annex for details.
As the market has grown, the industry composition of issuance has changed. However, in spite of significant variation between years, certain industries have remained dominant. Notably, industrial companies make up a substantial share of the total amount issued through corporate bonds throughout the analysed period, averaging 27% of annual issuance from 2000 to 2021, and reaching as much as 49% in 2021. Utility and consumer cyclical companies are also large issuers. Telecom companies, representing 7% of issuance in 2021, made up a significant part of issuance in the period from 2007 to 2012, averaging 29% of total annual issuance (Figure 1.12).
It should be mentioned that bonds issued by real estate companies are not included here since the analysis is restricted to non-financial companies and since OECD work on bond markets normally classifies real estate companies as financial companies. However, because of the real estate industry’s importance to the Swedish corporate bond market, this report devotes a separate sub-chapter to an analysis including real estate issuers (see Section 1.9). More generally, it bears mentioning that the financial industry represents a significant share of the Swedish bond market, accounting for as much as 77% of total issuance in 2021.
Figure 1.12. Industry composition of the Swedish non-financial bond market
![](/adobe/dynamicmedia/deliver/dm-aid--9280dd93-120c-48b0-8284-874f28c0f011/image14.png?quality=80&preferwebp=true)
Note: Real estate companies are not considered in the above graph. Please refer to Section 1.9 for an analysis including the real estate industry.
Source: OECD Capital Market Series dataset, Thomson Reuters Datastream, see Annex for details.
1.4. Corporate bond investors
Foreign investors are the largest owners of Swedish non-financial corporate bonds, holding 60% of total outstanding amounts at the end of 2020. This share has remained quite stable since 2003, when it increased from the lower levels seen in the first three years of the 2000s. The composition of domestic ownership, however, has changed. An important development is the growth of investment funds as owners since about 2012. In 2011, they held about 2% of total outstanding amounts, a figure that had grown to 17% by 2020 (Figure 1.13, Panel A). When counting only domestic ownership, they represent 43%, compared to 6% in 2011 (Figure 1.13, Panel B). As noted by Becker et al. (2020[10]), the aggregate statistics do not specify the composition of the foreign investor category. However, it is likely to include a substantial amount of investment funds. It is also notable that the share of foreign ownership decreased rather sharply in 2020 as markets globally were clouded by an uncertain outlook, falling from 65% at the end of 2019 to 60% in 2020 (representing a net outflow of over SEK 38 billion).
Monetary financial institutions (e.g. banks, money market funds and other credit institutions) represent 12% of domestic ownership, a decrease from 46% in 2009. This is significant and reflects the overhaul of the regulatory landscape, in particular for banks, after the 2008 financial crisis, notably the new Basel accords. These developments have made banks less willing to hold corporate bonds, reducing proprietary inventories which has led to a decrease in dealer intermediation (FSB, 2021[11]). Non-financial companies themselves represent 8% of domestic ownership of non-financial bonds. While direct retail (household) participation in bond markets is very low, retail investors are still exposed to the market through e.g. investment and pension funds. According to several market participants, retail investors represent a substantial share of investment fund ownership.
Figure 1.13. Ownership structure of outstanding Swedish non-financial bonds
![](/adobe/dynamicmedia/deliver/dm-aid--8ed7d22a-d534-4ef0-aa9e-3b807ab723c6/image15.png?quality=80&preferwebp=true)
Note: In Panel B, the “Other” category includes: the Swedish central bank, public administration, households and other financial intermediaries (original classification in raw data). In Panel A, the “Other” category includes all of these plus those shown in Panel B but not in Panel A (non‑financial companies and households). Data refer only to corporate bonds (not short-term securities). All values are as of year-end.
Source: Financial Accounts from Statistics Sweden (SCB).
1.5. Ownership structure of selected Swedish corporate bonds
With the aim of presenting more detailed information on the owners of Swedish non-financial corporate bonds, to gauge the level of ownership data available at the bond-level, and to distinguish who the foreign investors are, the following subsection looks at a sample of 100 of the largest bonds in Sweden. The size of bonds included in the analysis ranges from about USD 100 million to over USD 1 billion, for a sum of around USD 40 billion, representing 52% of total outstanding amounts as of 2021. As part of this analysis, over 300 bonds were reviewed, reflecting the scarce availability of information on the ownership records of individual corporate bonds. Indeed, the available ownership information for the sample of bonds included in the analysis is far from fully comprehensive. Only eight bonds have ownership data covering over 70% of the outstanding amount, and for almost half the coverage is below 20%.
Table 1.1. Ownership coverage, 100 largest outstanding Swedish non-financial corporate bonds
Ownership coverage |
Number of bonds |
---|---|
Over 70% |
8 |
[50%, 70%] |
13 |
[20%, 50%] |
33 |
Less than 20% |
46 |
Total number of bonds |
100 |
Source: Bloomberg, Refinitiv.
Panel A of Figure 1.14 shows that institutional investors own 22.5% of the total outstanding amount of bonds, while private corporations and the public sector own below 1%. The lack of publicly available information is evident from the fact that the owners of around 78% of the total outstanding amount are not reported. When classifying owners by domestic and foreign, the share of domestic institutional investors in total ownership is 7.0% and 15.5%, respectively, indicating that foreign institutional investors hold more than twice as much as their domestic counterparts (69% and 31% of total institutional investor ownership, respectively). Out of the USD 8.9 billion in the hands of institutional investors, 31% is made up by to Swedish investors, 12% by French investors and another 12% by US investors (Panel B of Figure 1.14).
