Diciembre y julio son los meses de más jaleo en la dirección financiera, el primero por los cierres de año y realización de presupuestos del año siguiente, y el segundo por la presentación de cuentas anuales e impuesto de sociedades. Todo ello aderezado con las prisas de tener que dejar todo cerrado porque vienen vacaciones y en un mes se paraliza todo.

Obviamente esto depende de cada organización, a más tamaño y complejidad más tiempo en preparación (antes se comienza).

En el caso de las empresas en las que trabajo, los preliminares de los presupuestos los empiezo ahora, en noviembre y a mitad de diciembre tenemos el borrador de los presupuestos del año siguiente. Lo normal es que en enero introduzcamos algún matiz, pero con cambios nada significativos.

Los presupuestos son una herramienta imprescindible en la dirección financiera por varios motivos:

1/ Guía

Ha sido pensado y calculado en función de las expectativas del año y el histórico de funcionamiento. se ha chequeado con diferentes responsables y tiene coherencia con toda la información contable de la empresa.

La coherencia es fundamental, cuando algo es incoherente saltan alarmas por varios sitios. La ventaja de la contabilidad es que es horizontal, partida doble, y evita los excesos de la unilateralidad de lo vertical.

2/ Dashboard

Cualquier dashboard efectivo y bueno debe hacer referencia a lo que sucede en relación a objetivos marcados. Nos tiene que decir «esto es lo importante» y «por qué» y «cómo» ha sucedido.

3/ Disciplina

Al tener una guía nos obliga a no desviarnos o no buscar excusas externas. Esto sobre todo es importante en el cálculo de tesorería (cash), que sintetiza todo el resto de variables (ventas, gastos, inversiones…).

Puede ocurrir que surjan diferencias entre la realidad y el presupuesto, incluso a veces que exista algún error en nuestras hipótesis, pero en este último caso, si hacemos el presupuesto varios años, cada año consecutivo se integran dichos errores y se minimizan las posibles divergencias de planteamiento.

4/ Estrategia

El presupuesto nos permite plasmar la estrategia de la empresa en números, y de paso hacer consciente el proceso estratégico en un documento con números concretos, ya que partes de este mucha veces se queda en la cabeza del CEO. Fuerza a que cada mánager sea consciente de la estrategia que lleva dentro el líder del proyecto.

Noviembre y diciembre, semanas de planificación exhaustiva. Vamos a planificar con precisión 2024.

Para seguir nuevas entradas utiliza el feed (rss) del blog en un agregador como feedly. No utilizo cookies de rastreo, ni newsletter, ni redes sociales. Escríbeme. Jorge.

Los autores del paper «Extracting Financial Data from Unstructured Sources: Leveraging Large Language Models» publicado hace unos días, hacen un ejercicio muy interesante; utilizan una aplicación de AI (ChatPDF) para extraer datos financieros, desperdigados por documentos PDF de empresas que cotizan en bolsa. Un problema común en la investigación y análisis de estados financieros o de otro tipo de información relacionada con las finanzas.

Según sus autores «los resultados de las pruebas piloto indican que el marco alcanza una tasa de precisión del 100% en un periodo de tiempo notablemente corto a la hora de extraer indicadores financieros clave.·

Yo no iría tan lejos, la herramienta tiene muchas limitaciones, pero es cierto que es un primer paso para las mejoras y evolución que vendrán. Como sabemos esto va muy rápido y en un año veremos cosas que parecían impensables.

La aplicación que utilizan es ChatPDF.com , yo la he utilizado en algunas ocasiones, básicamente para papers larguísimos, muy técnicos, en los que le pregunto cosas concretas para no tragarme todo el tocho infumable. Los resultados son mas o menos decentes, aunque limitados. En mi caso si a partir de ahí veo que me dice algo interesante, entonces profundizo, si no descarto el paper.

En este sentido, y utilizando los comandos que propone el paper (da comandos específicos), he comprendido que necesito conocimientos mínimos de dicho lenguaje que no tengo. Me refiero a entender la forma de «dialogar» con una máquina de LLM.

