---
category: "english"
media: "text"
outlet: "Czech National Bank"
url: "https://www.cnb.cz/export/sites/cnb/cs/menova-politika/.galleries/br_zapisy_z_jednani/2017/download/stanoviska_2017_06.pdf"
date: "2017-09-25"
headline: "Advisor's Opinion on Situation Report No. 6, 2017"
released: "2024-01-31"
byline: "Tomas Havranek"
genre: "advisor_opinion"
translated: "2026-08-27"
translation: "cnb-stanovisko-poradce-6sz-2017"
body_note: "An English translation of the Czech original, which remains the record of what was written: [Stanovisko poradce k 6. situační zprávě 2017](/komentare/cnb-stanovisko-poradce-6sz-2017/). Part I, the advisor's opinion, is here in full. The original document also carries Part II, prepared by the Financial Stability Department, which is in the [hosted PDF](/komentare/files/cnb-stanoviska-6sz-2017.pdf) and in the document on the CNB site. Numbered footnotes are collected at the end."
---

# Advisor's Opinion on Situation Report No. 6, 2017

OPINION ON SITUATION REPORT NO. 6[1]

## 1. Monetary policy recommendation

The August increase in rates supported the case that the exit from the exchange rate commitment was timed correctly. May the subsequent behavior of the key variables support the timing of the first rate increase in the same way. The koruna is stable, the inflation forecast is materializing, and the boom in the Czech economy is gathering strength. In such a comfortable situation it is possible to start preparing for the moment when this unprecedented alignment of favorable domestic and external conditions comes to an end, which is the main substance of this opinion. If we want to go on using interest rates as the main tool for stabilizing the economy, we will need a sufficient cushion of higher rates before the next recession. The average easing of monetary policy needed in response to a recession is 5.5 percentage points (I use here the US data presented in Reifschneider, 2016, since neither the Czech Republic nor the euro area has been through many standard rate-cutting cycles). So if we raised rates on average at every other monetary policy meeting, we would not reach that value until 2023.

The expectation that the current boom will end before 2023 is for me an additional argument for a steeper path of rate increases than the one presented in the Situation Report. Some cooling now will hardly hurt our economy (transmission from interbank rates to client rates also apparently weakened after the crisis; Havranek et al., 2016), and in harder times we will then be glad of more room for easing monetary conditions. On top of that, I believe the Czech economy is more overheated than the official GDP and inflation statistics suggest, since their construction does not fully reflect all the features of a modern economy.

Year-on-year GDP growth, despite its rising momentum, still does not match the euphoria observed among most firms and consumers. One reason may be that in today's digitalized economy GDP is getting worse and worse at capturing the value of products (see, e.g., Mokyr, 2014; Hulten and Nakamura, 2017).[2] It is equally possible that published inflation reflects the true purchasing power of money less and less[3] as household spending shifts into items that are not included in the consumption basket or carry too small a weight. Consumption is growing robustly, and yet we are surprised by its weaker momentum compared with the real growth of the wage bill. The classic “suspect” is property prices, whose weight in the official consumption basket, despite a recent increase, is still only 1.4% (while tobacco, for example, has a 5% weight). And it can be argued that the purchase of a home, especially a first home, is from the viewpoint of many households consumption rather than investment (Hampl and Havranek, 2017a).

An obvious option, then, is to raise rates by more than the mere 0.25 percentage point implied by the sensitivity scenario of the forecast (which is materializing better than the baseline scenario), but that would probably not be optimal in terms of communication, of the impact on the real economy (a sharper braking than necessary), or of financial stability (problems for existing borrowers if interest costs rose more steeply). I therefore consider it better to continue the normalization surely but slowly. A rate increase will also express confidence in the strength of the Czech economy after the end of the exchange rate commitment. I thus recommend raising the repo rate to 0.5%, the Lombard rate to 0.75%, and leaving the discount rate at 0.05%.

## 2. Assessment of the message of the Report and the forecast

The new Situation Report is, as usual, well put together. In retrospect I appreciate the construction of the sensitivity scenario for the previous Report, as well as the honest discussion there of the asymmetry contained in the baseline scenario of the current forecast: the effects of the ECB's quantitative easing are included, but the long-standing overboughtness of the koruna market is not, which leaves our exchange rate forecast biased toward stronger values. The result is a more gradual path of rate increases in the Situation Report than we consider likely. The problem with this asymmetry, which artificially deepens our dependence on ECB policy (since the koruna will appreciate less than the model assumes, we will be able to afford faster rate increases even during the ECB's QE), is one the professional community is aware of too, as was evident from the quarterly meeting with analysts.