Figure 1.14. Distribution of ownership by investor category
![](/adobe/dynamicmedia/deliver/dm-aid--33243f1d-babb-4b1f-9d10-dabbdfd2f238/image16.png?quality=80&preferwebp=true)
Source: Bloomberg, Refinitiv.
To gauge whether ownership patterns differ depending on data availability, Panel A of Figure 1.15 shows the ownership composition by investor category. The information is heavily skewed towards institutional investors, which is to be expected since they are normally required to disclose their portfolio composition whereas other investors are not. As seen in the figure, the higher the level of available information, the higher the share of institutional investor ownership recorded. For bonds where ownership data are available for less than 20% of the outstanding amount, institutional investors hold 7% of the outstanding amount whereas for bonds with over 70% of ownership information available they hold 88% of the outstanding amount. This suggests that bonds with scarce ownership data have a larger share of non‑institutional investors, although this is not possible to say with certainty since the unknown category by definition could include any type of owner, including institutional ones.
The level of ownership information available may also differ between bonds issued in the local market and those issued abroad. Corporate bonds with an ISIN starting with SE are issued locally and those starting with XS issued abroad. Panel B of Figure 1.15 plots the initial issuance amount against the ownership coverage of the outstanding amount. It could be expected that larger bonds would be more attractive to institutional investors which, following the logic above, should result in higher available ownership information given that many of these investors are required to disclose their holdings. However, the Swedish data show no such correlation. For bonds issued in the domestic market, there is no evident relationship between the size of the bond and the available ownership information. For bonds issued on foreign markets (ISIN starting with XS), the smaller the initial issuance amount the higher the level of ownership coverage, although it is naturally not possible to draw any causal conclusions from this (possibly spurious) correlation. Bonds issued on foreign markets are generally larger than domestic ones.
Figure 1.15. Ownership data coverage
![](/adobe/dynamicmedia/deliver/dm-aid--17e37386-e861-4a24-8bea-9d7ff0e5ff8b/image17.png?quality=80&preferwebp=true)
Source: Bloomberg, Refinitiv.
1.6. Covenant protection
Covenants are constraints placed on an issuer and are stipulated in the bond indenture (contract) at the time of issuance. They are bondholders’ main corporate governance tool and serve to ensure that issuers do not engage in activities that would reduce creditors’ claims or reduce the probability that they are repaid. A breach of covenant results in a so-called technical default (as opposed to a payment default, which occurs when an issuer does not make its interest/principal repayments in line with the contractual agreement). Previous OECD analysis (2021[4]) has shown a clear decrease in covenant protection over time for non‑investment grade bonds issued in the United States. While covenant data are not as accessible nor as accurate for Swedish bonds, Figure 1.16 shows how the prevalence of certain covenants in Swedish bond indentures has changed in the past decade. Certain trends are visible. For example, the prevalence of negative pledge covenants, preventing issuers from using encumbered assets as collateral for new borrowing (which would dilute the existing creditors’ protection), has decreased markedly since 2011. Contrarily, change in control covenants – under which a material change in a corporate ownership (definitions may vary) would trigger an obligation to repay the outstanding debt (“acceleration”) – have become much more common. The same is true for pari passu covenants, which guarantee that existing creditors are covered by potential additional guarantees the issuer offers to creditors in future borrowing. However, it should be noted that there are substantial differences in the prevalence of different covenants even between consecutive years, making it difficult to determine the exact trends. Many market participants have highlighted the lack of standardised bond terms as a particular challenge on the Swedish corporate bond market, which these differences might be an indication of.
Figure 1.16. Prevalence of different covenants in Swedish non-financial bond indentures
![](/adobe/dynamicmedia/deliver/dm-aid--a7be2025-fd3e-4c0e-94db-b49daa972c9e/image18.png?quality=80&preferwebp=true)
Note: Based on 861 bonds issued by Swedish non-financial companies between 2000 and 2021 with covenant indicators available. Only covenants that are included in at least one bond indenture are displayed.
Source: OECD Capital Market Series dataset, Thomson Reuters Datastream, see Annex for details.
1.7. Investment banks and market structure
Investment banks are important players on corporate bond markets, providing underwriting and advisory services. Their tasks are related to e.g. origination, distribution, risk bearing and certification, as well as general advice on pricing, timing of issuance and preparation of relevant documentation. As a general trend, the market concentration of investment banks underwriting non-financial bonds in Sweden has decreased somewhat in the past decade, although the trajectory has been uneven. In 2010, the top five banks had a market share of 73%, which had dropped to 57% in 2021 (Figure 1.17, Panel A). Notably, in 2021 the top five banks by market share were all Swedish or Nordic (Panel B). This is a relatively recent phenomenon. While local banks were always relatively high up in the league tables, up until 2017 the top five always included, and was often dominated by, foreign banks, notably from the United States, the United Kingdom, Germany and France. This increase in the share of local investment banks can also be observed in other regions, notably in a number of Asian jurisdictions (OECD, 2019[12]).
Figure 1.17. Composition of underwriters of non-financial bonds
![](/adobe/dynamicmedia/deliver/dm-aid--42e872dd-4f4d-4129-b43d-727dc2603a2a/image19.png?quality=80&preferwebp=true)
Note: Ranked based on gross proceeds. Panel A shows the share of the top five banks in any given year. However, the composition of the top five differs over time, meaning there can be changes in what banks are included.
