The analysis is founded on the lender of Italy’s Survey of Inflation and Growth Expectations, which runs on the sample around 1,000 industrial and service firms with at least 50 employees. The Survey has been conducted quarterly since 1999 and specifically collects firms’ point inflation expectations at several time horizons (Bank of Italy 2018).
Evidence on the inflation expectations of firms
In Italy, the inflation expectations formulated by firms have already been broadly consistent with those reported by professional forecasters (Figure 1). Differences are unlikely to reflect too little relevant information on the firms’ side because, because the start of the survey in 1999, respondents were given the newest inflation reading in Italy and in the euro area before these were surveyed. The difference around a half percentage point in the first portion of the period probably reflects different levels of learning about the brand new monetary framework.
Like professional forecasters and households, Italian firms disagree about future inflation (Mankiw et al. 2003, Dovern et al. 2012, Cavallo et al. 2015, Nishiguchi et al. 2015). This disagreement is pervasive even among virtually identical firms, suggesting it really is linked to very specific firm features, experiences or information sources (Coibion and Gorodnichenko 2015, Malmendier and Nagel 2016, Binder 2018). That is indirectly confirmed by the actual fact that disagreement is higher when price dynamics are more differentiated across HICP items so when consumer inflation is more volatile in the short run. Also, disagreement is larger when current inflation is further from the ECB price stability goal.
Figure 1 Inflation expectations of firms and consensus forecasts are broadly similar
Source: Dovern et al. (2012).
Note: Median 12-month ahead inflation expectations of informed firms and 12-month ahead Consensus forecasts.
The lingering risks of de-anchoring
When expectations are more anchored, agents put less weight on incoming data in order that their expectations are less sensitive to current developments (Bernanke 2007). A particular feature of the lender of Italy’s survey can help you quantify the weight that price-setters put on incoming data and, specifically, on the newest inflation reading and how it evolves as time passes.
Specifically, because the start of the survey respondents have already been made alert to current inflation before reporting their expectations. However, because the third quarter of 2012, a random subset around one-third of respondents had not been given this information. Both groups are named ‘Informed’ and ‘Uninformed’.
Figure 2 shows the median inflation expectations at 12 and two years of both groups. Although having less focus on current developments will not appear to be extreme, as suggested by the actual fact that in both groups inflation expectations move with current inflation, firms made alert to current inflation generally revise their expectations toward that value.
Figure 2 Firms revise their expectations with information regarding current developments
Note: Median 12- and 24- month ahead inflation expectations of informed and uninformed firms and yearly HICP inflation rate communicated to firms in the quarter.
In a recently available paper, Laura Bartiloro, Marco Bottone and I exploit the random nature of the provision of information and interpret the expectations reported by uninformed firms as the last expectations of a hypothetical population, and the ones reported by informed firms as the expectations of the same population updated with incoming information (Bartiloro, Bottone and Rosolia 2017).
We’re able to show that, under a typical representation of the firms’ learning process, the ratio of the cross-sectional dispersion of the expectations of informed and uninformedfirms could be taken as a way of measuring the effectiveness of prior expectations against incoming data. A joint reading of the index and the amount of prior expectations offers a rough assessment of the amount of anchoring of firms’ inflation expectations.
Figure 3 displays, for the 12- and 24-month horizons, the weight of the last expectations (left axis) and the expectations of uninformed firms (the last itself) and current inflation (the incoming information) (both right axis). As consumer prices quickly slowed up following the sovereign debt crisis, the weight assigned by firms with their prior expectations kept falling, implying an evergrowing role for incoming data to greatly help them form expectations.
Figure 3 The relative weight of prior expectations and incoming information varies as time passes
Note: The weight on the last may be the ratio of the typical deviation of the expectations of informed firms compared to that of uninformed ones. Standard deviations are computed excluding observations in the most notable and bottom 1% of the cross-sectional distribution of 12-month-ahead inflation expectations.
The downward revisions appear to have stopped when current inflation stabilised at exceptionally low levels, of which point the weight directed at new, lower prior inflation expectations rose again. As a result, the pick-up in inflation recorded since 2017 hasn’t yet been incorporated into firms’ expectations, still hovering around 1% at both 12 and two years, and with a higher weight in the updating process.
It really is interesting to notice that, especially at the 12-month horizon, when current inflation occasionally fell below zero, the weight placed on the (still positive) prior sharply increases. Similarly, when in the next quarter of 2017 inflation hit 2%, the weight on the low prior again increases considerably. Indeed, that inflation figure ended up being a temporary occurrence. In both cases, confronted with uncommon inflation readings, firms were cautious enough not revise their expectations as suggested by their recent updating strategy. Instead they chosen a temporary revision of the strategy itself.