Let us try to build the effects of the longer-lasting overboughtness experimentally into the baseline scenario of the November forecast. After all, our current inclusion of the effects of the ECB's quantitative easing is experimental too. The shadow rates used for this purpose can be estimated anywhere in the interval (−0.5; −8), as has been discussed several times at monetary policy meetings. Because the correct calculation of shadow rates is ambiguous, the post-crisis strength of interest rate parity is unclear (Borio et al., 2016), and shadow rates cannot even be arbitraged, this approach cannot give us more than a qualitative story. Let us attempt a similarly qualitative but explicit story in the forecast for the overboughtness as well, which will help us at least in external communication (we can hardly publish an exchange rate path until we believe it ourselves). Perhaps the Financial Markets Department could help here by taking part in the exchange rate forecast; another option may be to use bond yields instead of shadow rates, or to damp the shadow rates even more by expert judgment. A straightforward but only partial solution is to promote the sensitivity scenario to the baseline of the forecast, since we need to model the overboughtness over the entire forecast horizon (not just one or two quarters ahead).

## 3. My own topic: how to ease monetary conditions in the next recession

I do not, however, regard when and how to continue raising rates as the key monetary policy question. Even if we raised the repo rate by 0.5 percentage point today, or postponed the increase to the next meeting, the effects on inflation would be roughly similar and would show up only after a long time (the average of published estimates of transmission in the Czech Republic suggests that a change in rates has its greatest effect on prices after 15 months; Havranek and Rusnak, 2013). An unexpected rise in inflation can always be tamed quickly by selling off foreign exchange reserves. It will be harder to fulfill our primary mandate if we do not manage to raise rates to a sufficient level before the next recession (whether we have in mind the 5.5% mentioned above, based on the Fed's experience, or, say, “just” 4%). I see a considerable risk that even with the fast rate normalization recommended above, we will hit zero again next time. Three observations lead me to this: 1) the equilibrium real interest rate may lie below the 1% we assume, 2) even assuming a 1% equilibrium, the zero bound will probably constrain us 30 to 40% of the time in the future (Kiley and Roberts, 2017), 3) in two years' time the current boom in the US will be the longest in history, or at least since measurement began in 1854.

According to Goldman Sachs calculations, the statistical probability of a US recession within the next two years alone is 25%. It is possible that the future cycle will look completely different from what past statistics suggest, and that no crisis will come even within 5 years. Maybe it really is true that this time, after the Great Recession and with better regulation, things are different, that a golden decade awaits Europe, and that we will never need unconventional monetary policy tools again. But we cannot count on that, and it is appropriate to use the current, unusually calm situation to prepare a strategy for easing monetary policy in a crisis. Given that a crisis comes suddenly, and that preparing a strategy, together with implementing it, can take several years, it is better to have a plan ready “in the drawer”, including the specific conditions under which we would be willing to resort to unconventional tools, so that the countercyclical effect of monetary policy is as strong as possible.

In this section I build on the materials on unconventional monetary policy that the Monetary Department submitted to the Bank Board between 2009 and 2015. I focus on new findings that could not have been reflected in those materials. For the purposes of a strategy it would in any case be necessary to prepare a detailed feasibility study for each new tool, for which there is no room in an advisor's opinion. The tried and tested first choice is the exchange rate commitment. We know its advantages well (see, e.g., Bruha and Tonner, 2017), so in this material I will play devil's advocate and focus on its disadvantages, and conversely on the advantages of the other tools. Our first stay at the zero lower bound turned out well, but next time domestic and foreign experts will judge us much more strictly, precisely because the new tools now exist. One can think of at least five disadvantages of the exchange rate commitment:

1) It was often presented as the “nuclear button”, which we used at a time when our economy was threatened with a slide into a deflationary spiral. That need not be the situation in which we next arrive at the zero lower bound.