Source: Refinitiv.
When a bond is issued, the issuer normally assigns an independent trustee to supervise the implementation of the bond indenture (contract). The role of the trustee, typically a bank or a specialised trustee institution, is to ensure the contract is followed, including to review instances of covenant breaches.6 In terms of specialised trustee institutions, the Swedish market is dominated by two main players, Nordic Trustee and InterTrust, and by the former in particular. However, specialised trustees are typically only assigned for non-investment grade bonds in Sweden. For the lion’s share of investment grade bonds issued in the country, the assigned trustee is a bank. In many markets, the trustee often also plays the role of paying agent, handling collection and disbursements of principal and coupon payments. This is also the case in Nordic markets such as Norway. In Sweden, however, there are no paying agents. Instead, disbursements to bondholders are typically handled by the central securities depository, Euroclear (which only provides settlements in Euros and Swedish krona). It also bears mentioning that the role of the agent is not specifically regulated, which is the case in, for instance, Finland, Denmark and Norway.7
1.8. Secondary market liquidity
A well-functioning secondary market where investors can trade bonds is an important part of the bond market more broadly, including the primary market. A liquid secondary market ensures that the price‑finding mechanism functions properly and that bondholders can exit their investments before maturity as their circumstances change. However, it should be emphasised at the outset that bond markets are very different from equity markets in a number of regards. Firstly, electronic trading, while increasing, makes up a relatively small part of the market. In particular for larger bonds, phone‑based negotiations remain dominant (FSB, 2021[11]). This is contrary to equity markets where electronic trading is ubiquitous. Secondly, liquidity is generally significantly lower in bond markets than in equity markets. Trading is concentrated in the first days after a bond is issued, and then drops significantly. Evidence from the US market suggests that trading is mostly concentrated in the first 90 days after issuance (Mizrach, 2015[13]). The vast majority of outstanding bonds do not trade on any given day. Even in the United States, where bond markets are large and relatively liquid, and even for the most traded bonds, the number of trades per day is limited (Çelik, Demirtaş and Isaksson, 2015[14]). For the 1 000 most liquid corporate bonds in the US, only 1% of the issue size traded on a daily basis in 2015, compared to 0.16% for the remaining less-liquid instruments. This means that even for the most liquid corporate bonds, it takes 100 days to trade the full issue size (Mizrach, 2015[13]).
Partly, this has to do with the illiquid nature of the instrument. Corporate bonds are very different from equities. For many investors, these instruments serve the purpose of long-term liability matching, while offering the prospect of higher returns compared to other fixed income instruments such as government bonds. This is a reason why institutional investors such as insurance companies and pension funds are significant players in fixed income markets globally. Their long-term portfolio structures also help explain why liquidity is much lower in bond markets than in equity markets – the main investor type is not a liquidity supplier but rather soaks up liquidity through buy-and-hold strategies. However, as shown in Figure 1.13 and as will be discussed further below, the Swedish corporate bond market differs somewhat with respect to institutional investor holdings given that the share of pension funds and insurance companies is relatively low. However, low liquidity also has to do with trading mechanisms. As opposed to equities, bonds are typically traded in large increments, so-called “round-lot” trades of USD 1 million (bond trades below USD 100 000 are sometimes known as “odd lots”). For example, in the United States the average bond trade size from 2014‑16 was USD 1.2 million (a decrease from USD 2.0 million in 2006‑07). This is compared to an estimated USD 3 000 to 5 000 for equities, less than 0.5% of the size of an average bond trade (Bessembinder, Spatt and Venkataraman, 2020[15]). This is because most bond markets target large institutional investors, who trade in large minimum sizes. The retail component of the market (at least for direct investments) is typically small.
Finally, bond market trading is dominated by dealer intermediation by banks, which therefore play an important part in creating liquidity. However, in many places the growth in bond dealer balance sheets has not matched the growth of the corporate bond market. In addition, partly following more stringent regulation in the wake of the 2008 financial crisis as well as increased risk aversion, banks hold lower inventories of bonds and have reduced their activities as market makers, further reducing liquidity. Riskless principal trading – where the dealer finds both a buyer and a seller before going ahead with a trade – is a more common business model than regular principal trading (FSB, 2021[11]). In 2014, 95% of secondary market trading in Sweden was made up by trading between banks and their customers (Riksbanken, 2014[2]).
The annual volume traded on the Swedish market has outpaced the growth of the market more broadly. Figure 1.18 provides an overview of annual amounts traded in the Swedish market and the turnover ratio (measured as annual traded volume as a share of the outstanding amount at the end of the year). Both measures have increased over time. The turnover ratio is similar to that of other European countries. In the European Economic Area, the turnover ratio in 2020 was 46%, compared to 43% in Sweden (ESMA, 2021[16]).
Figure 1.18. Volume traded in the secondary corporate bond market
![](/adobe/dynamicmedia/deliver/dm-aid--3ce3fb7f-2afa-4a7b-899c-e2ebb29b4b15/image20.png?quality=80&preferwebp=true)
Note: Turnover ratio is calculated as the annual traded volume as a share of the outstanding amount of corporate bonds at the end of the year.
Source: Sveriges Riksbank (SELMA Statistics), OECD Capital Market Series dataset.
