This paper examines a mixed data sampling (MIDAS) approach to accounting research. MIDAS regression models parsimoniously incorporate variation embedded in existing economic data that are observed at much higher frequencies than accounting data. The additional source of data variation creates an opportunity to address new and important questions in accounting research. We develop and outline four new MIDAS models within the general framework that are simple to estimate and capture economic properties that are relevant to accounting research. We demonstrate the efficacy of our models with empirical applications to the January effect and to earnings response coefficients, which together illustrate the potential to expand the boundaries of accounting research by getting more out of high-frequency data.

We consider a single-period, pure-exchange setting where investors have mean-variance preferences. Investors are uncertain about the risk attitudes of other capital market participants and, thus, about the market's risk premium. Assuming that there are no strategic interactions among investors, we identify conditions under which investors choose to abstain from trading in the capital market due to the risk associated with an uncertain risk premium. In addition, we find that investors may prefer the disclosure of public information prior to trading, as this information reduces the risk premium and thus relaxes the condition for market participation. Therefore, the presence of investors with unknown risk attitudes suggests a beneficial role for public (accounting) information.

The paper uses structured machine learning regressions for nowcasting with panel data consisting of series sampled at different frequencies. Motivated by the problem of predicting corporate earnings for a large cross-section of firms with macroeconomic, financial, and news time series sampled at different frequencies, we focus on the sparse-group LASSO regularization which can take advantage of the mixed frequency time series panel data structures. Our empirical results show the superior performance of our machine learning panel data regression models over analysts' predictions, forecast combinations, firm-specific time series regression models, and standard machine learning methods.

The paper introduces structured machine learning regressions for heavy-tailed dependent panel data potentially sampled at different frequencies. We focus on the sparse-group LASSO regularization. This type of regularization can take advantage of the mixed frequency time series panel data structures and improve the quality of the estimates. We obtain oracle inequalities for the pooled and fixed effects sparse-group LASSO panel data estimators recognizing that financial and economic data can have fat tails. To that end, we leverage on a new Fuk-Nagaev concentration inequality for panel data consisting of heavy-tailed $\tau$-mixing processes.

Boards that compensate executives with cash face a costly ex post settling up problem: boards may compensate executives for gains which later fail to be realized. Using a novel linear mixed data sampling (MIDAS) method to exploit the variation in daily stock returns, we provide evidence of techniques that boards use to mitigate the ex post settling up problem. Using stock returns as a proxy for performance, we first show that cash compensation is more sensitive to returns earlier in the year than returns later in the year. Second, we show that while cash compensation is very sensitive to returns around the release of accounting earnings, this sensitivity dissipates over time. Finally, we reexamine the asymmetric sensitivity of compensation to unrealized gains and losses and show that compensation is more sensitive to early period returns for both positive and negative returns. Our findings suggest that boards, in an attempt to mitigate the ex post settling up problem, rely on the passage of time and more heavily weight performance earlier in the year when exercising discretion over cash compensation. In addition to these theoretical contributions, we provide an important methodological improvement for the study of executive compensation: the linear MIDAS model allows us to reject the implicit assumption in the literature that the sensitivity of compensation to returns is time invariant over the course of the compensation period.

This paper uses structured machine learning regressions for nowcasting with panel data consisting of series sampled at different frequencies. Motivated by the problem of predicting corporate earnings for a large cross-section of firms with macroeconomic, financial, and news time series sampled at different frequencies, we focus on the sparse-group LASSO regularization which can take advantage of the mixed frequency time series panel data structures. Our empirical results show the superior performance of our machine learning panel data regression models over analysts’ predictions, forecast combinations, firm-specific time series regression models, and standard machine learning methods.

This study develops and applies a model-implied measure of information imprecision. We define information imprecision as the degree of noise in investors' prior beliefs about the firm's asset value based on the information set that is currently available. We present a model of credit default swap (CDS) spreads in which the term structure is a function of information imprecision. We exploit observable CDS spreads with short and long maturities to extract an empirical measure of information imprecision. We then examine the moderating role of our measure in two settings. First, we show that the equity market response to credit rating changes increases in the level of information imprecision before the announcement. Second, we show that bond-market professionals' ability to charge a premium to smaller investors, relative to larger investors, increases in the issuing firm's information imprecision. This evidence illustrates the broad applicability of our model-implied measure of information imprecision.