Este es el esquema que han seguido:

Otro de los temas que comentan y que es relevante, es el potencial de los LLM como alternativa a la estandarización de datos, apartándose del enfoque predominante de de una taxonomía centralizada y detallada, como las iniciativas en curso para introducir XBRL (Business Reporting Standard) en el ámbito de la contabilidad de entes públicos. Este apunte no es menor, diría que tiene un alcance mucho mayor, dado que la adopción de la taxonomía XBRL es muy compleja y llevan muchísimos años intentando adoptarla. Aquí un post que escribí hace muchísimos años relacionado con el tema.

Volviendo a lo práctico inmediato, lo interesante del experimento es la posibilidad de descargar la información que queremos en formato CSV, porque desde ahí podemos tratar los datos, sin demasiado trabajo, en Excel o cualquier otro programa de tratamiento de datos, evitando meter datos a mano durante horas.

Estos son los resultados de precisión y rapidez:

El «prompt» es la orden que le das al chat «Dime esto, dame sta cifra…». Según esto tiene un 100% de precisión y da la información en tan solo 4 minutos, mientras que a mano un experto lo haríamos en 200 minutos (más de 3 horas para extraer la información). Eso si, habiendo hecho retoques («refined prompt»).

Aquí tengo que señalar, que los comandos (prompts) que dan en el estudio solo sirven para cuentas anuales que siguen las USGAAP de USA y no te sirven para, por ejemplo, la información de una cotizada española. Con lo cual ese tiempo de 4 minutos no es real, es mucho mayor,

Supongo que con el tiempo habrá prompts estandarizados para todo, incluso estudios que se dediquen a ello.

Lo cierto es que existen plugins sobre pdf que ya te hacen esta función, pero el recorrido del lenguaje LLM para mejorar y profundizar este tipo de solución es mucho más amplio y profundo.

Te animo a que entres y pruebes los diálogos o prompts que utilizan en el paper. La herramienta (chatpdf.com) es sencillísima y no requiere darse de alta. Puedes subir el pdf arrastrándolo o poniendo directamente la url donde esta subido dicho pdf.

Para seguir nuevas entradas utiliza el feed (rss) del blog en un agregador como feedly. No utilizo cookies de rastreo, ni newsletter, ni redes sociales. Escríbeme. Jorge.

# Everything comes from accounting

All economic information is based on accounting. It starts from an invoice, or a transaction, which is accounted for in a company, then aggregated at the national level to GDP, and so on.

All economic, monetary and fiscal policy decisions are born, in the first instance, from the accounting of one individual transacting with another individual.

Both microeconomics and macroeconomics can only be understood with accounting as a basis. One from the information of individuals and companies and the other from countries and groups.

# It is a relational language

Accounting relates movements, parts, and allows us to think horizontally about the materialisation or consequence of a movement; it is the double entry.

People who master and understand accounting well think in terms of flows, not static data; flows of relationships, which is how business, economics and reality work.

# Accounting craftsmanship is a powerful tool.

I have met very few professionals in economic management positions who really understand accounting. It is a very powerful tool because it allows you to shape your forecasts as reality changes. The problem is that when reality changes the template is no longer useful, or you have to make manual changes, and that’s where accounting knowledge comes in; making an accounting entry.

Accounting + business knowledge + creativity = total professional power.

One of the main risks to manage when a company grows is the currency risk or exchange rate risk. The size of the company is not important in this case, the problem is the same whether the company is large or small.

One of the CFO’s responsibilities is to hedge (cover) this change rate to minimize the risk of losses.

Here are the main platforms in Europe to do this easily:

/ Kantox – Spain/UK

/ Revolut – UK

/ CurrencyFair – UK

/ IbanFirst – Belgium

/ WorldFirst – UK

/ RationalFX – UK

/ Currencies Direct – UK

/ Moneycorp – UK

/ Wise – UK

Is there a relationship between digital banks and interest rate rises that we do not see in the same way in traditional banks? Do they pass these rises on better or worse to the end customer?

To do so, we are going to look at the conclusions of two papers that have just been published and whose conclusions are quite clear.

The first paper entitled «Destabilizing digital «Bank Walks» reaches the following main conclusions:

1. Digitalization lowers the value of the deposit franchise, which has important consequences for financial stability.

2. When the Fed funds rate increases, deposits flow out faster and the cost of deposits increases more in banks with a digital platform.

3. The authors suggest that their findings are important because they speak to the sensitivity of bank profitability to interest rate changes and overall financial stability.

