Machine learning for financial transaction classification across companies using character-level word embeddings of text fields

An important initial step in accounting is mapping financial transfers to the corresponding accounts. We devised machine-learning-based systems that automate this process. They use word embeddings with character-level features to process transaction texts. When considering 473 companies independently, our approach achieved an average top-1 accuracy of 80.50%, outperforming baselines that exclude the transaction texts or rely on a lexical bag-of-words text representation. We extended the approach to generalizes across companies and even across different corporate sectors. After standardization of the account structures and careful feature engineering, a single classifier trained on 44 companies from 28 sectors achieved a test accuracy of more than 80%. When trained on 43 companies and tested on the remaining one, the system achieved an average performance of 64.62%. This rate increased to nearly 70% when considering only the largest sector.     

Taxing Large Businesses: Cooperative Compliance in Action

Intereconomics - Cooperative compliance represents a shift in thinking for tax administrations, away from a deterrence approach where taxpayers are coerced to comply with tax rules and threatened...     

Reducing Demand for Litigation in Consumer Disputes—a Randomized Field Experiment with Social Information

It has long been argued that alternative dispute resolution is superior to traditional court litigation. The paper reviews traditional and behavioural arguments and findings for why litigation rates remain high. The paper then reports on a natural field experiment designed to test how to improve disputes between consumers and businesses in Denmark resolution by including social information into the dispute process. The experiment demonstrates that social information significantly affects litigants’ need for litigation, but that the effect on settlement rates is non-significant. The results are discussed along with suggestions for possible future research into improving the willingness to cooperate in consumer disputes.     

The cointegrated vector autoregressive model with general deterministic terms

In the cointegrated vector autoregression (CVAR) literature, deterministic terms have until now been analyzed on a case-by-case, or as-needed basis. We give a comprehensive unified treatment of deterministic terms in the additive model $X_t=\gamma Z_t+Y_t$, where $Z_t$ belongs to a large class of deterministic regressors and $Y_t$ is a zero-mean CVAR. We suggest an extended model that can be estimated by reduced rank regression, and give a condition for when the additive and extended models are asymptotically equivalent, as well as an algorithm for deriving the additive model parameters from the extended model parameters. We derive asymptotic properties of the maximum likelihood estimators and discuss tests for rank and tests on the deterministic terms. In particular, we give conditions under which the estimators are asymptotically (mixed) Gaussian, such that associated tests are ${\chi }^2$-distributed.     

Ambidexterity under digitalization: A tale of two decades of new media at a Swedish newspaper

The digitalization of the newspaper industry represents a significant challenge for incumbent companies to engage new technologies. Many companies in the industry have had to seek new markets through digital technologies to survive. This paper explores how one of the largest Swedish newspapers, Aftonbladet, has strategically embraced new media and new markets. We report a decade of engaged scholarship based on interviews and archival analysis that covers 20 years of strategic acts at the company. We consider this effort as a case of organizational ambidexterity under digitalization. The analysis seeks to extend theoretical understanding of the interrelationships between strategic intent and technological choice. The paper contributes to the understanding of ambidexterity under digitalization by theoretically framing it in terms of strategic acts. The research suggests that digitalization implies a more complex ambidexterity interrelationship between old and new markets and technologies. As digitalization enables the loosening of previously tight couplings, the clear theoretical distinction between old and new, and critically, the unproblematic transition, is brought into question. The paper suggests replacing the notion of an orderly shift from the old to the new with ambidexterity under digitalization as a duality of both old and new undergoing continual reconfiguration.     

Technologies in caregiving: professionals’ strategies for engaging with new technology

This article explores the adoption of new technology in organisations that provide senior citizen care. Inspired by Niklas Luhmann’s systems theory, we study how technology reduces complexity by identifying client needs and ensuring predictability in service delivery. However, how technologies are adopted in practice is not determined by technology since it is also structured by care-workers' continuous decision-making. Against this backdrop, we explore how technologies alter the conditions for decision-making in two settings of elderly care, and we describe how care workers seek to adapt technologies to their practical needs as well as conception of care ethics. Developing a systems theory approach, the article eschews a priori assumptions of technological constraint on care-workers’ professional autonomy, offering a more open-ended exploration of diversified strategies for coping with new technology. Our case studies show that employees develop diversified strategies for technology adoption, including both non-usage, heated resistance, excessive embrace, and creative adaption.     

