What type of planner are you?
Basically, we differentiate between indirect and direct liquidity planning approaches. In the case of indirect liquidity planning, the company result is adjusted by values that have no cash effect (such as amortizations), in order to obtain the operative cash flow. Often, not enough details are known about this value, because the projected basic income statement and the projected balance sheet often do not exist or if they do, are not granular enough. If this is the case, the finance division regularly opts for a direct liquidity planning. However, this means that many entrepreneurs are faced with the challenge of how to efficiently determine the planning values in an adequate quality.
Should a planning be based on experience, simple extrapolations of previous values, integrated planning approaches or is a completely different approach required? Which role do technological developments such as Predictive Analytics play in this context?
With the vast number of possibilities available, one question is often forgotten – which approach best suits my company? The aim must not be to find the best approach, but rather the most suitable one. As is so often the case, the question is how much you should invest in a robust liquidity planning process. Choosing a planning approach depends on many different factors: the company’s financial situation, the Treasury department’s organizational structures and workflows, the system landscape, the business model, etc. In reality, companies often overreach with big projects and technically complicated and expensive tools - and then wind up returning to Excel sheets.
The first question to address is whether conventional liquidity planning approaches still meet the desired standards or not, and whether they still reflect the liquidity requirements of companies in adequate quality and accuracy. Treasurers are often satisfied with vague plans. The reasons for this are manifold. Besides an inconsistent database, the planning approaches themselves are often the cause of inaccurate planning, such as:
In view of these challenges and deficiencies in the conventional planning approaches on the one hand, and current technological developments on the other hand, treasury departments increasingly introduce new liquidity planning approaches or expand their existing ones. In the following, we will discuss how such approaches might look like in reality and which approaches might be suitable for whom.
Integrated liquidity planning is based on expected liquidity-relevant events, which themselves base on the planning stage of the controlling. That is, controllers provide estimates and assumptions made based on controlling processes, for instance on future growth. Special effects are adjusted manually, for instance, by means of database normalization. Methodically, integrated planning uses the direct approach, however, it is based on a results-based financial planning. The objective is to reduce deviations from the direct cash flow statement and, at the same time, to eliminate manual planning efforts. As a result, there is an alignment of business planning (controlling) and cash flow planning (treasury) structures. Ideally, processes may even be aligned in relation to frequency, planning horizon and time slices.
Integrated liquidity planning is based on two elements: on the business planning on the one hand, and on data from accounts payables (AP) and accounts receivables (AR) on the other hand. The number of days of sales outstanding and the number of days payable outstanding are derived from historical AP/AR data. They allow the empirical establishment of the approximate arrears between income-effective bookings and payments. Either one single version (most likely) or various different versions are established for the business and the treasury planning, As a result, the inflowing and outflowing transaction data is identified based on the business planning. Business planning positions are then aligned with liquidity planning positions (at the level of categories), in order to achieve the highest possible correlation between business and liquidity planning.
Driver-based liquidity planning is based on driver models for the (semi-) automatic projection of data in relation to past events. One of the things paramount here is the alignment of business planning and treasury forecast. These same drivers are used as change factors. Actual data, ideally in the form of categorized actual cash flows (such as inflows from sales, payroll payments, taxes, etc.) are used as a data basis. In order to make future projections, the drivers are then linked to the actual data. Doing so, the planning horizon can be extended.
First of all, the main cash-flow drivers (such as leasing contracts, delivery orders etc.) and events (such as holidays) need to be identified. Thus, trends and developments concerning the selected drivers can be derived from different data sources. For a targeted projection, these driver trends are applied to detailed, current cash flows in the course of gathering data for the planning. To represent special effects (such as significant investments), it may make sense to perform an additional manual plausibility check and adjust the planning data. To assure the quality of the planning approach, an ex-ante back testing may be used to calibrate the approach. The main requirement for this is the availability of historical cash flows over the relevant period.