We examine the manner and extent to which firms evaluate performance relative to aspirational peer firms. Guided by the predictions of an agency model, we find that CEO compensation increases in the correlation between own and aspirational peer firm performances. In addition, we define and test conditions where aggregate peer performance, which has been the primary focus of prior relative performance evaluation studies of competitive peers, is expected to have an association with CEO compensation. These conditions are supported by our empirical results. Finally, we document that our results are more pronounced when the firm-peer relationship is one-way and the peer firm is in a different industry and therefore is more aspirational.

This paper introduces structured machine learning regressions for prediction and nowcasting with panel data consisting of series sampled at different frequencies. Motivated by the empirical problem of predicting corporate earnings for a large cross-section of firms with macroeconomic, financial, and news time series sampled at different frequencies, we focus on the sparse-group LASSO regularization. This type of regularization can take advantage of the mixed frequency time series panel data structures and we find that it empirically outperforms the unstructured machine learning methods. We obtain oracle inequalities for the pooled and fixed effects sparse-group LASSO panel data estimators recognizing that financial and economic data exhibit heavier than Gaussian tails. To that end, we leverage on a novel Fuk-Nagaev concentration inequality for panel data consisting of heavy-tailed $\tau$-mixing processes which may be of independent interest in other high-dimensional panel data settings.

This study develops and applies a model-implied measure of information imprecision. We define information imprecision as the degree of noise in investors' prior beliefs about the firm's asset value based on the information set that is currently available. We present a model of credit default swap (CDS) spreads in which the term structure is a function of information imprecision. We exploit observable CDS spreads with short and long maturities to extract an empirical measure of information imprecision. We then examine the moderating role of our measure in two settings. First, we show that the equity market response to credit rating changes increases in the level of information imprecision before the announcement. Second, we show that bond-market professionals' ability to charge a premium to smaller investors, relative to larger investors, increases in the issuing firm's information imprecision. This evidence illustrates the broad applicability of our model-implied measure of information imprecision.

Using a large panel from 46 countries over 20 years, we find that non-U.S. firms issue corporate bonds more frequently and at lower offering yields following an equity cross-listing on a U.S. exchange. Firms issue more bonds through public offerings instead of private placements and in foreign markets rather than at home, in both cases at significantly lower yields. Moreover, the debt-related benefits are concentrated among firms domiciled in countries with less private benefits of control, efficient debt enforcement, and developed bond markets, suggesting that equity cross-listings cannot completely offset the impact of weak home country institutions. The results support the notion that the monitoring, transparency, and visibility benefits brought about by equity cross-listings on U.S. exchanges are valuable to bond investors.

Prior studies attribute analysts’ forecast superiority over time-series forecasting models to their access to a large set of firm, industry, and macroeconomic information (an information advantage), which they use to update their forecasts on a daily, weekly or monthly basis (a timing advantage). This study leverages recently developed mixed data sampling (MIDAS) regression methods to synthesize a broad spectrum of high frequency data to construct forecasts of firm-level earnings. We compare the accuracy of these forecasts to those of analysts at short horizons of one quarter or less. We find that our MIDAS forecasts are more accurate and have forecast errors that are smaller than analysts’ when forecast dispersion is high and when the firm size is smaller. In addition, we find that combining our MIDAS forecasts with analysts’ forecasts systematically outperforms analysts alone, which indicates that our MIDAS models provide information orthogonal to analysts. Our results provide preliminary support for the potential to automate the process of forecasting firm-level earnings, or other accounting performance measures, on a high-frequency basis. The online appendix is available at https://doi.org/10.1287/mnsc.2017.2864 . This paper was accepted by Mary Barth, accounting.

We examine whether the contribution of firm-level accounting earnings to the informativeness of the aggregate is tilted towards earnings with specific financial reporting characteristics. Specifically, we investigate whether considering the smoothness of firm-level earnings increases the informativeness of aggregate earnings for future real GDP, and if so, whether macroeconomic forecasters use this information efficiently. Using recently-developed mixed data sampling methods, we find that the aggregate is tilted towards firms with smoother earnings and that this composition of aggregate earnings outperforms traditional weighting schemes. Further, this tilted aggregate has a stronger positive association with forecast revisions; in fact, analysts who utilize earnings the most in their forecasts appear to fully impound the informativeness of earnings smoothness. Our results synthesize and span parallel yet distinct streams of research on the role of accounting earnings in firm-level and macroeconomic outcomes and suggest an important role for financial reporting characteristics in the aggregate.

We examine the manner and extent to which firms evaluate performance relative to aspirational peer firms. Guided by the predictions of an agency model, we find that CEO compensation increases in the correlation between own and aspirational peer firm performances. In addition, we define and test conditions where aggregate peer performance, which has been the primary focus of prior relative performance evaluation studies of competitive peers, is expected to have an association with CEO compensation. These conditions are supported by our empirical results. Finally, we document that our results are more pronounced when the firm-peer relationship is one-way and the peer firm is in a different industry and therefore is more aspirational.