2) It does not allow fine-tuning of the economy. In theory the level of the exchange rate can be changed more often, but that erodes one of the main virtues of this policy, namely the stability and certainty it brought to the economy: a commitment that keeps being changed stops being a commitment. Besides, the effects of changes in the level of the exchange rate are somewhat harder to model than the effects of changes in interest rates (as shown by the theoretical prediction of exchange rate pass-through to prices, which was not fully borne out).

3) A side effect of the exchange rate commitment is negative rates in many segments of the market, often quite deep, over which monetary policy has no direct control.

4) A repeated weakening of the exchange rate (especially in response to a negative shock affecting all of Europe symmetrically) would be viewed bitterly by our trading partners.

5) A weakening of the koruna has negative psychological connotations. It is such an unpopular tool that its repeated use could indirectly threaten the independence of the CNB.[4]

If we find ourselves in a situation where a recession looms, interest rates are at zero, and the forecast implies a further need to ease monetary conditions, we have, besides the exchange rate commitment, three more technically feasible options. The first is well known and often used; the other two exploit recent innovations in FinTech and in economic theory (I ignore entirely classic quantitative easing a la the Fed and the ECB here, since in our conditions of surplus liquidity and a relatively shallow market I do not, in line with the Monetary Department's earlier materials, consider it useful; besides, buying bonds with policy rates at the zero bound would lead to negative rates on those bonds anyway).

## 3.1. Sitting it out with forward guidance

At the time of the decision to introduce the exchange rate commitment, the main dissenting current in the Bank Board, in the CNB's expert staff, and among the professional community pushed the view that the period of low inflation should be “sat out”: that is, that monetary conditions should not be eased any further, except perhaps for a commitment to keep rates at a low level (forward guidance). I consider such an approach optimal when the implied undershooting of the inflation target is short-lived and within the tolerance band. The faster we raise rates during the boom, the more likely it is that sitting it out will be enough for us in a crisis. Unfortunately it seems unlikely to me that we will get by with this approach, because by then we will probably still be well short of the needed cushion in policy rates, and the undershooting of the target will therefore be significant. Every such sitting out of a recession also means losses for society, and especially for low-income groups (inequality rises in a recession, see, e.g., Hampl and Havranek, 2017b). Forward guidance, in turn, is extremely effective in standard macroeconomic models, but not quite so effective in practice, or in some extensions of those models (McKay et al., 2016).

## 3.2. Direct support of consumption

The materials on unconventional tools for the Bank Board mentioned Friedman's concept of a “helicopter drop of money”, that is, sending money directly to consumers. The main argument against it is that it is a quasi-fiscal measure: to introduce such a tool the central bank would have to cooperate with the government, for example by financing tax relief, and would thereby endanger its independence. In recent years, however, a number of central banks have started to consider introducing a digital currency that would complement cash and give households direct access to the central bank's balance sheet (a good summary is provided by Bordo and Levin, 2017). If the central bank introduced this innovation (not at all necessarily blockchain-based), it could, when needed, stimulate household consumption directly in this way without any need for government cooperation.

A digital currency issued by the central bank is itself still, for a number of reasons, a fairly controversial idea (Hampl, 2017). But a digital currency is perhaps not even necessary for this purpose when the central bank administers the central register of accounts. Technically it is possible to send a certain amount each month to one of the accounts of every person listed in the register. If this intention is communicated sufficiently far in advance, all citizens will have time to open at least one account at any bank of their choosing, which gets them into the register. This would also achieve a positive side effect in the form of greater financial inclusion.

In a standard macroeconomic model, such “direct support of consumption”[5] does not have a large effect, mainly because the model treats consumption as a function of permanent income (not current income), which these relatively small transfers would change only a little. The question is whether citizens would not in fact consume a substantial part of such a transfer “in Keynesian fashion”. Gechert and Rannenberg (2017) show that by the typical estimate in the literature we can expect roughly half of the transfer to be saved.

This problem can be dealt with to a large extent precisely by using a limited type of digital currency, or rather a digital wallet, from which one could pay for consumer goods (but not, for example, buy investment assets) and into which no additional funds could be transferred, so that deposits would not flow out of accounts at commercial banks. The funds sent to these wallets could have a limited period of validity, to give citizens more motivation to consume. True, this does not deal with substitution, whereby consumers could save a larger percentage of their other income, but even so it should be a satisfactorily functioning tool (in theory it will always be effective as long as the transfers are large enough). Its great advantage is that it affects consumption directly, removing the uncertainty about transmission between several variables.

