Economic Outcomes

Inputs into the decisions of users of accounting information are not directly observable and are a notorious black box for researchers. The same holds for the complex process through which these decisions translate into market prices and economic outcomes. Therefore, a variety of proxies have been tested in the IFRS literature, such as (1) analysts’

information environment, (2) investors’ capital allocation decisions and portfolio choices, (3) equity markets, (4) debt markets, and (5) corporate investments and governance.

3.1.2.1 Analysts’ Information Environment

Information intermediaries, such as financial analysts, are considered to be prime users of accounting information and thus likely to adapt their behavior in response to IFRS adoption. For example, IFRS can impact an analyst’s decision to follow a firm and the composition of the portfolio of firms that person covers (i.e., variation at the analyst’s level), and therefore the number and mix of local versus foreign analysts following a firm (i.e., variation at the firm level). In turn, the number and type of analysts following a firm affect the production and dissemination of information, and coverage “shocks” have been documented to have consequences for firms (e.g., Irani and Oesch,2013). Moreover, the IFRS literature uses the properties of analysts’ forecasts (i.e., their accuracy,

10See Neel (2017) and De Georgeet al.(2016) as examples.

bias, and the dispersion across analysts) as proxies for the quality of a firm’s information environment (e.g., Bae et al.,2008).

Both the number and type of analysts following a firm and the properties of analyst forecasts are outcome measures that can be easily obtained from IBES. However, given that data-collection at IBES is opaque (e.g., Ljungqvist et al., 2009), it is unclear how the IFRS transition affected this process, which makes interpretations challenging.

For example, not all analysts report to IBES and so if IBES extended its scope of coverage to a region right when this region adopts IFRS, an increasing number of analysts following a firm displayed in IBES would not necessarily reflect a real increase in coverage. Moreover,

“actuals” (reported earnings by the firm) in IBES are “adjusted” to better match to what analysts forecasted (labelled “core earnings”), and it is unclear how IFRS impacted these adjustments. Similarly,

“consensus forecasts” during the transition periods reflect a mixture of IFRS and non-IFRS forecasts that different analysts provided.¹¹ Both issues can bias measures of analysts’ forecasting accuracy around IFRS adoption. Since IBES is the only broad-scale provider for international data on analysts, there is no possibility to benchmark the data quality.

3.1.2.2 Investors’ Capital Allocation and Portfolio Choices

Cross-country differences in accounting create information barriers for foreign investors and likely contribute to investors’ home bias (Karolyi and Stulz,2003). To test the effect of IFRS on investment decisions, the literature has examined (1) worldwide portfolio holdings of mutual funds,¹²(2) worldwide portfolio holdings of a wider range of institutional investors,¹³ (3) aggregated, country-level long-term equity investments

11See the IBES Summary History User Guide (Thomson Reuters,2013) for details.

No study has used information on the “accounting-bases” of IBES forecasts and actuals that became available in the “Company Level Footnote File” from early 2005 onwards.

12The Thomson Reuters International Mutual Funds database covers firm-level holdings of over 25,000 mutual funds from around the world. See Covriget al.(2007), DeFondet al.(2011), and Fanget al.(2015).

13The Thomson Financial Ownership database captures a wider set of investor types beyond mutual funds, such as pension funds, insurance companies, hedge funds, or private equity. See Florou and Pope (2012).

by U.S. investors,¹⁴and (4) data from the German Open Market serving as a proxy for retail investors’ trading.¹⁵Collectively, these sources cover holdings and trades for a wide range of investor types.

3.1.2.3 Equity Markets

The literature has developed a rich set of measures that capture the properties of stock returns, liquidity, and trading activities as well as the conditions for firms when raising equity in these markets (such as the cost of capital). Most measures share the limitation that they build on the assumptions that different markets have similar levels of efficiency in information processing and similar levels of private (insider) versus public information flows, both of which are unlikely to hold across countries (e.g., Frost et al.,2006; Morcket al.,2000).

