11.06.2019 - Studies
Inaccurate economic forecasts have deeper reasons than just uncertainty about the future. However, the incentives for this are quite different for companies on the one hand and for macro experts on the other.
Economic forecasts are often inaccurate. Complexity and un-predictable shocks regularly lead to forecasting errors. How-ever, as we discuss in this study, there are also other reasons that can explain incorrect forecasts. Both macro-experts and corporates are subject to certain incentives that can lead to distorted forecasts. However, these incentives have very different consequences.
Hardly any professional market participant will expect economic forecasts to accurately reflect future developments. Both GDP forecasts and profit forecasts of individual companies are inherently uncertain and will always produce a more or less large forecast error. The uncertainty usually increases with the length of the forecast horizon and the apparent accuracy with the precision of the formulation.
In the economic context, the complexity of forecasts results primarily from the fact that economic developments are always determined by human behavior, which is highly dynamic. And even if economic developments reveal certain regularities, the intensity of the cause-and-effect relationship will continue to change - even to the point of reversal. This makes economics as a social science fundamentally different from the natural sciences and related fields of activity, such as meteorology, where ever-increasing computing power is used to arrive at ever more accurate forecasts. In the case of economic forecasts, the progress achieved is rather modest. This confirms Friedrich August von Hayek’s description of such predictions as "arrogance of knowledge". Hayek warned against transferring the procedures of the natural sciences into economic models, since they are not suitable for adequately mapping the complexity in the diversity of the variables involved.1
But is the inaccuracy of economic forecasts always solely due to the underlying complexity and the associated uncertainty? Or are the makers of such forecasts subject to the incentive not to communicate their actual expectations regarding future developments unfiltered in the first place? If one follows the assumption that the forecasters have the exclusive aim of achieving a high forecast quality, there should be no systematic pattern in the forecasting errors to be found ex post. Positive and negative deviations in the forecast errors would have to balance each other out on average in the long term. However, the results of this analysis suggest that forecasts are systematically distorted in the economic context.
Companies tend to systematically underestimate their future sales and earnings situation. This is not purely coincidental, nor is the pessimism due to the lack of predictability of the expected order situation or the overall economic sentiment. Rather, corporate management strives to avoid negative capital market reactions that threaten to occur if a forecast is missed. Experience shows that investors react very sensitively to a failure to make profit forecasts and that share prices can suffer massively. Pessimistic forecasts are the result.
The findings in the field of economic or budget forecasts, are quite different. Economic and budget forecasters are subject to a certain forecast bias as well, albeit with the opposite sign. This form of forecasting is not about meeting capital market expectations. Rather, optimistic forecasts are intended to support public sentiment about the economy. This in turn can help to assert economic or party-political interests.
Both the permanent underestimation of corporate profits and the overoptimism in macroeconomic forecasts can have far-reaching consequences. Economic entities align their actions with their expectations. These expectations are in turn influenced by the assessments of others. Since economic forecasts are likely to have a major influence here, knowledge of the misguided incentives of forecasters and the resulting systematic misguided forecasts can help to avoid misallocations.
In order to prove the quality of macroeconomic forecasts, we use the economic forecasts of the International Monetary Fund (IMF). Additionally, we analyze the economic and budget forecasts from the respective national governments of the four largest euro area members, Germany, France, Italy and Spain. These forecasts are issued within the framework of the medium-term stability programs.2 We calculate the forecast error for the one-, two-, three-, and four-year forecast period as the difference between the forecast and actual values of real GDP growth or the government budget as a percentage of GDP. Positive values therefore indicate over-optimistic forecasts, while negative signs are an expression of pessimism.
Table 1 summarizes the calculations for the average forecast error for France, Germany, Italy, Spain, USA, euro area and the world over the period 2012-2018, based on the IMF real GDP growth projections. Except for Germany, where IMF forecasts were below actual real GDP growth over almost all forecast periods, forecasting errors are positive, which, according to our underlying definition, shows optimistic forecasts. In addition, forecast errors increase with the increasing length of the forecast horizon. The most optimistic distortions on average occur in the forecasts for Italy, France, the USA and the global economy.
Table 2 shows that government forecasts are also clearly overoptimistic. This applies both to real GDP growth and to the government budget.3 This also tends to increase with an increasing forecast window. On average for the period 2000-2018, the forecast error for the budget balance, with the exception of Germany, is significantly higher than that for the growth forecasts.
The fact that the forecast error increases in absolute values with the length of the forecast period is not surprising given the growing uncertainty. However, the one-sided direction of the error indicates that the incentive to "wishful thinking" increases with an increasing forecasting horizon. Finally, the verifiability of long-term forecasts is only given in the distant future, in which the respective government forecasters may no longer be responsible for the forecasts. From the status-quo perspective, it seems tempting to use the increasing uncertainty with the length of the time horizon to increase the favor of the electorate.