Despite the positive evolution in terms of turnover, liquidity remains low when compared to other markets outside the European Economic Area. A lack of transparency in pricing and trading has led to unreliable pricing (Wollert, 2020[17]). Part of the reduced transparency in Swedish bond markets is an effect of regulation, and more specifically by the implementation of the European Union’s directive MiFID II and MiFIR in 2018, as analysis by the Swedish Financial Supervisory Authority has shown. Prior to its implementation, since 2015 there had been national rules in place related to disclosure of price and volume of bond trades. For all bond trades, with the exception of trades exceeding SEK 50 million, this information was to be disclosed at latest by 09.00 (AM) the following day. MIFiD II and MiFIR have, on the face of it, more stringent disclosure requirements, mandating such disclosure both pre‑trade (orders) and post-trade (transactions) effectively in real time. For pre‑trade disclosure, the main rule states that buy and sell bids as well as order depth should be disclosed continuously by market operators and investment firms operating a trading venue during market hours.8 However, due to exceptions included in the directive, all investment firms trading in Swedish bonds are exempt from pre‑trade disclosure (it should be noted that pre‑trade transparency was very low also prior to the implementation of MiFIR). When it comes to post-trade disclosure, MiFIR lists three conditions under which disclosure of non-equity transactions may be deferred. They apply to transactions that are: 1) larger than normal market size; 2) related to instruments for which there is no liquid market; or 3) above an instrument-specific size. For trades subject to deferrals based on these criteria, information is to be disclosed at latest 19:00 (7:00 PM) the second working day after the trade. In practice, since only a handful of Swedish bonds are considered liquid under this regulation, essentially all trades are eligible for deferrals. In August 2019, only one Swedish ISIN bond was considered liquid. The effect of this has been reduced transparency on the Swedish bond markets. According to a survey of market participants, well above 60% find that MiFID II/MiFIR has decreased transparency, with just under 30% saying it is unchanged (Finansinspektionen, 2019[18]).9
The Financial Supervisory Authority’s analysis further shows that the reduction in transparency is due to fragmentation in data provision (such as turnover data, which was previously available in one place), owing to the lack of an entity compiling all published information (a “consolidated tape provider” or CTP). However, in July 2021, ESMA made available the first CTP data, which will be published biannually (ESMA, 2021[19]). In line with the objectives of MiFIR/MiFID II, trading on regulated venues has increased. Before their implementation in 2018, effectively all Swedish corporate bonds were traded OTC. By 2019 the share had fallen to below 40%, with a substantial increase particularly in systematic internalisers (executing orders against their own books or client orders), but also on trading venues (which includes both regulated markets, MTFs and OTFs).10 Data show that in practice transactions executed by systematic internalisers are published two days after close. For transactions executed on an MTF, the delay is normally four weeks or more for all types of bonds, although in Sweden this deferral only applies to sovereign and covered bonds and not for corporate bonds. However, most MTF trades in Swedish corporate bonds are executed on platforms that fall under other EU-country regulations. Different jurisdictions apply different exceptions, and the Swedish FSA has found discrepancies between the transaction data it receives and the publicly available data, meaning the latter gives an incomplete view of the market (Finansinspektionen, 2019[18]).
In response to this, the Financial Supervisory Authority tasked the Swedish Securities Markets Association (SSMA) with investigating how to improve transparency. This resulted in a self‑regulatory recommendation on bond market transparency, which applies voluntarily in addition to the mandatory rules set out in MiFID II and MiFIR. The recommendation entails e.g. the daily publication of aggregate transaction information executed on the Swedish market, which is to be made public through a single data provider. Notably, this (self) regulation is more stringent than the national rules applied prior to the implementation of the EU directive (Swedish Securities Markets Association, 2020[20]).
It bears mentioning that while increased transparency in bond markets is definitely beneficial to liquidity, a balance needs to be struck between transparency and dealer incentives to intermediate. Typically, if a trade leads a dealer to add bonds to its inventory, it will typically want to resell these bonds in the inter‑dealer market. Forcing dealers to show their hand may compromise their bargaining position in that market, in turn reducing their incentives to engage in such intermediation. This is particularly pertinent for high‑yield, illiquid bonds (Çelik, Demirtaş and Isaksson, 2015[14]). At the same time, it is difficult to establish which way the causality runs. Limited transparency may be required to maintain dealer incentives for markets with low liquidity, but low liquidity may equally be an effect of low transparency (making pricing difficult).
1.8.1. The Swedish corporate bond market during COVID‑19
The Swedish corporate bond market’s behaviour during the COVID‑19 crisis provides a useful case study, highlighting the key challenges still facing the market, notably pertaining to liquidity. This subsection provides an overview of the dynamics as the pandemic unfolded, including subsequent initiatives to improve market functioning, both from regulators and private sector representatives.