The implications of the findings in the paper are significant for financial stability and regulatory policy. Here are some possible implications:

1/ Regulators and bank supervisors need to be aware that banks may appear solvent based on the value of their deposit franchise, but digitalization can lower this value and lead to serious.

2/ The study suggests that digital banks may need to increase deposit rates to stem deposit flight during increases in the Fed funds rate, which could have implications for their profitability.

3/ The findings highlight concerns about the instability of deposit funding associated with the expansion and contraction of the Fed’s balance sheet during QE and QT episodes, which could have implications for financial stability.

4/ The study suggests that additional regulatory measures may be needed to achieve financial stability in economies with shadow financial markets, where digital banking is more prevalent.

5/ The authors suggest that their findings have implications for both insured and non-insured deposits, which could have broader implications for consumer protection and financial stability.

The other paper, entitled «Monetary Policy Transmission Through Online Banks» concludes that digital banks pass on interest rate rises to deposits much faster than traditional banks.

The conclusion should be: deposits come out faster but also pass on more profitability to the depositor in digital banks than in traditional banks.

In other words, digital banks suffer more from interest rate hikes but are more focused on their customers because they remunerate them accordingly.

The proliferation of AI tools in the last months is amazing. Today i bring you a tool call «There’s an AI for that», that is constantly updating (almost in real time).

What is interesting is that the categorisation is by task that is replaced; what kind of task can be done instead with a specific AI tool from the list.

In these moments there are 4.527 AIs for 1.188 tasks.

One curious thing is that AI tools started in 2015, and it has had tools since that date, which is a long time considering how young this technology is.

Another curiosity is that the website’s newsletter is written by ChatGPT.

Finally, very interesting is the «Jobs» section where the impact on each job is calculated based on the number of tasks covered with AI tools and the number of these.

You have here another AI tool agreggator.

In this amazing website we can navigate in a very easy and intuitive way to search for best AI tools for creators and founders.

It maintains a «Product Hunt» web style and you can search for the type of AI tools in a Google-like web browser or by tags. The latter is very useful, and they are divided by:

/ Use cases

/ Kind of tech

/ Learning curve

/ Pricing

The best part is that is made with Softr. Very useful for the user, very quick and cheap for the web master, who will know the enquiries of people.

Saas accounting tools («automated») for the self-employed and (micro) SMEs as of today, don’t still work properly. The problem with them is that they are sold as «automated» tools, but the accounting is very manual; there is a repetitive part that can be automated, but there is another part that needs to be done by a person who knows how to do the accounting.

In other words, although they have improved a lot, they do not solve the client’s problem.

The niche market for these solutions is still there, for example in Spain 98% of companies are SMEs, 90% being micro-SMEs (huge niche). You can outsource accounting while your company is very small, but after a certain point you have to internalise it as a strategic tool to use your information for decision making every week.

There is a dilemma; either the solutions are very «digital» (e.g. Holded, Quaderno) but very limited at the accounting level, or they are very «for accountants» (e.g. Sage, A3) but are difficult to fit for laypersons. The company that manages to do both will win the market.

It is interesting to see how some digital banks have completely pivoted their business model to Saas accounting (e.g. Quonto). The value proposition of many of the new neo-banks is small (they don’t add anything compared to the more traditional bank), but they are very good at attracting purely digital customers by putting the customer at the centre with a data-driven philosophy and they are turning to Saas accounting because there is a huge space to make money from business (B2C to B2B).

It is an interesting niche because it interconnects everything, but to be useful it needs professional accountants.

One of the problems for these companies is that local advisors are very competitive in price and provide a good service. You may find that the Saas accounting subscription is equal to the cost of an accountant… and you have to account for it yourself.

In one of the projects I worked on as fractional CFO we had this problem. The company was doing the accounting in a Northern European country on a Saas accounting solution in a standard package. When the company started to grow we needed the accounting up to date (and well done, accurate), we wanted to pay more for a more customised service but it was not available; moreover, they charged us more but did not provide the service. In the end we decided to look for a local consultant from a village and it was the best decision we made; he works with cloud software as a consultant, is efficient, very good value and manages a heavy accounting software to be connected to a lot of APIs.

Perhaps this is where the evolution of saas accounting will go, collaborating with local specialists in a hybrid model. I think Xero or Quickbooks are doing this quite well but in IFRS (Europe) enviroment doesn’t work. Europe = opportunity.