Compared with the IMF forecasts, the government forecasts of GDP growth for the four eurozone members with the exception of France show a higher level of optimism bias, albeit predominantly within the four-year forecast window. This over-optimism relative to the IMF forecasts almost balances out in shorter forecast windows. However, on average for the period 2012-2018, the forecasting error in government forecasts is significantly higher compared to the IMF forecasts for Italy and Spain (Figure 1).
To check the quality of management forecasts, it is reasonable to compare the sales and profits generated by the companies with the management forecasts as previously communicated. In Germany, listed companies are obliged to publish the expected development of key performance indicators for the current financial year as part of their quarterly and annual reporting. Assuming that management would always attempt to communicate these forecasts to the capital market in an unfiltered manner and that the only unknown factor in such forecasts would be macroeconomic development, in the long run the resulting forecasting error on average across all companies would have to roughly correspond to the level of the macroeconomic forecasting errors described above.
We use the annual and quarterly reports of German companies to test this hypothesis. Looking at the forecast errors, which are defined as the percentage deviation of the forecast from the actual result achieved, the picture differs from that of macroeconomic forecasts (Figure 2). More precisely, the forecast error always has a negative sign here. This shows that the companies’ management underestimate the actual development in each of the quarters under review in the period 2012 to 2018. The mean deviation is -1.9 %.
The evidence that the forecast error is probably not purely coincidentalis shown in Figure 3, which reports the average forecast errors for the individual quarters. The relative forecast error - in which positive and negative forecast errors offset each other - decreases quite continuously from quarter to quarter (Figure 3, left). While the relative forecast error at the beginning of the financial year was still -2.6%, it falls to -1.3% by the end of the third quarter. Figure 3 on the right also shows that management is in a better position to assess developments with a shorter forecast horizon. This is shown by the forecast error in absolute terms, i.e. the difference between forecast and realization, whereby the sign of the firm-specific forecast error is being neglected. The forecasting error is almost halved in this case.
In the management forecasts, the motivation goes back to the behavioral economics, according to which market participants punish negative surprises more than they reward positive ones even if of the same amount (loss aversion). Since management is aware of this behavioral bias, it is tempted to forecast sales and earnings expectations conservatively and not to communicate the true expectations to the capital market unfiltered. In that way, management uses a safety buffer that usually enables it to surprise positively or at least not to miss expectations.
There is no such incentive on the macro forecast side. The reasons for the over-optimism in IMF forecasts remain largely unexplored. Since the IMF is an independent body, it should not be subject to any apparent conflict of interest. Political-economic considerations suggest, however, that optimism may be an expression of the intention not to unnecessarily cloud the sentiment of economic actors by stirring up fears of recession which – on the wave of an emerging narrative – actually lead to a downturn. Rather, the aim is to spread optimism among economic actors. To what extent less confident IMF forecasts would actually be detrimental to growth cannot, however, be answered seriously.
The excessive optimism in government forecasts of the budget balance is due to a number of interdependent factors. On the one hand, our results show that estimates of real GDP growth (and inflation) are also regularly too optimistic on the government side.4 On the basis of these assumptions, it is not surprising that government budgets are also overestimated. However, this kind of regularity can only be seen in Italy and Spain, but not in France and Germany (Table 2).
Due to their short decision-making horizons, politicians are likely to have incentives to create a positive public perception of their past and future activities. Since their forecasts are not verifiable and the forecasters may no longer be responsible at the time when a review can be made, it is a dominant strategy in their view to overestimate revenues and underestimate expenditures. The same applies to the GDP growth forecasts of the respective governments, which tend to be even more overoptimistic than those of the IMF.5
Germany is a clear exception here. Since 2012, government forecasts have systematically underestimated both the budget balance and GDP growth (Figure 4). One explanation for the more cautious budget forecasts could be the debt brake in force since the 2011 financial year. Accordingly, the federal and state governments are to pursue the goal of a structurally balanced budget.6 Moreover, the underestimation of GDP growth since 2013 suggests equally underestimated tax revenues. On the expenditure side, the continued low level of interest rates – contrary to earlier expectations – should have contributed to falling interest payments. While the continuation of the favorable refinancing conditions appears realistic, the real economic risks could compensate for the recent over-pessimistic assessments in the opposite direction.