Well-functioning capital markets serve to provide an economy with increased resilience in times of crisis, allowing companies to access financing even as risk aversion increases, typically resulting in contractions of both bank credit and consumer demand, simultaneously increasing the need for financing and decreasing the availability of capital. In 2009, non-financial companies globally issued record amounts of bonds, and the same thing happened again in 2020 as the pandemic broke out. The exact same pattern can be seen for equity markets (of particular importance are secondary public offerings, i.e. already listed companies tapping equity markets) (OECD, 2021[4]). These developments are indicative of the ability of market-based financing, not least bonds, to help economies overcome periods of financial and general economic distress. However, a prerequisite for this is that markets are sufficiently flexible, liquid and deep. When they are not, they offer little crisis resilience. For this reason, the Swedish corporate bond market did not display the same dynamics as those seen globally during the COVID‑19 crisis. Figure 1.19 shows net credit flows to non-financial companies over time in Sweden. Panel A illustrates the same broader Swedish trend seen in previous figures, namely a shift from bank financing towards debt securities around 2011. It also gives an indication of the pro- or counter-cyclicality of different types of borrowing during crises. In 2009, as bank lending contracted sharply, debt securities actually increased. However, two points are critical to consider together with these data. First, the year after, in 2010, credit flows from debt securities contracted more than bank loans in absolute terms, despite being a much smaller market segment. Second, the Swedish bond market of 2009 was very different from the one of today. As illustrated in Figure 1.10, in those years the market was dominated by large, established issuers. The median issue size was significant and the number of issuers low. While market access for such companies is indeed important, it is not necessarily a good indicator of the extent to which a market provides resilience more broadly. The COVID‑19 crisis provides a better test of this, since by 2020 the market had expanded significantly to include a larger number of companies. Panel B below shows quarterly net credit flows in 2020 and 2021. This paints a markedly different picture, with debt securities contracting significantly in both the first (while bank lending remained positive and significant) and the second quarter of 2020. These flows only turned positive again in the third quarter, notably as the Riksbank’s corporate bond purchasing programme began in September, instilling a degree of trust in the continued functioning of the market.
Figure 1.19. Net credit flows to non-financial companies
![](/adobe/dynamicmedia/deliver/dm-aid--8f9e79bb-af34-4cff-bd12-8097aad5d78d/image21.png?quality=80&preferwebp=true)
Source: Financial Accounts from Statistics Sweden (SCB).
Figure 1.20 provides a more detailed account of primary market issuance during the COVID‑19 crisis. In the early months of the pandemic, issuance was significantly lower than its five‑year average for the market as a whole (Panel A). With the exception of July and September (when the central bank’s purchasing programme began and issuance increased sharply), this held true for the full year. February and November saw similar numbers as previous years. This is in sharp contrast to global developments, where total issuance markedly exceeded historical averages, in particular in the crucial months of March, April, May and June (OECD, 2021[4]). However, as Panels B and C reveal, the dynamics differed substantially between the investment grade and non-investment grade segments on the Swedish market. Notably, investment grade issuance was significantly higher than in previous years, in particular in February and March in the early stages of the pandemic and then again in September and November. Contrarily, non‑investment grade issuance was non-existent in the first three months of 2020. It remained below historical averages in every month, with the exception of April, July and September. However, issuance in these months was driven by a handful of large bonds by no more than one issuer per month.11 In total in 2020, there were no more than five non-investment grade bonds issued by Swedish companies.
Figure 1.20. Swedish non-financial bond issuances during the COVID‑19 pandemic
![](/adobe/dynamicmedia/deliver/dm-aid--5e6801f4-f795-4db0-ba7e-0966bc4747fa/image22.png?quality=80&preferwebp=true)
Note: Panel A includes unrated bonds. Not adjusted for inflation.
Source: OECD Capital Market Series dataset, Thomson Reuters Datastream, see Annex for details.
The significant impact of the COVID‑19 crisis can be seen even more clearly in the secondary market. Figure 1.21 below shows how an index of investment grade bonds issued in SEK developed during 2020. Panel A shows how the index fell very sharply in March, gradually recovering over the year, assisted by significant fiscal and monetary support at the national and supranational levels. Panel B shows the option‑adjusted spread, which increased 2.5 times from 31 January to 24 March. These developments are both indicative of a major sell-off and liquidity crunch. These dynamics were not unique to Sweden. The S&P 500 Investment Grade Corporate Bond Index (effectively the United States equivalent of the index presented below) fell even deeper and spreads increased more than for the Swedish index. However, this should be considered with two points in mind. Firstly, substantial parts of the Swedish market are unrated (as seen in Figure 1.3) or rated non-investment grade. Since the index does not reflect these bonds, it gives only a partial picture of the market which is likely skewed towards its strongest part. Secondly, the US index recovered much faster (both in terms of performance and spread) and more strongly after the initial downturn, indicating less uncertainty about the market’s continued functioning (again aided by substantial fiscal/monetary support).
A lack of liquidity was arguably the most pressing issue on the Swedish corporate bond market during the COVID-induced crisis. During times of financial distress, supply and demand on capital markets tend to be lopsided as large parts of the market rush to sell riskier assets and to buy very secure ones (a so-called flight-to-safety). These dynamics lead to e.g. the increased spread seen above, typically coupled with a widened bid-ask spread for most corporate bonds. However, in a resilient market, the price formation process should not break down entirely in such a scenario. When it does, it is indicative of a completely one‑sided market where prospective sellers are not certain they will be able to find a buyer at all for their investments. These tendencies were visible in the Swedish bond markets in 2020 (and indeed in Europe more generally), as roughly 30 investment funds, with aggregate assets under management equivalent to some SEK 120 billion, temporarily froze owing to lacking information on closing prices and volume trading (Riksbanken, 2021[5]). In addition to domestic investment funds, foreign investors also offloaded significant holdings of Swedish corporate bonds (Becker et al., 2020[10]).