The main disadvantage of direct support of consumption remains the impact on the independence of the CNB. Unlike the exchange rate commitment, this policy would certainly be popular, perhaps too popular; it could be perceived as support for the government, more strongly so than standard monetary easing. Direct support of consumption can be likened to a dividend paid to shareholders, since the central bank is a national institution in which every citizen of a democratic country in theory holds an equal share. While this comparison perhaps makes economic sense, and shows that, just as with the exchange rate commitment, no “spending” of public funds would be involved, it also hints, in its own way, at the danger to the future independence of the CNB (see the experience of the Swiss National Bank).

Direct support of consumption would have an immediate, one-hundred-percent impact on the CNB's financial result, since in accounting terms it would be a pure loss. An advantage over the exchange rate commitment is that the size of the losses would be known in advance; what is more, the counterparty would be domestic households, not financial institutions with mostly foreign owners. Accounting considerations are secondary, however, compared with monetary policy objectives; and the loss can be compensated over the long run by putting more emphasis on return in managing the foreign exchange reserves, above all by raising the share of equities in the investment tranche of the portfolio (Jiri Schwarz dealt with this topic in detail in the opinion on Situation Report No. 3, 2017). The endowments of the largest private universities in Western countries, which are of the same order of magnitude as the investment tranche of our reserves, report long-run returns of around 10%.

## 3.3. A bonus for cash withdrawals

The most standard nonstandard tool of monetary policy is negative rates, whose effects we can model well in theory. I covered the experience of the central banks that have implemented them in the opinion on Situation Report No. 2, 2016. Even now, that experience remains unimpressive, mainly because the potential of negative rates is limited: once they fall below a certain level, firms and consumers start withdrawing money from bank accounts and storing it in cash. In the opinion mentioned above I estimated this level of rates for the Czech Republic at −1.5%; Kolcunova (2017) more recently, and more rigorously, estimates −1%. The potential of having 4 “cuts” in hand does not outweigh the three drawbacks of negative rates: 1) problems for banks, which cannot fully pass negative interbank rates through to client rates on deposits, 2) problems for pension funds and insurers, which cannot earn a positive return on assets risk-free, and 3) unpopularity among savers, who mind negative nominal rates more than negative real rates, though it is the real ones that matter economically and that have been negative for a long time.

There is a fairly straightforward method (though one articulated in economic theory only recently) for removing the lower bound on negative rates and thereby turning them into a genuinely effective and useful tool. That method is a time-varying bonus for cash withdrawals (Agarwal and Kimball, 2015), set by the central bank so that the effective return on holding cash equals the monetary policy rate. So when a bank withdraws cash from the CNB in a period of negative rates, it receives somewhat more cash than is deducted from its account (and conversely, when it deposits cash, it is charged a fee). It can be ensured that with negative rates it is then equally profitable to hold money in an account and in cash, since the bonus for cash withdrawals gradually grows. There is thus no need to invalidate or stamp banknotes whose serial numbers are drawn in a lottery, as has occasionally been proposed, or to restrict the use of cash in other, often drastic ways (Rogoff, 2017). The solution is not to ban cash or reduce its volume, but on the contrary to expand it in a controlled way and leave the next act entirely to market mechanisms.

In this simple way a de facto exchange rate is created between cash and electronic money,[6] which gradually returns to parity once the monetary policy rate climbs out of negative territory (here I briefly summarize the arguments of Prof. Kimball of the University of Colorado, who, during last year's tour of the central banks of advanced countries, presented the technical nuances of this approach in detail at the CNB as well). It is then possible to respond to a crisis with deeply negative rates, which can rise back above zero all the faster in the recovery, in contrast to rates stuck at the zero bound for many years. It can be expected that, with such a policy in place, Gresham's law would assert itself at first: because cash will be worth less (at the ATM the consumer receives more cash than is deducted from the account), consumers will prefer to pay in cash. But it is likely that merchants would gradually restrict cash payments on their own initiative, or allow them only for small amounts. The final effect would thus, on the contrary, be to speed up the natural dying off of cash transactions, which would further increase the effectiveness of monetary policy.