As an equation, the driver-based projection of cash flows can be depicted as follows:
X = Y x A / B, whereby:
X = Cash forecast
Y = Cash flow history
A = Business planning
B = Historical controlling data (drivers)
The main difference between integrated liquidity planning and driver-based planning is that integrated liquidity planning is highly dependent on the frequency, the horizon and the temporal granularity of the business planning. In contrast, driver-based planning is more independent, because it is mainly based on historical cash flows.
Predictive cash forecasting is based on the forecasting of values determined with the help of an analytical projection of historical data using statistical regression methods. In contrast to other planning methods, which are based more on business intuition, predictive analytics methods allows a data-driven projection of internal and external influencing factors. In doing so, the model learns about the various factors and their effects on the cash flows in the past. The model calibrates its parameters by combining the parameters with the smallest deviation between forecast and actual data. The model is thus able to project future cash flows. Should the dynamics of the influencing factors change, the model needs to be trained accordingly. Complex interdependencies in the cash flow planning can thus be integrated into the data model and automatically generate planning data.
The specific procedures can be described as follows: The aim is the projection of a daily cash flow based on historical data using a framework specifically developed for this. The various models (such as ARIMA, Add. Regression etc.) are shaped and improved with the help of automatic parameter selection. Apart from drivers and events, the integration of the frequency of forecasts is also vital for risk assessment. The automatic selection of the optimal data granularity as well as the break-down of the time series into seasonality, trends and other units are further options provided by the predictive analytics approach. Vital for the model is the final step, the quantitative evaluation with the help of specifically defined error metrics (such as MAE, SMAPE).
As mentioned at the beginning, there is no single optimal approach to liquidity planning that is suitable for all companies. While there is no ‘best’ approach, it is nonetheless possible to identify the approach that is most suitable for a particular company – and this may include hybrids of the above-mentioned approaches.
The integrated planning method directly depends on the quality of the business planning, which is often influenced by business policies, while the liquidity planning method tries to make projections that are as realistic as possible. In exchange, the integrated planning approach is characterized by low implementation efforts because one can use already existing data sources. Operating expenses are also limited due to highly efficient processes and a strong potential for automation. In addition, because it is integrated in the business planning, very little coordination is required.
The driver-based planning is characterized by high-quality projections because of a prospective depiction of the business model as well as the use of categorized actual cash flows as bases for the planning. Merely the drivers may have to be adjusted depending on business developments. Implementation efforts are caused by the use of tools for the identification of categorized actual cash flows. In overall terms, the effort for the implementation depends on the complexity of the business model. As far as operating expenses are concerned, ongoing license costs and maintenance costs should be expected if tools are used for the identification of categorized actual cash flows. Should the general framework change, this would also cause changes in the drivers. Ultimately, driver-based planning causes higher coordination expenses than integrated planning, because there is no coordination with the business planning.
Predictive analytics methods are characterized by an automatic, empirical gathering of planning data. Like driver-based planning methods, they are characterized by high-quality projections because the planning is based on categorized actual cash flows. In turn, the tools required by this method cause increased implementation efforts. In addition, the configuration of parameters is complex, as many different influencing factors need to be taken into account. Apart from ongoing license costs and maintenance costs for tools used to identify categorized actual cash flows, the data model also has to be adjusted to any changes in the general framework. Ultimately, predictive analytics methods also require some coordination because, as in the case of driver-based planning, there is no coordination with the business planning.
This list of criteria should make it possible for every company to find the most suitable liquidity planning approach – depending on its size, business model, complexity, data availability, resources and business objectives. In addition, the most suitable solution in the context of liquidity planning can be anything from a ‘simple’ integrated planning method to the newest state-of-the-art predictive analytics method.
Source: KPMG Corporate Treasury News, Edition 98, January - February 2020
© 2020 KPMG AG Wirtschaftsprüfungsgesellschaft, ein Mitglied des KPMG-Netzwerks unabhängiger Mitgliedsfirmen, die KPMG International Cooperative (“KPMG International”), einer juristischen Person schweizerischen Rechts, angeschlossen sind. Alle Rechte vorbehalten.