Figure 1.21. S&P Sweden 1+ Year Investment Grade Corporate Bond Index
![](/adobe/dynamicmedia/deliver/dm-aid--a8ce5e27-3239-43dd-ba78-f3ac4fda5eca/image23.png?quality=80&preferwebp=true)
Note: The index includes investment grade bonds (as determined by ratings from S&P, Moody’s or Fitch) denominated in SEK with a minimum notional outstanding amount of SEK 250 million and a maturity of more than one year.
Source: S&P Global.
Notably, these adverse developments were to an extent an effect of the ownership structure on the Swedish market. As shown in Figure 1.13, investment funds represent a very large share of domestic ownership. Similar to banks, open-ended investment funds engage in liquidity transformation, since they offer essentially daily liquidity to investors on the basis of underlying assets that do not in fact have that level of liquidity. For this reason, they are often considered parts of the so-called “shadow banking” system (or “non-bank financial intermediation” in an EU context). There is evidence that in the EU as a whole, the share of non-liquid assets in bond funds’ total assets has increased substantially over time (ESRB, 2019[21]). The potential threat of this type of activity to financial stability is typically not visible as long as there is not a significant demand from investors to liquidate their assets simultaneously. This is, however, precisely what happened during the COVID‑19 crisis. Panel A of Figure 1.22 shows how market turnover increased over historical averages, roughly doubling in February and March of 2020. Panel B shows the Riksbank’s share in total secondary market turnover, following its intervention in the markets.
Figure 1.22. Turnover in the Swedish corporate bond market during the COVID‑19 pandemic
![](/adobe/dynamicmedia/deliver/dm-aid--99627b4c-8cc3-4f2f-8be0-e19d5057d0d7/image24.png?quality=80&preferwebp=true)
Note: Refers to spot contracts. Monthly figures are sums of daily data. In Panel B, secondary market trading is calculated as total turnover subtracted by transactions identified as primary market trading.
Source: Sveriges Riksbank (SELMA Statistics).
The FSA’s (Finansinspektionen) analysis concludes that corporate bond markets have not offered more flexible financing than banks in times of crises in Sweden, meaning they do not provide the same type of counter-cyclical financing as in the EU and US markets. However, issuers who could access bond markets in foreign currencies were able to benefit from greater flexibility. Over the past decade, the vast majority of SEK denominated bonds issued by Swedish companies (in the order of 90%) were not traded on any given day, compared to a figure of about 60% for foreign currency denominated bonds. The lack of resilience, in particular in the domestic currency market, can partly be explained by the underdeveloped market and in particular the lack of liquidity. However, it bears mentioning that Sweden has a stable and profitable banking system that exited the 2008 financial crisis (and the subsequent euro crisis) in a comparatively strong position. Notably, while bank lending was below trend in Sweden both following 2008 and the euro crisis, this was driven by a decrease in foreign currency lending, whereas domestic currency bank loans remained remarkably stable (Becker et al., 2020[10]).
Shortcomings in the Swedish bond market during the COVID‑19 crisis has prompted several regulators and public bodies to undertake initiatives to reform the market and make it more resilient, decreasing the degree of credit concentration in the banking sector. Notably, in 2021 the government tasked the FSA with investigating the need for additional liquidity management tools to deal with liquidity risk for investment funds. It reviewed three such tools: swing pricing (adjusting the fund unit value or sales and redemption price of the units up/down depending on the costs associated with the fund’s net flows); anti-dilution levies (a fee levied on investors when they sell/redeem fund units); and redemption gates (allowing funds to postpone redemptions above pre‑defined thresholds). Its assessment is that the application of anti‑dilutions levies is possible within the existing legal framework, along with a certain type of swing pricing.12 It also believes that the use of redemption gates should be allowed. The FSA’s position is that the conditions for applying these tools should be regulated “in legislation and related regulations”. Swing pricing has been identified as a priority area. It also recommends that a requirement that investment funds be open for redemption at least twice per month should be reflected in law, along with the longest allowed redemption time (Finansinspektionen, 2021[22]). In an article in the financial press, the Director General of the FSA has also urged market participants to improve their conduct, notably with respect to disclosure and transparency, and encourages the development of a self-regulatory regime by the SSMA. The article also stresses the need for investment funds to improve their liquidity management by designing their portfolios with scenarios of market pressure in mind, for example increasing the share of liquid assets or making clear to investors that they do not offer daily redemptions (Finansinspektionen, 2021[23]). These discussions are reminiscent of those that took place in the United States during the so-called “taper tantrum” in mid‑2013 when the substantial selling pressure had large effects on corporate bond prices and officials considered imposing exit fees on bond funds (Çelik, Demirtaş and Isaksson, 2015[14]).
Further, the Riksbank, the Swedish National Debt Office and the FSA have jointly called for the introduction of a Swedish standard for benchmark bonds. Their proposed standard for a benchmark bond includes: a minimum issuance of SEK 1 billion (roughly USD 100 million – as shown in Figure 1.10, in 2021 68% of the number of bonds issued were below this amount); a minimum of two bookrunners to broaden the investor base; and issued through syndicated public transactions in line with the Eurobond market standard. By increasing liquidity in the market, this would help diversify the investor base, notably to long‑term investors such as insurance companies and pension funds that contribute less to liquidity crunches in times of financial distress compared to the currently dominant investment funds (Finansinspektionen, 2022[24]). In addition to encouraging the development of such a standard, the central bank has outlined a number of recommendations of measures that can be taken by different market participants, which are summarised in Box 1.1.