Investor uncertainty about firm value drives investors' information collection and trading activities, as well as managers' disclosure choices. This study examines an important source of uncertainty that likely cannot be influenced by most managers and investors: uncertainty about government economic policy. We find that this uncertainty is associated with increased bid-ask spreads and decreased stock price reactions to earnings surprises. Managers respond to this uncertainty by increasing their voluntary disclosures, but these disclosures only partly mitigate the bid-ask spread increase. We conclude that government economic policy uncertainty is an important component of firms' information environments and managers' voluntary disclosure decisions.

Can we design statistical models to predict corporate earnings which either perform as well as, or even better than analysts? If we can, then we might consider automating the process, and notably apply it to small and international firms which typically have either sparse or no analyst coverage. There are at least two challenges: (1) analysts use real-time data whereas statistical models often rely on stale data and (2) analysts use potentially large set of observations whereas models often are frugal with data series. In this paper we introduce newly-developed mixed frequency regression methods that are able to synthesize rich real-time data and predict earnings out-of-sample. Our forecasts are shown to be systematically more accurate than analysts' consensus forecasts, reducing their forecast errors by 15% to 30% on average, depending on forecast horizon.

We develop and test measures of the horizon of firm uncertainty and of the horizon of managers’ corporate disclosures. The measures exploit information in the term structure of implied equity volatilities to gauge the relative extent to which the information underlying securities prices reflects long-term versus short-term uncertainty. We find that the horizon of firm uncertainty measure is associated with variables that are likely to capture the extent to which firms’ business models result in differing degrees of uncertainty about the long-term versus the short-term. The horizon of managers’ corporate disclosures measure allows us to characterize managers’ disclosures in terms of whether they provide information about long-term business strategies or are more oriented towards short-term operating results. We find that earnings announcements containing management forecasts have shorter disclosure horizons than earnings announcements not containing management forecasts.

We dissect the portion of stock price change of the fiscal year that is recognized in reported accounting earnings of the year. We call this portion earnings recognition timeliness (ERT). The emphasis in our dissection is on empirical identification of two fundamental precepts of financial accounting: (1) the matching principle, which is manifested in the recognition of expenses in the same period as the related benefits (i.e., sales revenue) accrue; and (2) recognition of expenses in the current period due to changes in expectations regarding earnings of future periods (we refer to these expenses as the expectations element of expenses). Although the expectations element has implicitly been at the core of much of the recent empirical literature on asymmetry in the earnings/return relation, it has not been explicitly identified. This recent literature is based on the premise that bad news about the future leads to more recognition of expenses in the current period (such as write-downs) whereas good news about the future tends to have a much lesser effect on expenses of the current period; asymmetry in the expenses/return relation is captured implicitly via the observation of asymmetry in the earnings/return relation (i.e., asymmetry in ERT). Since the ERT reflects the relation between sales revenue and returns, matched expenses and returns, as well as the relation between the expectations element of expenses and returns, a focus on the expectations element may lead to sharper inferences. Our straightforward empirical procedure permits a focus on this element.

This paper theoretically and empirically investigates how the risk of future adverse price changes created by the anticipated arrival of information influences risk-averse investors’ trading decisions in institutionally imperfect capital markets. Specifically, I examine how the selling activity of individual investors immediately following an earnings announcement is influenced by the trade-off between risk-sharing benefits of immediate trade and explicit transaction costs imposed on such trades. Consistent with my theoretically derived predictions, I find that investors’ current trading decisions are less sensitive to the incremental transaction costs created by short-term capital gains taxes on trading profits, as both the duration and intensity of the risk of future adverse price changes increase. This evidence is consistent with an incremental cost to investors that results from the revelation of precise information, which is commonly referred to as the Hirshleifer Effect (Hirshleifer, 1971; Verrecchia, 1982).

In 2004, a combined system test was performed in the H8 beam line at the CERN SPS with a setup reproducing the geometry of sectors of the ATLAS Muon Spectrometer, formed by three stations of Monitored Drift Tubes (MDT). The full ATLAS analysis chain was used to obtain the results presented in this paper. The basic design performances of the Muon Spectrometer were verified. The stability of MDT calibration constants, the alignment system using optical devices and high energy tracks, as well as the intrinsic sagitta resolution of the Muon Spectrometer were studied and found to agree with expectations. The reconstruction of muon tracks using the combined information from both the Inner Detector and the Muon Spectrometer are also presented.