Based on stock returns, research has tested for IFRS-related changes in how stock markets react to disclosure events, such as earnings an-nouncements of the firm (Landsman et al., 2012) or its competitors (Wang, 2014). The literature looked not only at first moments, that is, changes in the price level, but also into second, that is, variance and disagreement, and third moments of the return distribution, that is, skewness, or the frequency of extreme negative stock returns in the left-tail, called crash risk (DeFond et al., 2015). Other measures capture stock price synchronicity, that is, the extent to which IFRS causes stock prices to co-move more (less) closely with firm-specific (common) information (Kim and Shi,2012). In general, the wider the window over which returns are measured, the greater the risk of cap-turing improvements in firm-specific information flows at the time of IFRS adoption that are due to other contemporaneous innovations or

14Retrieved from reports by the U.S. Treasury Department, which, in theory, capture all types of long-term foreign equity holdings. See Khurana and Michas (2011) and Shima and Gordon (2011).

15Data on holdings and trading of retail investors are particularly difficult to acquire, as there are no corresponding disclosure requirements for individuals. The literature therefore has either exploited individual researcher’s access to propri-etary data of (online) brokerage services or trading data of stock market segments customized toward individual investors. See Brüggemannet al.(2012).

emerging channels of communications (e.g., the use of investor relations or earnings guidance).

Many studies use measures of liquidity and trading in equity markets (e.g., Daske et al., 2008, 2013), such as quoted bid-ask spreads, trading (volume or the percentage of zero return or trading days), the Amihud (2002) price-impact measure (absolute stock return divided by trading volume), and estimates of the total round-trip costs of transactions (including trading costs other than spreads) based on Lesmond et al.

(1999). Such measures have been used on a standalone basis (note the correct sign as some measures rather reflect illiquidity) or combined in joint liquidity scores (e.g., based on factor analysis; Christensenet al., 2013; Lang et al., 2012). Internationally, intra-day trading data was not available for many markets,¹⁶ and there is additional noise when using data for less developed markets.¹⁷ Still, liquidity measures in general, and bid-ask spreads in particular, have emerged over time as a frequently used outcome variable for testing information asymmetry, given the strong theoretical foundation, sensitivity to news, and large-scale availability at high frequencies. Other studies focus on equity market characteristics that are easily observable, such as firms’ cross-listed shares on a foreign exchange (e.g., Chen et al., 2015a) or the conditions of equity issuances (e.g., Hong et al.,2014).

Investors’ required returns or the firm’s cost of capital are not directly observable. Estimates of expected returns are notoriously imprecise and can be based on time-series of realized returns or ex-ante cost of capital using analyst forecasts. Many studies on disclosure effects rely on the latter estimates, given their conceptual appeal, forward-looking nature, and the fact that changes in cost of capital can be estimated in basis points, which allows for judging economic magnitudes (e.g., Daske, 2006; Hail and Leuz, 2006, 2009; Li, 2010). However, these measures produce substantial measurement errors (Easton, 2009), particularly in accounting regime change settings, such as IFRS (Easton, 2006),

16As a consequence, other established facets of liquidity from the market-microstructure literature could not be considered, such as effective bid-ask spreads (based on transactions) or measures for market depth.

17For example, for trading volume in Datastream there is an unsystematic treat-ment of missing data entries versus entries of the value “zero”.

because they necessarily rely on long-term growth assumptions and analyst forecasts as the main inputs into the models. There is also a debate to what extent the cost of capital reflectsexpected transparency changes.¹⁸

3.1.2.4 Debt Markets

Research has analyzed the impact of IFRS on (1) the pricing and terms of debt contracting, that is, volume of financing, yields or credit spreads, maturities, fees, and use and type debt covenants (e.g., Kim et al., 2011; Chen et al.,2015c); (2) the market structure for debt financing, that is, the mix of public versus private debt (Florou and Kosi,2015), the structure of loan syndicates (Brown, 2016), and the relationship between the pricing of debt instruments across market segments, for example, CDS and underlying financial instruments (Bhat et al.,2014, 2016); and (3) the use of accounting information in facilitating lending, in terms of credit relevance, that is, the ability of accounting numbers to explain firms’ default probabilities, reflected in ratings or CDS spreads (e.g., Florou et al.,2017; Kraft et al.,2020; Wu and Zhang, 2014), and contractibility, that is, the usage of accounting-based covenants in debt contracts (e.g., Ball et al.,2015; Chenet al.,2015c).