Box 1.1. The Riksbank’s action proposals for a better functioning corporate bond market
In its 2021 report Towards a better functioning corporate bond market, the Swedish Riksbank outlines a number of possible action points for different market participants (issuers, investors and banks). The nine recommendations are briefly summarised below.
Issuers
Companies issuing bonds may contribute to market development and liquidity by: issuing fewer but larger bonds, thus establishing a credit curve; involving more banks (bookrunners/arrangers) in the issue process, thereby expanding the investor base; obtaining credit ratings for more issues.
Investors
Investors, most notably investment funds, may: communicate more clearly to retail investors the liquidity risk and limitations of investing in fixed income funds (notably by reporting spread exposure1 in fact sheets, on websites, etc.); limiting the offering of daily redemptions (exchanges/fund platforms can play an important role in this regard); and by ensuring they are adequately managing liquidity risk, e.g. by increasing the share of liquid assets and possibly applying liquidity management tools such as swing pricing (see above).
Banks
As advisors and dealers, banks can: encourage issuers to e.g. obtain a credit rating and discourage the issuance of a large number of smaller bonds; publish price data in a timely manner (including for private placements); and finally possibly enter into resale agreements whereby issuers pay them to be more active in the markets, although the report notes that this may also bring disadvantages such as increased borrowing costs, since the pricing of such services may be difficult.
1. The value reduction of the fund as a share of its current value in a scenario where the interest rate spread between its holdings and government bonds doubles.
Source: Riksbanken (2021[5]), Towards a better functioning corporate bond market, https://www.riksbank.se/globalassets/media/rapporter/riksbanksstudie/engelska/2021/towards-a-better-functioning-corporate-bond-market.pdf.
1.9. The role of real estate companies in the Swedish corporate bond markets
The real estate sector represents an important part of the Swedish corporate bond market, accounting for almost half of total outstanding bonds issued in domestic currency by amount (Wollert, 2020[17]). This has raised concerns that a fall in real estate prices could reverberate through other sectors of the corporate bond market, threatening financial stability more generally (Riksbanken, 2021[5]). This is particularly relevant given the very steep growth in property valuations in Sweden since the mid‑1990s. Work within the OECD Capital Market Series, which is focused on non-financial companies, typically does not include the real estate sector. Real estate companies are instead classified as financials, given that they are effectively investment firms whose portfolios consist primarily of tangible assets. However, given the size, concentration and importance of real estate companies in Swedish bond markets, this subsection provides an overview of the sector’s use of corporate bond financing and how it compares to that of non-financial companies. The Annex provides details on the types of companies that are included in this analysis.
Real estate companies have gone from representing a negligible share of the Swedish bond market to becoming very substantial, accounting for 48% of total issuance and 39%13 of outstanding amounts in 2021 (Figure 1.23, Panels A and B).14 Issuance has been particularly high since 2017, averaging USD 11.5 billion annually between 2017 and 2021, up from an average of USD 2.9 billion from 2010 to 2016.
Figure 1.23. The real estate sector’s role in the Swedish corporate bond market
![](/adobe/dynamicmedia/deliver/dm-aid--90871983-8fc0-437e-bd71-a2f80134aa18/image25.png?quality=80&preferwebp=true)
Source: OECD Capital Market Series dataset, Thomson Reuters Datastream, see Annex for details.
Including real estate companies in the analysis significantly changes the industry composition showed in Figure 1.12. As shown in Figure 1.24, since 2019 real estate companies have made up roughly half of annual issuance, up from 20% on average between 2010 and 2015. This degree of concentration exposes the bond market as a whole to fluctuations within the dominant industry. Given the real estate sector’s particularly strong link to the financial sector (e.g. because of the amount of real estate collateral held by banks), this lack of diversification is a financial stability concern, especially when considering the prospect of significantly higher interest rates going forward and its likely impact on real estate prices.
It bears noting that real estate companies do not seem to issue foreign currency denominated bonds to a greater extent than non-financial companies. In fact, they have issued a larger share of SEK-denominated bonds in the last decade.
Figure 1.24. Industry composition of the Swedish corporate bond market
![](/adobe/dynamicmedia/deliver/dm-aid--8b7dba7e-e4e1-4ebf-9748-acb0c696f26c/image26.png?quality=80&preferwebp=true)
Note: The figure includes real estate companies.
Source: OECD Capital Market Series dataset, Thomson Reuters Datastream, see Annex for details.