But this still does not address the negatives of negative rates listed above: losses for banks, losses for pension funds, the unpopularity of the measure. Low interest rates will always be a problem for financial stability; what is specific to negative rates is the difficulty of passing them through to client rates on deposits, which threatens banks' high profitability and provokes their (and their analysts') strong and lasting resistance to this policy. In practice, though, we do observe negative client rates on large deposits. Bank stability can further be helped by giving each bank non-negative interest on reserves up to a volume corresponding, say, to the average wage multiplied by the number of clients. In other words, banks could in any situation offer consumers a current account with a non-negative interest rate, whose limit will suffice for standard use including direct debit payments.

We got a taste of negative rates (including the consequences for insurers and pension funds) across a large part of the financial market during the lifetime of the exchange rate commitment, and in some cases after its end as well. So it is likely that this effect will appear again in the next period when nonstandard tools are used here or in our neighborhood. The advantage of the bonus for cash withdrawals is control over these rates and a faster return to the equilibrium value after the crisis ends, so the overall impact on pension funds and savers may even be positive.

## 3.4. Implications of the new instruments for the inflation target

Because of the threat of a persistently binding lower bound on rates, a number of economists propose raising the inflation target toward 4%, that is, partly giving up on the objective of price stability on grounds of technical inability. The possibility of using direct support of consumption or the bonus for cash withdrawals implies, on the contrary, that a positive inflation target is redundant, since its main purpose is precisely to serve as a reserve against monetary policy being constrained at the zero bound. It is true that CPI growth overstates true inflation to some degree, since it does not take sufficient account of improvements in the quality of goods and of substitution, but the size of the overstatement is unclear (it certainly does not amount to 2 percentage points). On the other hand, the CPI may understate inflation, among other reasons because it excludes the faster-growing prices of claims on future consumption, above all property prices (Hampl and Havranek, 2017a). All told, then, we can work on the premise that the CPI measures the purchasing power of money roughly correctly; that is, in any case, how we all use it.

If the lower bound on rates does not constrain us, it is possible to target the purchasing power of money so that it stays, on a long-run average, entirely constant (which Bordo and Levin, 2017, call in this context “true price stability”).[7] Such an operating framework would be very pleasant for consumers and firms, in that it would turn the currency into a perfect store of value: a koruna today would have the same purchasing power as a koruna in 10 years, or indeed at any point until euro adoption. Removing this uncertainty would make it easier for households to save for retirement. Credible price-level targeting (of which targeting a constant purchasing power of money is a special case) has the further advantage that in a crisis it acts as a firm anchor that rules out a deflationary spiral; occasional mild deflation then does the economy no harm. A fall in prices triggers a substantially stronger monetary policy response than would occur under inflation targeting, where what matters for monetary policy is the future change in prices regardless of their past behavior. Switching to targeting a stable purchasing power of money would mean immediate costs for the economy, in particular some rise in unemployment (owing to nominal wage rigidity), but many economists judge that these costs would be only temporary. In any case such a change would have to come in the rising phase of the cycle, when unemployment is not such a large social problem.

The main disadvantage of targeting a rising price level in general is the difficulty of communicating it to the public (see the useful summary in Bohm et al., 2011), who would have trouble understanding why the central bank aims for 1% inflation one year and 3% the next. Targeting an unchanged purchasing power of money, by contrast, is even simpler to communicate than inflation targeting, because a constant price level is exactly what most people perceive as true price stability. The idea of such a policy change is bold by central banking standards, but probably less bold than introducing a de facto exchange rate between cash and electronic money in a crisis. So if we one day conclude that the bonus for cash withdrawals can be carried out, it would be useful to switch at the same time to targeting the present purchasing power of money. The first without the second is hard for the public to accept; the second without the first (with the exception of direct support of consumption) is technically infeasible. It would therefore be appropriate to join the two into a single package, which would at once expand the options of monetary policy in a crisis and bring our definition of price stability closer to its intuitive meaning.

While there is no need to write ourselves into the textbooks of central banking by being the first to introduce direct support of consumption or the bonus for cash withdrawals, on my reading of the literature both of these tools are usable already and would probably be effective enough in a crisis (especially in combination).[8] Assessing their legislative and implementation challenges in Czech conditions goes beyond the scope of an advisor's opinion on a Situation Report. That assessment, and possible implementation, would together take perhaps several years (including potential changes to legislation that might be needed to make direct support of consumption more comfortable to carry out), which is why I think it appropriate to start talking about a strategy for stimulating the economy in the next crisis before the first signs of a cyclical slowdown in growth appear.