Interestingly, Germany's recent experience offers a counter-evidence to earlier empirical evidence that countries with a fiscal rule - such as the Stability and Growth Pact in the euro area - are even more over-optimistic about growth and budget projections than countries without fiscal rules.7
The systematic forecast error in macro forecasts has two important consequences. Since the IMF forecasts serve as a basis for planning, especially for private actors, overestimations of economic growth can lead to misallocations. An overestimation of future growth could lead to a positive short-term growth impulse, but it increases the probability of negative surprises or even recessions. Systematic accumulation of misallocations can lead to structural problems in the medium to long term.8
The over-optimistic forecasts of national governments have indirect macroeconomic consequences, as they lead to excessive government deficits and the accumulation of government debt. If national governments lack currency sovereignty, as is the case in the euro area, the resulting deterioration in public finances translates into a deterioration in creditworthiness and hence to worsening of financing conditions in the private sector. This in turn increases the pressure on the monetary policy makers by prompting them to intervene in the market beyond their usual competencies.
In the context of company forecasts, there is a danger that the credibility of prospective financial reporting will suffer on a sustained basis. For example, the permanent underestimation of the future earnings situation has already led to a certain familiarization effect among shareholders. The asymmetric reactions of the capital market to an over- or underestimation of the communicated management forecasts are evidence that the receivers of the forecasts have already adjusted their expectations upwards in the course of a learning process. Given the persistence of forecast pessimism, it will not be an easy task for companies to dampen expectations to a realistic level in a credible and at the same time price-friendly manner.
Our analysis has shown that forecasts at the micro- and macroeconomic level are systematically distorted. While companies seem to tend to communicate their expectations to capital market participants with a certain degree of restraint, forecasters in the macroeconomic context are often too optimistic. The knowledge about the incentives to which the respective actors are subject when making their forecasts can help to adjust expectations to an appropriate degree on the part of the forecast recipients. In this way, both companies receiving financial capital and investors providing it can make better investment decisions.
1 Von Hayek, Friedrich A. (1974): Lecture to the memory of Alfred Nobel, December 11, 1974, under: www.nobelprize.org/prizes/economic-sciences/1974/hayek/lecture/.
2 The IMF forecasts represent one of the most comprehensive forecast databases in the world and cover a total of 189 IMF member countries. It enjoys widespread support and application among many official policy makers. However, the data are also used in the private sector. The medium-term forecasts of national governments, on the other hand, are mandatory for all EU members and are part of the Maastricht criteria and the Stability and Growth Pact.
3 The findings are consistent with the findings of Jonung, L. and Larch, M. (2006), "Improving fiscal policy in the EU: The case for independent forecasts", Economic Policy, 21(47): 491-534.
4 Increased inflation estimates lead to the transfer of taxpayers to higher tax rates.
5 There is empirical evidence for this so-called partisan bias. See Frendreis, J. and Tatalovich, R., (2000), „Accuracy and bias in macroeconomic forecasting by the administration, the CBO, and the Federal Reserve Board”, Polity, 32(4): 623-32.
6 Middes, K. and Wolf, G.B. ["Germany's even larger than expected fiscal surpluses: Is there a link with the constitutional debt brake?", Bruegel, May 13, 2019] come to this conclusion and take it as an opportunity to plead for an increase in expenditure.
7 See Frankel (2011), „Over-optimism in forecasts by official budget agencies and its implications”, Oxford Review of Economic Policy, 27(4): 536-562.
8 This is the conclusion of the latest study by Beaudry, P. and Willems, T. (2018), "On the macroeconomic consequences of over-optimism", IMF Working Paper No 18/122.
21.03.2019 - Companies
by Kai Lehmann
The information contained and opinions expressed in this document reflect the views of the author at the time of publication and are subject to change without prior notice. Forward-looking statements reflect the judgement and future expectations of the author. The opinions and expectations found in this document may differ from estimations found in other documents of Flossbach von Storch AG. The above information is provided for informational purposes only and without any obligation, whether contractual or otherwise. This document does not constitute an offer to sell, purchase or subscribe to securities or other assets. The information and estimates contained herein do not constitute investment advice or any other form of recommendation. All information has been compiled with care. However, no guarantee is given as to the accuracy and completeness of information and no liability is accepted. Past performance is not a reliable indicator of future performance. All authorial rights and other rights, titles and claims (including copyrights, brands, patents, intellectual property rights and other rights) to, for and from all the information in this publication are subject, without restriction, to the applicable provisions and property rights of the registered owners. You do not acquire any rights to the contents. Copyright for contents created and published by Flossbach von Storch AG remains solely with Flossbach von Storch AG. Such content may not be reproduced or used in full or in part without the written approval of Flossbach von Storch AG.
Reprinting or making the content publicly available – in particular by including it in third-party websites – together with reproduction on data storage devices of any kind requires the prior written consent of Flossbach von Storch AG.
© 2020 Flossbach von Storch. All rights reserved.