The median issue size by real estate companies has remained relatively constant since 2012, averaging USD 46 million between 2012 and 2020, and reaching USD 54 million in 2021 (Figure 1.25, Panel A). This is similar to the figure for non-financial companies at USD 60 million (Figure 1.10, Panel B). However, the distribution of issue sizes is weighted more towards smaller issues among real estate companies than is the case for non-financial firms. While the two smallest size categories (below USD 100 million) made up 68% of the total number of issues by non-financial companies in 2021, for real estate companies it was as high as 82%. In spite of this, the concentration of total issue amounts by the top ten real estate issuers (71% in 2021) is higher than for non-financial companies (51%) (Figure 1.25, Panel B; Figure 1.11). The difference is even starker when looking at the top three companies, which represented 47% of total issuance by the real estate sector in 2021, compared to 25% for the top three non-financial companies. This is partly a natural effect of a smaller sample of real estate companies, but it is also because many of the largest real estate companies issue several smaller bonds in any given year rather than issuing larger bonds. This can be seen when comparing the share of the three largest bonds in total issuance in a year with the share of the total amount issued by the three largest issuers. For example, between 2015 and 2021, the three largest real estate bonds made up an average of 19% of total issuance annually. However, the three largest issuers made up an average of 45% of annual issuance, meaning their total issuance is made up by a set of smaller bonds. This happens because companies want to issue bonds with different profiles (e.g. with respect to maturities to smooth their repayment schedule or in different currencies to attract different investors), and is not limited to the real estate sector. However, Swedish real estate companies issue a larger number of bonds per issuer than non-financial companies do. Between 2015 and 2021, non‑financial companies (those issuing bonds) issued an average of 2.2 bonds per year. The corresponding figure for real estate companies was more than twice as high at 4.5 bonds per issuer and year, indicating it is composed of active issuers to a greater extent than the broader non-financial group. Similarly, when looking at the median issue size (the bond) as a percentage of the median issuer size (the company) between 2015 and 2021, the average figure for real estate companies is lower (1.4%) than for non-financial companies (1.8%), meaning the former group of companies tends to issue smaller individual bonds relative to their size.
Figure 1.25. Size distribution and concentration of real estate issuers
![](/adobe/dynamicmedia/deliver/dm-aid--a60c0141-1488-4829-b09b-1a5ffc7829c6/image27.png?quality=80&preferwebp=true)
Source: OECD Capital Market Series dataset, Thomson Reuters Datastream, see Annex for details.
The real estate sector has converged with the non-financial sector over time both in terms of credit ratings and maturity profile. In 2021, the value‑weighted average maturity of real estate bonds was 7.8 years, compared to 8.4 years for the non-financial sector (Figure 1.26, Panel A). The groups are even closer in terms of credit ratings, with the average value‑weighted rating for real estate company bonds at 12.4 (about half a notch above BBB-) in 2021 compared to 12.3 for non-financial companies. Historically, the real estate sector had substantially higher ratings than the non-financial sector, in the order of three full notches on average between 2013 to 2016 (real estate company bonds had an average rating corresponding to A, compared to BBB for non-financial company bonds) (Figure 1.26, Panel B).
Figure 1.26. Maturity and credit rating profile of the real estate sector
![](/adobe/dynamicmedia/deliver/dm-aid--d8fa4e3f-8ac8-4cc7-b599-8a16f56b3dfd/image28.png?quality=80&preferwebp=true)
Note: Real estate time series starts in 2010 in Panel A because issuance before that year was very small. The same applies for rated bonds in Panel B (from 2013). The peak for non-financial companies in 2015 in Panel A is driven by a set of four large bonds (between USD 344 million to USD 1.05 billion) issued by state‑owned power company Vattenfall with maturities between 62 and 63 years.
Source: OECD Capital Market Series dataset, Thomson Reuters Datastream, see Annex for details.
Notes
← 1. Note that in the corresponding Swedish financial accounts reported by the ECB, intercompany loans are not reflected separately and are fully considered within the long-term loans category.
← 2. These figures are based on transactions involving US companies.
← 3. The spike in 2015 is driven by a series of large issuances maturing in 2077 and 2078 by state‑owned power company Vattenfall.
← 4. There were a number of sizeable domestic currency denominated bonds issued in 2007 and 2008, notably by manufacturing company Scania and state‑owned power company Vattenfall.
← 5. Notable USD denominated issues (>USD 500 million) without listing include Ellevio (2016), Atlas Copco (2007) and Stena (2014).
← 6. In practice, the incentives of the trustee to conduct any significant covenant compliance due diligence is rather weak, owing to a fixed fee structure and potential professional liability concerns. Their tasks are therefore often limited to administrative procedures. See e.g. (Çelik, Demirtaş and Isaksson, 2015[14]) for a more thorough review.
← 7. See Finish law 25.8.2017/574, and the regulation pertaining to the “repræsentant” in Danish law LBK nr 1 229 af 07 September 2016.
← 8. These requirements normally do not apply to investment firms trading only bilaterally. However, for so-called systematic internalisers (executing orders against their own books), it is mandatory to quote bids and offers. Such trading has increased since 2018, with corresponding reductions in OTC trading.
← 9. The survey includes market participants dealing in government bonds and covered bonds, in addition to corporate bonds.
← 10. An OTF is an Organised Trading Facility. MiFID II introduced OTFs as a new trading venue category, a multilateral system in which multiple third-party buying and selling interests in bonds, structured finance product, emissions allowances or derivatives are able to interact. An OTF is neither a regulated market (RM) nor a multilateral trading facility (MTF).
← 11. April and September: Verisure; June: Ellevio (two bonds); September: Volvo Cars.
← 12. The FSA calls this “adjusted sale and redemption price”, a method whereby the price of the fund units is adjusted up or down depending on net flows. Another type of swing pricing, which it calls “adjusted net asset value”, involves adjusting the value of the fund, which is not allowed in current fund-related legislation.
← 13. This share differs from that shown in the central bank’s staff memo (Wollert, 2020[17]). This is this is due to the following differences in methodology: 1) the central bank’s figures refer only to outstanding bonds issued in the domestic currency, while the present report also includes bond issued in foreign currency; and 2) possible differences in the industry classification used.
← 14. Throughout this subsection, “total issuance/outstanding amounts” refer to the sum of issuance/outstanding bonds by non-financial companies and real estate companies.