## 4. Specific comments and questions on the Report

- The small Situation Report, too, reads better after its facelift. Good work!

- The observed data on GDP and wage growth came in very substantially above the forecast. It surprises me how little this fundamental information changes the outlook for inflation and rates in the Situation Report.

- Our GDP growth this year (more than 8% annualized since the start of the year) is by the available data the highest in the world, which looks somewhat suspicious; and this comes on top of a large upward revision of the comparison base in previous years. Are there one-off factors, other than EET, the ramp-up of the Kodiaq, and the number of working days in the second quarter, that would help explain this year's unprecedented acceleration?

- The Situation Report assumes a quarter-on-quarter fall in GDP for the next quarter. Would it be possible to try to sketch roughly what the implied rate path would look like if we assumed that half of the reported growth in the first half of the year was fundamental and will continue at an annualized 4% pace? That is, that we expect only a partial regression to normal. That would still imply GDP growth of over 6% for 2017 as a whole.

- While the reader understands why the model exchange rate at the forecast horizon is even stronger than in the previous Report, it does not make an entirely good story, partly because the observed exchange rate is relatively stable (and that even while pricing in a hike beyond the forecast). Many central banks do not attempt an exchange rate forecast at all and assume a fixed value. I offer for consideration whether, given the problems with forecasting the exchange rate, the Report could not always contain a sensitivity scenario in which the koruna appreciates only at the pace of equilibrium appreciation (1.5% a year). Such a scenario will of course not be model-consistent, but it would seem useful to me.

- Given the overboughtness of the koruna and the weaker arbitrage in interest rate parity, I believe the fate of the ECB's quantitative easing is considerably less relevant for the path of our interest rates than it would be under normal circumstances. The autonomy of our monetary policy is greater than the current forecast implies.

- Would it be possible to try to quantify the impact on the NAIRU of the August tightening of the conditions for the long-term unemployed? It is a fairly significant measure (a one-third cut in benefits for inactive claimants).

- Can appreciation of the koruna, other things equal, also affect the NAIRU? Given the large number of job vacancies in the Czech Republic and the high unemployment in the southern wing of the euro area, a sufficiently strong koruna, and thus a Czech paycheck converted into euros, could start to attract labor from the south of Europe as well as the east.

- I would like to see in the Report a chart of the euro rates that actually enter, and have entered, the interest rate parity condition in our model. On p. 10 there is a chart of shadow rates, but in the model these are combined with market rates using a certain weight, and this weight changes over time. Changing the weight, meanwhile, is the same as if we changed the calculation of the shadow rates. Given the importance of the exchange rate forecast for our model, I argue for further strengthening transparency in this area.

- The reader of the Report may ask how much value the forecast of the financial account adds (the right-hand chart on p. 22) when only a small share of its items is forecast.

## References

- Agarwal, R. and Kimball, M., 2015: “Breaking Through the Zero Lower Bound,” IMF Working Papers 15/224.

- Bohm, J., J. Filacek, I. Kubicova and R. Zamazalova (2011): “Price-Level Targeting – A Real Alternative to Inflation Targeting?” CNB Research and Policy Notes 1/2011.

- Bordo, M. and A. Levin (2017): “Central Bank Digital Currency and the Future of Monetary Policy.” NBER Working Paper No. 23711.

- Borio, C., R. McCauley, P. McGuire and V. Sushko (2016): “Covered Interest Parity Lost: Understanding the Cross-Currency Basis.” BIS Quarterly Review, September 2016, 45-64.

- Bruha, J. and J. Tonner (2017): “The Exchange Rate Floor as an Instrument of Monetary Policy: An Ex-Post Assessment of the Czech Experience.” CNB Working Papers 4/2017.

- Gechert, S. and A. Rannenberg (2017): “Which Fiscal Multipliers Are Regime-Dependent? A Meta-Regression Analysis.” Journal of Economic Surveys, in press.

- Hall, R. and R. Reis (2017): “Achieving Price Stability by Manipulating the Central Bank’s Payment on Reserves.” NBER Working Paper No. 22761.

- Hampl, M. (2017): “Central Banks, Digital Currencies and Monetary Policy in Times of Elastic Money.” Lecture for the Official Monetary and Financial Institutions Forum Roundtable, London, July 11, 2017.