The response to monetary policy shocks
Firms are therefore been mindful of the economic environment, even if they’re apparently not completely alert to the latest developments. In addition they seem to put the info provided in to the proper context, as suggested by having less substantial updates in light of occasionally negative or unusually high inflation readings.
The unawareness of current HICP developments may partly be because they’re not relevant for all firms. Exporters may pay more focus on price developments in foreign markets, producers of intermediates may focus more on producer prices. To more thoroughly measure the degree of focus on macroeconomic developments and the capability to process incoming information, in a far more recent paper Marco Bottone and I directly ask if the inflation expectations of firms react to unanticipated monetary policy shocks which are undoubtedly highly relevant to all economic agents (Bottone and Rosolia 2018).
A confidential version of the info, such as the dates which firms filled in the questionnaire, can help you engineer a straightforward empirical exercise predicated on the comparison of the inflation expectations reported by firms interviewed in the times just before confirmed meeting of the ECB Governing Council to those of firms surveyed in the times just following it. These differences, one for every wave of the survey since 2002, are then linked to standard market-based measures of unanticipated monetary policy news predicated on the daily change of major market interest levels on the times the ECB issue its decisions (Kuttner 2001, Gurkaynak et al. 2005).
Figure 4 plots the difference in average inflation expectations at 12 and two years reported by firms interviewed before and following the meetings (vertical axis) against the daily change on Governing Council days of the three-months Overnight Index Swap (horizontal axis). The negative relationship shown in the figure is confirmed by empirical exercises where unanticipated changes along the word structure were jointly considered. It seems to hold even though the effective lower bound on interest levels is binding.
Figure 4 Inflation expectations react to unanticipated monetary policy news
Note: 1999 Q4 to 2008 Q1, only designed for 12-month-ahead inflation expectations, in red.
It really is obviously highly unlikely that firms themselves closely follow financial market developments on Governing Council days. Much more likely, the importance of the many ECB communications is conveyed through the media. The results therefore claim that firms know which events and sources to focus on, and have the capability to extract relevant information to update their expectations.
Firms understand current developments
Inflation expectations of Italian firms are no more deteriorating. Yet, persistent upward revisions require further sustained progress in moving towards the purchase price stability goal. Reassuringly, the data presented here shows that firms quickly understand developments and so are rarely misled by temporary shocks. Importantly, they understand and directly react to monetary policy decisions by revising their inflation expectations consistently with monetary policy impulses, even sometimes when the area for standard monetary policy tools is bound.
Authors’ note: The views expressed in this column are those of the writer and don’t necessarily reflect those of the lender of Italy.
Bartiloro, L, M Bottone, and A Rosolia (2017), “What does the heterogeneity of the inflation expectations of Italian firms reveal?”, Bank of Italy occasional papers 414.
Bernanke, B (2007), “Inflation Expectations and Inflation Forecasting”, speech at the Monetary Economics Workshop of the NBER Summer Institute, Cambridge, Massachusetts.
Binder, C (2018), “Inflation Expectations and the purchase price at the Pump”, Journal of Macroeconomics, forthcoming.
Bottone, M, A Rosolia (2018), “Monetary policy, firms’ inflation expectations and prices: evidence from daily firm-level data”, mimeo, Bank of Italy.
Carroll, C D (2003), “Macroeconomic Expectations of Households and Professional Forecasters”, Quarterly Journal of Economics 118(1): 269-298.
Christelis, D, D Georgarakos, T Jappelli, and M. van Rooij (2016), “Rely upon the Central Bank and Inflation Expectations”, CSEF working papers 458.
Coibion, O, Y Gorodnichenko (2015), “May be the Phillips curve alive and well in the end? Inflation expectations and the missing disinflation”, American Economic Journal: Macroeconomics 7(1): 197-232.
Dovern, J, U Fritsche, and J Slacalek (2012), “Disagreement Among Forecasters in G7 Countries”, Overview of Economics and Statistics 94(4): 1081-1096.
Draghi, M (2018), “Monetary policy in the euro area”, speech at the ECB Forum on Central Banking, Sintra, 19 June.
Gürkaynak, R S, B Sack, and E T Swanson (2005), “The Sensitivity of Long-Term INTEREST LEVELS to Economic News: Evidence and Implications for Macroeconomic Models”, American Economic Review 95(1): 435-436.
Kuttner, K N (2001), “Monetary Policy Surprises and INTEREST LEVELS: Evidence from the Fed Funds Futures Market”, Journal of Monetary Economics 47(3): 523-544.
Malmendier, U, S Nagel (2016), “Learning from inflation experiences”, Quarterly Journal of Economics 131(1): 53-87.
Mankiw, G N, R Reis, and J Wolfers (2003), “Disagreement about Inflation Expectations”, NBER working papers 9796.