A key challenge when using debt-market outcomes is that firms have multiple options for debt financing, and all relevant terms¹⁹ of a contract need to be negotiated (contrary to equity contracts, which are more standardized) and may substitute or complement each other.

Thus, the designs of debt contracts are endogenous, and, in theory, all terms should be modeled together and estimated simultaneously, which poses econometric challenges.

18Christensenet al.(2013, p. 152) argue that market participants “likely adjust market valuations or cost of capital estimates as soon as their expectations about future corporate transparency change, liquidity is less anticipatory because investors primarily worry about adverse selection and, hence, the level of transparency at the time they trade”. In contrast, De George et al. (2016, p. 935) suggest that

“from a theoretical perspective, it is unclear why investors decrease the premium for information risk even before the risk is attenuated and despite the significant uncertainty around IFRS implementation and its effect on reporting quality”.

19Such as volume, pricing, maturity, control rights, or collateral.

Many debt contracts are also private (such as loans or trade credit) and important contractual terms (such as the use of collateral, control rights, or the frequency of information exchange) are not observable for international lending relationships. Researchers mostly use data on debt issues as collected by DealScan. This data tends to be biased towards specific types of loans (especially syndicated loans), loans issued in countries where there are public sources for loan originations (especially the United States), loans with facility volumes that meet some disclosure threshold, or loans where arrangers and lenders voluntarily disclosed the contractual terms (e.g., to be included in league tables).²⁰One example for imperfections is that only around 10% of international debt issues have at least one recorded covenant, which probably represents the failure or inability of vendors to collect covenant information, rather than the debt being covenant free.²¹ Bank loan contracts, which are the key source of debt financing for many international firms, are rarely observable.

More information on the pricing of debt and credit risk under a rather fixed set of terms (bond contracts are rarely renegotiated) is available from public bond markets covered by Securities Data Company’s (SDC) Platinum. However, secondary corporate bond markets are often not very liquid internationally, because of the low number of bonds issued (many firms do not have a credit rating) and the low trading frequency after issuance (many bonds are privately placed with buy-and-hold investors). Researchers struggle with the small number of bonds, the stickiness of bond returns, and often missing controls for default risk (such as CDS contracts).

3.1.2.5 Effects Inside the Firm: Corporate Investments and Governance The last category of outcomes relates to effects inside the IFRS-adopting firm and can be broadly grouped into (1) firms’ internal investment

20Data is “primarily sourced from direct bank submissions from lenders, journalist news stories and relevant press releases. U.S. data also benefits from these sources in addition to regulatory filings with the SEC” (DealScan customer support).

21See Ballet al.(2015, p. 929). Furthermore, in theory, IFRS 7.18 mandates firms to report covenant violations during the past reporting period. However, in practice, this disclosure requirement is often ignored when covenants have been renegotiated.

decisions, such as corporates’ or subsidiaries’ investment cash flow sen-sitivities (Chenet al., 2013; Shroff et al., 2014), and (2) their corporate governance, such as incentive schemes in executive compensation (e.g., Ozkan et al., 2012) or managerial dismissals (Wu and Zhang, 2009, 2019).

Finally, in line with cost-benefit analyses, research has investigated IFRS reporting costs, often proxied by audit fees (De George et al., 2013; Kim et al., 2012). We discuss potential issues with these outcome variables along with the evidence in Subsection 3.4.