- Hampl, M. and T. Havranek (2017a): “Should Inflation Measures Used by Central Banks Incorporate House Prices? The Czech National Bank’s Approach.” CNB Research and Policy Notes 1/2017.

- Hampl, M. and T. Havranek (2017b): “Nerovnost neroste,” čnBlog, May 15, 2017.

- Havranek, T., Z. Irsova and J. Lesanovska (2016): “Bank Efficiency and Interest Rate Pass-Through: Evidence from Czech Loan Products,” Economic Modelling 54(1), 153-169.

- Havranek, T. and M. Rusnak (2013): “Transmission Lags of Monetary Policy: A Meta-Analysis.” International Journal of Central Banking 9(4), 39-75.

- Hulten, C. and L. Nakamura (2017): “Accounting for Growth in the Age of the Internet: The Importance of Output-Saving Technical Change.” NBER Working Paper No. 23315.

- Kiley, M. and J. Roberts (2017): “Monetary Policy in a Low Interest Rate World.” Brooking Papers on Economic Activity, in press.

- Kolcunova, D. (2017): “Estimating the Effective Lower Bound for the Czech National Bank's Policy Rate.” Master's thesis, Institute of Economic Studies, Charles University, September 2017.

- McKay, A., E. Nakamura and J. Steinsson (2016): “The Power of Forward Guidance Revisited.” American Economic Review 106(10), 3133-3158.

- Mokyr, J. (2014): “Secular Stagnation? Not in Your Life.” VoxEU, August 11, 2014.

- Reifschneider, D. (2016): “Gauging the Ability of the FOMC to Respond to Future Recessions.” Finance and Economics Discussion Series 2016-068, Board of Governors of the Federal Reserve System.

- Rogoff, K. (2017). “Dealing with Monetary Paralysis at the Zero Bound.” Journal of Economic Perspectives 31(3), 47-66.

## Notes

[1] I thank Vit Barta, Jan Filacek, Michal Franta, Kamil Galuscak, Dana Hajkova, Simona Malovana, Petr Polak, Jiri Schwarz, Milan Simacek, Jan Vlcek, and Michal Vodrazka for discussion of the topics in this opinion. The responsibility for its message is mine alone.

[2] Many new products and services in the digital economy can be used at low or zero cost. They therefore contribute little to the statistics, although they may significantly affect households' true standard of living.

[3] In many industries prices are rising only very mildly, but delivery times are lengthening markedly. Suppliers of prefabricated houses, for example, are not raising prices, but a person who orders a house from them now will not have it built for 3 years. The statistical price of the house is the same, but its value to the consumer is lower, because it can only be put to use later.

[4] The point of an independent central bank is that it takes very unpopular steps when needed. But if it has several tools available that are potentially similarly effective, the popularity of a tool is also a criterion it should keep in mind, so as not to endanger its standing in society needlessly, and with it the future ability to carry out important unpopular measures. Besides, in a crisis it is important to lift household sentiment too.

[5] It would probably not be good communication tactics to call such a policy “helicopter money” (see Bernanke's widespread nickname “Helicopter Ben”, earned on the basis of a single speech from 2002). For this tool and the next one I therefore use terms that seem to me similar in meaning but psychologically more encouraging than the established names.

[6] This exchange rate already exists in practice, but cash is usually more valuable to merchants, so they prefer the customer to pay that way. The bonus for cash withdrawals would turn the situation around: instead of restrictions on card payments we would more often see restrictions on cash payments (e.g. “cash payments up to CZK 100 only”).

[7] Hall and Reis (2017) show that the central bank can achieve a stable price level simply by indexing rates to expected prices. A summary is here: http://econlog.econlib.org/archives/2017/03/ricardo_reis_on.html.

[8] Why, then, has no central bank yet abandoned inflation targeting in favor of targeting a stable purchasing power of money, combined with occasional use of direct support of consumption and the bonus for cash withdrawals? The philosophy behind both tools is old, but they were brought into practically workable form only in the last two years. Unless insurmountable obstacles appear, the first to introduce them will probably be the central bank of a small advanced country (much as with inflation targeting), since the large central banks face stronger inertial pressure from the public and from institutional tradition. That is one reason it took 22 years from New Zealand's pioneering example before inflation targeting was officially adopted in the United States.