The cost of diabetes and obesity in Australia

Aims: To assess and compare the direct healthcare and non-healthcare costs and government subsidies by body weight and diabetes status.

Methods: The Australian Diabetes, Obesity and Lifestyle study collected health service utilization and health-related expenditure data at the 2011–2012 follow-up surveys. Costing data were available for 4,409 participants. Unit costs for 2016–2017 were used where available or were otherwise inflated to 2016–2017 dollars. Age- and sex-adjusted costs per person were estimated using generalized linear models.

Results: The annual total direct cost ranged from $1,998 per person with normal weight to $2,501 per person with obesity in participants without diabetes. For those with diabetes, total direct costs were $2,353 per person with normal weight, $3,263 per person with overweight, and $3,131 per person with obesity. Additional expenditure as government subsidies ranged from $5,649 per person with normal weight and no diabetes to $8,085 per person with overweight and diabetes. In general, direct costs and government subsidies were higher for overweight and obesity compared to normal weight, regardless of diabetes status, but were more noticeable in the diabetes sub-group. The annual total excess cost compared with normal weight people without diabetes was 26% for obesity alone and 46% for those with obesity and diabetes.

Limitations: Participants included in this study represented a healthier cohort than the Australian population. The relatively small sample of people with both obesity and diabetes prevented a more detailed analysis by obesity class.

Conclusion: Overweight and obesity are associated with increased costs, which are further increased in individuals who also have diabetes. Interventions to prevent overweight and obesity or reduce weight in people who are overweight or obese, and prevent diabetes, should reduce the financial burden.     

UNIVERSAL PRESCHOOL PROGRAMS AND LONG-TERM CHILD OUTCOMES: A SYSTEMATIC REVIEW

What are the long-term effects of universal preschool programs on child outcomes? We review 26 studies using natural experiments to estimate the effects of universal preschool programs for children aged 0–6 years on child outcomes measured from third grade to adulthood. Studies comparing universal preschool with a mix of parental, family, and private modes of care show mixed effects on test scores and on measures related to health, well-being, and behavior. All estimates for outcomes related to adequate primary and secondary school progression, years of schooling, highest degree completed, employment, and earnings indicate beneficial average effects of universal preschool programs. Three of the included studies calculate benefits-to-costs ratios and find ratios clearly above one. Universal preschool tends to be more beneficial for children with low socioeconomic status and there are not consistently different effects for boys or girls. Only three studies compare two alternative types of universal preschool programs in terms of long-term outcomes.     

Climate for conflict management,exposure to workplace bullying and work engagement: a moderated mediation analysis

The present study investigates a potential preventive factor in relation to workplace bullying. Specifically, we examine how climate for conflict management (CCM) may be related to less bullying, increased work engagement, as well as whether CCM is a moderator in the bullying engagement relationship. The study was based on a cross-sectional survey among employees in a transport company (N = 312). Hypotheses were tested simultaneously in a moderated mediation analysis which showed that bullying and job engagement were related (H1), CCM was related to less reports of bullying (H2), CCM was related to work engagement (H3) and that CCM was indirectly related to job engagement through bullying (H4), but only when CCM was weak (H5). That is, CCM moderated the relationship between bullying and work engagement in that this relationship only existed when CCM was low. The present study contributes to theory within this research field by showing that organizational measures may not only prevent bullying, but may also affect how employees react when subjected to bullying. Furthermore, the effect of climate in relation to bullying may be down to the narrow bandwidth facet of CCM. The study informs employers how they may act to prevent bullying while also reducing the potential negative outcomes of those cases of bullying that inevitably will show up from time to time.